Line data Source code
1 : // SPDX-License-Identifier: GPL-2.0
2 : /*
3 : * Copyright (C) 2007 Oracle. All rights reserved.
4 : */
5 :
6 : #include <linux/sched.h>
7 : #include "ctree.h"
8 : #include "disk-io.h"
9 : #include "print-tree.h"
10 : #include "transaction.h"
11 : #include "locking.h"
12 : #include "accessors.h"
13 : #include "messages.h"
14 : #include "delalloc-space.h"
15 : #include "subpage.h"
16 : #include "defrag.h"
17 : #include "file-item.h"
18 : #include "super.h"
19 :
20 : static struct kmem_cache *btrfs_inode_defrag_cachep;
21 :
22 : /*
23 : * When auto defrag is enabled we queue up these defrag structs to remember
24 : * which inodes need defragging passes.
25 : */
26 : struct inode_defrag {
27 : struct rb_node rb_node;
28 : /* Inode number */
29 : u64 ino;
30 : /*
31 : * Transid where the defrag was added, we search for extents newer than
32 : * this.
33 : */
34 : u64 transid;
35 :
36 : /* Root objectid */
37 : u64 root;
38 :
39 : /*
40 : * The extent size threshold for autodefrag.
41 : *
42 : * This value is different for compressed/non-compressed extents, thus
43 : * needs to be passed from higher layer.
44 : * (aka, inode_should_defrag())
45 : */
46 : u32 extent_thresh;
47 : };
48 :
49 : static int __compare_inode_defrag(struct inode_defrag *defrag1,
50 : struct inode_defrag *defrag2)
51 : {
52 0 : if (defrag1->root > defrag2->root)
53 : return 1;
54 0 : else if (defrag1->root < defrag2->root)
55 : return -1;
56 0 : else if (defrag1->ino > defrag2->ino)
57 : return 1;
58 0 : else if (defrag1->ino < defrag2->ino)
59 : return -1;
60 : else
61 : return 0;
62 : }
63 :
64 : /*
65 : * Pop a record for an inode into the defrag tree. The lock must be held
66 : * already.
67 : *
68 : * If you're inserting a record for an older transid than an existing record,
69 : * the transid already in the tree is lowered.
70 : *
71 : * If an existing record is found the defrag item you pass in is freed.
72 : */
73 0 : static int __btrfs_add_inode_defrag(struct btrfs_inode *inode,
74 : struct inode_defrag *defrag)
75 : {
76 0 : struct btrfs_fs_info *fs_info = inode->root->fs_info;
77 0 : struct inode_defrag *entry;
78 0 : struct rb_node **p;
79 0 : struct rb_node *parent = NULL;
80 0 : int ret;
81 :
82 0 : p = &fs_info->defrag_inodes.rb_node;
83 0 : while (*p) {
84 0 : parent = *p;
85 0 : entry = rb_entry(parent, struct inode_defrag, rb_node);
86 :
87 0 : ret = __compare_inode_defrag(defrag, entry);
88 : if (ret < 0)
89 0 : p = &parent->rb_left;
90 0 : else if (ret > 0)
91 0 : p = &parent->rb_right;
92 : else {
93 : /*
94 : * If we're reinserting an entry for an old defrag run,
95 : * make sure to lower the transid of our existing
96 : * record.
97 : */
98 0 : if (defrag->transid < entry->transid)
99 0 : entry->transid = defrag->transid;
100 0 : entry->extent_thresh = min(defrag->extent_thresh,
101 : entry->extent_thresh);
102 0 : return -EEXIST;
103 : }
104 : }
105 0 : set_bit(BTRFS_INODE_IN_DEFRAG, &inode->runtime_flags);
106 0 : rb_link_node(&defrag->rb_node, parent, p);
107 0 : rb_insert_color(&defrag->rb_node, &fs_info->defrag_inodes);
108 0 : return 0;
109 : }
110 :
111 0 : static inline int __need_auto_defrag(struct btrfs_fs_info *fs_info)
112 : {
113 0 : if (!btrfs_test_opt(fs_info, AUTO_DEFRAG))
114 : return 0;
115 :
116 0 : if (btrfs_fs_closing(fs_info))
117 0 : return 0;
118 :
119 : return 1;
120 : }
121 :
122 : /*
123 : * Insert a defrag record for this inode if auto defrag is enabled.
