Loading...
1// SPDX-License-Identifier: GPL-2.0
2/*
3 * fs/f2fs/segment.c
4 *
5 * Copyright (c) 2012 Samsung Electronics Co., Ltd.
6 * http://www.samsung.com/
7 */
8#include <linux/fs.h>
9#include <linux/f2fs_fs.h>
10#include <linux/bio.h>
11#include <linux/blkdev.h>
12#include <linux/sched/mm.h>
13#include <linux/prefetch.h>
14#include <linux/kthread.h>
15#include <linux/swap.h>
16#include <linux/timer.h>
17#include <linux/freezer.h>
18#include <linux/sched/signal.h>
19#include <linux/random.h>
20
21#include "f2fs.h"
22#include "segment.h"
23#include "node.h"
24#include "gc.h"
25#include "iostat.h"
26#include <trace/events/f2fs.h>
27
28#define __reverse_ffz(x) __reverse_ffs(~(x))
29
30static struct kmem_cache *discard_entry_slab;
31static struct kmem_cache *discard_cmd_slab;
32static struct kmem_cache *sit_entry_set_slab;
33static struct kmem_cache *revoke_entry_slab;
34
35static unsigned long __reverse_ulong(unsigned char *str)
36{
37 unsigned long tmp = 0;
38 int shift = 24, idx = 0;
39
40#if BITS_PER_LONG == 64
41 shift = 56;
42#endif
43 while (shift >= 0) {
44 tmp |= (unsigned long)str[idx++] << shift;
45 shift -= BITS_PER_BYTE;
46 }
47 return tmp;
48}
49
50/*
51 * __reverse_ffs is copied from include/asm-generic/bitops/__ffs.h since
52 * MSB and LSB are reversed in a byte by f2fs_set_bit.
53 */
54static inline unsigned long __reverse_ffs(unsigned long word)
55{
56 int num = 0;
57
58#if BITS_PER_LONG == 64
59 if ((word & 0xffffffff00000000UL) == 0)
60 num += 32;
61 else
62 word >>= 32;
63#endif
64 if ((word & 0xffff0000) == 0)
65 num += 16;
66 else
67 word >>= 16;
68
69 if ((word & 0xff00) == 0)
70 num += 8;
71 else
72 word >>= 8;
73
74 if ((word & 0xf0) == 0)
75 num += 4;
76 else
77 word >>= 4;
78
79 if ((word & 0xc) == 0)
80 num += 2;
81 else
82 word >>= 2;
83
84 if ((word & 0x2) == 0)
85 num += 1;
86 return num;
87}
88
89/*
90 * __find_rev_next(_zero)_bit is copied from lib/find_next_bit.c because
91 * f2fs_set_bit makes MSB and LSB reversed in a byte.
92 * @size must be integral times of unsigned long.
93 * Example:
94 * MSB <--> LSB
95 * f2fs_set_bit(0, bitmap) => 1000 0000
96 * f2fs_set_bit(7, bitmap) => 0000 0001
97 */
98static unsigned long __find_rev_next_bit(const unsigned long *addr,
99 unsigned long size, unsigned long offset)
100{
101 const unsigned long *p = addr + BIT_WORD(offset);
102 unsigned long result = size;
103 unsigned long tmp;
104
105 if (offset >= size)
106 return size;
107
108 size -= (offset & ~(BITS_PER_LONG - 1));
109 offset %= BITS_PER_LONG;
110
111 while (1) {
112 if (*p == 0)
113 goto pass;
114
115 tmp = __reverse_ulong((unsigned char *)p);
116
117 tmp &= ~0UL >> offset;
118 if (size < BITS_PER_LONG)
119 tmp &= (~0UL << (BITS_PER_LONG - size));
120 if (tmp)
121 goto found;
122pass:
123 if (size <= BITS_PER_LONG)
124 break;
125 size -= BITS_PER_LONG;
126 offset = 0;
127 p++;
128 }
129 return result;
130found:
131 return result - size + __reverse_ffs(tmp);
132}
133
134static unsigned long __find_rev_next_zero_bit(const unsigned long *addr,
135 unsigned long size, unsigned long offset)
136{
137 const unsigned long *p = addr + BIT_WORD(offset);
138 unsigned long result = size;
139 unsigned long tmp;
140
141 if (offset >= size)
142 return size;
143
144 size -= (offset & ~(BITS_PER_LONG - 1));
145 offset %= BITS_PER_LONG;
146
147 while (1) {
148 if (*p == ~0UL)
149 goto pass;
150
151 tmp = __reverse_ulong((unsigned char *)p);
152
153 if (offset)
154 tmp |= ~0UL << (BITS_PER_LONG - offset);
155 if (size < BITS_PER_LONG)
156 tmp |= ~0UL >> size;
157 if (tmp != ~0UL)
158 goto found;
159pass:
160 if (size <= BITS_PER_LONG)
161 break;
162 size -= BITS_PER_LONG;
163 offset = 0;
164 p++;
165 }
166 return result;
167found:
168 return result - size + __reverse_ffz(tmp);
169}
170
171bool f2fs_need_SSR(struct f2fs_sb_info *sbi)
172{
173 int node_secs = get_blocktype_secs(sbi, F2FS_DIRTY_NODES);
174 int dent_secs = get_blocktype_secs(sbi, F2FS_DIRTY_DENTS);
175 int imeta_secs = get_blocktype_secs(sbi, F2FS_DIRTY_IMETA);
176
177 if (f2fs_lfs_mode(sbi))
178 return false;
179 if (sbi->gc_mode == GC_URGENT_HIGH)
180 return true;
181 if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED)))
182 return true;
183
184 return free_sections(sbi) <= (node_secs + 2 * dent_secs + imeta_secs +
185 SM_I(sbi)->min_ssr_sections + reserved_sections(sbi));
186}
187
188void f2fs_abort_atomic_write(struct inode *inode, bool clean)
189{
190 struct f2fs_inode_info *fi = F2FS_I(inode);
191
192 if (!f2fs_is_atomic_file(inode))
193 return;
194
195 if (clean)
196 truncate_inode_pages_final(inode->i_mapping);
197
198 release_atomic_write_cnt(inode);
199 clear_inode_flag(inode, FI_ATOMIC_COMMITTED);
200 clear_inode_flag(inode, FI_ATOMIC_REPLACE);
201 clear_inode_flag(inode, FI_ATOMIC_FILE);
202 if (is_inode_flag_set(inode, FI_ATOMIC_DIRTIED)) {
203 clear_inode_flag(inode, FI_ATOMIC_DIRTIED);
204 f2fs_mark_inode_dirty_sync(inode, true);
205 }
206 stat_dec_atomic_inode(inode);
207
208 F2FS_I(inode)->atomic_write_task = NULL;
209
210 if (clean) {
211 f2fs_i_size_write(inode, fi->original_i_size);
212 fi->original_i_size = 0;
213 }
214 /* avoid stale dirty inode during eviction */
215 sync_inode_metadata(inode, 0);
216}
217
218static int __replace_atomic_write_block(struct inode *inode, pgoff_t index,
219 block_t new_addr, block_t *old_addr, bool recover)
220{
221 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
222 struct dnode_of_data dn;
223 struct node_info ni;
224 int err;
225
226retry:
227 set_new_dnode(&dn, inode, NULL, NULL, 0);
228 err = f2fs_get_dnode_of_data(&dn, index, ALLOC_NODE);
229 if (err) {
230 if (err == -ENOMEM) {
231 f2fs_io_schedule_timeout(DEFAULT_IO_TIMEOUT);
232 goto retry;
233 }
234 return err;
235 }
236
237 err = f2fs_get_node_info(sbi, dn.nid, &ni, false);
238 if (err) {
239 f2fs_put_dnode(&dn);
240 return err;
241 }
242
243 if (recover) {
244 /* dn.data_blkaddr is always valid */
245 if (!__is_valid_data_blkaddr(new_addr)) {
246 if (new_addr == NULL_ADDR)
247 dec_valid_block_count(sbi, inode, 1);
248 f2fs_invalidate_blocks(sbi, dn.data_blkaddr);
249 f2fs_update_data_blkaddr(&dn, new_addr);
250 } else {
251 f2fs_replace_block(sbi, &dn, dn.data_blkaddr,
252 new_addr, ni.version, true, true);
253 }
254 } else {
255 blkcnt_t count = 1;
256
257 err = inc_valid_block_count(sbi, inode, &count, true);
258 if (err) {
259 f2fs_put_dnode(&dn);
260 return err;
261 }
262
263 *old_addr = dn.data_blkaddr;
264 f2fs_truncate_data_blocks_range(&dn, 1);
265 dec_valid_block_count(sbi, F2FS_I(inode)->cow_inode, count);
266
267 f2fs_replace_block(sbi, &dn, dn.data_blkaddr, new_addr,
268 ni.version, true, false);
269 }
270
271 f2fs_put_dnode(&dn);
272
273 trace_f2fs_replace_atomic_write_block(inode, F2FS_I(inode)->cow_inode,
274 index, old_addr ? *old_addr : 0, new_addr, recover);
275 return 0;
276}
277
278static void __complete_revoke_list(struct inode *inode, struct list_head *head,
279 bool revoke)
280{
281 struct revoke_entry *cur, *tmp;
282 pgoff_t start_index = 0;
283 bool truncate = is_inode_flag_set(inode, FI_ATOMIC_REPLACE);
284
285 list_for_each_entry_safe(cur, tmp, head, list) {
286 if (revoke) {
287 __replace_atomic_write_block(inode, cur->index,
288 cur->old_addr, NULL, true);
289 } else if (truncate) {
290 f2fs_truncate_hole(inode, start_index, cur->index);
291 start_index = cur->index + 1;
292 }
293
294 list_del(&cur->list);
295 kmem_cache_free(revoke_entry_slab, cur);
296 }
297
298 if (!revoke && truncate)
299 f2fs_do_truncate_blocks(inode, start_index * PAGE_SIZE, false);
300}
301
302static int __f2fs_commit_atomic_write(struct inode *inode)
303{
304 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
305 struct f2fs_inode_info *fi = F2FS_I(inode);
306 struct inode *cow_inode = fi->cow_inode;
307 struct revoke_entry *new;
308 struct list_head revoke_list;
309 block_t blkaddr;
310 struct dnode_of_data dn;
311 pgoff_t len = DIV_ROUND_UP(i_size_read(inode), PAGE_SIZE);
312 pgoff_t off = 0, blen, index;
313 int ret = 0, i;
314
315 INIT_LIST_HEAD(&revoke_list);
316
317 while (len) {
318 blen = min_t(pgoff_t, ADDRS_PER_BLOCK(cow_inode), len);
319
320 set_new_dnode(&dn, cow_inode, NULL, NULL, 0);
321 ret = f2fs_get_dnode_of_data(&dn, off, LOOKUP_NODE_RA);
322 if (ret && ret != -ENOENT) {
323 goto out;
324 } else if (ret == -ENOENT) {
325 ret = 0;
326 if (dn.max_level == 0)
327 goto out;
328 goto next;
329 }
330
331 blen = min((pgoff_t)ADDRS_PER_PAGE(dn.node_page, cow_inode),
332 len);
333 index = off;
334 for (i = 0; i < blen; i++, dn.ofs_in_node++, index++) {
335 blkaddr = f2fs_data_blkaddr(&dn);
336
337 if (!__is_valid_data_blkaddr(blkaddr)) {
338 continue;
339 } else if (!f2fs_is_valid_blkaddr(sbi, blkaddr,
340 DATA_GENERIC_ENHANCE)) {
341 f2fs_put_dnode(&dn);
342 ret = -EFSCORRUPTED;
343 goto out;
344 }
345
346 new = f2fs_kmem_cache_alloc(revoke_entry_slab, GFP_NOFS,
347 true, NULL);
348
349 ret = __replace_atomic_write_block(inode, index, blkaddr,
350 &new->old_addr, false);
351 if (ret) {
352 f2fs_put_dnode(&dn);
353 kmem_cache_free(revoke_entry_slab, new);
354 goto out;
355 }
356
357 f2fs_update_data_blkaddr(&dn, NULL_ADDR);
358 new->index = index;
359 list_add_tail(&new->list, &revoke_list);
360 }
361 f2fs_put_dnode(&dn);
362next:
363 off += blen;
364 len -= blen;
365 }
366
367out:
368 if (ret) {
369 sbi->revoked_atomic_block += fi->atomic_write_cnt;
370 } else {
371 sbi->committed_atomic_block += fi->atomic_write_cnt;
372 set_inode_flag(inode, FI_ATOMIC_COMMITTED);
373 if (is_inode_flag_set(inode, FI_ATOMIC_DIRTIED)) {
374 clear_inode_flag(inode, FI_ATOMIC_DIRTIED);
375 f2fs_mark_inode_dirty_sync(inode, true);
376 }
377 }
378
379 __complete_revoke_list(inode, &revoke_list, ret ? true : false);
380
381 return ret;
382}
383
384int f2fs_commit_atomic_write(struct inode *inode)
385{
386 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
387 struct f2fs_inode_info *fi = F2FS_I(inode);
388 int err;
389
390 err = filemap_write_and_wait_range(inode->i_mapping, 0, LLONG_MAX);
391 if (err)
392 return err;
393
394 f2fs_down_write(&fi->i_gc_rwsem[WRITE]);
395 f2fs_lock_op(sbi);
396
397 err = __f2fs_commit_atomic_write(inode);
398
399 f2fs_unlock_op(sbi);
400 f2fs_up_write(&fi->i_gc_rwsem[WRITE]);
401
402 return err;
403}
404
405/*
406 * This function balances dirty node and dentry pages.
407 * In addition, it controls garbage collection.
408 */
409void f2fs_balance_fs(struct f2fs_sb_info *sbi, bool need)
410{
411 if (f2fs_cp_error(sbi))
412 return;
413
414 if (time_to_inject(sbi, FAULT_CHECKPOINT))
415 f2fs_stop_checkpoint(sbi, false, STOP_CP_REASON_FAULT_INJECT);
416
417 /* balance_fs_bg is able to be pending */
418 if (need && excess_cached_nats(sbi))
419 f2fs_balance_fs_bg(sbi, false);
420
421 if (!f2fs_is_checkpoint_ready(sbi))
422 return;
423
424 /*
425 * We should do GC or end up with checkpoint, if there are so many dirty
426 * dir/node pages without enough free segments.
427 */
428 if (has_enough_free_secs(sbi, 0, 0))
429 return;
430
431 if (test_opt(sbi, GC_MERGE) && sbi->gc_thread &&
432 sbi->gc_thread->f2fs_gc_task) {
433 DEFINE_WAIT(wait);
434
435 prepare_to_wait(&sbi->gc_thread->fggc_wq, &wait,
436 TASK_UNINTERRUPTIBLE);
437 wake_up(&sbi->gc_thread->gc_wait_queue_head);
438 io_schedule();
439 finish_wait(&sbi->gc_thread->fggc_wq, &wait);
440 } else {
441 struct f2fs_gc_control gc_control = {
442 .victim_segno = NULL_SEGNO,
443 .init_gc_type = BG_GC,
444 .no_bg_gc = true,
445 .should_migrate_blocks = false,
446 .err_gc_skipped = false,
447 .nr_free_secs = 1 };
448 f2fs_down_write(&sbi->gc_lock);
449 stat_inc_gc_call_count(sbi, FOREGROUND);
450 f2fs_gc(sbi, &gc_control);
451 }
452}
453
454static inline bool excess_dirty_threshold(struct f2fs_sb_info *sbi)
455{
456 int factor = f2fs_rwsem_is_locked(&sbi->cp_rwsem) ? 3 : 2;
457 unsigned int dents = get_pages(sbi, F2FS_DIRTY_DENTS);
458 unsigned int qdata = get_pages(sbi, F2FS_DIRTY_QDATA);
459 unsigned int nodes = get_pages(sbi, F2FS_DIRTY_NODES);
460 unsigned int meta = get_pages(sbi, F2FS_DIRTY_META);
461 unsigned int imeta = get_pages(sbi, F2FS_DIRTY_IMETA);
462 unsigned int threshold =
463 SEGS_TO_BLKS(sbi, (factor * DEFAULT_DIRTY_THRESHOLD));
464 unsigned int global_threshold = threshold * 3 / 2;
465
466 if (dents >= threshold || qdata >= threshold ||
467 nodes >= threshold || meta >= threshold ||
468 imeta >= threshold)
469 return true;
470 return dents + qdata + nodes + meta + imeta > global_threshold;
471}
472
473void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
474{
475 if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
476 return;
477
478 /* try to shrink extent cache when there is no enough memory */
479 if (!f2fs_available_free_memory(sbi, READ_EXTENT_CACHE))
480 f2fs_shrink_read_extent_tree(sbi,
481 READ_EXTENT_CACHE_SHRINK_NUMBER);
482
483 /* try to shrink age extent cache when there is no enough memory */
484 if (!f2fs_available_free_memory(sbi, AGE_EXTENT_CACHE))
485 f2fs_shrink_age_extent_tree(sbi,
486 AGE_EXTENT_CACHE_SHRINK_NUMBER);
487
488 /* check the # of cached NAT entries */
489 if (!f2fs_available_free_memory(sbi, NAT_ENTRIES))
490 f2fs_try_to_free_nats(sbi, NAT_ENTRY_PER_BLOCK);
491
492 if (!f2fs_available_free_memory(sbi, FREE_NIDS))
493 f2fs_try_to_free_nids(sbi, MAX_FREE_NIDS);
494 else
495 f2fs_build_free_nids(sbi, false, false);
496
497 if (excess_dirty_nats(sbi) || excess_dirty_threshold(sbi) ||
498 excess_prefree_segs(sbi) || !f2fs_space_for_roll_forward(sbi))
499 goto do_sync;
500
501 /* there is background inflight IO or foreground operation recently */
502 if (is_inflight_io(sbi, REQ_TIME) ||
503 (!f2fs_time_over(sbi, REQ_TIME) && f2fs_rwsem_is_locked(&sbi->cp_rwsem)))
504 return;
505
506 /* exceed periodical checkpoint timeout threshold */
507 if (f2fs_time_over(sbi, CP_TIME))
508 goto do_sync;
509
510 /* checkpoint is the only way to shrink partial cached entries */
511 if (f2fs_available_free_memory(sbi, NAT_ENTRIES) &&
512 f2fs_available_free_memory(sbi, INO_ENTRIES))
513 return;
514
515do_sync:
516 if (test_opt(sbi, DATA_FLUSH) && from_bg) {
517 struct blk_plug plug;
518
519 mutex_lock(&sbi->flush_lock);
520
521 blk_start_plug(&plug);
522 f2fs_sync_dirty_inodes(sbi, FILE_INODE, false);
523 blk_finish_plug(&plug);
524
525 mutex_unlock(&sbi->flush_lock);
526 }
527 stat_inc_cp_call_count(sbi, BACKGROUND);
528 f2fs_sync_fs(sbi->sb, 1);
529}
530
531static int __submit_flush_wait(struct f2fs_sb_info *sbi,
532 struct block_device *bdev)
533{
534 int ret = blkdev_issue_flush(bdev);
535
536 trace_f2fs_issue_flush(bdev, test_opt(sbi, NOBARRIER),
537 test_opt(sbi, FLUSH_MERGE), ret);
538 if (!ret)
539 f2fs_update_iostat(sbi, NULL, FS_FLUSH_IO, 0);
540 return ret;
541}
542
543static int submit_flush_wait(struct f2fs_sb_info *sbi, nid_t ino)
544{
545 int ret = 0;
546 int i;
547
548 if (!f2fs_is_multi_device(sbi))
549 return __submit_flush_wait(sbi, sbi->sb->s_bdev);
550
551 for (i = 0; i < sbi->s_ndevs; i++) {
552 if (!f2fs_is_dirty_device(sbi, ino, i, FLUSH_INO))
553 continue;
554 ret = __submit_flush_wait(sbi, FDEV(i).bdev);
555 if (ret)
556 break;
557 }
558 return ret;
559}
560
561static int issue_flush_thread(void *data)
562{
563 struct f2fs_sb_info *sbi = data;
564 struct flush_cmd_control *fcc = SM_I(sbi)->fcc_info;
565 wait_queue_head_t *q = &fcc->flush_wait_queue;
566repeat:
567 if (kthread_should_stop())
568 return 0;
569
570 if (!llist_empty(&fcc->issue_list)) {
571 struct flush_cmd *cmd, *next;
572 int ret;
573
574 fcc->dispatch_list = llist_del_all(&fcc->issue_list);
575 fcc->dispatch_list = llist_reverse_order(fcc->dispatch_list);
576
577 cmd = llist_entry(fcc->dispatch_list, struct flush_cmd, llnode);
578
579 ret = submit_flush_wait(sbi, cmd->ino);
580 atomic_inc(&fcc->issued_flush);
581
582 llist_for_each_entry_safe(cmd, next,
583 fcc->dispatch_list, llnode) {
584 cmd->ret = ret;
585 complete(&cmd->wait);
586 }
587 fcc->dispatch_list = NULL;
588 }
589
590 wait_event_interruptible(*q,
591 kthread_should_stop() || !llist_empty(&fcc->issue_list));
592 goto repeat;
593}
594
595int f2fs_issue_flush(struct f2fs_sb_info *sbi, nid_t ino)
596{
597 struct flush_cmd_control *fcc = SM_I(sbi)->fcc_info;
598 struct flush_cmd cmd;
599 int ret;
600
601 if (test_opt(sbi, NOBARRIER))
602 return 0;
603
604 if (!test_opt(sbi, FLUSH_MERGE)) {
605 atomic_inc(&fcc->queued_flush);
606 ret = submit_flush_wait(sbi, ino);
607 atomic_dec(&fcc->queued_flush);
608 atomic_inc(&fcc->issued_flush);
609 return ret;
610 }
611
612 if (atomic_inc_return(&fcc->queued_flush) == 1 ||
613 f2fs_is_multi_device(sbi)) {
614 ret = submit_flush_wait(sbi, ino);
615 atomic_dec(&fcc->queued_flush);
616
617 atomic_inc(&fcc->issued_flush);
618 return ret;
619 }
620
621 cmd.ino = ino;
622 init_completion(&cmd.wait);
623
624 llist_add(&cmd.llnode, &fcc->issue_list);
625
626 /*
627 * update issue_list before we wake up issue_flush thread, this
628 * smp_mb() pairs with another barrier in ___wait_event(), see
629 * more details in comments of waitqueue_active().
630 */
631 smp_mb();
632
633 if (waitqueue_active(&fcc->flush_wait_queue))
634 wake_up(&fcc->flush_wait_queue);
635
636 if (fcc->f2fs_issue_flush) {
637 wait_for_completion(&cmd.wait);
638 atomic_dec(&fcc->queued_flush);
639 } else {
640 struct llist_node *list;
641
642 list = llist_del_all(&fcc->issue_list);
643 if (!list) {
644 wait_for_completion(&cmd.wait);
645 atomic_dec(&fcc->queued_flush);
646 } else {
647 struct flush_cmd *tmp, *next;
648
649 ret = submit_flush_wait(sbi, ino);
650
651 llist_for_each_entry_safe(tmp, next, list, llnode) {
652 if (tmp == &cmd) {
653 cmd.ret = ret;
654 atomic_dec(&fcc->queued_flush);
655 continue;
656 }
657 tmp->ret = ret;
658 complete(&tmp->wait);
659 }
660 }
661 }
662
663 return cmd.ret;
664}
665
666int f2fs_create_flush_cmd_control(struct f2fs_sb_info *sbi)
667{
668 dev_t dev = sbi->sb->s_bdev->bd_dev;
669 struct flush_cmd_control *fcc;
670
671 if (SM_I(sbi)->fcc_info) {
672 fcc = SM_I(sbi)->fcc_info;
673 if (fcc->f2fs_issue_flush)
674 return 0;
675 goto init_thread;
676 }
677
678 fcc = f2fs_kzalloc(sbi, sizeof(struct flush_cmd_control), GFP_KERNEL);
679 if (!fcc)
680 return -ENOMEM;
681 atomic_set(&fcc->issued_flush, 0);
682 atomic_set(&fcc->queued_flush, 0);
683 init_waitqueue_head(&fcc->flush_wait_queue);
684 init_llist_head(&fcc->issue_list);
685 SM_I(sbi)->fcc_info = fcc;
686 if (!test_opt(sbi, FLUSH_MERGE))
687 return 0;
688
689init_thread:
690 fcc->f2fs_issue_flush = kthread_run(issue_flush_thread, sbi,
691 "f2fs_flush-%u:%u", MAJOR(dev), MINOR(dev));
692 if (IS_ERR(fcc->f2fs_issue_flush)) {
693 int err = PTR_ERR(fcc->f2fs_issue_flush);
694
695 fcc->f2fs_issue_flush = NULL;
696 return err;
697 }
698
699 return 0;
700}
701
702void f2fs_destroy_flush_cmd_control(struct f2fs_sb_info *sbi, bool free)
703{
704 struct flush_cmd_control *fcc = SM_I(sbi)->fcc_info;
705
706 if (fcc && fcc->f2fs_issue_flush) {
707 struct task_struct *flush_thread = fcc->f2fs_issue_flush;
708
709 fcc->f2fs_issue_flush = NULL;
710 kthread_stop(flush_thread);
711 }
712 if (free) {
713 kfree(fcc);
714 SM_I(sbi)->fcc_info = NULL;
715 }
716}
717
718int f2fs_flush_device_cache(struct f2fs_sb_info *sbi)
719{
720 int ret = 0, i;
721
722 if (!f2fs_is_multi_device(sbi))
723 return 0;
724
725 if (test_opt(sbi, NOBARRIER))
726 return 0;
727
728 for (i = 1; i < sbi->s_ndevs; i++) {
729 int count = DEFAULT_RETRY_IO_COUNT;
730
731 if (!f2fs_test_bit(i, (char *)&sbi->dirty_device))
732 continue;
733
734 do {
735 ret = __submit_flush_wait(sbi, FDEV(i).bdev);
736 if (ret)
737 f2fs_io_schedule_timeout(DEFAULT_IO_TIMEOUT);
738 } while (ret && --count);
739
740 if (ret) {
741 f2fs_stop_checkpoint(sbi, false,
742 STOP_CP_REASON_FLUSH_FAIL);
743 break;
744 }
745
746 spin_lock(&sbi->dev_lock);
747 f2fs_clear_bit(i, (char *)&sbi->dirty_device);
748 spin_unlock(&sbi->dev_lock);
749 }
750
751 return ret;
752}
753
754static void __locate_dirty_segment(struct f2fs_sb_info *sbi, unsigned int segno,
755 enum dirty_type dirty_type)
756{
757 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
758
759 /* need not be added */
760 if (IS_CURSEG(sbi, segno))
761 return;
762
763 if (!test_and_set_bit(segno, dirty_i->dirty_segmap[dirty_type]))
764 dirty_i->nr_dirty[dirty_type]++;
765
766 if (dirty_type == DIRTY) {
767 struct seg_entry *sentry = get_seg_entry(sbi, segno);
768 enum dirty_type t = sentry->type;
769
770 if (unlikely(t >= DIRTY)) {
771 f2fs_bug_on(sbi, 1);
772 return;
773 }
774 if (!test_and_set_bit(segno, dirty_i->dirty_segmap[t]))
775 dirty_i->nr_dirty[t]++;
776
777 if (__is_large_section(sbi)) {
778 unsigned int secno = GET_SEC_FROM_SEG(sbi, segno);
779 block_t valid_blocks =
780 get_valid_blocks(sbi, segno, true);
781
782 f2fs_bug_on(sbi,
783 (!is_sbi_flag_set(sbi, SBI_CP_DISABLED) &&
784 !valid_blocks) ||
785 valid_blocks == CAP_BLKS_PER_SEC(sbi));
786
787 if (!IS_CURSEC(sbi, secno))
788 set_bit(secno, dirty_i->dirty_secmap);
789 }
790 }
791}
792
793static void __remove_dirty_segment(struct f2fs_sb_info *sbi, unsigned int segno,
794 enum dirty_type dirty_type)
795{
796 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
797 block_t valid_blocks;
798
799 if (test_and_clear_bit(segno, dirty_i->dirty_segmap[dirty_type]))
800 dirty_i->nr_dirty[dirty_type]--;
801
802 if (dirty_type == DIRTY) {
803 struct seg_entry *sentry = get_seg_entry(sbi, segno);
804 enum dirty_type t = sentry->type;
805
806 if (test_and_clear_bit(segno, dirty_i->dirty_segmap[t]))
807 dirty_i->nr_dirty[t]--;
808
809 valid_blocks = get_valid_blocks(sbi, segno, true);
810 if (valid_blocks == 0) {
811 clear_bit(GET_SEC_FROM_SEG(sbi, segno),
812 dirty_i->victim_secmap);
813#ifdef CONFIG_F2FS_CHECK_FS
814 clear_bit(segno, SIT_I(sbi)->invalid_segmap);
815#endif
816 }
817 if (__is_large_section(sbi)) {
818 unsigned int secno = GET_SEC_FROM_SEG(sbi, segno);
819
820 if (!valid_blocks ||
821 valid_blocks == CAP_BLKS_PER_SEC(sbi)) {
822 clear_bit(secno, dirty_i->dirty_secmap);
823 return;
824 }
825
826 if (!IS_CURSEC(sbi, secno))
827 set_bit(secno, dirty_i->dirty_secmap);
828 }
829 }
830}
831
832/*
833 * Should not occur error such as -ENOMEM.
834 * Adding dirty entry into seglist is not critical operation.
835 * If a given segment is one of current working segments, it won't be added.
836 */
837static void locate_dirty_segment(struct f2fs_sb_info *sbi, unsigned int segno)
838{
839 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
840 unsigned short valid_blocks, ckpt_valid_blocks;
841 unsigned int usable_blocks;
842
843 if (segno == NULL_SEGNO || IS_CURSEG(sbi, segno))
844 return;
845
846 usable_blocks = f2fs_usable_blks_in_seg(sbi, segno);
847 mutex_lock(&dirty_i->seglist_lock);
848
849 valid_blocks = get_valid_blocks(sbi, segno, false);
850 ckpt_valid_blocks = get_ckpt_valid_blocks(sbi, segno, false);
851
852 if (valid_blocks == 0 && (!is_sbi_flag_set(sbi, SBI_CP_DISABLED) ||
853 ckpt_valid_blocks == usable_blocks)) {
854 __locate_dirty_segment(sbi, segno, PRE);
855 __remove_dirty_segment(sbi, segno, DIRTY);
856 } else if (valid_blocks < usable_blocks) {
857 __locate_dirty_segment(sbi, segno, DIRTY);
858 } else {
859 /* Recovery routine with SSR needs this */
860 __remove_dirty_segment(sbi, segno, DIRTY);
861 }
862
863 mutex_unlock(&dirty_i->seglist_lock);
864}
865
866/* This moves currently empty dirty blocks to prefree. Must hold seglist_lock */
867void f2fs_dirty_to_prefree(struct f2fs_sb_info *sbi)
868{
869 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
870 unsigned int segno;
871
872 mutex_lock(&dirty_i->seglist_lock);
873 for_each_set_bit(segno, dirty_i->dirty_segmap[DIRTY], MAIN_SEGS(sbi)) {
874 if (get_valid_blocks(sbi, segno, false))
875 continue;
876 if (IS_CURSEG(sbi, segno))
877 continue;
878 __locate_dirty_segment(sbi, segno, PRE);
879 __remove_dirty_segment(sbi, segno, DIRTY);
880 }
881 mutex_unlock(&dirty_i->seglist_lock);
882}
883
884block_t f2fs_get_unusable_blocks(struct f2fs_sb_info *sbi)
885{
886 int ovp_hole_segs =
887 (overprovision_segments(sbi) - reserved_segments(sbi));
888 block_t ovp_holes = SEGS_TO_BLKS(sbi, ovp_hole_segs);
889 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
890 block_t holes[2] = {0, 0}; /* DATA and NODE */
891 block_t unusable;
892 struct seg_entry *se;
893 unsigned int segno;
894
895 mutex_lock(&dirty_i->seglist_lock);
896 for_each_set_bit(segno, dirty_i->dirty_segmap[DIRTY], MAIN_SEGS(sbi)) {
897 se = get_seg_entry(sbi, segno);
898 if (IS_NODESEG(se->type))
899 holes[NODE] += f2fs_usable_blks_in_seg(sbi, segno) -
900 se->valid_blocks;
901 else
902 holes[DATA] += f2fs_usable_blks_in_seg(sbi, segno) -
903 se->valid_blocks;
904 }
905 mutex_unlock(&dirty_i->seglist_lock);
906
907 unusable = max(holes[DATA], holes[NODE]);
908 if (unusable > ovp_holes)
909 return unusable - ovp_holes;
910 return 0;
911}
912
913int f2fs_disable_cp_again(struct f2fs_sb_info *sbi, block_t unusable)
914{
915 int ovp_hole_segs =
916 (overprovision_segments(sbi) - reserved_segments(sbi));
917
918 if (F2FS_OPTION(sbi).unusable_cap_perc == 100)
919 return 0;
920 if (unusable > F2FS_OPTION(sbi).unusable_cap)
921 return -EAGAIN;
922 if (is_sbi_flag_set(sbi, SBI_CP_DISABLED_QUICK) &&
923 dirty_segments(sbi) > ovp_hole_segs)
924 return -EAGAIN;
925 if (has_not_enough_free_secs(sbi, 0, 0))
926 return -EAGAIN;
927 return 0;
928}
929
930/* This is only used by SBI_CP_DISABLED */
931static unsigned int get_free_segment(struct f2fs_sb_info *sbi)
932{
933 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
934 unsigned int segno = 0;
935
936 mutex_lock(&dirty_i->seglist_lock);
937 for_each_set_bit(segno, dirty_i->dirty_segmap[DIRTY], MAIN_SEGS(sbi)) {
938 if (get_valid_blocks(sbi, segno, false))
939 continue;
940 if (get_ckpt_valid_blocks(sbi, segno, false))
941 continue;
942 mutex_unlock(&dirty_i->seglist_lock);
943 return segno;
944 }
945 mutex_unlock(&dirty_i->seglist_lock);
946 return NULL_SEGNO;
947}
948
949static struct discard_cmd *__create_discard_cmd(struct f2fs_sb_info *sbi,
950 struct block_device *bdev, block_t lstart,
951 block_t start, block_t len)
952{
953 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
954 struct list_head *pend_list;
955 struct discard_cmd *dc;
956
957 f2fs_bug_on(sbi, !len);
958
959 pend_list = &dcc->pend_list[plist_idx(len)];
960
961 dc = f2fs_kmem_cache_alloc(discard_cmd_slab, GFP_NOFS, true, NULL);
962 INIT_LIST_HEAD(&dc->list);
963 dc->bdev = bdev;
964 dc->di.lstart = lstart;
965 dc->di.start = start;
966 dc->di.len = len;
967 dc->ref = 0;
968 dc->state = D_PREP;
969 dc->queued = 0;
970 dc->error = 0;
971 init_completion(&dc->wait);
972 list_add_tail(&dc->list, pend_list);
973 spin_lock_init(&dc->lock);
974 dc->bio_ref = 0;
975 atomic_inc(&dcc->discard_cmd_cnt);
976 dcc->undiscard_blks += len;
977
978 return dc;
979}
980
981static bool f2fs_check_discard_tree(struct f2fs_sb_info *sbi)
982{
983#ifdef CONFIG_F2FS_CHECK_FS
984 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
985 struct rb_node *cur = rb_first_cached(&dcc->root), *next;
986 struct discard_cmd *cur_dc, *next_dc;
987
988 while (cur) {
989 next = rb_next(cur);
990 if (!next)
991 return true;
992
993 cur_dc = rb_entry(cur, struct discard_cmd, rb_node);
994 next_dc = rb_entry(next, struct discard_cmd, rb_node);
995
996 if (cur_dc->di.lstart + cur_dc->di.len > next_dc->di.lstart) {
997 f2fs_info(sbi, "broken discard_rbtree, "
998 "cur(%u, %u) next(%u, %u)",
999 cur_dc->di.lstart, cur_dc->di.len,
1000 next_dc->di.lstart, next_dc->di.len);
1001 return false;
1002 }
1003 cur = next;
1004 }
1005#endif
1006 return true;
1007}
1008
1009static struct discard_cmd *__lookup_discard_cmd(struct f2fs_sb_info *sbi,
1010 block_t blkaddr)
1011{
1012 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1013 struct rb_node *node = dcc->root.rb_root.rb_node;
1014 struct discard_cmd *dc;
1015
1016 while (node) {
1017 dc = rb_entry(node, struct discard_cmd, rb_node);
1018
1019 if (blkaddr < dc->di.lstart)
1020 node = node->rb_left;
1021 else if (blkaddr >= dc->di.lstart + dc->di.len)
1022 node = node->rb_right;
1023 else
1024 return dc;
1025 }
1026 return NULL;
1027}
1028
1029static struct discard_cmd *__lookup_discard_cmd_ret(struct rb_root_cached *root,
1030 block_t blkaddr,
1031 struct discard_cmd **prev_entry,
1032 struct discard_cmd **next_entry,
1033 struct rb_node ***insert_p,
1034 struct rb_node **insert_parent)
1035{
1036 struct rb_node **pnode = &root->rb_root.rb_node;
1037 struct rb_node *parent = NULL, *tmp_node;
1038 struct discard_cmd *dc;
1039
1040 *insert_p = NULL;
1041 *insert_parent = NULL;
1042 *prev_entry = NULL;
1043 *next_entry = NULL;
1044
1045 if (RB_EMPTY_ROOT(&root->rb_root))
1046 return NULL;
1047
1048 while (*pnode) {
1049 parent = *pnode;
1050 dc = rb_entry(*pnode, struct discard_cmd, rb_node);
1051
1052 if (blkaddr < dc->di.lstart)
1053 pnode = &(*pnode)->rb_left;
1054 else if (blkaddr >= dc->di.lstart + dc->di.len)
1055 pnode = &(*pnode)->rb_right;
1056 else
1057 goto lookup_neighbors;
1058 }
1059
1060 *insert_p = pnode;
1061 *insert_parent = parent;
1062
1063 dc = rb_entry(parent, struct discard_cmd, rb_node);
1064 tmp_node = parent;
1065 if (parent && blkaddr > dc->di.lstart)
1066 tmp_node = rb_next(parent);
1067 *next_entry = rb_entry_safe(tmp_node, struct discard_cmd, rb_node);
1068
1069 tmp_node = parent;
1070 if (parent && blkaddr < dc->di.lstart)
1071 tmp_node = rb_prev(parent);
1072 *prev_entry = rb_entry_safe(tmp_node, struct discard_cmd, rb_node);
1073 return NULL;
1074
1075lookup_neighbors:
1076 /* lookup prev node for merging backward later */
1077 tmp_node = rb_prev(&dc->rb_node);
1078 *prev_entry = rb_entry_safe(tmp_node, struct discard_cmd, rb_node);
1079
1080 /* lookup next node for merging frontward later */
1081 tmp_node = rb_next(&dc->rb_node);
1082 *next_entry = rb_entry_safe(tmp_node, struct discard_cmd, rb_node);
1083 return dc;
1084}
1085
1086static void __detach_discard_cmd(struct discard_cmd_control *dcc,
1087 struct discard_cmd *dc)
1088{
1089 if (dc->state == D_DONE)
1090 atomic_sub(dc->queued, &dcc->queued_discard);
1091
1092 list_del(&dc->list);
1093 rb_erase_cached(&dc->rb_node, &dcc->root);
1094 dcc->undiscard_blks -= dc->di.len;
1095
1096 kmem_cache_free(discard_cmd_slab, dc);
1097
1098 atomic_dec(&dcc->discard_cmd_cnt);
1099}
1100
1101static void __remove_discard_cmd(struct f2fs_sb_info *sbi,
1102 struct discard_cmd *dc)
1103{
1104 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1105 unsigned long flags;
1106
1107 trace_f2fs_remove_discard(dc->bdev, dc->di.start, dc->di.len);
1108
1109 spin_lock_irqsave(&dc->lock, flags);
1110 if (dc->bio_ref) {
1111 spin_unlock_irqrestore(&dc->lock, flags);
1112 return;
1113 }
1114 spin_unlock_irqrestore(&dc->lock, flags);
1115
1116 f2fs_bug_on(sbi, dc->ref);
1117
1118 if (dc->error == -EOPNOTSUPP)
1119 dc->error = 0;
1120
1121 if (dc->error)
1122 f2fs_info_ratelimited(sbi,
1123 "Issue discard(%u, %u, %u) failed, ret: %d",
1124 dc->di.lstart, dc->di.start, dc->di.len, dc->error);
1125 __detach_discard_cmd(dcc, dc);
1126}
1127
1128static void f2fs_submit_discard_endio(struct bio *bio)
1129{
1130 struct discard_cmd *dc = (struct discard_cmd *)bio->bi_private;
1131 unsigned long flags;
1132
1133 spin_lock_irqsave(&dc->lock, flags);
1134 if (!dc->error)
1135 dc->error = blk_status_to_errno(bio->bi_status);
1136 dc->bio_ref--;
1137 if (!dc->bio_ref && dc->state == D_SUBMIT) {
1138 dc->state = D_DONE;
1139 complete_all(&dc->wait);
1140 }
1141 spin_unlock_irqrestore(&dc->lock, flags);
1142 bio_put(bio);
1143}
1144
1145static void __check_sit_bitmap(struct f2fs_sb_info *sbi,
1146 block_t start, block_t end)
1147{
1148#ifdef CONFIG_F2FS_CHECK_FS
1149 struct seg_entry *sentry;
1150 unsigned int segno;
1151 block_t blk = start;
1152 unsigned long offset, size, *map;
1153
1154 while (blk < end) {
1155 segno = GET_SEGNO(sbi, blk);
1156 sentry = get_seg_entry(sbi, segno);
1157 offset = GET_BLKOFF_FROM_SEG0(sbi, blk);
1158
1159 if (end < START_BLOCK(sbi, segno + 1))
1160 size = GET_BLKOFF_FROM_SEG0(sbi, end);
1161 else
1162 size = BLKS_PER_SEG(sbi);
1163 map = (unsigned long *)(sentry->cur_valid_map);
1164 offset = __find_rev_next_bit(map, size, offset);
1165 f2fs_bug_on(sbi, offset != size);
1166 blk = START_BLOCK(sbi, segno + 1);
1167 }
1168#endif
1169}
1170
1171static void __init_discard_policy(struct f2fs_sb_info *sbi,
1172 struct discard_policy *dpolicy,
1173 int discard_type, unsigned int granularity)
1174{
1175 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1176
1177 /* common policy */
1178 dpolicy->type = discard_type;
1179 dpolicy->sync = true;
1180 dpolicy->ordered = false;
1181 dpolicy->granularity = granularity;
1182
1183 dpolicy->max_requests = dcc->max_discard_request;
1184 dpolicy->io_aware_gran = dcc->discard_io_aware_gran;
1185 dpolicy->timeout = false;
1186
1187 if (discard_type == DPOLICY_BG) {
1188 dpolicy->min_interval = dcc->min_discard_issue_time;
1189 dpolicy->mid_interval = dcc->mid_discard_issue_time;
1190 dpolicy->max_interval = dcc->max_discard_issue_time;
1191 if (dcc->discard_io_aware == DPOLICY_IO_AWARE_ENABLE)
1192 dpolicy->io_aware = true;
1193 else if (dcc->discard_io_aware == DPOLICY_IO_AWARE_DISABLE)
1194 dpolicy->io_aware = false;
1195 dpolicy->sync = false;
1196 dpolicy->ordered = true;
1197 if (utilization(sbi) > dcc->discard_urgent_util) {
1198 dpolicy->granularity = MIN_DISCARD_GRANULARITY;
1199 if (atomic_read(&dcc->discard_cmd_cnt))
1200 dpolicy->max_interval =
1201 dcc->min_discard_issue_time;
1202 }
1203 } else if (discard_type == DPOLICY_FORCE) {
1204 dpolicy->min_interval = dcc->min_discard_issue_time;
1205 dpolicy->mid_interval = dcc->mid_discard_issue_time;
1206 dpolicy->max_interval = dcc->max_discard_issue_time;
1207 dpolicy->io_aware = false;
1208 } else if (discard_type == DPOLICY_FSTRIM) {
1209 dpolicy->io_aware = false;
1210 } else if (discard_type == DPOLICY_UMOUNT) {
1211 dpolicy->io_aware = false;
1212 /* we need to issue all to keep CP_TRIMMED_FLAG */
1213 dpolicy->granularity = MIN_DISCARD_GRANULARITY;
1214 dpolicy->timeout = true;
1215 }
1216}
1217
1218static void __update_discard_tree_range(struct f2fs_sb_info *sbi,
1219 struct block_device *bdev, block_t lstart,
1220 block_t start, block_t len);
1221
1222#ifdef CONFIG_BLK_DEV_ZONED
1223static void __submit_zone_reset_cmd(struct f2fs_sb_info *sbi,
1224 struct discard_cmd *dc, blk_opf_t flag,
1225 struct list_head *wait_list,
1226 unsigned int *issued)
1227{
1228 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1229 struct block_device *bdev = dc->bdev;
1230 struct bio *bio = bio_alloc(bdev, 0, REQ_OP_ZONE_RESET | flag, GFP_NOFS);
1231 unsigned long flags;
1232
1233 trace_f2fs_issue_reset_zone(bdev, dc->di.start);
1234
1235 spin_lock_irqsave(&dc->lock, flags);
1236 dc->state = D_SUBMIT;
1237 dc->bio_ref++;
1238 spin_unlock_irqrestore(&dc->lock, flags);
1239
1240 if (issued)
1241 (*issued)++;
1242
1243 atomic_inc(&dcc->queued_discard);
1244 dc->queued++;
1245 list_move_tail(&dc->list, wait_list);
1246
1247 /* sanity check on discard range */
1248 __check_sit_bitmap(sbi, dc->di.lstart, dc->di.lstart + dc->di.len);
1249
1250 bio->bi_iter.bi_sector = SECTOR_FROM_BLOCK(dc->di.start);
1251 bio->bi_private = dc;
1252 bio->bi_end_io = f2fs_submit_discard_endio;
1253 submit_bio(bio);
1254
1255 atomic_inc(&dcc->issued_discard);
1256 f2fs_update_iostat(sbi, NULL, FS_ZONE_RESET_IO, dc->di.len * F2FS_BLKSIZE);
1257}
1258#endif
1259
1260/* this function is copied from blkdev_issue_discard from block/blk-lib.c */
1261static int __submit_discard_cmd(struct f2fs_sb_info *sbi,
1262 struct discard_policy *dpolicy,
1263 struct discard_cmd *dc, int *issued)
1264{
1265 struct block_device *bdev = dc->bdev;
1266 unsigned int max_discard_blocks =
1267 SECTOR_TO_BLOCK(bdev_max_discard_sectors(bdev));
1268 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1269 struct list_head *wait_list = (dpolicy->type == DPOLICY_FSTRIM) ?
1270 &(dcc->fstrim_list) : &(dcc->wait_list);
1271 blk_opf_t flag = dpolicy->sync ? REQ_SYNC : 0;
1272 block_t lstart, start, len, total_len;
1273 int err = 0;
1274
1275 if (dc->state != D_PREP)
1276 return 0;
1277
1278 if (is_sbi_flag_set(sbi, SBI_NEED_FSCK))
1279 return 0;
1280
1281#ifdef CONFIG_BLK_DEV_ZONED
1282 if (f2fs_sb_has_blkzoned(sbi) && bdev_is_zoned(bdev)) {
1283 int devi = f2fs_bdev_index(sbi, bdev);
1284
1285 if (devi < 0)
1286 return -EINVAL;
1287
1288 if (f2fs_blkz_is_seq(sbi, devi, dc->di.start)) {
1289 __submit_zone_reset_cmd(sbi, dc, flag,
1290 wait_list, issued);
1291 return 0;
1292 }
1293 }
1294#endif
1295
1296 /*
1297 * stop issuing discard for any of below cases:
1298 * 1. device is conventional zone, but it doesn't support discard.
1299 * 2. device is regulare device, after snapshot it doesn't support
1300 * discard.
