summaryrefslogtreecommitdiff
path: root/fs/btrfs/dev-replace.c
blob: 4a0243cb9d9730401f967c46ede2ec3a62a58556 (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
// SPDX-License-Identifier: GPL-2.0
/*
 * Copyright (C) STRATO AG 2012.  All rights reserved.
 */

#include <linux/sched.h>
#include <linux/bio.h>
#include <linux/slab.h>
#include <linux/blkdev.h>
#include <linux/kthread.h>
#include <linux/math64.h>
#include "misc.h"
#include "ctree.h"
#include "extent_map.h"
#include "disk-io.h"
#include "transaction.h"
#include "print-tree.h"
#include "volumes.h"
#include "async-thread.h"
#include "check-integrity.h"
#include "rcu-string.h"
#include "dev-replace.h"
#include "sysfs.h"

/*
 * Device replace overview
 *
 * [Objective]
 * To copy all extents (both new and on-disk) from source device to target
 * device, while still keeping the filesystem read-write.
 *
 * [Method]
 * There are two main methods involved:
 *
 * - Write duplication
 *
 *   All new writes will be written to both target and source devices, so even
 *   if replace gets canceled, sources device still contans up-to-date data.
 *
 *   Location:		handle_ops_on_dev_replace() from __btrfs_map_block()
 *   Start:		btrfs_dev_replace_start()
 *   End:		btrfs_dev_replace_finishing()
 *   Content:		Latest data/metadata
 *
 * - Copy existing extents
 *
 *   This happens by re-using scrub facility, as scrub also iterates through
 *   existing extents from commit root.
 *
 *   Location:		scrub_write_block_to_dev_replace() from
 *   			scrub_block_complete()
 *   Content:		Data/meta from commit root.
 *
 * Due to the content difference, we need to avoid nocow write when dev-replace
 * is happening.  This is done by marking the block group read-only and waiting
 * for NOCOW writes.
 *
 * After replace is done, the finishing part is done by swapping the target and
 * source devices.
 *
 *   Location:		btrfs_dev_replace_update_device_in_mapping_tree() from
 *   			btrfs_dev_replace_finishing()
 */

static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
				       int scrub_ret);
static int btrfs_dev_replace_kthread(void *data);

int btrfs_init_dev_replace(struct btrfs_fs_info *fs_info)
{
	struct btrfs_key key;
	struct btrfs_root *dev_root = fs_info->dev_root;
	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
	struct extent_buffer *eb;
	int slot;
	int ret = 0;
	struct btrfs_path *path = NULL;
	int item_size;
	struct btrfs_dev_replace_item *ptr;
	u64 src_devid;

	path = btrfs_alloc_path();
	if (!path) {
		ret = -ENOMEM;
		goto out;
	}

	key.objectid = 0;
	key.type = BTRFS_DEV_REPLACE_KEY;
	key.offset = 0;
	ret = btrfs_search_slot(NULL, dev_root, &key, path, 0, 0);
	if (ret) {
no_valid_dev_replace_entry_found:
		ret = 0;
		dev_replace->replace_state =
			BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED;
		dev_replace->cont_reading_from_srcdev_mode =
		    BTRFS_DEV_REPLACE_ITEM_CONT_READING_FROM_SRCDEV_MODE_ALWAYS;
		dev_replace->time_started = 0;
		dev_replace->time_stopped = 0;
		atomic64_set(&dev_replace->num_write_errors, 0);
		atomic64_set(&dev_replace->num_uncorrectable_read_errors, 0);
		dev_replace->cursor_left = 0;
		dev_replace->committed_cursor_left = 0;
		dev_replace->cursor_left_last_write_of_item = 0;
		dev_replace->cursor_right = 0;
		dev_replace->srcdev = NULL;
		dev_replace->tgtdev = NULL;
		dev_replace->is_valid = 0;
		dev_replace->item_needs_writeback = 0;
		goto out;
	}
	slot = path->slots[0];
	eb = path->nodes[0];
	item_size = btrfs_item_size_nr(eb, slot);
	ptr = btrfs_item_ptr(eb, slot, struct btrfs_dev_replace_item);

	if (item_size != sizeof(struct btrfs_dev_replace_item)) {
		btrfs_warn(fs_info,
			"dev_replace entry found has unexpected size, ignore entry");
		goto no_valid_dev_replace_entry_found;
	}

