summaryrefslogtreecommitdiff
path: root/fs/bcachefs/fs-io-pagecache.c
blob: ff664fd0d8ef80e8b4816d7c430e87d41759b498 (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
// SPDX-License-Identifier: GPL-2.0
#ifndef NO_BCACHEFS_FS

#include "bcachefs.h"
#include "btree_iter.h"
#include "extents.h"
#include "fs-io.h"
#include "fs-io-pagecache.h"
#include "subvolume.h"

#include <linux/pagevec.h>
#include <linux/writeback.h>

int bch2_filemap_get_contig_folios_d(struct address_space *mapping,
				     loff_t start, u64 end,
				     fgf_t fgp_flags, gfp_t gfp,
				     folios *fs)
{
	struct folio *f;
	u64 pos = start;
	int ret = 0;

	while (pos < end) {
		if ((u64) pos >= (u64) start + (1ULL << 20))
			fgp_flags &= ~FGP_CREAT;

		ret = darray_make_room_gfp(fs, 1, gfp & GFP_KERNEL);
		if (ret)
			break;

		f = __filemap_get_folio(mapping, pos >> PAGE_SHIFT, fgp_flags, gfp);
		if (IS_ERR_OR_NULL(f))
			break;

		BUG_ON(fs->nr && folio_pos(f) != pos);

		pos = folio_end_pos(f);
		darray_push(fs, f);
	}

	if (!fs->nr && !ret && (fgp_flags & FGP_CREAT))
		ret = -ENOMEM;

	return fs->nr ? 0 : ret;
}

/* pagecache_block must be held */
int bch2_write_invalidate_inode_pages_range(struct address_space *mapping,
					    loff_t start, loff_t end)
{
	int ret;

	/*
	 * XXX: the way this is currently implemented, we can spin if a process
	 * is continually redirtying a specific page
	 */
	do {
		if (!mapping->nrpages)
			return 0;

		ret = filemap_write_and_wait_range(mapping, start, end);
		if (ret)
			break;

		if (!mapping->nrpages)
			return 0;

		ret = invalidate_inode_pages2_range(mapping,
				start >> PAGE_SHIFT,
				end >> PAGE_SHIFT);
	} while (ret == -EBUSY);

	return ret;
}

#if 0
/* Useful for debug tracing: */
static const char * const bch2_folio_sector_states[] = {
#define x(n)	#n,
	BCH_FOLIO_SECTOR_STATE()
#undef x
	NULL
};
#endif

static inline enum bch_folio_sector_state
folio_sector_dirty(enum bch_folio_sector_state state)
{
	switch (state) {
	case SECTOR_unallocated:
		return SECTOR_dirty;
	case SECTOR_reserved:
		return SECTOR_dirty_reserved;
	default:
		return state;
	}
}

static inline enum bch_folio_sector_state
folio_sector_undirty(enum bch_folio_sector_state state)
{
	switch (state) {
	case SECTOR_dirty:
		return SECTOR_unallocated;
	case SECTOR_dirty_reserved:
		return SECTOR_reserved;
	default:
		return state;
	}
}

static inline enum bch_folio_sector_state
folio_sector_reserve(enum bch_folio_sector_state state)
{
	switch (state) {
	case SECTOR_unallocated:
		return SECTOR_reserved;
	case SECTOR_dirty:
		return SECTOR_dirty_reserved;
	default:
		return state;
	}
}

/* for newly allocated folios: */
struct bch_folio *__bch2_folio_create(struct folio *folio, gfp_t gfp)
{
	struct bch_folio *s;

	s = kzalloc(sizeof(*s) +
		    sizeof(struct bch_folio_sector) *
		    folio_sectors(folio), gfp);
	if (!s)
		return NULL;

	spin_lock_init(&s->lock);
	folio_attach_private(folio, s);
	return s;
}

struct bch_folio *bch2_folio_create(struct folio *folio, gfp_t gfp)
{
	return bch2_folio(folio) ?: __bch2_folio_create(folio, gfp);
}

static unsigned bkey_to_sector_state(struct bkey_s_c k)
{
	if (bkey_extent_is_reservation(k))
		return SECTOR_reserved;
	if (bkey_extent_is_allocation(k.k))
		return SECTOR_allocated;
	return SECTOR_unallocated;
}

static void __bch2_folio_set(struct folio *folio,
			     unsigned pg_offset, unsigned pg_len,
			     unsigned nr_ptrs, unsigned state)
{
	struct bch_folio *s = bch2_folio(folio);
	unsigned i, sectors = folio_sectors(folio);