124 : */
125 0 : int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans,
126 : struct btrfs_inode *inode, u32 extent_thresh)
127 : {
128 0 : struct btrfs_root *root = inode->root;
129 0 : struct btrfs_fs_info *fs_info = root->fs_info;
130 0 : struct inode_defrag *defrag;
131 0 : u64 transid;
132 0 : int ret;
133 :
134 0 : if (!__need_auto_defrag(fs_info))
135 : return 0;
136 :
137 0 : if (test_bit(BTRFS_INODE_IN_DEFRAG, &inode->runtime_flags))
138 : return 0;
139 :
140 0 : if (trans)
141 0 : transid = trans->transid;
142 : else
143 0 : transid = inode->root->last_trans;
144 :
145 0 : defrag = kmem_cache_zalloc(btrfs_inode_defrag_cachep, GFP_NOFS);
146 0 : if (!defrag)
147 : return -ENOMEM;
148 :
149 0 : defrag->ino = btrfs_ino(inode);
150 0 : defrag->transid = transid;
151 0 : defrag->root = root->root_key.objectid;
152 0 : defrag->extent_thresh = extent_thresh;
153 :
154 0 : spin_lock(&fs_info->defrag_inodes_lock);
155 0 : if (!test_bit(BTRFS_INODE_IN_DEFRAG, &inode->runtime_flags)) {
156 : /*
157 : * If we set IN_DEFRAG flag and evict the inode from memory,
158 : * and then re-read this inode, this new inode doesn't have
159 : * IN_DEFRAG flag. At the case, we may find the existed defrag.
160 : */
161 0 : ret = __btrfs_add_inode_defrag(inode, defrag);
162 0 : if (ret)
163 0 : kmem_cache_free(btrfs_inode_defrag_cachep, defrag);
164 : } else {
165 0 : kmem_cache_free(btrfs_inode_defrag_cachep, defrag);
166 : }
167 0 : spin_unlock(&fs_info->defrag_inodes_lock);
168 0 : return 0;
169 : }
170 :
171 : /*
172 : * Pick the defragable inode that we want, if it doesn't exist, we will get the
173 : * next one.
174 : */
175 0 : static struct inode_defrag *btrfs_pick_defrag_inode(
176 : struct btrfs_fs_info *fs_info, u64 root, u64 ino)
177 : {
178 0 : struct inode_defrag *entry = NULL;
179 0 : struct inode_defrag tmp;
180 0 : struct rb_node *p;
181 0 : struct rb_node *parent = NULL;
182 0 : int ret;
183 :
184 0 : tmp.ino = ino;
185 0 : tmp.root = root;
186 :
187 0 : spin_lock(&fs_info->defrag_inodes_lock);
188 0 : p = fs_info->defrag_inodes.rb_node;
189 0 : while (p) {
190 0 : parent = p;
191 0 : entry = rb_entry(parent, struct inode_defrag, rb_node);
192 :
193 0 : ret = __compare_inode_defrag(&tmp, entry);
194 : if (ret < 0)
195 0 : p = parent->rb_left;
196 0 : else if (ret > 0)
197 0 : p = parent->rb_right;
198 : else
199 0 : goto out;
200 : }
201 :
202 0 : if (parent && __compare_inode_defrag(&tmp, entry) > 0) {
203 0 : parent = rb_next(parent);
204 0 : if (parent)
205 : entry = rb_entry(parent, struct inode_defrag, rb_node);
206 : else
207 : entry = NULL;
208 : }
209 0 : out:
210 0 : if (entry)
211 0 : rb_erase(parent, &fs_info->defrag_inodes);
212 0 : spin_unlock(&fs_info->defrag_inodes_lock);
213 0 : return entry;
214 : }
215 :
216 0 : void btrfs_cleanup_defrag_inodes(struct btrfs_fs_info *fs_info)
217 : {
218 0 : struct inode_defrag *defrag;
219 0 : struct rb_node *node;
220 :
221 0 : spin_lock(&fs_info->defrag_inodes_lock);
222 0 : node = rb_first(&fs_info->defrag_inodes);
223 0 : while (node) {
224 0 : rb_erase(node, &fs_info->defrag_inodes);
225 0 : defrag = rb_entry(node, struct inode_defrag, rb_node);
226 0 : kmem_cache_free(btrfs_inode_defrag_cachep, defrag);
227 :
228 0 : cond_resched_lock(&fs_info->defrag_inodes_lock);
229 :
230 0 : node = rb_first(&fs_info->defrag_inodes);
231 : }
232 0 : spin_unlock(&fs_info->defrag_inodes_lock);
233 0 : }
234 :
235 : #define BTRFS_DEFRAG_BATCH 1024
236 :
237 0 : static int __btrfs_run_defrag_inode(struct btrfs_fs_info *fs_info,
238 : struct inode_defrag *defrag)
239 : {
240 0 : struct btrfs_root *inode_root;
241 0 : struct inode *inode;
242 0 : struct btrfs_ioctl_defrag_range_args range;
243 0 : int ret = 0;
244 0 : u64 cur = 0;
245 :
246 0 : again:
247 0 : if (test_bit(BTRFS_FS_STATE_REMOUNTING, &fs_info->fs_state))
248 0 : goto cleanup;
249 0 : if (!__need_auto_defrag(fs_info))
250 0 : goto cleanup;
251 :
252 : /* Get the inode */
253 0 : inode_root = btrfs_get_fs_root(fs_info, defrag->root, true);
254 0 : if (IS_ERR(inode_root)) {
255 0 : ret = PTR_ERR(inode_root);