	src_devid = btrfs_dev_replace_src_devid(eb, ptr);
	dev_replace->cont_reading_from_srcdev_mode =
		btrfs_dev_replace_cont_reading_from_srcdev_mode(eb, ptr);
	dev_replace->replace_state = btrfs_dev_replace_replace_state(eb, ptr);
	dev_replace->time_started = btrfs_dev_replace_time_started(eb, ptr);
	dev_replace->time_stopped =
		btrfs_dev_replace_time_stopped(eb, ptr);
	atomic64_set(&dev_replace->num_write_errors,
		     btrfs_dev_replace_num_write_errors(eb, ptr));
	atomic64_set(&dev_replace->num_uncorrectable_read_errors,
		     btrfs_dev_replace_num_uncorrectable_read_errors(eb, ptr));
	dev_replace->cursor_left = btrfs_dev_replace_cursor_left(eb, ptr);
	dev_replace->committed_cursor_left = dev_replace->cursor_left;
	dev_replace->cursor_left_last_write_of_item = dev_replace->cursor_left;
	dev_replace->cursor_right = btrfs_dev_replace_cursor_right(eb, ptr);
	dev_replace->is_valid = 1;

	dev_replace->item_needs_writeback = 0;
	switch (dev_replace->replace_state) {
	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
		dev_replace->srcdev = NULL;
		dev_replace->tgtdev = NULL;
		break;
	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
		dev_replace->srcdev = btrfs_find_device(fs_info->fs_devices,
						src_devid, NULL, NULL, true);
		dev_replace->tgtdev = btrfs_find_device(fs_info->fs_devices,
							BTRFS_DEV_REPLACE_DEVID,
							NULL, NULL, true);
		/*
		 * allow 'btrfs dev replace_cancel' if src/tgt device is
		 * missing
		 */
		if (!dev_replace->srcdev &&
		    !btrfs_test_opt(fs_info, DEGRADED)) {
			ret = -EIO;
			btrfs_warn(fs_info,
			   "cannot mount because device replace operation is ongoing and");
			btrfs_warn(fs_info,
			   "srcdev (devid %llu) is missing, need to run 'btrfs dev scan'?",
			   src_devid);
		}
		if (!dev_replace->tgtdev &&
		    !btrfs_test_opt(fs_info, DEGRADED)) {
			ret = -EIO;
			btrfs_warn(fs_info,
			   "cannot mount because device replace operation is ongoing and");
			btrfs_warn(fs_info,
			   "tgtdev (devid %llu) is missing, need to run 'btrfs dev scan'?",
				BTRFS_DEV_REPLACE_DEVID);
		}
		if (dev_replace->tgtdev) {
			if (dev_replace->srcdev) {
				dev_replace->tgtdev->total_bytes =
					dev_replace->srcdev->total_bytes;
				dev_replace->tgtdev->disk_total_bytes =
					dev_replace->srcdev->disk_total_bytes;
				dev_replace->tgtdev->commit_total_bytes =
					dev_replace->srcdev->commit_total_bytes;
				dev_replace->tgtdev->bytes_used =
					dev_replace->srcdev->bytes_used;
				dev_replace->tgtdev->commit_bytes_used =
					dev_replace->srcdev->commit_bytes_used;
			}
			set_bit(BTRFS_DEV_STATE_REPLACE_TGT,
				&dev_replace->tgtdev->dev_state);

			WARN_ON(fs_info->fs_devices->rw_devices == 0);
			dev_replace->tgtdev->io_width = fs_info->sectorsize;
			dev_replace->tgtdev->io_align = fs_info->sectorsize;
			dev_replace->tgtdev->sector_size = fs_info->sectorsize;
			dev_replace->tgtdev->fs_info = fs_info;
			set_bit(BTRFS_DEV_STATE_IN_FS_METADATA,
				&dev_replace->tgtdev->dev_state);
		}
		break;
	}

out:
	btrfs_free_path(path);
	return ret;
}

/*
 * Initialize a new device for device replace target from a given source dev
 * and path.
 *
 * Return 0 and new device in @device_out, otherwise return < 0
 */
static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
				  const char *device_path,
				  struct btrfs_device *srcdev,
				  struct btrfs_device **device_out)
{
	struct btrfs_device *device;
	struct block_device *bdev;
	struct rcu_string *name;
	u64 devid = BTRFS_DEV_REPLACE_DEVID;
	int ret = 0;

	*device_out = NULL;
	if (srcdev->fs_devices->seeding) {
		btrfs_err(fs_info, "the filesystem is a seed filesystem!");
		return -EINVAL;
	}

	bdev = blkdev_get_by_path(device_path, FMODE_WRITE | FMODE_EXCL,
				  fs_info->bdev_holder);
	if (IS_ERR(bdev)) {
		btrfs_err(fs_info, "target device %s is invalid!", device_path);
		return PTR_ERR(bdev);
	}

	sync_blockdev(bdev);

	list_for_each_entry(device, &fs_info->fs_devices->devices, dev_list) {
		if (device->bdev == bdev) {
			btrfs_err(fs_info,
				  "target device is in the filesystem!");
			ret = -EEXIST;
			goto error;
		}
	}


	if (i_size_read(bdev->bd_inode) <
	    btrfs_device_get_total_bytes(srcdev)) {
		btrfs_err(fs_info,
			  "target device is smaller than source device!");
		ret = -EINVAL;
		goto error;
	}