	BUG_ON(pg_offset >= sectors);
	BUG_ON(pg_offset + pg_len > sectors);

	spin_lock(&s->lock);

	for (i = pg_offset; i < pg_offset + pg_len; i++) {
		s->s[i].nr_replicas	= nr_ptrs;
		bch2_folio_sector_set(folio, s, i, state);
	}

	if (i == sectors)
		s->uptodate = true;

	spin_unlock(&s->lock);
}

/*
 * Initialize bch_folio state (allocated/unallocated, nr_replicas) from the
 * extents btree:
 */
int bch2_folio_set(struct bch_fs *c, subvol_inum inum,
		   struct folio **fs, unsigned nr_folios)
{
	struct btree_trans *trans;
	struct btree_iter iter;
	struct bkey_s_c k;
	struct bch_folio *s;
	u64 offset = folio_sector(fs[0]);
	unsigned folio_idx;
	u32 snapshot;
	bool need_set = false;
	int ret;

	for (folio_idx = 0; folio_idx < nr_folios; folio_idx++) {
		s = bch2_folio_create(fs[folio_idx], GFP_KERNEL);
		if (!s)
			return -ENOMEM;

		need_set |= !s->uptodate;
	}

	if (!need_set)
		return 0;

	folio_idx = 0;
	trans = bch2_trans_get(c);
retry:
	bch2_trans_begin(trans);

	ret = bch2_subvolume_get_snapshot(trans, inum.subvol, &snapshot);
	if (ret)
		goto err;

	for_each_btree_key_norestart(trans, iter, BTREE_ID_extents,
			   SPOS(inum.inum, offset, snapshot),
			   BTREE_ITER_SLOTS, k, ret) {
		unsigned nr_ptrs = bch2_bkey_nr_ptrs_fully_allocated(k);
		unsigned state = bkey_to_sector_state(k);

		while (folio_idx < nr_folios) {
			struct folio *folio = fs[folio_idx];
			u64 folio_start	= folio_sector(folio);
			u64 folio_end	= folio_end_sector(folio);
			unsigned folio_offset = max(bkey_start_offset(k.k), folio_start) -
				folio_start;
			unsigned folio_len = min(k.k->p.offset, folio_end) -
				folio_offset - folio_start;

			BUG_ON(k.k->p.offset < folio_start);
			BUG_ON(bkey_start_offset(k.k) > folio_end);

			if (!bch2_folio(folio)->uptodate)
				__bch2_folio_set(folio, folio_offset, folio_len, nr_ptrs, state);

			if (k.k->p.offset < folio_end)
				break;
			folio_idx++;
		}

		if (folio_idx == nr_folios)
			break;
	}

	offset = iter.pos.offset;
	bch2_trans_iter_exit(trans, &iter);
err:
	if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
		goto retry;
	bch2_trans_put(trans);

	return ret;
}

void bch2_bio_page_state_set(struct bio *bio, struct bkey_s_c k)
{
	struct bvec_iter iter;
	struct folio_vec fv;
	unsigned nr_ptrs = k.k->type == KEY_TYPE_reflink_v
		? 0 : bch2_bkey_nr_ptrs_fully_allocated(k);
	unsigned state = bkey_to_sector_state(k);

	bio_for_each_folio(fv, bio, iter)
		__bch2_folio_set(fv.fv_folio,
				 fv.fv_offset >> 9,
				 fv.fv_len >> 9,
				 nr_ptrs, state);
}

void bch2_mark_pagecache_unallocated(struct bch_inode_info *inode,
				     u64 start, u64 end)
{
	pgoff_t index = start >> PAGE_SECTORS_SHIFT;
	pgoff_t end_index = (end - 1) >> PAGE_SECTORS_SHIFT;
	struct folio_batch fbatch;
	unsigned i, j;