256 0 : goto cleanup;
257 : }
258 :
259 0 : inode = btrfs_iget(fs_info->sb, defrag->ino, inode_root);
260 0 : btrfs_put_root(inode_root);
261 0 : if (IS_ERR(inode)) {
262 0 : ret = PTR_ERR(inode);
263 0 : goto cleanup;
264 : }
265 :
266 0 : if (cur >= i_size_read(inode)) {
267 0 : iput(inode);
268 0 : goto cleanup;
269 : }
270 :
271 : /* Do a chunk of defrag */
272 0 : clear_bit(BTRFS_INODE_IN_DEFRAG, &BTRFS_I(inode)->runtime_flags);
273 0 : memset(&range, 0, sizeof(range));
274 0 : range.len = (u64)-1;
275 0 : range.start = cur;
276 0 : range.extent_thresh = defrag->extent_thresh;
277 :
278 0 : sb_start_write(fs_info->sb);
279 0 : ret = btrfs_defrag_file(inode, NULL, &range, defrag->transid,
280 : BTRFS_DEFRAG_BATCH);
281 0 : sb_end_write(fs_info->sb);
282 0 : iput(inode);
283 :
284 0 : if (ret < 0)
285 0 : goto cleanup;
286 :
287 0 : cur = max(cur + fs_info->sectorsize, range.start);
288 0 : goto again;
289 :
290 0 : cleanup:
291 0 : kmem_cache_free(btrfs_inode_defrag_cachep, defrag);
292 0 : return ret;
293 : }
294 :
295 : /*
296 : * Run through the list of inodes in the FS that need defragging.
297 : */
298 0 : int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info)
299 : {
300 0 : struct inode_defrag *defrag;
301 0 : u64 first_ino = 0;
302 0 : u64 root_objectid = 0;
303 :
304 0 : atomic_inc(&fs_info->defrag_running);
305 0 : while (1) {
306 : /* Pause the auto defragger. */
307 0 : if (test_bit(BTRFS_FS_STATE_REMOUNTING, &fs_info->fs_state))
308 : break;
309 :
310 0 : if (!__need_auto_defrag(fs_info))
311 : break;
312 :
313 : /* find an inode to defrag */
314 0 : defrag = btrfs_pick_defrag_inode(fs_info, root_objectid, first_ino);
315 0 : if (!defrag) {
316 0 : if (root_objectid || first_ino) {
317 0 : root_objectid = 0;
318 0 : first_ino = 0;
319 0 : continue;
320 : } else {
321 : break;
322 : }
323 : }
324 :
325 0 : first_ino = defrag->ino + 1;
326 0 : root_objectid = defrag->root;
327 :
328 0 : __btrfs_run_defrag_inode(fs_info, defrag);
329 : }
330 0 : atomic_dec(&fs_info->defrag_running);
331 :
332 : /*
333 : * During unmount, we use the transaction_wait queue to wait for the
334 : * defragger to stop.
335 : */
336 0 : wake_up(&fs_info->transaction_wait);
337 0 : return 0;
338 : }
339 :
340 : /*
341 : * Defrag all the leaves in a given btree.
342 : * Read all the leaves and try to get key order to
343 : * better reflect disk order
344 : */
345 :
346 0 : int btrfs_defrag_leaves(struct btrfs_trans_handle *trans,
347 : struct btrfs_root *root)
348 : {
349 0 : struct btrfs_path *path = NULL;
350 0 : struct btrfs_key key;
351 0 : int ret = 0;
352 0 : int wret;
353 0 : int level;
354 0 : int next_key_ret = 0;
355 0 : u64 last_ret = 0;
356 :
357 0 : if (!test_bit(BTRFS_ROOT_SHAREABLE, &root->state))
358 0 : goto out;
359 :
360 0 : path = btrfs_alloc_path();
361 0 : if (!path) {
362 0 : ret = -ENOMEM;
363 0 : goto out;
364 : }
365 :
366 0 : level = btrfs_header_level(root->node);
367 :
368 0 : if (level == 0)
369 0 : goto out;
370 :
371 0 : if (root->defrag_progress.objectid == 0) {
372 0 : struct extent_buffer *root_node;
373 0 : u32 nritems;
374 :
375 0 : root_node = btrfs_lock_root_node(root);
376 0 : nritems = btrfs_header_nritems(root_node);
377 0 : root->defrag_max.objectid = 0;
378 : /* from above we know this is not a leaf */
379 0 : btrfs_node_key_to_cpu(root_node, &root->defrag_max,
380 0 : nritems - 1);
381 0 : btrfs_tree_unlock(root_node);
382 0 : free_extent_buffer(root_node);
383 0 : memset(&key, 0, sizeof(key));
384 : } else {
385 0 : memcpy(&key, &root->defrag_progress, sizeof(key));
386 : }
387 :
388 0 : path->keep_locks = 1;
389 :
390 0 : ret = btrfs_search_forward(root, &key, path, BTRFS_OLDEST_GENERATION);
391 0 : if (ret < 0)
392 0 : goto out;
393 0 : if (ret > 0) {
394 0 : ret = 0;
395 0 : goto out;
396 : }
397 0 : btrfs_release_path(path);
398 : /*
399 : * We don't need a lock on a leaf. btrfs_realloc_node() will lock all
400 : * leafs from path->nodes[1], so set lowest_level to 1 to avoid later
401 : * a deadlock (attempting to write lock an already write locked leaf).