	device = btrfs_alloc_device(NULL, &devid, NULL);
	if (IS_ERR(device)) {
		ret = PTR_ERR(device);
		goto error;
	}

	name = rcu_string_strdup(device_path, GFP_KERNEL);
	if (!name) {
		btrfs_free_device(device);
		ret = -ENOMEM;
		goto error;
	}
	rcu_assign_pointer(device->name, name);

	set_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
	device->generation = 0;
	device->io_width = fs_info->sectorsize;
	device->io_align = fs_info->sectorsize;
	device->sector_size = fs_info->sectorsize;
	device->total_bytes = btrfs_device_get_total_bytes(srcdev);
	device->disk_total_bytes = btrfs_device_get_disk_total_bytes(srcdev);
	device->bytes_used = btrfs_device_get_bytes_used(srcdev);
	device->commit_total_bytes = srcdev->commit_total_bytes;
	device->commit_bytes_used = device->bytes_used;
	device->fs_info = fs_info;
	device->bdev = bdev;
	set_bit(BTRFS_DEV_STATE_IN_FS_METADATA, &device->dev_state);
	set_bit(BTRFS_DEV_STATE_REPLACE_TGT, &device->dev_state);
	device->mode = FMODE_EXCL;
	device->dev_stats_valid = 1;
	set_blocksize(device->bdev, BTRFS_BDEV_BLOCKSIZE);
	device->fs_devices = fs_info->fs_devices;

	mutex_lock(&fs_info->fs_devices->device_list_mutex);
	list_add(&device->dev_list, &fs_info->fs_devices->devices);
	fs_info->fs_devices->num_devices++;
	fs_info->fs_devices->open_devices++;
	mutex_unlock(&fs_info->fs_devices->device_list_mutex);

	*device_out = device;
	return 0;

error:
	blkdev_put(bdev, FMODE_EXCL);
	return ret;
}

/*
 * called from commit_transaction. Writes changed device replace state to
 * disk.
 */
int btrfs_run_dev_replace(struct btrfs_trans_handle *trans)
{
	struct btrfs_fs_info *fs_info = trans->fs_info;
	int ret;
	struct btrfs_root *dev_root = fs_info->dev_root;
	struct btrfs_path *path;
	struct btrfs_key key;
	struct extent_buffer *eb;
	struct btrfs_dev_replace_item *ptr;
	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;

	down_read(&dev_replace->rwsem);
	if (!dev_replace->is_valid ||
	    !dev_replace->item_needs_writeback) {
		up_read(&dev_replace->rwsem);
		return 0;
	}
	up_read(&dev_replace->rwsem);

	key.objectid = 0;
	key.type = BTRFS_DEV_REPLACE_KEY;
	key.offset = 0;

	path = btrfs_alloc_path();
	if (!path) {
		ret = -ENOMEM;
		goto out;
	}
	ret = btrfs_search_slot(trans, dev_root, &key, path, -1, 1);
	if (ret < 0) {
		btrfs_warn(fs_info,
			   "error %d while searching for dev_replace item!",
			   ret);
		goto out;
	}

	if (ret == 0 &&
	    btrfs_item_size_nr(path->nodes[0], path->slots[0]) < sizeof(*ptr)) {
		/*
		 * need to delete old one and insert a new one.
		 * Since no attempt is made to recover any old state, if the
		 * dev_replace state is 'running', the data on the target
		 * drive is lost.
		 * It would be possible to recover the state: just make sure
		 * that the beginning of the item is never changed and always
		 * contains all the essential information. Then read this
		 * minimal set of information and use it as a base for the
		 * new state.
		 */
		ret = btrfs_del_item(trans, dev_root, path);
		if (ret != 0) {
			btrfs_warn(fs_info,
				   "delete too small dev_replace item failed %d!",
				   ret);
			goto out;
		}
		ret = 1;
	}

	if (ret == 1) {
		/* need to insert a new item */
		btrfs_release_path(path);
		ret = btrfs_insert_empty_item(trans, dev_root, path,
					      &key, sizeof(*ptr));
		if (ret < 0) {
			btrfs_warn(fs_info,
				   "insert dev_replace item failed %d!", ret);
			goto out;
		}
	}

	eb = path->nodes[0];
	ptr = btrfs_item_ptr(eb, path->slots[0],
			     struct btrfs_dev_replace_item);