	if (end <= start)
		return;

	folio_batch_init(&fbatch);

	while (filemap_get_folios(inode->v.i_mapping,
				  &index, end_index, &fbatch)) {
		for (i = 0; i < folio_batch_count(&fbatch); i++) {
			struct folio *folio = fbatch.folios[i];
			u64 folio_start = folio_sector(folio);
			u64 folio_end = folio_end_sector(folio);
			unsigned folio_offset = max(start, folio_start) - folio_start;
			unsigned folio_len = min(end, folio_end) - folio_offset - folio_start;
			struct bch_folio *s;

			BUG_ON(end <= folio_start);

			folio_lock(folio);
			s = bch2_folio(folio);

			if (s) {
				spin_lock(&s->lock);
				for (j = folio_offset; j < folio_offset + folio_len; j++)
					s->s[j].nr_replicas = 0;
				spin_unlock(&s->lock);
			}

			folio_unlock(folio);
		}
		folio_batch_release(&fbatch);
		cond_resched();
	}
}

void bch2_mark_pagecache_reserved(struct bch_inode_info *inode,
				  u64 start, u64 end)
{
	struct bch_fs *c = inode->v.i_sb->s_fs_info;
	pgoff_t index = start >> PAGE_SECTORS_SHIFT;
	pgoff_t end_index = (end - 1) >> PAGE_SECTORS_SHIFT;
	struct folio_batch fbatch;
	s64 i_sectors_delta = 0;
	unsigned i, j;

	if (end <= start)
		return;

	folio_batch_init(&fbatch);

	while (filemap_get_folios(inode->v.i_mapping,
				  &index, end_index, &fbatch)) {
		for (i = 0; i < folio_batch_count(&fbatch); i++) {
			struct folio *folio = fbatch.folios[i];
			u64 folio_start = folio_sector(folio);
			u64 folio_end = folio_end_sector(folio);
			unsigned folio_offset = max(start, folio_start) - folio_start;
			unsigned folio_len = min(end, folio_end) - folio_offset - folio_start;
			struct bch_folio *s;

			BUG_ON(end <= folio_start);

			folio_lock(folio);
			s = bch2_folio(folio);

			if (s) {
				spin_lock(&s->lock);
				for (j = folio_offset; j < folio_offset + folio_len; j++) {
					i_sectors_delta -= s->s[j].state == SECTOR_dirty;
					bch2_folio_sector_set(folio, s, j,
						folio_sector_reserve(s->s[j].state));
				}
				spin_unlock(&s->lock);
			}

			folio_unlock(folio);
		}
		folio_batch_release(&fbatch);
		cond_resched();
	}

	bch2_i_sectors_acct(c, inode, NULL, i_sectors_delta);
}

static inline unsigned sectors_to_reserve(struct bch_folio_sector *s,
					  unsigned nr_replicas)
{
	return max(0, (int) nr_replicas -
		   s->nr_replicas -
		   s->replicas_reserved);
}

int bch2_get_folio_disk_reservation(struct bch_fs *c,
				struct bch_inode_info *inode,
				struct folio *folio, bool check_enospc)
{
	struct bch_folio *s = bch2_folio_create(folio, 0);
	unsigned nr_replicas = inode_nr_replicas(c, inode);
	struct disk_reservation disk_res = { 0 };
	unsigned i, sectors = folio_sectors(folio), disk_res_sectors = 0;
	int ret;

	if (!s)
		return -ENOMEM;

	for (i = 0; i < sectors; i++)
		disk_res_sectors += sectors_to_reserve(&s->s[i], nr_replicas);

	if (!disk_res_sectors)
		return 0;

	ret = bch2_disk_reservation_get(c, &disk_res,
					disk_res_sectors, 1,
					!check_enospc
					? BCH_DISK_RESERVATION_NOFAIL
					: 0);
	if (unlikely(ret))
		return ret;

	for (i = 0; i < sectors; i++)
		s->s[i].replicas_reserved +=
			sectors_to_reserve(&s->s[i], nr_replicas);

