summaryrefslogtreecommitdiffstats
path: root/fs/bcachefs/btree_update.c
blob: 8e47e260eba59b09097fd07b1d9063a0e665a812 (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
// SPDX-License-Identifier: GPL-2.0

#include "bcachefs.h"
#include "btree_update.h"
#include "btree_iter.h"
#include "btree_journal_iter.h"
#include "btree_locking.h"
#include "buckets.h"
#include "debug.h"
#include "errcode.h"
#include "error.h"
#include "extents.h"
#include "keylist.h"
#include "snapshot.h"
#include "trace.h"

static inline int btree_insert_entry_cmp(const struct btree_insert_entry *l,
					 const struct btree_insert_entry *r)
{
	return   cmp_int(l->btree_id,	r->btree_id) ?:
		 cmp_int(l->cached,	r->cached) ?:
		 -cmp_int(l->level,	r->level) ?:
		 bpos_cmp(l->k->k.p,	r->k->k.p);
}

static int __must_check
bch2_trans_update_by_path(struct btree_trans *, btree_path_idx_t,
			  struct bkey_i *, enum btree_update_flags,
			  unsigned long ip);

static noinline int extent_front_merge(struct btree_trans *trans,
				       struct btree_iter *iter,
				       struct bkey_s_c k,
				       struct bkey_i **insert,
				       enum btree_update_flags flags)
{
	struct bch_fs *c = trans->c;
	struct bkey_i *update;
	int ret;

	if (unlikely(trans->journal_replay_not_finished))
		return 0;

	update = bch2_bkey_make_mut_noupdate(trans, k);
	ret = PTR_ERR_OR_ZERO(update);
	if (ret)
		return ret;

	if (!bch2_bkey_merge(c, bkey_i_to_s(update), bkey_i_to_s_c(*insert)))
		return 0;

	ret =   bch2_key_has_snapshot_overwrites(trans, iter->btree_id, k.k->p) ?:
		bch2_key_has_snapshot_overwrites(trans, iter->btree_id, (*insert)->k.p);
	if (ret < 0)
		return ret;
	if (ret)
		return 0;

	ret = bch2_btree_delete_at(trans, iter, flags);
	if (ret)
		return ret;

	*insert = update;
	return 0;
}

static noinline int extent_back_merge(struct btree_trans *trans,
				      struct btree_iter *iter,
				      struct bkey_i *insert,
				      struct bkey_s_c k)
{
	struct bch_fs *c = trans->c;
	int ret;

	if (unlikely(trans->journal_replay_not_finished))
		return 0;

	ret =   bch2_key_has_snapshot_overwrites(trans, iter->btree_id, insert->k.p) ?:
		bch2_key_has_snapshot_overwrites(trans, iter->btree_id, k.k->p);
	if (ret < 0)
		return ret;
	if (ret)
		return 0;

	bch2_bkey_merge(c, bkey_i_to_s(insert), k);
	return 0;
}

/*
 * When deleting, check if we need to emit a whiteout (because we're overwriting
 * something in an ancestor snapshot)
 */
static int need_whiteout_for_snapshot(struct btree_trans *trans,
				      enum btree_id btree_id, struct bpos pos)
{
	struct btree_iter iter;
	struct bkey_s_c k;
	u32 snapshot = pos.snapshot;
	int ret;

	if (!bch2_snapshot_parent(trans->c, pos.snapshot))
		return 0;

	pos.snapshot++;

	for_each_btree_key_norestart(trans, iter, btree_id, pos,
			   BTREE_ITER_ALL_SNAPSHOTS|
			   BTREE_ITER_NOPRESERVE, k, ret) {
		if (!bkey_eq(k.k->p, pos))
			break;

		if (bch2_snapshot_is_ancestor(trans->c, snapshot,
					      k.k->p.snapshot)) {
			ret = !bkey_whiteout(k.k);
			break;
		}
	}
	bch2_trans_iter_exit(trans, &iter);