	down_write(&dev_replace->rwsem);
	if (dev_replace->srcdev)
		btrfs_set_dev_replace_src_devid(eb, ptr,
			dev_replace->srcdev->devid);
	else
		btrfs_set_dev_replace_src_devid(eb, ptr, (u64)-1);
	btrfs_set_dev_replace_cont_reading_from_srcdev_mode(eb, ptr,
		dev_replace->cont_reading_from_srcdev_mode);
	btrfs_set_dev_replace_replace_state(eb, ptr,
		dev_replace->replace_state);
	btrfs_set_dev_replace_time_started(eb, ptr, dev_replace->time_started);
	btrfs_set_dev_replace_time_stopped(eb, ptr, dev_replace->time_stopped);
	btrfs_set_dev_replace_num_write_errors(eb, ptr,
		atomic64_read(&dev_replace->num_write_errors));
	btrfs_set_dev_replace_num_uncorrectable_read_errors(eb, ptr,
		atomic64_read(&dev_replace->num_uncorrectable_read_errors));
	dev_replace->cursor_left_last_write_of_item =
		dev_replace->cursor_left;
	btrfs_set_dev_replace_cursor_left(eb, ptr,
		dev_replace->cursor_left_last_write_of_item);
	btrfs_set_dev_replace_cursor_right(eb, ptr,
		dev_replace->cursor_right);
	dev_replace->item_needs_writeback = 0;
	up_write(&dev_replace->rwsem);

	btrfs_mark_buffer_dirty(eb);

out:
	btrfs_free_path(path);

	return ret;
}

static char* btrfs_dev_name(struct btrfs_device *device)
{
	if (!device || test_bit(BTRFS_DEV_STATE_MISSING, &device->dev_state))
		return "<missing disk>";
	else
		return rcu_str_deref(device->name);
}

static int btrfs_dev_replace_start(struct btrfs_fs_info *fs_info,
		const char *tgtdev_name, u64 srcdevid, const char *srcdev_name,
		int read_src)
{
	struct btrfs_root *root = fs_info->dev_root;
	struct btrfs_trans_handle *trans;
	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
	int ret;
	struct btrfs_device *tgt_device = NULL;
	struct btrfs_device *src_device = NULL;

	src_device = btrfs_find_device_by_devspec(fs_info, srcdevid,
						  srcdev_name);
	if (IS_ERR(src_device))
		return PTR_ERR(src_device);

	if (btrfs_pinned_by_swapfile(fs_info, src_device)) {
		btrfs_warn_in_rcu(fs_info,
	  "cannot replace device %s (devid %llu) due to active swapfile",
			btrfs_dev_name(src_device), src_device->devid);
		return -ETXTBSY;
	}

	/*
	 * Here we commit the transaction to make sure commit_total_bytes
	 * of all the devices are updated.
	 */
	trans = btrfs_attach_transaction(root);
	if (!IS_ERR(trans)) {
		ret = btrfs_commit_transaction(trans);
		if (ret)
			return ret;
	} else if (PTR_ERR(trans) != -ENOENT) {
		return PTR_ERR(trans);
	}

	ret = btrfs_init_dev_replace_tgtdev(fs_info, tgtdev_name,
					    src_device, &tgt_device);
	if (ret)
		return ret;

	down_write(&dev_replace->rwsem);
	switch (dev_replace->replace_state) {
	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
		break;
	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
		ASSERT(0);
		ret = BTRFS_IOCTL_DEV_REPLACE_RESULT_ALREADY_STARTED;
		up_write(&dev_replace->rwsem);
		goto leave;
	}

	dev_replace->cont_reading_from_srcdev_mode = read_src;
	dev_replace->srcdev = src_device;
	dev_replace->tgtdev = tgt_device;

	btrfs_info_in_rcu(fs_info,
		      "dev_replace from %s (devid %llu) to %s started",
		      btrfs_dev_name(src_device),
		      src_device->devid,
		      rcu_str_deref(tgt_device->name));

	/*
	 * from now on, the writes to the srcdev are all duplicated to
	 * go to the tgtdev as well (refer to btrfs_map_block()).
	 */
	dev_replace->replace_state = BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED;
	dev_replace->time_started = ktime_get_real_seconds();
	dev_replace->cursor_left = 0;
	dev_replace->committed_cursor_left = 0;
	dev_replace->cursor_left_last_write_of_item = 0;
	dev_replace->cursor_right = 0;
	dev_replace->is_valid = 1;
	dev_replace->item_needs_writeback = 1;
	atomic64_set(&dev_replace->num_write_errors, 0);
	atomic64_set(&dev_replace->num_uncorrectable_read_errors, 0);
	up_write(&dev_replace->rwsem);

	ret = btrfs_sysfs_add_device(tgt_device);
	if (ret)
		btrfs_err(fs_info, "kobj add dev failed %d", ret);

	btrfs_wait_ordered_roots(fs_info, U64_MAX, 0, (u64)-1);

	/* Commit dev_replace state and reserve 1 item for it. */
	trans = btrfs_start_transaction(root, 1);
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
		down_write(&dev_replace->rwsem);
		dev_replace->replace_state =
			BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED;
		dev_replace->srcdev = NULL;
		dev_replace->tgtdev = NULL;
		up_write(&dev_replace->rwsem);
		goto leave;
	}

	ret = btrfs_commit_transaction(trans);
	WARN_ON(ret);