	return 0;
}

void bch2_folio_reservation_put(struct bch_fs *c,
			struct bch_inode_info *inode,
			struct bch2_folio_reservation *res)
{
	bch2_disk_reservation_put(c, &res->disk);
	bch2_quota_reservation_put(c, inode, &res->quota);
}

int bch2_folio_reservation_get(struct bch_fs *c,
			struct bch_inode_info *inode,
			struct folio *folio,
			struct bch2_folio_reservation *res,
			unsigned offset, unsigned len)
{
	struct bch_folio *s = bch2_folio_create(folio, 0);
	unsigned i, disk_sectors = 0, quota_sectors = 0;
	int ret;

	if (!s)
		return -ENOMEM;

	BUG_ON(!s->uptodate);

	for (i = round_down(offset, block_bytes(c)) >> 9;
	     i < round_up(offset + len, block_bytes(c)) >> 9;
	     i++) {
		disk_sectors += sectors_to_reserve(&s->s[i],
						res->disk.nr_replicas);
		quota_sectors += s->s[i].state == SECTOR_unallocated;
	}

	if (disk_sectors) {
		ret = bch2_disk_reservation_add(c, &res->disk, disk_sectors, 0);
		if (unlikely(ret))
			return ret;
	}

	if (quota_sectors) {
		ret = bch2_quota_reservation_add(c, inode, &res->quota,
						 quota_sectors, true);
		if (unlikely(ret)) {
			struct disk_reservation tmp = {
				.sectors = disk_sectors
			};

			bch2_disk_reservation_put(c, &tmp);
			res->disk.sectors -= disk_sectors;
			return ret;
		}
	}

	return 0;
}

static void bch2_clear_folio_bits(struct folio *folio)
{
	struct bch_inode_info *inode = to_bch_ei(folio->mapping->host);
	struct bch_fs *c = inode->v.i_sb->s_fs_info;
	struct bch_folio *s = bch2_folio(folio);
	struct disk_reservation disk_res = { 0 };
	int i, sectors = folio_sectors(folio), dirty_sectors = 0;

	if (!s)
		return;

	EBUG_ON(!folio_test_locked(folio));
	EBUG_ON(folio_test_writeback(folio));

	for (i = 0; i < sectors; i++) {
		disk_res.sectors += s->s[i].replicas_reserved;
		s->s[i].replicas_reserved = 0;

		dirty_sectors -= s->s[i].state == SECTOR_dirty;
		bch2_folio_sector_set(folio, s, i, folio_sector_undirty(s->s[i].state));
	}

	bch2_disk_reservation_put(c, &disk_res);

	bch2_i_sectors_acct(c, inode, NULL, dirty_sectors);

	bch2_folio_release(folio);
}

void bch2_set_folio_dirty(struct bch_fs *c,
			  struct bch_inode_info *inode,
			  struct folio *folio,
			  struct bch2_folio_reservation *res,
			  unsigned offset, unsigned len)
{
	struct bch_folio *s = bch2_folio(folio);
	unsigned i, dirty_sectors = 0;

	WARN_ON((u64) folio_pos(folio) + offset + len >
		round_up((u64) i_size_read(&inode->v), block_bytes(c)));

	BUG_ON(!s->uptodate);

	spin_lock(&s->lock);

	for (i = round_down(offset, block_bytes(c)) >> 9;
	     i < round_up(offset + len, block_bytes(c)) >> 9;
	     i++) {
		unsigned sectors = sectors_to_reserve(&s->s[i],
						res->disk.nr_replicas);

		/*
		 * This can happen if we race with the error path in
		 * bch2_writepage_io_done():
		 */
		sectors = min_t(unsigned, sectors, res->disk.sectors);

		s->s[i].replicas_reserved += sectors;
		res->disk.sectors -= sectors;

		dirty_sectors += s->s[i].state == SECTOR_unallocated;

		bch2_folio_sector_set(folio, s, i, folio_sector_dirty(s->s[i].state));
	}

	spin_unlock(&s->lock);

	bch2_i_sectors_acct(c, inode, &res->quota, dirty_sectors);

	if (!folio_test_dirty(folio))
		filemap_dirty_folio(inode->v.i_mapping, folio);
}

vm_fault_t bch2_page_fault(struct vm_fault *vmf)
{
	struct file *file = vmf->vma->vm_file;
	struct address_space *mapping = file->f_mapping;
	struct address_space *fdm = faults_disabled_mapping();
	struct bch_inode_info *inode = file_bch_inode(file);
	vm_fault_t ret;

	if (fdm == mapping)
		return VM_FAULT_SIGBUS;