	return ret;
}

int __bch2_insert_snapshot_whiteouts(struct btree_trans *trans,
				   enum btree_id id,
				   struct bpos old_pos,
				   struct bpos new_pos)
{
	struct bch_fs *c = trans->c;
	struct btree_iter old_iter, new_iter = { NULL };
	struct bkey_s_c old_k, new_k;
	snapshot_id_list s;
	struct bkey_i *update;
	int ret = 0;

	if (!bch2_snapshot_has_children(c, old_pos.snapshot))
		return 0;

	darray_init(&s);

	bch2_trans_iter_init(trans, &old_iter, id, old_pos,
			     BTREE_ITER_NOT_EXTENTS|
			     BTREE_ITER_ALL_SNAPSHOTS);
	while ((old_k = bch2_btree_iter_prev(&old_iter)).k &&
	       !(ret = bkey_err(old_k)) &&
	       bkey_eq(old_pos, old_k.k->p)) {
		struct bpos whiteout_pos =
			SPOS(new_pos.inode, new_pos.offset, old_k.k->p.snapshot);;

		if (!bch2_snapshot_is_ancestor(c, old_k.k->p.snapshot, old_pos.snapshot) ||
		    snapshot_list_has_ancestor(c, &s, old_k.k->p.snapshot))
			continue;

		new_k = bch2_bkey_get_iter(trans, &new_iter, id, whiteout_pos,
					   BTREE_ITER_NOT_EXTENTS|
					   BTREE_ITER_INTENT);
		ret = bkey_err(new_k);
		if (ret)
			break;

		if (new_k.k->type == KEY_TYPE_deleted) {
			update = bch2_trans_kmalloc(trans, sizeof(struct bkey_i));
			ret = PTR_ERR_OR_ZERO(update);
			if (ret)
				break;

			bkey_init(&update->k);
			update->k.p		= whiteout_pos;
			update->k.type		= KEY_TYPE_whiteout;

			ret = bch2_trans_update(trans, &new_iter, update,
						BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
		}
		bch2_trans_iter_exit(trans, &new_iter);

		ret = snapshot_list_add(c, &s, old_k.k->p.snapshot);
		if (ret)
			break;
	}
	bch2_trans_iter_exit(trans, &new_iter);
	bch2_trans_iter_exit(trans, &old_iter);
	darray_exit(&s);

	return ret;
}

int bch2_trans_update_extent_overwrite(struct btree_trans *trans,
				       struct btree_iter *iter,
				       enum btree_update_flags flags,
				       struct bkey_s_c old,
				       struct bkey_s_c new)
{
	enum btree_id btree_id = iter->btree_id;
	struct bkey_i *update;
	struct bpos new_start = bkey_start_pos(new.k);
	unsigned front_split = bkey_lt(bkey_start_pos(old.k), new_start);
	unsigned back_split  = bkey_gt(old.k->p, new.k->p);
	unsigned middle_split = (front_split || back_split) &&
		old.k->p.snapshot != new.k->p.snapshot;
	unsigned nr_splits = front_split + back_split + middle_split;
	int ret = 0, compressed_sectors;

	/*
	 * If we're going to be splitting a compressed extent, note it
	 * so that __bch2_trans_commit() can increase our disk
	 * reservation:
	 */
	if (nr_splits > 1 &&
	    (compressed_sectors = bch2_bkey_sectors_compressed(old)))
		trans->extra_disk_res += compressed_sectors * (nr_splits - 1);

	if (front_split) {
		update = bch2_bkey_make_mut_noupdate(trans, old);
		if ((ret = PTR_ERR_OR_ZERO(update)))
			return ret;

		bch2_cut_back(new_start, update);

		ret =   bch2_insert_snapshot_whiteouts(trans, btree_id,
					old.k->p, update->k.p) ?:
			bch2_btree_insert_nonextent(trans, btree_id, update,
					BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|flags);
		if (ret)
			return ret;
	}