1301 */
1302 if (!bdev_max_discard_sectors(bdev))
1303 return -EOPNOTSUPP;
1304
1305 trace_f2fs_issue_discard(bdev, dc->di.start, dc->di.len);
1306
1307 lstart = dc->di.lstart;
1308 start = dc->di.start;
1309 len = dc->di.len;
1310 total_len = len;
1311
1312 dc->di.len = 0;
1313
1314 while (total_len && *issued < dpolicy->max_requests && !err) {
1315 struct bio *bio = NULL;
1316 unsigned long flags;
1317 bool last = true;
1318
1319 if (len > max_discard_blocks) {
1320 len = max_discard_blocks;
1321 last = false;
1322 }
1323
1324 (*issued)++;
1325 if (*issued == dpolicy->max_requests)
1326 last = true;
1327
1328 dc->di.len += len;
1329
1330 if (time_to_inject(sbi, FAULT_DISCARD)) {
1331 err = -EIO;
1332 } else {
1333 err = __blkdev_issue_discard(bdev,
1334 SECTOR_FROM_BLOCK(start),
1335 SECTOR_FROM_BLOCK(len),
1336 GFP_NOFS, &bio);
1337 }
1338 if (err) {
1339 spin_lock_irqsave(&dc->lock, flags);
1340 if (dc->state == D_PARTIAL)
1341 dc->state = D_SUBMIT;
1342 spin_unlock_irqrestore(&dc->lock, flags);
1343
1344 break;
1345 }
1346
1347 f2fs_bug_on(sbi, !bio);
1348
1349 /*
1350 * should keep before submission to avoid D_DONE
1351 * right away
1352 */
1353 spin_lock_irqsave(&dc->lock, flags);
1354 if (last)
1355 dc->state = D_SUBMIT;
1356 else
1357 dc->state = D_PARTIAL;
1358 dc->bio_ref++;
1359 spin_unlock_irqrestore(&dc->lock, flags);
1360
1361 atomic_inc(&dcc->queued_discard);
1362 dc->queued++;
1363 list_move_tail(&dc->list, wait_list);
1364
1365 /* sanity check on discard range */
1366 __check_sit_bitmap(sbi, lstart, lstart + len);
1367
1368 bio->bi_private = dc;
1369 bio->bi_end_io = f2fs_submit_discard_endio;
1370 bio->bi_opf |= flag;
1371 submit_bio(bio);
1372
1373 atomic_inc(&dcc->issued_discard);
1374
1375 f2fs_update_iostat(sbi, NULL, FS_DISCARD_IO, len * F2FS_BLKSIZE);
1376
1377 lstart += len;
1378 start += len;
1379 total_len -= len;
1380 len = total_len;
1381 }
1382
1383 if (!err && len) {
1384 dcc->undiscard_blks -= len;
1385 __update_discard_tree_range(sbi, bdev, lstart, start, len);
1386 }
1387 return err;
1388}
1389
1390static void __insert_discard_cmd(struct f2fs_sb_info *sbi,
1391 struct block_device *bdev, block_t lstart,
1392 block_t start, block_t len)
1393{
1394 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1395 struct rb_node **p = &dcc->root.rb_root.rb_node;
1396 struct rb_node *parent = NULL;
1397 struct discard_cmd *dc;
1398 bool leftmost = true;
1399
1400 /* look up rb tree to find parent node */
1401 while (*p) {
1402 parent = *p;
1403 dc = rb_entry(parent, struct discard_cmd, rb_node);
1404
1405 if (lstart < dc->di.lstart) {
1406 p = &(*p)->rb_left;
1407 } else if (lstart >= dc->di.lstart + dc->di.len) {
1408 p = &(*p)->rb_right;
1409 leftmost = false;
1410 } else {
1411 /* Let's skip to add, if exists */
1412 return;
1413 }
1414 }
1415
1416 dc = __create_discard_cmd(sbi, bdev, lstart, start, len);
1417
1418 rb_link_node(&dc->rb_node, parent, p);
1419 rb_insert_color_cached(&dc->rb_node, &dcc->root, leftmost);
1420}
1421
1422static void __relocate_discard_cmd(struct discard_cmd_control *dcc,
1423 struct discard_cmd *dc)
1424{
1425 list_move_tail(&dc->list, &dcc->pend_list[plist_idx(dc->di.len)]);
1426}
1427
1428static void __punch_discard_cmd(struct f2fs_sb_info *sbi,
1429 struct discard_cmd *dc, block_t blkaddr)
1430{
1431 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1432 struct discard_info di = dc->di;
1433 bool modified = false;
1434
1435 if (dc->state == D_DONE || dc->di.len == 1) {
1436 __remove_discard_cmd(sbi, dc);
1437 return;
1438 }
1439
1440 dcc->undiscard_blks -= di.len;
1441
1442 if (blkaddr > di.lstart) {
1443 dc->di.len = blkaddr - dc->di.lstart;
1444 dcc->undiscard_blks += dc->di.len;
1445 __relocate_discard_cmd(dcc, dc);
1446 modified = true;
1447 }
1448
1449 if (blkaddr < di.lstart + di.len - 1) {
1450 if (modified) {
1451 __insert_discard_cmd(sbi, dc->bdev, blkaddr + 1,
1452 di.start + blkaddr + 1 - di.lstart,
1453 di.lstart + di.len - 1 - blkaddr);
1454 } else {
1455 dc->di.lstart++;
1456 dc->di.len--;
1457 dc->di.start++;
1458 dcc->undiscard_blks += dc->di.len;
1459 __relocate_discard_cmd(dcc, dc);
1460 }
1461 }
1462}
1463
1464static void __update_discard_tree_range(struct f2fs_sb_info *sbi,
1465 struct block_device *bdev, block_t lstart,
1466 block_t start, block_t len)
1467{
1468 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1469 struct discard_cmd *prev_dc = NULL, *next_dc = NULL;
1470 struct discard_cmd *dc;
1471 struct discard_info di = {0};
1472 struct rb_node **insert_p = NULL, *insert_parent = NULL;
1473 unsigned int max_discard_blocks =
1474 SECTOR_TO_BLOCK(bdev_max_discard_sectors(bdev));
1475 block_t end = lstart + len;
1476
1477 dc = __lookup_discard_cmd_ret(&dcc->root, lstart,
1478 &prev_dc, &next_dc, &insert_p, &insert_parent);
1479 if (dc)
1480 prev_dc = dc;
1481
1482 if (!prev_dc) {
1483 di.lstart = lstart;
1484 di.len = next_dc ? next_dc->di.lstart - lstart : len;
1485 di.len = min(di.len, len);
1486 di.start = start;
1487 }
1488
1489 while (1) {
1490 struct rb_node *node;
1491 bool merged = false;
1492 struct discard_cmd *tdc = NULL;
1493
1494 if (prev_dc) {
1495 di.lstart = prev_dc->di.lstart + prev_dc->di.len;
1496 if (di.lstart < lstart)
1497 di.lstart = lstart;
1498 if (di.lstart >= end)
1499 break;
1500
1501 if (!next_dc || next_dc->di.lstart > end)
1502 di.len = end - di.lstart;
1503 else
1504 di.len = next_dc->di.lstart - di.lstart;
1505 di.start = start + di.lstart - lstart;
1506 }
1507
1508 if (!di.len)
1509 goto next;
1510
1511 if (prev_dc && prev_dc->state == D_PREP &&
1512 prev_dc->bdev == bdev &&
1513 __is_discard_back_mergeable(&di, &prev_dc->di,
1514 max_discard_blocks)) {
1515 prev_dc->di.len += di.len;
1516 dcc->undiscard_blks += di.len;
1517 __relocate_discard_cmd(dcc, prev_dc);
1518 di = prev_dc->di;
1519 tdc = prev_dc;
1520 merged = true;
1521 }
1522
1523 if (next_dc && next_dc->state == D_PREP &&
1524 next_dc->bdev == bdev &&
1525 __is_discard_front_mergeable(&di, &next_dc->di,
1526 max_discard_blocks)) {
1527 next_dc->di.lstart = di.lstart;
1528 next_dc->di.len += di.len;
1529 next_dc->di.start = di.start;
1530 dcc->undiscard_blks += di.len;
1531 __relocate_discard_cmd(dcc, next_dc);
1532 if (tdc)
1533 __remove_discard_cmd(sbi, tdc);
1534 merged = true;
1535 }
1536
1537 if (!merged)
1538 __insert_discard_cmd(sbi, bdev,
1539 di.lstart, di.start, di.len);
1540 next:
1541 prev_dc = next_dc;
1542 if (!prev_dc)
1543 break;
1544
1545 node = rb_next(&prev_dc->rb_node);
1546 next_dc = rb_entry_safe(node, struct discard_cmd, rb_node);
1547 }
1548}
1549
1550#ifdef CONFIG_BLK_DEV_ZONED
1551static void __queue_zone_reset_cmd(struct f2fs_sb_info *sbi,
1552 struct block_device *bdev, block_t blkstart, block_t lblkstart,
1553 block_t blklen)
1554{
1555 trace_f2fs_queue_reset_zone(bdev, blkstart);
1556
1557 mutex_lock(&SM_I(sbi)->dcc_info->cmd_lock);
1558 __insert_discard_cmd(sbi, bdev, lblkstart, blkstart, blklen);
1559 mutex_unlock(&SM_I(sbi)->dcc_info->cmd_lock);
1560}
1561#endif
1562
1563static void __queue_discard_cmd(struct f2fs_sb_info *sbi,
1564 struct block_device *bdev, block_t blkstart, block_t blklen)
1565{
1566 block_t lblkstart = blkstart;
1567
1568 if (!f2fs_bdev_support_discard(bdev))
1569 return;
1570
1571 trace_f2fs_queue_discard(bdev, blkstart, blklen);
1572
1573 if (f2fs_is_multi_device(sbi)) {
1574 int devi = f2fs_target_device_index(sbi, blkstart);
1575
1576 blkstart -= FDEV(devi).start_blk;
1577 }
1578 mutex_lock(&SM_I(sbi)->dcc_info->cmd_lock);
1579 __update_discard_tree_range(sbi, bdev, lblkstart, blkstart, blklen);
1580 mutex_unlock(&SM_I(sbi)->dcc_info->cmd_lock);
1581}
1582
1583static void __issue_discard_cmd_orderly(struct f2fs_sb_info *sbi,
1584 struct discard_policy *dpolicy, int *issued)
1585{
1586 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1587 struct discard_cmd *prev_dc = NULL, *next_dc = NULL;
1588 struct rb_node **insert_p = NULL, *insert_parent = NULL;
1589 struct discard_cmd *dc;
1590 struct blk_plug plug;
1591 bool io_interrupted = false;
1592
1593 mutex_lock(&dcc->cmd_lock);
1594 dc = __lookup_discard_cmd_ret(&dcc->root, dcc->next_pos,
1595 &prev_dc, &next_dc, &insert_p, &insert_parent);
1596 if (!dc)
1597 dc = next_dc;
1598
1599 blk_start_plug(&plug);
1600
1601 while (dc) {
1602 struct rb_node *node;
1603 int err = 0;
1604
1605 if (dc->state != D_PREP)
1606 goto next;
1607
1608 if (dpolicy->io_aware && !is_idle(sbi, DISCARD_TIME)) {
1609 io_interrupted = true;
1610 break;
1611 }
1612
1613 dcc->next_pos = dc->di.lstart + dc->di.len;
1614 err = __submit_discard_cmd(sbi, dpolicy, dc, issued);
1615
1616 if (*issued >= dpolicy->max_requests)
1617 break;
1618next:
1619 node = rb_next(&dc->rb_node);
1620 if (err)
1621 __remove_discard_cmd(sbi, dc);
1622 dc = rb_entry_safe(node, struct discard_cmd, rb_node);
1623 }
1624
1625 blk_finish_plug(&plug);
1626
1627 if (!dc)
1628 dcc->next_pos = 0;
1629
1630 mutex_unlock(&dcc->cmd_lock);
1631
1632 if (!(*issued) && io_interrupted)
1633 *issued = -1;
1634}
1635static unsigned int __wait_all_discard_cmd(struct f2fs_sb_info *sbi,
1636 struct discard_policy *dpolicy);
1637
1638static int __issue_discard_cmd(struct f2fs_sb_info *sbi,
1639 struct discard_policy *dpolicy)
1640{
1641 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1642 struct list_head *pend_list;
1643 struct discard_cmd *dc, *tmp;
1644 struct blk_plug plug;
1645 int i, issued;
1646 bool io_interrupted = false;
1647
1648 if (dpolicy->timeout)
1649 f2fs_update_time(sbi, UMOUNT_DISCARD_TIMEOUT);
1650
1651retry:
1652 issued = 0;
1653 for (i = MAX_PLIST_NUM - 1; i >= 0; i--) {
1654 if (dpolicy->timeout &&
1655 f2fs_time_over(sbi, UMOUNT_DISCARD_TIMEOUT))
1656 break;
1657
1658 if (i + 1 < dpolicy->granularity)
1659 break;
1660
1661 if (i + 1 < dcc->max_ordered_discard && dpolicy->ordered) {
1662 __issue_discard_cmd_orderly(sbi, dpolicy, &issued);
1663 return issued;
1664 }
1665
1666 pend_list = &dcc->pend_list[i];
1667
1668 mutex_lock(&dcc->cmd_lock);
1669 if (list_empty(pend_list))
1670 goto next;
1671 if (unlikely(dcc->rbtree_check))
1672 f2fs_bug_on(sbi, !f2fs_check_discard_tree(sbi));
1673 blk_start_plug(&plug);
1674 list_for_each_entry_safe(dc, tmp, pend_list, list) {
1675 f2fs_bug_on(sbi, dc->state != D_PREP);
1676
1677 if (dpolicy->timeout &&
1678 f2fs_time_over(sbi, UMOUNT_DISCARD_TIMEOUT))
1679 break;
1680
1681 if (dpolicy->io_aware && i < dpolicy->io_aware_gran &&
1682 !is_idle(sbi, DISCARD_TIME)) {
1683 io_interrupted = true;
1684 break;
1685 }
1686
1687 __submit_discard_cmd(sbi, dpolicy, dc, &issued);
1688
1689 if (issued >= dpolicy->max_requests)
1690 break;
1691 }
1692 blk_finish_plug(&plug);
1693next:
1694 mutex_unlock(&dcc->cmd_lock);
1695
1696 if (issued >= dpolicy->max_requests || io_interrupted)
1697 break;
1698 }
1699
1700 if (dpolicy->type == DPOLICY_UMOUNT && issued) {
1701 __wait_all_discard_cmd(sbi, dpolicy);
1702 goto retry;
1703 }
1704
1705 if (!issued && io_interrupted)
1706 issued = -1;
1707
1708 return issued;
1709}
1710
1711static bool __drop_discard_cmd(struct f2fs_sb_info *sbi)
1712{
1713 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1714 struct list_head *pend_list;
1715 struct discard_cmd *dc, *tmp;
1716 int i;
1717 bool dropped = false;
1718
1719 mutex_lock(&dcc->cmd_lock);
1720 for (i = MAX_PLIST_NUM - 1; i >= 0; i--) {
1721 pend_list = &dcc->pend_list[i];
1722 list_for_each_entry_safe(dc, tmp, pend_list, list) {
1723 f2fs_bug_on(sbi, dc->state != D_PREP);
1724 __remove_discard_cmd(sbi, dc);
1725 dropped = true;
1726 }
1727 }
1728 mutex_unlock(&dcc->cmd_lock);
1729
1730 return dropped;
1731}
1732
1733void f2fs_drop_discard_cmd(struct f2fs_sb_info *sbi)
1734{
1735 __drop_discard_cmd(sbi);
1736}
1737
1738static unsigned int __wait_one_discard_bio(struct f2fs_sb_info *sbi,
1739 struct discard_cmd *dc)
1740{
1741 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1742 unsigned int len = 0;
1743
1744 wait_for_completion_io(&dc->wait);
1745 mutex_lock(&dcc->cmd_lock);
1746 f2fs_bug_on(sbi, dc->state != D_DONE);
1747 dc->ref--;
1748 if (!dc->ref) {
1749 if (!dc->error)
1750 len = dc->di.len;
1751 __remove_discard_cmd(sbi, dc);
1752 }
1753 mutex_unlock(&dcc->cmd_lock);
1754
1755 return len;
1756}
1757
1758static unsigned int __wait_discard_cmd_range(struct f2fs_sb_info *sbi,
1759 struct discard_policy *dpolicy,
1760 block_t start, block_t end)
1761{
1762 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1763 struct list_head *wait_list = (dpolicy->type == DPOLICY_FSTRIM) ?
1764 &(dcc->fstrim_list) : &(dcc->wait_list);
1765 struct discard_cmd *dc = NULL, *iter, *tmp;
1766 unsigned int trimmed = 0;
1767
1768next:
1769 dc = NULL;
1770
1771 mutex_lock(&dcc->cmd_lock);
1772 list_for_each_entry_safe(iter, tmp, wait_list, list) {
1773 if (iter->di.lstart + iter->di.len <= start ||
1774 end <= iter->di.lstart)
1775 continue;
1776 if (iter->di.len < dpolicy->granularity)
1777 continue;
1778 if (iter->state == D_DONE && !iter->ref) {
1779 wait_for_completion_io(&iter->wait);
1780 if (!iter->error)
1781 trimmed += iter->di.len;
1782 __remove_discard_cmd(sbi, iter);
1783 } else {
1784 iter->ref++;
1785 dc = iter;
1786 break;
1787 }
1788 }
1789 mutex_unlock(&dcc->cmd_lock);
1790
1791 if (dc) {
1792 trimmed += __wait_one_discard_bio(sbi, dc);
1793 goto next;
1794 }
1795
1796 return trimmed;
1797}
1798
1799static unsigned int __wait_all_discard_cmd(struct f2fs_sb_info *sbi,
1800 struct discard_policy *dpolicy)
1801{
1802 struct discard_policy dp;
1803 unsigned int discard_blks;
1804
1805 if (dpolicy)
1806 return __wait_discard_cmd_range(sbi, dpolicy, 0, UINT_MAX);
1807
1808 /* wait all */
1809 __init_discard_policy(sbi, &dp, DPOLICY_FSTRIM, MIN_DISCARD_GRANULARITY);
1810 discard_blks = __wait_discard_cmd_range(sbi, &dp, 0, UINT_MAX);
1811 __init_discard_policy(sbi, &dp, DPOLICY_UMOUNT, MIN_DISCARD_GRANULARITY);
1812 discard_blks += __wait_discard_cmd_range(sbi, &dp, 0, UINT_MAX);
1813
1814 return discard_blks;
1815}
1816
1817/* This should be covered by global mutex, &sit_i->sentry_lock */
1818static void f2fs_wait_discard_bio(struct f2fs_sb_info *sbi, block_t blkaddr)
1819{
1820 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1821 struct discard_cmd *dc;
1822 bool need_wait = false;
1823
1824 mutex_lock(&dcc->cmd_lock);
1825 dc = __lookup_discard_cmd(sbi, blkaddr);
1826#ifdef CONFIG_BLK_DEV_ZONED
1827 if (dc && f2fs_sb_has_blkzoned(sbi) && bdev_is_zoned(dc->bdev)) {
1828 int devi = f2fs_bdev_index(sbi, dc->bdev);
1829
1830 if (devi < 0) {
1831 mutex_unlock(&dcc->cmd_lock);
1832 return;
1833 }
1834
1835 if (f2fs_blkz_is_seq(sbi, devi, dc->di.start)) {
1836 /* force submit zone reset */
1837 if (dc->state == D_PREP)
1838 __submit_zone_reset_cmd(sbi, dc, REQ_SYNC,
1839 &dcc->wait_list, NULL);
1840 dc->ref++;
1841 mutex_unlock(&dcc->cmd_lock);
1842 /* wait zone reset */
1843 __wait_one_discard_bio(sbi, dc);
1844 return;
1845 }
1846 }
1847#endif
1848 if (dc) {
1849 if (dc->state == D_PREP) {
1850 __punch_discard_cmd(sbi, dc, blkaddr);
1851 } else {
1852 dc->ref++;
1853 need_wait = true;
1854 }
1855 }
1856 mutex_unlock(&dcc->cmd_lock);
1857
1858 if (need_wait)
1859 __wait_one_discard_bio(sbi, dc);
1860}
1861
1862void f2fs_stop_discard_thread(struct f2fs_sb_info *sbi)
1863{
1864 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1865
1866 if (dcc && dcc->f2fs_issue_discard) {
1867 struct task_struct *discard_thread = dcc->f2fs_issue_discard;
1868
1869 dcc->f2fs_issue_discard = NULL;
1870 kthread_stop(discard_thread);
1871 }
1872}
1873
1874/**
1875 * f2fs_issue_discard_timeout() - Issue all discard cmd within UMOUNT_DISCARD_TIMEOUT
1876 * @sbi: the f2fs_sb_info data for discard cmd to issue
1877 *
1878 * When UMOUNT_DISCARD_TIMEOUT is exceeded, all remaining discard commands will be dropped
1879 *
1880 * Return true if issued all discard cmd or no discard cmd need issue, otherwise return false.
1881 */
1882bool f2fs_issue_discard_timeout(struct f2fs_sb_info *sbi)
1883{
1884 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1885 struct discard_policy dpolicy;
1886 bool dropped;
1887
1888 if (!atomic_read(&dcc->discard_cmd_cnt))
1889 return true;
1890
1891 __init_discard_policy(sbi, &dpolicy, DPOLICY_UMOUNT,
1892 dcc->discard_granularity);
1893 __issue_discard_cmd(sbi, &dpolicy);
1894 dropped = __drop_discard_cmd(sbi);
1895
1896 /* just to make sure there is no pending discard commands */
1897 __wait_all_discard_cmd(sbi, NULL);
1898
1899 f2fs_bug_on(sbi, atomic_read(&dcc->discard_cmd_cnt));
1900 return !dropped;
1901}
1902
1903static int issue_discard_thread(void *data)
1904{
1905 struct f2fs_sb_info *sbi = data;
1906 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1907 wait_queue_head_t *q = &dcc->discard_wait_queue;
1908 struct discard_policy dpolicy;
1909 unsigned int wait_ms = dcc->min_discard_issue_time;
1910 int issued;
1911
1912 set_freezable();
1913
1914 do {
1915 wait_event_freezable_timeout(*q,
1916 kthread_should_stop() || dcc->discard_wake,
1917 msecs_to_jiffies(wait_ms));
1918
1919 if (sbi->gc_mode == GC_URGENT_HIGH ||
1920 !f2fs_available_free_memory(sbi, DISCARD_CACHE))
1921 __init_discard_policy(sbi, &dpolicy, DPOLICY_FORCE,
1922 MIN_DISCARD_GRANULARITY);
1923 else
1924 __init_discard_policy(sbi, &dpolicy, DPOLICY_BG,
1925 dcc->discard_granularity);
1926
1927 if (dcc->discard_wake)
1928 dcc->discard_wake = false;
1929
1930 /* clean up pending candidates before going to sleep */
1931 if (atomic_read(&dcc->queued_discard))
1932 __wait_all_discard_cmd(sbi, NULL);
1933
1934 if (f2fs_readonly(sbi->sb))
1935 continue;
1936 if (kthread_should_stop())
1937 return 0;
1938 if (is_sbi_flag_set(sbi, SBI_NEED_FSCK) ||
1939 !atomic_read(&dcc->discard_cmd_cnt)) {
1940 wait_ms = dpolicy.max_interval;
1941 continue;
1942 }
1943
1944 sb_start_intwrite(sbi->sb);
1945
1946 issued = __issue_discard_cmd(sbi, &dpolicy);
1947 if (issued > 0) {
1948 __wait_all_discard_cmd(sbi, &dpolicy);
1949 wait_ms = dpolicy.min_interval;
1950 } else if (issued == -1) {
1951 wait_ms = f2fs_time_to_wait(sbi, DISCARD_TIME);
1952 if (!wait_ms)
1953 wait_ms = dpolicy.mid_interval;
1954 } else {
1955 wait_ms = dpolicy.max_interval;
1956 }
1957 if (!atomic_read(&dcc->discard_cmd_cnt))
1958 wait_ms = dpolicy.max_interval;
1959
1960 sb_end_intwrite(sbi->sb);
1961
1962 } while (!kthread_should_stop());
1963 return 0;
1964}
1965
1966#ifdef CONFIG_BLK_DEV_ZONED
1967static int __f2fs_issue_discard_zone(struct f2fs_sb_info *sbi,
1968 struct block_device *bdev, block_t blkstart, block_t blklen)
1969{
1970 sector_t sector, nr_sects;
1971 block_t lblkstart = blkstart;
1972 int devi = 0;
1973 u64 remainder = 0;
1974
1975 if (f2fs_is_multi_device(sbi)) {
1976 devi = f2fs_target_device_index(sbi, blkstart);
1977 if (blkstart < FDEV(devi).start_blk ||
1978 blkstart > FDEV(devi).end_blk) {
1979 f2fs_err(sbi, "Invalid block %x", blkstart);
1980 return -EIO;
1981 }
1982 blkstart -= FDEV(devi).start_blk;
1983 }
1984
1985 /* For sequential zones, reset the zone write pointer */
1986 if (f2fs_blkz_is_seq(sbi, devi, blkstart)) {
1987 sector = SECTOR_FROM_BLOCK(blkstart);
1988 nr_sects = SECTOR_FROM_BLOCK(blklen);
1989 div64_u64_rem(sector, bdev_zone_sectors(bdev), &remainder);
1990
1991 if (remainder || nr_sects != bdev_zone_sectors(bdev)) {
1992 f2fs_err(sbi, "(%d) %s: Unaligned zone reset attempted (block %x + %x)",
1993 devi, sbi->s_ndevs ? FDEV(devi).path : "",
1994 blkstart, blklen);
1995 return -EIO;
1996 }
1997
1998 if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING))) {
1999 unsigned int nofs_flags;
2000 int ret;
2001
2002 trace_f2fs_issue_reset_zone(bdev, blkstart);
2003 nofs_flags = memalloc_nofs_save();
2004 ret = blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
2005 sector, nr_sects);
2006 memalloc_nofs_restore(nofs_flags);
2007 return ret;
2008 }
2009
2010 __queue_zone_reset_cmd(sbi, bdev, blkstart, lblkstart, blklen);
2011 return 0;
2012 }
2013
2014 /* For conventional zones, use regular discard if supported */
2015 __queue_discard_cmd(sbi, bdev, lblkstart, blklen);
2016 return 0;
2017}
2018#endif
2019
2020static int __issue_discard_async(struct f2fs_sb_info *sbi,
2021 struct block_device *bdev, block_t blkstart, block_t blklen)
2022{
2023#ifdef CONFIG_BLK_DEV_ZONED
2024 if (f2fs_sb_has_blkzoned(sbi) && bdev_is_zoned(bdev))
2025 return __f2fs_issue_discard_zone(sbi, bdev, blkstart, blklen);
2026#endif
2027 __queue_discard_cmd(sbi, bdev, blkstart, blklen);
2028 return 0;
2029}
2030
2031static int f2fs_issue_discard(struct f2fs_sb_info *sbi,
2032 block_t blkstart, block_t blklen)
2033{
2034 sector_t start = blkstart, len = 0;
2035 struct block_device *bdev;
2036 struct seg_entry *se;
2037 unsigned int offset;
2038 block_t i;
2039 int err = 0;
2040
2041 bdev = f2fs_target_device(sbi, blkstart, NULL);
2042
2043 for (i = blkstart; i < blkstart + blklen; i++, len++) {
2044 if (i != start) {
2045 struct block_device *bdev2 =
2046 f2fs_target_device(sbi, i, NULL);
2047
2048 if (bdev2 != bdev) {
2049 err = __issue_discard_async(sbi, bdev,
2050 start, len);
2051 if (err)
2052 return err;
2053 bdev = bdev2;
2054 start = i;
2055 len = 0;
2056 }
2057 }
2058
2059 se = get_seg_entry(sbi, GET_SEGNO(sbi, i));
2060 offset = GET_BLKOFF_FROM_SEG0(sbi, i);
2061
2062 if (f2fs_block_unit_discard(sbi) &&
2063 !f2fs_test_and_set_bit(offset, se->discard_map))
2064 sbi->discard_blks--;
2065 }
2066
2067 if (len)
2068 err = __issue_discard_async(sbi, bdev, start, len);
2069 return err;
2070}
2071
2072static bool add_discard_addrs(struct f2fs_sb_info *sbi, struct cp_control *cpc,
2073 bool check_only)
2074{
2075 int entries = SIT_VBLOCK_MAP_SIZE / sizeof(unsigned long);
2076 struct seg_entry *se = get_seg_entry(sbi, cpc->trim_start);
2077 unsigned long *cur_map = (unsigned long *)se->cur_valid_map;
2078 unsigned long *ckpt_map = (unsigned long *)se->ckpt_valid_map;
2079 unsigned long *discard_map = (unsigned long *)se->discard_map;
2080 unsigned long *dmap = SIT_I(sbi)->tmp_map;
2081 unsigned int start = 0, end = -1;
2082 bool force = (cpc->reason & CP_DISCARD);
2083 struct discard_entry *de = NULL;
2084 struct list_head *head = &SM_I(sbi)->dcc_info->entry_list;
2085 int i;
2086
2087 if (se->valid_blocks == BLKS_PER_SEG(sbi) ||
2088 !f2fs_hw_support_discard(sbi) ||
2089 !f2fs_block_unit_discard(sbi))
2090 return false;
2091
2092 if (!force) {
2093 if (!f2fs_realtime_discard_enable(sbi) || !se->valid_blocks ||
2094 SM_I(sbi)->dcc_info->nr_discards >=
2095 SM_I(sbi)->dcc_info->max_discards)
2096 return false;
2097 }
2098
2099 /* SIT_VBLOCK_MAP_SIZE should be multiple of sizeof(unsigned long) */
2100 for (i = 0; i < entries; i++)
2101 dmap[i] = force ? ~ckpt_map[i] & ~discard_map[i] :
2102 (cur_map[i] ^ ckpt_map[i]) & ckpt_map[i];
2103
2104 while (force || SM_I(sbi)->dcc_info->nr_discards <=
2105 SM_I(sbi)->dcc_info->max_discards) {
2106 start = __find_rev_next_bit(dmap, BLKS_PER_SEG(sbi), end + 1);
2107 if (start >= BLKS_PER_SEG(sbi))
2108 break;
2109
2110 end = __find_rev_next_zero_bit(dmap,
2111 BLKS_PER_SEG(sbi), start + 1);
2112 if (force && start && end != BLKS_PER_SEG(sbi) &&
2113 (end - start) < cpc->trim_minlen)
2114 continue;
2115
2116 if (check_only)
2117 return true;
2118
2119 if (!de) {
2120 de = f2fs_kmem_cache_alloc(discard_entry_slab,
2121 GFP_F2FS_ZERO, true, NULL);
2122 de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start);
2123 list_add_tail(&de->list, head);
2124 }
2125
2126 for (i = start; i < end; i++)
2127 __set_bit_le(i, (void *)de->discard_map);
2128
2129 SM_I(sbi)->dcc_info->nr_discards += end - start;
2130 }
2131 return false;
2132}
2133
2134static void release_discard_addr(struct discard_entry *entry)
2135{
2136 list_del(&entry->list);
2137 kmem_cache_free(discard_entry_slab, entry);
2138}
2139
2140void f2fs_release_discard_addrs(struct f2fs_sb_info *sbi)
2141{
2142 struct list_head *head = &(SM_I(sbi)->dcc_info->entry_list);
2143 struct discard_entry *entry, *this;
2144
2145 /* drop caches */
2146 list_for_each_entry_safe(entry, this, head, list)
2147 release_discard_addr(entry);
2148}
2149
2150/*
2151 * Should call f2fs_clear_prefree_segments after checkpoint is done.
2152 */
2153static void set_prefree_as_free_segments(struct f2fs_sb_info *sbi)
2154{
2155 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
2156 unsigned int segno;
2157
2158 mutex_lock(&dirty_i->seglist_lock);
2159 for_each_set_bit(segno, dirty_i->dirty_segmap[PRE], MAIN_SEGS(sbi))
2160 __set_test_and_free(sbi, segno, false);
2161 mutex_unlock(&dirty_i->seglist_lock);
2162}
2163
2164void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
2165 struct cp_control *cpc)
2166{
2167 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
2168 struct list_head *head = &dcc->entry_list;
2169 struct discard_entry *entry, *this;
2170 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
2171 unsigned long *prefree_map = dirty_i->dirty_segmap[PRE];
2172 unsigned int start = 0, end = -1;
2173 unsigned int secno, start_segno;
2174 bool force = (cpc->reason & CP_DISCARD);
2175 bool section_alignment = F2FS_OPTION(sbi).discard_unit ==
2176 DISCARD_UNIT_SECTION;
2177
2178 if (f2fs_lfs_mode(sbi) && __is_large_section(sbi))
2179 section_alignment = true;
2180
2181 mutex_lock(&dirty_i->seglist_lock);
2182
2183 while (1) {
2184 int i;
2185
2186 if (section_alignment && end != -1)
2187 end--;
2188 start = find_next_bit(prefree_map, MAIN_SEGS(sbi), end + 1);
2189 if (start >= MAIN_SEGS(sbi))
2190 break;
2191 end = find_next_zero_bit(prefree_map, MAIN_SEGS(sbi),
2192 start + 1);
2193
2194 if (section_alignment) {
2195 start = rounddown(start, SEGS_PER_SEC(sbi));
2196 end = roundup(end, SEGS_PER_SEC(sbi));
2197 }
2198
2199 for (i = start; i < end; i++) {
2200 if (test_and_clear_bit(i, prefree_map))
2201 dirty_i->nr_dirty[PRE]--;
2202 }
2203
2204 if (!f2fs_realtime_discard_enable(sbi))
2205 continue;
2206
2207 if (force && start >= cpc->trim_start &&
2208 (end - 1) <= cpc->trim_end)
2209 continue;
2210
2211 /* Should cover 2MB zoned device for zone-based reset */
2212 if (!f2fs_sb_has_blkzoned(sbi) &&
2213 (!f2fs_lfs_mode(sbi) || !__is_large_section(sbi))) {
2214 f2fs_issue_discard(sbi, START_BLOCK(sbi, start),
2215 SEGS_TO_BLKS(sbi, end - start));
2216 continue;
2217 }
2218next:
2219 secno = GET_SEC_FROM_SEG(sbi, start);
2220 start_segno = GET_SEG_FROM_SEC(sbi, secno);
2221 if (!IS_CURSEC(sbi, secno) &&
2222 !get_valid_blocks(sbi, start, true))
2223 f2fs_issue_discard(sbi, START_BLOCK(sbi, start_segno),
2224 BLKS_PER_SEC(sbi));
2225
2226 start = start_segno + SEGS_PER_SEC(sbi);
2227 if (start < end)
2228 goto next;
2229 else
2230 end = start - 1;
2231 }
2232 mutex_unlock(&dirty_i->seglist_lock);
2233
2234 if (!f2fs_block_unit_discard(sbi))
2235 goto wakeup;
2236
2237 /* send small discards */
2238 list_for_each_entry_safe(entry, this, head, list) {
2239 unsigned int cur_pos = 0, next_pos, len, total_len = 0;
2240 bool is_valid = test_bit_le(0, entry->discard_map);
2241
2242find_next:
2243 if (is_valid) {
2244 next_pos = find_next_zero_bit_le(entry->discard_map,
2245 BLKS_PER_SEG(sbi), cur_pos);
2246 len = next_pos - cur_pos;
2247
2248 if (f2fs_sb_has_blkzoned(sbi) ||
2249 (force && len < cpc->trim_minlen))
2250 goto skip;
2251
2252 f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
2253 len);
2254 total_len += len;
2255 } else {
2256 next_pos = find_next_bit_le(entry->discard_map,
2257 BLKS_PER_SEG(sbi), cur_pos);
2258 }
2259skip:
2260 cur_pos = next_pos;
2261 is_valid = !is_valid;
2262
2263 if (cur_pos < BLKS_PER_SEG(sbi))
2264 goto find_next;
2265
2266 release_discard_addr(entry);
2267 dcc->nr_discards -= total_len;
2268 }
2269
2270wakeup:
2271 wake_up_discard_thread(sbi, false);
2272}
2273
2274int f2fs_start_discard_thread(struct f2fs_sb_info *sbi)
2275{
2276 dev_t dev = sbi->sb->s_bdev->bd_dev;
2277 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
2278 int err = 0;
2279
2280 if (f2fs_sb_has_readonly(sbi)) {
2281 f2fs_info(sbi,
2282 "Skip to start discard thread for readonly image");
2283 return 0;
2284 }
2285
2286 if (!f2fs_realtime_discard_enable(sbi))
2287 return 0;
2288
2289 dcc->f2fs_issue_discard = kthread_run(issue_discard_thread, sbi,
2290 "f2fs_discard-%u:%u", MAJOR(dev), MINOR(dev));
2291 if (IS_ERR(dcc->f2fs_issue_discard)) {
2292 err = PTR_ERR(dcc->f2fs_issue_discard);
2293 dcc->f2fs_issue_discard = NULL;
2294 }
2295
2296 return err;
2297}
2298
2299static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
2300{
2301 struct discard_cmd_control *dcc;
2302 int err = 0, i;
2303
2304 if (SM_I(sbi)->dcc_info) {
2305 dcc = SM_I(sbi)->dcc_info;
2306 goto init_thread;
2307 }
2308
2309 dcc = f2fs_kzalloc(sbi, sizeof(struct discard_cmd_control), GFP_KERNEL);
2310 if (!dcc)
2311 return -ENOMEM;
2312
2313 dcc->discard_io_aware_gran = MAX_PLIST_NUM;
2314 dcc->discard_granularity = DEFAULT_DISCARD_GRANULARITY;
2315 dcc->max_ordered_discard = DEFAULT_MAX_ORDERED_DISCARD_GRANULARITY;
2316 dcc->discard_io_aware = DPOLICY_IO_AWARE_ENABLE;
2317 if (F2FS_OPTION(sbi).discard_unit == DISCARD_UNIT_SEGMENT)
2318 dcc->discard_granularity = BLKS_PER_SEG(sbi);
2319 else if (F2FS_OPTION(sbi).discard_unit == DISCARD_UNIT_SECTION)
2320 dcc->discard_granularity = BLKS_PER_SEC(sbi);
2321
2322 INIT_LIST_HEAD(&dcc->entry_list);
2323 for (i = 0; i < MAX_PLIST_NUM; i++)
2324 INIT_LIST_HEAD(&dcc->pend_list[i]);
2325 INIT_LIST_HEAD(&dcc->wait_list);
2326 INIT_LIST_HEAD(&dcc->fstrim_list);
2327 mutex_init(&dcc->cmd_lock);
2328 atomic_set(&dcc->issued_discard, 0);
2329 atomic_set(&dcc->queued_discard, 0);
2330 atomic_set(&dcc->discard_cmd_cnt, 0);
2331 dcc->nr_discards = 0;
2332 dcc->max_discards = SEGS_TO_BLKS(sbi, MAIN_SEGS(sbi));
2333 dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
2334 dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
2335 dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
2336 dcc->max_discard_issue_time = DEF_MAX_DISCARD_ISSUE_TIME;
2337 dcc->discard_urgent_util = DEF_DISCARD_URGENT_UTIL;
2338 dcc->undiscard_blks = 0;
2339 dcc->next_pos = 0;
2340 dcc->root = RB_ROOT_CACHED;
2341 dcc->rbtree_check = false;
2342
2343 init_waitqueue_head(&dcc->discard_wait_queue);
2344 SM_I(sbi)->dcc_info = dcc;
2345init_thread:
2346 err = f2fs_start_discard_thread(sbi);
2347 if (err) {
2348 kfree(dcc);
2349 SM_I(sbi)->dcc_info = NULL;
2350 }
2351
2352 return err;
2353}
2354
2355static void destroy_discard_cmd_control(struct f2fs_sb_info *sbi)
2356{
2357 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
2358
2359 if (!dcc)
2360 return;
2361
2362 f2fs_stop_discard_thread(sbi);
2363
2364 /*
2365 * Recovery can cache discard commands, so in error path of
2366 * fill_super(), it needs to give a chance to handle them.
2367 */
2368 f2fs_issue_discard_timeout(sbi);
2369
2370 kfree(dcc);
2371 SM_I(sbi)->dcc_info = NULL;
2372}
2373
2374static bool __mark_sit_entry_dirty(struct f2fs_sb_info *sbi, unsigned int segno)
2375{
2376 struct sit_info *sit_i = SIT_I(sbi);
2377
2378 if (!__test_and_set_bit(segno, sit_i->dirty_sentries_bitmap)) {
2379 sit_i->dirty_sentries++;
2380 return false;
2381 }
2382
2383 return true;
2384}
2385
2386static void __set_sit_entry_type(struct f2fs_sb_info *sbi, int type,
2387 unsigned int segno, int modified)
2388{
2389 struct seg_entry *se = get_seg_entry(sbi, segno);
2390
2391 se->type = type;
2392 if (modified)
2393 __mark_sit_entry_dirty(sbi, segno);
2394}
2395
2396static inline unsigned long long get_segment_mtime(struct f2fs_sb_info *sbi,
2397 block_t blkaddr)
2398{
2399 unsigned int segno = GET_SEGNO(sbi, blkaddr);
2400
2401 if (segno == NULL_SEGNO)
2402 return 0;
2403 return get_seg_entry(sbi, segno)->mtime;
2404}
2405
2406static void update_segment_mtime(struct f2fs_sb_info *sbi, block_t blkaddr,
2407 unsigned long long old_mtime)
2408{
2409 struct seg_entry *se;
2410 unsigned int segno = GET_SEGNO(sbi, blkaddr);
2411 unsigned long long ctime = get_mtime(sbi, false);
2412 unsigned long long mtime = old_mtime ? old_mtime : ctime;
2413
2414 if (segno == NULL_SEGNO)
2415 return;
2416
2417 se = get_seg_entry(sbi, segno);
2418
2419 if (!se->mtime)
2420 se->mtime = mtime;
2421 else
2422 se->mtime = div_u64(se->mtime * se->valid_blocks + mtime,
2423 se->valid_blocks + 1);
2424
2425 if (ctime > SIT_I(sbi)->max_mtime)
2426 SIT_I(sbi)->max_mtime = ctime;
2427}
2428
2429static void update_sit_entry(struct f2fs_sb_info *sbi, block_t blkaddr, int del)
2430{
2431 struct seg_entry *se;
2432 unsigned int segno, offset;
2433 long int new_vblocks;
2434 bool exist;
2435#ifdef CONFIG_F2FS_CHECK_FS
2436 bool mir_exist;
2437#endif
2438
2439 segno = GET_SEGNO(sbi, blkaddr);
2440 if (segno == NULL_SEGNO)
2441 return;
2442
2443 se = get_seg_entry(sbi, segno);
2444 new_vblocks = se->valid_blocks + del;
2445 offset = GET_BLKOFF_FROM_SEG0(sbi, blkaddr);
2446
2447 f2fs_bug_on(sbi, (new_vblocks < 0 ||
2448 (new_vblocks > f2fs_usable_blks_in_seg(sbi, segno))));
2449
2450 se->valid_blocks = new_vblocks;
2451
2452 /* Update valid block bitmap */
2453 if (del > 0) {
2454 exist = f2fs_test_and_set_bit(offset, se->cur_valid_map);
2455#ifdef CONFIG_F2FS_CHECK_FS
2456 mir_exist = f2fs_test_and_set_bit(offset,
2457 se->cur_valid_map_mir);
2458 if (unlikely(exist != mir_exist)) {
2459 f2fs_err(sbi, "Inconsistent error when setting bitmap, blk:%u, old bit:%d",
2460 blkaddr, exist);
2461 f2fs_bug_on(sbi, 1);
2462 }
2463#endif
2464 if (unlikely(exist)) {
2465 f2fs_err(sbi, "Bitmap was wrongly set, blk:%u",
2466 blkaddr);
2467 f2fs_bug_on(sbi, 1);
2468 se->valid_blocks--;
2469 del = 0;
2470 }
2471
2472 if (f2fs_block_unit_discard(sbi) &&
2473 !f2fs_test_and_set_bit(offset, se->discard_map))
2474 sbi->discard_blks--;
2475
2476 /*
2477 * SSR should never reuse block which is checkpointed
2478 * or newly invalidated.
2479 */
2480 if (!is_sbi_flag_set(sbi, SBI_CP_DISABLED)) {
2481 if (!f2fs_test_and_set_bit(offset, se->ckpt_valid_map))
2482 se->ckpt_valid_blocks++;
2483 }
2484 } else {
2485 exist = f2fs_test_and_clear_bit(offset, se->cur_valid_map);
2486#ifdef CONFIG_F2FS_CHECK_FS
2487 mir_exist = f2fs_test_and_clear_bit(offset,
2488 se->cur_valid_map_mir);
2489 if (unlikely(exist != mir_exist)) {
2490 f2fs_err(sbi, "Inconsistent error when clearing bitmap, blk:%u, old bit:%d",
2491 blkaddr, exist);
2492 f2fs_bug_on(sbi, 1);
2493 }
2494#endif
2495 if (unlikely(!exist)) {
2496 f2fs_err(sbi, "Bitmap was wrongly cleared, blk:%u",
2497 blkaddr);
2498 f2fs_bug_on(sbi, 1);
2499 se->valid_blocks++;
2500 del = 0;
2501 } else if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED))) {
2502 /*
2503 * If checkpoints are off, we must not reuse data that
2504 * was used in the previous checkpoint. If it was used
2505 * before, we must track that to know how much space we
2506 * really have.
2507 */
2508 if (f2fs_test_bit(offset, se->ckpt_valid_map)) {
2509 spin_lock(&sbi->stat_lock);
2510 sbi->unusable_block_count++;
2511 spin_unlock(&sbi->stat_lock);
2512 }
2513 }
2514
2515 if (f2fs_block_unit_discard(sbi) &&
2516 f2fs_test_and_clear_bit(offset, se->discard_map))
2517 sbi->discard_blks++;
2518 }
2519 if (!f2fs_test_bit(offset, se->ckpt_valid_map))
2520 se->ckpt_valid_blocks += del;
2521
2522 __mark_sit_entry_dirty(sbi, segno);
2523
2524 /* update total number of valid blocks to be written in ckpt area */
2525 SIT_I(sbi)->written_valid_blocks += del;
2526
2527 if (__is_large_section(sbi))
2528 get_sec_entry(sbi, segno)->valid_blocks += del;
2529}
2530
2531void f2fs_invalidate_blocks(struct f2fs_sb_info *sbi, block_t addr)
2532{
2533 unsigned int segno = GET_SEGNO(sbi, addr);
2534 struct sit_info *sit_i = SIT_I(sbi);
2535
2536 f2fs_bug_on(sbi, addr == NULL_ADDR);
2537 if (addr == NEW_ADDR || addr == COMPRESS_ADDR)
2538 return;
2539
2540 f2fs_invalidate_internal_cache(sbi, addr);
2541
2542 /* add it into sit main buffer */
2543 down_write(&sit_i->sentry_lock);
2544
2545 update_segment_mtime(sbi, addr, 0);
2546 update_sit_entry(sbi, addr, -1);
2547
2548 /* add it into dirty seglist */
2549 locate_dirty_segment(sbi, segno);
2550
2551 up_write(&sit_i->sentry_lock);
2552}
2553
2554bool f2fs_is_checkpointed_data(struct f2fs_sb_info *sbi, block_t blkaddr)
2555{
2556 struct sit_info *sit_i = SIT_I(sbi);
2557 unsigned int segno, offset;
2558 struct seg_entry *se;
2559 bool is_cp = false;
2560
2561 if (!__is_valid_data_blkaddr(blkaddr))
2562 return true;
2563
2564 down_read(&sit_i->sentry_lock);
2565
2566 segno = GET_SEGNO(sbi, blkaddr);
2567 se = get_seg_entry(sbi, segno);
2568 offset = GET_BLKOFF_FROM_SEG0(sbi, blkaddr);
2569
2570 if (f2fs_test_bit(offset, se->ckpt_valid_map))
2571 is_cp = true;
2572
2573 up_read(&sit_i->sentry_lock);
2574
2575 return is_cp;
2576}
2577
2578static unsigned short f2fs_curseg_valid_blocks(struct f2fs_sb_info *sbi, int type)
2579{
2580 struct curseg_info *curseg = CURSEG_I(sbi, type);
2581
2582 if (sbi->ckpt->alloc_type[type] == SSR)
2583 return BLKS_PER_SEG(sbi);
2584 return curseg->next_blkoff;
2585}
2586
2587/*
2588 * Calculate the number of current summary pages for writing
2589 */
2590int f2fs_npages_for_summary_flush(struct f2fs_sb_info *sbi, bool for_ra)
2591{
2592 int valid_sum_count = 0;
2593 int i, sum_in_page;
2594
2595 for (i = CURSEG_HOT_DATA; i <= CURSEG_COLD_DATA; i++) {
2596 if (sbi->ckpt->alloc_type[i] != SSR && for_ra)
2597 valid_sum_count +=
2598 le16_to_cpu(F2FS_CKPT(sbi)->cur_data_blkoff[i]);
2599 else
2600 valid_sum_count += f2fs_curseg_valid_blocks(sbi, i);
2601 }
2602
2603 sum_in_page = (PAGE_SIZE - 2 * SUM_JOURNAL_SIZE -
2604 SUM_FOOTER_SIZE) / SUMMARY_SIZE;
2605 if (valid_sum_count <= sum_in_page)
2606 return 1;
2607 else if ((valid_sum_count - sum_in_page) <=
2608 (PAGE_SIZE - SUM_FOOTER_SIZE) / SUMMARY_SIZE)
2609 return 2;
2610 return 3;
2611}
2612
2613/*
2614 * Caller should put this summary page
2615 */
2616struct page *f2fs_get_sum_page(struct f2fs_sb_info *sbi, unsigned int segno)
2617{
2618 if (unlikely(f2fs_cp_error(sbi)))
2619 return ERR_PTR(-EIO);
2620 return f2fs_get_meta_page_retry(sbi, GET_SUM_BLOCK(sbi, segno));
2621}
2622
2623void f2fs_update_meta_page(struct f2fs_sb_info *sbi,
2624 void *src, block_t blk_addr)
2625{
2626 struct page *page = f2fs_grab_meta_page(sbi, blk_addr);
2627
2628 memcpy(page_address(page), src, PAGE_SIZE);
2629 set_page_dirty(page);
2630 f2fs_put_page(page, 1);
2631}
2632
2633static void write_sum_page(struct f2fs_sb_info *sbi,
2634 struct f2fs_summary_block *sum_blk, block_t blk_addr)
2635{
2636 f2fs_update_meta_page(sbi, (void *)sum_blk, blk_addr);
2637}
2638
2639static void write_current_sum_page(struct f2fs_sb_info *sbi,
2640 int type, block_t blk_addr)
2641{
2642 struct curseg_info *curseg = CURSEG_I(sbi, type);
2643 struct page *page = f2fs_grab_meta_page(sbi, blk_addr);
2644 struct f2fs_summary_block *src = curseg->sum_blk;
2645 struct f2fs_summary_block *dst;
2646
2647 dst = (struct f2fs_summary_block *)page_address(page);
2648 memset(dst, 0, PAGE_SIZE);
2649
2650 mutex_lock(&curseg->curseg_mutex);
2651
2652 down_read(&curseg->journal_rwsem);
2653 memcpy(&dst->journal, curseg->journal, SUM_JOURNAL_SIZE);
2654 up_read(&curseg->journal_rwsem);
2655
2656 memcpy(dst->entries, src->entries, SUM_ENTRY_SIZE);
2657 memcpy(&dst->footer, &src->footer, SUM_FOOTER_SIZE);
2658
2659 mutex_unlock(&curseg->curseg_mutex);
2660
2661 set_page_dirty(page);
2662 f2fs_put_page(page, 1);
2663}
2664
2665static int is_next_segment_free(struct f2fs_sb_info *sbi,
2666 struct curseg_info *curseg)
2667{
2668 unsigned int segno = curseg->segno + 1;
2669 struct free_segmap_info *free_i = FREE_I(sbi);
2670
2671 if (segno < MAIN_SEGS(sbi) && segno % SEGS_PER_SEC(sbi))
2672 return !test_bit(segno, free_i->free_segmap);
2673 return 0;
2674}
2675
2676/*
2677 * Find a new segment from the free segments bitmap to right order
2678 * This function should be returned with success, otherwise BUG
2679 */
2680static int get_new_segment(struct f2fs_sb_info *sbi,
2681 unsigned int *newseg, bool new_sec, bool pinning)
2682{
2683 struct free_segmap_info *free_i = FREE_I(sbi);
2684 unsigned int segno, secno, zoneno;
2685 unsigned int total_zones = MAIN_SECS(sbi) / sbi->secs_per_zone;
2686 unsigned int hint = GET_SEC_FROM_SEG(sbi, *newseg);
2687 unsigned int old_zoneno = GET_ZONE_FROM_SEG(sbi, *newseg);
2688 bool init = true;
2689 int i;
2690 int ret = 0;
2691
2692 spin_lock(&free_i->segmap_lock);
2693
2694 if (time_to_inject(sbi, FAULT_NO_SEGMENT)) {
2695 ret = -ENOSPC;
2696 goto out_unlock;
2697 }
2698
2699 if (!new_sec && ((*newseg + 1) % SEGS_PER_SEC(sbi))) {
2700 segno = find_next_zero_bit(free_i->free_segmap,
2701 GET_SEG_FROM_SEC(sbi, hint + 1), *newseg + 1);
2702 if (segno < GET_SEG_FROM_SEC(sbi, hint + 1))
2703 goto got_it;
2704 }
2705
2706#ifdef CONFIG_BLK_DEV_ZONED
2707 /*
2708 * If we format f2fs on zoned storage, let's try to get pinned sections
2709 * from beginning of the storage, which should be a conventional one.
2710 */
2711 if (f2fs_sb_has_blkzoned(sbi)) {
2712 /* Prioritize writing to conventional zones */
2713 if (sbi->blkzone_alloc_policy == BLKZONE_ALLOC_PRIOR_CONV || pinning)
2714 segno = 0;
2715 else
2716 segno = max(sbi->first_zoned_segno, *newseg);
2717 hint = GET_SEC_FROM_SEG(sbi, segno);
2718 }
2719#endif
2720
2721find_other_zone:
2722 secno = find_next_zero_bit(free_i->free_secmap, MAIN_SECS(sbi), hint);
2723
2724#ifdef CONFIG_BLK_DEV_ZONED
2725 if (secno >= MAIN_SECS(sbi) && f2fs_sb_has_blkzoned(sbi)) {
2726 /* Write only to sequential zones */
2727 if (sbi->blkzone_alloc_policy == BLKZONE_ALLOC_ONLY_SEQ) {
2728 hint = GET_SEC_FROM_SEG(sbi, sbi->first_zoned_segno);
2729 secno = find_next_zero_bit(free_i->free_secmap, MAIN_SECS(sbi), hint);
2730 } else
2731 secno = find_first_zero_bit(free_i->free_secmap,
2732 MAIN_SECS(sbi));
2733 if (secno >= MAIN_SECS(sbi)) {
2734 ret = -ENOSPC;
2735 f2fs_bug_on(sbi, 1);
2736 goto out_unlock;
2737 }
2738 }
2739#endif
2740
2741 if (secno >= MAIN_SECS(sbi)) {
2742 secno = find_first_zero_bit(free_i->free_secmap,
2743 MAIN_SECS(sbi));
2744 if (secno >= MAIN_SECS(sbi)) {
2745 ret = -ENOSPC;
2746 f2fs_bug_on(sbi, 1);
2747 goto out_unlock;
2748 }
2749 }
2750 segno = GET_SEG_FROM_SEC(sbi, secno);
2751 zoneno = GET_ZONE_FROM_SEC(sbi, secno);
2752
2753 /* give up on finding another zone */
2754 if (!init)
2755 goto got_it;
2756 if (sbi->secs_per_zone == 1)
2757 goto got_it;
2758 if (zoneno == old_zoneno)
2759 goto got_it;
2760 for (i = 0; i < NR_CURSEG_TYPE; i++)
2761 if (CURSEG_I(sbi, i)->zone == zoneno)
2762 break;
2763
2764 if (i < NR_CURSEG_TYPE) {
2765 /* zone is in user, try another */
2766 if (zoneno + 1 >= total_zones)
2767 hint = 0;
2768 else
2769 hint = (zoneno + 1) * sbi->secs_per_zone;
2770 init = false;
2771 goto find_other_zone;
2772 }
2773got_it:
2774 /* set it as dirty segment in free segmap */
2775 f2fs_bug_on(sbi, test_bit(segno, free_i->free_segmap));
2776
2777 /* no free section in conventional zone */
2778 if (new_sec && pinning &&
2779 !f2fs_valid_pinned_area(sbi, START_BLOCK(sbi, segno))) {
2780 ret = -EAGAIN;
2781 goto out_unlock;
2782 }
2783 __set_inuse(sbi, segno);
2784 *newseg = segno;
2785out_unlock:
2786 spin_unlock(&free_i->segmap_lock);
2787
2788 if (ret == -ENOSPC)
2789 f2fs_stop_checkpoint(sbi, false, STOP_CP_REASON_NO_SEGMENT);
2790 return ret;
2791}
2792
2793static void reset_curseg(struct f2fs_sb_info *sbi, int type, int modified)
2794{
2795 struct curseg_info *curseg = CURSEG_I(sbi, type);
2796 struct summary_footer *sum_footer;
2797 unsigned short seg_type = curseg->seg_type;
2798
2799 /* only happen when get_new_segment() fails */
2800 if (curseg->next_segno == NULL_SEGNO)
2801 return;
2802
2803 curseg->inited = true;
2804 curseg->segno = curseg->next_segno;
2805 curseg->zone = GET_ZONE_FROM_SEG(sbi, curseg->segno);
2806 curseg->next_blkoff = 0;
2807 curseg->next_segno = NULL_SEGNO;
2808
2809 sum_footer = &(curseg->sum_blk->footer);
2810 memset(sum_footer, 0, sizeof(struct summary_footer));
2811
2812 sanity_check_seg_type(sbi, seg_type);
2813
2814 if (IS_DATASEG(seg_type))
2815 SET_SUM_TYPE(sum_footer, SUM_TYPE_DATA);
2816 if (IS_NODESEG(seg_type))
2817 SET_SUM_TYPE(sum_footer, SUM_TYPE_NODE);
2818 __set_sit_entry_type(sbi, seg_type, curseg->segno, modified);
2819}
2820
2821static unsigned int __get_next_segno(struct f2fs_sb_info *sbi, int type)
2822{
2823 struct curseg_info *curseg = CURSEG_I(sbi, type);
2824 unsigned short seg_type = curseg->seg_type;
2825
2826 sanity_check_seg_type(sbi, seg_type);
2827 if (__is_large_section(sbi)) {
2828 if (f2fs_need_rand_seg(sbi)) {
2829 unsigned int hint = GET_SEC_FROM_SEG(sbi, curseg->segno);
2830
2831 if (GET_SEC_FROM_SEG(sbi, curseg->segno + 1) != hint)
2832 return curseg->segno;
2833 return get_random_u32_inclusive(curseg->segno + 1,
2834 GET_SEG_FROM_SEC(sbi, hint + 1) - 1);
2835 }
2836 return curseg->segno;
2837 } else if (f2fs_need_rand_seg(sbi)) {
2838 return get_random_u32_below(MAIN_SECS(sbi) * SEGS_PER_SEC(sbi));
2839 }
2840
2841 /* inmem log may not locate on any segment after mount */
2842 if (!curseg->inited)
2843 return 0;
2844
2845 if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED)))
2846 return 0;
2847
2848 if (seg_type == CURSEG_HOT_DATA || IS_NODESEG(seg_type))
2849 return 0;
2850
2851 if (SIT_I(sbi)->last_victim[ALLOC_NEXT])
2852 return SIT_I(sbi)->last_victim[ALLOC_NEXT];
2853
2854 /* find segments from 0 to reuse freed segments */
2855 if (F2FS_OPTION(sbi).alloc_mode == ALLOC_MODE_REUSE)
2856 return 0;
2857
2858 return curseg->segno;
2859}
2860
2861/*
2862 * Allocate a current working segment.
2863 * This function always allocates a free segment in LFS manner.