402 : */
403 0 : path->lowest_level = 1;
404 0 : wret = btrfs_search_slot(trans, root, &key, path, 0, 1);
405 :
406 0 : if (wret < 0) {
407 0 : ret = wret;
408 0 : goto out;
409 : }
410 0 : if (!path->nodes[1]) {
411 0 : ret = 0;
412 0 : goto out;
413 : }
414 : /*
415 : * The node at level 1 must always be locked when our path has
416 : * keep_locks set and lowest_level is 1, regardless of the value of
417 : * path->slots[1].
418 : */
419 0 : BUG_ON(path->locks[1] == 0);
420 0 : ret = btrfs_realloc_node(trans, root,
421 : path->nodes[1], 0,
422 : &last_ret,
423 : &root->defrag_progress);
424 0 : if (ret) {
425 0 : WARN_ON(ret == -EAGAIN);
426 0 : goto out;
427 : }
428 : /*
429 : * Now that we reallocated the node we can find the next key. Note that
430 : * btrfs_find_next_key() can release our path and do another search
431 : * without COWing, this is because even with path->keep_locks = 1,
432 : * btrfs_search_slot() / ctree.c:unlock_up() does not keeps a lock on a
433 : * node when path->slots[node_level - 1] does not point to the last
434 : * item or a slot beyond the last item (ctree.c:unlock_up()). Therefore
435 : * we search for the next key after reallocating our node.
436 : */
437 0 : path->slots[1] = btrfs_header_nritems(path->nodes[1]);
438 0 : next_key_ret = btrfs_find_next_key(root, path, &key, 1,
439 : BTRFS_OLDEST_GENERATION);
440 0 : if (next_key_ret == 0) {
441 0 : memcpy(&root->defrag_progress, &key, sizeof(key));
442 0 : ret = -EAGAIN;
443 : }
444 0 : out:
445 0 : btrfs_free_path(path);
446 0 : if (ret == -EAGAIN) {
447 0 : if (root->defrag_max.objectid > root->defrag_progress.objectid)
448 0 : goto done;
449 0 : if (root->defrag_max.type > root->defrag_progress.type)
450 0 : goto done;
451 0 : if (root->defrag_max.offset > root->defrag_progress.offset)
452 0 : goto done;
453 : ret = 0;
454 : }
455 0 : done:
456 0 : if (ret != -EAGAIN)
457 0 : memset(&root->defrag_progress, 0,
458 : sizeof(root->defrag_progress));
459 :
460 0 : return ret;
461 : }
462 :
463 : /*
464 : * Defrag specific helper to get an extent map.
465 : *
466 : * Differences between this and btrfs_get_extent() are:
467 : *
468 : * - No extent_map will be added to inode->extent_tree
469 : * To reduce memory usage in the long run.
470 : *
471 : * - Extra optimization to skip file extents older than @newer_than
472 : * By using btrfs_search_forward() we can skip entire file ranges that
473 : * have extents created in past transactions, because btrfs_search_forward()
474 : * will not visit leaves and nodes with a generation smaller than given
475 : * minimal generation threshold (@newer_than).
476 : *
477 : * Return valid em if we find a file extent matching the requirement.
478 : * Return NULL if we can not find a file extent matching the requirement.
479 : *
480 : * Return ERR_PTR() for error.
481 : */
482 0 : static struct extent_map *defrag_get_extent(struct btrfs_inode *inode,
483 : u64 start, u64 newer_than)
484 : {
485 0 : struct btrfs_root *root = inode->root;
486 0 : struct btrfs_file_extent_item *fi;
487 0 : struct btrfs_path path = { 0 };
488 0 : struct extent_map *em;
489 0 : struct btrfs_key key;
490 0 : u64 ino = btrfs_ino(inode);
491 0 : int ret;
492 :
493 0 : em = alloc_extent_map();
494 0 : if (!em) {
495 0 : ret = -ENOMEM;
496 0 : goto err;
497 : }
498 :
499 0 : key.objectid = ino;
500 0 : key.type = BTRFS_EXTENT_DATA_KEY;
501 0 : key.offset = start;
502 :
503 0 : if (newer_than) {
504 0 : ret = btrfs_search_forward(root, &key, &path, newer_than);
505 0 : if (ret < 0)
506 0 : goto err;
507 : /* Can't find anything newer */
508 0 : if (ret > 0)
509 0 : goto not_found;
510 : } else {
511 0 : ret = btrfs_search_slot(NULL, root, &key, &path, 0, 0);
512 0 : if (ret < 0)
513 0 : goto err;
514 : }
515 0 : if (path.slots[0] >= btrfs_header_nritems(path.nodes[0])) {
516 : /*
517 : * If btrfs_search_slot() makes path to point beyond nritems,
518 : * we should not have an empty leaf, as this inode must at
519 : * least have its INODE_ITEM.