	/* Lock ordering: */
	if (fdm > mapping) {
		struct bch_inode_info *fdm_host = to_bch_ei(fdm->host);

		if (bch2_pagecache_add_tryget(inode))
			goto got_lock;

		bch2_pagecache_block_put(fdm_host);

		bch2_pagecache_add_get(inode);
		bch2_pagecache_add_put(inode);

		bch2_pagecache_block_get(fdm_host);

		/* Signal that lock has been dropped: */
		set_fdm_dropped_locks();
		return VM_FAULT_SIGBUS;
	}

	bch2_pagecache_add_get(inode);
got_lock:
	ret = filemap_fault(vmf);
	bch2_pagecache_add_put(inode);

	return ret;
}

vm_fault_t bch2_page_mkwrite(struct vm_fault *vmf)
{
	struct folio *folio = page_folio(vmf->page);
	struct file *file = vmf->vma->vm_file;
	struct bch_inode_info *inode = file_bch_inode(file);
	struct address_space *mapping = file->f_mapping;
	struct bch_fs *c = inode->v.i_sb->s_fs_info;
	struct bch2_folio_reservation res;
	unsigned len;
	loff_t isize;
	vm_fault_t ret;

	bch2_folio_reservation_init(c, inode, &res);

	sb_start_pagefault(inode->v.i_sb);
	file_update_time(file);

	/*
	 * Not strictly necessary, but helps avoid dio writes livelocking in
	 * bch2_write_invalidate_inode_pages_range() - can drop this if/when we get
	 * a bch2_write_invalidate_inode_pages_range() that works without dropping
	 * page lock before invalidating page
	 */
	bch2_pagecache_add_get(inode);

	folio_lock(folio);
	isize = i_size_read(&inode->v);

	if (folio->mapping != mapping || folio_pos(folio) >= isize) {
		folio_unlock(folio);
		ret = VM_FAULT_NOPAGE;
		goto out;
	}

	len = min_t(loff_t, folio_size(folio), isize - folio_pos(folio));

	if (bch2_folio_set(c, inode_inum(inode), &folio, 1) ?:
	    bch2_folio_reservation_get(c, inode, folio, &res, 0, len)) {
		folio_unlock(folio);
		ret = VM_FAULT_SIGBUS;
		goto out;
	}

	bch2_set_folio_dirty(c, inode, folio, &res, 0, len);
	bch2_folio_reservation_put(c, inode, &res);

	folio_wait_stable(folio);
	ret = VM_FAULT_LOCKED;
out:
	bch2_pagecache_add_put(inode);
	sb_end_pagefault(inode->v.i_sb);

	return ret;
}

void bch2_invalidate_folio(struct folio *folio, size_t offset, size_t length)
{
	if (offset || length < folio_size(folio))
		return;

	bch2_clear_folio_bits(folio);
}

bool bch2_release_folio(struct folio *folio, gfp_t gfp_mask)
{
	if (folio_test_dirty(folio) || folio_test_writeback(folio))
		return false;

	bch2_clear_folio_bits(folio);
	return true;
}

/* fseek: */

static int folio_data_offset(struct folio *folio, loff_t pos,
			     unsigned min_replicas)
{
	struct bch_folio *s = bch2_folio(folio);
	unsigned i, sectors = folio_sectors(folio);

	if (s)
		for (i = folio_pos_to_s(folio, pos); i < sectors; i++)
			if (s->s[i].state >= SECTOR_dirty &&
			    s->s[i].nr_replicas + s->s[i].replicas_reserved >= min_replicas)
				return i << SECTOR_SHIFT;