2864 */
2865static int new_curseg(struct f2fs_sb_info *sbi, int type, bool new_sec)
2866{
2867 struct curseg_info *curseg = CURSEG_I(sbi, type);
2868 unsigned int segno = curseg->segno;
2869 bool pinning = type == CURSEG_COLD_DATA_PINNED;
2870 int ret;
2871
2872 if (curseg->inited)
2873 write_sum_page(sbi, curseg->sum_blk, GET_SUM_BLOCK(sbi, segno));
2874
2875 segno = __get_next_segno(sbi, type);
2876 ret = get_new_segment(sbi, &segno, new_sec, pinning);
2877 if (ret) {
2878 if (ret == -ENOSPC)
2879 curseg->segno = NULL_SEGNO;
2880 return ret;
2881 }
2882
2883 curseg->next_segno = segno;
2884 reset_curseg(sbi, type, 1);
2885 curseg->alloc_type = LFS;
2886 if (F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_BLK)
2887 curseg->fragment_remained_chunk =
2888 get_random_u32_inclusive(1, sbi->max_fragment_chunk);
2889 return 0;
2890}
2891
2892static int __next_free_blkoff(struct f2fs_sb_info *sbi,
2893 int segno, block_t start)
2894{
2895 struct seg_entry *se = get_seg_entry(sbi, segno);
2896 int entries = SIT_VBLOCK_MAP_SIZE / sizeof(unsigned long);
2897 unsigned long *target_map = SIT_I(sbi)->tmp_map;
2898 unsigned long *ckpt_map = (unsigned long *)se->ckpt_valid_map;
2899 unsigned long *cur_map = (unsigned long *)se->cur_valid_map;
2900 int i;
2901
2902 for (i = 0; i < entries; i++)
2903 target_map[i] = ckpt_map[i] | cur_map[i];
2904
2905 return __find_rev_next_zero_bit(target_map, BLKS_PER_SEG(sbi), start);
2906}
2907
2908static int f2fs_find_next_ssr_block(struct f2fs_sb_info *sbi,
2909 struct curseg_info *seg)
2910{
2911 return __next_free_blkoff(sbi, seg->segno, seg->next_blkoff + 1);
2912}
2913
2914bool f2fs_segment_has_free_slot(struct f2fs_sb_info *sbi, int segno)
2915{
2916 return __next_free_blkoff(sbi, segno, 0) < BLKS_PER_SEG(sbi);
2917}
2918
2919/*
2920 * This function always allocates a used segment(from dirty seglist) by SSR
2921 * manner, so it should recover the existing segment information of valid blocks
2922 */
2923static int change_curseg(struct f2fs_sb_info *sbi, int type)
2924{
2925 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
2926 struct curseg_info *curseg = CURSEG_I(sbi, type);
2927 unsigned int new_segno = curseg->next_segno;
2928 struct f2fs_summary_block *sum_node;
2929 struct page *sum_page;
2930
2931 if (curseg->inited)
2932 write_sum_page(sbi, curseg->sum_blk, GET_SUM_BLOCK(sbi, curseg->segno));
2933
2934 __set_test_and_inuse(sbi, new_segno);
2935
2936 mutex_lock(&dirty_i->seglist_lock);
2937 __remove_dirty_segment(sbi, new_segno, PRE);
2938 __remove_dirty_segment(sbi, new_segno, DIRTY);
2939 mutex_unlock(&dirty_i->seglist_lock);
2940
2941 reset_curseg(sbi, type, 1);
2942 curseg->alloc_type = SSR;
2943 curseg->next_blkoff = __next_free_blkoff(sbi, curseg->segno, 0);
2944
2945 sum_page = f2fs_get_sum_page(sbi, new_segno);
2946 if (IS_ERR(sum_page)) {
2947 /* GC won't be able to use stale summary pages by cp_error */
2948 memset(curseg->sum_blk, 0, SUM_ENTRY_SIZE);
2949 return PTR_ERR(sum_page);
2950 }
2951 sum_node = (struct f2fs_summary_block *)page_address(sum_page);
2952 memcpy(curseg->sum_blk, sum_node, SUM_ENTRY_SIZE);
2953 f2fs_put_page(sum_page, 1);
2954 return 0;
2955}
2956
2957static int get_ssr_segment(struct f2fs_sb_info *sbi, int type,
2958 int alloc_mode, unsigned long long age);
2959
2960static int get_atssr_segment(struct f2fs_sb_info *sbi, int type,
2961 int target_type, int alloc_mode,
2962 unsigned long long age)
2963{
2964 struct curseg_info *curseg = CURSEG_I(sbi, type);
2965 int ret = 0;
2966
2967 curseg->seg_type = target_type;
2968
2969 if (get_ssr_segment(sbi, type, alloc_mode, age)) {
2970 struct seg_entry *se = get_seg_entry(sbi, curseg->next_segno);
2971
2972 curseg->seg_type = se->type;
2973 ret = change_curseg(sbi, type);
2974 } else {
2975 /* allocate cold segment by default */
2976 curseg->seg_type = CURSEG_COLD_DATA;
2977 ret = new_curseg(sbi, type, true);
2978 }
2979 stat_inc_seg_type(sbi, curseg);
2980 return ret;
2981}
2982
2983static int __f2fs_init_atgc_curseg(struct f2fs_sb_info *sbi, bool force)
2984{
2985 struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_ALL_DATA_ATGC);
2986 int ret = 0;
2987
2988 if (!sbi->am.atgc_enabled && !force)
2989 return 0;
2990
2991 f2fs_down_read(&SM_I(sbi)->curseg_lock);
2992
2993 mutex_lock(&curseg->curseg_mutex);
2994 down_write(&SIT_I(sbi)->sentry_lock);
2995
2996 ret = get_atssr_segment(sbi, CURSEG_ALL_DATA_ATGC,
2997 CURSEG_COLD_DATA, SSR, 0);
2998
2999 up_write(&SIT_I(sbi)->sentry_lock);
3000 mutex_unlock(&curseg->curseg_mutex);
3001
3002 f2fs_up_read(&SM_I(sbi)->curseg_lock);
3003 return ret;
3004}
3005
3006int f2fs_init_inmem_curseg(struct f2fs_sb_info *sbi)
3007{
3008 return __f2fs_init_atgc_curseg(sbi, false);
3009}
3010
3011int f2fs_reinit_atgc_curseg(struct f2fs_sb_info *sbi)
3012{
3013 int ret;
3014
3015 if (!test_opt(sbi, ATGC))
3016 return 0;
3017 if (sbi->am.atgc_enabled)
3018 return 0;
3019 if (le64_to_cpu(F2FS_CKPT(sbi)->elapsed_time) <
3020 sbi->am.age_threshold)
3021 return 0;
3022
3023 ret = __f2fs_init_atgc_curseg(sbi, true);
3024 if (!ret) {
3025 sbi->am.atgc_enabled = true;
3026 f2fs_info(sbi, "reenabled age threshold GC");
3027 }
3028 return ret;
3029}
3030
3031static void __f2fs_save_inmem_curseg(struct f2fs_sb_info *sbi, int type)
3032{
3033 struct curseg_info *curseg = CURSEG_I(sbi, type);
3034
3035 mutex_lock(&curseg->curseg_mutex);
3036 if (!curseg->inited)
3037 goto out;
3038
3039 if (get_valid_blocks(sbi, curseg->segno, false)) {
3040 write_sum_page(sbi, curseg->sum_blk,
3041 GET_SUM_BLOCK(sbi, curseg->segno));
3042 } else {
3043 mutex_lock(&DIRTY_I(sbi)->seglist_lock);
3044 __set_test_and_free(sbi, curseg->segno, true);
3045 mutex_unlock(&DIRTY_I(sbi)->seglist_lock);
3046 }
3047out:
3048 mutex_unlock(&curseg->curseg_mutex);
3049}
3050
3051void f2fs_save_inmem_curseg(struct f2fs_sb_info *sbi)
3052{
3053 __f2fs_save_inmem_curseg(sbi, CURSEG_COLD_DATA_PINNED);
3054
3055 if (sbi->am.atgc_enabled)
3056 __f2fs_save_inmem_curseg(sbi, CURSEG_ALL_DATA_ATGC);
3057}
3058
3059static void __f2fs_restore_inmem_curseg(struct f2fs_sb_info *sbi, int type)
3060{
3061 struct curseg_info *curseg = CURSEG_I(sbi, type);
3062
3063 mutex_lock(&curseg->curseg_mutex);
3064 if (!curseg->inited)
3065 goto out;
3066 if (get_valid_blocks(sbi, curseg->segno, false))
3067 goto out;
3068
3069 mutex_lock(&DIRTY_I(sbi)->seglist_lock);
3070 __set_test_and_inuse(sbi, curseg->segno);
3071 mutex_unlock(&DIRTY_I(sbi)->seglist_lock);
3072out:
3073 mutex_unlock(&curseg->curseg_mutex);
3074}
3075
3076void f2fs_restore_inmem_curseg(struct f2fs_sb_info *sbi)
3077{
3078 __f2fs_restore_inmem_curseg(sbi, CURSEG_COLD_DATA_PINNED);
3079
3080 if (sbi->am.atgc_enabled)
3081 __f2fs_restore_inmem_curseg(sbi, CURSEG_ALL_DATA_ATGC);
3082}
3083
3084static int get_ssr_segment(struct f2fs_sb_info *sbi, int type,
3085 int alloc_mode, unsigned long long age)
3086{
3087 struct curseg_info *curseg = CURSEG_I(sbi, type);
3088 unsigned segno = NULL_SEGNO;
3089 unsigned short seg_type = curseg->seg_type;
3090 int i, cnt;
3091 bool reversed = false;
3092
3093 sanity_check_seg_type(sbi, seg_type);
3094
3095 /* f2fs_need_SSR() already forces to do this */
3096 if (!f2fs_get_victim(sbi, &segno, BG_GC, seg_type,
3097 alloc_mode, age, false)) {
3098 curseg->next_segno = segno;
3099 return 1;
3100 }
3101
3102 /* For node segments, let's do SSR more intensively */
3103 if (IS_NODESEG(seg_type)) {
3104 if (seg_type >= CURSEG_WARM_NODE) {
3105 reversed = true;
3106 i = CURSEG_COLD_NODE;
3107 } else {
3108 i = CURSEG_HOT_NODE;
3109 }
3110 cnt = NR_CURSEG_NODE_TYPE;
3111 } else {
3112 if (seg_type >= CURSEG_WARM_DATA) {
3113 reversed = true;
3114 i = CURSEG_COLD_DATA;
3115 } else {
3116 i = CURSEG_HOT_DATA;
3117 }
3118 cnt = NR_CURSEG_DATA_TYPE;
3119 }
3120
3121 for (; cnt-- > 0; reversed ? i-- : i++) {
3122 if (i == seg_type)
3123 continue;
3124 if (!f2fs_get_victim(sbi, &segno, BG_GC, i,
3125 alloc_mode, age, false)) {
3126 curseg->next_segno = segno;
3127 return 1;
3128 }
3129 }
3130
3131 /* find valid_blocks=0 in dirty list */
3132 if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED))) {
3133 segno = get_free_segment(sbi);
3134 if (segno != NULL_SEGNO) {
3135 curseg->next_segno = segno;
3136 return 1;
3137 }
3138 }
3139 return 0;
3140}
3141
3142static bool need_new_seg(struct f2fs_sb_info *sbi, int type)
3143{
3144 struct curseg_info *curseg = CURSEG_I(sbi, type);
3145
3146 if (!is_set_ckpt_flags(sbi, CP_CRC_RECOVERY_FLAG) &&
3147 curseg->seg_type == CURSEG_WARM_NODE)
3148 return true;
3149 if (curseg->alloc_type == LFS && is_next_segment_free(sbi, curseg) &&
3150 likely(!is_sbi_flag_set(sbi, SBI_CP_DISABLED)))
3151 return true;
3152 if (!f2fs_need_SSR(sbi) || !get_ssr_segment(sbi, type, SSR, 0))
3153 return true;
3154 return false;
3155}
3156
3157int f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type,
3158 unsigned int start, unsigned int end)
3159{
3160 struct curseg_info *curseg = CURSEG_I(sbi, type);
3161 unsigned int segno;
3162 int ret = 0;
3163
3164 f2fs_down_read(&SM_I(sbi)->curseg_lock);
3165 mutex_lock(&curseg->curseg_mutex);
3166 down_write(&SIT_I(sbi)->sentry_lock);
3167
3168 segno = CURSEG_I(sbi, type)->segno;
3169 if (segno < start || segno > end)
3170 goto unlock;
3171
3172 if (f2fs_need_SSR(sbi) && get_ssr_segment(sbi, type, SSR, 0))
3173 ret = change_curseg(sbi, type);
3174 else
3175 ret = new_curseg(sbi, type, true);
3176
3177 stat_inc_seg_type(sbi, curseg);
3178
3179 locate_dirty_segment(sbi, segno);
3180unlock:
3181 up_write(&SIT_I(sbi)->sentry_lock);
3182
3183 if (segno != curseg->segno)
3184 f2fs_notice(sbi, "For resize: curseg of type %d: %u ==> %u",
3185 type, segno, curseg->segno);
3186
3187 mutex_unlock(&curseg->curseg_mutex);
3188 f2fs_up_read(&SM_I(sbi)->curseg_lock);
3189 return ret;
3190}
3191
3192static int __allocate_new_segment(struct f2fs_sb_info *sbi, int type,
3193 bool new_sec, bool force)
3194{
3195 struct curseg_info *curseg = CURSEG_I(sbi, type);
3196 unsigned int old_segno;
3197 int err = 0;
3198
3199 if (type == CURSEG_COLD_DATA_PINNED && !curseg->inited)
3200 goto allocate;
3201
3202 if (!force && curseg->inited &&
3203 !curseg->next_blkoff &&
3204 !get_valid_blocks(sbi, curseg->segno, new_sec) &&
3205 !get_ckpt_valid_blocks(sbi, curseg->segno, new_sec))
3206 return 0;
3207
3208allocate:
3209 old_segno = curseg->segno;
3210 err = new_curseg(sbi, type, true);
3211 if (err)
3212 return err;
3213 stat_inc_seg_type(sbi, curseg);
3214 locate_dirty_segment(sbi, old_segno);
3215 return 0;
3216}
3217
3218int f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force)
3219{
3220 int ret;
3221
3222 f2fs_down_read(&SM_I(sbi)->curseg_lock);
3223 down_write(&SIT_I(sbi)->sentry_lock);
3224 ret = __allocate_new_segment(sbi, type, true, force);
3225 up_write(&SIT_I(sbi)->sentry_lock);
3226 f2fs_up_read(&SM_I(sbi)->curseg_lock);
3227
3228 return ret;
3229}
3230
3231int f2fs_allocate_pinning_section(struct f2fs_sb_info *sbi)
3232{
3233 int err;
3234 bool gc_required = true;
3235
3236retry:
3237 f2fs_lock_op(sbi);
3238 err = f2fs_allocate_new_section(sbi, CURSEG_COLD_DATA_PINNED, false);
3239 f2fs_unlock_op(sbi);
3240
3241 if (f2fs_sb_has_blkzoned(sbi) && err == -EAGAIN && gc_required) {
3242 f2fs_down_write(&sbi->gc_lock);
3243 err = f2fs_gc_range(sbi, 0, GET_SEGNO(sbi, FDEV(0).end_blk),
3244 true, ZONED_PIN_SEC_REQUIRED_COUNT);
3245 f2fs_up_write(&sbi->gc_lock);
3246
3247 gc_required = false;
3248 if (!err)
3249 goto retry;
3250 }
3251
3252 return err;
3253}
3254
3255int f2fs_allocate_new_segments(struct f2fs_sb_info *sbi)
3256{
3257 int i;
3258 int err = 0;
3259
3260 f2fs_down_read(&SM_I(sbi)->curseg_lock);
3261 down_write(&SIT_I(sbi)->sentry_lock);
3262 for (i = CURSEG_HOT_DATA; i <= CURSEG_COLD_DATA; i++)
3263 err += __allocate_new_segment(sbi, i, false, false);
3264 up_write(&SIT_I(sbi)->sentry_lock);
3265 f2fs_up_read(&SM_I(sbi)->curseg_lock);
3266
3267 return err;
3268}
3269
3270bool f2fs_exist_trim_candidates(struct f2fs_sb_info *sbi,
3271 struct cp_control *cpc)
3272{
3273 __u64 trim_start = cpc->trim_start;
3274 bool has_candidate = false;
3275
3276 down_write(&SIT_I(sbi)->sentry_lock);
3277 for (; cpc->trim_start <= cpc->trim_end; cpc->trim_start++) {
3278 if (add_discard_addrs(sbi, cpc, true)) {
3279 has_candidate = true;
3280 break;
3281 }
3282 }
3283 up_write(&SIT_I(sbi)->sentry_lock);
3284
3285 cpc->trim_start = trim_start;
3286 return has_candidate;
3287}
3288
3289static unsigned int __issue_discard_cmd_range(struct f2fs_sb_info *sbi,
3290 struct discard_policy *dpolicy,
3291 unsigned int start, unsigned int end)
3292{
3293 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
3294 struct discard_cmd *prev_dc = NULL, *next_dc = NULL;
3295 struct rb_node **insert_p = NULL, *insert_parent = NULL;
3296 struct discard_cmd *dc;
3297 struct blk_plug plug;
3298 int issued;
3299 unsigned int trimmed = 0;
3300
3301next:
3302 issued = 0;
3303
3304 mutex_lock(&dcc->cmd_lock);
3305 if (unlikely(dcc->rbtree_check))
3306 f2fs_bug_on(sbi, !f2fs_check_discard_tree(sbi));
3307
3308 dc = __lookup_discard_cmd_ret(&dcc->root, start,
3309 &prev_dc, &next_dc, &insert_p, &insert_parent);
3310 if (!dc)
3311 dc = next_dc;
3312
3313 blk_start_plug(&plug);
3314
3315 while (dc && dc->di.lstart <= end) {
3316 struct rb_node *node;
3317 int err = 0;
3318
3319 if (dc->di.len < dpolicy->granularity)
3320 goto skip;
3321
3322 if (dc->state != D_PREP) {
3323 list_move_tail(&dc->list, &dcc->fstrim_list);
3324 goto skip;
3325 }
3326
3327 err = __submit_discard_cmd(sbi, dpolicy, dc, &issued);
3328
3329 if (issued >= dpolicy->max_requests) {
3330 start = dc->di.lstart + dc->di.len;
3331
3332 if (err)
3333 __remove_discard_cmd(sbi, dc);
3334
3335 blk_finish_plug(&plug);
3336 mutex_unlock(&dcc->cmd_lock);
3337 trimmed += __wait_all_discard_cmd(sbi, NULL);
3338 f2fs_io_schedule_timeout(DEFAULT_IO_TIMEOUT);
3339 goto next;
3340 }
3341skip:
3342 node = rb_next(&dc->rb_node);
3343 if (err)
3344 __remove_discard_cmd(sbi, dc);
3345 dc = rb_entry_safe(node, struct discard_cmd, rb_node);
3346
3347 if (fatal_signal_pending(current))
3348 break;
3349 }
3350
3351 blk_finish_plug(&plug);
3352 mutex_unlock(&dcc->cmd_lock);
3353
3354 return trimmed;
3355}
3356
3357int f2fs_trim_fs(struct f2fs_sb_info *sbi, struct fstrim_range *range)
3358{
3359 __u64 start = F2FS_BYTES_TO_BLK(range->start);
3360 __u64 end = start + F2FS_BYTES_TO_BLK(range->len) - 1;
3361 unsigned int start_segno, end_segno;
3362 block_t start_block, end_block;
3363 struct cp_control cpc;
3364 struct discard_policy dpolicy;
3365 unsigned long long trimmed = 0;
3366 int err = 0;
3367 bool need_align = f2fs_lfs_mode(sbi) && __is_large_section(sbi);
3368
3369 if (start >= MAX_BLKADDR(sbi) || range->len < sbi->blocksize)
3370 return -EINVAL;
3371
3372 if (end < MAIN_BLKADDR(sbi))
3373 goto out;
3374
3375 if (is_sbi_flag_set(sbi, SBI_NEED_FSCK)) {
3376 f2fs_warn(sbi, "Found FS corruption, run fsck to fix.");
3377 return -EFSCORRUPTED;
3378 }
3379
3380 /* start/end segment number in main_area */
3381 start_segno = (start <= MAIN_BLKADDR(sbi)) ? 0 : GET_SEGNO(sbi, start);
3382 end_segno = (end >= MAX_BLKADDR(sbi)) ? MAIN_SEGS(sbi) - 1 :
3383 GET_SEGNO(sbi, end);
3384 if (need_align) {
3385 start_segno = rounddown(start_segno, SEGS_PER_SEC(sbi));
3386 end_segno = roundup(end_segno + 1, SEGS_PER_SEC(sbi)) - 1;
3387 }
3388
3389 cpc.reason = CP_DISCARD;
3390 cpc.trim_minlen = max_t(__u64, 1, F2FS_BYTES_TO_BLK(range->minlen));
3391 cpc.trim_start = start_segno;
3392 cpc.trim_end = end_segno;
3393
3394 if (sbi->discard_blks == 0)
3395 goto out;
3396
3397 f2fs_down_write(&sbi->gc_lock);
3398 stat_inc_cp_call_count(sbi, TOTAL_CALL);
3399 err = f2fs_write_checkpoint(sbi, &cpc);
3400 f2fs_up_write(&sbi->gc_lock);
3401 if (err)
3402 goto out;
3403
3404 /*
3405 * We filed discard candidates, but actually we don't need to wait for
3406 * all of them, since they'll be issued in idle time along with runtime
3407 * discard option. User configuration looks like using runtime discard
3408 * or periodic fstrim instead of it.
3409 */
3410 if (f2fs_realtime_discard_enable(sbi))
3411 goto out;
3412
3413 start_block = START_BLOCK(sbi, start_segno);
3414 end_block = START_BLOCK(sbi, end_segno + 1);
3415
3416 __init_discard_policy(sbi, &dpolicy, DPOLICY_FSTRIM, cpc.trim_minlen);
3417 trimmed = __issue_discard_cmd_range(sbi, &dpolicy,
3418 start_block, end_block);
3419
3420 trimmed += __wait_discard_cmd_range(sbi, &dpolicy,
3421 start_block, end_block);
3422out:
3423 if (!err)
3424 range->len = F2FS_BLK_TO_BYTES(trimmed);
3425 return err;
3426}
3427
3428int f2fs_rw_hint_to_seg_type(struct f2fs_sb_info *sbi, enum rw_hint hint)
3429{
3430 if (F2FS_OPTION(sbi).active_logs == 2)
3431 return CURSEG_HOT_DATA;
3432 else if (F2FS_OPTION(sbi).active_logs == 4)
3433 return CURSEG_COLD_DATA;
3434
3435 /* active_log == 6 */
3436 switch (hint) {
3437 case WRITE_LIFE_SHORT:
3438 return CURSEG_HOT_DATA;
3439 case WRITE_LIFE_EXTREME:
3440 return CURSEG_COLD_DATA;
3441 default:
3442 return CURSEG_WARM_DATA;
3443 }
3444}
3445
3446/*
3447 * This returns write hints for each segment type. This hints will be
3448 * passed down to block layer as below by default.
3449 *
3450 * User F2FS Block
3451 * ---- ---- -----
3452 * META WRITE_LIFE_NONE|REQ_META
3453 * HOT_NODE WRITE_LIFE_NONE
3454 * WARM_NODE WRITE_LIFE_MEDIUM
3455 * COLD_NODE WRITE_LIFE_LONG
3456 * ioctl(COLD) COLD_DATA WRITE_LIFE_EXTREME
3457 * extension list " "
3458 *
3459 * -- buffered io
3460 * COLD_DATA WRITE_LIFE_EXTREME
3461 * HOT_DATA WRITE_LIFE_SHORT
3462 * WARM_DATA WRITE_LIFE_NOT_SET
3463 *
3464 * -- direct io
3465 * WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME
3466 * WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT
3467 * WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET
3468 * WRITE_LIFE_NONE " WRITE_LIFE_NONE
3469 * WRITE_LIFE_MEDIUM " WRITE_LIFE_MEDIUM
3470 * WRITE_LIFE_LONG " WRITE_LIFE_LONG
3471 */
3472enum rw_hint f2fs_io_type_to_rw_hint(struct f2fs_sb_info *sbi,
3473 enum page_type type, enum temp_type temp)
3474{
3475 switch (type) {
3476 case DATA:
3477 switch (temp) {
3478 case WARM:
3479 return WRITE_LIFE_NOT_SET;
3480 case HOT:
3481 return WRITE_LIFE_SHORT;
3482 case COLD:
3483 return WRITE_LIFE_EXTREME;
3484 default:
3485 return WRITE_LIFE_NONE;
3486 }
3487 case NODE:
3488 switch (temp) {
3489 case WARM:
3490 return WRITE_LIFE_MEDIUM;
3491 case HOT:
3492 return WRITE_LIFE_NONE;
3493 case COLD:
3494 return WRITE_LIFE_LONG;
3495 default:
3496 return WRITE_LIFE_NONE;
3497 }
3498 case META:
3499 return WRITE_LIFE_NONE;
3500 default:
3501 return WRITE_LIFE_NONE;
3502 }
3503}
3504
3505static int __get_segment_type_2(struct f2fs_io_info *fio)
3506{
3507 if (fio->type == DATA)
3508 return CURSEG_HOT_DATA;
3509 else
3510 return CURSEG_HOT_NODE;
3511}
3512
3513static int __get_segment_type_4(struct f2fs_io_info *fio)
3514{
3515 if (fio->type == DATA) {
3516 struct inode *inode = fio->page->mapping->host;
3517
3518 if (S_ISDIR(inode->i_mode))
3519 return CURSEG_HOT_DATA;
3520 else
3521 return CURSEG_COLD_DATA;
3522 } else {
3523 if (IS_DNODE(fio->page) && is_cold_node(fio->page))
3524 return CURSEG_WARM_NODE;
3525 else
3526 return CURSEG_COLD_NODE;
3527 }
3528}
3529
3530static int __get_age_segment_type(struct inode *inode, pgoff_t pgofs)
3531{
3532 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
3533 struct extent_info ei = {};
3534
3535 if (f2fs_lookup_age_extent_cache(inode, pgofs, &ei)) {
3536 if (!ei.age)
3537 return NO_CHECK_TYPE;
3538 if (ei.age <= sbi->hot_data_age_threshold)
3539 return CURSEG_HOT_DATA;
3540 if (ei.age <= sbi->warm_data_age_threshold)
3541 return CURSEG_WARM_DATA;
3542 return CURSEG_COLD_DATA;
3543 }
3544 return NO_CHECK_TYPE;
3545}
3546
3547static int __get_segment_type_6(struct f2fs_io_info *fio)
3548{
3549 if (fio->type == DATA) {
3550 struct inode *inode = fio->page->mapping->host;
3551 int type;
3552
3553 if (is_inode_flag_set(inode, FI_ALIGNED_WRITE))
3554 return CURSEG_COLD_DATA_PINNED;
3555
3556 if (page_private_gcing(fio->page)) {
3557 if (fio->sbi->am.atgc_enabled &&
3558 (fio->io_type == FS_DATA_IO) &&
3559 (fio->sbi->gc_mode != GC_URGENT_HIGH) &&
3560 __is_valid_data_blkaddr(fio->old_blkaddr) &&
3561 !is_inode_flag_set(inode, FI_OPU_WRITE))
3562 return CURSEG_ALL_DATA_ATGC;
3563 else
3564 return CURSEG_COLD_DATA;
3565 }
3566 if (file_is_cold(inode) || f2fs_need_compress_data(inode))
3567 return CURSEG_COLD_DATA;
3568
3569 type = __get_age_segment_type(inode,
3570 page_folio(fio->page)->index);
3571 if (type != NO_CHECK_TYPE)
3572 return type;
3573
3574 if (file_is_hot(inode) ||
3575 is_inode_flag_set(inode, FI_HOT_DATA) ||
3576 f2fs_is_cow_file(inode))
3577 return CURSEG_HOT_DATA;
3578 return f2fs_rw_hint_to_seg_type(F2FS_I_SB(inode),
3579 inode->i_write_hint);
3580 } else {
3581 if (IS_DNODE(fio->page))
3582 return is_cold_node(fio->page) ? CURSEG_WARM_NODE :
3583 CURSEG_HOT_NODE;
3584 return CURSEG_COLD_NODE;
3585 }
3586}
3587
3588enum temp_type f2fs_get_segment_temp(struct f2fs_sb_info *sbi,
3589 enum log_type type)
3590{
3591 struct curseg_info *curseg = CURSEG_I(sbi, type);
3592 enum temp_type temp = COLD;
3593
3594 switch (curseg->seg_type) {
3595 case CURSEG_HOT_NODE:
3596 case CURSEG_HOT_DATA:
3597 temp = HOT;
3598 break;
3599 case CURSEG_WARM_NODE:
3600 case CURSEG_WARM_DATA:
3601 temp = WARM;
3602 break;
3603 case CURSEG_COLD_NODE:
3604 case CURSEG_COLD_DATA:
3605 temp = COLD;
3606 break;
3607 default:
3608 f2fs_bug_on(sbi, 1);
3609 }
3610
3611 return temp;
3612}
3613
3614static int __get_segment_type(struct f2fs_io_info *fio)
3615{
3616 enum log_type type = CURSEG_HOT_DATA;
3617
3618 switch (F2FS_OPTION(fio->sbi).active_logs) {
3619 case 2:
3620 type = __get_segment_type_2(fio);
3621 break;
3622 case 4:
3623 type = __get_segment_type_4(fio);
3624 break;
3625 case 6:
3626 type = __get_segment_type_6(fio);
3627 break;
3628 default:
3629 f2fs_bug_on(fio->sbi, true);
3630 }
3631
3632 fio->temp = f2fs_get_segment_temp(fio->sbi, type);
3633
3634 return type;
3635}
3636
3637static void f2fs_randomize_chunk(struct f2fs_sb_info *sbi,
3638 struct curseg_info *seg)
3639{
3640 /* To allocate block chunks in different sizes, use random number */
3641 if (--seg->fragment_remained_chunk > 0)
3642 return;
3643
3644 seg->fragment_remained_chunk =
3645 get_random_u32_inclusive(1, sbi->max_fragment_chunk);
3646 seg->next_blkoff +=
3647 get_random_u32_inclusive(1, sbi->max_fragment_hole);
3648}
3649
3650static void reset_curseg_fields(struct curseg_info *curseg)
3651{
3652 curseg->inited = false;
3653 curseg->segno = NULL_SEGNO;
3654 curseg->next_segno = 0;
3655}
3656
3657int f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
3658 block_t old_blkaddr, block_t *new_blkaddr,
3659 struct f2fs_summary *sum, int type,
3660 struct f2fs_io_info *fio)
3661{
3662 struct sit_info *sit_i = SIT_I(sbi);
3663 struct curseg_info *curseg = CURSEG_I(sbi, type);
3664 unsigned long long old_mtime;
3665 bool from_gc = (type == CURSEG_ALL_DATA_ATGC);
3666 struct seg_entry *se = NULL;
3667 bool segment_full = false;
3668 int ret = 0;
3669
3670 f2fs_down_read(&SM_I(sbi)->curseg_lock);
3671
3672 mutex_lock(&curseg->curseg_mutex);
3673 down_write(&sit_i->sentry_lock);
3674
3675 if (curseg->segno == NULL_SEGNO) {
3676 ret = -ENOSPC;
3677 goto out_err;
3678 }
3679
3680 if (from_gc) {
3681 f2fs_bug_on(sbi, GET_SEGNO(sbi, old_blkaddr) == NULL_SEGNO);
3682 se = get_seg_entry(sbi, GET_SEGNO(sbi, old_blkaddr));
3683 sanity_check_seg_type(sbi, se->type);
3684 f2fs_bug_on(sbi, IS_NODESEG(se->type));
3685 }
3686 *new_blkaddr = NEXT_FREE_BLKADDR(sbi, curseg);
3687
3688 f2fs_bug_on(sbi, curseg->next_blkoff >= BLKS_PER_SEG(sbi));
3689
3690 f2fs_wait_discard_bio(sbi, *new_blkaddr);
3691
3692 curseg->sum_blk->entries[curseg->next_blkoff] = *sum;
3693 if (curseg->alloc_type == SSR) {
3694 curseg->next_blkoff = f2fs_find_next_ssr_block(sbi, curseg);
3695 } else {
3696 curseg->next_blkoff++;
3697 if (F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_BLK)
3698 f2fs_randomize_chunk(sbi, curseg);
3699 }
3700 if (curseg->next_blkoff >= f2fs_usable_blks_in_seg(sbi, curseg->segno))
3701 segment_full = true;
3702 stat_inc_block_count(sbi, curseg);
3703
3704 if (from_gc) {
3705 old_mtime = get_segment_mtime(sbi, old_blkaddr);
3706 } else {
3707 update_segment_mtime(sbi, old_blkaddr, 0);
3708 old_mtime = 0;
3709 }
3710 update_segment_mtime(sbi, *new_blkaddr, old_mtime);
3711
3712 /*
3713 * SIT information should be updated before segment allocation,
3714 * since SSR needs latest valid block information.
3715 */
3716 update_sit_entry(sbi, *new_blkaddr, 1);
3717 update_sit_entry(sbi, old_blkaddr, -1);
3718
3719 /*
3720 * If the current segment is full, flush it out and replace it with a
3721 * new segment.
3722 */
3723 if (segment_full) {
3724 if (type == CURSEG_COLD_DATA_PINNED &&
3725 !((curseg->segno + 1) % sbi->segs_per_sec)) {
3726 write_sum_page(sbi, curseg->sum_blk,
3727 GET_SUM_BLOCK(sbi, curseg->segno));
3728 reset_curseg_fields(curseg);
3729 goto skip_new_segment;
3730 }
3731
3732 if (from_gc) {
3733 ret = get_atssr_segment(sbi, type, se->type,
3734 AT_SSR, se->mtime);
3735 } else {
3736 if (need_new_seg(sbi, type))
3737 ret = new_curseg(sbi, type, false);
3738 else
3739 ret = change_curseg(sbi, type);
3740 stat_inc_seg_type(sbi, curseg);
3741 }
3742
3743 if (ret)
3744 goto out_err;
3745 }
3746
3747skip_new_segment:
3748 /*
3749 * segment dirty status should be updated after segment allocation,
3750 * so we just need to update status only one time after previous
3751 * segment being closed.
3752 */
3753 locate_dirty_segment(sbi, GET_SEGNO(sbi, old_blkaddr));
3754 locate_dirty_segment(sbi, GET_SEGNO(sbi, *new_blkaddr));
3755
3756 if (IS_DATASEG(curseg->seg_type))
3757 atomic64_inc(&sbi->allocated_data_blocks);
3758
3759 up_write(&sit_i->sentry_lock);
3760
3761 if (page && IS_NODESEG(curseg->seg_type)) {
3762 fill_node_footer_blkaddr(page, NEXT_FREE_BLKADDR(sbi, curseg));
3763
3764 f2fs_inode_chksum_set(sbi, page);
3765 }
3766
3767 if (fio) {
3768 struct f2fs_bio_info *io;
3769
3770 INIT_LIST_HEAD(&fio->list);
3771 fio->in_list = 1;
3772 io = sbi->write_io[fio->type] + fio->temp;
3773 spin_lock(&io->io_lock);
3774 list_add_tail(&fio->list, &io->io_list);
3775 spin_unlock(&io->io_lock);
3776 }
3777
3778 mutex_unlock(&curseg->curseg_mutex);
3779 f2fs_up_read(&SM_I(sbi)->curseg_lock);
3780 return 0;
3781
3782out_err:
3783 *new_blkaddr = NULL_ADDR;
3784 up_write(&sit_i->sentry_lock);
3785 mutex_unlock(&curseg->curseg_mutex);
3786 f2fs_up_read(&SM_I(sbi)->curseg_lock);
3787 return ret;
3788}
3789
3790void f2fs_update_device_state(struct f2fs_sb_info *sbi, nid_t ino,
3791 block_t blkaddr, unsigned int blkcnt)
3792{
3793 if (!f2fs_is_multi_device(sbi))
3794 return;
3795
3796 while (1) {
3797 unsigned int devidx = f2fs_target_device_index(sbi, blkaddr);
3798 unsigned int blks = FDEV(devidx).end_blk - blkaddr + 1;
3799
3800 /* update device state for fsync */
3801 f2fs_set_dirty_device(sbi, ino, devidx, FLUSH_INO);
3802
3803 /* update device state for checkpoint */
3804 if (!f2fs_test_bit(devidx, (char *)&sbi->dirty_device)) {
3805 spin_lock(&sbi->dev_lock);
3806 f2fs_set_bit(devidx, (char *)&sbi->dirty_device);
3807 spin_unlock(&sbi->dev_lock);
3808 }
3809
3810 if (blkcnt <= blks)
3811 break;
3812 blkcnt -= blks;
3813 blkaddr += blks;
3814 }
3815}
3816
3817static int log_type_to_seg_type(enum log_type type)
3818{
3819 int seg_type = CURSEG_COLD_DATA;
3820
3821 switch (type) {
3822 case CURSEG_HOT_DATA:
3823 case CURSEG_WARM_DATA:
3824 case CURSEG_COLD_DATA:
3825 case CURSEG_HOT_NODE:
3826 case CURSEG_WARM_NODE:
3827 case CURSEG_COLD_NODE:
3828 seg_type = (int)type;
3829 break;
3830 case CURSEG_COLD_DATA_PINNED:
3831 case CURSEG_ALL_DATA_ATGC:
3832 seg_type = CURSEG_COLD_DATA;
3833 break;
3834 default:
3835 break;
3836 }
3837 return seg_type;
3838}
3839
3840static void do_write_page(struct f2fs_summary *sum, struct f2fs_io_info *fio)
3841{
3842 enum log_type type = __get_segment_type(fio);
3843 int seg_type = log_type_to_seg_type(type);
3844 bool keep_order = (f2fs_lfs_mode(fio->sbi) &&
3845 seg_type == CURSEG_COLD_DATA);
3846
3847 if (keep_order)
3848 f2fs_down_read(&fio->sbi->io_order_lock);
3849
3850 if (f2fs_allocate_data_block(fio->sbi, fio->page, fio->old_blkaddr,
3851 &fio->new_blkaddr, sum, type, fio)) {
3852 if (fscrypt_inode_uses_fs_layer_crypto(fio->page->mapping->host))
3853 fscrypt_finalize_bounce_page(&fio->encrypted_page);
3854 end_page_writeback(fio->page);
3855 if (f2fs_in_warm_node_list(fio->sbi, fio->page))
3856 f2fs_del_fsync_node_entry(fio->sbi, fio->page);
3857 goto out;
3858 }
3859 if (GET_SEGNO(fio->sbi, fio->old_blkaddr) != NULL_SEGNO)
3860 f2fs_invalidate_internal_cache(fio->sbi, fio->old_blkaddr);
3861
3862 /* writeout dirty page into bdev */
3863 f2fs_submit_page_write(fio);
3864
3865 f2fs_update_device_state(fio->sbi, fio->ino, fio->new_blkaddr, 1);
3866out:
3867 if (keep_order)
3868 f2fs_up_read(&fio->sbi->io_order_lock);
3869}
3870
3871void f2fs_do_write_meta_page(struct f2fs_sb_info *sbi, struct folio *folio,
3872 enum iostat_type io_type)
3873{
3874 struct f2fs_io_info fio = {
3875 .sbi = sbi,
3876 .type = META,
3877 .temp = HOT,
3878 .op = REQ_OP_WRITE,
3879 .op_flags = REQ_SYNC | REQ_META | REQ_PRIO,
3880 .old_blkaddr = folio->index,
3881 .new_blkaddr = folio->index,
3882 .page = folio_page(folio, 0),
3883 .encrypted_page = NULL,
3884 .in_list = 0,
3885 };
3886
3887 if (unlikely(folio->index >= MAIN_BLKADDR(sbi)))
3888 fio.op_flags &= ~REQ_META;
3889
3890 folio_start_writeback(folio);
3891 f2fs_submit_page_write(&fio);
3892
3893 stat_inc_meta_count(sbi, folio->index);
3894 f2fs_update_iostat(sbi, NULL, io_type, F2FS_BLKSIZE);
3895}
3896
3897void f2fs_do_write_node_page(unsigned int nid, struct f2fs_io_info *fio)
3898{
3899 struct f2fs_summary sum;
3900
3901 set_summary(&sum, nid, 0, 0);
3902 do_write_page(&sum, fio);
3903
3904 f2fs_update_iostat(fio->sbi, NULL, fio->io_type, F2FS_BLKSIZE);
3905}
3906
3907void f2fs_outplace_write_data(struct dnode_of_data *dn,
3908 struct f2fs_io_info *fio)
3909{
3910 struct f2fs_sb_info *sbi = fio->sbi;
3911 struct f2fs_summary sum;
3912
3913 f2fs_bug_on(sbi, dn->data_blkaddr == NULL_ADDR);
3914 if (fio->io_type == FS_DATA_IO || fio->io_type == FS_CP_DATA_IO)
3915 f2fs_update_age_extent_cache(dn);
3916 set_summary(&sum, dn->nid, dn->ofs_in_node, fio->version);
3917 do_write_page(&sum, fio);
3918 f2fs_update_data_blkaddr(dn, fio->new_blkaddr);
3919
3920 f2fs_update_iostat(sbi, dn->inode, fio->io_type, F2FS_BLKSIZE);
3921}
3922
3923int f2fs_inplace_write_data(struct f2fs_io_info *fio)
3924{
3925 int err;
3926 struct f2fs_sb_info *sbi = fio->sbi;
3927 unsigned int segno;
3928
3929 fio->new_blkaddr = fio->old_blkaddr;
3930 /* i/o temperature is needed for passing down write hints */
3931 __get_segment_type(fio);
3932
3933 segno = GET_SEGNO(sbi, fio->new_blkaddr);
3934
3935 if (!IS_DATASEG(get_seg_entry(sbi, segno)->type)) {
3936 set_sbi_flag(sbi, SBI_NEED_FSCK);
3937 f2fs_warn(sbi, "%s: incorrect segment(%u) type, run fsck to fix.",
3938 __func__, segno);
3939 err = -EFSCORRUPTED;
3940 f2fs_handle_error(sbi, ERROR_INCONSISTENT_SUM_TYPE);
3941 goto drop_bio;
3942 }
3943
3944 if (f2fs_cp_error(sbi)) {
3945 err = -EIO;
3946 goto drop_bio;
3947 }
3948
3949 if (fio->meta_gc)
3950 f2fs_truncate_meta_inode_pages(sbi, fio->new_blkaddr, 1);
3951
3952 stat_inc_inplace_blocks(fio->sbi);
3953
3954 if (fio->bio && !IS_F2FS_IPU_NOCACHE(sbi))
3955 err = f2fs_merge_page_bio(fio);
3956 else
3957 err = f2fs_submit_page_bio(fio);
3958 if (!err) {
3959 f2fs_update_device_state(fio->sbi, fio->ino,
3960 fio->new_blkaddr, 1);
3961 f2fs_update_iostat(fio->sbi, fio->page->mapping->host,
3962 fio->io_type, F2FS_BLKSIZE);
3963 }
3964
3965 return err;
3966drop_bio:
3967 if (fio->bio && *(fio->bio)) {
3968 struct bio *bio = *(fio->bio);
3969
3970 bio->bi_status = BLK_STS_IOERR;
3971 bio_endio(bio);
3972 *(fio->bio) = NULL;
3973 }
3974 return err;
3975}
3976
3977static inline int __f2fs_get_curseg(struct f2fs_sb_info *sbi,
3978 unsigned int segno)
3979{
3980 int i;
3981
3982 for (i = CURSEG_HOT_DATA; i < NO_CHECK_TYPE; i++) {
3983 if (CURSEG_I(sbi, i)->segno == segno)
3984 break;
3985 }
3986 return i;
3987}
3988
3989void f2fs_do_replace_block(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
3990 block_t old_blkaddr, block_t new_blkaddr,
3991 bool recover_curseg, bool recover_newaddr,
3992 bool from_gc)
3993{
3994 struct sit_info *sit_i = SIT_I(sbi);
3995 struct curseg_info *curseg;
3996 unsigned int segno, old_cursegno;
3997 struct seg_entry *se;
3998 int type;
3999 unsigned short old_blkoff;
4000 unsigned char old_alloc_type;
4001
4002 segno = GET_SEGNO(sbi, new_blkaddr);
4003 se = get_seg_entry(sbi, segno);
4004 type = se->type;
4005
4006 f2fs_down_write(&SM_I(sbi)->curseg_lock);
4007
4008 if (!recover_curseg) {
4009 /* for recovery flow */
4010 if (se->valid_blocks == 0 && !IS_CURSEG(sbi, segno)) {
4011 if (old_blkaddr == NULL_ADDR)
4012 type = CURSEG_COLD_DATA;
4013 else
4014 type = CURSEG_WARM_DATA;
4015 }
4016 } else {
4017 if (IS_CURSEG(sbi, segno)) {
4018 /* se->type is volatile as SSR allocation */
4019 type = __f2fs_get_curseg(sbi, segno);
4020 f2fs_bug_on(sbi, type == NO_CHECK_TYPE);
4021 } else {
4022 type = CURSEG_WARM_DATA;
4023 }
4024 }
4025
4026 curseg = CURSEG_I(sbi, type);
4027 f2fs_bug_on(sbi, !IS_DATASEG(curseg->seg_type));
4028
4029 mutex_lock(&curseg->curseg_mutex);
4030 down_write(&sit_i->sentry_lock);
4031
4032 old_cursegno = curseg->segno;
4033 old_blkoff = curseg->next_blkoff;
4034 old_alloc_type = curseg->alloc_type;
4035
4036 /* change the current segment */
4037 if (segno != curseg->segno) {
4038 curseg->next_segno = segno;
4039 if (change_curseg(sbi, type))
4040 goto out_unlock;
4041 }
4042
4043 curseg->next_blkoff = GET_BLKOFF_FROM_SEG0(sbi, new_blkaddr);
4044 curseg->sum_blk->entries[curseg->next_blkoff] = *sum;
4045
4046 if (!recover_curseg || recover_newaddr) {
4047 if (!from_gc)
4048 update_segment_mtime(sbi, new_blkaddr, 0);
4049 update_sit_entry(sbi, new_blkaddr, 1);
4050 }
4051 if (GET_SEGNO(sbi, old_blkaddr) != NULL_SEGNO) {
4052 f2fs_invalidate_internal_cache(sbi, old_blkaddr);
4053 if (!from_gc)
4054 update_segment_mtime(sbi, old_blkaddr, 0);
4055 update_sit_entry(sbi, old_blkaddr, -1);
4056 }
4057
4058 locate_dirty_segment(sbi, GET_SEGNO(sbi, old_blkaddr));
4059 locate_dirty_segment(sbi, GET_SEGNO(sbi, new_blkaddr));
4060
4061 locate_dirty_segment(sbi, old_cursegno);
4062
4063 if (recover_curseg) {
4064 if (old_cursegno != curseg->segno) {
4065 curseg->next_segno = old_cursegno;
4066 if (change_curseg(sbi, type))
4067 goto out_unlock;
4068 }
4069 curseg->next_blkoff = old_blkoff;
4070 curseg->alloc_type = old_alloc_type;
4071 }
4072
4073out_unlock:
4074 up_write(&sit_i->sentry_lock);
4075 mutex_unlock(&curseg->curseg_mutex);
4076 f2fs_up_write(&SM_I(sbi)->curseg_lock);
4077}
4078
4079void f2fs_replace_block(struct f2fs_sb_info *sbi, struct dnode_of_data *dn,
4080 block_t old_addr, block_t new_addr,
4081 unsigned char version, bool recover_curseg,
4082 bool recover_newaddr)
4083{
4084 struct f2fs_summary sum;
4085
4086 set_summary(&sum, dn->nid, dn->ofs_in_node, version);
4087
4088 f2fs_do_replace_block(sbi, &sum, old_addr, new_addr,
4089 recover_curseg, recover_newaddr, false);
4090
4091 f2fs_update_data_blkaddr(dn, new_addr);
4092}
4093
4094void f2fs_wait_on_page_writeback(struct page *page,
4095 enum page_type type, bool ordered, bool locked)
4096{
4097 if (folio_test_writeback(page_folio(page))) {
4098 struct f2fs_sb_info *sbi = F2FS_P_SB(page);
4099
4100 /* submit cached LFS IO */
4101 f2fs_submit_merged_write_cond(sbi, NULL, page, 0, type);
4102 /* submit cached IPU IO */
4103 f2fs_submit_merged_ipu_write(sbi, NULL, page);
4104 if (ordered) {
4105 wait_on_page_writeback(page);
4106 f2fs_bug_on(sbi, locked &&
4107 folio_test_writeback(page_folio(page)));
4108 } else {
4109 wait_for_stable_page(page);
4110 }
4111 }
4112}
4113
4114void f2fs_wait_on_block_writeback(struct inode *inode, block_t blkaddr)
4115{
4116 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
4117 struct page *cpage;
4118
4119 if (!f2fs_meta_inode_gc_required(inode))
4120 return;
4121
4122 if (!__is_valid_data_blkaddr(blkaddr))
4123 return;
4124
4125 cpage = find_lock_page(META_MAPPING(sbi), blkaddr);
4126 if (cpage) {
4127 f2fs_wait_on_page_writeback(cpage, DATA, true, true);
4128 f2fs_put_page(cpage, 1);
4129 }
4130}
4131
4132void f2fs_wait_on_block_writeback_range(struct inode *inode, block_t blkaddr,
4133 block_t len)
4134{
4135 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
4136 block_t i;
4137
4138 if (!f2fs_meta_inode_gc_required(inode))
4139 return;
4140
4141 for (i = 0; i < len; i++)
4142 f2fs_wait_on_block_writeback(inode, blkaddr + i);
4143
4144 f2fs_truncate_meta_inode_pages(sbi, blkaddr, len);
4145}
4146
4147static int read_compacted_summaries(struct f2fs_sb_info *sbi)
4148{
4149 struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
4150 struct curseg_info *seg_i;
4151 unsigned char *kaddr;
4152 struct page *page;
4153 block_t start;
4154 int i, j, offset;
4155
4156 start = start_sum_block(sbi);
4157
4158 page = f2fs_get_meta_page(sbi, start++);
4159 if (IS_ERR(page))
4160 return PTR_ERR(page);
4161 kaddr = (unsigned char *)page_address(page);
4162
4163 /* Step 1: restore nat cache */
4164 seg_i = CURSEG_I(sbi, CURSEG_HOT_DATA);
4165 memcpy(seg_i->journal, kaddr, SUM_JOURNAL_SIZE);
4166
4167 /* Step 2: restore sit cache */
4168 seg_i = CURSEG_I(sbi, CURSEG_COLD_DATA);
4169 memcpy(seg_i->journal, kaddr + SUM_JOURNAL_SIZE, SUM_JOURNAL_SIZE);
4170 offset = 2 * SUM_JOURNAL_SIZE;
4171
4172 /* Step 3: restore summary entries */
4173 for (i = CURSEG_HOT_DATA; i <= CURSEG_COLD_DATA; i++) {
4174 unsigned short blk_off;
4175 unsigned int segno;
4176
4177 seg_i = CURSEG_I(sbi, i);
4178 segno = le32_to_cpu(ckpt->cur_data_segno[i]);
4179 blk_off = le16_to_cpu(ckpt->cur_data_blkoff[i]);
4180 seg_i->next_segno = segno;
4181 reset_curseg(sbi, i, 0);
4182 seg_i->alloc_type = ckpt->alloc_type[i];
4183 seg_i->next_blkoff = blk_off;
4184
4185 if (seg_i->alloc_type == SSR)
4186 blk_off = BLKS_PER_SEG(sbi);
4187
4188 for (j = 0; j < blk_off; j++) {
4189 struct f2fs_summary *s;
4190
4191 s = (struct f2fs_summary *)(kaddr + offset);
4192 seg_i->sum_blk->entries[j] = *s;
4193 offset += SUMMARY_SIZE;
4194 if (offset + SUMMARY_SIZE <= PAGE_SIZE -
4195 SUM_FOOTER_SIZE)
4196 continue;
4197
4198 f2fs_put_page(page, 1);
4199 page = NULL;
4200
4201 page = f2fs_get_meta_page(sbi, start++);
4202 if (IS_ERR(page))
4203 return PTR_ERR(page);
4204 kaddr = (unsigned char *)page_address(page);
4205 offset = 0;
4206 }
4207 }
4208 f2fs_put_page(page, 1);
4209 return 0;
4210}
4211
4212static int read_normal_summaries(struct f2fs_sb_info *sbi, int type)
4213{
4214 struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
4215 struct f2fs_summary_block *sum;
4216 struct curseg_info *curseg;
4217 struct page *new;
4218 unsigned short blk_off;
4219 unsigned int segno = 0;
4220 block_t blk_addr = 0;
4221 int err = 0;
4222
4223 /* get segment number and block addr */
4224 if (IS_DATASEG(type)) {
4225 segno = le32_to_cpu(ckpt->cur_data_segno[type]);
4226 blk_off = le16_to_cpu(ckpt->cur_data_blkoff[type -
4227 CURSEG_HOT_DATA]);
4228 if (__exist_node_summaries(sbi))
4229 blk_addr = sum_blk_addr(sbi, NR_CURSEG_PERSIST_TYPE, type);
4230 else
4231 blk_addr = sum_blk_addr(sbi, NR_CURSEG_DATA_TYPE, type);
4232 } else {
4233 segno = le32_to_cpu(ckpt->cur_node_segno[type -
4234 CURSEG_HOT_NODE]);
4235 blk_off = le16_to_cpu(ckpt->cur_node_blkoff[type -
4236 CURSEG_HOT_NODE]);
4237 if (__exist_node_summaries(sbi))
4238 blk_addr = sum_blk_addr(sbi, NR_CURSEG_NODE_TYPE,
4239 type - CURSEG_HOT_NODE);
4240 else
4241 blk_addr = GET_SUM_BLOCK(sbi, segno);
4242 }
4243
4244 new = f2fs_get_meta_page(sbi, blk_addr);
4245 if (IS_ERR(new))
4246 return PTR_ERR(new);
4247 sum = (struct f2fs_summary_block *)page_address(new);
4248
4249 if (IS_NODESEG(type)) {
4250 if (__exist_node_summaries(sbi)) {
4251 struct f2fs_summary *ns = &sum->entries[0];
4252 int i;
4253
4254 for (i = 0; i < BLKS_PER_SEG(sbi); i++, ns++) {
4255 ns->version = 0;
4256 ns->ofs_in_node = 0;
4257 }
4258 } else {
4259 err = f2fs_restore_node_summary(sbi, segno, sum);
4260 if (err)
4261 goto out;
4262 }
4263 }
4264
4265 /* set uncompleted segment to curseg */
4266 curseg = CURSEG_I(sbi, type);
4267 mutex_lock(&curseg->curseg_mutex);
4268
4269 /* update journal info */
4270 down_write(&curseg->journal_rwsem);
4271 memcpy(curseg->journal, &sum->journal, SUM_JOURNAL_SIZE);
4272 up_write(&curseg->journal_rwsem);
4273
4274 memcpy(curseg->sum_blk->entries, sum->entries, SUM_ENTRY_SIZE);
4275 memcpy(&curseg->sum_blk->footer, &sum->footer, SUM_FOOTER_SIZE);
4276 curseg->next_segno = segno;
4277 reset_curseg(sbi, type, 0);
4278 curseg->alloc_type = ckpt->alloc_type[type];
4279 curseg->next_blkoff = blk_off;
4280 mutex_unlock(&curseg->curseg_mutex);
4281out:
4282 f2fs_put_page(new, 1);
4283 return err;
4284}
4285
4286static int restore_curseg_summaries(struct f2fs_sb_info *sbi)
4287{
4288 struct f2fs_journal *sit_j = CURSEG_I(sbi, CURSEG_COLD_DATA)->journal;
4289 struct f2fs_journal *nat_j = CURSEG_I(sbi, CURSEG_HOT_DATA)->journal;
4290 int type = CURSEG_HOT_DATA;
4291 int err;
4292
4293 if (is_set_ckpt_flags(sbi, CP_COMPACT_SUM_FLAG)) {
4294 int npages = f2fs_npages_for_summary_flush(sbi, true);
4295
4296 if (npages >= 2)
4297 f2fs_ra_meta_pages(sbi, start_sum_block(sbi), npages,
4298 META_CP, true);
4299
4300 /* restore for compacted data summary */
4301 err = read_compacted_summaries(sbi);
4302 if (err)
4303 return err;
4304 type = CURSEG_HOT_NODE;
4305 }
4306
4307 if (__exist_node_summaries(sbi))
4308 f2fs_ra_meta_pages(sbi,
4309 sum_blk_addr(sbi, NR_CURSEG_PERSIST_TYPE, type),
4310 NR_CURSEG_PERSIST_TYPE - type, META_CP, true);
4311
4312 for (; type <= CURSEG_COLD_NODE; type++) {
4313 err = read_normal_summaries(sbi, type);
4314 if (err)
4315 return err;
4316 }
4317
4318 /* sanity check for summary blocks */
4319 if (nats_in_cursum(nat_j) > NAT_JOURNAL_ENTRIES ||
4320 sits_in_cursum(sit_j) > SIT_JOURNAL_ENTRIES) {
4321 f2fs_err(sbi, "invalid journal entries nats %u sits %u",
4322 nats_in_cursum(nat_j), sits_in_cursum(sit_j));
4323 return -EINVAL;
4324 }
4325
4326 return 0;
4327}
4328
4329static void write_compacted_summaries(struct f2fs_sb_info *sbi, block_t blkaddr)
4330{
4331 struct page *page;
4332 unsigned char *kaddr;
4333 struct f2fs_summary *summary;
4334 struct curseg_info *seg_i;
4335 int written_size = 0;
4336 int i, j;
4337
4338 page = f2fs_grab_meta_page(sbi, blkaddr++);
4339 kaddr = (unsigned char *)page_address(page);
4340 memset(kaddr, 0, PAGE_SIZE);
4341
4342 /* Step 1: write nat cache */
4343 seg_i = CURSEG_I(sbi, CURSEG_HOT_DATA);
4344 memcpy(kaddr, seg_i->journal, SUM_JOURNAL_SIZE);
4345 written_size += SUM_JOURNAL_SIZE;
4346
4347 /* Step 2: write sit cache */
4348 seg_i = CURSEG_I(sbi, CURSEG_COLD_DATA);
4349 memcpy(kaddr + written_size, seg_i->journal, SUM_JOURNAL_SIZE);
4350 written_size += SUM_JOURNAL_SIZE;
4351
4352 /* Step 3: write summary entries */
4353 for (i = CURSEG_HOT_DATA; i <= CURSEG_COLD_DATA; i++) {
4354 seg_i = CURSEG_I(sbi, i);
4355 for (j = 0; j < f2fs_curseg_valid_blocks(sbi, i); j++) {
4356 if (!page) {
4357 page = f2fs_grab_meta_page(sbi, blkaddr++);
4358 kaddr = (unsigned char *)page_address(page);
4359 memset(kaddr, 0, PAGE_SIZE);
4360 written_size = 0;
4361 }
4362 summary = (struct f2fs_summary *)(kaddr + written_size);
4363 *summary = seg_i->sum_blk->entries[j];
4364 written_size += SUMMARY_SIZE;
4365
4366 if (written_size + SUMMARY_SIZE <= PAGE_SIZE -
4367 SUM_FOOTER_SIZE)
4368 continue;
4369
4370 set_page_dirty(page);
4371 f2fs_put_page(page, 1);
4372 page = NULL;
4373 }
4374 }
4375 if (page) {
4376 set_page_dirty(page);
4377 f2fs_put_page(page, 1);
4378 }
4379}
4380
4381static void write_normal_summaries(struct f2fs_sb_info *sbi,
4382 block_t blkaddr, int type)
4383{
4384 int i, end;
4385
4386 if (IS_DATASEG(type))
4387 end = type + NR_CURSEG_DATA_TYPE;
4388 else
4389 end = type + NR_CURSEG_NODE_TYPE;
4390
4391 for (i = type; i < end; i++)
4392 write_current_sum_page(sbi, i, blkaddr + (i - type));
4393}
4394
4395void f2fs_write_data_summaries(struct f2fs_sb_info *sbi, block_t start_blk)
4396{
4397 if (is_set_ckpt_flags(sbi, CP_COMPACT_SUM_FLAG))
4398 write_compacted_summaries(sbi, start_blk);
4399 else
4400 write_normal_summaries(sbi, start_blk, CURSEG_HOT_DATA);
4401}
4402
4403void f2fs_write_node_summaries(struct f2fs_sb_info *sbi, block_t start_blk)
4404{
4405 write_normal_summaries(sbi, start_blk, CURSEG_HOT_NODE);
4406}
4407
4408int f2fs_lookup_journal_in_cursum(struct f2fs_journal *journal, int type,
4409 unsigned int val, int alloc)
4410{
4411 int i;
4412
4413 if (type == NAT_JOURNAL) {
4414 for (i = 0; i < nats_in_cursum(journal); i++) {
4415 if (le32_to_cpu(nid_in_journal(journal, i)) == val)
4416 return i;
4417 }
4418 if (alloc && __has_cursum_space(journal, 1, NAT_JOURNAL))
4419 return update_nats_in_cursum(journal, 1);
4420 } else if (type == SIT_JOURNAL) {
4421 for (i = 0; i < sits_in_cursum(journal); i++)
4422 if (le32_to_cpu(segno_in_journal(journal, i)) == val)
4423 return i;
4424 if (alloc && __has_cursum_space(journal, 1, SIT_JOURNAL))
4425 return update_sits_in_cursum(journal, 1);
4426 }
4427 return -1;
4428}
4429
4430static struct page *get_current_sit_page(struct f2fs_sb_info *sbi,
4431 unsigned int segno)
4432{
4433 return f2fs_get_meta_page(sbi, current_sit_addr(sbi, segno));
4434}
4435
4436static struct page *get_next_sit_page(struct f2fs_sb_info *sbi,
4437 unsigned int start)
4438{
4439 struct sit_info *sit_i = SIT_I(sbi);
4440 struct page *page;
4441 pgoff_t src_off, dst_off;
4442
4443 src_off = current_sit_addr(sbi, start);
4444 dst_off = next_sit_addr(sbi, src_off);
4445
4446 page = f2fs_grab_meta_page(sbi, dst_off);
4447 seg_info_to_sit_page(sbi, page, start);
4448
4449 set_page_dirty(page);
4450 set_to_next_sit(sit_i, start);
4451
4452 return page;
4453}
4454
4455static struct sit_entry_set *grab_sit_entry_set(void)
4456{
4457 struct sit_entry_set *ses =
4458 f2fs_kmem_cache_alloc(sit_entry_set_slab,
4459 GFP_NOFS, true, NULL);
4460
4461 ses->entry_cnt = 0;
4462 INIT_LIST_HEAD(&ses->set_list);
4463 return ses;
4464}
4465
4466static void release_sit_entry_set(struct sit_entry_set *ses)
4467{
4468 list_del(&ses->set_list);
4469 kmem_cache_free(sit_entry_set_slab, ses);
4470}
4471
4472static void adjust_sit_entry_set(struct sit_entry_set *ses,
4473 struct list_head *head)
4474{
4475 struct sit_entry_set *next = ses;
4476
4477 if (list_is_last(&ses->set_list, head))
4478 return;
4479
4480 list_for_each_entry_continue(next, head, set_list)
4481 if (ses->entry_cnt <= next->entry_cnt) {
4482 list_move_tail(&ses->set_list, &next->set_list);
4483 return;
4484 }
4485
4486 list_move_tail(&ses->set_list, head);
4487}
4488
4489static void add_sit_entry(unsigned int segno, struct list_head *head)
4490{
4491 struct sit_entry_set *ses;
4492 unsigned int start_segno = START_SEGNO(segno);
4493
4494 list_for_each_entry(ses, head, set_list) {
4495 if (ses->start_segno == start_segno) {
4496 ses->entry_cnt++;
4497 adjust_sit_entry_set(ses, head);
4498 return;
4499 }
4500 }
4501
4502 ses = grab_sit_entry_set();
4503
4504 ses->start_segno = start_segno;
4505 ses->entry_cnt++;
4506 list_add(&ses->set_list, head);
4507}
4508
4509static void add_sits_in_set(struct f2fs_sb_info *sbi)
4510{
4511 struct f2fs_sm_info *sm_info = SM_I(sbi);
4512 struct list_head *set_list = &sm_info->sit_entry_set;
4513 unsigned long *bitmap = SIT_I(sbi)->dirty_sentries_bitmap;
4514 unsigned int segno;
4515
4516 for_each_set_bit(segno, bitmap, MAIN_SEGS(sbi))
4517 add_sit_entry(segno, set_list);
4518}
4519
4520static void remove_sits_in_journal(struct f2fs_sb_info *sbi)
4521{
4522 struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_COLD_DATA);
4523 struct f2fs_journal *journal = curseg->journal;
4524 int i;
4525
4526 down_write(&curseg->journal_rwsem);
4527 for (i = 0; i < sits_in_cursum(journal); i++) {
4528 unsigned int segno;
4529 bool dirtied;
4530
4531 segno = le32_to_cpu(segno_in_journal(journal, i));
4532 dirtied = __mark_sit_entry_dirty(sbi, segno);
4533
4534 if (!dirtied)
4535 add_sit_entry(segno, &SM_I(sbi)->sit_entry_set);
4536 }
4537 update_sits_in_cursum(journal, -i);
4538 up_write(&curseg->journal_rwsem);
4539}
4540
4541/*
4542 * CP calls this function, which flushes SIT entries including sit_journal,
4543 * and moves prefree segs to free segs.