	/* If we're overwriting in a different snapshot - middle split: */
	if (middle_split) {
		update = bch2_bkey_make_mut_noupdate(trans, old);
		if ((ret = PTR_ERR_OR_ZERO(update)))
			return ret;

		bch2_cut_front(new_start, update);
		bch2_cut_back(new.k->p, update);

		ret =   bch2_insert_snapshot_whiteouts(trans, btree_id,
					old.k->p, update->k.p) ?:
			bch2_btree_insert_nonextent(trans, btree_id, update,
					  BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|flags);
		if (ret)
			return ret;
	}

	if (bkey_le(old.k->p, new.k->p)) {
		update = bch2_trans_kmalloc(trans, sizeof(*update));
		if ((ret = PTR_ERR_OR_ZERO(update)))
			return ret;

		bkey_init(&update->k);
		update->k.p = old.k->p;
		update->k.p.snapshot = new.k->p.snapshot;

		if (new.k->p.snapshot != old.k->p.snapshot) {
			update->k.type = KEY_TYPE_whiteout;
		} else if (btree_type_has_snapshots(btree_id)) {
			ret = need_whiteout_for_snapshot(trans, btree_id, update->k.p);
			if (ret < 0)
				return ret;
			if (ret)
				update->k.type = KEY_TYPE_whiteout;
		}

		ret = bch2_btree_insert_nonextent(trans, btree_id, update,
					  BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|flags);
		if (ret)
			return ret;
	}

	if (back_split) {
		update = bch2_bkey_make_mut_noupdate(trans, old);
		if ((ret = PTR_ERR_OR_ZERO(update)))
			return ret;

		bch2_cut_front(new.k->p, update);

		ret = bch2_trans_update_by_path(trans, iter->path, update,
					  BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|
					  flags, _RET_IP_);
		if (ret)
			return ret;
	}

	return 0;
}

static int bch2_trans_update_extent(struct btree_trans *trans,
				    struct btree_iter *orig_iter,
				    struct bkey_i *insert,
				    enum btree_update_flags flags)
{
	struct btree_iter iter;
	struct bkey_s_c k;
	enum btree_id btree_id = orig_iter->btree_id;
	int ret = 0;

	bch2_trans_iter_init(trans, &iter, btree_id, bkey_start_pos(&insert->k),
			     BTREE_ITER_INTENT|
			     BTREE_ITER_WITH_UPDATES|
			     BTREE_ITER_NOT_EXTENTS);
	k = bch2_btree_iter_peek_upto(&iter, POS(insert->k.p.inode, U64_MAX));
	if ((ret = bkey_err(k)))
		goto err;
	if (!k.k)
		goto out;

	if (bkey_eq(k.k->p, bkey_start_pos(&insert->k))) {
		if (bch2_bkey_maybe_mergable(k.k, &insert->k)) {
			ret = extent_front_merge(trans, &iter, k, &insert, flags);
			if (ret)
				goto err;
		}

		goto next;
	}

	while (bkey_gt(insert->k.p, bkey_start_pos(k.k))) {
		bool done = bkey_lt(insert->k.p, k.k->p);

		ret = bch2_trans_update_extent_overwrite(trans, &iter, flags, k, bkey_i_to_s_c(insert));
		if (ret)
			goto err;

		if (done)
			goto out;
next:
		bch2_btree_iter_advance(&iter);
		k = bch2_btree_iter_peek_upto(&iter, POS(insert->k.p.inode, U64_MAX));
		if ((ret = bkey_err(k)))
			goto err;
		if (!k.k)
			goto out;
	}

	if (bch2_bkey_maybe_mergable(&insert->k, k.k)) {
		ret = extent_back_merge(trans, &iter, insert, k);
		if (ret)
			goto err;
	}
out:
	if (!bkey_deleted(&insert->k))
		ret = bch2_btree_insert_nonextent(trans, btree_id, insert, flags);
err:
	bch2_trans_iter_exit(trans, &iter);