520 : */
521 0 : ASSERT(btrfs_header_nritems(path.nodes[0]));
522 0 : path.slots[0] = btrfs_header_nritems(path.nodes[0]) - 1;
523 : }
524 0 : btrfs_item_key_to_cpu(path.nodes[0], &key, path.slots[0]);
525 : /* Perfect match, no need to go one slot back */
526 0 : if (key.objectid == ino && key.type == BTRFS_EXTENT_DATA_KEY &&
527 0 : key.offset == start)
528 0 : goto iterate;
529 :
530 : /* We didn't find a perfect match, needs to go one slot back */
531 0 : if (path.slots[0] > 0) {
532 0 : btrfs_item_key_to_cpu(path.nodes[0], &key, path.slots[0]);
533 0 : if (key.objectid == ino && key.type == BTRFS_EXTENT_DATA_KEY)
534 0 : path.slots[0]--;
535 : }
536 :
537 0 : iterate:
538 : /* Iterate through the path to find a file extent covering @start */
539 0 : while (true) {
540 0 : u64 extent_end;
541 :
542 0 : if (path.slots[0] >= btrfs_header_nritems(path.nodes[0]))
543 0 : goto next;
544 :
545 0 : btrfs_item_key_to_cpu(path.nodes[0], &key, path.slots[0]);
546 :
547 : /*
548 : * We may go one slot back to INODE_REF/XATTR item, then
549 : * need to go forward until we reach an EXTENT_DATA.
550 : * But we should still has the correct ino as key.objectid.
551 : */
552 0 : if (WARN_ON(key.objectid < ino) || key.type < BTRFS_EXTENT_DATA_KEY)
553 0 : goto next;
554 :
555 : /* It's beyond our target range, definitely not extent found */
556 0 : if (key.objectid > ino || key.type > BTRFS_EXTENT_DATA_KEY)
557 0 : goto not_found;
558 :
559 : /*
560 : * | |<- File extent ->|
561 : * \- start
562 : *
563 : * This means there is a hole between start and key.offset.
564 : */
565 0 : if (key.offset > start) {
566 0 : em->start = start;
567 0 : em->orig_start = start;
568 0 : em->block_start = EXTENT_MAP_HOLE;
569 0 : em->len = key.offset - start;
570 0 : break;
571 : }
572 :
573 0 : fi = btrfs_item_ptr(path.nodes[0], path.slots[0],
574 : struct btrfs_file_extent_item);
575 0 : extent_end = btrfs_file_extent_end(&path);
576 :
577 : /*
578 : * |<- file extent ->| |
579 : * \- start
580 : *
581 : * We haven't reached start, search next slot.
582 : */
583 0 : if (extent_end <= start)
584 0 : goto next;
585 :
586 : /* Now this extent covers @start, convert it to em */
587 0 : btrfs_extent_item_to_extent_map(inode, &path, fi, em);
588 0 : break;
589 0 : next:
590 0 : ret = btrfs_next_item(root, &path);
591 0 : if (ret < 0)
592 0 : goto err;
593 0 : if (ret > 0)
594 0 : goto not_found;
595 : }
596 0 : btrfs_release_path(&path);
597 0 : return em;
598 :
599 0 : not_found:
600 0 : btrfs_release_path(&path);
601 0 : free_extent_map(em);
602 0 : return NULL;
603 :
604 0 : err:
605 0 : btrfs_release_path(&path);
606 0 : free_extent_map(em);
607 0 : return ERR_PTR(ret);
608 : }
609 :
610 0 : static struct extent_map *defrag_lookup_extent(struct inode *inode, u64 start,
611 : u64 newer_than, bool locked)
612 : {
613 0 : struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
614 0 : struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
615 0 : struct extent_map *em;
616 0 : const u32 sectorsize = BTRFS_I(inode)->root->fs_info->sectorsize;
617 :
618 : /*
619 : * Hopefully we have this extent in the tree already, try without the
620 : * full extent lock.
621 : */
622 0 : read_lock(&em_tree->lock);
623 0 : em = lookup_extent_mapping(em_tree, start, sectorsize);
624 0 : read_unlock(&em_tree->lock);
625 :
626 : /*
627 : * We can get a merged extent, in that case, we need to re-search
628 : * tree to get the original em for defrag.
629 : *
630 : * If @newer_than is 0 or em::generation < newer_than, we can trust
631 : * this em, as either we don't care about the generation, or the
632 : * merged extent map will be rejected anyway.