4544 */
4545void f2fs_flush_sit_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc)
4546{
4547 struct sit_info *sit_i = SIT_I(sbi);
4548 unsigned long *bitmap = sit_i->dirty_sentries_bitmap;
4549 struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_COLD_DATA);
4550 struct f2fs_journal *journal = curseg->journal;
4551 struct sit_entry_set *ses, *tmp;
4552 struct list_head *head = &SM_I(sbi)->sit_entry_set;
4553 bool to_journal = !is_sbi_flag_set(sbi, SBI_IS_RESIZEFS);
4554 struct seg_entry *se;
4555
4556 down_write(&sit_i->sentry_lock);
4557
4558 if (!sit_i->dirty_sentries)
4559 goto out;
4560
4561 /*
4562 * add and account sit entries of dirty bitmap in sit entry
4563 * set temporarily
4564 */
4565 add_sits_in_set(sbi);
4566
4567 /*
4568 * if there are no enough space in journal to store dirty sit
4569 * entries, remove all entries from journal and add and account
4570 * them in sit entry set.
4571 */
4572 if (!__has_cursum_space(journal, sit_i->dirty_sentries, SIT_JOURNAL) ||
4573 !to_journal)
4574 remove_sits_in_journal(sbi);
4575
4576 /*
4577 * there are two steps to flush sit entries:
4578 * #1, flush sit entries to journal in current cold data summary block.
4579 * #2, flush sit entries to sit page.
4580 */
4581 list_for_each_entry_safe(ses, tmp, head, set_list) {
4582 struct page *page = NULL;
4583 struct f2fs_sit_block *raw_sit = NULL;
4584 unsigned int start_segno = ses->start_segno;
4585 unsigned int end = min(start_segno + SIT_ENTRY_PER_BLOCK,
4586 (unsigned long)MAIN_SEGS(sbi));
4587 unsigned int segno = start_segno;
4588
4589 if (to_journal &&
4590 !__has_cursum_space(journal, ses->entry_cnt, SIT_JOURNAL))
4591 to_journal = false;
4592
4593 if (to_journal) {
4594 down_write(&curseg->journal_rwsem);
4595 } else {
4596 page = get_next_sit_page(sbi, start_segno);
4597 raw_sit = page_address(page);
4598 }
4599
4600 /* flush dirty sit entries in region of current sit set */
4601 for_each_set_bit_from(segno, bitmap, end) {
4602 int offset, sit_offset;
4603
4604 se = get_seg_entry(sbi, segno);
4605#ifdef CONFIG_F2FS_CHECK_FS
4606 if (memcmp(se->cur_valid_map, se->cur_valid_map_mir,
4607 SIT_VBLOCK_MAP_SIZE))
4608 f2fs_bug_on(sbi, 1);
4609#endif
4610
4611 /* add discard candidates */
4612 if (!(cpc->reason & CP_DISCARD)) {
4613 cpc->trim_start = segno;
4614 add_discard_addrs(sbi, cpc, false);
4615 }
4616
4617 if (to_journal) {
4618 offset = f2fs_lookup_journal_in_cursum(journal,
4619 SIT_JOURNAL, segno, 1);
4620 f2fs_bug_on(sbi, offset < 0);
4621 segno_in_journal(journal, offset) =
4622 cpu_to_le32(segno);
4623 seg_info_to_raw_sit(se,
4624 &sit_in_journal(journal, offset));
4625 check_block_count(sbi, segno,
4626 &sit_in_journal(journal, offset));
4627 } else {
4628 sit_offset = SIT_ENTRY_OFFSET(sit_i, segno);
4629 seg_info_to_raw_sit(se,
4630 &raw_sit->entries[sit_offset]);
4631 check_block_count(sbi, segno,
4632 &raw_sit->entries[sit_offset]);
4633 }
4634
4635 __clear_bit(segno, bitmap);
4636 sit_i->dirty_sentries--;
4637 ses->entry_cnt--;
4638 }
4639
4640 if (to_journal)
4641 up_write(&curseg->journal_rwsem);
4642 else
4643 f2fs_put_page(page, 1);
4644
4645 f2fs_bug_on(sbi, ses->entry_cnt);
4646 release_sit_entry_set(ses);
4647 }
4648
4649 f2fs_bug_on(sbi, !list_empty(head));
4650 f2fs_bug_on(sbi, sit_i->dirty_sentries);
4651out:
4652 if (cpc->reason & CP_DISCARD) {
4653 __u64 trim_start = cpc->trim_start;
4654
4655 for (; cpc->trim_start <= cpc->trim_end; cpc->trim_start++)
4656 add_discard_addrs(sbi, cpc, false);
4657
4658 cpc->trim_start = trim_start;
4659 }
4660 up_write(&sit_i->sentry_lock);
4661
4662 set_prefree_as_free_segments(sbi);
4663}
4664
4665static int build_sit_info(struct f2fs_sb_info *sbi)
4666{
4667 struct f2fs_super_block *raw_super = F2FS_RAW_SUPER(sbi);
4668 struct sit_info *sit_i;
4669 unsigned int sit_segs, start;
4670 char *src_bitmap, *bitmap;
4671 unsigned int bitmap_size, main_bitmap_size, sit_bitmap_size;
4672 unsigned int discard_map = f2fs_block_unit_discard(sbi) ? 1 : 0;
4673
4674 /* allocate memory for SIT information */
4675 sit_i = f2fs_kzalloc(sbi, sizeof(struct sit_info), GFP_KERNEL);
4676 if (!sit_i)
4677 return -ENOMEM;
4678
4679 SM_I(sbi)->sit_info = sit_i;
4680
4681 sit_i->sentries =
4682 f2fs_kvzalloc(sbi, array_size(sizeof(struct seg_entry),
4683 MAIN_SEGS(sbi)),
4684 GFP_KERNEL);
4685 if (!sit_i->sentries)
4686 return -ENOMEM;
4687
4688 main_bitmap_size = f2fs_bitmap_size(MAIN_SEGS(sbi));
4689 sit_i->dirty_sentries_bitmap = f2fs_kvzalloc(sbi, main_bitmap_size,
4690 GFP_KERNEL);
4691 if (!sit_i->dirty_sentries_bitmap)
4692 return -ENOMEM;
4693
4694#ifdef CONFIG_F2FS_CHECK_FS
4695 bitmap_size = MAIN_SEGS(sbi) * SIT_VBLOCK_MAP_SIZE * (3 + discard_map);
4696#else
4697 bitmap_size = MAIN_SEGS(sbi) * SIT_VBLOCK_MAP_SIZE * (2 + discard_map);
4698#endif
4699 sit_i->bitmap = f2fs_kvzalloc(sbi, bitmap_size, GFP_KERNEL);
4700 if (!sit_i->bitmap)
4701 return -ENOMEM;
4702
4703 bitmap = sit_i->bitmap;
4704
4705 for (start = 0; start < MAIN_SEGS(sbi); start++) {
4706 sit_i->sentries[start].cur_valid_map = bitmap;
4707 bitmap += SIT_VBLOCK_MAP_SIZE;
4708
4709 sit_i->sentries[start].ckpt_valid_map = bitmap;
4710 bitmap += SIT_VBLOCK_MAP_SIZE;
4711
4712#ifdef CONFIG_F2FS_CHECK_FS
4713 sit_i->sentries[start].cur_valid_map_mir = bitmap;
4714 bitmap += SIT_VBLOCK_MAP_SIZE;
4715#endif
4716
4717 if (discard_map) {
4718 sit_i->sentries[start].discard_map = bitmap;
4719 bitmap += SIT_VBLOCK_MAP_SIZE;
4720 }
4721 }
4722
4723 sit_i->tmp_map = f2fs_kzalloc(sbi, SIT_VBLOCK_MAP_SIZE, GFP_KERNEL);
4724 if (!sit_i->tmp_map)
4725 return -ENOMEM;
4726
4727 if (__is_large_section(sbi)) {
4728 sit_i->sec_entries =
4729 f2fs_kvzalloc(sbi, array_size(sizeof(struct sec_entry),
4730 MAIN_SECS(sbi)),
4731 GFP_KERNEL);
4732 if (!sit_i->sec_entries)
4733 return -ENOMEM;
4734 }
4735
4736 /* get information related with SIT */
4737 sit_segs = le32_to_cpu(raw_super->segment_count_sit) >> 1;
4738
4739 /* setup SIT bitmap from ckeckpoint pack */
4740 sit_bitmap_size = __bitmap_size(sbi, SIT_BITMAP);
4741 src_bitmap = __bitmap_ptr(sbi, SIT_BITMAP);
4742
4743 sit_i->sit_bitmap = kmemdup(src_bitmap, sit_bitmap_size, GFP_KERNEL);
4744 if (!sit_i->sit_bitmap)
4745 return -ENOMEM;
4746
4747#ifdef CONFIG_F2FS_CHECK_FS
4748 sit_i->sit_bitmap_mir = kmemdup(src_bitmap,
4749 sit_bitmap_size, GFP_KERNEL);
4750 if (!sit_i->sit_bitmap_mir)
4751 return -ENOMEM;
4752
4753 sit_i->invalid_segmap = f2fs_kvzalloc(sbi,
4754 main_bitmap_size, GFP_KERNEL);
4755 if (!sit_i->invalid_segmap)
4756 return -ENOMEM;
4757#endif
4758
4759 sit_i->sit_base_addr = le32_to_cpu(raw_super->sit_blkaddr);
4760 sit_i->sit_blocks = SEGS_TO_BLKS(sbi, sit_segs);
4761 sit_i->written_valid_blocks = 0;
4762 sit_i->bitmap_size = sit_bitmap_size;
4763 sit_i->dirty_sentries = 0;
4764 sit_i->sents_per_block = SIT_ENTRY_PER_BLOCK;
4765 sit_i->elapsed_time = le64_to_cpu(sbi->ckpt->elapsed_time);
4766 sit_i->mounted_time = ktime_get_boottime_seconds();
4767 init_rwsem(&sit_i->sentry_lock);
4768 return 0;
4769}
4770
4771static int build_free_segmap(struct f2fs_sb_info *sbi)
4772{
4773 struct free_segmap_info *free_i;
4774 unsigned int bitmap_size, sec_bitmap_size;
4775
4776 /* allocate memory for free segmap information */
4777 free_i = f2fs_kzalloc(sbi, sizeof(struct free_segmap_info), GFP_KERNEL);
4778 if (!free_i)
4779 return -ENOMEM;
4780
4781 SM_I(sbi)->free_info = free_i;
4782
4783 bitmap_size = f2fs_bitmap_size(MAIN_SEGS(sbi));
4784 free_i->free_segmap = f2fs_kvmalloc(sbi, bitmap_size, GFP_KERNEL);
4785 if (!free_i->free_segmap)
4786 return -ENOMEM;
4787
4788 sec_bitmap_size = f2fs_bitmap_size(MAIN_SECS(sbi));
4789 free_i->free_secmap = f2fs_kvmalloc(sbi, sec_bitmap_size, GFP_KERNEL);
4790 if (!free_i->free_secmap)
4791 return -ENOMEM;
4792
4793 /* set all segments as dirty temporarily */
4794 memset(free_i->free_segmap, 0xff, bitmap_size);
4795 memset(free_i->free_secmap, 0xff, sec_bitmap_size);
4796
4797 /* init free segmap information */
4798 free_i->start_segno = GET_SEGNO_FROM_SEG0(sbi, MAIN_BLKADDR(sbi));
4799 free_i->free_segments = 0;
4800 free_i->free_sections = 0;
4801 spin_lock_init(&free_i->segmap_lock);
4802 return 0;
4803}
4804
4805static int build_curseg(struct f2fs_sb_info *sbi)
4806{
4807 struct curseg_info *array;
4808 int i;
4809
4810 array = f2fs_kzalloc(sbi, array_size(NR_CURSEG_TYPE,
4811 sizeof(*array)), GFP_KERNEL);
4812 if (!array)
4813 return -ENOMEM;
4814
4815 SM_I(sbi)->curseg_array = array;
4816
4817 for (i = 0; i < NO_CHECK_TYPE; i++) {
4818 mutex_init(&array[i].curseg_mutex);
4819 array[i].sum_blk = f2fs_kzalloc(sbi, PAGE_SIZE, GFP_KERNEL);
4820 if (!array[i].sum_blk)
4821 return -ENOMEM;
4822 init_rwsem(&array[i].journal_rwsem);
4823 array[i].journal = f2fs_kzalloc(sbi,
4824 sizeof(struct f2fs_journal), GFP_KERNEL);
4825 if (!array[i].journal)
4826 return -ENOMEM;
4827 array[i].seg_type = log_type_to_seg_type(i);
4828 reset_curseg_fields(&array[i]);
4829 }
4830 return restore_curseg_summaries(sbi);
4831}
4832
4833static int build_sit_entries(struct f2fs_sb_info *sbi)
4834{
4835 struct sit_info *sit_i = SIT_I(sbi);
4836 struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_COLD_DATA);
4837 struct f2fs_journal *journal = curseg->journal;
4838 struct seg_entry *se;
4839 struct f2fs_sit_entry sit;
4840 int sit_blk_cnt = SIT_BLK_CNT(sbi);
4841 unsigned int i, start, end;
4842 unsigned int readed, start_blk = 0;
4843 int err = 0;
4844 block_t sit_valid_blocks[2] = {0, 0};
4845
4846 do {
4847 readed = f2fs_ra_meta_pages(sbi, start_blk, BIO_MAX_VECS,
4848 META_SIT, true);
4849
4850 start = start_blk * sit_i->sents_per_block;
4851 end = (start_blk + readed) * sit_i->sents_per_block;
4852
4853 for (; start < end && start < MAIN_SEGS(sbi); start++) {
4854 struct f2fs_sit_block *sit_blk;
4855 struct page *page;
4856
4857 se = &sit_i->sentries[start];
4858 page = get_current_sit_page(sbi, start);
4859 if (IS_ERR(page))
4860 return PTR_ERR(page);
4861 sit_blk = (struct f2fs_sit_block *)page_address(page);
4862 sit = sit_blk->entries[SIT_ENTRY_OFFSET(sit_i, start)];
4863 f2fs_put_page(page, 1);
4864
4865 err = check_block_count(sbi, start, &sit);
4866 if (err)
4867 return err;
4868 seg_info_from_raw_sit(se, &sit);
4869
4870 if (se->type >= NR_PERSISTENT_LOG) {
4871 f2fs_err(sbi, "Invalid segment type: %u, segno: %u",
4872 se->type, start);
4873 f2fs_handle_error(sbi,
4874 ERROR_INCONSISTENT_SUM_TYPE);
4875 return -EFSCORRUPTED;
4876 }
4877
4878 sit_valid_blocks[SE_PAGETYPE(se)] += se->valid_blocks;
4879
4880 if (!f2fs_block_unit_discard(sbi))
4881 goto init_discard_map_done;
4882
4883 /* build discard map only one time */
4884 if (is_set_ckpt_flags(sbi, CP_TRIMMED_FLAG)) {
4885 memset(se->discard_map, 0xff,
4886 SIT_VBLOCK_MAP_SIZE);
4887 goto init_discard_map_done;
4888 }
4889 memcpy(se->discard_map, se->cur_valid_map,
4890 SIT_VBLOCK_MAP_SIZE);
4891 sbi->discard_blks += BLKS_PER_SEG(sbi) -
4892 se->valid_blocks;
4893init_discard_map_done:
4894 if (__is_large_section(sbi))
4895 get_sec_entry(sbi, start)->valid_blocks +=
4896 se->valid_blocks;
4897 }
4898 start_blk += readed;
4899 } while (start_blk < sit_blk_cnt);
4900
4901 down_read(&curseg->journal_rwsem);
4902 for (i = 0; i < sits_in_cursum(journal); i++) {
4903 unsigned int old_valid_blocks;
4904
4905 start = le32_to_cpu(segno_in_journal(journal, i));
4906 if (start >= MAIN_SEGS(sbi)) {
4907 f2fs_err(sbi, "Wrong journal entry on segno %u",
4908 start);
4909 err = -EFSCORRUPTED;
4910 f2fs_handle_error(sbi, ERROR_CORRUPTED_JOURNAL);
4911 break;
4912 }
4913
4914 se = &sit_i->sentries[start];
4915 sit = sit_in_journal(journal, i);
4916
4917 old_valid_blocks = se->valid_blocks;
4918
4919 sit_valid_blocks[SE_PAGETYPE(se)] -= old_valid_blocks;
4920
4921 err = check_block_count(sbi, start, &sit);
4922 if (err)
4923 break;
4924 seg_info_from_raw_sit(se, &sit);
4925
4926 if (se->type >= NR_PERSISTENT_LOG) {
4927 f2fs_err(sbi, "Invalid segment type: %u, segno: %u",
4928 se->type, start);
4929 err = -EFSCORRUPTED;
4930 f2fs_handle_error(sbi, ERROR_INCONSISTENT_SUM_TYPE);
4931 break;
4932 }
4933
4934 sit_valid_blocks[SE_PAGETYPE(se)] += se->valid_blocks;
4935
4936 if (f2fs_block_unit_discard(sbi)) {
4937 if (is_set_ckpt_flags(sbi, CP_TRIMMED_FLAG)) {
4938 memset(se->discard_map, 0xff, SIT_VBLOCK_MAP_SIZE);
4939 } else {
4940 memcpy(se->discard_map, se->cur_valid_map,
4941 SIT_VBLOCK_MAP_SIZE);
4942 sbi->discard_blks += old_valid_blocks;
4943 sbi->discard_blks -= se->valid_blocks;
4944 }
4945 }
4946
4947 if (__is_large_section(sbi)) {
4948 get_sec_entry(sbi, start)->valid_blocks +=
4949 se->valid_blocks;
4950 get_sec_entry(sbi, start)->valid_blocks -=
4951 old_valid_blocks;
4952 }
4953 }
4954 up_read(&curseg->journal_rwsem);
4955
4956 if (err)
4957 return err;
4958
4959 if (sit_valid_blocks[NODE] != valid_node_count(sbi)) {
4960 f2fs_err(sbi, "SIT is corrupted node# %u vs %u",
4961 sit_valid_blocks[NODE], valid_node_count(sbi));
4962 f2fs_handle_error(sbi, ERROR_INCONSISTENT_NODE_COUNT);
4963 return -EFSCORRUPTED;
4964 }
4965
4966 if (sit_valid_blocks[DATA] + sit_valid_blocks[NODE] >
4967 valid_user_blocks(sbi)) {
4968 f2fs_err(sbi, "SIT is corrupted data# %u %u vs %u",
4969 sit_valid_blocks[DATA], sit_valid_blocks[NODE],
4970 valid_user_blocks(sbi));
4971 f2fs_handle_error(sbi, ERROR_INCONSISTENT_BLOCK_COUNT);
4972 return -EFSCORRUPTED;
4973 }
4974
4975 return 0;
4976}
4977
4978static void init_free_segmap(struct f2fs_sb_info *sbi)
4979{
4980 unsigned int start;
4981 int type;
4982 struct seg_entry *sentry;
4983
4984 for (start = 0; start < MAIN_SEGS(sbi); start++) {
4985 if (f2fs_usable_blks_in_seg(sbi, start) == 0)
4986 continue;
4987 sentry = get_seg_entry(sbi, start);
4988 if (!sentry->valid_blocks)
4989 __set_free(sbi, start);
4990 else
4991 SIT_I(sbi)->written_valid_blocks +=
4992 sentry->valid_blocks;
4993 }
4994
4995 /* set use the current segments */
4996 for (type = CURSEG_HOT_DATA; type <= CURSEG_COLD_NODE; type++) {
4997 struct curseg_info *curseg_t = CURSEG_I(sbi, type);
4998
4999 __set_test_and_inuse(sbi, curseg_t->segno);
5000 }
5001}
5002
5003static void init_dirty_segmap(struct f2fs_sb_info *sbi)
5004{
5005 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
5006 struct free_segmap_info *free_i = FREE_I(sbi);
5007 unsigned int segno = 0, offset = 0, secno;
5008 block_t valid_blocks, usable_blks_in_seg;
5009
5010 while (1) {
5011 /* find dirty segment based on free segmap */
5012 segno = find_next_inuse(free_i, MAIN_SEGS(sbi), offset);
5013 if (segno >= MAIN_SEGS(sbi))
5014 break;
5015 offset = segno + 1;
5016 valid_blocks = get_valid_blocks(sbi, segno, false);
5017 usable_blks_in_seg = f2fs_usable_blks_in_seg(sbi, segno);
5018 if (valid_blocks == usable_blks_in_seg || !valid_blocks)
5019 continue;
5020 if (valid_blocks > usable_blks_in_seg) {
5021 f2fs_bug_on(sbi, 1);
5022 continue;
5023 }
5024 mutex_lock(&dirty_i->seglist_lock);
5025 __locate_dirty_segment(sbi, segno, DIRTY);
5026 mutex_unlock(&dirty_i->seglist_lock);
5027 }
5028
5029 if (!__is_large_section(sbi))
5030 return;
5031
5032 mutex_lock(&dirty_i->seglist_lock);
5033 for (segno = 0; segno < MAIN_SEGS(sbi); segno += SEGS_PER_SEC(sbi)) {
5034 valid_blocks = get_valid_blocks(sbi, segno, true);
5035 secno = GET_SEC_FROM_SEG(sbi, segno);
5036
5037 if (!valid_blocks || valid_blocks == CAP_BLKS_PER_SEC(sbi))
5038 continue;
5039 if (IS_CURSEC(sbi, secno))
5040 continue;
5041 set_bit(secno, dirty_i->dirty_secmap);
5042 }
5043 mutex_unlock(&dirty_i->seglist_lock);
5044}
5045
5046static int init_victim_secmap(struct f2fs_sb_info *sbi)
5047{
5048 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
5049 unsigned int bitmap_size = f2fs_bitmap_size(MAIN_SECS(sbi));
5050
5051 dirty_i->victim_secmap = f2fs_kvzalloc(sbi, bitmap_size, GFP_KERNEL);
5052 if (!dirty_i->victim_secmap)
5053 return -ENOMEM;
5054
5055 dirty_i->pinned_secmap = f2fs_kvzalloc(sbi, bitmap_size, GFP_KERNEL);
5056 if (!dirty_i->pinned_secmap)
5057 return -ENOMEM;
5058
5059 dirty_i->pinned_secmap_cnt = 0;
5060 dirty_i->enable_pin_section = true;
5061 return 0;
5062}
5063
5064static int build_dirty_segmap(struct f2fs_sb_info *sbi)
5065{
5066 struct dirty_seglist_info *dirty_i;
5067 unsigned int bitmap_size, i;
5068
5069 /* allocate memory for dirty segments list information */
5070 dirty_i = f2fs_kzalloc(sbi, sizeof(struct dirty_seglist_info),
5071 GFP_KERNEL);
5072 if (!dirty_i)
5073 return -ENOMEM;
5074
5075 SM_I(sbi)->dirty_info = dirty_i;
5076 mutex_init(&dirty_i->seglist_lock);
5077
5078 bitmap_size = f2fs_bitmap_size(MAIN_SEGS(sbi));
5079
5080 for (i = 0; i < NR_DIRTY_TYPE; i++) {
5081 dirty_i->dirty_segmap[i] = f2fs_kvzalloc(sbi, bitmap_size,
5082 GFP_KERNEL);
5083 if (!dirty_i->dirty_segmap[i])
5084 return -ENOMEM;
5085 }
5086
5087 if (__is_large_section(sbi)) {
5088 bitmap_size = f2fs_bitmap_size(MAIN_SECS(sbi));
5089 dirty_i->dirty_secmap = f2fs_kvzalloc(sbi,
5090 bitmap_size, GFP_KERNEL);
5091 if (!dirty_i->dirty_secmap)
5092 return -ENOMEM;
5093 }
5094
5095 init_dirty_segmap(sbi);
5096 return init_victim_secmap(sbi);
5097}
5098
5099static int sanity_check_curseg(struct f2fs_sb_info *sbi)
5100{
5101 int i;
5102
5103 /*
5104 * In LFS/SSR curseg, .next_blkoff should point to an unused blkaddr;
5105 * In LFS curseg, all blkaddr after .next_blkoff should be unused.
5106 */
5107 for (i = 0; i < NR_PERSISTENT_LOG; i++) {
5108 struct curseg_info *curseg = CURSEG_I(sbi, i);
5109 struct seg_entry *se = get_seg_entry(sbi, curseg->segno);
5110 unsigned int blkofs = curseg->next_blkoff;
5111
5112 if (f2fs_sb_has_readonly(sbi) &&
5113 i != CURSEG_HOT_DATA && i != CURSEG_HOT_NODE)
5114 continue;
5115
5116 sanity_check_seg_type(sbi, curseg->seg_type);
5117
5118 if (curseg->alloc_type != LFS && curseg->alloc_type != SSR) {
5119 f2fs_err(sbi,
5120 "Current segment has invalid alloc_type:%d",
5121 curseg->alloc_type);
5122 f2fs_handle_error(sbi, ERROR_INVALID_CURSEG);
5123 return -EFSCORRUPTED;
5124 }
5125
5126 if (f2fs_test_bit(blkofs, se->cur_valid_map))
5127 goto out;
5128
5129 if (curseg->alloc_type == SSR)
5130 continue;
5131
5132 for (blkofs += 1; blkofs < BLKS_PER_SEG(sbi); blkofs++) {
5133 if (!f2fs_test_bit(blkofs, se->cur_valid_map))
5134 continue;
5135out:
5136 f2fs_err(sbi,
5137 "Current segment's next free block offset is inconsistent with bitmap, logtype:%u, segno:%u, type:%u, next_blkoff:%u, blkofs:%u",
5138 i, curseg->segno, curseg->alloc_type,
5139 curseg->next_blkoff, blkofs);
5140 f2fs_handle_error(sbi, ERROR_INVALID_CURSEG);
5141 return -EFSCORRUPTED;
5142 }
5143 }
5144 return 0;
5145}
5146
5147#ifdef CONFIG_BLK_DEV_ZONED
5148static int check_zone_write_pointer(struct f2fs_sb_info *sbi,
5149 struct f2fs_dev_info *fdev,
5150 struct blk_zone *zone)
5151{
5152 unsigned int zone_segno;
5153 block_t zone_block, valid_block_cnt;
5154 unsigned int log_sectors_per_block = sbi->log_blocksize - SECTOR_SHIFT;
5155 int ret;
5156 unsigned int nofs_flags;
5157
5158 if (zone->type != BLK_ZONE_TYPE_SEQWRITE_REQ)
5159 return 0;
5160
5161 zone_block = fdev->start_blk + (zone->start >> log_sectors_per_block);
5162 zone_segno = GET_SEGNO(sbi, zone_block);
5163
5164 /*
5165 * Skip check of zones cursegs point to, since
5166 * fix_curseg_write_pointer() checks them.
5167 */
5168 if (zone_segno >= MAIN_SEGS(sbi))
5169 return 0;
5170
5171 /*
5172 * Get # of valid block of the zone.
5173 */
5174 valid_block_cnt = get_valid_blocks(sbi, zone_segno, true);
5175 if (IS_CURSEC(sbi, GET_SEC_FROM_SEG(sbi, zone_segno))) {
5176 f2fs_notice(sbi, "Open zones: valid block[0x%x,0x%x] cond[%s]",
5177 zone_segno, valid_block_cnt,
5178 blk_zone_cond_str(zone->cond));
5179 return 0;
5180 }
5181
5182 if ((!valid_block_cnt && zone->cond == BLK_ZONE_COND_EMPTY) ||
5183 (valid_block_cnt && zone->cond == BLK_ZONE_COND_FULL))
5184 return 0;
5185
5186 if (!valid_block_cnt) {
5187 f2fs_notice(sbi, "Zone without valid block has non-zero write "
5188 "pointer. Reset the write pointer: cond[%s]",
5189 blk_zone_cond_str(zone->cond));
5190 ret = __f2fs_issue_discard_zone(sbi, fdev->bdev, zone_block,
5191 zone->len >> log_sectors_per_block);
5192 if (ret)
5193 f2fs_err(sbi, "Discard zone failed: %s (errno=%d)",
5194 fdev->path, ret);
5195 return ret;
5196 }
5197
5198 /*
5199 * If there are valid blocks and the write pointer doesn't match
5200 * with them, we need to report the inconsistency and fill
5201 * the zone till the end to close the zone. This inconsistency
5202 * does not cause write error because the zone will not be
5203 * selected for write operation until it get discarded.
5204 */
5205 f2fs_notice(sbi, "Valid blocks are not aligned with write "
5206 "pointer: valid block[0x%x,0x%x] cond[%s]",
5207 zone_segno, valid_block_cnt, blk_zone_cond_str(zone->cond));
5208
5209 nofs_flags = memalloc_nofs_save();
5210 ret = blkdev_zone_mgmt(fdev->bdev, REQ_OP_ZONE_FINISH,
5211 zone->start, zone->len);
5212 memalloc_nofs_restore(nofs_flags);
5213 if (ret == -EOPNOTSUPP) {
5214 ret = blkdev_issue_zeroout(fdev->bdev, zone->wp,
5215 zone->len - (zone->wp - zone->start),
5216 GFP_NOFS, 0);
5217 if (ret)
5218 f2fs_err(sbi, "Fill up zone failed: %s (errno=%d)",
5219 fdev->path, ret);
5220 } else if (ret) {
5221 f2fs_err(sbi, "Finishing zone failed: %s (errno=%d)",
5222 fdev->path, ret);
5223 }
5224
5225 return ret;
5226}
5227
5228static struct f2fs_dev_info *get_target_zoned_dev(struct f2fs_sb_info *sbi,
5229 block_t zone_blkaddr)
5230{
5231 int i;
5232
5233 for (i = 0; i < sbi->s_ndevs; i++) {
5234 if (!bdev_is_zoned(FDEV(i).bdev))
5235 continue;
5236 if (sbi->s_ndevs == 1 || (FDEV(i).start_blk <= zone_blkaddr &&
5237 zone_blkaddr <= FDEV(i).end_blk))
5238 return &FDEV(i);
5239 }
5240
5241 return NULL;
5242}
5243
5244static int report_one_zone_cb(struct blk_zone *zone, unsigned int idx,
5245 void *data)
5246{
5247 memcpy(data, zone, sizeof(struct blk_zone));
5248 return 0;
5249}
5250
5251static int do_fix_curseg_write_pointer(struct f2fs_sb_info *sbi, int type)
5252{
5253 struct curseg_info *cs = CURSEG_I(sbi, type);
5254 struct f2fs_dev_info *zbd;
5255 struct blk_zone zone;
5256 unsigned int cs_section, wp_segno, wp_blkoff, wp_sector_off;
5257 block_t cs_zone_block, wp_block;
5258 unsigned int log_sectors_per_block = sbi->log_blocksize - SECTOR_SHIFT;
5259 sector_t zone_sector;
5260 int err;
5261
5262 cs_section = GET_SEC_FROM_SEG(sbi, cs->segno);
5263 cs_zone_block = START_BLOCK(sbi, GET_SEG_FROM_SEC(sbi, cs_section));
5264
5265 zbd = get_target_zoned_dev(sbi, cs_zone_block);
5266 if (!zbd)
5267 return 0;
5268
5269 /* report zone for the sector the curseg points to */
5270 zone_sector = (sector_t)(cs_zone_block - zbd->start_blk)
5271 << log_sectors_per_block;
5272 err = blkdev_report_zones(zbd->bdev, zone_sector, 1,
5273 report_one_zone_cb, &zone);
5274 if (err != 1) {
5275 f2fs_err(sbi, "Report zone failed: %s errno=(%d)",
5276 zbd->path, err);
5277 return err;
5278 }
5279
5280 if (zone.type != BLK_ZONE_TYPE_SEQWRITE_REQ)
5281 return 0;
5282
5283 /*
5284 * When safely unmounted in the previous mount, we could use current
5285 * segments. Otherwise, allocate new sections.
5286 */
5287 if (is_set_ckpt_flags(sbi, CP_UMOUNT_FLAG)) {
5288 wp_block = zbd->start_blk + (zone.wp >> log_sectors_per_block);
5289 wp_segno = GET_SEGNO(sbi, wp_block);
5290 wp_blkoff = wp_block - START_BLOCK(sbi, wp_segno);
5291 wp_sector_off = zone.wp & GENMASK(log_sectors_per_block - 1, 0);
5292
5293 if (cs->segno == wp_segno && cs->next_blkoff == wp_blkoff &&
5294 wp_sector_off == 0)
5295 return 0;
5296
5297 f2fs_notice(sbi, "Unaligned curseg[%d] with write pointer: "
5298 "curseg[0x%x,0x%x] wp[0x%x,0x%x]", type, cs->segno,
5299 cs->next_blkoff, wp_segno, wp_blkoff);
5300 }
5301
5302 /* Allocate a new section if it's not new. */
5303 if (cs->next_blkoff ||
5304 cs->segno != GET_SEG_FROM_SEC(sbi, GET_ZONE_FROM_SEC(sbi, cs_section))) {
5305 unsigned int old_segno = cs->segno, old_blkoff = cs->next_blkoff;
5306
5307 f2fs_allocate_new_section(sbi, type, true);
5308 f2fs_notice(sbi, "Assign new section to curseg[%d]: "
5309 "[0x%x,0x%x] -> [0x%x,0x%x]",
5310 type, old_segno, old_blkoff,
5311 cs->segno, cs->next_blkoff);
5312 }
5313
5314 /* check consistency of the zone curseg pointed to */
5315 if (check_zone_write_pointer(sbi, zbd, &zone))
5316 return -EIO;
5317
5318 /* check newly assigned zone */
5319 cs_section = GET_SEC_FROM_SEG(sbi, cs->segno);
5320 cs_zone_block = START_BLOCK(sbi, GET_SEG_FROM_SEC(sbi, cs_section));
5321
5322 zbd = get_target_zoned_dev(sbi, cs_zone_block);
5323 if (!zbd)
5324 return 0;
5325
5326 zone_sector = (sector_t)(cs_zone_block - zbd->start_blk)
5327 << log_sectors_per_block;
5328 err = blkdev_report_zones(zbd->bdev, zone_sector, 1,
5329 report_one_zone_cb, &zone);
5330 if (err != 1) {
5331 f2fs_err(sbi, "Report zone failed: %s errno=(%d)",
5332 zbd->path, err);
5333 return err;
5334 }
5335
5336 if (zone.type != BLK_ZONE_TYPE_SEQWRITE_REQ)
5337 return 0;
5338
5339 if (zone.wp != zone.start) {
5340 f2fs_notice(sbi,
5341 "New zone for curseg[%d] is not yet discarded. "
5342 "Reset the zone: curseg[0x%x,0x%x]",
5343 type, cs->segno, cs->next_blkoff);
5344 err = __f2fs_issue_discard_zone(sbi, zbd->bdev, cs_zone_block,
5345 zone.len >> log_sectors_per_block);
5346 if (err) {
5347 f2fs_err(sbi, "Discard zone failed: %s (errno=%d)",
5348 zbd->path, err);
5349 return err;
5350 }
5351 }
5352
5353 return 0;
5354}
5355
5356static int fix_curseg_write_pointer(struct f2fs_sb_info *sbi)
5357{
5358 int i, ret;
5359
5360 for (i = 0; i < NR_PERSISTENT_LOG; i++) {
5361 ret = do_fix_curseg_write_pointer(sbi, i);
5362 if (ret)
5363 return ret;
5364 }
5365
5366 return 0;
5367}
5368
5369struct check_zone_write_pointer_args {
5370 struct f2fs_sb_info *sbi;
5371 struct f2fs_dev_info *fdev;
5372};
5373
5374static int check_zone_write_pointer_cb(struct blk_zone *zone, unsigned int idx,
5375 void *data)
5376{
5377 struct check_zone_write_pointer_args *args;
5378
5379 args = (struct check_zone_write_pointer_args *)data;
5380
5381 return check_zone_write_pointer(args->sbi, args->fdev, zone);
5382}
5383
5384static int check_write_pointer(struct f2fs_sb_info *sbi)
5385{
5386 int i, ret;
5387 struct check_zone_write_pointer_args args;
5388
5389 for (i = 0; i < sbi->s_ndevs; i++) {
5390 if (!bdev_is_zoned(FDEV(i).bdev))
5391 continue;
5392
5393 args.sbi = sbi;
5394 args.fdev = &FDEV(i);
5395 ret = blkdev_report_zones(FDEV(i).bdev, 0, BLK_ALL_ZONES,
5396 check_zone_write_pointer_cb, &args);
5397 if (ret < 0)
5398 return ret;
5399 }
5400
5401 return 0;
5402}
5403
5404int f2fs_check_and_fix_write_pointer(struct f2fs_sb_info *sbi)
5405{
5406 int ret;
5407
5408 if (!f2fs_sb_has_blkzoned(sbi) || f2fs_readonly(sbi->sb))
5409 return 0;
5410
5411 f2fs_notice(sbi, "Checking entire write pointers");
5412 ret = fix_curseg_write_pointer(sbi);
5413 if (!ret)
5414 ret = check_write_pointer(sbi);
5415 return ret;
5416}
5417
5418/*
5419 * Return the number of usable blocks in a segment. The number of blocks
5420 * returned is always equal to the number of blocks in a segment for
5421 * segments fully contained within a sequential zone capacity or a
5422 * conventional zone. For segments partially contained in a sequential
5423 * zone capacity, the number of usable blocks up to the zone capacity
5424 * is returned. 0 is returned in all other cases.
5425 */
5426static inline unsigned int f2fs_usable_zone_blks_in_seg(
5427 struct f2fs_sb_info *sbi, unsigned int segno)
5428{
5429 block_t seg_start, sec_start_blkaddr, sec_cap_blkaddr;
5430 unsigned int secno;
5431
5432 if (!sbi->unusable_blocks_per_sec)
5433 return BLKS_PER_SEG(sbi);
5434
5435 secno = GET_SEC_FROM_SEG(sbi, segno);
5436 seg_start = START_BLOCK(sbi, segno);
5437 sec_start_blkaddr = START_BLOCK(sbi, GET_SEG_FROM_SEC(sbi, secno));
5438 sec_cap_blkaddr = sec_start_blkaddr + CAP_BLKS_PER_SEC(sbi);
5439
5440 /*
5441 * If segment starts before zone capacity and spans beyond
5442 * zone capacity, then usable blocks are from seg start to
5443 * zone capacity. If the segment starts after the zone capacity,
5444 * then there are no usable blocks.