	return ret;
}

static noinline int flush_new_cached_update(struct btree_trans *trans,
					    struct btree_insert_entry *i,
					    enum btree_update_flags flags,
					    unsigned long ip)
{
	struct bkey k;
	int ret;

	btree_path_idx_t path_idx =
		bch2_path_get(trans, i->btree_id, i->old_k.p, 1, 0,
			      BTREE_ITER_INTENT, _THIS_IP_);
	ret = bch2_btree_path_traverse(trans, path_idx, 0);
	if (ret)
		goto out;

	struct btree_path *btree_path = trans->paths + path_idx;

	/*
	 * The old key in the insert entry might actually refer to an existing
	 * key in the btree that has been deleted from cache and not yet
	 * flushed. Check for this and skip the flush so we don't run triggers
	 * against a stale key.
	 */
	bch2_btree_path_peek_slot_exact(btree_path, &k);
	if (!bkey_deleted(&k))
		goto out;

	i->key_cache_already_flushed = true;
	i->flags |= BTREE_TRIGGER_NORUN;

	btree_path_set_should_be_locked(btree_path);
	ret = bch2_trans_update_by_path(trans, path_idx, i->k, flags, ip);
out:
	bch2_path_put(trans, path_idx, true);
	return ret;
}

static int __must_check
bch2_trans_update_by_path(struct btree_trans *trans, btree_path_idx_t path_idx,
			  struct bkey_i *k, enum btree_update_flags flags,
			  unsigned long ip)
{
	struct bch_fs *c = trans->c;
	struct btree_insert_entry *i, n;
	int cmp;

	struct btree_path *path = trans->paths + path_idx;
	EBUG_ON(!path->should_be_locked);
	EBUG_ON(trans->nr_updates >= trans->nr_paths);
	EBUG_ON(!bpos_eq(k->k.p, path->pos));

	n = (struct btree_insert_entry) {
		.flags		= flags,
		.bkey_type	= __btree_node_type(path->level, path->btree_id),
		.btree_id	= path->btree_id,
		.level		= path->level,
		.cached		= path->cached,
		.path		= path_idx,
		.k		= k,
		.ip_allocated	= ip,
	};

#ifdef CONFIG_BCACHEFS_DEBUG
	trans_for_each_update(trans, i)
		BUG_ON(i != trans->updates &&
		       btree_insert_entry_cmp(i - 1, i) >= 0);
#endif

	/*
	 * Pending updates are kept sorted: first, find position of new update,
	 * then delete/trim any updates the new update overwrites:
	 */
	for (i = trans->updates; i < trans->updates + trans->nr_updates; i++) {
		cmp = btree_insert_entry_cmp(&n, i);
		if (cmp <= 0)
			break;
	}

	if (!cmp && i < trans->updates + trans->nr_updates) {
		EBUG_ON(i->insert_trigger_run || i->overwrite_trigger_run);

		bch2_path_put(trans, i->path, true);
		i->flags	= n.flags;
		i->cached	= n.cached;
		i->k		= n.k;
		i->path		= n.path;
		i->ip_allocated	= n.ip_allocated;
	} else {
		array_insert_item(trans->updates, trans->nr_updates,
				  i - trans->updates, n);

		i->old_v = bch2_btree_path_peek_slot_exact(path, &i->old_k).v;
		i->old_btree_u64s = !bkey_deleted(&i->old_k) ? i->old_k.u64s : 0;

		if (unlikely(trans->journal_replay_not_finished)) {
			struct bkey_i *j_k =
				bch2_journal_keys_peek_slot(c, n.btree_id, n.level, k->k.p);