633 : */
634 0 : if (em && test_bit(EXTENT_FLAG_MERGED, &em->flags) &&
635 0 : newer_than && em->generation >= newer_than) {
636 0 : free_extent_map(em);
637 0 : em = NULL;
638 : }
639 :
640 0 : if (!em) {
641 0 : struct extent_state *cached = NULL;
642 0 : u64 end = start + sectorsize - 1;
643 :
644 : /* Get the big lock and read metadata off disk. */
645 0 : if (!locked)
646 0 : lock_extent(io_tree, start, end, &cached);
647 0 : em = defrag_get_extent(BTRFS_I(inode), start, newer_than);
648 0 : if (!locked)
649 0 : unlock_extent(io_tree, start, end, &cached);
650 :
651 0 : if (IS_ERR(em))
652 0 : return NULL;
653 : }
654 :
655 : return em;
656 : }
657 :
658 0 : static u32 get_extent_max_capacity(const struct btrfs_fs_info *fs_info,
659 : const struct extent_map *em)
660 : {
661 0 : if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags))
662 : return BTRFS_MAX_COMPRESSED;
663 0 : return fs_info->max_extent_size;
664 : }
665 :
666 0 : static bool defrag_check_next_extent(struct inode *inode, struct extent_map *em,
667 : u32 extent_thresh, u64 newer_than, bool locked)
668 : {
669 0 : struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
670 0 : struct extent_map *next;
671 0 : bool ret = false;
672 :
673 : /* This is the last extent */
674 0 : if (em->start + em->len >= i_size_read(inode))
675 : return false;
676 :
677 : /*
678 : * Here we need to pass @newer_then when checking the next extent, or
679 : * we will hit a case we mark current extent for defrag, but the next
680 : * one will not be a target.
681 : * This will just cause extra IO without really reducing the fragments.
682 : */
683 0 : next = defrag_lookup_extent(inode, em->start + em->len, newer_than, locked);
684 : /* No more em or hole */
685 0 : if (!next || next->block_start >= EXTENT_MAP_LAST_BYTE)
686 0 : goto out;
687 0 : if (test_bit(EXTENT_FLAG_PREALLOC, &next->flags))
688 0 : goto out;
689 : /*
690 : * If the next extent is at its max capacity, defragging current extent
691 : * makes no sense, as the total number of extents won't change.
692 : */
693 0 : if (next->len >= get_extent_max_capacity(fs_info, em))
694 0 : goto out;
695 : /* Skip older extent */
696 0 : if (next->generation < newer_than)
697 0 : goto out;
698 : /* Also check extent size */
699 0 : if (next->len >= extent_thresh)
700 0 : goto out;
701 :
702 : ret = true;
703 0 : out:
704 0 : free_extent_map(next);
705 0 : return ret;
706 : }
707 :
708 : /*
709 : * Prepare one page to be defragged.
710 : *
711 : * This will ensure:
712 : *
713 : * - Returned page is locked and has been set up properly.
714 : * - No ordered extent exists in the page.
715 : * - The page is uptodate.
716 : *
717 : * NOTE: Caller should also wait for page writeback after the cluster is
718 : * prepared, here we don't do writeback wait for each page.
719 : */
720 0 : static struct page *defrag_prepare_one_page(struct btrfs_inode *inode, pgoff_t index)
721 : {
722 0 : struct address_space *mapping = inode->vfs_inode.i_mapping;
723 0 : gfp_t mask = btrfs_alloc_write_mask(mapping);
724 0 : u64 page_start = (u64)index << PAGE_SHIFT;
725 0 : u64 page_end = page_start + PAGE_SIZE - 1;
726 0 : struct extent_state *cached_state = NULL;
727 0 : struct page *page;
728 0 : int ret;
729 :
730 : again:
731 0 : page = find_or_create_page(mapping, index, mask);
732 0 : if (!page)
733 : return ERR_PTR(-ENOMEM);
734 :
735 : /*
736 : * Since we can defragment files opened read-only, we can encounter
737 : * transparent huge pages here (see CONFIG_READ_ONLY_THP_FOR_FS). We
738 : * can't do I/O using huge pages yet, so return an error for now.
739 : * Filesystem transparent huge pages are typically only used for
740 : * executables that explicitly enable them, so this isn't very
741 : * restrictive.
742 : */
743 0 : if (PageCompound(page)) {
744 0 : unlock_page(page);
745 0 : put_page(page);
746 0 : return ERR_PTR(-ETXTBSY);
747 : }
748 :
749 0 : ret = set_page_extent_mapped(page);
750 0 : if (ret < 0) {
751 0 : unlock_page(page);
752 0 : put_page(page);
753 0 : return ERR_PTR(ret);
754 : }
755 :
756 : /* Wait for any existing ordered extent in the range */
757 0 : while (1) {
758 0 : struct btrfs_ordered_extent *ordered;
759 :
760 0 : lock_extent(&inode->io_tree, page_start, page_end, &cached_state);
761 0 : ordered = btrfs_lookup_ordered_range(inode, page_start, PAGE_SIZE);
762 0 : unlock_extent(&inode->io_tree, page_start, page_end,
763 : &cached_state);
764 0 : if (!ordered)
765 : break;
766 :
767 0 : unlock_page(page);
768 0 : btrfs_start_ordered_extent(ordered);
769 0 : btrfs_put_ordered_extent(ordered);
770 0 : lock_page(page);
771 : /*
772 : * We unlocked the page above, so we need check if it was
773 : * released or not.