	return -1;
}

loff_t bch2_seek_pagecache_data(struct inode *vinode,
				loff_t start_offset,
				loff_t end_offset,
				unsigned min_replicas,
				bool nonblock)
{
	struct folio_batch fbatch;
	pgoff_t start_index	= start_offset >> PAGE_SHIFT;
	pgoff_t end_index	= end_offset >> PAGE_SHIFT;
	pgoff_t index		= start_index;
	unsigned i;
	loff_t ret;
	int offset;

	folio_batch_init(&fbatch);

	while (filemap_get_folios(vinode->i_mapping,
				  &index, end_index, &fbatch)) {
		for (i = 0; i < folio_batch_count(&fbatch); i++) {
			struct folio *folio = fbatch.folios[i];

			if (!nonblock) {
				folio_lock(folio);
			} else if (!folio_trylock(folio)) {
				folio_batch_release(&fbatch);
				return -EAGAIN;
			}

			offset = folio_data_offset(folio,
					max(folio_pos(folio), start_offset),
					min_replicas);
			if (offset >= 0) {
				ret = clamp(folio_pos(folio) + offset,
					    start_offset, end_offset);
				folio_unlock(folio);
				folio_batch_release(&fbatch);
				return ret;
			}
			folio_unlock(folio);
		}
		folio_batch_release(&fbatch);
		cond_resched();
	}

	return end_offset;
}

/*
 * Search for a hole in a folio.
 *
 * The filemap layer returns -ENOENT if no folio exists, so reuse the same error
 * code to indicate a pagecache hole exists at the returned offset. Otherwise
 * return 0 if the folio is filled with data, or an error code. This function
 * can return -EAGAIN if nonblock is specified.
 */
static int folio_hole_offset(struct address_space *mapping, loff_t *offset,
			      unsigned min_replicas, bool nonblock)
{
	struct folio *folio;
	struct bch_folio *s;
	unsigned i, sectors;
	int ret = -ENOENT;

	folio = __filemap_get_folio(mapping, *offset >> PAGE_SHIFT,
				    FGP_LOCK|(nonblock ? FGP_NOWAIT : 0), 0);
	if (IS_ERR(folio))
		return PTR_ERR(folio);

	s = bch2_folio(folio);
	if (!s)
		goto unlock;

	sectors = folio_sectors(folio);
	for (i = folio_pos_to_s(folio, *offset); i < sectors; i++)
		if (s->s[i].state < SECTOR_dirty ||
		    s->s[i].nr_replicas + s->s[i].replicas_reserved < min_replicas) {
			*offset = max(*offset,
				      folio_pos(folio) + (i << SECTOR_SHIFT));
			goto unlock;
		}

	*offset = folio_end_pos(folio);
	ret = 0;
unlock:
	folio_unlock(folio);
	folio_put(folio);
	return ret;
}

loff_t bch2_seek_pagecache_hole(struct inode *vinode,
				loff_t start_offset,
				loff_t end_offset,
				unsigned min_replicas,
				bool nonblock)
{
	struct address_space *mapping = vinode->i_mapping;
	loff_t offset = start_offset;
	loff_t ret = 0;

	while (!ret && offset < end_offset)
		ret = folio_hole_offset(mapping, &offset, min_replicas, nonblock);

	if (ret && ret != -ENOENT)
		return ret;
	return min(offset, end_offset);
}

int bch2_clamp_data_hole(struct inode *inode,
			 u64 *hole_start,
			 u64 *hole_end,
			 unsigned min_replicas,
			 bool nonblock)
{
	loff_t ret;

	ret = bch2_seek_pagecache_hole(inode,
		*hole_start << 9, *hole_end << 9, min_replicas, nonblock) >> 9;
	if (ret < 0)
		return ret;

	*hole_start = ret;

	if (*hole_start == *hole_end)
		return 0;

	ret = bch2_seek_pagecache_data(inode,
		*hole_start << 9, *hole_end << 9, min_replicas, nonblock) >> 9;
	if (ret < 0)
		return ret;

	*hole_end = ret;
	return 0;
}

#endif /* NO_BCACHEFS_FS */