5445 */
5446 if (seg_start >= sec_cap_blkaddr)
5447 return 0;
5448 if (seg_start + BLKS_PER_SEG(sbi) > sec_cap_blkaddr)
5449 return sec_cap_blkaddr - seg_start;
5450
5451 return BLKS_PER_SEG(sbi);
5452}
5453#else
5454int f2fs_check_and_fix_write_pointer(struct f2fs_sb_info *sbi)
5455{
5456 return 0;
5457}
5458
5459static inline unsigned int f2fs_usable_zone_blks_in_seg(struct f2fs_sb_info *sbi,
5460 unsigned int segno)
5461{
5462 return 0;
5463}
5464
5465#endif
5466unsigned int f2fs_usable_blks_in_seg(struct f2fs_sb_info *sbi,
5467 unsigned int segno)
5468{
5469 if (f2fs_sb_has_blkzoned(sbi))
5470 return f2fs_usable_zone_blks_in_seg(sbi, segno);
5471
5472 return BLKS_PER_SEG(sbi);
5473}
5474
5475unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi)
5476{
5477 if (f2fs_sb_has_blkzoned(sbi))
5478 return CAP_SEGS_PER_SEC(sbi);
5479
5480 return SEGS_PER_SEC(sbi);
5481}
5482
5483unsigned long long f2fs_get_section_mtime(struct f2fs_sb_info *sbi,
5484 unsigned int segno)
5485{
5486 unsigned int usable_segs_per_sec = f2fs_usable_segs_in_sec(sbi);
5487 unsigned int secno = 0, start = 0;
5488 unsigned int total_valid_blocks = 0;
5489 unsigned long long mtime = 0;
5490 unsigned int i = 0;
5491
5492 secno = GET_SEC_FROM_SEG(sbi, segno);
5493 start = GET_SEG_FROM_SEC(sbi, secno);
5494
5495 if (!__is_large_section(sbi))
5496 return get_seg_entry(sbi, start + i)->mtime;
5497
5498 for (i = 0; i < usable_segs_per_sec; i++) {
5499 /* for large section, only check the mtime of valid segments */
5500 struct seg_entry *se = get_seg_entry(sbi, start+i);
5501
5502 mtime += se->mtime * se->valid_blocks;
5503 total_valid_blocks += se->valid_blocks;
5504 }
5505
5506 if (total_valid_blocks == 0)
5507 return INVALID_MTIME;
5508
5509 return div_u64(mtime, total_valid_blocks);
5510}
5511
5512/*
5513 * Update min, max modified time for cost-benefit GC algorithm
5514 */
5515static void init_min_max_mtime(struct f2fs_sb_info *sbi)
5516{
5517 struct sit_info *sit_i = SIT_I(sbi);
5518 unsigned int segno;
5519
5520 down_write(&sit_i->sentry_lock);
5521
5522 sit_i->min_mtime = ULLONG_MAX;
5523
5524 for (segno = 0; segno < MAIN_SEGS(sbi); segno += SEGS_PER_SEC(sbi)) {
5525 unsigned long long mtime = 0;
5526
5527 mtime = f2fs_get_section_mtime(sbi, segno);
5528
5529 if (sit_i->min_mtime > mtime)
5530 sit_i->min_mtime = mtime;
5531 }
5532 sit_i->max_mtime = get_mtime(sbi, false);
5533 sit_i->dirty_max_mtime = 0;
5534 up_write(&sit_i->sentry_lock);
5535}
5536
5537int f2fs_build_segment_manager(struct f2fs_sb_info *sbi)
5538{
5539 struct f2fs_super_block *raw_super = F2FS_RAW_SUPER(sbi);
5540 struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
5541 struct f2fs_sm_info *sm_info;
5542 int err;
5543
5544 sm_info = f2fs_kzalloc(sbi, sizeof(struct f2fs_sm_info), GFP_KERNEL);
5545 if (!sm_info)
5546 return -ENOMEM;
5547
5548 /* init sm info */
5549 sbi->sm_info = sm_info;
5550 sm_info->seg0_blkaddr = le32_to_cpu(raw_super->segment0_blkaddr);
5551 sm_info->main_blkaddr = le32_to_cpu(raw_super->main_blkaddr);
5552 sm_info->segment_count = le32_to_cpu(raw_super->segment_count);
5553 sm_info->reserved_segments = le32_to_cpu(ckpt->rsvd_segment_count);
5554 sm_info->ovp_segments = le32_to_cpu(ckpt->overprov_segment_count);
5555 sm_info->main_segments = le32_to_cpu(raw_super->segment_count_main);
5556 sm_info->ssa_blkaddr = le32_to_cpu(raw_super->ssa_blkaddr);
5557 sm_info->rec_prefree_segments = sm_info->main_segments *
5558 DEF_RECLAIM_PREFREE_SEGMENTS / 100;
5559 if (sm_info->rec_prefree_segments > DEF_MAX_RECLAIM_PREFREE_SEGMENTS)
5560 sm_info->rec_prefree_segments = DEF_MAX_RECLAIM_PREFREE_SEGMENTS;
5561
5562 if (!f2fs_lfs_mode(sbi))
5563 sm_info->ipu_policy = BIT(F2FS_IPU_FSYNC);
5564 sm_info->min_ipu_util = DEF_MIN_IPU_UTIL;
5565 sm_info->min_fsync_blocks = DEF_MIN_FSYNC_BLOCKS;
5566 sm_info->min_seq_blocks = BLKS_PER_SEG(sbi);
5567 sm_info->min_hot_blocks = DEF_MIN_HOT_BLOCKS;
5568 sm_info->min_ssr_sections = reserved_sections(sbi);
5569
5570 INIT_LIST_HEAD(&sm_info->sit_entry_set);
5571
5572 init_f2fs_rwsem(&sm_info->curseg_lock);
5573
5574 err = f2fs_create_flush_cmd_control(sbi);
5575 if (err)
5576 return err;
5577
5578 err = create_discard_cmd_control(sbi);
5579 if (err)
5580 return err;
5581
5582 err = build_sit_info(sbi);
5583 if (err)
5584 return err;
5585 err = build_free_segmap(sbi);
5586 if (err)
5587 return err;
5588 err = build_curseg(sbi);
5589 if (err)
5590 return err;
5591
5592 /* reinit free segmap based on SIT */
5593 err = build_sit_entries(sbi);
5594 if (err)
5595 return err;
5596
5597 init_free_segmap(sbi);
5598 err = build_dirty_segmap(sbi);
5599 if (err)
5600 return err;
5601
5602 err = sanity_check_curseg(sbi);
5603 if (err)
5604 return err;
5605
5606 init_min_max_mtime(sbi);
5607 return 0;
5608}
5609
5610static void discard_dirty_segmap(struct f2fs_sb_info *sbi,
5611 enum dirty_type dirty_type)
5612{
5613 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
5614
5615 mutex_lock(&dirty_i->seglist_lock);
5616 kvfree(dirty_i->dirty_segmap[dirty_type]);
5617 dirty_i->nr_dirty[dirty_type] = 0;
5618 mutex_unlock(&dirty_i->seglist_lock);
5619}
5620
5621static void destroy_victim_secmap(struct f2fs_sb_info *sbi)
5622{
5623 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
5624
5625 kvfree(dirty_i->pinned_secmap);
5626 kvfree(dirty_i->victim_secmap);
5627}
5628
5629static void destroy_dirty_segmap(struct f2fs_sb_info *sbi)
5630{
5631 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
5632 int i;
5633
5634 if (!dirty_i)
5635 return;
5636
5637 /* discard pre-free/dirty segments list */
5638 for (i = 0; i < NR_DIRTY_TYPE; i++)
5639 discard_dirty_segmap(sbi, i);
5640
5641 if (__is_large_section(sbi)) {
5642 mutex_lock(&dirty_i->seglist_lock);
5643 kvfree(dirty_i->dirty_secmap);
5644 mutex_unlock(&dirty_i->seglist_lock);
5645 }
5646
5647 destroy_victim_secmap(sbi);
5648 SM_I(sbi)->dirty_info = NULL;
5649 kfree(dirty_i);
5650}
5651
5652static void destroy_curseg(struct f2fs_sb_info *sbi)
5653{
5654 struct curseg_info *array = SM_I(sbi)->curseg_array;
5655 int i;
5656
5657 if (!array)
5658 return;
5659 SM_I(sbi)->curseg_array = NULL;
5660 for (i = 0; i < NR_CURSEG_TYPE; i++) {
5661 kfree(array[i].sum_blk);
5662 kfree(array[i].journal);
5663 }
5664 kfree(array);
5665}
5666
5667static void destroy_free_segmap(struct f2fs_sb_info *sbi)
5668{
5669 struct free_segmap_info *free_i = SM_I(sbi)->free_info;
5670
5671 if (!free_i)
5672 return;
5673 SM_I(sbi)->free_info = NULL;
5674 kvfree(free_i->free_segmap);
5675 kvfree(free_i->free_secmap);
5676 kfree(free_i);
5677}
5678
5679static void destroy_sit_info(struct f2fs_sb_info *sbi)
5680{
5681 struct sit_info *sit_i = SIT_I(sbi);
5682
5683 if (!sit_i)
5684 return;
5685
5686 if (sit_i->sentries)
5687 kvfree(sit_i->bitmap);
5688 kfree(sit_i->tmp_map);
5689
5690 kvfree(sit_i->sentries);
5691 kvfree(sit_i->sec_entries);
5692 kvfree(sit_i->dirty_sentries_bitmap);
5693
5694 SM_I(sbi)->sit_info = NULL;
5695 kvfree(sit_i->sit_bitmap);
5696#ifdef CONFIG_F2FS_CHECK_FS
5697 kvfree(sit_i->sit_bitmap_mir);
5698 kvfree(sit_i->invalid_segmap);
5699#endif
5700 kfree(sit_i);
5701}
5702
5703void f2fs_destroy_segment_manager(struct f2fs_sb_info *sbi)
5704{
5705 struct f2fs_sm_info *sm_info = SM_I(sbi);
5706
5707 if (!sm_info)
5708 return;
5709 f2fs_destroy_flush_cmd_control(sbi, true);
5710 destroy_discard_cmd_control(sbi);
5711 destroy_dirty_segmap(sbi);
5712 destroy_curseg(sbi);
5713 destroy_free_segmap(sbi);
5714 destroy_sit_info(sbi);
5715 sbi->sm_info = NULL;
5716 kfree(sm_info);
5717}
5718
5719int __init f2fs_create_segment_manager_caches(void)
5720{
5721 discard_entry_slab = f2fs_kmem_cache_create("f2fs_discard_entry",
5722 sizeof(struct discard_entry));
5723 if (!discard_entry_slab)
5724 goto fail;
5725
5726 discard_cmd_slab = f2fs_kmem_cache_create("f2fs_discard_cmd",
5727 sizeof(struct discard_cmd));
5728 if (!discard_cmd_slab)
5729 goto destroy_discard_entry;
5730
5731 sit_entry_set_slab = f2fs_kmem_cache_create("f2fs_sit_entry_set",
5732 sizeof(struct sit_entry_set));
5733 if (!sit_entry_set_slab)
5734 goto destroy_discard_cmd;
5735
5736 revoke_entry_slab = f2fs_kmem_cache_create("f2fs_revoke_entry",
5737 sizeof(struct revoke_entry));
5738 if (!revoke_entry_slab)
5739 goto destroy_sit_entry_set;
5740 return 0;
5741
5742destroy_sit_entry_set:
5743 kmem_cache_destroy(sit_entry_set_slab);
5744destroy_discard_cmd:
5745 kmem_cache_destroy(discard_cmd_slab);
5746destroy_discard_entry:
5747 kmem_cache_destroy(discard_entry_slab);
5748fail:
5749 return -ENOMEM;
5750}
5751
5752void f2fs_destroy_segment_manager_caches(void)
5753{
5754 kmem_cache_destroy(sit_entry_set_slab);
5755 kmem_cache_destroy(discard_cmd_slab);
5756 kmem_cache_destroy(discard_entry_slab);
5757 kmem_cache_destroy(revoke_entry_slab);
5758}
1// SPDX-License-Identifier: GPL-2.0
2/*
3 * fs/f2fs/segment.c
4 *
5 * Copyright (c) 2012 Samsung Electronics Co., Ltd.
6 * http://www.samsung.com/
7 */
8#include <linux/fs.h>
9#include <linux/f2fs_fs.h>
10#include <linux/bio.h>
11#include <linux/blkdev.h>
12#include <linux/sched/mm.h>
13#include <linux/prefetch.h>
14#include <linux/kthread.h>
15#include <linux/swap.h>
16#include <linux/timer.h>
17#include <linux/freezer.h>
18#include <linux/sched/signal.h>
19#include <linux/random.h>
20
21#include "f2fs.h"
22#include "segment.h"
23#include "node.h"
24#include "gc.h"
25#include "iostat.h"
26#include <trace/events/f2fs.h>
27
28#define __reverse_ffz(x) __reverse_ffs(~(x))
29
30static struct kmem_cache *discard_entry_slab;
31static struct kmem_cache *discard_cmd_slab;
32static struct kmem_cache *sit_entry_set_slab;
33static struct kmem_cache *revoke_entry_slab;
34
35static unsigned long __reverse_ulong(unsigned char *str)
36{
37 unsigned long tmp = 0;
38 int shift = 24, idx = 0;
39
40#if BITS_PER_LONG == 64
41 shift = 56;
42#endif
43 while (shift >= 0) {
44 tmp |= (unsigned long)str[idx++] << shift;
45 shift -= BITS_PER_BYTE;
46 }
47 return tmp;
48}
49
50/*
51 * __reverse_ffs is copied from include/asm-generic/bitops/__ffs.h since
52 * MSB and LSB are reversed in a byte by f2fs_set_bit.
53 */
54static inline unsigned long __reverse_ffs(unsigned long word)
55{
56 int num = 0;
57
58#if BITS_PER_LONG == 64
59 if ((word & 0xffffffff00000000UL) == 0)
60 num += 32;
61 else
62 word >>= 32;
63#endif
64 if ((word & 0xffff0000) == 0)
65 num += 16;
66 else
67 word >>= 16;
68
69 if ((word & 0xff00) == 0)
70 num += 8;
71 else
72 word >>= 8;
73
74 if ((word & 0xf0) == 0)
75 num += 4;
76 else
77 word >>= 4;
78
79 if ((word & 0xc) == 0)
80 num += 2;
81 else
82 word >>= 2;
83
84 if ((word & 0x2) == 0)
85 num += 1;
86 return num;
87}
88
89/*
90 * __find_rev_next(_zero)_bit is copied from lib/find_next_bit.c because
91 * f2fs_set_bit makes MSB and LSB reversed in a byte.
92 * @size must be integral times of unsigned long.
93 * Example:
94 * MSB <--> LSB
95 * f2fs_set_bit(0, bitmap) => 1000 0000
96 * f2fs_set_bit(7, bitmap) => 0000 0001
97 */
98static unsigned long __find_rev_next_bit(const unsigned long *addr,
99 unsigned long size, unsigned long offset)
100{
101 const unsigned long *p = addr + BIT_WORD(offset);
102 unsigned long result = size;
103 unsigned long tmp;
104
105 if (offset >= size)
106 return size;
107
108 size -= (offset & ~(BITS_PER_LONG - 1));
109 offset %= BITS_PER_LONG;
110
111 while (1) {
112 if (*p == 0)
113 goto pass;
114
115 tmp = __reverse_ulong((unsigned char *)p);
116
117 tmp &= ~0UL >> offset;
118 if (size < BITS_PER_LONG)
119 tmp &= (~0UL << (BITS_PER_LONG - size));
120 if (tmp)
121 goto found;
122pass:
123 if (size <= BITS_PER_LONG)
124 break;
125 size -= BITS_PER_LONG;
126 offset = 0;
127 p++;
128 }
129 return result;
130found:
131 return result - size + __reverse_ffs(tmp);
132}
133
134static unsigned long __find_rev_next_zero_bit(const unsigned long *addr,
135 unsigned long size, unsigned long offset)
136{
137 const unsigned long *p = addr + BIT_WORD(offset);
138 unsigned long result = size;
139 unsigned long tmp;
140
141 if (offset >= size)
142 return size;
143
144 size -= (offset & ~(BITS_PER_LONG - 1));
145 offset %= BITS_PER_LONG;
146
147 while (1) {
148 if (*p == ~0UL)
149 goto pass;
150
151 tmp = __reverse_ulong((unsigned char *)p);
152
153 if (offset)
154 tmp |= ~0UL << (BITS_PER_LONG - offset);
155 if (size < BITS_PER_LONG)
156 tmp |= ~0UL >> size;
157 if (tmp != ~0UL)
158 goto found;
159pass:
160 if (size <= BITS_PER_LONG)
161 break;
162 size -= BITS_PER_LONG;
163 offset = 0;
164 p++;
165 }
166 return result;
167found:
168 return result - size + __reverse_ffz(tmp);
169}
170
171bool f2fs_need_SSR(struct f2fs_sb_info *sbi)
172{
173 int node_secs = get_blocktype_secs(sbi, F2FS_DIRTY_NODES);
174 int dent_secs = get_blocktype_secs(sbi, F2FS_DIRTY_DENTS);
175 int imeta_secs = get_blocktype_secs(sbi, F2FS_DIRTY_IMETA);
176
177 if (f2fs_lfs_mode(sbi))
178 return false;
179 if (sbi->gc_mode == GC_URGENT_HIGH)
180 return true;
181 if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED)))
182 return true;
183
184 return free_sections(sbi) <= (node_secs + 2 * dent_secs + imeta_secs +
185 SM_I(sbi)->min_ssr_sections + reserved_sections(sbi));
186}
187
188void f2fs_abort_atomic_write(struct inode *inode, bool clean)
189{
190 struct f2fs_inode_info *fi = F2FS_I(inode);
191
192 if (!f2fs_is_atomic_file(inode))
193 return;
194
195 release_atomic_write_cnt(inode);
196 clear_inode_flag(inode, FI_ATOMIC_COMMITTED);
197 clear_inode_flag(inode, FI_ATOMIC_REPLACE);
198 clear_inode_flag(inode, FI_ATOMIC_FILE);
199 stat_dec_atomic_inode(inode);
200
201 F2FS_I(inode)->atomic_write_task = NULL;
202
203 if (clean) {
204 truncate_inode_pages_final(inode->i_mapping);
205 f2fs_i_size_write(inode, fi->original_i_size);
206 fi->original_i_size = 0;
207 }
208 /* avoid stale dirty inode during eviction */
209 sync_inode_metadata(inode, 0);
210}
211
212static int __replace_atomic_write_block(struct inode *inode, pgoff_t index,
213 block_t new_addr, block_t *old_addr, bool recover)
214{
215 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
216 struct dnode_of_data dn;
217 struct node_info ni;
218 int err;
219
220retry:
221 set_new_dnode(&dn, inode, NULL, NULL, 0);
222 err = f2fs_get_dnode_of_data(&dn, index, ALLOC_NODE);
223 if (err) {
224 if (err == -ENOMEM) {
225 f2fs_io_schedule_timeout(DEFAULT_IO_TIMEOUT);
226 goto retry;
227 }
228 return err;
229 }
230
231 err = f2fs_get_node_info(sbi, dn.nid, &ni, false);
232 if (err) {
233 f2fs_put_dnode(&dn);
234 return err;
235 }
236
237 if (recover) {
238 /* dn.data_blkaddr is always valid */
239 if (!__is_valid_data_blkaddr(new_addr)) {
240 if (new_addr == NULL_ADDR)
241 dec_valid_block_count(sbi, inode, 1);
242 f2fs_invalidate_blocks(sbi, dn.data_blkaddr);
243 f2fs_update_data_blkaddr(&dn, new_addr);
244 } else {
245 f2fs_replace_block(sbi, &dn, dn.data_blkaddr,
246 new_addr, ni.version, true, true);
247 }
248 } else {
249 blkcnt_t count = 1;
250
251 err = inc_valid_block_count(sbi, inode, &count);
252 if (err) {
253 f2fs_put_dnode(&dn);
254 return err;
255 }
256
257 *old_addr = dn.data_blkaddr;
258 f2fs_truncate_data_blocks_range(&dn, 1);
259 dec_valid_block_count(sbi, F2FS_I(inode)->cow_inode, count);
260
261 f2fs_replace_block(sbi, &dn, dn.data_blkaddr, new_addr,
262 ni.version, true, false);
263 }
264
265 f2fs_put_dnode(&dn);
266
267 trace_f2fs_replace_atomic_write_block(inode, F2FS_I(inode)->cow_inode,
268 index, old_addr ? *old_addr : 0, new_addr, recover);
269 return 0;
270}
271
272static void __complete_revoke_list(struct inode *inode, struct list_head *head,
273 bool revoke)
274{
275 struct revoke_entry *cur, *tmp;
276 pgoff_t start_index = 0;
277 bool truncate = is_inode_flag_set(inode, FI_ATOMIC_REPLACE);
278
279 list_for_each_entry_safe(cur, tmp, head, list) {
280 if (revoke) {
281 __replace_atomic_write_block(inode, cur->index,
282 cur->old_addr, NULL, true);
283 } else if (truncate) {
284 f2fs_truncate_hole(inode, start_index, cur->index);
285 start_index = cur->index + 1;
286 }
287
288 list_del(&cur->list);
289 kmem_cache_free(revoke_entry_slab, cur);
290 }
291
292 if (!revoke && truncate)
293 f2fs_do_truncate_blocks(inode, start_index * PAGE_SIZE, false);
294}
295
296static int __f2fs_commit_atomic_write(struct inode *inode)
297{
298 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
299 struct f2fs_inode_info *fi = F2FS_I(inode);
300 struct inode *cow_inode = fi->cow_inode;
301 struct revoke_entry *new;
302 struct list_head revoke_list;
303 block_t blkaddr;
304 struct dnode_of_data dn;
305 pgoff_t len = DIV_ROUND_UP(i_size_read(inode), PAGE_SIZE);
306 pgoff_t off = 0, blen, index;
307 int ret = 0, i;
308
309 INIT_LIST_HEAD(&revoke_list);
310
311 while (len) {
312 blen = min_t(pgoff_t, ADDRS_PER_BLOCK(cow_inode), len);
313
314 set_new_dnode(&dn, cow_inode, NULL, NULL, 0);
315 ret = f2fs_get_dnode_of_data(&dn, off, LOOKUP_NODE_RA);
316 if (ret && ret != -ENOENT) {
317 goto out;
318 } else if (ret == -ENOENT) {
319 ret = 0;
320 if (dn.max_level == 0)
321 goto out;
322 goto next;
323 }
324
325 blen = min((pgoff_t)ADDRS_PER_PAGE(dn.node_page, cow_inode),
326 len);
327 index = off;
328 for (i = 0; i < blen; i++, dn.ofs_in_node++, index++) {
329 blkaddr = f2fs_data_blkaddr(&dn);
330
331 if (!__is_valid_data_blkaddr(blkaddr)) {
332 continue;
333 } else if (!f2fs_is_valid_blkaddr(sbi, blkaddr,
334 DATA_GENERIC_ENHANCE)) {
335 f2fs_put_dnode(&dn);
336 ret = -EFSCORRUPTED;
337 f2fs_handle_error(sbi,
338 ERROR_INVALID_BLKADDR);
339 goto out;
340 }
341
342 new = f2fs_kmem_cache_alloc(revoke_entry_slab, GFP_NOFS,
343 true, NULL);
344
345 ret = __replace_atomic_write_block(inode, index, blkaddr,
346 &new->old_addr, false);
347 if (ret) {
348 f2fs_put_dnode(&dn);
349 kmem_cache_free(revoke_entry_slab, new);
350 goto out;
351 }
352
353 f2fs_update_data_blkaddr(&dn, NULL_ADDR);
354 new->index = index;
355 list_add_tail(&new->list, &revoke_list);
356 }
357 f2fs_put_dnode(&dn);
358next:
359 off += blen;
360 len -= blen;
361 }
362
363out:
364 if (ret) {
365 sbi->revoked_atomic_block += fi->atomic_write_cnt;
366 } else {
367 sbi->committed_atomic_block += fi->atomic_write_cnt;
368 set_inode_flag(inode, FI_ATOMIC_COMMITTED);
369 }
370
371 __complete_revoke_list(inode, &revoke_list, ret ? true : false);
372
373 return ret;
374}
375
376int f2fs_commit_atomic_write(struct inode *inode)
377{
378 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
379 struct f2fs_inode_info *fi = F2FS_I(inode);
380 int err;
381
382 err = filemap_write_and_wait_range(inode->i_mapping, 0, LLONG_MAX);
383 if (err)
384 return err;
385
386 f2fs_down_write(&fi->i_gc_rwsem[WRITE]);
387 f2fs_lock_op(sbi);
388
389 err = __f2fs_commit_atomic_write(inode);
390
391 f2fs_unlock_op(sbi);
392 f2fs_up_write(&fi->i_gc_rwsem[WRITE]);
393
394 return err;
395}
396
397/*
398 * This function balances dirty node and dentry pages.
399 * In addition, it controls garbage collection.
400 */
401void f2fs_balance_fs(struct f2fs_sb_info *sbi, bool need)
402{
403 if (time_to_inject(sbi, FAULT_CHECKPOINT))
404 f2fs_stop_checkpoint(sbi, false, STOP_CP_REASON_FAULT_INJECT);
405
406 /* balance_fs_bg is able to be pending */
407 if (need && excess_cached_nats(sbi))
408 f2fs_balance_fs_bg(sbi, false);
409
410 if (!f2fs_is_checkpoint_ready(sbi))
411 return;
412
413 /*
414 * We should do GC or end up with checkpoint, if there are so many dirty
415 * dir/node pages without enough free segments.
416 */
417 if (has_enough_free_secs(sbi, 0, 0))
418 return;
419
420 if (test_opt(sbi, GC_MERGE) && sbi->gc_thread &&
421 sbi->gc_thread->f2fs_gc_task) {
422 DEFINE_WAIT(wait);
423
424 prepare_to_wait(&sbi->gc_thread->fggc_wq, &wait,
425 TASK_UNINTERRUPTIBLE);
426 wake_up(&sbi->gc_thread->gc_wait_queue_head);
427 io_schedule();
428 finish_wait(&sbi->gc_thread->fggc_wq, &wait);
429 } else {
430 struct f2fs_gc_control gc_control = {
431 .victim_segno = NULL_SEGNO,
432 .init_gc_type = BG_GC,
433 .no_bg_gc = true,
434 .should_migrate_blocks = false,
435 .err_gc_skipped = false,
436 .nr_free_secs = 1 };
437 f2fs_down_write(&sbi->gc_lock);
438 stat_inc_gc_call_count(sbi, FOREGROUND);
439 f2fs_gc(sbi, &gc_control);
440 }
441}
442
443static inline bool excess_dirty_threshold(struct f2fs_sb_info *sbi)
444{
445 int factor = f2fs_rwsem_is_locked(&sbi->cp_rwsem) ? 3 : 2;
446 unsigned int dents = get_pages(sbi, F2FS_DIRTY_DENTS);
447 unsigned int qdata = get_pages(sbi, F2FS_DIRTY_QDATA);
448 unsigned int nodes = get_pages(sbi, F2FS_DIRTY_NODES);
449 unsigned int meta = get_pages(sbi, F2FS_DIRTY_META);
450 unsigned int imeta = get_pages(sbi, F2FS_DIRTY_IMETA);
451 unsigned int threshold = sbi->blocks_per_seg * factor *
452 DEFAULT_DIRTY_THRESHOLD;
453 unsigned int global_threshold = threshold * 3 / 2;
454
455 if (dents >= threshold || qdata >= threshold ||
456 nodes >= threshold || meta >= threshold ||
457 imeta >= threshold)
458 return true;
459 return dents + qdata + nodes + meta + imeta > global_threshold;
460}
461
462void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
463{
464 if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
465 return;
466
467 /* try to shrink extent cache when there is no enough memory */
468 if (!f2fs_available_free_memory(sbi, READ_EXTENT_CACHE))
469 f2fs_shrink_read_extent_tree(sbi,
470 READ_EXTENT_CACHE_SHRINK_NUMBER);
471
472 /* try to shrink age extent cache when there is no enough memory */
473 if (!f2fs_available_free_memory(sbi, AGE_EXTENT_CACHE))
474 f2fs_shrink_age_extent_tree(sbi,
475 AGE_EXTENT_CACHE_SHRINK_NUMBER);
476
477 /* check the # of cached NAT entries */
478 if (!f2fs_available_free_memory(sbi, NAT_ENTRIES))
479 f2fs_try_to_free_nats(sbi, NAT_ENTRY_PER_BLOCK);
480
481 if (!f2fs_available_free_memory(sbi, FREE_NIDS))
482 f2fs_try_to_free_nids(sbi, MAX_FREE_NIDS);
483 else
484 f2fs_build_free_nids(sbi, false, false);
485
486 if (excess_dirty_nats(sbi) || excess_dirty_threshold(sbi) ||
487 excess_prefree_segs(sbi) || !f2fs_space_for_roll_forward(sbi))
488 goto do_sync;
489
490 /* there is background inflight IO or foreground operation recently */
491 if (is_inflight_io(sbi, REQ_TIME) ||
492 (!f2fs_time_over(sbi, REQ_TIME) && f2fs_rwsem_is_locked(&sbi->cp_rwsem)))
493 return;
494
495 /* exceed periodical checkpoint timeout threshold */
496 if (f2fs_time_over(sbi, CP_TIME))
497 goto do_sync;
498
499 /* checkpoint is the only way to shrink partial cached entries */
500 if (f2fs_available_free_memory(sbi, NAT_ENTRIES) &&
501 f2fs_available_free_memory(sbi, INO_ENTRIES))
502 return;
503
504do_sync:
505 if (test_opt(sbi, DATA_FLUSH) && from_bg) {
506 struct blk_plug plug;
507
508 mutex_lock(&sbi->flush_lock);
509
510 blk_start_plug(&plug);
511 f2fs_sync_dirty_inodes(sbi, FILE_INODE, false);
512 blk_finish_plug(&plug);
513
514 mutex_unlock(&sbi->flush_lock);
515 }
516 stat_inc_cp_call_count(sbi, BACKGROUND);
517 f2fs_sync_fs(sbi->sb, 1);
518}
519
520static int __submit_flush_wait(struct f2fs_sb_info *sbi,
521 struct block_device *bdev)
522{
523 int ret = blkdev_issue_flush(bdev);
524
525 trace_f2fs_issue_flush(bdev, test_opt(sbi, NOBARRIER),
526 test_opt(sbi, FLUSH_MERGE), ret);
527 if (!ret)
528 f2fs_update_iostat(sbi, NULL, FS_FLUSH_IO, 0);
529 return ret;
530}
531
532static int submit_flush_wait(struct f2fs_sb_info *sbi, nid_t ino)
533{
534 int ret = 0;
535 int i;
536
537 if (!f2fs_is_multi_device(sbi))
538 return __submit_flush_wait(sbi, sbi->sb->s_bdev);
539
540 for (i = 0; i < sbi->s_ndevs; i++) {
541 if (!f2fs_is_dirty_device(sbi, ino, i, FLUSH_INO))
542 continue;
543 ret = __submit_flush_wait(sbi, FDEV(i).bdev);
544 if (ret)
545 break;
546 }
547 return ret;
548}
549
550static int issue_flush_thread(void *data)
551{
552 struct f2fs_sb_info *sbi = data;
553 struct flush_cmd_control *fcc = SM_I(sbi)->fcc_info;
554 wait_queue_head_t *q = &fcc->flush_wait_queue;
555repeat:
556 if (kthread_should_stop())
557 return 0;
558
559 if (!llist_empty(&fcc->issue_list)) {
560 struct flush_cmd *cmd, *next;
561 int ret;
562
563 fcc->dispatch_list = llist_del_all(&fcc->issue_list);
564 fcc->dispatch_list = llist_reverse_order(fcc->dispatch_list);
565
566 cmd = llist_entry(fcc->dispatch_list, struct flush_cmd, llnode);
567
568 ret = submit_flush_wait(sbi, cmd->ino);
569 atomic_inc(&fcc->issued_flush);
570
571 llist_for_each_entry_safe(cmd, next,
572 fcc->dispatch_list, llnode) {
573 cmd->ret = ret;
574 complete(&cmd->wait);
575 }
576 fcc->dispatch_list = NULL;
577 }
578
579 wait_event_interruptible(*q,
580 kthread_should_stop() || !llist_empty(&fcc->issue_list));
581 goto repeat;
582}
583
584int f2fs_issue_flush(struct f2fs_sb_info *sbi, nid_t ino)
585{
586 struct flush_cmd_control *fcc = SM_I(sbi)->fcc_info;
587 struct flush_cmd cmd;
588 int ret;
589
590 if (test_opt(sbi, NOBARRIER))
591 return 0;
592
593 if (!test_opt(sbi, FLUSH_MERGE)) {
594 atomic_inc(&fcc->queued_flush);
595 ret = submit_flush_wait(sbi, ino);
596 atomic_dec(&fcc->queued_flush);
597 atomic_inc(&fcc->issued_flush);
598 return ret;
599 }
600
601 if (atomic_inc_return(&fcc->queued_flush) == 1 ||
602 f2fs_is_multi_device(sbi)) {
603 ret = submit_flush_wait(sbi, ino);
604 atomic_dec(&fcc->queued_flush);
605
606 atomic_inc(&fcc->issued_flush);
607 return ret;
608 }
609
610 cmd.ino = ino;
611 init_completion(&cmd.wait);
612
613 llist_add(&cmd.llnode, &fcc->issue_list);
614
615 /*
616 * update issue_list before we wake up issue_flush thread, this
617 * smp_mb() pairs with another barrier in ___wait_event(), see
618 * more details in comments of waitqueue_active().
619 */
620 smp_mb();
621
622 if (waitqueue_active(&fcc->flush_wait_queue))
623 wake_up(&fcc->flush_wait_queue);
624
625 if (fcc->f2fs_issue_flush) {
626 wait_for_completion(&cmd.wait);
627 atomic_dec(&fcc->queued_flush);
628 } else {
629 struct llist_node *list;
630
631 list = llist_del_all(&fcc->issue_list);
632 if (!list) {
633 wait_for_completion(&cmd.wait);
634 atomic_dec(&fcc->queued_flush);
635 } else {
636 struct flush_cmd *tmp, *next;
637
638 ret = submit_flush_wait(sbi, ino);
639
640 llist_for_each_entry_safe(tmp, next, list, llnode) {
641 if (tmp == &cmd) {
642 cmd.ret = ret;
643 atomic_dec(&fcc->queued_flush);
644 continue;
645 }
646 tmp->ret = ret;
647 complete(&tmp->wait);
648 }
649 }
650 }
651
652 return cmd.ret;
653}
654
655int f2fs_create_flush_cmd_control(struct f2fs_sb_info *sbi)
656{
657 dev_t dev = sbi->sb->s_bdev->bd_dev;
658 struct flush_cmd_control *fcc;
659
660 if (SM_I(sbi)->fcc_info) {
661 fcc = SM_I(sbi)->fcc_info;
662 if (fcc->f2fs_issue_flush)
663 return 0;
664 goto init_thread;
665 }
666
667 fcc = f2fs_kzalloc(sbi, sizeof(struct flush_cmd_control), GFP_KERNEL);
668 if (!fcc)
669 return -ENOMEM;
670 atomic_set(&fcc->issued_flush, 0);
671 atomic_set(&fcc->queued_flush, 0);
672 init_waitqueue_head(&fcc->flush_wait_queue);
673 init_llist_head(&fcc->issue_list);
674 SM_I(sbi)->fcc_info = fcc;
675 if (!test_opt(sbi, FLUSH_MERGE))
676 return 0;
677
678init_thread:
679 fcc->f2fs_issue_flush = kthread_run(issue_flush_thread, sbi,
680 "f2fs_flush-%u:%u", MAJOR(dev), MINOR(dev));
681 if (IS_ERR(fcc->f2fs_issue_flush)) {
682 int err = PTR_ERR(fcc->f2fs_issue_flush);
683
684 fcc->f2fs_issue_flush = NULL;
685 return err;
686 }
687
688 return 0;
689}
690
691void f2fs_destroy_flush_cmd_control(struct f2fs_sb_info *sbi, bool free)
692{
693 struct flush_cmd_control *fcc = SM_I(sbi)->fcc_info;
694
695 if (fcc && fcc->f2fs_issue_flush) {
696 struct task_struct *flush_thread = fcc->f2fs_issue_flush;
697
698 fcc->f2fs_issue_flush = NULL;
699 kthread_stop(flush_thread);
700 }
701 if (free) {
702 kfree(fcc);
703 SM_I(sbi)->fcc_info = NULL;
704 }
705}
706
707int f2fs_flush_device_cache(struct f2fs_sb_info *sbi)
708{
709 int ret = 0, i;
710
711 if (!f2fs_is_multi_device(sbi))
712 return 0;
713
714 if (test_opt(sbi, NOBARRIER))
715 return 0;
716
717 for (i = 1; i < sbi->s_ndevs; i++) {
718 int count = DEFAULT_RETRY_IO_COUNT;
719
720 if (!f2fs_test_bit(i, (char *)&sbi->dirty_device))
721 continue;
722
723 do {
724 ret = __submit_flush_wait(sbi, FDEV(i).bdev);
725 if (ret)
726 f2fs_io_schedule_timeout(DEFAULT_IO_TIMEOUT);
727 } while (ret && --count);
728
729 if (ret) {
730 f2fs_stop_checkpoint(sbi, false,
731 STOP_CP_REASON_FLUSH_FAIL);
732 break;
733 }
734
735 spin_lock(&sbi->dev_lock);
736 f2fs_clear_bit(i, (char *)&sbi->dirty_device);
737 spin_unlock(&sbi->dev_lock);
738 }
739
740 return ret;
741}
742
743static void __locate_dirty_segment(struct f2fs_sb_info *sbi, unsigned int segno,
744 enum dirty_type dirty_type)
745{
746 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
747
748 /* need not be added */
749 if (IS_CURSEG(sbi, segno))
750 return;
751
752 if (!test_and_set_bit(segno, dirty_i->dirty_segmap[dirty_type]))
753 dirty_i->nr_dirty[dirty_type]++;
754
755 if (dirty_type == DIRTY) {
756 struct seg_entry *sentry = get_seg_entry(sbi, segno);
757 enum dirty_type t = sentry->type;
758
759 if (unlikely(t >= DIRTY)) {
760 f2fs_bug_on(sbi, 1);
761 return;
762 }
763 if (!test_and_set_bit(segno, dirty_i->dirty_segmap[t]))
764 dirty_i->nr_dirty[t]++;
765
766 if (__is_large_section(sbi)) {
767 unsigned int secno = GET_SEC_FROM_SEG(sbi, segno);
768 block_t valid_blocks =
769 get_valid_blocks(sbi, segno, true);
770
771 f2fs_bug_on(sbi, unlikely(!valid_blocks ||
772 valid_blocks == CAP_BLKS_PER_SEC(sbi)));
773
774 if (!IS_CURSEC(sbi, secno))
775 set_bit(secno, dirty_i->dirty_secmap);
776 }
777 }
778}
779
780static void __remove_dirty_segment(struct f2fs_sb_info *sbi, unsigned int segno,
781 enum dirty_type dirty_type)
782{
783 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
784 block_t valid_blocks;
785
786 if (test_and_clear_bit(segno, dirty_i->dirty_segmap[dirty_type]))
787 dirty_i->nr_dirty[dirty_type]--;
788
789 if (dirty_type == DIRTY) {
790 struct seg_entry *sentry = get_seg_entry(sbi, segno);
791 enum dirty_type t = sentry->type;
792
793 if (test_and_clear_bit(segno, dirty_i->dirty_segmap[t]))
794 dirty_i->nr_dirty[t]--;
795
796 valid_blocks = get_valid_blocks(sbi, segno, true);
797 if (valid_blocks == 0) {
798 clear_bit(GET_SEC_FROM_SEG(sbi, segno),
799 dirty_i->victim_secmap);
800#ifdef CONFIG_F2FS_CHECK_FS
801 clear_bit(segno, SIT_I(sbi)->invalid_segmap);
802#endif
803 }
804 if (__is_large_section(sbi)) {
805 unsigned int secno = GET_SEC_FROM_SEG(sbi, segno);
806
807 if (!valid_blocks ||
808 valid_blocks == CAP_BLKS_PER_SEC(sbi)) {
809 clear_bit(secno, dirty_i->dirty_secmap);
810 return;
811 }
812
813 if (!IS_CURSEC(sbi, secno))
814 set_bit(secno, dirty_i->dirty_secmap);
815 }
816 }
817}
818
819/*
820 * Should not occur error such as -ENOMEM.
821 * Adding dirty entry into seglist is not critical operation.
822 * If a given segment is one of current working segments, it won't be added.
823 */
824static void locate_dirty_segment(struct f2fs_sb_info *sbi, unsigned int segno)
825{
826 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
827 unsigned short valid_blocks, ckpt_valid_blocks;
828 unsigned int usable_blocks;
829
830 if (segno == NULL_SEGNO || IS_CURSEG(sbi, segno))
831 return;
832
833 usable_blocks = f2fs_usable_blks_in_seg(sbi, segno);
834 mutex_lock(&dirty_i->seglist_lock);
835
836 valid_blocks = get_valid_blocks(sbi, segno, false);
837 ckpt_valid_blocks = get_ckpt_valid_blocks(sbi, segno, false);
838
839 if (valid_blocks == 0 && (!is_sbi_flag_set(sbi, SBI_CP_DISABLED) ||
840 ckpt_valid_blocks == usable_blocks)) {
841 __locate_dirty_segment(sbi, segno, PRE);
842 __remove_dirty_segment(sbi, segno, DIRTY);
843 } else if (valid_blocks < usable_blocks) {
844 __locate_dirty_segment(sbi, segno, DIRTY);
845 } else {
846 /* Recovery routine with SSR needs this */
847 __remove_dirty_segment(sbi, segno, DIRTY);
848 }
849
850 mutex_unlock(&dirty_i->seglist_lock);
851}
852
853/* This moves currently empty dirty blocks to prefree. Must hold seglist_lock */
854void f2fs_dirty_to_prefree(struct f2fs_sb_info *sbi)
855{
856 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
857 unsigned int segno;
858
859 mutex_lock(&dirty_i->seglist_lock);
860 for_each_set_bit(segno, dirty_i->dirty_segmap[DIRTY], MAIN_SEGS(sbi)) {
861 if (get_valid_blocks(sbi, segno, false))
862 continue;
863 if (IS_CURSEG(sbi, segno))
864 continue;
865 __locate_dirty_segment(sbi, segno, PRE);
866 __remove_dirty_segment(sbi, segno, DIRTY);
867 }
868 mutex_unlock(&dirty_i->seglist_lock);
869}
870
871block_t f2fs_get_unusable_blocks(struct f2fs_sb_info *sbi)
872{
873 int ovp_hole_segs =
874 (overprovision_segments(sbi) - reserved_segments(sbi));
875 block_t ovp_holes = ovp_hole_segs << sbi->log_blocks_per_seg;
876 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
877 block_t holes[2] = {0, 0}; /* DATA and NODE */
878 block_t unusable;
879 struct seg_entry *se;
880 unsigned int segno;
881
882 mutex_lock(&dirty_i->seglist_lock);
883 for_each_set_bit(segno, dirty_i->dirty_segmap[DIRTY], MAIN_SEGS(sbi)) {
884 se = get_seg_entry(sbi, segno);
885 if (IS_NODESEG(se->type))
886 holes[NODE] += f2fs_usable_blks_in_seg(sbi, segno) -
887 se->valid_blocks;
888 else
889 holes[DATA] += f2fs_usable_blks_in_seg(sbi, segno) -
890 se->valid_blocks;
891 }
892 mutex_unlock(&dirty_i->seglist_lock);
893
894 unusable = max(holes[DATA], holes[NODE]);
895 if (unusable > ovp_holes)
896 return unusable - ovp_holes;
897 return 0;
898}
899
900int f2fs_disable_cp_again(struct f2fs_sb_info *sbi, block_t unusable)
901{
902 int ovp_hole_segs =
903 (overprovision_segments(sbi) - reserved_segments(sbi));
904 if (unusable > F2FS_OPTION(sbi).unusable_cap)
905 return -EAGAIN;
906 if (is_sbi_flag_set(sbi, SBI_CP_DISABLED_QUICK) &&
907 dirty_segments(sbi) > ovp_hole_segs)
908 return -EAGAIN;
909 return 0;
910}
911
912/* This is only used by SBI_CP_DISABLED */
913static unsigned int get_free_segment(struct f2fs_sb_info *sbi)
914{
915 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
916 unsigned int segno = 0;
917
918 mutex_lock(&dirty_i->seglist_lock);
919 for_each_set_bit(segno, dirty_i->dirty_segmap[DIRTY], MAIN_SEGS(sbi)) {
920 if (get_valid_blocks(sbi, segno, false))
921 continue;
922 if (get_ckpt_valid_blocks(sbi, segno, false))
923 continue;
924 mutex_unlock(&dirty_i->seglist_lock);
925 return segno;
926 }
927 mutex_unlock(&dirty_i->seglist_lock);
928 return NULL_SEGNO;
929}
930
931static struct discard_cmd *__create_discard_cmd(struct f2fs_sb_info *sbi,
932 struct block_device *bdev, block_t lstart,
933 block_t start, block_t len)
934{
935 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
936 struct list_head *pend_list;
937 struct discard_cmd *dc;
938
939 f2fs_bug_on(sbi, !len);
940
941 pend_list = &dcc->pend_list[plist_idx(len)];
942
943 dc = f2fs_kmem_cache_alloc(discard_cmd_slab, GFP_NOFS, true, NULL);
944 INIT_LIST_HEAD(&dc->list);
945 dc->bdev = bdev;
946 dc->di.lstart = lstart;
947 dc->di.start = start;
948 dc->di.len = len;
949 dc->ref = 0;
950 dc->state = D_PREP;
951 dc->queued = 0;
952 dc->error = 0;
953 init_completion(&dc->wait);
954 list_add_tail(&dc->list, pend_list);
955 spin_lock_init(&dc->lock);
956 dc->bio_ref = 0;
957 atomic_inc(&dcc->discard_cmd_cnt);
958 dcc->undiscard_blks += len;
959
960 return dc;
961}
962
963static bool f2fs_check_discard_tree(struct f2fs_sb_info *sbi)
964{
965#ifdef CONFIG_F2FS_CHECK_FS
966 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
967 struct rb_node *cur = rb_first_cached(&dcc->root), *next;
968 struct discard_cmd *cur_dc, *next_dc;
969
970 while (cur) {
971 next = rb_next(cur);
972 if (!next)
973 return true;
974
975 cur_dc = rb_entry(cur, struct discard_cmd, rb_node);
976 next_dc = rb_entry(next, struct discard_cmd, rb_node);
977
978 if (cur_dc->di.lstart + cur_dc->di.len > next_dc->di.lstart) {
979 f2fs_info(sbi, "broken discard_rbtree, "
980 "cur(%u, %u) next(%u, %u)",
981 cur_dc->di.lstart, cur_dc->di.len,
982 next_dc->di.lstart, next_dc->di.len);
983 return false;
984 }
985 cur = next;
986 }
987#endif
988 return true;
989}
990
991static struct discard_cmd *__lookup_discard_cmd(struct f2fs_sb_info *sbi,
992 block_t blkaddr)
993{
994 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
995 struct rb_node *node = dcc->root.rb_root.rb_node;
996 struct discard_cmd *dc;
997
998 while (node) {
999 dc = rb_entry(node, struct discard_cmd, rb_node);
1000
1001 if (blkaddr < dc->di.lstart)
1002 node = node->rb_left;
1003 else if (blkaddr >= dc->di.lstart + dc->di.len)
1004 node = node->rb_right;
1005 else
1006 return dc;
1007 }
1008 return NULL;
1009}
1010
1011static struct discard_cmd *__lookup_discard_cmd_ret(struct rb_root_cached *root,
1012 block_t blkaddr,
1013 struct discard_cmd **prev_entry,
1014 struct discard_cmd **next_entry,
1015 struct rb_node ***insert_p,
1016 struct rb_node **insert_parent)
1017{
1018 struct rb_node **pnode = &root->rb_root.rb_node;
1019 struct rb_node *parent = NULL, *tmp_node;
1020 struct discard_cmd *dc;
1021
1022 *insert_p = NULL;
1023 *insert_parent = NULL;
1024 *prev_entry = NULL;
1025 *next_entry = NULL;
1026
1027 if (RB_EMPTY_ROOT(&root->rb_root))
1028 return NULL;
1029
1030 while (*pnode) {
1031 parent = *pnode;
1032 dc = rb_entry(*pnode, struct discard_cmd, rb_node);
1033
1034 if (blkaddr < dc->di.lstart)
1035 pnode = &(*pnode)->rb_left;
1036 else if (blkaddr >= dc->di.lstart + dc->di.len)
1037 pnode = &(*pnode)->rb_right;
1038 else
1039 goto lookup_neighbors;
1040 }
1041
1042 *insert_p = pnode;
1043 *insert_parent = parent;
1044
1045 dc = rb_entry(parent, struct discard_cmd, rb_node);
1046 tmp_node = parent;
1047 if (parent && blkaddr > dc->di.lstart)
1048 tmp_node = rb_next(parent);
1049 *next_entry = rb_entry_safe(tmp_node, struct discard_cmd, rb_node);
1050
1051 tmp_node = parent;
1052 if (parent && blkaddr < dc->di.lstart)
1053 tmp_node = rb_prev(parent);
1054 *prev_entry = rb_entry_safe(tmp_node, struct discard_cmd, rb_node);
1055 return NULL;
1056
1057lookup_neighbors:
1058 /* lookup prev node for merging backward later */
1059 tmp_node = rb_prev(&dc->rb_node);
1060 *prev_entry = rb_entry_safe(tmp_node, struct discard_cmd, rb_node);
1061
1062 /* lookup next node for merging frontward later */
1063 tmp_node = rb_next(&dc->rb_node);
1064 *next_entry = rb_entry_safe(tmp_node, struct discard_cmd, rb_node);
1065 return dc;
1066}
1067
1068static void __detach_discard_cmd(struct discard_cmd_control *dcc,
1069 struct discard_cmd *dc)
1070{
1071 if (dc->state == D_DONE)
1072 atomic_sub(dc->queued, &dcc->queued_discard);
1073
1074 list_del(&dc->list);
1075 rb_erase_cached(&dc->rb_node, &dcc->root);
1076 dcc->undiscard_blks -= dc->di.len;
1077
1078 kmem_cache_free(discard_cmd_slab, dc);
1079
1080 atomic_dec(&dcc->discard_cmd_cnt);
1081}
1082
1083static void __remove_discard_cmd(struct f2fs_sb_info *sbi,
1084 struct discard_cmd *dc)
1085{
1086 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1087 unsigned long flags;
1088
1089 trace_f2fs_remove_discard(dc->bdev, dc->di.start, dc->di.len);
1090
1091 spin_lock_irqsave(&dc->lock, flags);
1092 if (dc->bio_ref) {
1093 spin_unlock_irqrestore(&dc->lock, flags);
1094 return;
1095 }
1096 spin_unlock_irqrestore(&dc->lock, flags);
1097
1098 f2fs_bug_on(sbi, dc->ref);
1099
1100 if (dc->error == -EOPNOTSUPP)
1101 dc->error = 0;
1102
1103 if (dc->error)
1104 printk_ratelimited(
1105 "%sF2FS-fs (%s): Issue discard(%u, %u, %u) failed, ret: %d",
1106 KERN_INFO, sbi->sb->s_id,
1107 dc->di.lstart, dc->di.start, dc->di.len, dc->error);
1108 __detach_discard_cmd(dcc, dc);
1109}
1110
1111static void f2fs_submit_discard_endio(struct bio *bio)
1112{
1113 struct discard_cmd *dc = (struct discard_cmd *)bio->bi_private;
1114 unsigned long flags;
1115
1116 spin_lock_irqsave(&dc->lock, flags);
1117 if (!dc->error)
1118 dc->error = blk_status_to_errno(bio->bi_status);
1119 dc->bio_ref--;
1120 if (!dc->bio_ref && dc->state == D_SUBMIT) {
1121 dc->state = D_DONE;
1122 complete_all(&dc->wait);
1123 }
1124 spin_unlock_irqrestore(&dc->lock, flags);
1125 bio_put(bio);
1126}
1127
1128static void __check_sit_bitmap(struct f2fs_sb_info *sbi,
1129 block_t start, block_t end)
1130{
1131#ifdef CONFIG_F2FS_CHECK_FS
1132 struct seg_entry *sentry;
1133 unsigned int segno;
1134 block_t blk = start;
1135 unsigned long offset, size, max_blocks = sbi->blocks_per_seg;
1136 unsigned long *map;
1137
1138 while (blk < end) {
1139 segno = GET_SEGNO(sbi, blk);
1140 sentry = get_seg_entry(sbi, segno);
1141 offset = GET_BLKOFF_FROM_SEG0(sbi, blk);
1142
1143 if (end < START_BLOCK(sbi, segno + 1))
1144 size = GET_BLKOFF_FROM_SEG0(sbi, end);
1145 else
1146 size = max_blocks;
1147 map = (unsigned long *)(sentry->cur_valid_map);
1148 offset = __find_rev_next_bit(map, size, offset);
1149 f2fs_bug_on(sbi, offset != size);
1150 blk = START_BLOCK(sbi, segno + 1);
1151 }
1152#endif
1153}
1154
1155static void __init_discard_policy(struct f2fs_sb_info *sbi,
1156 struct discard_policy *dpolicy,
1157 int discard_type, unsigned int granularity)
1158{
1159 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1160
1161 /* common policy */
1162 dpolicy->type = discard_type;
1163 dpolicy->sync = true;
1164 dpolicy->ordered = false;
1165 dpolicy->granularity = granularity;
1166
1167 dpolicy->max_requests = dcc->max_discard_request;
1168 dpolicy->io_aware_gran = dcc->discard_io_aware_gran;
1169 dpolicy->timeout = false;
1170
1171 if (discard_type == DPOLICY_BG) {
1172 dpolicy->min_interval = dcc->min_discard_issue_time;
1173 dpolicy->mid_interval = dcc->mid_discard_issue_time;
1174 dpolicy->max_interval = dcc->max_discard_issue_time;
1175 if (dcc->discard_io_aware == DPOLICY_IO_AWARE_ENABLE)
1176 dpolicy->io_aware = true;
1177 else if (dcc->discard_io_aware == DPOLICY_IO_AWARE_DISABLE)
1178 dpolicy->io_aware = false;
1179 dpolicy->sync = false;
1180 dpolicy->ordered = true;
1181 if (utilization(sbi) > dcc->discard_urgent_util) {
1182 dpolicy->granularity = MIN_DISCARD_GRANULARITY;
1183 if (atomic_read(&dcc->discard_cmd_cnt))
1184 dpolicy->max_interval =
1185 dcc->min_discard_issue_time;
1186 }
1187 } else if (discard_type == DPOLICY_FORCE) {
1188 dpolicy->min_interval = dcc->min_discard_issue_time;
1189 dpolicy->mid_interval = dcc->mid_discard_issue_time;
1190 dpolicy->max_interval = dcc->max_discard_issue_time;
1191 dpolicy->io_aware = false;
1192 } else if (discard_type == DPOLICY_FSTRIM) {
1193 dpolicy->io_aware = false;
1194 } else if (discard_type == DPOLICY_UMOUNT) {
1195 dpolicy->io_aware = false;
1196 /* we need to issue all to keep CP_TRIMMED_FLAG */
1197 dpolicy->granularity = MIN_DISCARD_GRANULARITY;
1198 dpolicy->timeout = true;
1199 }
1200}
1201
1202static void __update_discard_tree_range(struct f2fs_sb_info *sbi,
1203 struct block_device *bdev, block_t lstart,
1204 block_t start, block_t len);
1205
1206#ifdef CONFIG_BLK_DEV_ZONED
1207static void __submit_zone_reset_cmd(struct f2fs_sb_info *sbi,
1208 struct discard_cmd *dc, blk_opf_t flag,
1209 struct list_head *wait_list,
1210 unsigned int *issued)
1211{
1212 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1213 struct block_device *bdev = dc->bdev;
1214 struct bio *bio = bio_alloc(bdev, 0, REQ_OP_ZONE_RESET | flag, GFP_NOFS);
1215 unsigned long flags;
1216
1217 trace_f2fs_issue_reset_zone(bdev, dc->di.start);
1218
1219 spin_lock_irqsave(&dc->lock, flags);
1220 dc->state = D_SUBMIT;
1221 dc->bio_ref++;
1222 spin_unlock_irqrestore(&dc->lock, flags);
1223
1224 if (issued)
1225 (*issued)++;
1226
1227 atomic_inc(&dcc->queued_discard);
1228 dc->queued++;
1229 list_move_tail(&dc->list, wait_list);
1230
1231 /* sanity check on discard range */
1232 __check_sit_bitmap(sbi, dc->di.lstart, dc->di.lstart + dc->di.len);
1233
1234 bio->bi_iter.bi_sector = SECTOR_FROM_BLOCK(dc->di.start);
1235 bio->bi_private = dc;
1236 bio->bi_end_io = f2fs_submit_discard_endio;
1237 submit_bio(bio);
1238
1239 atomic_inc(&dcc->issued_discard);
1240 f2fs_update_iostat(sbi, NULL, FS_ZONE_RESET_IO, dc->di.len * F2FS_BLKSIZE);
1241}
1242#endif
1243
1244/* this function is copied from blkdev_issue_discard from block/blk-lib.c */
1245static int __submit_discard_cmd(struct f2fs_sb_info *sbi,
1246 struct discard_policy *dpolicy,
1247 struct discard_cmd *dc, int *issued)
1248{
1249 struct block_device *bdev = dc->bdev;
1250 unsigned int max_discard_blocks =
1251 SECTOR_TO_BLOCK(bdev_max_discard_sectors(bdev));
1252 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1253 struct list_head *wait_list = (dpolicy->type == DPOLICY_FSTRIM) ?