			if (j_k) {
				i->old_k = j_k->k;
				i->old_v = &j_k->v;
			}
		}
	}

	__btree_path_get(trans->paths + i->path, true);

	/*
	 * If a key is present in the key cache, it must also exist in the
	 * btree - this is necessary for cache coherency. When iterating over
	 * a btree that's cached in the key cache, the btree iter code checks
	 * the key cache - but the key has to exist in the btree for that to
	 * work:
	 */
	if (path->cached && !i->old_btree_u64s)
		return flush_new_cached_update(trans, i, flags, ip);

	return 0;
}

static noinline int bch2_trans_update_get_key_cache(struct btree_trans *trans,
						    struct btree_iter *iter,
						    struct btree_path *path)
{
	struct btree_path *key_cache_path = btree_iter_key_cache_path(trans, iter);

	if (!key_cache_path ||
	    !key_cache_path->should_be_locked ||
	    !bpos_eq(key_cache_path->pos, iter->pos)) {
		struct bkey_cached *ck;
		int ret;

		if (!iter->key_cache_path)
			iter->key_cache_path =
				bch2_path_get(trans, path->btree_id, path->pos, 1, 0,
					      BTREE_ITER_INTENT|
					      BTREE_ITER_CACHED, _THIS_IP_);

		iter->key_cache_path =
			bch2_btree_path_set_pos(trans, iter->key_cache_path, path->pos,
						iter->flags & BTREE_ITER_INTENT,
						_THIS_IP_);

		ret = bch2_btree_path_traverse(trans, iter->key_cache_path, BTREE_ITER_CACHED);
		if (unlikely(ret))
			return ret;

		ck = (void *) trans->paths[iter->key_cache_path].l[0].b;

		if (test_bit(BKEY_CACHED_DIRTY, &ck->flags)) {
			trace_and_count(trans->c, trans_restart_key_cache_raced, trans, _RET_IP_);
			return btree_trans_restart(trans, BCH_ERR_transaction_restart_key_cache_raced);
		}

		btree_path_set_should_be_locked(trans->paths + iter->key_cache_path);
	}

	return 0;
}

int __must_check bch2_trans_update(struct btree_trans *trans, struct btree_iter *iter,
				   struct bkey_i *k, enum btree_update_flags flags)
{
	btree_path_idx_t path_idx = iter->update_path ?: iter->path;
	int ret;

	if (iter->flags & BTREE_ITER_IS_EXTENTS)
		return bch2_trans_update_extent(trans, iter, k, flags);

	if (bkey_deleted(&k->k) &&
	    !(flags & BTREE_UPDATE_KEY_CACHE_RECLAIM) &&
	    (iter->flags & BTREE_ITER_FILTER_SNAPSHOTS)) {
		ret = need_whiteout_for_snapshot(trans, iter->btree_id, k->k.p);
		if (unlikely(ret < 0))
			return ret;

		if (ret)
			k->k.type = KEY_TYPE_whiteout;
	}

	/*
	 * Ensure that updates to cached btrees go to the key cache:
	 */
	struct btree_path *path = trans->paths + path_idx;
	if (!(flags & BTREE_UPDATE_KEY_CACHE_RECLAIM) &&
	    !path->cached &&
	    !path->level &&
	    btree_id_cached(trans->c, path->btree_id)) {
		ret = bch2_trans_update_get_key_cache(trans, iter, path);
		if (ret)
			return ret;

		path_idx = iter->key_cache_path;
	}

	return bch2_trans_update_by_path(trans, path_idx, k, flags, _RET_IP_);
}

int bch2_btree_insert_clone_trans(struct btree_trans *trans,
				  enum btree_id btree,
				  struct bkey_i *k)
{
	struct bkey_i *n = bch2_trans_kmalloc(trans, bkey_bytes(&k->k));
	int ret = PTR_ERR_OR_ZERO(n);
	if (ret)
		return ret;

	bkey_copy(n, k);
	return bch2_btree_insert_trans(trans, btree, n, 0);
}

struct jset_entry *__bch2_trans_jset_entry_alloc(struct btree_trans *trans, unsigned u64s)
{
	unsigned new_top = trans->journal_entries_u64s + u64s;
	unsigned old_size = trans->journal_entries_size;