	/* the disk copy procedure reuses the scrub code */
	ret = btrfs_scrub_dev(fs_info, src_device->devid, 0,
			      btrfs_device_get_total_bytes(src_device),
			      &dev_replace->scrub_progress, 0, 1);

	ret = btrfs_dev_replace_finishing(fs_info, ret);
	if (ret == -EINPROGRESS)
		ret = BTRFS_IOCTL_DEV_REPLACE_RESULT_SCRUB_INPROGRESS;

	return ret;

leave:
	btrfs_destroy_dev_replace_tgtdev(tgt_device);
	return ret;
}

int btrfs_dev_replace_by_ioctl(struct btrfs_fs_info *fs_info,
			    struct btrfs_ioctl_dev_replace_args *args)
{
	int ret;

	switch (args->start.cont_reading_from_srcdev_mode) {
	case BTRFS_IOCTL_DEV_REPLACE_CONT_READING_FROM_SRCDEV_MODE_ALWAYS:
	case BTRFS_IOCTL_DEV_REPLACE_CONT_READING_FROM_SRCDEV_MODE_AVOID:
		break;
	default:
		return -EINVAL;
	}

	if ((args->start.srcdevid == 0 && args->start.srcdev_name[0] == '\0') ||
	    args->start.tgtdev_name[0] == '\0')
		return -EINVAL;

	ret = btrfs_dev_replace_start(fs_info, args->start.tgtdev_name,
					args->start.srcdevid,
					args->start.srcdev_name,
					args->start.cont_reading_from_srcdev_mode);
	args->result = ret;
	/* don't warn if EINPROGRESS, someone else might be running scrub */
	if (ret == BTRFS_IOCTL_DEV_REPLACE_RESULT_SCRUB_INPROGRESS ||
	    ret == BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR)
		return 0;

	return ret;
}

/*
 * blocked until all in-flight bios operations are finished.
 */
static void btrfs_rm_dev_replace_blocked(struct btrfs_fs_info *fs_info)
{
	set_bit(BTRFS_FS_STATE_DEV_REPLACING, &fs_info->fs_state);
	wait_event(fs_info->dev_replace.replace_wait, !percpu_counter_sum(
		   &fs_info->dev_replace.bio_counter));
}

/*
 * we have removed target device, it is safe to allow new bios request.
 */
static void btrfs_rm_dev_replace_unblocked(struct btrfs_fs_info *fs_info)
{
	clear_bit(BTRFS_FS_STATE_DEV_REPLACING, &fs_info->fs_state);
	wake_up(&fs_info->dev_replace.replace_wait);
}

/*
 * When finishing the device replace, before swapping the source device with the
 * target device we must update the chunk allocation state in the target device,
 * as it is empty because replace works by directly copying the chunks and not
 * through the normal chunk allocation path.
 */
static int btrfs_set_target_alloc_state(struct btrfs_device *srcdev,
					struct btrfs_device *tgtdev)
{
	struct extent_state *cached_state = NULL;
	u64 start = 0;
	u64 found_start;
	u64 found_end;
	int ret = 0;

	lockdep_assert_held(&srcdev->fs_info->chunk_mutex);

	while (!find_first_extent_bit(&srcdev->alloc_state, start,
				      &found_start, &found_end,
				      CHUNK_ALLOCATED, &cached_state)) {
		ret = set_extent_bits(&tgtdev->alloc_state, found_start,
				      found_end, CHUNK_ALLOCATED);
		if (ret)
			break;
		start = found_end + 1;
	}

	free_extent_state(cached_state);
	return ret;
}

static void btrfs_dev_replace_update_device_in_mapping_tree(
						struct btrfs_fs_info *fs_info,
						struct btrfs_device *srcdev,
						struct btrfs_device *tgtdev)
{
	struct extent_map_tree *em_tree = &fs_info->mapping_tree;
	struct extent_map *em;
	struct map_lookup *map;
	u64 start = 0;
	int i;

	write_lock(&em_tree->lock);
	do {
		em = lookup_extent_mapping(em_tree, start, (u64)-1);
		if (!em)
			break;
		map = em->map_lookup;
		for (i = 0; i < map->num_stripes; i++)
			if (srcdev == map->stripes[i].dev)
				map->stripes[i].dev = tgtdev;
		start = em->start + em->len;
		free_extent_map(em);
	} while (start);
	write_unlock(&em_tree->lock);
}

static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
				       int scrub_ret)
{
	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
	struct btrfs_device *tgt_device;
	struct btrfs_device *src_device;
	struct btrfs_root *root = fs_info->tree_root;
	u8 uuid_tmp[BTRFS_UUID_SIZE];
	struct btrfs_trans_handle *trans;
	int ret = 0;

	/* don't allow cancel or unmount to disturb the finishing procedure */
	mutex_lock(&dev_replace->lock_finishing_cancel_unmount);

	down_read(&dev_replace->rwsem);
	/* was the operation canceled, or is it finished? */
	if (dev_replace->replace_state !=
	    BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED) {
		up_read(&dev_replace->rwsem);
		mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
		return 0;
	}

	tgt_device = dev_replace->tgtdev;
	src_device = dev_replace->srcdev;
	up_read(&dev_replace->rwsem);