1254 &(dcc->fstrim_list) : &(dcc->wait_list);
1255 blk_opf_t flag = dpolicy->sync ? REQ_SYNC : 0;
1256 block_t lstart, start, len, total_len;
1257 int err = 0;
1258
1259 if (dc->state != D_PREP)
1260 return 0;
1261
1262 if (is_sbi_flag_set(sbi, SBI_NEED_FSCK))
1263 return 0;
1264
1265#ifdef CONFIG_BLK_DEV_ZONED
1266 if (f2fs_sb_has_blkzoned(sbi) && bdev_is_zoned(bdev)) {
1267 int devi = f2fs_bdev_index(sbi, bdev);
1268
1269 if (devi < 0)
1270 return -EINVAL;
1271
1272 if (f2fs_blkz_is_seq(sbi, devi, dc->di.start)) {
1273 __submit_zone_reset_cmd(sbi, dc, flag,
1274 wait_list, issued);
1275 return 0;
1276 }
1277 }
1278#endif
1279
1280 trace_f2fs_issue_discard(bdev, dc->di.start, dc->di.len);
1281
1282 lstart = dc->di.lstart;
1283 start = dc->di.start;
1284 len = dc->di.len;
1285 total_len = len;
1286
1287 dc->di.len = 0;
1288
1289 while (total_len && *issued < dpolicy->max_requests && !err) {
1290 struct bio *bio = NULL;
1291 unsigned long flags;
1292 bool last = true;
1293
1294 if (len > max_discard_blocks) {
1295 len = max_discard_blocks;
1296 last = false;
1297 }
1298
1299 (*issued)++;
1300 if (*issued == dpolicy->max_requests)
1301 last = true;
1302
1303 dc->di.len += len;
1304
1305 if (time_to_inject(sbi, FAULT_DISCARD)) {
1306 err = -EIO;
1307 } else {
1308 err = __blkdev_issue_discard(bdev,
1309 SECTOR_FROM_BLOCK(start),
1310 SECTOR_FROM_BLOCK(len),
1311 GFP_NOFS, &bio);
1312 }
1313 if (err) {
1314 spin_lock_irqsave(&dc->lock, flags);
1315 if (dc->state == D_PARTIAL)
1316 dc->state = D_SUBMIT;
1317 spin_unlock_irqrestore(&dc->lock, flags);
1318
1319 break;
1320 }
1321
1322 f2fs_bug_on(sbi, !bio);
1323
1324 /*
1325 * should keep before submission to avoid D_DONE
1326 * right away
1327 */
1328 spin_lock_irqsave(&dc->lock, flags);
1329 if (last)
1330 dc->state = D_SUBMIT;
1331 else
1332 dc->state = D_PARTIAL;
1333 dc->bio_ref++;
1334 spin_unlock_irqrestore(&dc->lock, flags);
1335
1336 atomic_inc(&dcc->queued_discard);
1337 dc->queued++;
1338 list_move_tail(&dc->list, wait_list);
1339
1340 /* sanity check on discard range */
1341 __check_sit_bitmap(sbi, lstart, lstart + len);
1342
1343 bio->bi_private = dc;
1344 bio->bi_end_io = f2fs_submit_discard_endio;
1345 bio->bi_opf |= flag;
1346 submit_bio(bio);
1347
1348 atomic_inc(&dcc->issued_discard);
1349
1350 f2fs_update_iostat(sbi, NULL, FS_DISCARD_IO, len * F2FS_BLKSIZE);
1351
1352 lstart += len;
1353 start += len;
1354 total_len -= len;
1355 len = total_len;
1356 }
1357
1358 if (!err && len) {
1359 dcc->undiscard_blks -= len;
1360 __update_discard_tree_range(sbi, bdev, lstart, start, len);
1361 }
1362 return err;
1363}
1364
1365static void __insert_discard_cmd(struct f2fs_sb_info *sbi,
1366 struct block_device *bdev, block_t lstart,
1367 block_t start, block_t len)
1368{
1369 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1370 struct rb_node **p = &dcc->root.rb_root.rb_node;
1371 struct rb_node *parent = NULL;
1372 struct discard_cmd *dc;
1373 bool leftmost = true;
1374
1375 /* look up rb tree to find parent node */
1376 while (*p) {
1377 parent = *p;
1378 dc = rb_entry(parent, struct discard_cmd, rb_node);
1379
1380 if (lstart < dc->di.lstart) {
1381 p = &(*p)->rb_left;
1382 } else if (lstart >= dc->di.lstart + dc->di.len) {
1383 p = &(*p)->rb_right;
1384 leftmost = false;
1385 } else {
1386 /* Let's skip to add, if exists */
1387 return;
1388 }
1389 }
1390
1391 dc = __create_discard_cmd(sbi, bdev, lstart, start, len);
1392
1393 rb_link_node(&dc->rb_node, parent, p);
1394 rb_insert_color_cached(&dc->rb_node, &dcc->root, leftmost);
1395}
1396
1397static void __relocate_discard_cmd(struct discard_cmd_control *dcc,
1398 struct discard_cmd *dc)
1399{
1400 list_move_tail(&dc->list, &dcc->pend_list[plist_idx(dc->di.len)]);
1401}
1402
1403static void __punch_discard_cmd(struct f2fs_sb_info *sbi,
1404 struct discard_cmd *dc, block_t blkaddr)
1405{
1406 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1407 struct discard_info di = dc->di;
1408 bool modified = false;
1409
1410 if (dc->state == D_DONE || dc->di.len == 1) {
1411 __remove_discard_cmd(sbi, dc);
1412 return;
1413 }
1414
1415 dcc->undiscard_blks -= di.len;
1416
1417 if (blkaddr > di.lstart) {
1418 dc->di.len = blkaddr - dc->di.lstart;
1419 dcc->undiscard_blks += dc->di.len;
1420 __relocate_discard_cmd(dcc, dc);
1421 modified = true;
1422 }
1423
1424 if (blkaddr < di.lstart + di.len - 1) {
1425 if (modified) {
1426 __insert_discard_cmd(sbi, dc->bdev, blkaddr + 1,
1427 di.start + blkaddr + 1 - di.lstart,
1428 di.lstart + di.len - 1 - blkaddr);
1429 } else {
1430 dc->di.lstart++;
1431 dc->di.len--;
1432 dc->di.start++;
1433 dcc->undiscard_blks += dc->di.len;
1434 __relocate_discard_cmd(dcc, dc);
1435 }
1436 }
1437}
1438
1439static void __update_discard_tree_range(struct f2fs_sb_info *sbi,
1440 struct block_device *bdev, block_t lstart,
1441 block_t start, block_t len)
1442{
1443 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1444 struct discard_cmd *prev_dc = NULL, *next_dc = NULL;
1445 struct discard_cmd *dc;
1446 struct discard_info di = {0};
1447 struct rb_node **insert_p = NULL, *insert_parent = NULL;
1448 unsigned int max_discard_blocks =
1449 SECTOR_TO_BLOCK(bdev_max_discard_sectors(bdev));
1450 block_t end = lstart + len;
1451
1452 dc = __lookup_discard_cmd_ret(&dcc->root, lstart,
1453 &prev_dc, &next_dc, &insert_p, &insert_parent);
1454 if (dc)
1455 prev_dc = dc;
1456
1457 if (!prev_dc) {
1458 di.lstart = lstart;
1459 di.len = next_dc ? next_dc->di.lstart - lstart : len;
1460 di.len = min(di.len, len);
1461 di.start = start;
1462 }
1463
1464 while (1) {
1465 struct rb_node *node;
1466 bool merged = false;
1467 struct discard_cmd *tdc = NULL;
1468
1469 if (prev_dc) {
1470 di.lstart = prev_dc->di.lstart + prev_dc->di.len;
1471 if (di.lstart < lstart)
1472 di.lstart = lstart;
1473 if (di.lstart >= end)
1474 break;
1475
1476 if (!next_dc || next_dc->di.lstart > end)
1477 di.len = end - di.lstart;
1478 else
1479 di.len = next_dc->di.lstart - di.lstart;
1480 di.start = start + di.lstart - lstart;
1481 }
1482
1483 if (!di.len)
1484 goto next;
1485
1486 if (prev_dc && prev_dc->state == D_PREP &&
1487 prev_dc->bdev == bdev &&
1488 __is_discard_back_mergeable(&di, &prev_dc->di,
1489 max_discard_blocks)) {
1490 prev_dc->di.len += di.len;
1491 dcc->undiscard_blks += di.len;
1492 __relocate_discard_cmd(dcc, prev_dc);
1493 di = prev_dc->di;
1494 tdc = prev_dc;
1495 merged = true;
1496 }
1497
1498 if (next_dc && next_dc->state == D_PREP &&
1499 next_dc->bdev == bdev &&
1500 __is_discard_front_mergeable(&di, &next_dc->di,
1501 max_discard_blocks)) {
1502 next_dc->di.lstart = di.lstart;
1503 next_dc->di.len += di.len;
1504 next_dc->di.start = di.start;
1505 dcc->undiscard_blks += di.len;
1506 __relocate_discard_cmd(dcc, next_dc);
1507 if (tdc)
1508 __remove_discard_cmd(sbi, tdc);
1509 merged = true;
1510 }
1511
1512 if (!merged)
1513 __insert_discard_cmd(sbi, bdev,
1514 di.lstart, di.start, di.len);
1515 next:
1516 prev_dc = next_dc;
1517 if (!prev_dc)
1518 break;
1519
1520 node = rb_next(&prev_dc->rb_node);
1521 next_dc = rb_entry_safe(node, struct discard_cmd, rb_node);
1522 }
1523}
1524
1525#ifdef CONFIG_BLK_DEV_ZONED
1526static void __queue_zone_reset_cmd(struct f2fs_sb_info *sbi,
1527 struct block_device *bdev, block_t blkstart, block_t lblkstart,
1528 block_t blklen)
1529{
1530 trace_f2fs_queue_reset_zone(bdev, blkstart);
1531
1532 mutex_lock(&SM_I(sbi)->dcc_info->cmd_lock);
1533 __insert_discard_cmd(sbi, bdev, lblkstart, blkstart, blklen);
1534 mutex_unlock(&SM_I(sbi)->dcc_info->cmd_lock);
1535}
1536#endif
1537
1538static void __queue_discard_cmd(struct f2fs_sb_info *sbi,
1539 struct block_device *bdev, block_t blkstart, block_t blklen)
1540{
1541 block_t lblkstart = blkstart;
1542
1543 if (!f2fs_bdev_support_discard(bdev))
1544 return;
1545
1546 trace_f2fs_queue_discard(bdev, blkstart, blklen);
1547
1548 if (f2fs_is_multi_device(sbi)) {
1549 int devi = f2fs_target_device_index(sbi, blkstart);
1550
1551 blkstart -= FDEV(devi).start_blk;
1552 }
1553 mutex_lock(&SM_I(sbi)->dcc_info->cmd_lock);
1554 __update_discard_tree_range(sbi, bdev, lblkstart, blkstart, blklen);
1555 mutex_unlock(&SM_I(sbi)->dcc_info->cmd_lock);
1556}
1557
1558static void __issue_discard_cmd_orderly(struct f2fs_sb_info *sbi,
1559 struct discard_policy *dpolicy, int *issued)
1560{
1561 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1562 struct discard_cmd *prev_dc = NULL, *next_dc = NULL;
1563 struct rb_node **insert_p = NULL, *insert_parent = NULL;
1564 struct discard_cmd *dc;
1565 struct blk_plug plug;
1566 bool io_interrupted = false;
1567
1568 mutex_lock(&dcc->cmd_lock);
1569 dc = __lookup_discard_cmd_ret(&dcc->root, dcc->next_pos,
1570 &prev_dc, &next_dc, &insert_p, &insert_parent);
1571 if (!dc)
1572 dc = next_dc;
1573
1574 blk_start_plug(&plug);
1575
1576 while (dc) {
1577 struct rb_node *node;
1578 int err = 0;
1579
1580 if (dc->state != D_PREP)
1581 goto next;
1582
1583 if (dpolicy->io_aware && !is_idle(sbi, DISCARD_TIME)) {
1584 io_interrupted = true;
1585 break;
1586 }
1587
1588 dcc->next_pos = dc->di.lstart + dc->di.len;
1589 err = __submit_discard_cmd(sbi, dpolicy, dc, issued);
1590
1591 if (*issued >= dpolicy->max_requests)
1592 break;
1593next:
1594 node = rb_next(&dc->rb_node);
1595 if (err)
1596 __remove_discard_cmd(sbi, dc);
1597 dc = rb_entry_safe(node, struct discard_cmd, rb_node);
1598 }
1599
1600 blk_finish_plug(&plug);
1601
1602 if (!dc)
1603 dcc->next_pos = 0;
1604
1605 mutex_unlock(&dcc->cmd_lock);
1606
1607 if (!(*issued) && io_interrupted)
1608 *issued = -1;
1609}
1610static unsigned int __wait_all_discard_cmd(struct f2fs_sb_info *sbi,
1611 struct discard_policy *dpolicy);
1612
1613static int __issue_discard_cmd(struct f2fs_sb_info *sbi,
1614 struct discard_policy *dpolicy)
1615{
1616 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1617 struct list_head *pend_list;
1618 struct discard_cmd *dc, *tmp;
1619 struct blk_plug plug;
1620 int i, issued;
1621 bool io_interrupted = false;
1622
1623 if (dpolicy->timeout)
1624 f2fs_update_time(sbi, UMOUNT_DISCARD_TIMEOUT);
1625
1626retry:
1627 issued = 0;
1628 for (i = MAX_PLIST_NUM - 1; i >= 0; i--) {
1629 if (dpolicy->timeout &&
1630 f2fs_time_over(sbi, UMOUNT_DISCARD_TIMEOUT))
1631 break;
1632
1633 if (i + 1 < dpolicy->granularity)
1634 break;
1635
1636 if (i + 1 < dcc->max_ordered_discard && dpolicy->ordered) {
1637 __issue_discard_cmd_orderly(sbi, dpolicy, &issued);
1638 return issued;
1639 }
1640
1641 pend_list = &dcc->pend_list[i];
1642
1643 mutex_lock(&dcc->cmd_lock);
1644 if (list_empty(pend_list))
1645 goto next;
1646 if (unlikely(dcc->rbtree_check))
1647 f2fs_bug_on(sbi, !f2fs_check_discard_tree(sbi));
1648 blk_start_plug(&plug);
1649 list_for_each_entry_safe(dc, tmp, pend_list, list) {
1650 f2fs_bug_on(sbi, dc->state != D_PREP);
1651
1652 if (dpolicy->timeout &&
1653 f2fs_time_over(sbi, UMOUNT_DISCARD_TIMEOUT))
1654 break;
1655
1656 if (dpolicy->io_aware && i < dpolicy->io_aware_gran &&
1657 !is_idle(sbi, DISCARD_TIME)) {
1658 io_interrupted = true;
1659 break;
1660 }
1661
1662 __submit_discard_cmd(sbi, dpolicy, dc, &issued);
1663
1664 if (issued >= dpolicy->max_requests)
1665 break;
1666 }
1667 blk_finish_plug(&plug);
1668next:
1669 mutex_unlock(&dcc->cmd_lock);
1670
1671 if (issued >= dpolicy->max_requests || io_interrupted)
1672 break;
1673 }
1674
1675 if (dpolicy->type == DPOLICY_UMOUNT && issued) {
1676 __wait_all_discard_cmd(sbi, dpolicy);
1677 goto retry;
1678 }
1679
1680 if (!issued && io_interrupted)
1681 issued = -1;
1682
1683 return issued;
1684}
1685
1686static bool __drop_discard_cmd(struct f2fs_sb_info *sbi)
1687{
1688 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1689 struct list_head *pend_list;
1690 struct discard_cmd *dc, *tmp;
1691 int i;
1692 bool dropped = false;
1693
1694 mutex_lock(&dcc->cmd_lock);
1695 for (i = MAX_PLIST_NUM - 1; i >= 0; i--) {
1696 pend_list = &dcc->pend_list[i];
1697 list_for_each_entry_safe(dc, tmp, pend_list, list) {
1698 f2fs_bug_on(sbi, dc->state != D_PREP);
1699 __remove_discard_cmd(sbi, dc);
1700 dropped = true;
1701 }
1702 }
1703 mutex_unlock(&dcc->cmd_lock);
1704
1705 return dropped;
1706}
1707
1708void f2fs_drop_discard_cmd(struct f2fs_sb_info *sbi)
1709{
1710 __drop_discard_cmd(sbi);
1711}
1712
1713static unsigned int __wait_one_discard_bio(struct f2fs_sb_info *sbi,
1714 struct discard_cmd *dc)
1715{
1716 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1717 unsigned int len = 0;
1718
1719 wait_for_completion_io(&dc->wait);
1720 mutex_lock(&dcc->cmd_lock);
1721 f2fs_bug_on(sbi, dc->state != D_DONE);
1722 dc->ref--;
1723 if (!dc->ref) {
1724 if (!dc->error)
1725 len = dc->di.len;
1726 __remove_discard_cmd(sbi, dc);
1727 }
1728 mutex_unlock(&dcc->cmd_lock);
1729
1730 return len;
1731}
1732
1733static unsigned int __wait_discard_cmd_range(struct f2fs_sb_info *sbi,
1734 struct discard_policy *dpolicy,
1735 block_t start, block_t end)
1736{
1737 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1738 struct list_head *wait_list = (dpolicy->type == DPOLICY_FSTRIM) ?
1739 &(dcc->fstrim_list) : &(dcc->wait_list);
1740 struct discard_cmd *dc = NULL, *iter, *tmp;
1741 unsigned int trimmed = 0;
1742
1743next:
1744 dc = NULL;
1745
1746 mutex_lock(&dcc->cmd_lock);
1747 list_for_each_entry_safe(iter, tmp, wait_list, list) {
1748 if (iter->di.lstart + iter->di.len <= start ||
1749 end <= iter->di.lstart)
1750 continue;
1751 if (iter->di.len < dpolicy->granularity)
1752 continue;
1753 if (iter->state == D_DONE && !iter->ref) {
1754 wait_for_completion_io(&iter->wait);
1755 if (!iter->error)
1756 trimmed += iter->di.len;
1757 __remove_discard_cmd(sbi, iter);
1758 } else {
1759 iter->ref++;
1760 dc = iter;
1761 break;
1762 }
1763 }
1764 mutex_unlock(&dcc->cmd_lock);
1765
1766 if (dc) {
1767 trimmed += __wait_one_discard_bio(sbi, dc);
1768 goto next;
1769 }
1770
1771 return trimmed;
1772}
1773
1774static unsigned int __wait_all_discard_cmd(struct f2fs_sb_info *sbi,
1775 struct discard_policy *dpolicy)
1776{
1777 struct discard_policy dp;
1778 unsigned int discard_blks;
1779
1780 if (dpolicy)
1781 return __wait_discard_cmd_range(sbi, dpolicy, 0, UINT_MAX);
1782
1783 /* wait all */
1784 __init_discard_policy(sbi, &dp, DPOLICY_FSTRIM, MIN_DISCARD_GRANULARITY);
1785 discard_blks = __wait_discard_cmd_range(sbi, &dp, 0, UINT_MAX);
1786 __init_discard_policy(sbi, &dp, DPOLICY_UMOUNT, MIN_DISCARD_GRANULARITY);
1787 discard_blks += __wait_discard_cmd_range(sbi, &dp, 0, UINT_MAX);
1788
1789 return discard_blks;
1790}
1791
1792/* This should be covered by global mutex, &sit_i->sentry_lock */
1793static void f2fs_wait_discard_bio(struct f2fs_sb_info *sbi, block_t blkaddr)
1794{
1795 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1796 struct discard_cmd *dc;
1797 bool need_wait = false;
1798
1799 mutex_lock(&dcc->cmd_lock);
1800 dc = __lookup_discard_cmd(sbi, blkaddr);
1801#ifdef CONFIG_BLK_DEV_ZONED
1802 if (dc && f2fs_sb_has_blkzoned(sbi) && bdev_is_zoned(dc->bdev)) {
1803 int devi = f2fs_bdev_index(sbi, dc->bdev);
1804
1805 if (devi < 0) {
1806 mutex_unlock(&dcc->cmd_lock);
1807 return;
1808 }
1809
1810 if (f2fs_blkz_is_seq(sbi, devi, dc->di.start)) {
1811 /* force submit zone reset */
1812 if (dc->state == D_PREP)
1813 __submit_zone_reset_cmd(sbi, dc, REQ_SYNC,
1814 &dcc->wait_list, NULL);
1815 dc->ref++;
1816 mutex_unlock(&dcc->cmd_lock);
1817 /* wait zone reset */
1818 __wait_one_discard_bio(sbi, dc);
1819 return;
1820 }
1821 }
1822#endif
1823 if (dc) {
1824 if (dc->state == D_PREP) {
1825 __punch_discard_cmd(sbi, dc, blkaddr);
1826 } else {
1827 dc->ref++;
1828 need_wait = true;
1829 }
1830 }
1831 mutex_unlock(&dcc->cmd_lock);
1832
1833 if (need_wait)
1834 __wait_one_discard_bio(sbi, dc);
1835}
1836
1837void f2fs_stop_discard_thread(struct f2fs_sb_info *sbi)
1838{
1839 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1840
1841 if (dcc && dcc->f2fs_issue_discard) {
1842 struct task_struct *discard_thread = dcc->f2fs_issue_discard;
1843
1844 dcc->f2fs_issue_discard = NULL;
1845 kthread_stop(discard_thread);
1846 }
1847}
1848
1849/**
1850 * f2fs_issue_discard_timeout() - Issue all discard cmd within UMOUNT_DISCARD_TIMEOUT
1851 * @sbi: the f2fs_sb_info data for discard cmd to issue
1852 *
1853 * When UMOUNT_DISCARD_TIMEOUT is exceeded, all remaining discard commands will be dropped
1854 *
1855 * Return true if issued all discard cmd or no discard cmd need issue, otherwise return false.
1856 */
1857bool f2fs_issue_discard_timeout(struct f2fs_sb_info *sbi)
1858{
1859 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1860 struct discard_policy dpolicy;
1861 bool dropped;
1862
1863 if (!atomic_read(&dcc->discard_cmd_cnt))
1864 return true;
1865
1866 __init_discard_policy(sbi, &dpolicy, DPOLICY_UMOUNT,
1867 dcc->discard_granularity);
1868 __issue_discard_cmd(sbi, &dpolicy);
1869 dropped = __drop_discard_cmd(sbi);
1870
1871 /* just to make sure there is no pending discard commands */
1872 __wait_all_discard_cmd(sbi, NULL);
1873
1874 f2fs_bug_on(sbi, atomic_read(&dcc->discard_cmd_cnt));
1875 return !dropped;
1876}
1877
1878static int issue_discard_thread(void *data)
1879{
1880 struct f2fs_sb_info *sbi = data;
1881 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
1882 wait_queue_head_t *q = &dcc->discard_wait_queue;
1883 struct discard_policy dpolicy;
1884 unsigned int wait_ms = dcc->min_discard_issue_time;
1885 int issued;
1886
1887 set_freezable();
1888
1889 do {
1890 wait_event_freezable_timeout(*q,
1891 kthread_should_stop() || dcc->discard_wake,
1892 msecs_to_jiffies(wait_ms));
1893
1894 if (sbi->gc_mode == GC_URGENT_HIGH ||
1895 !f2fs_available_free_memory(sbi, DISCARD_CACHE))
1896 __init_discard_policy(sbi, &dpolicy, DPOLICY_FORCE,
1897 MIN_DISCARD_GRANULARITY);
1898 else
1899 __init_discard_policy(sbi, &dpolicy, DPOLICY_BG,
1900 dcc->discard_granularity);
1901
1902 if (dcc->discard_wake)
1903 dcc->discard_wake = false;
1904
1905 /* clean up pending candidates before going to sleep */
1906 if (atomic_read(&dcc->queued_discard))
1907 __wait_all_discard_cmd(sbi, NULL);
1908
1909 if (f2fs_readonly(sbi->sb))
1910 continue;
1911 if (kthread_should_stop())
1912 return 0;
1913 if (is_sbi_flag_set(sbi, SBI_NEED_FSCK) ||
1914 !atomic_read(&dcc->discard_cmd_cnt)) {
1915 wait_ms = dpolicy.max_interval;
1916 continue;
1917 }
1918
1919 sb_start_intwrite(sbi->sb);
1920
1921 issued = __issue_discard_cmd(sbi, &dpolicy);
1922 if (issued > 0) {
1923 __wait_all_discard_cmd(sbi, &dpolicy);
1924 wait_ms = dpolicy.min_interval;
1925 } else if (issued == -1) {
1926 wait_ms = f2fs_time_to_wait(sbi, DISCARD_TIME);
1927 if (!wait_ms)
1928 wait_ms = dpolicy.mid_interval;
1929 } else {
1930 wait_ms = dpolicy.max_interval;
1931 }
1932 if (!atomic_read(&dcc->discard_cmd_cnt))
1933 wait_ms = dpolicy.max_interval;
1934
1935 sb_end_intwrite(sbi->sb);
1936
1937 } while (!kthread_should_stop());
1938 return 0;
1939}
1940
1941#ifdef CONFIG_BLK_DEV_ZONED
1942static int __f2fs_issue_discard_zone(struct f2fs_sb_info *sbi,
1943 struct block_device *bdev, block_t blkstart, block_t blklen)
1944{
1945 sector_t sector, nr_sects;
1946 block_t lblkstart = blkstart;
1947 int devi = 0;
1948 u64 remainder = 0;
1949
1950 if (f2fs_is_multi_device(sbi)) {
1951 devi = f2fs_target_device_index(sbi, blkstart);
1952 if (blkstart < FDEV(devi).start_blk ||
1953 blkstart > FDEV(devi).end_blk) {
1954 f2fs_err(sbi, "Invalid block %x", blkstart);
1955 return -EIO;
1956 }
1957 blkstart -= FDEV(devi).start_blk;
1958 }
1959
1960 /* For sequential zones, reset the zone write pointer */
1961 if (f2fs_blkz_is_seq(sbi, devi, blkstart)) {
1962 sector = SECTOR_FROM_BLOCK(blkstart);
1963 nr_sects = SECTOR_FROM_BLOCK(blklen);
1964 div64_u64_rem(sector, bdev_zone_sectors(bdev), &remainder);
1965
1966 if (remainder || nr_sects != bdev_zone_sectors(bdev)) {
1967 f2fs_err(sbi, "(%d) %s: Unaligned zone reset attempted (block %x + %x)",
1968 devi, sbi->s_ndevs ? FDEV(devi).path : "",
1969 blkstart, blklen);
1970 return -EIO;
1971 }
1972
1973 if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING))) {
1974 trace_f2fs_issue_reset_zone(bdev, blkstart);
1975 return blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
1976 sector, nr_sects, GFP_NOFS);
1977 }
1978
1979 __queue_zone_reset_cmd(sbi, bdev, blkstart, lblkstart, blklen);
1980 return 0;
1981 }
1982
1983 /* For conventional zones, use regular discard if supported */
1984 __queue_discard_cmd(sbi, bdev, lblkstart, blklen);
1985 return 0;
1986}
1987#endif
1988
1989static int __issue_discard_async(struct f2fs_sb_info *sbi,
1990 struct block_device *bdev, block_t blkstart, block_t blklen)
1991{
1992#ifdef CONFIG_BLK_DEV_ZONED
1993 if (f2fs_sb_has_blkzoned(sbi) && bdev_is_zoned(bdev))
1994 return __f2fs_issue_discard_zone(sbi, bdev, blkstart, blklen);
1995#endif
1996 __queue_discard_cmd(sbi, bdev, blkstart, blklen);
1997 return 0;
1998}
1999
2000static int f2fs_issue_discard(struct f2fs_sb_info *sbi,
2001 block_t blkstart, block_t blklen)
2002{
2003 sector_t start = blkstart, len = 0;
2004 struct block_device *bdev;
2005 struct seg_entry *se;
2006 unsigned int offset;
2007 block_t i;
2008 int err = 0;
2009
2010 bdev = f2fs_target_device(sbi, blkstart, NULL);
2011
2012 for (i = blkstart; i < blkstart + blklen; i++, len++) {
2013 if (i != start) {
2014 struct block_device *bdev2 =
2015 f2fs_target_device(sbi, i, NULL);
2016
2017 if (bdev2 != bdev) {
2018 err = __issue_discard_async(sbi, bdev,
2019 start, len);
2020 if (err)
2021 return err;
2022 bdev = bdev2;
2023 start = i;
2024 len = 0;
2025 }
2026 }
2027
2028 se = get_seg_entry(sbi, GET_SEGNO(sbi, i));
2029 offset = GET_BLKOFF_FROM_SEG0(sbi, i);
2030
2031 if (f2fs_block_unit_discard(sbi) &&
2032 !f2fs_test_and_set_bit(offset, se->discard_map))
2033 sbi->discard_blks--;
2034 }
2035
2036 if (len)
2037 err = __issue_discard_async(sbi, bdev, start, len);
2038 return err;
2039}
2040
2041static bool add_discard_addrs(struct f2fs_sb_info *sbi, struct cp_control *cpc,
2042 bool check_only)
2043{
2044 int entries = SIT_VBLOCK_MAP_SIZE / sizeof(unsigned long);
2045 int max_blocks = sbi->blocks_per_seg;
2046 struct seg_entry *se = get_seg_entry(sbi, cpc->trim_start);
2047 unsigned long *cur_map = (unsigned long *)se->cur_valid_map;
2048 unsigned long *ckpt_map = (unsigned long *)se->ckpt_valid_map;
2049 unsigned long *discard_map = (unsigned long *)se->discard_map;
2050 unsigned long *dmap = SIT_I(sbi)->tmp_map;
2051 unsigned int start = 0, end = -1;
2052 bool force = (cpc->reason & CP_DISCARD);
2053 struct discard_entry *de = NULL;
2054 struct list_head *head = &SM_I(sbi)->dcc_info->entry_list;
2055 int i;
2056
2057 if (se->valid_blocks == max_blocks || !f2fs_hw_support_discard(sbi) ||
2058 !f2fs_block_unit_discard(sbi))
2059 return false;
2060
2061 if (!force) {
2062 if (!f2fs_realtime_discard_enable(sbi) || !se->valid_blocks ||
2063 SM_I(sbi)->dcc_info->nr_discards >=
2064 SM_I(sbi)->dcc_info->max_discards)
2065 return false;
2066 }
2067
2068 /* SIT_VBLOCK_MAP_SIZE should be multiple of sizeof(unsigned long) */
2069 for (i = 0; i < entries; i++)
2070 dmap[i] = force ? ~ckpt_map[i] & ~discard_map[i] :
2071 (cur_map[i] ^ ckpt_map[i]) & ckpt_map[i];
2072
2073 while (force || SM_I(sbi)->dcc_info->nr_discards <=
2074 SM_I(sbi)->dcc_info->max_discards) {
2075 start = __find_rev_next_bit(dmap, max_blocks, end + 1);
2076 if (start >= max_blocks)
2077 break;
2078
2079 end = __find_rev_next_zero_bit(dmap, max_blocks, start + 1);
2080 if (force && start && end != max_blocks
2081 && (end - start) < cpc->trim_minlen)
2082 continue;
2083
2084 if (check_only)
2085 return true;
2086
2087 if (!de) {
2088 de = f2fs_kmem_cache_alloc(discard_entry_slab,
2089 GFP_F2FS_ZERO, true, NULL);
2090 de->start_blkaddr = START_BLOCK(sbi, cpc->trim_start);
2091 list_add_tail(&de->list, head);
2092 }
2093
2094 for (i = start; i < end; i++)
2095 __set_bit_le(i, (void *)de->discard_map);
2096
2097 SM_I(sbi)->dcc_info->nr_discards += end - start;
2098 }
2099 return false;
2100}
2101
2102static void release_discard_addr(struct discard_entry *entry)
2103{
2104 list_del(&entry->list);
2105 kmem_cache_free(discard_entry_slab, entry);
2106}
2107
2108void f2fs_release_discard_addrs(struct f2fs_sb_info *sbi)
2109{
2110 struct list_head *head = &(SM_I(sbi)->dcc_info->entry_list);
2111 struct discard_entry *entry, *this;
2112
2113 /* drop caches */
2114 list_for_each_entry_safe(entry, this, head, list)
2115 release_discard_addr(entry);
2116}
2117
2118/*
2119 * Should call f2fs_clear_prefree_segments after checkpoint is done.
2120 */
2121static void set_prefree_as_free_segments(struct f2fs_sb_info *sbi)
2122{
2123 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
2124 unsigned int segno;
2125
2126 mutex_lock(&dirty_i->seglist_lock);
2127 for_each_set_bit(segno, dirty_i->dirty_segmap[PRE], MAIN_SEGS(sbi))
2128 __set_test_and_free(sbi, segno, false);
2129 mutex_unlock(&dirty_i->seglist_lock);
2130}
2131
2132void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
2133 struct cp_control *cpc)
2134{
2135 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
2136 struct list_head *head = &dcc->entry_list;
2137 struct discard_entry *entry, *this;
2138 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
2139 unsigned long *prefree_map = dirty_i->dirty_segmap[PRE];
2140 unsigned int start = 0, end = -1;
2141 unsigned int secno, start_segno;
2142 bool force = (cpc->reason & CP_DISCARD);
2143 bool section_alignment = F2FS_OPTION(sbi).discard_unit ==
2144 DISCARD_UNIT_SECTION;
2145
2146 if (f2fs_lfs_mode(sbi) && __is_large_section(sbi))
2147 section_alignment = true;
2148
2149 mutex_lock(&dirty_i->seglist_lock);
2150
2151 while (1) {
2152 int i;
2153
2154 if (section_alignment && end != -1)
2155 end--;
2156 start = find_next_bit(prefree_map, MAIN_SEGS(sbi), end + 1);
2157 if (start >= MAIN_SEGS(sbi))
2158 break;
2159 end = find_next_zero_bit(prefree_map, MAIN_SEGS(sbi),
2160 start + 1);
2161
2162 if (section_alignment) {
2163 start = rounddown(start, sbi->segs_per_sec);
2164 end = roundup(end, sbi->segs_per_sec);
2165 }
2166
2167 for (i = start; i < end; i++) {
2168 if (test_and_clear_bit(i, prefree_map))
2169 dirty_i->nr_dirty[PRE]--;
2170 }
2171
2172 if (!f2fs_realtime_discard_enable(sbi))
2173 continue;
2174
2175 if (force && start >= cpc->trim_start &&
2176 (end - 1) <= cpc->trim_end)
2177 continue;
2178
2179 /* Should cover 2MB zoned device for zone-based reset */
2180 if (!f2fs_sb_has_blkzoned(sbi) &&
2181 (!f2fs_lfs_mode(sbi) || !__is_large_section(sbi))) {
2182 f2fs_issue_discard(sbi, START_BLOCK(sbi, start),
2183 (end - start) << sbi->log_blocks_per_seg);
2184 continue;
2185 }
2186next:
2187 secno = GET_SEC_FROM_SEG(sbi, start);
2188 start_segno = GET_SEG_FROM_SEC(sbi, secno);
2189 if (!IS_CURSEC(sbi, secno) &&
2190 !get_valid_blocks(sbi, start, true))
2191 f2fs_issue_discard(sbi, START_BLOCK(sbi, start_segno),
2192 sbi->segs_per_sec << sbi->log_blocks_per_seg);
2193
2194 start = start_segno + sbi->segs_per_sec;
2195 if (start < end)
2196 goto next;
2197 else
2198 end = start - 1;
2199 }
2200 mutex_unlock(&dirty_i->seglist_lock);
2201
2202 if (!f2fs_block_unit_discard(sbi))
2203 goto wakeup;
2204
2205 /* send small discards */
2206 list_for_each_entry_safe(entry, this, head, list) {
2207 unsigned int cur_pos = 0, next_pos, len, total_len = 0;
2208 bool is_valid = test_bit_le(0, entry->discard_map);
2209
2210find_next:
2211 if (is_valid) {
2212 next_pos = find_next_zero_bit_le(entry->discard_map,
2213 sbi->blocks_per_seg, cur_pos);
2214 len = next_pos - cur_pos;
2215
2216 if (f2fs_sb_has_blkzoned(sbi) ||
2217 (force && len < cpc->trim_minlen))
2218 goto skip;
2219
2220 f2fs_issue_discard(sbi, entry->start_blkaddr + cur_pos,
2221 len);
2222 total_len += len;
2223 } else {
2224 next_pos = find_next_bit_le(entry->discard_map,
2225 sbi->blocks_per_seg, cur_pos);
2226 }
2227skip:
2228 cur_pos = next_pos;
2229 is_valid = !is_valid;
2230
2231 if (cur_pos < sbi->blocks_per_seg)
2232 goto find_next;
2233
2234 release_discard_addr(entry);
2235 dcc->nr_discards -= total_len;
2236 }
2237
2238wakeup:
2239 wake_up_discard_thread(sbi, false);
2240}
2241
2242int f2fs_start_discard_thread(struct f2fs_sb_info *sbi)
2243{
2244 dev_t dev = sbi->sb->s_bdev->bd_dev;
2245 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
2246 int err = 0;
2247
2248 if (!f2fs_realtime_discard_enable(sbi))
2249 return 0;
2250
2251 dcc->f2fs_issue_discard = kthread_run(issue_discard_thread, sbi,
2252 "f2fs_discard-%u:%u", MAJOR(dev), MINOR(dev));
2253 if (IS_ERR(dcc->f2fs_issue_discard)) {
2254 err = PTR_ERR(dcc->f2fs_issue_discard);
2255 dcc->f2fs_issue_discard = NULL;
2256 }
2257
2258 return err;
2259}
2260
2261static int create_discard_cmd_control(struct f2fs_sb_info *sbi)
2262{
2263 struct discard_cmd_control *dcc;
2264 int err = 0, i;
2265
2266 if (SM_I(sbi)->dcc_info) {
2267 dcc = SM_I(sbi)->dcc_info;
2268 goto init_thread;
2269 }
2270
2271 dcc = f2fs_kzalloc(sbi, sizeof(struct discard_cmd_control), GFP_KERNEL);
2272 if (!dcc)
2273 return -ENOMEM;
2274
2275 dcc->discard_io_aware_gran = MAX_PLIST_NUM;
2276 dcc->discard_granularity = DEFAULT_DISCARD_GRANULARITY;
2277 dcc->max_ordered_discard = DEFAULT_MAX_ORDERED_DISCARD_GRANULARITY;
2278 dcc->discard_io_aware = DPOLICY_IO_AWARE_ENABLE;
2279 if (F2FS_OPTION(sbi).discard_unit == DISCARD_UNIT_SEGMENT)
2280 dcc->discard_granularity = sbi->blocks_per_seg;
2281 else if (F2FS_OPTION(sbi).discard_unit == DISCARD_UNIT_SECTION)
2282 dcc->discard_granularity = BLKS_PER_SEC(sbi);
2283
2284 INIT_LIST_HEAD(&dcc->entry_list);
2285 for (i = 0; i < MAX_PLIST_NUM; i++)
2286 INIT_LIST_HEAD(&dcc->pend_list[i]);
2287 INIT_LIST_HEAD(&dcc->wait_list);
2288 INIT_LIST_HEAD(&dcc->fstrim_list);
2289 mutex_init(&dcc->cmd_lock);
2290 atomic_set(&dcc->issued_discard, 0);
2291 atomic_set(&dcc->queued_discard, 0);
2292 atomic_set(&dcc->discard_cmd_cnt, 0);
2293 dcc->nr_discards = 0;
2294 dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
2295 dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST;
2296 dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME;
2297 dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME;
2298 dcc->max_discard_issue_time = DEF_MAX_DISCARD_ISSUE_TIME;
2299 dcc->discard_urgent_util = DEF_DISCARD_URGENT_UTIL;
2300 dcc->undiscard_blks = 0;
2301 dcc->next_pos = 0;
2302 dcc->root = RB_ROOT_CACHED;
2303 dcc->rbtree_check = false;
2304
2305 init_waitqueue_head(&dcc->discard_wait_queue);
2306 SM_I(sbi)->dcc_info = dcc;
2307init_thread:
2308 err = f2fs_start_discard_thread(sbi);
2309 if (err) {
2310 kfree(dcc);
2311 SM_I(sbi)->dcc_info = NULL;
2312 }
2313
2314 return err;
2315}
2316
2317static void destroy_discard_cmd_control(struct f2fs_sb_info *sbi)
2318{
2319 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
2320
2321 if (!dcc)
2322 return;
2323
2324 f2fs_stop_discard_thread(sbi);
2325
2326 /*
2327 * Recovery can cache discard commands, so in error path of
2328 * fill_super(), it needs to give a chance to handle them.
2329 */
2330 f2fs_issue_discard_timeout(sbi);
2331
2332 kfree(dcc);
2333 SM_I(sbi)->dcc_info = NULL;
2334}
2335
2336static bool __mark_sit_entry_dirty(struct f2fs_sb_info *sbi, unsigned int segno)
2337{
2338 struct sit_info *sit_i = SIT_I(sbi);
2339
2340 if (!__test_and_set_bit(segno, sit_i->dirty_sentries_bitmap)) {
2341 sit_i->dirty_sentries++;
2342 return false;
2343 }
2344
2345 return true;
2346}
2347
2348static void __set_sit_entry_type(struct f2fs_sb_info *sbi, int type,
2349 unsigned int segno, int modified)
2350{
2351 struct seg_entry *se = get_seg_entry(sbi, segno);
2352
2353 se->type = type;
2354 if (modified)
2355 __mark_sit_entry_dirty(sbi, segno);
2356}
2357
2358static inline unsigned long long get_segment_mtime(struct f2fs_sb_info *sbi,
2359 block_t blkaddr)
2360{
2361 unsigned int segno = GET_SEGNO(sbi, blkaddr);
2362
2363 if (segno == NULL_SEGNO)
2364 return 0;
2365 return get_seg_entry(sbi, segno)->mtime;
2366}
2367
2368static void update_segment_mtime(struct f2fs_sb_info *sbi, block_t blkaddr,
2369 unsigned long long old_mtime)
2370{
2371 struct seg_entry *se;
2372 unsigned int segno = GET_SEGNO(sbi, blkaddr);
2373 unsigned long long ctime = get_mtime(sbi, false);
2374 unsigned long long mtime = old_mtime ? old_mtime : ctime;
2375
2376 if (segno == NULL_SEGNO)
2377 return;
2378
2379 se = get_seg_entry(sbi, segno);
2380
2381 if (!se->mtime)
2382 se->mtime = mtime;
2383 else
2384 se->mtime = div_u64(se->mtime * se->valid_blocks + mtime,
2385 se->valid_blocks + 1);
2386
2387 if (ctime > SIT_I(sbi)->max_mtime)
2388 SIT_I(sbi)->max_mtime = ctime;
2389}
2390
2391static void update_sit_entry(struct f2fs_sb_info *sbi, block_t blkaddr, int del)
2392{
2393 struct seg_entry *se;
2394 unsigned int segno, offset;
2395 long int new_vblocks;
2396 bool exist;
2397#ifdef CONFIG_F2FS_CHECK_FS
2398 bool mir_exist;
2399#endif
2400
2401 segno = GET_SEGNO(sbi, blkaddr);
2402
2403 se = get_seg_entry(sbi, segno);
2404 new_vblocks = se->valid_blocks + del;
2405 offset = GET_BLKOFF_FROM_SEG0(sbi, blkaddr);
2406
2407 f2fs_bug_on(sbi, (new_vblocks < 0 ||
2408 (new_vblocks > f2fs_usable_blks_in_seg(sbi, segno))));
2409
2410 se->valid_blocks = new_vblocks;
2411
2412 /* Update valid block bitmap */
2413 if (del > 0) {
2414 exist = f2fs_test_and_set_bit(offset, se->cur_valid_map);
2415#ifdef CONFIG_F2FS_CHECK_FS
2416 mir_exist = f2fs_test_and_set_bit(offset,
2417 se->cur_valid_map_mir);
2418 if (unlikely(exist != mir_exist)) {
2419 f2fs_err(sbi, "Inconsistent error when setting bitmap, blk:%u, old bit:%d",
2420 blkaddr, exist);
2421 f2fs_bug_on(sbi, 1);
2422 }
2423#endif
2424 if (unlikely(exist)) {
2425 f2fs_err(sbi, "Bitmap was wrongly set, blk:%u",
2426 blkaddr);
2427 f2fs_bug_on(sbi, 1);
2428 se->valid_blocks--;
2429 del = 0;
2430 }
2431
2432 if (f2fs_block_unit_discard(sbi) &&
2433 !f2fs_test_and_set_bit(offset, se->discard_map))
2434 sbi->discard_blks--;
2435
2436 /*
2437 * SSR should never reuse block which is checkpointed
2438 * or newly invalidated.
2439 */
2440 if (!is_sbi_flag_set(sbi, SBI_CP_DISABLED)) {
2441 if (!f2fs_test_and_set_bit(offset, se->ckpt_valid_map))
2442 se->ckpt_valid_blocks++;
2443 }
2444 } else {
2445 exist = f2fs_test_and_clear_bit(offset, se->cur_valid_map);
2446#ifdef CONFIG_F2FS_CHECK_FS
2447 mir_exist = f2fs_test_and_clear_bit(offset,
2448 se->cur_valid_map_mir);
2449 if (unlikely(exist != mir_exist)) {
2450 f2fs_err(sbi, "Inconsistent error when clearing bitmap, blk:%u, old bit:%d",
2451 blkaddr, exist);
2452 f2fs_bug_on(sbi, 1);
2453 }
2454#endif
2455 if (unlikely(!exist)) {
2456 f2fs_err(sbi, "Bitmap was wrongly cleared, blk:%u",
2457 blkaddr);
2458 f2fs_bug_on(sbi, 1);
2459 se->valid_blocks++;
2460 del = 0;
2461 } else if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED))) {
2462 /*
2463 * If checkpoints are off, we must not reuse data that
2464 * was used in the previous checkpoint. If it was used
2465 * before, we must track that to know how much space we
2466 * really have.
2467 */
2468 if (f2fs_test_bit(offset, se->ckpt_valid_map)) {
2469 spin_lock(&sbi->stat_lock);
2470 sbi->unusable_block_count++;
2471 spin_unlock(&sbi->stat_lock);
2472 }
2473 }
2474
2475 if (f2fs_block_unit_discard(sbi) &&
2476 f2fs_test_and_clear_bit(offset, se->discard_map))
2477 sbi->discard_blks++;
2478 }
2479 if (!f2fs_test_bit(offset, se->ckpt_valid_map))
2480 se->ckpt_valid_blocks += del;
2481
2482 __mark_sit_entry_dirty(sbi, segno);
2483
2484 /* update total number of valid blocks to be written in ckpt area */
2485 SIT_I(sbi)->written_valid_blocks += del;
2486
2487 if (__is_large_section(sbi))
2488 get_sec_entry(sbi, segno)->valid_blocks += del;
2489}
2490
2491void f2fs_invalidate_blocks(struct f2fs_sb_info *sbi, block_t addr)
2492{
2493 unsigned int segno = GET_SEGNO(sbi, addr);
2494 struct sit_info *sit_i = SIT_I(sbi);
2495
2496 f2fs_bug_on(sbi, addr == NULL_ADDR);
2497 if (addr == NEW_ADDR || addr == COMPRESS_ADDR)
2498 return;
2499
2500 f2fs_invalidate_internal_cache(sbi, addr);
2501
2502 /* add it into sit main buffer */
2503 down_write(&sit_i->sentry_lock);
2504
2505 update_segment_mtime(sbi, addr, 0);
2506 update_sit_entry(sbi, addr, -1);
2507
2508 /* add it into dirty seglist */
2509 locate_dirty_segment(sbi, segno);
2510
2511 up_write(&sit_i->sentry_lock);
2512}
2513
2514bool f2fs_is_checkpointed_data(struct f2fs_sb_info *sbi, block_t blkaddr)
2515{
2516 struct sit_info *sit_i = SIT_I(sbi);
2517 unsigned int segno, offset;
2518 struct seg_entry *se;
2519 bool is_cp = false;
2520
2521 if (!__is_valid_data_blkaddr(blkaddr))
2522 return true;
2523
2524 down_read(&sit_i->sentry_lock);
2525
2526 segno = GET_SEGNO(sbi, blkaddr);
2527 se = get_seg_entry(sbi, segno);
2528 offset = GET_BLKOFF_FROM_SEG0(sbi, blkaddr);
2529
2530 if (f2fs_test_bit(offset, se->ckpt_valid_map))
2531 is_cp = true;
2532
2533 up_read(&sit_i->sentry_lock);
2534
2535 return is_cp;
2536}
2537
2538static unsigned short f2fs_curseg_valid_blocks(struct f2fs_sb_info *sbi, int type)
2539{
2540 struct curseg_info *curseg = CURSEG_I(sbi, type);
2541
2542 if (sbi->ckpt->alloc_type[type] == SSR)
2543 return sbi->blocks_per_seg;
2544 return curseg->next_blkoff;
2545}
2546
2547/*
2548 * Calculate the number of current summary pages for writing
2549 */
2550int f2fs_npages_for_summary_flush(struct f2fs_sb_info *sbi, bool for_ra)
2551{
2552 int valid_sum_count = 0;
2553 int i, sum_in_page;
2554
2555 for (i = CURSEG_HOT_DATA; i <= CURSEG_COLD_DATA; i++) {
2556 if (sbi->ckpt->alloc_type[i] != SSR && for_ra)
2557 valid_sum_count +=
2558 le16_to_cpu(F2FS_CKPT(sbi)->cur_data_blkoff[i]);
2559 else
2560 valid_sum_count += f2fs_curseg_valid_blocks(sbi, i);
2561 }
2562
2563 sum_in_page = (PAGE_SIZE - 2 * SUM_JOURNAL_SIZE -
2564 SUM_FOOTER_SIZE) / SUMMARY_SIZE;
2565 if (valid_sum_count <= sum_in_page)
2566 return 1;
2567 else if ((valid_sum_count - sum_in_page) <=
2568 (PAGE_SIZE - SUM_FOOTER_SIZE) / SUMMARY_SIZE)
2569 return 2;
2570 return 3;
2571}
2572
2573/*
2574 * Caller should put this summary page
2575 */
2576struct page *f2fs_get_sum_page(struct f2fs_sb_info *sbi, unsigned int segno)
2577{
2578 if (unlikely(f2fs_cp_error(sbi)))
2579 return ERR_PTR(-EIO);
2580 return f2fs_get_meta_page_retry(sbi, GET_SUM_BLOCK(sbi, segno));
2581}
2582
2583void f2fs_update_meta_page(struct f2fs_sb_info *sbi,
2584 void *src, block_t blk_addr)
2585{
2586 struct page *page = f2fs_grab_meta_page(sbi, blk_addr);
2587
2588 memcpy(page_address(page), src, PAGE_SIZE);
2589 set_page_dirty(page);
2590 f2fs_put_page(page, 1);
2591}
2592
2593static void write_sum_page(struct f2fs_sb_info *sbi,
2594 struct f2fs_summary_block *sum_blk, block_t blk_addr)
2595{
2596 f2fs_update_meta_page(sbi, (void *)sum_blk, blk_addr);
2597}
2598
2599static void write_current_sum_page(struct f2fs_sb_info *sbi,
2600 int type, block_t blk_addr)
2601{
2602 struct curseg_info *curseg = CURSEG_I(sbi, type);
2603 struct page *page = f2fs_grab_meta_page(sbi, blk_addr);
2604 struct f2fs_summary_block *src = curseg->sum_blk;
2605 struct f2fs_summary_block *dst;
2606
2607 dst = (struct f2fs_summary_block *)page_address(page);
2608 memset(dst, 0, PAGE_SIZE);
2609
2610 mutex_lock(&curseg->curseg_mutex);
2611
2612 down_read(&curseg->journal_rwsem);
2613 memcpy(&dst->journal, curseg->journal, SUM_JOURNAL_SIZE);
2614 up_read(&curseg->journal_rwsem);
2615
2616 memcpy(dst->entries, src->entries, SUM_ENTRY_SIZE);
2617 memcpy(&dst->footer, &src->footer, SUM_FOOTER_SIZE);
2618
2619 mutex_unlock(&curseg->curseg_mutex);
2620
2621 set_page_dirty(page);
2622 f2fs_put_page(page, 1);
2623}
2624
2625static int is_next_segment_free(struct f2fs_sb_info *sbi,
2626 struct curseg_info *curseg, int type)
2627{
2628 unsigned int segno = curseg->segno + 1;
2629 struct free_segmap_info *free_i = FREE_I(sbi);
2630
2631 if (segno < MAIN_SEGS(sbi) && segno % sbi->segs_per_sec)
2632 return !test_bit(segno, free_i->free_segmap);
2633 return 0;
2634}
2635
2636/*
2637 * Find a new segment from the free segments bitmap to right order
2638 * This function should be returned with success, otherwise BUG
2639 */
2640static void get_new_segment(struct f2fs_sb_info *sbi,
2641 unsigned int *newseg, bool new_sec, int dir)
2642{
2643 struct free_segmap_info *free_i = FREE_I(sbi);
2644 unsigned int segno, secno, zoneno;
2645 unsigned int total_zones = MAIN_SECS(sbi) / sbi->secs_per_zone;
2646 unsigned int hint = GET_SEC_FROM_SEG(sbi, *newseg);
2647 unsigned int old_zoneno = GET_ZONE_FROM_SEG(sbi, *newseg);
2648 unsigned int left_start = hint;
2649 bool init = true;
2650 int go_left = 0;
2651 int i;
2652
2653 spin_lock(&free_i->segmap_lock);
2654
2655 if (!new_sec && ((*newseg + 1) % sbi->segs_per_sec)) {
2656 segno = find_next_zero_bit(free_i->free_segmap,
2657 GET_SEG_FROM_SEC(sbi, hint + 1), *newseg + 1);
2658 if (segno < GET_SEG_FROM_SEC(sbi, hint + 1))
2659 goto got_it;
2660 }
2661find_other_zone:
2662 secno = find_next_zero_bit(free_i->free_secmap, MAIN_SECS(sbi), hint);
2663 if (secno >= MAIN_SECS(sbi)) {
2664 if (dir == ALLOC_RIGHT) {
2665 secno = find_first_zero_bit(free_i->free_secmap,
2666 MAIN_SECS(sbi));
2667 f2fs_bug_on(sbi, secno >= MAIN_SECS(sbi));
2668 } else {
2669 go_left = 1;
2670 left_start = hint - 1;
2671 }
2672 }
2673 if (go_left == 0)
2674 goto skip_left;
2675
2676 while (test_bit(left_start, free_i->free_secmap)) {
2677 if (left_start > 0) {
2678 left_start--;
2679 continue;
2680 }
2681 left_start = find_first_zero_bit(free_i->free_secmap,
2682 MAIN_SECS(sbi));
2683 f2fs_bug_on(sbi, left_start >= MAIN_SECS(sbi));
2684 break;
2685 }
2686 secno = left_start;
2687skip_left:
2688 segno = GET_SEG_FROM_SEC(sbi, secno);
2689 zoneno = GET_ZONE_FROM_SEC(sbi, secno);
2690
2691 /* give up on finding another zone */
2692 if (!init)
2693 goto got_it;
2694 if (sbi->secs_per_zone == 1)
2695 goto got_it;
2696 if (zoneno == old_zoneno)
2697 goto got_it;
2698 if (dir == ALLOC_LEFT) {
2699 if (!go_left && zoneno + 1 >= total_zones)
2700 goto got_it;
2701 if (go_left && zoneno == 0)
2702 goto got_it;
2703 }
2704 for (i = 0; i < NR_CURSEG_TYPE; i++)
2705 if (CURSEG_I(sbi, i)->zone == zoneno)
2706 break;
2707
2708 if (i < NR_CURSEG_TYPE) {
2709 /* zone is in user, try another */
2710 if (go_left)
2711 hint = zoneno * sbi->secs_per_zone - 1;
2712 else if (zoneno + 1 >= total_zones)
2713 hint = 0;
2714 else
2715 hint = (zoneno + 1) * sbi->secs_per_zone;
2716 init = false;
2717 goto find_other_zone;
2718 }
2719got_it:
2720 /* set it as dirty segment in free segmap */
2721 f2fs_bug_on(sbi, test_bit(segno, free_i->free_segmap));
2722 __set_inuse(sbi, segno);
2723 *newseg = segno;
2724 spin_unlock(&free_i->segmap_lock);
2725}
2726
2727static void reset_curseg(struct f2fs_sb_info *sbi, int type, int modified)
2728{
2729 struct curseg_info *curseg = CURSEG_I(sbi, type);
2730 struct summary_footer *sum_footer;
2731 unsigned short seg_type = curseg->seg_type;
2732
2733 curseg->inited = true;
2734 curseg->segno = curseg->next_segno;
2735 curseg->zone = GET_ZONE_FROM_SEG(sbi, curseg->segno);
2736 curseg->next_blkoff = 0;
2737 curseg->next_segno = NULL_SEGNO;
2738
2739 sum_footer = &(curseg->sum_blk->footer);
2740 memset(sum_footer, 0, sizeof(struct summary_footer));
2741
2742 sanity_check_seg_type(sbi, seg_type);
2743
2744 if (IS_DATASEG(seg_type))
2745 SET_SUM_TYPE(sum_footer, SUM_TYPE_DATA);
2746 if (IS_NODESEG(seg_type))
2747 SET_SUM_TYPE(sum_footer, SUM_TYPE_NODE);
2748 __set_sit_entry_type(sbi, seg_type, curseg->segno, modified);
2749}
2750
2751static unsigned int __get_next_segno(struct f2fs_sb_info *sbi, int type)
2752{
2753 struct curseg_info *curseg = CURSEG_I(sbi, type);
2754 unsigned short seg_type = curseg->seg_type;
2755
2756 sanity_check_seg_type(sbi, seg_type);
2757 if (f2fs_need_rand_seg(sbi))
2758 return get_random_u32_below(MAIN_SECS(sbi) * sbi->segs_per_sec);
2759
2760 /* if segs_per_sec is large than 1, we need to keep original policy. */
2761 if (__is_large_section(sbi))
2762 return curseg->segno;
2763
2764 /* inmem log may not locate on any segment after mount */
2765 if (!curseg->inited)
2766 return 0;
2767
2768 if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED)))
2769 return 0;
2770
2771 if (test_opt(sbi, NOHEAP) &&
2772 (seg_type == CURSEG_HOT_DATA || IS_NODESEG(seg_type)))
2773 return 0;
2774
2775 if (SIT_I(sbi)->last_victim[ALLOC_NEXT])
2776 return SIT_I(sbi)->last_victim[ALLOC_NEXT];
2777
2778 /* find segments from 0 to reuse freed segments */
2779 if (F2FS_OPTION(sbi).alloc_mode == ALLOC_MODE_REUSE)
2780 return 0;
2781
2782 return curseg->segno;
2783}
2784
2785/*
2786 * Allocate a current working segment.