	if (new_top > trans->journal_entries_size) {
		trans->journal_entries_size = roundup_pow_of_two(new_top);

		btree_trans_stats(trans)->journal_entries_size = trans->journal_entries_size;
	}

	struct jset_entry *n =
		bch2_trans_kmalloc_nomemzero(trans,
				trans->journal_entries_size * sizeof(u64));
	if (IS_ERR(n))
		return ERR_CAST(n);

	if (trans->journal_entries)
		memcpy(n, trans->journal_entries, old_size * sizeof(u64));
	trans->journal_entries = n;

	struct jset_entry *e = btree_trans_journal_entries_top(trans);
	trans->journal_entries_u64s = new_top;
	return e;
}

int bch2_bkey_get_empty_slot(struct btree_trans *trans, struct btree_iter *iter,
			     enum btree_id btree, struct bpos end)
{
	struct bkey_s_c k;
	int ret = 0;

	bch2_trans_iter_init(trans, iter, btree, POS_MAX, BTREE_ITER_INTENT);
	k = bch2_btree_iter_prev(iter);
	ret = bkey_err(k);
	if (ret)
		goto err;

	bch2_btree_iter_advance(iter);
	k = bch2_btree_iter_peek_slot(iter);
	ret = bkey_err(k);
	if (ret)
		goto err;

	BUG_ON(k.k->type != KEY_TYPE_deleted);

	if (bkey_gt(k.k->p, end)) {
		ret = -BCH_ERR_ENOSPC_btree_slot;
		goto err;
	}

	return 0;
err:
	bch2_trans_iter_exit(trans, iter);
	return ret;
}

void bch2_trans_commit_hook(struct btree_trans *trans,
			    struct btree_trans_commit_hook *h)
{
	h->next = trans->hooks;
	trans->hooks = h;
}

int bch2_btree_insert_nonextent(struct btree_trans *trans,
				enum btree_id btree, struct bkey_i *k,
				enum btree_update_flags flags)
{
	struct btree_iter iter;
	int ret;

	bch2_trans_iter_init(trans, &iter, btree, k->k.p,
			     BTREE_ITER_CACHED|
			     BTREE_ITER_NOT_EXTENTS|
			     BTREE_ITER_INTENT);
	ret   = bch2_btree_iter_traverse(&iter) ?:
		bch2_trans_update(trans, &iter, k, flags);
	bch2_trans_iter_exit(trans, &iter);
	return ret;
}

int bch2_btree_insert_trans(struct btree_trans *trans, enum btree_id id,
			    struct bkey_i *k, enum btree_update_flags flags)
{
	struct btree_iter iter;
	int ret;

	bch2_trans_iter_init(trans, &iter, id, bkey_start_pos(&k->k),
			     BTREE_ITER_CACHED|
			     BTREE_ITER_INTENT);
	ret   = bch2_btree_iter_traverse(&iter) ?:
		bch2_trans_update(trans, &iter, k, flags);
	bch2_trans_iter_exit(trans, &iter);
	return ret;
}

/**
 * bch2_btree_insert - insert keys into the extent btree
 * @c:			pointer to struct bch_fs
 * @id:			btree to insert into
 * @k:			key to insert
 * @disk_res:		must be non-NULL whenever inserting or potentially
 *			splitting data extents
 * @flags:		transaction commit flags
 *
 * Returns:		0 on success, error code on failure
 */
int bch2_btree_insert(struct bch_fs *c, enum btree_id id, struct bkey_i *k,
		      struct disk_reservation *disk_res, int flags)
{
	return bch2_trans_do(c, disk_res, NULL, flags,
			     bch2_btree_insert_trans(trans, id, k, 0));
}