774 : */
775 0 : if (page->mapping != mapping || !PagePrivate(page)) {
776 0 : unlock_page(page);
777 0 : put_page(page);
778 0 : goto again;
779 : }
780 : }
781 :
782 : /*
783 : * Now the page range has no ordered extent any more. Read the page to
784 : * make it uptodate.
785 : */
786 0 : if (!PageUptodate(page)) {
787 0 : btrfs_read_folio(NULL, page_folio(page));
788 0 : lock_page(page);
789 0 : if (page->mapping != mapping || !PagePrivate(page)) {
790 0 : unlock_page(page);
791 0 : put_page(page);
792 0 : goto again;
793 : }
794 0 : if (!PageUptodate(page)) {
795 0 : unlock_page(page);
796 0 : put_page(page);
797 0 : return ERR_PTR(-EIO);
798 : }
799 : }
800 : return page;
801 : }
802 :
803 : struct defrag_target_range {
804 : struct list_head list;
805 : u64 start;
806 : u64 len;
807 : };
808 :
809 : /*
810 : * Collect all valid target extents.
811 : *
812 : * @start: file offset to lookup
813 : * @len: length to lookup
814 : * @extent_thresh: file extent size threshold, any extent size >= this value
815 : * will be ignored
816 : * @newer_than: only defrag extents newer than this value
817 : * @do_compress: whether the defrag is doing compression
818 : * if true, @extent_thresh will be ignored and all regular
819 : * file extents meeting @newer_than will be targets.
820 : * @locked: if the range has already held extent lock
821 : * @target_list: list of targets file extents
822 : */
823 0 : static int defrag_collect_targets(struct btrfs_inode *inode,
824 : u64 start, u64 len, u32 extent_thresh,
825 : u64 newer_than, bool do_compress,
826 : bool locked, struct list_head *target_list,
827 : u64 *last_scanned_ret)
828 : {
829 0 : struct btrfs_fs_info *fs_info = inode->root->fs_info;
830 0 : bool last_is_target = false;
831 0 : u64 cur = start;
832 0 : int ret = 0;
833 :
834 0 : while (cur < start + len) {
835 0 : struct extent_map *em;
836 0 : struct defrag_target_range *new;
837 0 : bool next_mergeable = true;
838 0 : u64 range_len;
839 :
840 0 : last_is_target = false;
841 0 : em = defrag_lookup_extent(&inode->vfs_inode, cur, newer_than, locked);
842 0 : if (!em)
843 : break;
844 :
845 : /*
846 : * If the file extent is an inlined one, we may still want to
847 : * defrag it (fallthrough) if it will cause a regular extent.
848 : * This is for users who want to convert inline extents to
849 : * regular ones through max_inline= mount option.
850 : */
851 0 : if (em->block_start == EXTENT_MAP_INLINE &&
852 0 : em->len <= inode->root->fs_info->max_inline)
853 0 : goto next;
854 :
855 : /* Skip hole/delalloc/preallocated extents */
856 0 : if (em->block_start == EXTENT_MAP_HOLE ||
857 0 : em->block_start == EXTENT_MAP_DELALLOC ||
858 0 : test_bit(EXTENT_FLAG_PREALLOC, &em->flags))
859 0 : goto next;
860 :
861 : /* Skip older extent */
862 0 : if (em->generation < newer_than)
863 0 : goto next;
864 :
865 : /* This em is under writeback, no need to defrag */
866 0 : if (em->generation == (u64)-1)
867 0 : goto next;
868 :
869 : /*
870 : * Our start offset might be in the middle of an existing extent
871 : * map, so take that into account.
872 : */
873 0 : range_len = em->len - (cur - em->start);
874 : /*
875 : * If this range of the extent map is already flagged for delalloc,
876 : * skip it, because:
877 : *
878 : * 1) We could deadlock later, when trying to reserve space for
879 : * delalloc, because in case we can't immediately reserve space
880 : * the flusher can start delalloc and wait for the respective
881 : * ordered extents to complete. The deadlock would happen
882 : * because we do the space reservation while holding the range
883 : * locked, and starting writeback, or finishing an ordered
884 : * extent, requires locking the range;
885 : *
886 : * 2) If there's delalloc there, it means there's dirty pages for
887 : * which writeback has not started yet (we clean the delalloc
888 : * flag when starting writeback and after creating an ordered
889 : * extent). If we mark pages in an adjacent range for defrag,
890 : * then we will have a larger contiguous range for delalloc,
891 : * very likely resulting in a larger extent after writeback is
892 : * triggered (except in a case of free space fragmentation).