	/*
	 * flush all outstanding I/O and inode extent mappings before the
	 * copy operation is declared as being finished
	 */
	ret = btrfs_start_delalloc_roots(fs_info, U64_MAX);
	if (ret) {
		mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
		return ret;
	}
	btrfs_wait_ordered_roots(fs_info, U64_MAX, 0, (u64)-1);

	/*
	 * We have to use this loop approach because at this point src_device
	 * has to be available for transaction commit to complete, yet new
	 * chunks shouldn't be allocated on the device.
	 */
	while (1) {
		trans = btrfs_start_transaction(root, 0);
		if (IS_ERR(trans)) {
			mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
			return PTR_ERR(trans);
		}
		ret = btrfs_commit_transaction(trans);
		WARN_ON(ret);

		/* Prevent write_all_supers() during the finishing procedure */
		mutex_lock(&fs_info->fs_devices->device_list_mutex);
		/* Prevent new chunks being allocated on the source device */
		mutex_lock(&fs_info->chunk_mutex);

		if (!list_empty(&src_device->post_commit_list)) {
			mutex_unlock(&fs_info->fs_devices->device_list_mutex);
			mutex_unlock(&fs_info->chunk_mutex);
		} else {
			break;
		}
	}

	down_write(&dev_replace->rwsem);
	dev_replace->replace_state =
		scrub_ret ? BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED
			  : BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED;
	dev_replace->tgtdev = NULL;
	dev_replace->srcdev = NULL;
	dev_replace->time_stopped = ktime_get_real_seconds();
	dev_replace->item_needs_writeback = 1;

	/*
	 * Update allocation state in the new device and replace the old device
	 * with the new one in the mapping tree.
	 */
	if (!scrub_ret) {
		scrub_ret = btrfs_set_target_alloc_state(src_device, tgt_device);
		if (scrub_ret)
			goto error;
		btrfs_dev_replace_update_device_in_mapping_tree(fs_info,
								src_device,
								tgt_device);
	} else {
		if (scrub_ret != -ECANCELED)
			btrfs_err_in_rcu(fs_info,
				 "btrfs_scrub_dev(%s, %llu, %s) failed %d",
				 btrfs_dev_name(src_device),
				 src_device->devid,
				 rcu_str_deref(tgt_device->name), scrub_ret);
error:
		up_write(&dev_replace->rwsem);
		mutex_unlock(&fs_info->chunk_mutex);
		mutex_unlock(&fs_info->fs_devices->device_list_mutex);
		btrfs_rm_dev_replace_blocked(fs_info);
		if (tgt_device)
			btrfs_destroy_dev_replace_tgtdev(tgt_device);
		btrfs_rm_dev_replace_unblocked(fs_info);
		mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);

		return scrub_ret;
	}

	btrfs_info_in_rcu(fs_info,
			  "dev_replace from %s (devid %llu) to %s finished",
			  btrfs_dev_name(src_device),
			  src_device->devid,
			  rcu_str_deref(tgt_device->name));
	clear_bit(BTRFS_DEV_STATE_REPLACE_TGT, &tgt_device->dev_state);
	tgt_device->devid = src_device->devid;
	src_device->devid = BTRFS_DEV_REPLACE_DEVID;
	memcpy(uuid_tmp, tgt_device->uuid, sizeof(uuid_tmp));
	memcpy(tgt_device->uuid, src_device->uuid, sizeof(tgt_device->uuid));
	memcpy(src_device->uuid, uuid_tmp, sizeof(src_device->uuid));
	btrfs_device_set_total_bytes(tgt_device, src_device->total_bytes);
	btrfs_device_set_disk_total_bytes(tgt_device,
					  src_device->disk_total_bytes);
	btrfs_device_set_bytes_used(tgt_device, src_device->bytes_used);
	tgt_device->commit_bytes_used = src_device->bytes_used;

	btrfs_assign_next_active_device(src_device, tgt_device);

	list_add(&tgt_device->dev_alloc_list, &fs_info->fs_devices->alloc_list);
	fs_info->fs_devices->rw_devices++;

	up_write(&dev_replace->rwsem);
	btrfs_rm_dev_replace_blocked(fs_info);

	btrfs_rm_dev_replace_remove_srcdev(src_device);

	btrfs_rm_dev_replace_unblocked(fs_info);

	/*
	 * Increment dev_stats_ccnt so that btrfs_run_dev_stats() will
	 * update on-disk dev stats value during commit transaction
	 */
	atomic_inc(&tgt_device->dev_stats_ccnt);