int bch2_btree_delete_extent_at(struct btree_trans *trans, struct btree_iter *iter,
				unsigned len, unsigned update_flags)
{
	struct bkey_i *k;

	k = bch2_trans_kmalloc(trans, sizeof(*k));
	if (IS_ERR(k))
		return PTR_ERR(k);

	bkey_init(&k->k);
	k->k.p = iter->pos;
	bch2_key_resize(&k->k, len);
	return bch2_trans_update(trans, iter, k, update_flags);
}

int bch2_btree_delete_at(struct btree_trans *trans,
			 struct btree_iter *iter, unsigned update_flags)
{
	return bch2_btree_delete_extent_at(trans, iter, 0, update_flags);
}

int bch2_btree_delete(struct btree_trans *trans,
		      enum btree_id btree, struct bpos pos,
		      unsigned update_flags)
{
	struct btree_iter iter;
	int ret;

	bch2_trans_iter_init(trans, &iter, btree, pos,
			     BTREE_ITER_CACHED|
			     BTREE_ITER_INTENT);
	ret   = bch2_btree_iter_traverse(&iter) ?:
		bch2_btree_delete_at(trans, &iter, update_flags);
	bch2_trans_iter_exit(trans, &iter);

	return ret;
}

int bch2_btree_delete_range_trans(struct btree_trans *trans, enum btree_id id,
				  struct bpos start, struct bpos end,
				  unsigned update_flags,
				  u64 *journal_seq)
{
	u32 restart_count = trans->restart_count;
	struct btree_iter iter;
	struct bkey_s_c k;
	int ret = 0;

	bch2_trans_iter_init(trans, &iter, id, start, BTREE_ITER_INTENT);
	while ((k = bch2_btree_iter_peek_upto(&iter, end)).k) {
		struct disk_reservation disk_res =
			bch2_disk_reservation_init(trans->c, 0);
		struct bkey_i delete;

		ret = bkey_err(k);
		if (ret)
			goto err;

		bkey_init(&delete.k);

		/*
		 * This could probably be more efficient for extents:
		 */

		/*
		 * For extents, iter.pos won't necessarily be the same as
		 * bkey_start_pos(k.k) (for non extents they always will be the
		 * same). It's important that we delete starting from iter.pos
		 * because the range we want to delete could start in the middle
		 * of k.
		 *
		 * (bch2_btree_iter_peek() does guarantee that iter.pos >=
		 * bkey_start_pos(k.k)).
		 */
		delete.k.p = iter.pos;

		if (iter.flags & BTREE_ITER_IS_EXTENTS)
			bch2_key_resize(&delete.k,
					bpos_min(end, k.k->p).offset -
					iter.pos.offset);

		ret   = bch2_trans_update(trans, &iter, &delete, update_flags) ?:
			bch2_trans_commit(trans, &disk_res, journal_seq,
					  BCH_TRANS_COMMIT_no_enospc);
		bch2_disk_reservation_put(trans->c, &disk_res);
err:
		/*
		 * the bch2_trans_begin() call is in a weird place because we
		 * need to call it after every transaction commit, to avoid path
		 * overflow, but don't want to call it if the delete operation
		 * is a no-op and we have no work to do:
		 */
		bch2_trans_begin(trans);

		if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
			ret = 0;
		if (ret)
			break;
	}
	bch2_trans_iter_exit(trans, &iter);

	return ret ?: trans_was_restarted(trans, restart_count);
}

/*
 * bch_btree_delete_range - delete everything within a given range
 *
 * Range is a half open interval - [start, end)
 */
int bch2_btree_delete_range(struct bch_fs *c, enum btree_id id,
			    struct bpos start, struct bpos end,
			    unsigned update_flags,
			    u64 *journal_seq)
{
	int ret = bch2_trans_run(c,
			bch2_btree_delete_range_trans(trans, id, start, end,
						      update_flags, journal_seq));
	if (ret == -BCH_ERR_transaction_restart_nested)
		ret = 0;
	return ret;
}