2787 * This function always allocates a free segment in LFS manner.
2788 */
2789static void new_curseg(struct f2fs_sb_info *sbi, int type, bool new_sec)
2790{
2791 struct curseg_info *curseg = CURSEG_I(sbi, type);
2792 unsigned short seg_type = curseg->seg_type;
2793 unsigned int segno = curseg->segno;
2794 int dir = ALLOC_LEFT;
2795
2796 if (curseg->inited)
2797 write_sum_page(sbi, curseg->sum_blk,
2798 GET_SUM_BLOCK(sbi, segno));
2799 if (seg_type == CURSEG_WARM_DATA || seg_type == CURSEG_COLD_DATA)
2800 dir = ALLOC_RIGHT;
2801
2802 if (test_opt(sbi, NOHEAP))
2803 dir = ALLOC_RIGHT;
2804
2805 segno = __get_next_segno(sbi, type);
2806 get_new_segment(sbi, &segno, new_sec, dir);
2807 curseg->next_segno = segno;
2808 reset_curseg(sbi, type, 1);
2809 curseg->alloc_type = LFS;
2810 if (F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_BLK)
2811 curseg->fragment_remained_chunk =
2812 get_random_u32_inclusive(1, sbi->max_fragment_chunk);
2813}
2814
2815static int __next_free_blkoff(struct f2fs_sb_info *sbi,
2816 int segno, block_t start)
2817{
2818 struct seg_entry *se = get_seg_entry(sbi, segno);
2819 int entries = SIT_VBLOCK_MAP_SIZE / sizeof(unsigned long);
2820 unsigned long *target_map = SIT_I(sbi)->tmp_map;
2821 unsigned long *ckpt_map = (unsigned long *)se->ckpt_valid_map;
2822 unsigned long *cur_map = (unsigned long *)se->cur_valid_map;
2823 int i;
2824
2825 for (i = 0; i < entries; i++)
2826 target_map[i] = ckpt_map[i] | cur_map[i];
2827
2828 return __find_rev_next_zero_bit(target_map, sbi->blocks_per_seg, start);
2829}
2830
2831static int f2fs_find_next_ssr_block(struct f2fs_sb_info *sbi,
2832 struct curseg_info *seg)
2833{
2834 return __next_free_blkoff(sbi, seg->segno, seg->next_blkoff + 1);
2835}
2836
2837bool f2fs_segment_has_free_slot(struct f2fs_sb_info *sbi, int segno)
2838{
2839 return __next_free_blkoff(sbi, segno, 0) < sbi->blocks_per_seg;
2840}
2841
2842/*
2843 * This function always allocates a used segment(from dirty seglist) by SSR
2844 * manner, so it should recover the existing segment information of valid blocks
2845 */
2846static void change_curseg(struct f2fs_sb_info *sbi, int type)
2847{
2848 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
2849 struct curseg_info *curseg = CURSEG_I(sbi, type);
2850 unsigned int new_segno = curseg->next_segno;
2851 struct f2fs_summary_block *sum_node;
2852 struct page *sum_page;
2853
2854 write_sum_page(sbi, curseg->sum_blk, GET_SUM_BLOCK(sbi, curseg->segno));
2855
2856 __set_test_and_inuse(sbi, new_segno);
2857
2858 mutex_lock(&dirty_i->seglist_lock);
2859 __remove_dirty_segment(sbi, new_segno, PRE);
2860 __remove_dirty_segment(sbi, new_segno, DIRTY);
2861 mutex_unlock(&dirty_i->seglist_lock);
2862
2863 reset_curseg(sbi, type, 1);
2864 curseg->alloc_type = SSR;
2865 curseg->next_blkoff = __next_free_blkoff(sbi, curseg->segno, 0);
2866
2867 sum_page = f2fs_get_sum_page(sbi, new_segno);
2868 if (IS_ERR(sum_page)) {
2869 /* GC won't be able to use stale summary pages by cp_error */
2870 memset(curseg->sum_blk, 0, SUM_ENTRY_SIZE);
2871 return;
2872 }
2873 sum_node = (struct f2fs_summary_block *)page_address(sum_page);
2874 memcpy(curseg->sum_blk, sum_node, SUM_ENTRY_SIZE);
2875 f2fs_put_page(sum_page, 1);
2876}
2877
2878static int get_ssr_segment(struct f2fs_sb_info *sbi, int type,
2879 int alloc_mode, unsigned long long age);
2880
2881static void get_atssr_segment(struct f2fs_sb_info *sbi, int type,
2882 int target_type, int alloc_mode,
2883 unsigned long long age)
2884{
2885 struct curseg_info *curseg = CURSEG_I(sbi, type);
2886
2887 curseg->seg_type = target_type;
2888
2889 if (get_ssr_segment(sbi, type, alloc_mode, age)) {
2890 struct seg_entry *se = get_seg_entry(sbi, curseg->next_segno);
2891
2892 curseg->seg_type = se->type;
2893 change_curseg(sbi, type);
2894 } else {
2895 /* allocate cold segment by default */
2896 curseg->seg_type = CURSEG_COLD_DATA;
2897 new_curseg(sbi, type, true);
2898 }
2899 stat_inc_seg_type(sbi, curseg);
2900}
2901
2902static void __f2fs_init_atgc_curseg(struct f2fs_sb_info *sbi)
2903{
2904 struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_ALL_DATA_ATGC);
2905
2906 if (!sbi->am.atgc_enabled)
2907 return;
2908
2909 f2fs_down_read(&SM_I(sbi)->curseg_lock);
2910
2911 mutex_lock(&curseg->curseg_mutex);
2912 down_write(&SIT_I(sbi)->sentry_lock);
2913
2914 get_atssr_segment(sbi, CURSEG_ALL_DATA_ATGC, CURSEG_COLD_DATA, SSR, 0);
2915
2916 up_write(&SIT_I(sbi)->sentry_lock);
2917 mutex_unlock(&curseg->curseg_mutex);
2918
2919 f2fs_up_read(&SM_I(sbi)->curseg_lock);
2920
2921}
2922void f2fs_init_inmem_curseg(struct f2fs_sb_info *sbi)
2923{
2924 __f2fs_init_atgc_curseg(sbi);
2925}
2926
2927static void __f2fs_save_inmem_curseg(struct f2fs_sb_info *sbi, int type)
2928{
2929 struct curseg_info *curseg = CURSEG_I(sbi, type);
2930
2931 mutex_lock(&curseg->curseg_mutex);
2932 if (!curseg->inited)
2933 goto out;
2934
2935 if (get_valid_blocks(sbi, curseg->segno, false)) {
2936 write_sum_page(sbi, curseg->sum_blk,
2937 GET_SUM_BLOCK(sbi, curseg->segno));
2938 } else {
2939 mutex_lock(&DIRTY_I(sbi)->seglist_lock);
2940 __set_test_and_free(sbi, curseg->segno, true);
2941 mutex_unlock(&DIRTY_I(sbi)->seglist_lock);
2942 }
2943out:
2944 mutex_unlock(&curseg->curseg_mutex);
2945}
2946
2947void f2fs_save_inmem_curseg(struct f2fs_sb_info *sbi)
2948{
2949 __f2fs_save_inmem_curseg(sbi, CURSEG_COLD_DATA_PINNED);
2950
2951 if (sbi->am.atgc_enabled)
2952 __f2fs_save_inmem_curseg(sbi, CURSEG_ALL_DATA_ATGC);
2953}
2954
2955static void __f2fs_restore_inmem_curseg(struct f2fs_sb_info *sbi, int type)
2956{
2957 struct curseg_info *curseg = CURSEG_I(sbi, type);
2958
2959 mutex_lock(&curseg->curseg_mutex);
2960 if (!curseg->inited)
2961 goto out;
2962 if (get_valid_blocks(sbi, curseg->segno, false))
2963 goto out;
2964
2965 mutex_lock(&DIRTY_I(sbi)->seglist_lock);
2966 __set_test_and_inuse(sbi, curseg->segno);
2967 mutex_unlock(&DIRTY_I(sbi)->seglist_lock);
2968out:
2969 mutex_unlock(&curseg->curseg_mutex);
2970}
2971
2972void f2fs_restore_inmem_curseg(struct f2fs_sb_info *sbi)
2973{
2974 __f2fs_restore_inmem_curseg(sbi, CURSEG_COLD_DATA_PINNED);
2975
2976 if (sbi->am.atgc_enabled)
2977 __f2fs_restore_inmem_curseg(sbi, CURSEG_ALL_DATA_ATGC);
2978}
2979
2980static int get_ssr_segment(struct f2fs_sb_info *sbi, int type,
2981 int alloc_mode, unsigned long long age)
2982{
2983 struct curseg_info *curseg = CURSEG_I(sbi, type);
2984 unsigned segno = NULL_SEGNO;
2985 unsigned short seg_type = curseg->seg_type;
2986 int i, cnt;
2987 bool reversed = false;
2988
2989 sanity_check_seg_type(sbi, seg_type);
2990
2991 /* f2fs_need_SSR() already forces to do this */
2992 if (!f2fs_get_victim(sbi, &segno, BG_GC, seg_type, alloc_mode, age)) {
2993 curseg->next_segno = segno;
2994 return 1;
2995 }
2996
2997 /* For node segments, let's do SSR more intensively */
2998 if (IS_NODESEG(seg_type)) {
2999 if (seg_type >= CURSEG_WARM_NODE) {
3000 reversed = true;
3001 i = CURSEG_COLD_NODE;
3002 } else {
3003 i = CURSEG_HOT_NODE;
3004 }
3005 cnt = NR_CURSEG_NODE_TYPE;
3006 } else {
3007 if (seg_type >= CURSEG_WARM_DATA) {
3008 reversed = true;
3009 i = CURSEG_COLD_DATA;
3010 } else {
3011 i = CURSEG_HOT_DATA;
3012 }
3013 cnt = NR_CURSEG_DATA_TYPE;
3014 }
3015
3016 for (; cnt-- > 0; reversed ? i-- : i++) {
3017 if (i == seg_type)
3018 continue;
3019 if (!f2fs_get_victim(sbi, &segno, BG_GC, i, alloc_mode, age)) {
3020 curseg->next_segno = segno;
3021 return 1;
3022 }
3023 }
3024
3025 /* find valid_blocks=0 in dirty list */
3026 if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED))) {
3027 segno = get_free_segment(sbi);
3028 if (segno != NULL_SEGNO) {
3029 curseg->next_segno = segno;
3030 return 1;
3031 }
3032 }
3033 return 0;
3034}
3035
3036static bool need_new_seg(struct f2fs_sb_info *sbi, int type)
3037{
3038 struct curseg_info *curseg = CURSEG_I(sbi, type);
3039
3040 if (!is_set_ckpt_flags(sbi, CP_CRC_RECOVERY_FLAG) &&
3041 curseg->seg_type == CURSEG_WARM_NODE)
3042 return true;
3043 if (curseg->alloc_type == LFS &&
3044 is_next_segment_free(sbi, curseg, type) &&
3045 likely(!is_sbi_flag_set(sbi, SBI_CP_DISABLED)))
3046 return true;
3047 if (!f2fs_need_SSR(sbi) || !get_ssr_segment(sbi, type, SSR, 0))
3048 return true;
3049 return false;
3050}
3051
3052void f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type,
3053 unsigned int start, unsigned int end)
3054{
3055 struct curseg_info *curseg = CURSEG_I(sbi, type);
3056 unsigned int segno;
3057
3058 f2fs_down_read(&SM_I(sbi)->curseg_lock);
3059 mutex_lock(&curseg->curseg_mutex);
3060 down_write(&SIT_I(sbi)->sentry_lock);
3061
3062 segno = CURSEG_I(sbi, type)->segno;
3063 if (segno < start || segno > end)
3064 goto unlock;
3065
3066 if (f2fs_need_SSR(sbi) && get_ssr_segment(sbi, type, SSR, 0))
3067 change_curseg(sbi, type);
3068 else
3069 new_curseg(sbi, type, true);
3070
3071 stat_inc_seg_type(sbi, curseg);
3072
3073 locate_dirty_segment(sbi, segno);
3074unlock:
3075 up_write(&SIT_I(sbi)->sentry_lock);
3076
3077 if (segno != curseg->segno)
3078 f2fs_notice(sbi, "For resize: curseg of type %d: %u ==> %u",
3079 type, segno, curseg->segno);
3080
3081 mutex_unlock(&curseg->curseg_mutex);
3082 f2fs_up_read(&SM_I(sbi)->curseg_lock);
3083}
3084
3085static void __allocate_new_segment(struct f2fs_sb_info *sbi, int type,
3086 bool new_sec, bool force)
3087{
3088 struct curseg_info *curseg = CURSEG_I(sbi, type);
3089 unsigned int old_segno;
3090
3091 if (!force && curseg->inited &&
3092 !curseg->next_blkoff &&
3093 !get_valid_blocks(sbi, curseg->segno, new_sec) &&
3094 !get_ckpt_valid_blocks(sbi, curseg->segno, new_sec))
3095 return;
3096
3097 old_segno = curseg->segno;
3098 new_curseg(sbi, type, true);
3099 stat_inc_seg_type(sbi, curseg);
3100 locate_dirty_segment(sbi, old_segno);
3101}
3102
3103void f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force)
3104{
3105 f2fs_down_read(&SM_I(sbi)->curseg_lock);
3106 down_write(&SIT_I(sbi)->sentry_lock);
3107 __allocate_new_segment(sbi, type, true, force);
3108 up_write(&SIT_I(sbi)->sentry_lock);
3109 f2fs_up_read(&SM_I(sbi)->curseg_lock);
3110}
3111
3112void f2fs_allocate_new_segments(struct f2fs_sb_info *sbi)
3113{
3114 int i;
3115
3116 f2fs_down_read(&SM_I(sbi)->curseg_lock);
3117 down_write(&SIT_I(sbi)->sentry_lock);
3118 for (i = CURSEG_HOT_DATA; i <= CURSEG_COLD_DATA; i++)
3119 __allocate_new_segment(sbi, i, false, false);
3120 up_write(&SIT_I(sbi)->sentry_lock);
3121 f2fs_up_read(&SM_I(sbi)->curseg_lock);
3122}
3123
3124bool f2fs_exist_trim_candidates(struct f2fs_sb_info *sbi,
3125 struct cp_control *cpc)
3126{
3127 __u64 trim_start = cpc->trim_start;
3128 bool has_candidate = false;
3129
3130 down_write(&SIT_I(sbi)->sentry_lock);
3131 for (; cpc->trim_start <= cpc->trim_end; cpc->trim_start++) {
3132 if (add_discard_addrs(sbi, cpc, true)) {
3133 has_candidate = true;
3134 break;
3135 }
3136 }
3137 up_write(&SIT_I(sbi)->sentry_lock);
3138
3139 cpc->trim_start = trim_start;
3140 return has_candidate;
3141}
3142
3143static unsigned int __issue_discard_cmd_range(struct f2fs_sb_info *sbi,
3144 struct discard_policy *dpolicy,
3145 unsigned int start, unsigned int end)
3146{
3147 struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
3148 struct discard_cmd *prev_dc = NULL, *next_dc = NULL;
3149 struct rb_node **insert_p = NULL, *insert_parent = NULL;
3150 struct discard_cmd *dc;
3151 struct blk_plug plug;
3152 int issued;
3153 unsigned int trimmed = 0;
3154
3155next:
3156 issued = 0;
3157
3158 mutex_lock(&dcc->cmd_lock);
3159 if (unlikely(dcc->rbtree_check))
3160 f2fs_bug_on(sbi, !f2fs_check_discard_tree(sbi));
3161
3162 dc = __lookup_discard_cmd_ret(&dcc->root, start,
3163 &prev_dc, &next_dc, &insert_p, &insert_parent);
3164 if (!dc)
3165 dc = next_dc;
3166
3167 blk_start_plug(&plug);
3168
3169 while (dc && dc->di.lstart <= end) {
3170 struct rb_node *node;
3171 int err = 0;
3172
3173 if (dc->di.len < dpolicy->granularity)
3174 goto skip;
3175
3176 if (dc->state != D_PREP) {
3177 list_move_tail(&dc->list, &dcc->fstrim_list);
3178 goto skip;
3179 }
3180
3181 err = __submit_discard_cmd(sbi, dpolicy, dc, &issued);
3182
3183 if (issued >= dpolicy->max_requests) {
3184 start = dc->di.lstart + dc->di.len;
3185
3186 if (err)
3187 __remove_discard_cmd(sbi, dc);
3188
3189 blk_finish_plug(&plug);
3190 mutex_unlock(&dcc->cmd_lock);
3191 trimmed += __wait_all_discard_cmd(sbi, NULL);
3192 f2fs_io_schedule_timeout(DEFAULT_IO_TIMEOUT);
3193 goto next;
3194 }
3195skip:
3196 node = rb_next(&dc->rb_node);
3197 if (err)
3198 __remove_discard_cmd(sbi, dc);
3199 dc = rb_entry_safe(node, struct discard_cmd, rb_node);
3200
3201 if (fatal_signal_pending(current))
3202 break;
3203 }
3204
3205 blk_finish_plug(&plug);
3206 mutex_unlock(&dcc->cmd_lock);
3207
3208 return trimmed;
3209}
3210
3211int f2fs_trim_fs(struct f2fs_sb_info *sbi, struct fstrim_range *range)
3212{
3213 __u64 start = F2FS_BYTES_TO_BLK(range->start);
3214 __u64 end = start + F2FS_BYTES_TO_BLK(range->len) - 1;
3215 unsigned int start_segno, end_segno;
3216 block_t start_block, end_block;
3217 struct cp_control cpc;
3218 struct discard_policy dpolicy;
3219 unsigned long long trimmed = 0;
3220 int err = 0;
3221 bool need_align = f2fs_lfs_mode(sbi) && __is_large_section(sbi);
3222
3223 if (start >= MAX_BLKADDR(sbi) || range->len < sbi->blocksize)
3224 return -EINVAL;
3225
3226 if (end < MAIN_BLKADDR(sbi))
3227 goto out;
3228
3229 if (is_sbi_flag_set(sbi, SBI_NEED_FSCK)) {
3230 f2fs_warn(sbi, "Found FS corruption, run fsck to fix.");
3231 return -EFSCORRUPTED;
3232 }
3233
3234 /* start/end segment number in main_area */
3235 start_segno = (start <= MAIN_BLKADDR(sbi)) ? 0 : GET_SEGNO(sbi, start);
3236 end_segno = (end >= MAX_BLKADDR(sbi)) ? MAIN_SEGS(sbi) - 1 :
3237 GET_SEGNO(sbi, end);
3238 if (need_align) {
3239 start_segno = rounddown(start_segno, sbi->segs_per_sec);
3240 end_segno = roundup(end_segno + 1, sbi->segs_per_sec) - 1;
3241 }
3242
3243 cpc.reason = CP_DISCARD;
3244 cpc.trim_minlen = max_t(__u64, 1, F2FS_BYTES_TO_BLK(range->minlen));
3245 cpc.trim_start = start_segno;
3246 cpc.trim_end = end_segno;
3247
3248 if (sbi->discard_blks == 0)
3249 goto out;
3250
3251 f2fs_down_write(&sbi->gc_lock);
3252 stat_inc_cp_call_count(sbi, TOTAL_CALL);
3253 err = f2fs_write_checkpoint(sbi, &cpc);
3254 f2fs_up_write(&sbi->gc_lock);
3255 if (err)
3256 goto out;
3257
3258 /*
3259 * We filed discard candidates, but actually we don't need to wait for
3260 * all of them, since they'll be issued in idle time along with runtime
3261 * discard option. User configuration looks like using runtime discard
3262 * or periodic fstrim instead of it.
3263 */
3264 if (f2fs_realtime_discard_enable(sbi))
3265 goto out;
3266
3267 start_block = START_BLOCK(sbi, start_segno);
3268 end_block = START_BLOCK(sbi, end_segno + 1);
3269
3270 __init_discard_policy(sbi, &dpolicy, DPOLICY_FSTRIM, cpc.trim_minlen);
3271 trimmed = __issue_discard_cmd_range(sbi, &dpolicy,
3272 start_block, end_block);
3273
3274 trimmed += __wait_discard_cmd_range(sbi, &dpolicy,
3275 start_block, end_block);
3276out:
3277 if (!err)
3278 range->len = F2FS_BLK_TO_BYTES(trimmed);
3279 return err;
3280}
3281
3282int f2fs_rw_hint_to_seg_type(enum rw_hint hint)
3283{
3284 switch (hint) {
3285 case WRITE_LIFE_SHORT:
3286 return CURSEG_HOT_DATA;
3287 case WRITE_LIFE_EXTREME:
3288 return CURSEG_COLD_DATA;
3289 default:
3290 return CURSEG_WARM_DATA;
3291 }
3292}
3293
3294static int __get_segment_type_2(struct f2fs_io_info *fio)
3295{
3296 if (fio->type == DATA)
3297 return CURSEG_HOT_DATA;
3298 else
3299 return CURSEG_HOT_NODE;
3300}
3301
3302static int __get_segment_type_4(struct f2fs_io_info *fio)
3303{
3304 if (fio->type == DATA) {
3305 struct inode *inode = fio->page->mapping->host;
3306
3307 if (S_ISDIR(inode->i_mode))
3308 return CURSEG_HOT_DATA;
3309 else
3310 return CURSEG_COLD_DATA;
3311 } else {
3312 if (IS_DNODE(fio->page) && is_cold_node(fio->page))
3313 return CURSEG_WARM_NODE;
3314 else
3315 return CURSEG_COLD_NODE;
3316 }
3317}
3318
3319static int __get_age_segment_type(struct inode *inode, pgoff_t pgofs)
3320{
3321 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
3322 struct extent_info ei = {};
3323
3324 if (f2fs_lookup_age_extent_cache(inode, pgofs, &ei)) {
3325 if (!ei.age)
3326 return NO_CHECK_TYPE;
3327 if (ei.age <= sbi->hot_data_age_threshold)
3328 return CURSEG_HOT_DATA;
3329 if (ei.age <= sbi->warm_data_age_threshold)
3330 return CURSEG_WARM_DATA;
3331 return CURSEG_COLD_DATA;
3332 }
3333 return NO_CHECK_TYPE;
3334}
3335
3336static int __get_segment_type_6(struct f2fs_io_info *fio)
3337{
3338 if (fio->type == DATA) {
3339 struct inode *inode = fio->page->mapping->host;
3340 int type;
3341
3342 if (is_inode_flag_set(inode, FI_ALIGNED_WRITE))
3343 return CURSEG_COLD_DATA_PINNED;
3344
3345 if (page_private_gcing(fio->page)) {
3346 if (fio->sbi->am.atgc_enabled &&
3347 (fio->io_type == FS_DATA_IO) &&
3348 (fio->sbi->gc_mode != GC_URGENT_HIGH))
3349 return CURSEG_ALL_DATA_ATGC;
3350 else
3351 return CURSEG_COLD_DATA;
3352 }
3353 if (file_is_cold(inode) || f2fs_need_compress_data(inode))
3354 return CURSEG_COLD_DATA;
3355
3356 type = __get_age_segment_type(inode, fio->page->index);
3357 if (type != NO_CHECK_TYPE)
3358 return type;
3359
3360 if (file_is_hot(inode) ||
3361 is_inode_flag_set(inode, FI_HOT_DATA) ||
3362 f2fs_is_cow_file(inode))
3363 return CURSEG_HOT_DATA;
3364 return f2fs_rw_hint_to_seg_type(inode->i_write_hint);
3365 } else {
3366 if (IS_DNODE(fio->page))
3367 return is_cold_node(fio->page) ? CURSEG_WARM_NODE :
3368 CURSEG_HOT_NODE;
3369 return CURSEG_COLD_NODE;
3370 }
3371}
3372
3373static int __get_segment_type(struct f2fs_io_info *fio)
3374{
3375 int type = 0;
3376
3377 switch (F2FS_OPTION(fio->sbi).active_logs) {
3378 case 2:
3379 type = __get_segment_type_2(fio);
3380 break;
3381 case 4:
3382 type = __get_segment_type_4(fio);
3383 break;
3384 case 6:
3385 type = __get_segment_type_6(fio);
3386 break;
3387 default:
3388 f2fs_bug_on(fio->sbi, true);
3389 }
3390
3391 if (IS_HOT(type))
3392 fio->temp = HOT;
3393 else if (IS_WARM(type))
3394 fio->temp = WARM;
3395 else
3396 fio->temp = COLD;
3397 return type;
3398}
3399
3400static void f2fs_randomize_chunk(struct f2fs_sb_info *sbi,
3401 struct curseg_info *seg)
3402{
3403 /* To allocate block chunks in different sizes, use random number */
3404 if (--seg->fragment_remained_chunk > 0)
3405 return;
3406
3407 seg->fragment_remained_chunk =
3408 get_random_u32_inclusive(1, sbi->max_fragment_chunk);
3409 seg->next_blkoff +=
3410 get_random_u32_inclusive(1, sbi->max_fragment_hole);
3411}
3412
3413void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
3414 block_t old_blkaddr, block_t *new_blkaddr,
3415 struct f2fs_summary *sum, int type,
3416 struct f2fs_io_info *fio)
3417{
3418 struct sit_info *sit_i = SIT_I(sbi);
3419 struct curseg_info *curseg = CURSEG_I(sbi, type);
3420 unsigned long long old_mtime;
3421 bool from_gc = (type == CURSEG_ALL_DATA_ATGC);
3422 struct seg_entry *se = NULL;
3423 bool segment_full = false;
3424
3425 f2fs_down_read(&SM_I(sbi)->curseg_lock);
3426
3427 mutex_lock(&curseg->curseg_mutex);
3428 down_write(&sit_i->sentry_lock);
3429
3430 if (from_gc) {
3431 f2fs_bug_on(sbi, GET_SEGNO(sbi, old_blkaddr) == NULL_SEGNO);
3432 se = get_seg_entry(sbi, GET_SEGNO(sbi, old_blkaddr));
3433 sanity_check_seg_type(sbi, se->type);
3434 f2fs_bug_on(sbi, IS_NODESEG(se->type));
3435 }
3436 *new_blkaddr = NEXT_FREE_BLKADDR(sbi, curseg);
3437
3438 f2fs_bug_on(sbi, curseg->next_blkoff >= sbi->blocks_per_seg);
3439
3440 f2fs_wait_discard_bio(sbi, *new_blkaddr);
3441
3442 curseg->sum_blk->entries[curseg->next_blkoff] = *sum;
3443 if (curseg->alloc_type == SSR) {
3444 curseg->next_blkoff = f2fs_find_next_ssr_block(sbi, curseg);
3445 } else {
3446 curseg->next_blkoff++;
3447 if (F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_BLK)
3448 f2fs_randomize_chunk(sbi, curseg);
3449 }
3450 if (curseg->next_blkoff >= f2fs_usable_blks_in_seg(sbi, curseg->segno))
3451 segment_full = true;
3452 stat_inc_block_count(sbi, curseg);
3453
3454 if (from_gc) {
3455 old_mtime = get_segment_mtime(sbi, old_blkaddr);
3456 } else {
3457 update_segment_mtime(sbi, old_blkaddr, 0);
3458 old_mtime = 0;
3459 }
3460 update_segment_mtime(sbi, *new_blkaddr, old_mtime);
3461
3462 /*
3463 * SIT information should be updated before segment allocation,
3464 * since SSR needs latest valid block information.
3465 */
3466 update_sit_entry(sbi, *new_blkaddr, 1);
3467 if (GET_SEGNO(sbi, old_blkaddr) != NULL_SEGNO)
3468 update_sit_entry(sbi, old_blkaddr, -1);
3469
3470 /*
3471 * If the current segment is full, flush it out and replace it with a
3472 * new segment.
3473 */
3474 if (segment_full) {
3475 if (from_gc) {
3476 get_atssr_segment(sbi, type, se->type,
3477 AT_SSR, se->mtime);
3478 } else {
3479 if (need_new_seg(sbi, type))
3480 new_curseg(sbi, type, false);
3481 else
3482 change_curseg(sbi, type);
3483 stat_inc_seg_type(sbi, curseg);
3484 }
3485 }
3486 /*
3487 * segment dirty status should be updated after segment allocation,
3488 * so we just need to update status only one time after previous
3489 * segment being closed.
3490 */
3491 locate_dirty_segment(sbi, GET_SEGNO(sbi, old_blkaddr));
3492 locate_dirty_segment(sbi, GET_SEGNO(sbi, *new_blkaddr));
3493
3494 if (IS_DATASEG(type))
3495 atomic64_inc(&sbi->allocated_data_blocks);
3496
3497 up_write(&sit_i->sentry_lock);
3498
3499 if (page && IS_NODESEG(type)) {
3500 fill_node_footer_blkaddr(page, NEXT_FREE_BLKADDR(sbi, curseg));
3501
3502 f2fs_inode_chksum_set(sbi, page);
3503 }
3504
3505 if (fio) {
3506 struct f2fs_bio_info *io;
3507
3508 if (F2FS_IO_ALIGNED(sbi))
3509 fio->retry = 0;
3510
3511 INIT_LIST_HEAD(&fio->list);
3512 fio->in_list = 1;
3513 io = sbi->write_io[fio->type] + fio->temp;
3514 spin_lock(&io->io_lock);
3515 list_add_tail(&fio->list, &io->io_list);
3516 spin_unlock(&io->io_lock);
3517 }
3518
3519 mutex_unlock(&curseg->curseg_mutex);
3520
3521 f2fs_up_read(&SM_I(sbi)->curseg_lock);
3522}
3523
3524void f2fs_update_device_state(struct f2fs_sb_info *sbi, nid_t ino,
3525 block_t blkaddr, unsigned int blkcnt)
3526{
3527 if (!f2fs_is_multi_device(sbi))
3528 return;
3529
3530 while (1) {
3531 unsigned int devidx = f2fs_target_device_index(sbi, blkaddr);
3532 unsigned int blks = FDEV(devidx).end_blk - blkaddr + 1;
3533
3534 /* update device state for fsync */
3535 f2fs_set_dirty_device(sbi, ino, devidx, FLUSH_INO);
3536
3537 /* update device state for checkpoint */
3538 if (!f2fs_test_bit(devidx, (char *)&sbi->dirty_device)) {
3539 spin_lock(&sbi->dev_lock);
3540 f2fs_set_bit(devidx, (char *)&sbi->dirty_device);
3541 spin_unlock(&sbi->dev_lock);
3542 }
3543
3544 if (blkcnt <= blks)
3545 break;
3546 blkcnt -= blks;
3547 blkaddr += blks;
3548 }
3549}
3550
3551static void do_write_page(struct f2fs_summary *sum, struct f2fs_io_info *fio)
3552{
3553 int type = __get_segment_type(fio);
3554 bool keep_order = (f2fs_lfs_mode(fio->sbi) && type == CURSEG_COLD_DATA);
3555
3556 if (keep_order)
3557 f2fs_down_read(&fio->sbi->io_order_lock);
3558reallocate:
3559 f2fs_allocate_data_block(fio->sbi, fio->page, fio->old_blkaddr,
3560 &fio->new_blkaddr, sum, type, fio);
3561 if (GET_SEGNO(fio->sbi, fio->old_blkaddr) != NULL_SEGNO)
3562 f2fs_invalidate_internal_cache(fio->sbi, fio->old_blkaddr);
3563
3564 /* writeout dirty page into bdev */
3565 f2fs_submit_page_write(fio);
3566 if (fio->retry) {
3567 fio->old_blkaddr = fio->new_blkaddr;
3568 goto reallocate;
3569 }
3570
3571 f2fs_update_device_state(fio->sbi, fio->ino, fio->new_blkaddr, 1);
3572
3573 if (keep_order)
3574 f2fs_up_read(&fio->sbi->io_order_lock);
3575}
3576
3577void f2fs_do_write_meta_page(struct f2fs_sb_info *sbi, struct page *page,
3578 enum iostat_type io_type)
3579{
3580 struct f2fs_io_info fio = {
3581 .sbi = sbi,
3582 .type = META,
3583 .temp = HOT,
3584 .op = REQ_OP_WRITE,
3585 .op_flags = REQ_SYNC | REQ_META | REQ_PRIO,
3586 .old_blkaddr = page->index,
3587 .new_blkaddr = page->index,
3588 .page = page,
3589 .encrypted_page = NULL,
3590 .in_list = 0,
3591 };
3592
3593 if (unlikely(page->index >= MAIN_BLKADDR(sbi)))
3594 fio.op_flags &= ~REQ_META;
3595
3596 set_page_writeback(page);
3597 f2fs_submit_page_write(&fio);
3598
3599 stat_inc_meta_count(sbi, page->index);
3600 f2fs_update_iostat(sbi, NULL, io_type, F2FS_BLKSIZE);
3601}
3602
3603void f2fs_do_write_node_page(unsigned int nid, struct f2fs_io_info *fio)
3604{
3605 struct f2fs_summary sum;
3606
3607 set_summary(&sum, nid, 0, 0);
3608 do_write_page(&sum, fio);
3609
3610 f2fs_update_iostat(fio->sbi, NULL, fio->io_type, F2FS_BLKSIZE);
3611}
3612
3613void f2fs_outplace_write_data(struct dnode_of_data *dn,
3614 struct f2fs_io_info *fio)
3615{
3616 struct f2fs_sb_info *sbi = fio->sbi;
3617 struct f2fs_summary sum;
3618
3619 f2fs_bug_on(sbi, dn->data_blkaddr == NULL_ADDR);
3620 if (fio->io_type == FS_DATA_IO || fio->io_type == FS_CP_DATA_IO)
3621 f2fs_update_age_extent_cache(dn);
3622 set_summary(&sum, dn->nid, dn->ofs_in_node, fio->version);
3623 do_write_page(&sum, fio);
3624 f2fs_update_data_blkaddr(dn, fio->new_blkaddr);
3625
3626 f2fs_update_iostat(sbi, dn->inode, fio->io_type, F2FS_BLKSIZE);
3627}
3628
3629int f2fs_inplace_write_data(struct f2fs_io_info *fio)
3630{
3631 int err;
3632 struct f2fs_sb_info *sbi = fio->sbi;
3633 unsigned int segno;
3634
3635 fio->new_blkaddr = fio->old_blkaddr;
3636 /* i/o temperature is needed for passing down write hints */
3637 __get_segment_type(fio);
3638
3639 segno = GET_SEGNO(sbi, fio->new_blkaddr);
3640
3641 if (!IS_DATASEG(get_seg_entry(sbi, segno)->type)) {
3642 set_sbi_flag(sbi, SBI_NEED_FSCK);
3643 f2fs_warn(sbi, "%s: incorrect segment(%u) type, run fsck to fix.",
3644 __func__, segno);
3645 err = -EFSCORRUPTED;
3646 f2fs_handle_error(sbi, ERROR_INCONSISTENT_SUM_TYPE);
3647 goto drop_bio;
3648 }
3649
3650 if (f2fs_cp_error(sbi)) {
3651 err = -EIO;
3652 goto drop_bio;
3653 }
3654
3655 if (fio->post_read)
3656 invalidate_mapping_pages(META_MAPPING(sbi),
3657 fio->new_blkaddr, fio->new_blkaddr);
3658
3659 stat_inc_inplace_blocks(fio->sbi);
3660
3661 if (fio->bio && !IS_F2FS_IPU_NOCACHE(sbi))
3662 err = f2fs_merge_page_bio(fio);
3663 else
3664 err = f2fs_submit_page_bio(fio);
3665 if (!err) {
3666 f2fs_update_device_state(fio->sbi, fio->ino,
3667 fio->new_blkaddr, 1);
3668 f2fs_update_iostat(fio->sbi, fio->page->mapping->host,
3669 fio->io_type, F2FS_BLKSIZE);
3670 }
3671
3672 return err;
3673drop_bio:
3674 if (fio->bio && *(fio->bio)) {
3675 struct bio *bio = *(fio->bio);
3676
3677 bio->bi_status = BLK_STS_IOERR;
3678 bio_endio(bio);
3679 *(fio->bio) = NULL;
3680 }
3681 return err;
3682}
3683
3684static inline int __f2fs_get_curseg(struct f2fs_sb_info *sbi,
3685 unsigned int segno)
3686{
3687 int i;
3688
3689 for (i = CURSEG_HOT_DATA; i < NO_CHECK_TYPE; i++) {
3690 if (CURSEG_I(sbi, i)->segno == segno)
3691 break;
3692 }
3693 return i;
3694}
3695
3696void f2fs_do_replace_block(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
3697 block_t old_blkaddr, block_t new_blkaddr,
3698 bool recover_curseg, bool recover_newaddr,
3699 bool from_gc)
3700{
3701 struct sit_info *sit_i = SIT_I(sbi);
3702 struct curseg_info *curseg;
3703 unsigned int segno, old_cursegno;
3704 struct seg_entry *se;
3705 int type;
3706 unsigned short old_blkoff;
3707 unsigned char old_alloc_type;
3708
3709 segno = GET_SEGNO(sbi, new_blkaddr);
3710 se = get_seg_entry(sbi, segno);
3711 type = se->type;
3712
3713 f2fs_down_write(&SM_I(sbi)->curseg_lock);
3714
3715 if (!recover_curseg) {
3716 /* for recovery flow */
3717 if (se->valid_blocks == 0 && !IS_CURSEG(sbi, segno)) {
3718 if (old_blkaddr == NULL_ADDR)
3719 type = CURSEG_COLD_DATA;
3720 else
3721 type = CURSEG_WARM_DATA;
3722 }
3723 } else {
3724 if (IS_CURSEG(sbi, segno)) {
3725 /* se->type is volatile as SSR allocation */
3726 type = __f2fs_get_curseg(sbi, segno);
3727 f2fs_bug_on(sbi, type == NO_CHECK_TYPE);
3728 } else {
3729 type = CURSEG_WARM_DATA;
3730 }
3731 }
3732
3733 f2fs_bug_on(sbi, !IS_DATASEG(type));
3734 curseg = CURSEG_I(sbi, type);
3735
3736 mutex_lock(&curseg->curseg_mutex);
3737 down_write(&sit_i->sentry_lock);
3738
3739 old_cursegno = curseg->segno;
3740 old_blkoff = curseg->next_blkoff;
3741 old_alloc_type = curseg->alloc_type;
3742
3743 /* change the current segment */
3744 if (segno != curseg->segno) {
3745 curseg->next_segno = segno;
3746 change_curseg(sbi, type);
3747 }
3748
3749 curseg->next_blkoff = GET_BLKOFF_FROM_SEG0(sbi, new_blkaddr);
3750 curseg->sum_blk->entries[curseg->next_blkoff] = *sum;
3751
3752 if (!recover_curseg || recover_newaddr) {
3753 if (!from_gc)
3754 update_segment_mtime(sbi, new_blkaddr, 0);
3755 update_sit_entry(sbi, new_blkaddr, 1);
3756 }
3757 if (GET_SEGNO(sbi, old_blkaddr) != NULL_SEGNO) {
3758 f2fs_invalidate_internal_cache(sbi, old_blkaddr);
3759 if (!from_gc)
3760 update_segment_mtime(sbi, old_blkaddr, 0);
3761 update_sit_entry(sbi, old_blkaddr, -1);
3762 }
3763
3764 locate_dirty_segment(sbi, GET_SEGNO(sbi, old_blkaddr));
3765 locate_dirty_segment(sbi, GET_SEGNO(sbi, new_blkaddr));
3766
3767 locate_dirty_segment(sbi, old_cursegno);
3768
3769 if (recover_curseg) {
3770 if (old_cursegno != curseg->segno) {
3771 curseg->next_segno = old_cursegno;
3772 change_curseg(sbi, type);
3773 }
3774 curseg->next_blkoff = old_blkoff;
3775 curseg->alloc_type = old_alloc_type;
3776 }
3777
3778 up_write(&sit_i->sentry_lock);
3779 mutex_unlock(&curseg->curseg_mutex);
3780 f2fs_up_write(&SM_I(sbi)->curseg_lock);
3781}
3782
3783void f2fs_replace_block(struct f2fs_sb_info *sbi, struct dnode_of_data *dn,
3784 block_t old_addr, block_t new_addr,
3785 unsigned char version, bool recover_curseg,
3786 bool recover_newaddr)
3787{
3788 struct f2fs_summary sum;
3789
3790 set_summary(&sum, dn->nid, dn->ofs_in_node, version);
3791
3792 f2fs_do_replace_block(sbi, &sum, old_addr, new_addr,
3793 recover_curseg, recover_newaddr, false);
3794
3795 f2fs_update_data_blkaddr(dn, new_addr);
3796}
3797
3798void f2fs_wait_on_page_writeback(struct page *page,
3799 enum page_type type, bool ordered, bool locked)
3800{
3801 if (PageWriteback(page)) {
3802 struct f2fs_sb_info *sbi = F2FS_P_SB(page);
3803
3804 /* submit cached LFS IO */
3805 f2fs_submit_merged_write_cond(sbi, NULL, page, 0, type);
3806 /* submit cached IPU IO */
3807 f2fs_submit_merged_ipu_write(sbi, NULL, page);
3808 if (ordered) {
3809 wait_on_page_writeback(page);
3810 f2fs_bug_on(sbi, locked && PageWriteback(page));
3811 } else {
3812 wait_for_stable_page(page);
3813 }
3814 }
3815}
3816
3817void f2fs_wait_on_block_writeback(struct inode *inode, block_t blkaddr)
3818{
3819 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
3820 struct page *cpage;
3821
3822 if (!f2fs_post_read_required(inode))
3823 return;
3824
3825 if (!__is_valid_data_blkaddr(blkaddr))
3826 return;
3827
3828 cpage = find_lock_page(META_MAPPING(sbi), blkaddr);
3829 if (cpage) {
3830 f2fs_wait_on_page_writeback(cpage, DATA, true, true);
3831 f2fs_put_page(cpage, 1);
3832 }
3833}
3834
3835void f2fs_wait_on_block_writeback_range(struct inode *inode, block_t blkaddr,
3836 block_t len)
3837{
3838 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
3839 block_t i;
3840
3841 if (!f2fs_post_read_required(inode))
3842 return;
3843
3844 for (i = 0; i < len; i++)
3845 f2fs_wait_on_block_writeback(inode, blkaddr + i);
3846
3847 invalidate_mapping_pages(META_MAPPING(sbi), blkaddr, blkaddr + len - 1);
3848}
3849
3850static int read_compacted_summaries(struct f2fs_sb_info *sbi)
3851{
3852 struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
3853 struct curseg_info *seg_i;
3854 unsigned char *kaddr;
3855 struct page *page;
3856 block_t start;
3857 int i, j, offset;
3858
3859 start = start_sum_block(sbi);
3860
3861 page = f2fs_get_meta_page(sbi, start++);
3862 if (IS_ERR(page))
3863 return PTR_ERR(page);
3864 kaddr = (unsigned char *)page_address(page);
3865
3866 /* Step 1: restore nat cache */
3867 seg_i = CURSEG_I(sbi, CURSEG_HOT_DATA);
3868 memcpy(seg_i->journal, kaddr, SUM_JOURNAL_SIZE);
3869
3870 /* Step 2: restore sit cache */
3871 seg_i = CURSEG_I(sbi, CURSEG_COLD_DATA);
3872 memcpy(seg_i->journal, kaddr + SUM_JOURNAL_SIZE, SUM_JOURNAL_SIZE);
3873 offset = 2 * SUM_JOURNAL_SIZE;
3874
3875 /* Step 3: restore summary entries */
3876 for (i = CURSEG_HOT_DATA; i <= CURSEG_COLD_DATA; i++) {
3877 unsigned short blk_off;
3878 unsigned int segno;
3879
3880 seg_i = CURSEG_I(sbi, i);
3881 segno = le32_to_cpu(ckpt->cur_data_segno[i]);
3882 blk_off = le16_to_cpu(ckpt->cur_data_blkoff[i]);
3883 seg_i->next_segno = segno;
3884 reset_curseg(sbi, i, 0);
3885 seg_i->alloc_type = ckpt->alloc_type[i];
3886 seg_i->next_blkoff = blk_off;
3887
3888 if (seg_i->alloc_type == SSR)
3889 blk_off = sbi->blocks_per_seg;
3890
3891 for (j = 0; j < blk_off; j++) {
3892 struct f2fs_summary *s;
3893
3894 s = (struct f2fs_summary *)(kaddr + offset);
3895 seg_i->sum_blk->entries[j] = *s;
3896 offset += SUMMARY_SIZE;
3897 if (offset + SUMMARY_SIZE <= PAGE_SIZE -
3898 SUM_FOOTER_SIZE)
3899 continue;
3900
3901 f2fs_put_page(page, 1);
3902 page = NULL;
3903
3904 page = f2fs_get_meta_page(sbi, start++);
3905 if (IS_ERR(page))
3906 return PTR_ERR(page);
3907 kaddr = (unsigned char *)page_address(page);
3908 offset = 0;
3909 }
3910 }
3911 f2fs_put_page(page, 1);
3912 return 0;
3913}
3914
3915static int read_normal_summaries(struct f2fs_sb_info *sbi, int type)
3916{
3917 struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
3918 struct f2fs_summary_block *sum;
3919 struct curseg_info *curseg;
3920 struct page *new;
3921 unsigned short blk_off;
3922 unsigned int segno = 0;
3923 block_t blk_addr = 0;
3924 int err = 0;
3925
3926 /* get segment number and block addr */
3927 if (IS_DATASEG(type)) {
3928 segno = le32_to_cpu(ckpt->cur_data_segno[type]);
3929 blk_off = le16_to_cpu(ckpt->cur_data_blkoff[type -
3930 CURSEG_HOT_DATA]);
3931 if (__exist_node_summaries(sbi))
3932 blk_addr = sum_blk_addr(sbi, NR_CURSEG_PERSIST_TYPE, type);
3933 else
3934 blk_addr = sum_blk_addr(sbi, NR_CURSEG_DATA_TYPE, type);
3935 } else {
3936 segno = le32_to_cpu(ckpt->cur_node_segno[type -
3937 CURSEG_HOT_NODE]);
3938 blk_off = le16_to_cpu(ckpt->cur_node_blkoff[type -
3939 CURSEG_HOT_NODE]);
3940 if (__exist_node_summaries(sbi))
3941 blk_addr = sum_blk_addr(sbi, NR_CURSEG_NODE_TYPE,
3942 type - CURSEG_HOT_NODE);
3943 else
3944 blk_addr = GET_SUM_BLOCK(sbi, segno);
3945 }
3946
3947 new = f2fs_get_meta_page(sbi, blk_addr);
3948 if (IS_ERR(new))
3949 return PTR_ERR(new);
3950 sum = (struct f2fs_summary_block *)page_address(new);
3951
3952 if (IS_NODESEG(type)) {
3953 if (__exist_node_summaries(sbi)) {
3954 struct f2fs_summary *ns = &sum->entries[0];
3955 int i;
3956
3957 for (i = 0; i < sbi->blocks_per_seg; i++, ns++) {
3958 ns->version = 0;
3959 ns->ofs_in_node = 0;
3960 }
3961 } else {
3962 err = f2fs_restore_node_summary(sbi, segno, sum);
3963 if (err)
3964 goto out;
3965 }
3966 }
3967
3968 /* set uncompleted segment to curseg */
3969 curseg = CURSEG_I(sbi, type);
3970 mutex_lock(&curseg->curseg_mutex);
3971
3972 /* update journal info */
3973 down_write(&curseg->journal_rwsem);
3974 memcpy(curseg->journal, &sum->journal, SUM_JOURNAL_SIZE);
3975 up_write(&curseg->journal_rwsem);
3976
3977 memcpy(curseg->sum_blk->entries, sum->entries, SUM_ENTRY_SIZE);
3978 memcpy(&curseg->sum_blk->footer, &sum->footer, SUM_FOOTER_SIZE);
3979 curseg->next_segno = segno;
3980 reset_curseg(sbi, type, 0);
3981 curseg->alloc_type = ckpt->alloc_type[type];
3982 curseg->next_blkoff = blk_off;
3983 mutex_unlock(&curseg->curseg_mutex);
3984out:
3985 f2fs_put_page(new, 1);
3986 return err;
3987}
3988
3989static int restore_curseg_summaries(struct f2fs_sb_info *sbi)
3990{
3991 struct f2fs_journal *sit_j = CURSEG_I(sbi, CURSEG_COLD_DATA)->journal;
3992 struct f2fs_journal *nat_j = CURSEG_I(sbi, CURSEG_HOT_DATA)->journal;
3993 int type = CURSEG_HOT_DATA;
3994 int err;
3995
3996 if (is_set_ckpt_flags(sbi, CP_COMPACT_SUM_FLAG)) {
3997 int npages = f2fs_npages_for_summary_flush(sbi, true);
3998
3999 if (npages >= 2)
4000 f2fs_ra_meta_pages(sbi, start_sum_block(sbi), npages,
4001 META_CP, true);
4002
4003 /* restore for compacted data summary */
4004 err = read_compacted_summaries(sbi);
4005 if (err)
4006 return err;
4007 type = CURSEG_HOT_NODE;
4008 }
4009
4010 if (__exist_node_summaries(sbi))
4011 f2fs_ra_meta_pages(sbi,
4012 sum_blk_addr(sbi, NR_CURSEG_PERSIST_TYPE, type),
4013 NR_CURSEG_PERSIST_TYPE - type, META_CP, true);
4014
4015 for (; type <= CURSEG_COLD_NODE; type++) {
4016 err = read_normal_summaries(sbi, type);
4017 if (err)
4018 return err;
4019 }
4020
4021 /* sanity check for summary blocks */
4022 if (nats_in_cursum(nat_j) > NAT_JOURNAL_ENTRIES ||
4023 sits_in_cursum(sit_j) > SIT_JOURNAL_ENTRIES) {
4024 f2fs_err(sbi, "invalid journal entries nats %u sits %u",
4025 nats_in_cursum(nat_j), sits_in_cursum(sit_j));
4026 return -EINVAL;
4027 }
4028
4029 return 0;
4030}
4031
4032static void write_compacted_summaries(struct f2fs_sb_info *sbi, block_t blkaddr)
4033{
4034 struct page *page;
4035 unsigned char *kaddr;
4036 struct f2fs_summary *summary;
4037 struct curseg_info *seg_i;
4038 int written_size = 0;
4039 int i, j;
4040
4041 page = f2fs_grab_meta_page(sbi, blkaddr++);
4042 kaddr = (unsigned char *)page_address(page);
4043 memset(kaddr, 0, PAGE_SIZE);
4044
4045 /* Step 1: write nat cache */
4046 seg_i = CURSEG_I(sbi, CURSEG_HOT_DATA);
4047 memcpy(kaddr, seg_i->journal, SUM_JOURNAL_SIZE);
4048 written_size += SUM_JOURNAL_SIZE;
4049
4050 /* Step 2: write sit cache */
4051 seg_i = CURSEG_I(sbi, CURSEG_COLD_DATA);
4052 memcpy(kaddr + written_size, seg_i->journal, SUM_JOURNAL_SIZE);
4053 written_size += SUM_JOURNAL_SIZE;
4054
4055 /* Step 3: write summary entries */
4056 for (i = CURSEG_HOT_DATA; i <= CURSEG_COLD_DATA; i++) {
4057 seg_i = CURSEG_I(sbi, i);
4058 for (j = 0; j < f2fs_curseg_valid_blocks(sbi, i); j++) {
4059 if (!page) {
4060 page = f2fs_grab_meta_page(sbi, blkaddr++);
4061 kaddr = (unsigned char *)page_address(page);
4062 memset(kaddr, 0, PAGE_SIZE);
4063 written_size = 0;
4064 }
4065 summary = (struct f2fs_summary *)(kaddr + written_size);
4066 *summary = seg_i->sum_blk->entries[j];
4067 written_size += SUMMARY_SIZE;
4068
4069 if (written_size + SUMMARY_SIZE <= PAGE_SIZE -
4070 SUM_FOOTER_SIZE)
4071 continue;
4072
4073 set_page_dirty(page);
4074 f2fs_put_page(page, 1);
4075 page = NULL;
4076 }
4077 }
4078 if (page) {
4079 set_page_dirty(page);
4080 f2fs_put_page(page, 1);
4081 }
4082}
4083
4084static void write_normal_summaries(struct f2fs_sb_info *sbi,
4085 block_t blkaddr, int type)
4086{
4087 int i, end;
4088
4089 if (IS_DATASEG(type))
4090 end = type + NR_CURSEG_DATA_TYPE;
4091 else
4092 end = type + NR_CURSEG_NODE_TYPE;
4093
4094 for (i = type; i < end; i++)
4095 write_current_sum_page(sbi, i, blkaddr + (i - type));
4096}
4097
4098void f2fs_write_data_summaries(struct f2fs_sb_info *sbi, block_t start_blk)
4099{
4100 if (is_set_ckpt_flags(sbi, CP_COMPACT_SUM_FLAG))
4101 write_compacted_summaries(sbi, start_blk);
4102 else
4103 write_normal_summaries(sbi, start_blk, CURSEG_HOT_DATA);
4104}
4105
4106void f2fs_write_node_summaries(struct f2fs_sb_info *sbi, block_t start_blk)
4107{
4108 write_normal_summaries(sbi, start_blk, CURSEG_HOT_NODE);
4109}
4110
4111int f2fs_lookup_journal_in_cursum(struct f2fs_journal *journal, int type,
4112 unsigned int val, int alloc)
4113{
4114 int i;
4115
4116 if (type == NAT_JOURNAL) {
4117 for (i = 0; i < nats_in_cursum(journal); i++) {
4118 if (le32_to_cpu(nid_in_journal(journal, i)) == val)
4119 return i;
4120 }
4121 if (alloc && __has_cursum_space(journal, 1, NAT_JOURNAL))
4122 return update_nats_in_cursum(journal, 1);
4123 } else if (type == SIT_JOURNAL) {
4124 for (i = 0; i < sits_in_cursum(journal); i++)
4125 if (le32_to_cpu(segno_in_journal(journal, i)) == val)
4126 return i;
4127 if (alloc && __has_cursum_space(journal, 1, SIT_JOURNAL))
4128 return update_sits_in_cursum(journal, 1);
4129 }
4130 return -1;
4131}
4132
4133static struct page *get_current_sit_page(struct f2fs_sb_info *sbi,
4134 unsigned int segno)
4135{
4136 return f2fs_get_meta_page(sbi, current_sit_addr(sbi, segno));
4137}
4138
4139static struct page *get_next_sit_page(struct f2fs_sb_info *sbi,
4140 unsigned int start)
4141{
4142 struct sit_info *sit_i = SIT_I(sbi);
4143 struct page *page;
4144 pgoff_t src_off, dst_off;
4145
4146 src_off = current_sit_addr(sbi, start);
4147 dst_off = next_sit_addr(sbi, src_off);
4148
4149 page = f2fs_grab_meta_page(sbi, dst_off);
4150 seg_info_to_sit_page(sbi, page, start);
4151
4152 set_page_dirty(page);
4153 set_to_next_sit(sit_i, start);
4154
4155 return page;
4156}
4157
4158static struct sit_entry_set *grab_sit_entry_set(void)
4159{
4160 struct sit_entry_set *ses =
4161 f2fs_kmem_cache_alloc(sit_entry_set_slab,
4162 GFP_NOFS, true, NULL);
4163
4164 ses->entry_cnt = 0;
4165 INIT_LIST_HEAD(&ses->set_list);
4166 return ses;
4167}
4168
4169static void release_sit_entry_set(struct sit_entry_set *ses)
4170{
4171 list_del(&ses->set_list);
4172 kmem_cache_free(sit_entry_set_slab, ses);
4173}
4174
4175static void adjust_sit_entry_set(struct sit_entry_set *ses,
4176 struct list_head *head)
4177{
4178 struct sit_entry_set *next = ses;
4179
4180 if (list_is_last(&ses->set_list, head))
4181 return;
4182
4183 list_for_each_entry_continue(next, head, set_list)
4184 if (ses->entry_cnt <= next->entry_cnt) {
4185 list_move_tail(&ses->set_list, &next->set_list);
4186 return;
4187 }
4188
4189 list_move_tail(&ses->set_list, head);
4190}
4191
4192static void add_sit_entry(unsigned int segno, struct list_head *head)
4193{
4194 struct sit_entry_set *ses;
4195 unsigned int start_segno = START_SEGNO(segno);
4196
4197 list_for_each_entry(ses, head, set_list) {
4198 if (ses->start_segno == start_segno) {
4199 ses->entry_cnt++;
4200 adjust_sit_entry_set(ses, head);
4201 return;
4202 }
4203 }
4204
4205 ses = grab_sit_entry_set();
4206
4207 ses->start_segno = start_segno;
4208 ses->entry_cnt++;
4209 list_add(&ses->set_list, head);
4210}
4211
4212static void add_sits_in_set(struct f2fs_sb_info *sbi)
4213{
4214 struct f2fs_sm_info *sm_info = SM_I(sbi);
4215 struct list_head *set_list = &sm_info->sit_entry_set;
4216 unsigned long *bitmap = SIT_I(sbi)->dirty_sentries_bitmap;
4217 unsigned int segno;
4218
4219 for_each_set_bit(segno, bitmap, MAIN_SEGS(sbi))
4220 add_sit_entry(segno, set_list);
4221}
4222
4223static void remove_sits_in_journal(struct f2fs_sb_info *sbi)
4224{
4225 struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_COLD_DATA);
4226 struct f2fs_journal *journal = curseg->journal;
4227 int i;
4228
4229 down_write(&curseg->journal_rwsem);
4230 for (i = 0; i < sits_in_cursum(journal); i++) {
4231 unsigned int segno;
4232 bool dirtied;
4233
4234 segno = le32_to_cpu(segno_in_journal(journal, i));
4235 dirtied = __mark_sit_entry_dirty(sbi, segno);
4236
4237 if (!dirtied)
4238 add_sit_entry(segno, &SM_I(sbi)->sit_entry_set);
4239 }
4240 update_sits_in_cursum(journal, -i);
4241 up_write(&curseg->journal_rwsem);
4242}
4243
4244/*
4245 * CP calls this function, which flushes SIT entries including sit_journal,
4246 * and moves prefree segs to free segs.