	/*
	 * this is again a consistent state where no dev_replace procedure
	 * is running, the target device is part of the filesystem, the
	 * source device is not part of the filesystem anymore and its 1st
	 * superblock is scratched out so that it is no longer marked to
	 * belong to this filesystem.
	 */
	mutex_unlock(&fs_info->chunk_mutex);
	mutex_unlock(&fs_info->fs_devices->device_list_mutex);

	/* replace the sysfs entry */
	btrfs_sysfs_remove_device(src_device);
	btrfs_sysfs_update_devid(tgt_device);
	if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &src_device->dev_state))
		btrfs_scratch_superblocks(fs_info, src_device->bdev,
					  src_device->name->str);

	/* write back the superblocks */
	trans = btrfs_start_transaction(root, 0);
	if (!IS_ERR(trans))
		btrfs_commit_transaction(trans);

	mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);

	btrfs_rm_dev_replace_free_srcdev(src_device);

	return 0;
}

/*
 * Read progress of device replace status according to the state and last
 * stored position. The value format is the same as for
 * btrfs_dev_replace::progress_1000
 */
static u64 btrfs_dev_replace_progress(struct btrfs_fs_info *fs_info)
{
	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
	u64 ret = 0;

	switch (dev_replace->replace_state) {
	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
		ret = 0;
		break;
	case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
		ret = 1000;
		break;
	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
		ret = div64_u64(dev_replace->cursor_left,
				div_u64(btrfs_device_get_total_bytes(
						dev_replace->srcdev), 1000));
		break;
	}

	return ret;
}

void btrfs_dev_replace_status(struct btrfs_fs_info *fs_info,
			      struct btrfs_ioctl_dev_replace_args *args)
{
	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;

	down_read(&dev_replace->rwsem);
	/* even if !dev_replace_is_valid, the values are good enough for
	 * the replace_status ioctl */
	args->result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
	args->status.replace_state = dev_replace->replace_state;
	args->status.time_started = dev_replace->time_started;
	args->status.time_stopped = dev_replace->time_stopped;
	args->status.num_write_errors =
		atomic64_read(&dev_replace->num_write_errors);
	args->status.num_uncorrectable_read_errors =
		atomic64_read(&dev_replace->num_uncorrectable_read_errors);
	args->status.progress_1000 = btrfs_dev_replace_progress(fs_info);
	up_read(&dev_replace->rwsem);
}

int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info)
{
	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
	struct btrfs_device *tgt_device = NULL;
	struct btrfs_device *src_device = NULL;
	struct btrfs_trans_handle *trans;
	struct btrfs_root *root = fs_info->tree_root;
	int result;
	int ret;

	if (sb_rdonly(fs_info->sb))
		return -EROFS;

	mutex_lock(&dev_replace->lock_finishing_cancel_unmount);
	down_write(&dev_replace->rwsem);
	switch (dev_replace->replace_state) {
	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
		result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NOT_STARTED;
		up_write(&dev_replace->rwsem);
		break;
	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
		tgt_device = dev_replace->tgtdev;
		src_device = dev_replace->srcdev;
		up_write(&dev_replace->rwsem);
		ret = btrfs_scrub_cancel(fs_info);
		if (ret < 0) {
			result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NOT_STARTED;
		} else {
			result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
			/*
			 * btrfs_dev_replace_finishing() will handle the
			 * cleanup part
			 */
			btrfs_info_in_rcu(fs_info,
				"dev_replace from %s (devid %llu) to %s canceled",
				btrfs_dev_name(src_device), src_device->devid,
				btrfs_dev_name(tgt_device));
		}
		break;
	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
		/*
		 * Scrub doing the replace isn't running so we need to do the
		 * cleanup step of btrfs_dev_replace_finishing() here
		 */
		result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
		tgt_device = dev_replace->tgtdev;
		src_device = dev_replace->srcdev;
		dev_replace->tgtdev = NULL;
		dev_replace->srcdev = NULL;
		dev_replace->replace_state =
				BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED;
		dev_replace->time_stopped = ktime_get_real_seconds();
		dev_replace->item_needs_writeback = 1;

		up_write(&dev_replace->rwsem);

		/* Scrub for replace must not be running in suspended state */
		ret = btrfs_scrub_cancel(fs_info);
		ASSERT(ret != -ENOTCONN);

		trans = btrfs_start_transaction(root, 0);
		if (IS_ERR(trans)) {
			mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
			return PTR_ERR(trans);
		}
		ret = btrfs_commit_transaction(trans);
		WARN_ON(ret);

		btrfs_info_in_rcu(fs_info,
		"suspended dev_replace from %s (devid %llu) to %s canceled",
			btrfs_dev_name(src_device), src_device->devid,
			btrfs_dev_name(tgt_device));

		if (tgt_device)
			btrfs_destroy_dev_replace_tgtdev(tgt_device);
		break;
	default:
		up_write(&dev_replace->rwsem);
		result = -EINVAL;
	}

	mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
	return result;
}

void btrfs_dev_replace_suspend_for_unmount(struct btrfs_fs_info *fs_info)
{
	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;

	mutex_lock(&dev_replace->lock_finishing_cancel_unmount);
	down_write(&dev_replace->rwsem);

	switch (dev_replace->replace_state) {
	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
		break;
	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
		dev_replace->replace_state =
			BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED;
		dev_replace->time_stopped = ktime_get_real_seconds();
		dev_replace->item_needs_writeback = 1;
		btrfs_info(fs_info, "suspending dev_replace for unmount");
		break;
	}

	up_write(&dev_replace->rwsem);
	mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
}