int bch2_btree_bit_mod(struct btree_trans *trans, enum btree_id btree,
		       struct bpos pos, bool set)
{
	struct bkey_i *k = bch2_trans_kmalloc(trans, sizeof(*k));
	int ret = PTR_ERR_OR_ZERO(k);
	if (ret)
		return ret;

	bkey_init(&k->k);
	k->k.type = set ? KEY_TYPE_set : KEY_TYPE_deleted;
	k->k.p = pos;

	struct btree_iter iter;
	bch2_trans_iter_init(trans, &iter, btree, pos, BTREE_ITER_INTENT);

	ret   = bch2_btree_iter_traverse(&iter) ?:
		bch2_trans_update(trans, &iter, k, 0);
	bch2_trans_iter_exit(trans, &iter);
	return ret;
}

int bch2_btree_bit_mod_buffered(struct btree_trans *trans, enum btree_id btree,
				struct bpos pos, bool set)
{
	struct bkey_i k;

	bkey_init(&k.k);
	k.k.type = set ? KEY_TYPE_set : KEY_TYPE_deleted;
	k.k.p = pos;

	return bch2_trans_update_buffered(trans, btree, &k);
}

static int __bch2_trans_log_msg(struct btree_trans *trans, struct printbuf *buf, unsigned u64s)
{
	struct jset_entry *e = bch2_trans_jset_entry_alloc(trans, jset_u64s(u64s));
	int ret = PTR_ERR_OR_ZERO(e);
	if (ret)
		return ret;

	struct jset_entry_log *l = container_of(e, struct jset_entry_log, entry);
	journal_entry_init(e, BCH_JSET_ENTRY_log, 0, 1, u64s);
	memcpy(l->d, buf->buf, buf->pos);
	return 0;
}

__printf(3, 0)
static int
__bch2_fs_log_msg(struct bch_fs *c, unsigned commit_flags, const char *fmt,
		  va_list args)
{
	struct printbuf buf = PRINTBUF;
	prt_vprintf(&buf, fmt, args);

	unsigned u64s = DIV_ROUND_UP(buf.pos, sizeof(u64));
	prt_chars(&buf, '\0', u64s * sizeof(u64) - buf.pos);

	int ret = buf.allocation_failure ? -BCH_ERR_ENOMEM_trans_log_msg : 0;
	if (ret)
		goto err;

	if (!test_bit(JOURNAL_STARTED, &c->journal.flags)) {
		ret = darray_make_room(&c->journal.early_journal_entries, jset_u64s(u64s));
		if (ret)
			goto err;

		struct jset_entry_log *l = (void *) &darray_top(c->journal.early_journal_entries);
		journal_entry_init(&l->entry, BCH_JSET_ENTRY_log, 0, 1, u64s);
		memcpy(l->d, buf.buf, buf.pos);
		c->journal.early_journal_entries.nr += jset_u64s(u64s);
	} else {
		ret = bch2_trans_do(c, NULL, NULL,
			BCH_TRANS_COMMIT_lazy_rw|commit_flags,
			__bch2_trans_log_msg(trans, &buf, u64s));
	}
err:
	printbuf_exit(&buf);
	return ret;
}

__printf(2, 3)
int bch2_fs_log_msg(struct bch_fs *c, const char *fmt, ...)
{
	va_list args;
	int ret;

	va_start(args, fmt);
	ret = __bch2_fs_log_msg(c, 0, fmt, args);
	va_end(args);
	return ret;
}

/*
 * Use for logging messages during recovery to enable reserved space and avoid
 * blocking.
 */
__printf(2, 3)
int bch2_journal_log_msg(struct bch_fs *c, const char *fmt, ...)
{
	va_list args;
	int ret;

	va_start(args, fmt);
	ret = __bch2_fs_log_msg(c, BCH_WATERMARK_reclaim, fmt, args);
	va_end(args);
	return ret;
}