4247 */
4248void f2fs_flush_sit_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc)
4249{
4250 struct sit_info *sit_i = SIT_I(sbi);
4251 unsigned long *bitmap = sit_i->dirty_sentries_bitmap;
4252 struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_COLD_DATA);
4253 struct f2fs_journal *journal = curseg->journal;
4254 struct sit_entry_set *ses, *tmp;
4255 struct list_head *head = &SM_I(sbi)->sit_entry_set;
4256 bool to_journal = !is_sbi_flag_set(sbi, SBI_IS_RESIZEFS);
4257 struct seg_entry *se;
4258
4259 down_write(&sit_i->sentry_lock);
4260
4261 if (!sit_i->dirty_sentries)
4262 goto out;
4263
4264 /*
4265 * add and account sit entries of dirty bitmap in sit entry
4266 * set temporarily
4267 */
4268 add_sits_in_set(sbi);
4269
4270 /*
4271 * if there are no enough space in journal to store dirty sit
4272 * entries, remove all entries from journal and add and account
4273 * them in sit entry set.
4274 */
4275 if (!__has_cursum_space(journal, sit_i->dirty_sentries, SIT_JOURNAL) ||
4276 !to_journal)
4277 remove_sits_in_journal(sbi);
4278
4279 /*
4280 * there are two steps to flush sit entries:
4281 * #1, flush sit entries to journal in current cold data summary block.
4282 * #2, flush sit entries to sit page.
4283 */
4284 list_for_each_entry_safe(ses, tmp, head, set_list) {
4285 struct page *page = NULL;
4286 struct f2fs_sit_block *raw_sit = NULL;
4287 unsigned int start_segno = ses->start_segno;
4288 unsigned int end = min(start_segno + SIT_ENTRY_PER_BLOCK,
4289 (unsigned long)MAIN_SEGS(sbi));
4290 unsigned int segno = start_segno;
4291
4292 if (to_journal &&
4293 !__has_cursum_space(journal, ses->entry_cnt, SIT_JOURNAL))
4294 to_journal = false;
4295
4296 if (to_journal) {
4297 down_write(&curseg->journal_rwsem);
4298 } else {
4299 page = get_next_sit_page(sbi, start_segno);
4300 raw_sit = page_address(page);
4301 }
4302
4303 /* flush dirty sit entries in region of current sit set */
4304 for_each_set_bit_from(segno, bitmap, end) {
4305 int offset, sit_offset;
4306
4307 se = get_seg_entry(sbi, segno);
4308#ifdef CONFIG_F2FS_CHECK_FS
4309 if (memcmp(se->cur_valid_map, se->cur_valid_map_mir,
4310 SIT_VBLOCK_MAP_SIZE))
4311 f2fs_bug_on(sbi, 1);
4312#endif
4313
4314 /* add discard candidates */
4315 if (!(cpc->reason & CP_DISCARD)) {
4316 cpc->trim_start = segno;
4317 add_discard_addrs(sbi, cpc, false);
4318 }
4319
4320 if (to_journal) {
4321 offset = f2fs_lookup_journal_in_cursum(journal,
4322 SIT_JOURNAL, segno, 1);
4323 f2fs_bug_on(sbi, offset < 0);
4324 segno_in_journal(journal, offset) =
4325 cpu_to_le32(segno);
4326 seg_info_to_raw_sit(se,
4327 &sit_in_journal(journal, offset));
4328 check_block_count(sbi, segno,
4329 &sit_in_journal(journal, offset));
4330 } else {
4331 sit_offset = SIT_ENTRY_OFFSET(sit_i, segno);
4332 seg_info_to_raw_sit(se,
4333 &raw_sit->entries[sit_offset]);
4334 check_block_count(sbi, segno,
4335 &raw_sit->entries[sit_offset]);
4336 }
4337
4338 __clear_bit(segno, bitmap);
4339 sit_i->dirty_sentries--;
4340 ses->entry_cnt--;
4341 }
4342
4343 if (to_journal)
4344 up_write(&curseg->journal_rwsem);
4345 else
4346 f2fs_put_page(page, 1);
4347
4348 f2fs_bug_on(sbi, ses->entry_cnt);
4349 release_sit_entry_set(ses);
4350 }
4351
4352 f2fs_bug_on(sbi, !list_empty(head));
4353 f2fs_bug_on(sbi, sit_i->dirty_sentries);
4354out:
4355 if (cpc->reason & CP_DISCARD) {
4356 __u64 trim_start = cpc->trim_start;
4357
4358 for (; cpc->trim_start <= cpc->trim_end; cpc->trim_start++)
4359 add_discard_addrs(sbi, cpc, false);
4360
4361 cpc->trim_start = trim_start;
4362 }
4363 up_write(&sit_i->sentry_lock);
4364
4365 set_prefree_as_free_segments(sbi);
4366}
4367
4368static int build_sit_info(struct f2fs_sb_info *sbi)
4369{
4370 struct f2fs_super_block *raw_super = F2FS_RAW_SUPER(sbi);
4371 struct sit_info *sit_i;
4372 unsigned int sit_segs, start;
4373 char *src_bitmap, *bitmap;
4374 unsigned int bitmap_size, main_bitmap_size, sit_bitmap_size;
4375 unsigned int discard_map = f2fs_block_unit_discard(sbi) ? 1 : 0;
4376
4377 /* allocate memory for SIT information */
4378 sit_i = f2fs_kzalloc(sbi, sizeof(struct sit_info), GFP_KERNEL);
4379 if (!sit_i)
4380 return -ENOMEM;
4381
4382 SM_I(sbi)->sit_info = sit_i;
4383
4384 sit_i->sentries =
4385 f2fs_kvzalloc(sbi, array_size(sizeof(struct seg_entry),
4386 MAIN_SEGS(sbi)),
4387 GFP_KERNEL);
4388 if (!sit_i->sentries)
4389 return -ENOMEM;
4390
4391 main_bitmap_size = f2fs_bitmap_size(MAIN_SEGS(sbi));
4392 sit_i->dirty_sentries_bitmap = f2fs_kvzalloc(sbi, main_bitmap_size,
4393 GFP_KERNEL);
4394 if (!sit_i->dirty_sentries_bitmap)
4395 return -ENOMEM;
4396
4397#ifdef CONFIG_F2FS_CHECK_FS
4398 bitmap_size = MAIN_SEGS(sbi) * SIT_VBLOCK_MAP_SIZE * (3 + discard_map);
4399#else
4400 bitmap_size = MAIN_SEGS(sbi) * SIT_VBLOCK_MAP_SIZE * (2 + discard_map);
4401#endif
4402 sit_i->bitmap = f2fs_kvzalloc(sbi, bitmap_size, GFP_KERNEL);
4403 if (!sit_i->bitmap)
4404 return -ENOMEM;
4405
4406 bitmap = sit_i->bitmap;
4407
4408 for (start = 0; start < MAIN_SEGS(sbi); start++) {
4409 sit_i->sentries[start].cur_valid_map = bitmap;
4410 bitmap += SIT_VBLOCK_MAP_SIZE;
4411
4412 sit_i->sentries[start].ckpt_valid_map = bitmap;
4413 bitmap += SIT_VBLOCK_MAP_SIZE;
4414
4415#ifdef CONFIG_F2FS_CHECK_FS
4416 sit_i->sentries[start].cur_valid_map_mir = bitmap;
4417 bitmap += SIT_VBLOCK_MAP_SIZE;
4418#endif
4419
4420 if (discard_map) {
4421 sit_i->sentries[start].discard_map = bitmap;
4422 bitmap += SIT_VBLOCK_MAP_SIZE;
4423 }
4424 }
4425
4426 sit_i->tmp_map = f2fs_kzalloc(sbi, SIT_VBLOCK_MAP_SIZE, GFP_KERNEL);
4427 if (!sit_i->tmp_map)
4428 return -ENOMEM;
4429
4430 if (__is_large_section(sbi)) {
4431 sit_i->sec_entries =
4432 f2fs_kvzalloc(sbi, array_size(sizeof(struct sec_entry),
4433 MAIN_SECS(sbi)),
4434 GFP_KERNEL);
4435 if (!sit_i->sec_entries)
4436 return -ENOMEM;
4437 }
4438
4439 /* get information related with SIT */
4440 sit_segs = le32_to_cpu(raw_super->segment_count_sit) >> 1;
4441
4442 /* setup SIT bitmap from ckeckpoint pack */
4443 sit_bitmap_size = __bitmap_size(sbi, SIT_BITMAP);
4444 src_bitmap = __bitmap_ptr(sbi, SIT_BITMAP);
4445
4446 sit_i->sit_bitmap = kmemdup(src_bitmap, sit_bitmap_size, GFP_KERNEL);
4447 if (!sit_i->sit_bitmap)
4448 return -ENOMEM;
4449
4450#ifdef CONFIG_F2FS_CHECK_FS
4451 sit_i->sit_bitmap_mir = kmemdup(src_bitmap,
4452 sit_bitmap_size, GFP_KERNEL);
4453 if (!sit_i->sit_bitmap_mir)
4454 return -ENOMEM;
4455
4456 sit_i->invalid_segmap = f2fs_kvzalloc(sbi,
4457 main_bitmap_size, GFP_KERNEL);
4458 if (!sit_i->invalid_segmap)
4459 return -ENOMEM;
4460#endif
4461
4462 sit_i->sit_base_addr = le32_to_cpu(raw_super->sit_blkaddr);
4463 sit_i->sit_blocks = sit_segs << sbi->log_blocks_per_seg;
4464 sit_i->written_valid_blocks = 0;
4465 sit_i->bitmap_size = sit_bitmap_size;
4466 sit_i->dirty_sentries = 0;
4467 sit_i->sents_per_block = SIT_ENTRY_PER_BLOCK;
4468 sit_i->elapsed_time = le64_to_cpu(sbi->ckpt->elapsed_time);
4469 sit_i->mounted_time = ktime_get_boottime_seconds();
4470 init_rwsem(&sit_i->sentry_lock);
4471 return 0;
4472}
4473
4474static int build_free_segmap(struct f2fs_sb_info *sbi)
4475{
4476 struct free_segmap_info *free_i;
4477 unsigned int bitmap_size, sec_bitmap_size;
4478
4479 /* allocate memory for free segmap information */
4480 free_i = f2fs_kzalloc(sbi, sizeof(struct free_segmap_info), GFP_KERNEL);
4481 if (!free_i)
4482 return -ENOMEM;
4483
4484 SM_I(sbi)->free_info = free_i;
4485
4486 bitmap_size = f2fs_bitmap_size(MAIN_SEGS(sbi));
4487 free_i->free_segmap = f2fs_kvmalloc(sbi, bitmap_size, GFP_KERNEL);
4488 if (!free_i->free_segmap)
4489 return -ENOMEM;
4490
4491 sec_bitmap_size = f2fs_bitmap_size(MAIN_SECS(sbi));
4492 free_i->free_secmap = f2fs_kvmalloc(sbi, sec_bitmap_size, GFP_KERNEL);
4493 if (!free_i->free_secmap)
4494 return -ENOMEM;
4495
4496 /* set all segments as dirty temporarily */
4497 memset(free_i->free_segmap, 0xff, bitmap_size);
4498 memset(free_i->free_secmap, 0xff, sec_bitmap_size);
4499
4500 /* init free segmap information */
4501 free_i->start_segno = GET_SEGNO_FROM_SEG0(sbi, MAIN_BLKADDR(sbi));
4502 free_i->free_segments = 0;
4503 free_i->free_sections = 0;
4504 spin_lock_init(&free_i->segmap_lock);
4505 return 0;
4506}
4507
4508static int build_curseg(struct f2fs_sb_info *sbi)
4509{
4510 struct curseg_info *array;
4511 int i;
4512
4513 array = f2fs_kzalloc(sbi, array_size(NR_CURSEG_TYPE,
4514 sizeof(*array)), GFP_KERNEL);
4515 if (!array)
4516 return -ENOMEM;
4517
4518 SM_I(sbi)->curseg_array = array;
4519
4520 for (i = 0; i < NO_CHECK_TYPE; i++) {
4521 mutex_init(&array[i].curseg_mutex);
4522 array[i].sum_blk = f2fs_kzalloc(sbi, PAGE_SIZE, GFP_KERNEL);
4523 if (!array[i].sum_blk)
4524 return -ENOMEM;
4525 init_rwsem(&array[i].journal_rwsem);
4526 array[i].journal = f2fs_kzalloc(sbi,
4527 sizeof(struct f2fs_journal), GFP_KERNEL);
4528 if (!array[i].journal)
4529 return -ENOMEM;
4530 if (i < NR_PERSISTENT_LOG)
4531 array[i].seg_type = CURSEG_HOT_DATA + i;
4532 else if (i == CURSEG_COLD_DATA_PINNED)
4533 array[i].seg_type = CURSEG_COLD_DATA;
4534 else if (i == CURSEG_ALL_DATA_ATGC)
4535 array[i].seg_type = CURSEG_COLD_DATA;
4536 array[i].segno = NULL_SEGNO;
4537 array[i].next_blkoff = 0;
4538 array[i].inited = false;
4539 }
4540 return restore_curseg_summaries(sbi);
4541}
4542
4543static int build_sit_entries(struct f2fs_sb_info *sbi)
4544{
4545 struct sit_info *sit_i = SIT_I(sbi);
4546 struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_COLD_DATA);
4547 struct f2fs_journal *journal = curseg->journal;
4548 struct seg_entry *se;
4549 struct f2fs_sit_entry sit;
4550 int sit_blk_cnt = SIT_BLK_CNT(sbi);
4551 unsigned int i, start, end;
4552 unsigned int readed, start_blk = 0;
4553 int err = 0;
4554 block_t sit_valid_blocks[2] = {0, 0};
4555
4556 do {
4557 readed = f2fs_ra_meta_pages(sbi, start_blk, BIO_MAX_VECS,
4558 META_SIT, true);
4559
4560 start = start_blk * sit_i->sents_per_block;
4561 end = (start_blk + readed) * sit_i->sents_per_block;
4562
4563 for (; start < end && start < MAIN_SEGS(sbi); start++) {
4564 struct f2fs_sit_block *sit_blk;
4565 struct page *page;
4566
4567 se = &sit_i->sentries[start];
4568 page = get_current_sit_page(sbi, start);
4569 if (IS_ERR(page))
4570 return PTR_ERR(page);
4571 sit_blk = (struct f2fs_sit_block *)page_address(page);
4572 sit = sit_blk->entries[SIT_ENTRY_OFFSET(sit_i, start)];
4573 f2fs_put_page(page, 1);
4574
4575 err = check_block_count(sbi, start, &sit);
4576 if (err)
4577 return err;
4578 seg_info_from_raw_sit(se, &sit);
4579
4580 if (se->type >= NR_PERSISTENT_LOG) {
4581 f2fs_err(sbi, "Invalid segment type: %u, segno: %u",
4582 se->type, start);
4583 f2fs_handle_error(sbi,
4584 ERROR_INCONSISTENT_SUM_TYPE);
4585 return -EFSCORRUPTED;
4586 }
4587
4588 sit_valid_blocks[SE_PAGETYPE(se)] += se->valid_blocks;
4589
4590 if (f2fs_block_unit_discard(sbi)) {
4591 /* build discard map only one time */
4592 if (is_set_ckpt_flags(sbi, CP_TRIMMED_FLAG)) {
4593 memset(se->discard_map, 0xff,
4594 SIT_VBLOCK_MAP_SIZE);
4595 } else {
4596 memcpy(se->discard_map,
4597 se->cur_valid_map,
4598 SIT_VBLOCK_MAP_SIZE);
4599 sbi->discard_blks +=
4600 sbi->blocks_per_seg -
4601 se->valid_blocks;
4602 }
4603 }
4604
4605 if (__is_large_section(sbi))
4606 get_sec_entry(sbi, start)->valid_blocks +=
4607 se->valid_blocks;
4608 }
4609 start_blk += readed;
4610 } while (start_blk < sit_blk_cnt);
4611
4612 down_read(&curseg->journal_rwsem);
4613 for (i = 0; i < sits_in_cursum(journal); i++) {
4614 unsigned int old_valid_blocks;
4615
4616 start = le32_to_cpu(segno_in_journal(journal, i));
4617 if (start >= MAIN_SEGS(sbi)) {
4618 f2fs_err(sbi, "Wrong journal entry on segno %u",
4619 start);
4620 err = -EFSCORRUPTED;
4621 f2fs_handle_error(sbi, ERROR_CORRUPTED_JOURNAL);
4622 break;
4623 }
4624
4625 se = &sit_i->sentries[start];
4626 sit = sit_in_journal(journal, i);
4627
4628 old_valid_blocks = se->valid_blocks;
4629
4630 sit_valid_blocks[SE_PAGETYPE(se)] -= old_valid_blocks;
4631
4632 err = check_block_count(sbi, start, &sit);
4633 if (err)
4634 break;
4635 seg_info_from_raw_sit(se, &sit);
4636
4637 if (se->type >= NR_PERSISTENT_LOG) {
4638 f2fs_err(sbi, "Invalid segment type: %u, segno: %u",
4639 se->type, start);
4640 err = -EFSCORRUPTED;
4641 f2fs_handle_error(sbi, ERROR_INCONSISTENT_SUM_TYPE);
4642 break;
4643 }
4644
4645 sit_valid_blocks[SE_PAGETYPE(se)] += se->valid_blocks;
4646
4647 if (f2fs_block_unit_discard(sbi)) {
4648 if (is_set_ckpt_flags(sbi, CP_TRIMMED_FLAG)) {
4649 memset(se->discard_map, 0xff, SIT_VBLOCK_MAP_SIZE);
4650 } else {
4651 memcpy(se->discard_map, se->cur_valid_map,
4652 SIT_VBLOCK_MAP_SIZE);
4653 sbi->discard_blks += old_valid_blocks;
4654 sbi->discard_blks -= se->valid_blocks;
4655 }
4656 }
4657
4658 if (__is_large_section(sbi)) {
4659 get_sec_entry(sbi, start)->valid_blocks +=
4660 se->valid_blocks;
4661 get_sec_entry(sbi, start)->valid_blocks -=
4662 old_valid_blocks;
4663 }
4664 }
4665 up_read(&curseg->journal_rwsem);
4666
4667 if (err)
4668 return err;
4669
4670 if (sit_valid_blocks[NODE] != valid_node_count(sbi)) {
4671 f2fs_err(sbi, "SIT is corrupted node# %u vs %u",
4672 sit_valid_blocks[NODE], valid_node_count(sbi));
4673 f2fs_handle_error(sbi, ERROR_INCONSISTENT_NODE_COUNT);
4674 return -EFSCORRUPTED;
4675 }
4676
4677 if (sit_valid_blocks[DATA] + sit_valid_blocks[NODE] >
4678 valid_user_blocks(sbi)) {
4679 f2fs_err(sbi, "SIT is corrupted data# %u %u vs %u",
4680 sit_valid_blocks[DATA], sit_valid_blocks[NODE],
4681 valid_user_blocks(sbi));
4682 f2fs_handle_error(sbi, ERROR_INCONSISTENT_BLOCK_COUNT);
4683 return -EFSCORRUPTED;
4684 }
4685
4686 return 0;
4687}
4688
4689static void init_free_segmap(struct f2fs_sb_info *sbi)
4690{
4691 unsigned int start;
4692 int type;
4693 struct seg_entry *sentry;
4694
4695 for (start = 0; start < MAIN_SEGS(sbi); start++) {
4696 if (f2fs_usable_blks_in_seg(sbi, start) == 0)
4697 continue;
4698 sentry = get_seg_entry(sbi, start);
4699 if (!sentry->valid_blocks)
4700 __set_free(sbi, start);
4701 else
4702 SIT_I(sbi)->written_valid_blocks +=
4703 sentry->valid_blocks;
4704 }
4705
4706 /* set use the current segments */
4707 for (type = CURSEG_HOT_DATA; type <= CURSEG_COLD_NODE; type++) {
4708 struct curseg_info *curseg_t = CURSEG_I(sbi, type);
4709
4710 __set_test_and_inuse(sbi, curseg_t->segno);
4711 }
4712}
4713
4714static void init_dirty_segmap(struct f2fs_sb_info *sbi)
4715{
4716 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
4717 struct free_segmap_info *free_i = FREE_I(sbi);
4718 unsigned int segno = 0, offset = 0, secno;
4719 block_t valid_blocks, usable_blks_in_seg;
4720
4721 while (1) {
4722 /* find dirty segment based on free segmap */
4723 segno = find_next_inuse(free_i, MAIN_SEGS(sbi), offset);
4724 if (segno >= MAIN_SEGS(sbi))
4725 break;
4726 offset = segno + 1;
4727 valid_blocks = get_valid_blocks(sbi, segno, false);
4728 usable_blks_in_seg = f2fs_usable_blks_in_seg(sbi, segno);
4729 if (valid_blocks == usable_blks_in_seg || !valid_blocks)
4730 continue;
4731 if (valid_blocks > usable_blks_in_seg) {
4732 f2fs_bug_on(sbi, 1);
4733 continue;
4734 }
4735 mutex_lock(&dirty_i->seglist_lock);
4736 __locate_dirty_segment(sbi, segno, DIRTY);
4737 mutex_unlock(&dirty_i->seglist_lock);
4738 }
4739
4740 if (!__is_large_section(sbi))
4741 return;
4742
4743 mutex_lock(&dirty_i->seglist_lock);
4744 for (segno = 0; segno < MAIN_SEGS(sbi); segno += sbi->segs_per_sec) {
4745 valid_blocks = get_valid_blocks(sbi, segno, true);
4746 secno = GET_SEC_FROM_SEG(sbi, segno);
4747
4748 if (!valid_blocks || valid_blocks == CAP_BLKS_PER_SEC(sbi))
4749 continue;
4750 if (IS_CURSEC(sbi, secno))
4751 continue;
4752 set_bit(secno, dirty_i->dirty_secmap);
4753 }
4754 mutex_unlock(&dirty_i->seglist_lock);
4755}
4756
4757static int init_victim_secmap(struct f2fs_sb_info *sbi)
4758{
4759 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
4760 unsigned int bitmap_size = f2fs_bitmap_size(MAIN_SECS(sbi));
4761
4762 dirty_i->victim_secmap = f2fs_kvzalloc(sbi, bitmap_size, GFP_KERNEL);
4763 if (!dirty_i->victim_secmap)
4764 return -ENOMEM;
4765
4766 dirty_i->pinned_secmap = f2fs_kvzalloc(sbi, bitmap_size, GFP_KERNEL);
4767 if (!dirty_i->pinned_secmap)
4768 return -ENOMEM;
4769
4770 dirty_i->pinned_secmap_cnt = 0;
4771 dirty_i->enable_pin_section = true;
4772 return 0;
4773}
4774
4775static int build_dirty_segmap(struct f2fs_sb_info *sbi)
4776{
4777 struct dirty_seglist_info *dirty_i;
4778 unsigned int bitmap_size, i;
4779
4780 /* allocate memory for dirty segments list information */
4781 dirty_i = f2fs_kzalloc(sbi, sizeof(struct dirty_seglist_info),
4782 GFP_KERNEL);
4783 if (!dirty_i)
4784 return -ENOMEM;
4785
4786 SM_I(sbi)->dirty_info = dirty_i;
4787 mutex_init(&dirty_i->seglist_lock);
4788
4789 bitmap_size = f2fs_bitmap_size(MAIN_SEGS(sbi));
4790
4791 for (i = 0; i < NR_DIRTY_TYPE; i++) {
4792 dirty_i->dirty_segmap[i] = f2fs_kvzalloc(sbi, bitmap_size,
4793 GFP_KERNEL);
4794 if (!dirty_i->dirty_segmap[i])
4795 return -ENOMEM;
4796 }
4797
4798 if (__is_large_section(sbi)) {
4799 bitmap_size = f2fs_bitmap_size(MAIN_SECS(sbi));
4800 dirty_i->dirty_secmap = f2fs_kvzalloc(sbi,
4801 bitmap_size, GFP_KERNEL);
4802 if (!dirty_i->dirty_secmap)
4803 return -ENOMEM;
4804 }
4805
4806 init_dirty_segmap(sbi);
4807 return init_victim_secmap(sbi);
4808}
4809
4810static int sanity_check_curseg(struct f2fs_sb_info *sbi)
4811{
4812 int i;
4813
4814 /*
4815 * In LFS/SSR curseg, .next_blkoff should point to an unused blkaddr;
4816 * In LFS curseg, all blkaddr after .next_blkoff should be unused.
4817 */
4818 for (i = 0; i < NR_PERSISTENT_LOG; i++) {
4819 struct curseg_info *curseg = CURSEG_I(sbi, i);
4820 struct seg_entry *se = get_seg_entry(sbi, curseg->segno);
4821 unsigned int blkofs = curseg->next_blkoff;
4822
4823 if (f2fs_sb_has_readonly(sbi) &&
4824 i != CURSEG_HOT_DATA && i != CURSEG_HOT_NODE)
4825 continue;
4826
4827 sanity_check_seg_type(sbi, curseg->seg_type);
4828
4829 if (curseg->alloc_type != LFS && curseg->alloc_type != SSR) {
4830 f2fs_err(sbi,
4831 "Current segment has invalid alloc_type:%d",
4832 curseg->alloc_type);
4833 f2fs_handle_error(sbi, ERROR_INVALID_CURSEG);
4834 return -EFSCORRUPTED;
4835 }
4836
4837 if (f2fs_test_bit(blkofs, se->cur_valid_map))
4838 goto out;
4839
4840 if (curseg->alloc_type == SSR)
4841 continue;
4842
4843 for (blkofs += 1; blkofs < sbi->blocks_per_seg; blkofs++) {
4844 if (!f2fs_test_bit(blkofs, se->cur_valid_map))
4845 continue;
4846out:
4847 f2fs_err(sbi,
4848 "Current segment's next free block offset is inconsistent with bitmap, logtype:%u, segno:%u, type:%u, next_blkoff:%u, blkofs:%u",
4849 i, curseg->segno, curseg->alloc_type,
4850 curseg->next_blkoff, blkofs);
4851 f2fs_handle_error(sbi, ERROR_INVALID_CURSEG);
4852 return -EFSCORRUPTED;
4853 }
4854 }
4855 return 0;
4856}
4857
4858#ifdef CONFIG_BLK_DEV_ZONED
4859
4860static int check_zone_write_pointer(struct f2fs_sb_info *sbi,
4861 struct f2fs_dev_info *fdev,
4862 struct blk_zone *zone)
4863{
4864 unsigned int zone_segno;
4865 block_t zone_block, valid_block_cnt;
4866 unsigned int log_sectors_per_block = sbi->log_blocksize - SECTOR_SHIFT;
4867 int ret;
4868
4869 if (zone->type != BLK_ZONE_TYPE_SEQWRITE_REQ)
4870 return 0;
4871
4872 zone_block = fdev->start_blk + (zone->start >> log_sectors_per_block);
4873 zone_segno = GET_SEGNO(sbi, zone_block);
4874
4875 /*
4876 * Skip check of zones cursegs point to, since
4877 * fix_curseg_write_pointer() checks them.
4878 */
4879 if (zone_segno >= MAIN_SEGS(sbi) ||
4880 IS_CURSEC(sbi, GET_SEC_FROM_SEG(sbi, zone_segno)))
4881 return 0;
4882
4883 /*
4884 * Get # of valid block of the zone.
4885 */
4886 valid_block_cnt = get_valid_blocks(sbi, zone_segno, true);
4887
4888 if ((!valid_block_cnt && zone->cond == BLK_ZONE_COND_EMPTY) ||
4889 (valid_block_cnt && zone->cond == BLK_ZONE_COND_FULL))
4890 return 0;
4891
4892 if (!valid_block_cnt) {
4893 f2fs_notice(sbi, "Zone without valid block has non-zero write "
4894 "pointer. Reset the write pointer: cond[0x%x]",
4895 zone->cond);
4896 ret = __f2fs_issue_discard_zone(sbi, fdev->bdev, zone_block,
4897 zone->len >> log_sectors_per_block);
4898 if (ret)
4899 f2fs_err(sbi, "Discard zone failed: %s (errno=%d)",
4900 fdev->path, ret);
4901 return ret;
4902 }
4903
4904 /*
4905 * If there are valid blocks and the write pointer doesn't match
4906 * with them, we need to report the inconsistency and fill
4907 * the zone till the end to close the zone. This inconsistency
4908 * does not cause write error because the zone will not be
4909 * selected for write operation until it get discarded.
4910 */
4911 f2fs_notice(sbi, "Valid blocks are not aligned with write "
4912 "pointer: valid block[0x%x,0x%x] cond[0x%x]",
4913 zone_segno, valid_block_cnt, zone->cond);
4914
4915 ret = blkdev_zone_mgmt(fdev->bdev, REQ_OP_ZONE_FINISH,
4916 zone->start, zone->len, GFP_NOFS);
4917 if (ret == -EOPNOTSUPP) {
4918 ret = blkdev_issue_zeroout(fdev->bdev, zone->wp,
4919 zone->len - (zone->wp - zone->start),
4920 GFP_NOFS, 0);
4921 if (ret)
4922 f2fs_err(sbi, "Fill up zone failed: %s (errno=%d)",
4923 fdev->path, ret);
4924 } else if (ret) {
4925 f2fs_err(sbi, "Finishing zone failed: %s (errno=%d)",
4926 fdev->path, ret);
4927 }
4928
4929 return ret;
4930}
4931
4932static struct f2fs_dev_info *get_target_zoned_dev(struct f2fs_sb_info *sbi,
4933 block_t zone_blkaddr)
4934{
4935 int i;
4936
4937 for (i = 0; i < sbi->s_ndevs; i++) {
4938 if (!bdev_is_zoned(FDEV(i).bdev))
4939 continue;
4940 if (sbi->s_ndevs == 1 || (FDEV(i).start_blk <= zone_blkaddr &&
4941 zone_blkaddr <= FDEV(i).end_blk))
4942 return &FDEV(i);
4943 }
4944
4945 return NULL;
4946}
4947
4948static int report_one_zone_cb(struct blk_zone *zone, unsigned int idx,
4949 void *data)
4950{
4951 memcpy(data, zone, sizeof(struct blk_zone));
4952 return 0;
4953}
4954
4955static int fix_curseg_write_pointer(struct f2fs_sb_info *sbi, int type)
4956{
4957 struct curseg_info *cs = CURSEG_I(sbi, type);
4958 struct f2fs_dev_info *zbd;
4959 struct blk_zone zone;
4960 unsigned int cs_section, wp_segno, wp_blkoff, wp_sector_off;
4961 block_t cs_zone_block, wp_block;
4962 unsigned int log_sectors_per_block = sbi->log_blocksize - SECTOR_SHIFT;
4963 sector_t zone_sector;
4964 int err;
4965
4966 cs_section = GET_SEC_FROM_SEG(sbi, cs->segno);
4967 cs_zone_block = START_BLOCK(sbi, GET_SEG_FROM_SEC(sbi, cs_section));
4968
4969 zbd = get_target_zoned_dev(sbi, cs_zone_block);
4970 if (!zbd)
4971 return 0;
4972
4973 /* report zone for the sector the curseg points to */
4974 zone_sector = (sector_t)(cs_zone_block - zbd->start_blk)
4975 << log_sectors_per_block;
4976 err = blkdev_report_zones(zbd->bdev, zone_sector, 1,
4977 report_one_zone_cb, &zone);
4978 if (err != 1) {
4979 f2fs_err(sbi, "Report zone failed: %s errno=(%d)",
4980 zbd->path, err);
4981 return err;
4982 }
4983
4984 if (zone.type != BLK_ZONE_TYPE_SEQWRITE_REQ)
4985 return 0;
4986
4987 /*
4988 * When safely unmounted in the previous mount, we could use current
4989 * segments. Otherwise, allocate new sections.
4990 */
4991 if (is_set_ckpt_flags(sbi, CP_UMOUNT_FLAG)) {
4992 wp_block = zbd->start_blk + (zone.wp >> log_sectors_per_block);
4993 wp_segno = GET_SEGNO(sbi, wp_block);
4994 wp_blkoff = wp_block - START_BLOCK(sbi, wp_segno);
4995 wp_sector_off = zone.wp & GENMASK(log_sectors_per_block - 1, 0);
4996
4997 if (cs->segno == wp_segno && cs->next_blkoff == wp_blkoff &&
4998 wp_sector_off == 0)
4999 return 0;
5000
5001 f2fs_notice(sbi, "Unaligned curseg[%d] with write pointer: "
5002 "curseg[0x%x,0x%x] wp[0x%x,0x%x]", type, cs->segno,
5003 cs->next_blkoff, wp_segno, wp_blkoff);
5004 }
5005
5006 /* Allocate a new section if it's not new. */
5007 if (cs->next_blkoff) {
5008 unsigned int old_segno = cs->segno, old_blkoff = cs->next_blkoff;
5009
5010 f2fs_allocate_new_section(sbi, type, true);
5011 f2fs_notice(sbi, "Assign new section to curseg[%d]: "
5012 "[0x%x,0x%x] -> [0x%x,0x%x]",
5013 type, old_segno, old_blkoff,
5014 cs->segno, cs->next_blkoff);
5015 }
5016
5017 /* check consistency of the zone curseg pointed to */
5018 if (check_zone_write_pointer(sbi, zbd, &zone))
5019 return -EIO;
5020
5021 /* check newly assigned zone */
5022 cs_section = GET_SEC_FROM_SEG(sbi, cs->segno);
5023 cs_zone_block = START_BLOCK(sbi, GET_SEG_FROM_SEC(sbi, cs_section));
5024
5025 zbd = get_target_zoned_dev(sbi, cs_zone_block);
5026 if (!zbd)
5027 return 0;
5028
5029 zone_sector = (sector_t)(cs_zone_block - zbd->start_blk)
5030 << log_sectors_per_block;
5031 err = blkdev_report_zones(zbd->bdev, zone_sector, 1,
5032 report_one_zone_cb, &zone);
5033 if (err != 1) {
5034 f2fs_err(sbi, "Report zone failed: %s errno=(%d)",
5035 zbd->path, err);
5036 return err;
5037 }
5038
5039 if (zone.type != BLK_ZONE_TYPE_SEQWRITE_REQ)
5040 return 0;
5041
5042 if (zone.wp != zone.start) {
5043 f2fs_notice(sbi,
5044 "New zone for curseg[%d] is not yet discarded. "
5045 "Reset the zone: curseg[0x%x,0x%x]",
5046 type, cs->segno, cs->next_blkoff);
5047 err = __f2fs_issue_discard_zone(sbi, zbd->bdev, cs_zone_block,
5048 zone.len >> log_sectors_per_block);
5049 if (err) {
5050 f2fs_err(sbi, "Discard zone failed: %s (errno=%d)",
5051 zbd->path, err);
5052 return err;
5053 }
5054 }
5055
5056 return 0;
5057}
5058
5059int f2fs_fix_curseg_write_pointer(struct f2fs_sb_info *sbi)
5060{
5061 int i, ret;
5062
5063 for (i = 0; i < NR_PERSISTENT_LOG; i++) {
5064 ret = fix_curseg_write_pointer(sbi, i);
5065 if (ret)
5066 return ret;
5067 }
5068
5069 return 0;
5070}
5071
5072struct check_zone_write_pointer_args {
5073 struct f2fs_sb_info *sbi;
5074 struct f2fs_dev_info *fdev;
5075};
5076
5077static int check_zone_write_pointer_cb(struct blk_zone *zone, unsigned int idx,
5078 void *data)
5079{
5080 struct check_zone_write_pointer_args *args;
5081
5082 args = (struct check_zone_write_pointer_args *)data;
5083
5084 return check_zone_write_pointer(args->sbi, args->fdev, zone);
5085}
5086
5087int f2fs_check_write_pointer(struct f2fs_sb_info *sbi)
5088{
5089 int i, ret;
5090 struct check_zone_write_pointer_args args;
5091
5092 for (i = 0; i < sbi->s_ndevs; i++) {
5093 if (!bdev_is_zoned(FDEV(i).bdev))
5094 continue;
5095
5096 args.sbi = sbi;
5097 args.fdev = &FDEV(i);
5098 ret = blkdev_report_zones(FDEV(i).bdev, 0, BLK_ALL_ZONES,
5099 check_zone_write_pointer_cb, &args);
5100 if (ret < 0)
5101 return ret;
5102 }
5103
5104 return 0;
5105}
5106
5107/*
5108 * Return the number of usable blocks in a segment. The number of blocks
5109 * returned is always equal to the number of blocks in a segment for
5110 * segments fully contained within a sequential zone capacity or a
5111 * conventional zone. For segments partially contained in a sequential
5112 * zone capacity, the number of usable blocks up to the zone capacity
5113 * is returned. 0 is returned in all other cases.
5114 */
5115static inline unsigned int f2fs_usable_zone_blks_in_seg(
5116 struct f2fs_sb_info *sbi, unsigned int segno)
5117{
5118 block_t seg_start, sec_start_blkaddr, sec_cap_blkaddr;
5119 unsigned int secno;
5120
5121 if (!sbi->unusable_blocks_per_sec)
5122 return sbi->blocks_per_seg;
5123
5124 secno = GET_SEC_FROM_SEG(sbi, segno);
5125 seg_start = START_BLOCK(sbi, segno);
5126 sec_start_blkaddr = START_BLOCK(sbi, GET_SEG_FROM_SEC(sbi, secno));
5127 sec_cap_blkaddr = sec_start_blkaddr + CAP_BLKS_PER_SEC(sbi);
5128
5129 /*
5130 * If segment starts before zone capacity and spans beyond
5131 * zone capacity, then usable blocks are from seg start to
5132 * zone capacity. If the segment starts after the zone capacity,
5133 * then there are no usable blocks.
5134 */
5135 if (seg_start >= sec_cap_blkaddr)
5136 return 0;
5137 if (seg_start + sbi->blocks_per_seg > sec_cap_blkaddr)
5138 return sec_cap_blkaddr - seg_start;
5139
5140 return sbi->blocks_per_seg;
5141}
5142#else
5143int f2fs_fix_curseg_write_pointer(struct f2fs_sb_info *sbi)
5144{
5145 return 0;
5146}
5147
5148int f2fs_check_write_pointer(struct f2fs_sb_info *sbi)
5149{
5150 return 0;
5151}
5152
5153static inline unsigned int f2fs_usable_zone_blks_in_seg(struct f2fs_sb_info *sbi,
5154 unsigned int segno)
5155{
5156 return 0;
5157}
5158
5159#endif
5160unsigned int f2fs_usable_blks_in_seg(struct f2fs_sb_info *sbi,
5161 unsigned int segno)
5162{
5163 if (f2fs_sb_has_blkzoned(sbi))
5164 return f2fs_usable_zone_blks_in_seg(sbi, segno);
5165
5166 return sbi->blocks_per_seg;
5167}
5168
5169unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi,
5170 unsigned int segno)
5171{
5172 if (f2fs_sb_has_blkzoned(sbi))
5173 return CAP_SEGS_PER_SEC(sbi);
5174
5175 return sbi->segs_per_sec;
5176}
5177
5178/*
5179 * Update min, max modified time for cost-benefit GC algorithm
5180 */
5181static void init_min_max_mtime(struct f2fs_sb_info *sbi)
5182{
5183 struct sit_info *sit_i = SIT_I(sbi);
5184 unsigned int segno;
5185
5186 down_write(&sit_i->sentry_lock);
5187
5188 sit_i->min_mtime = ULLONG_MAX;
5189
5190 for (segno = 0; segno < MAIN_SEGS(sbi); segno += sbi->segs_per_sec) {
5191 unsigned int i;
5192 unsigned long long mtime = 0;
5193
5194 for (i = 0; i < sbi->segs_per_sec; i++)
5195 mtime += get_seg_entry(sbi, segno + i)->mtime;
5196
5197 mtime = div_u64(mtime, sbi->segs_per_sec);
5198
5199 if (sit_i->min_mtime > mtime)
5200 sit_i->min_mtime = mtime;
5201 }
5202 sit_i->max_mtime = get_mtime(sbi, false);
5203 sit_i->dirty_max_mtime = 0;
5204 up_write(&sit_i->sentry_lock);
5205}
5206
5207int f2fs_build_segment_manager(struct f2fs_sb_info *sbi)
5208{
5209 struct f2fs_super_block *raw_super = F2FS_RAW_SUPER(sbi);
5210 struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
5211 struct f2fs_sm_info *sm_info;
5212 int err;
5213
5214 sm_info = f2fs_kzalloc(sbi, sizeof(struct f2fs_sm_info), GFP_KERNEL);
5215 if (!sm_info)
5216 return -ENOMEM;
5217
5218 /* init sm info */
5219 sbi->sm_info = sm_info;
5220 sm_info->seg0_blkaddr = le32_to_cpu(raw_super->segment0_blkaddr);
5221 sm_info->main_blkaddr = le32_to_cpu(raw_super->main_blkaddr);
5222 sm_info->segment_count = le32_to_cpu(raw_super->segment_count);
5223 sm_info->reserved_segments = le32_to_cpu(ckpt->rsvd_segment_count);
5224 sm_info->ovp_segments = le32_to_cpu(ckpt->overprov_segment_count);
5225 sm_info->main_segments = le32_to_cpu(raw_super->segment_count_main);
5226 sm_info->ssa_blkaddr = le32_to_cpu(raw_super->ssa_blkaddr);
5227 sm_info->rec_prefree_segments = sm_info->main_segments *
5228 DEF_RECLAIM_PREFREE_SEGMENTS / 100;
5229 if (sm_info->rec_prefree_segments > DEF_MAX_RECLAIM_PREFREE_SEGMENTS)
5230 sm_info->rec_prefree_segments = DEF_MAX_RECLAIM_PREFREE_SEGMENTS;
5231
5232 if (!f2fs_lfs_mode(sbi))
5233 sm_info->ipu_policy = BIT(F2FS_IPU_FSYNC);
5234 sm_info->min_ipu_util = DEF_MIN_IPU_UTIL;
5235 sm_info->min_fsync_blocks = DEF_MIN_FSYNC_BLOCKS;
5236 sm_info->min_seq_blocks = sbi->blocks_per_seg;
5237 sm_info->min_hot_blocks = DEF_MIN_HOT_BLOCKS;
5238 sm_info->min_ssr_sections = reserved_sections(sbi);
5239
5240 INIT_LIST_HEAD(&sm_info->sit_entry_set);
5241
5242 init_f2fs_rwsem(&sm_info->curseg_lock);
5243
5244 err = f2fs_create_flush_cmd_control(sbi);
5245 if (err)
5246 return err;
5247
5248 err = create_discard_cmd_control(sbi);
5249 if (err)
5250 return err;
5251
5252 err = build_sit_info(sbi);
5253 if (err)
5254 return err;
5255 err = build_free_segmap(sbi);
5256 if (err)
5257 return err;
5258 err = build_curseg(sbi);
5259 if (err)
5260 return err;
5261
5262 /* reinit free segmap based on SIT */
5263 err = build_sit_entries(sbi);
5264 if (err)
5265 return err;
5266
5267 init_free_segmap(sbi);
5268 err = build_dirty_segmap(sbi);
5269 if (err)
5270 return err;
5271
5272 err = sanity_check_curseg(sbi);
5273 if (err)
5274 return err;
5275
5276 init_min_max_mtime(sbi);
5277 return 0;
5278}
5279
5280static void discard_dirty_segmap(struct f2fs_sb_info *sbi,
5281 enum dirty_type dirty_type)
5282{
5283 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
5284
5285 mutex_lock(&dirty_i->seglist_lock);
5286 kvfree(dirty_i->dirty_segmap[dirty_type]);
5287 dirty_i->nr_dirty[dirty_type] = 0;
5288 mutex_unlock(&dirty_i->seglist_lock);
5289}
5290
5291static void destroy_victim_secmap(struct f2fs_sb_info *sbi)
5292{
5293 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
5294
5295 kvfree(dirty_i->pinned_secmap);
5296 kvfree(dirty_i->victim_secmap);
5297}
5298
5299static void destroy_dirty_segmap(struct f2fs_sb_info *sbi)
5300{
5301 struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
5302 int i;
5303
5304 if (!dirty_i)
5305 return;
5306
5307 /* discard pre-free/dirty segments list */
5308 for (i = 0; i < NR_DIRTY_TYPE; i++)
5309 discard_dirty_segmap(sbi, i);
5310
5311 if (__is_large_section(sbi)) {
5312 mutex_lock(&dirty_i->seglist_lock);
5313 kvfree(dirty_i->dirty_secmap);
5314 mutex_unlock(&dirty_i->seglist_lock);
5315 }
5316
5317 destroy_victim_secmap(sbi);
5318 SM_I(sbi)->dirty_info = NULL;
5319 kfree(dirty_i);
5320}
5321
5322static void destroy_curseg(struct f2fs_sb_info *sbi)
5323{
5324 struct curseg_info *array = SM_I(sbi)->curseg_array;
5325 int i;
5326
5327 if (!array)
5328 return;
5329 SM_I(sbi)->curseg_array = NULL;
5330 for (i = 0; i < NR_CURSEG_TYPE; i++) {
5331 kfree(array[i].sum_blk);
5332 kfree(array[i].journal);
5333 }
5334 kfree(array);
5335}
5336
5337static void destroy_free_segmap(struct f2fs_sb_info *sbi)
5338{
5339 struct free_segmap_info *free_i = SM_I(sbi)->free_info;
5340
5341 if (!free_i)
5342 return;
5343 SM_I(sbi)->free_info = NULL;
5344 kvfree(free_i->free_segmap);
5345 kvfree(free_i->free_secmap);
5346 kfree(free_i);
5347}
5348
5349static void destroy_sit_info(struct f2fs_sb_info *sbi)
5350{
5351 struct sit_info *sit_i = SIT_I(sbi);
5352
5353 if (!sit_i)
5354 return;
5355
5356 if (sit_i->sentries)
5357 kvfree(sit_i->bitmap);
5358 kfree(sit_i->tmp_map);
5359
5360 kvfree(sit_i->sentries);
5361 kvfree(sit_i->sec_entries);
5362 kvfree(sit_i->dirty_sentries_bitmap);
5363
5364 SM_I(sbi)->sit_info = NULL;
5365 kvfree(sit_i->sit_bitmap);
5366#ifdef CONFIG_F2FS_CHECK_FS
5367 kvfree(sit_i->sit_bitmap_mir);
5368 kvfree(sit_i->invalid_segmap);
5369#endif
5370 kfree(sit_i);
5371}
5372
5373void f2fs_destroy_segment_manager(struct f2fs_sb_info *sbi)
5374{
5375 struct f2fs_sm_info *sm_info = SM_I(sbi);
5376
5377 if (!sm_info)
5378 return;
5379 f2fs_destroy_flush_cmd_control(sbi, true);
5380 destroy_discard_cmd_control(sbi);
5381 destroy_dirty_segmap(sbi);
5382 destroy_curseg(sbi);
5383 destroy_free_segmap(sbi);
5384 destroy_sit_info(sbi);
5385 sbi->sm_info = NULL;
5386 kfree(sm_info);
5387}
5388
5389int __init f2fs_create_segment_manager_caches(void)
5390{
5391 discard_entry_slab = f2fs_kmem_cache_create("f2fs_discard_entry",
5392 sizeof(struct discard_entry));
5393 if (!discard_entry_slab)
5394 goto fail;
5395
5396 discard_cmd_slab = f2fs_kmem_cache_create("f2fs_discard_cmd",
5397 sizeof(struct discard_cmd));
5398 if (!discard_cmd_slab)
5399 goto destroy_discard_entry;
5400
5401 sit_entry_set_slab = f2fs_kmem_cache_create("f2fs_sit_entry_set",
5402 sizeof(struct sit_entry_set));
5403 if (!sit_entry_set_slab)
5404 goto destroy_discard_cmd;
5405
5406 revoke_entry_slab = f2fs_kmem_cache_create("f2fs_revoke_entry",
5407 sizeof(struct revoke_entry));
5408 if (!revoke_entry_slab)
5409 goto destroy_sit_entry_set;
5410 return 0;
5411
5412destroy_sit_entry_set:
5413 kmem_cache_destroy(sit_entry_set_slab);
5414destroy_discard_cmd:
5415 kmem_cache_destroy(discard_cmd_slab);
5416destroy_discard_entry:
5417 kmem_cache_destroy(discard_entry_slab);
5418fail:
5419 return -ENOMEM;
5420}
5421
5422void f2fs_destroy_segment_manager_caches(void)
5423{
5424 kmem_cache_destroy(sit_entry_set_slab);
5425 kmem_cache_destroy(discard_cmd_slab);
5426 kmem_cache_destroy(discard_entry_slab);
5427 kmem_cache_destroy(revoke_entry_slab);
5428}