/* resume dev_replace procedure that was interrupted by unmount */
int btrfs_resume_dev_replace_async(struct btrfs_fs_info *fs_info)
{
	struct task_struct *task;
	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;

	down_write(&dev_replace->rwsem);

	switch (dev_replace->replace_state) {
	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
		up_write(&dev_replace->rwsem);
		return 0;
	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
		break;
	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
		dev_replace->replace_state =
			BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED;
		break;
	}
	if (!dev_replace->tgtdev || !dev_replace->tgtdev->bdev) {
		btrfs_info(fs_info,
			   "cannot continue dev_replace, tgtdev is missing");
		btrfs_info(fs_info,
			   "you may cancel the operation after 'mount -o degraded'");
		dev_replace->replace_state =
					BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED;
		up_write(&dev_replace->rwsem);
		return 0;
	}
	up_write(&dev_replace->rwsem);

	/*
	 * This could collide with a paused balance, but the exclusive op logic
	 * should never allow both to start and pause. We don't want to allow
	 * dev-replace to start anyway.
	 */
	if (!btrfs_exclop_start(fs_info, BTRFS_EXCLOP_DEV_REPLACE)) {
		down_write(&dev_replace->rwsem);
		dev_replace->replace_state =
					BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED;
		up_write(&dev_replace->rwsem);
		btrfs_info(fs_info,
		"cannot resume dev-replace, other exclusive operation running");
		return 0;
	}

	task = kthread_run(btrfs_dev_replace_kthread, fs_info, "btrfs-devrepl");
	return PTR_ERR_OR_ZERO(task);
}

static int btrfs_dev_replace_kthread(void *data)
{
	struct btrfs_fs_info *fs_info = data;
	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
	u64 progress;
	int ret;

	progress = btrfs_dev_replace_progress(fs_info);
	progress = div_u64(progress, 10);
	btrfs_info_in_rcu(fs_info,
		"continuing dev_replace from %s (devid %llu) to target %s @%u%%",
		btrfs_dev_name(dev_replace->srcdev),
		dev_replace->srcdev->devid,
		btrfs_dev_name(dev_replace->tgtdev),
		(unsigned int)progress);

	ret = btrfs_scrub_dev(fs_info, dev_replace->srcdev->devid,
			      dev_replace->committed_cursor_left,
			      btrfs_device_get_total_bytes(dev_replace->srcdev),
			      &dev_replace->scrub_progress, 0, 1);
	ret = btrfs_dev_replace_finishing(fs_info, ret);
	WARN_ON(ret && ret != -ECANCELED);

	btrfs_exclop_finish(fs_info);
	return 0;
}

int __pure btrfs_dev_replace_is_ongoing(struct btrfs_dev_replace *dev_replace)
{
	if (!dev_replace->is_valid)
		return 0;

	switch (dev_replace->replace_state) {
	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
		return 0;
	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
		/*
		 * return true even if tgtdev is missing (this is
		 * something that can happen if the dev_replace
		 * procedure is suspended by an umount and then
		 * the tgtdev is missing (or "btrfs dev scan") was
		 * not called and the filesystem is remounted
		 * in degraded state. This does not stop the
		 * dev_replace procedure. It needs to be canceled
		 * manually if the cancellation is wanted.
		 */
		break;
	}
	return 1;
}

void btrfs_bio_counter_inc_noblocked(struct btrfs_fs_info *fs_info)
{
	percpu_counter_inc(&fs_info->dev_replace.bio_counter);
}

void btrfs_bio_counter_sub(struct btrfs_fs_info *fs_info, s64 amount)
{
	percpu_counter_sub(&fs_info->dev_replace.bio_counter, amount);
	cond_wake_up_nomb(&fs_info->dev_replace.replace_wait);
}

void btrfs_bio_counter_inc_blocked(struct btrfs_fs_info *fs_info)
{
	while (1) {
		percpu_counter_inc(&fs_info->dev_replace.bio_counter);
		if (likely(!test_bit(BTRFS_FS_STATE_DEV_REPLACING,
				     &fs_info->fs_state)))
			break;

		btrfs_bio_counter_dec(fs_info);
		wait_event(fs_info->dev_replace.replace_wait,
			   !test_bit(BTRFS_FS_STATE_DEV_REPLACING,
				     &fs_info->fs_state));
	}
}