893 : */
894 0 : if (test_range_bit(&inode->io_tree, cur, cur + range_len - 1,
895 : EXTENT_DELALLOC, 0, NULL))
896 0 : goto next;
897 :
898 : /*
899 : * For do_compress case, we want to compress all valid file
900 : * extents, thus no @extent_thresh or mergeable check.
901 : */
902 0 : if (do_compress)
903 0 : goto add;
904 :
905 : /* Skip too large extent */
906 0 : if (range_len >= extent_thresh)
907 0 : goto next;
908 :
909 : /*
910 : * Skip extents already at its max capacity, this is mostly for
911 : * compressed extents, which max cap is only 128K.
912 : */
913 0 : if (em->len >= get_extent_max_capacity(fs_info, em))
914 0 : goto next;
915 :
916 : /*
917 : * Normally there are no more extents after an inline one, thus
918 : * @next_mergeable will normally be false and not defragged.
919 : * So if an inline extent passed all above checks, just add it
920 : * for defrag, and be converted to regular extents.
921 : */
922 0 : if (em->block_start == EXTENT_MAP_INLINE)
923 0 : goto add;
924 :
925 0 : next_mergeable = defrag_check_next_extent(&inode->vfs_inode, em,
926 : extent_thresh, newer_than, locked);
927 0 : if (!next_mergeable) {
928 0 : struct defrag_target_range *last;
929 :
930 : /* Empty target list, no way to merge with last entry */
931 0 : if (list_empty(target_list))
932 0 : goto next;
933 0 : last = list_entry(target_list->prev,
934 : struct defrag_target_range, list);
935 : /* Not mergeable with last entry */
936 0 : if (last->start + last->len != cur)
937 0 : goto next;
938 :
939 : /* Mergeable, fall through to add it to @target_list. */
940 : }
941 :
942 0 : add:
943 0 : last_is_target = true;
944 0 : range_len = min(extent_map_end(em), start + len) - cur;
945 : /*
946 : * This one is a good target, check if it can be merged into
947 : * last range of the target list.
948 : */
949 0 : if (!list_empty(target_list)) {
950 0 : struct defrag_target_range *last;
951 :
952 0 : last = list_entry(target_list->prev,
953 : struct defrag_target_range, list);
954 0 : ASSERT(last->start + last->len <= cur);
955 0 : if (last->start + last->len == cur) {
956 : /* Mergeable, enlarge the last entry */
957 0 : last->len += range_len;
958 0 : goto next;
959 : }
960 : /* Fall through to allocate a new entry */
961 : }
962 :
963 : /* Allocate new defrag_target_range */
964 0 : new = kmalloc(sizeof(*new), GFP_NOFS);
965 0 : if (!new) {
966 0 : free_extent_map(em);
967 0 : ret = -ENOMEM;
968 0 : break;
969 : }
970 0 : new->start = cur;
971 0 : new->len = range_len;
972 0 : list_add_tail(&new->list, target_list);
973 :
974 0 : next:
975 0 : cur = extent_map_end(em);
976 0 : free_extent_map(em);
977 : }
978 0 : if (ret < 0) {
979 0 : struct defrag_target_range *entry;
980 0 : struct defrag_target_range *tmp;
981 :
982 0 : list_for_each_entry_safe(entry, tmp, target_list, list) {
983 0 : list_del_init(&entry->list);
984 0 : kfree(entry);
985 : }
986 : }
987 0 : if (!ret && last_scanned_ret) {
988 : /*
989 : * If the last extent is not a target, the caller can skip to
990 : * the end of that extent.
991 : * Otherwise, we can only go the end of the specified range.
992 : */
993 0 : if (!last_is_target)
994 0 : *last_scanned_ret = max(cur, *last_scanned_ret);
995 : else
996 0 : *last_scanned_ret = max(start + len, *last_scanned_ret);
997 : }
998 0 : return ret;
999 : }
1000 :
1001 : #define CLUSTER_SIZE (SZ_256K)
1002 : static_assert(PAGE_ALIGNED(CLUSTER_SIZE));
1003 :
1004 : /*
1005 : * Defrag one contiguous target range.
1006 : *
1007 : * @inode: target inode
1008 : * @target: target range to defrag
1009 : * @pages: locked pages covering the defrag range
1010 : * @nr_pages: number of locked pages
1011 : *
1012 : * Caller should ensure:
1013 : *
1014 : * - Pages are prepared
1015 : * Pages should be locked, no ordered extent in the pages range,
1016 : * no writeback.
1017 : *
1018 : * - Extent bits are locked
1019 : */
1020 0 : static int defrag_one_locked_target(struct btrfs_inode *inode,
1021 : struct defrag_target_range *target,
1022 : struct page **pages, int nr_pages,
1023 : struct extent_state **cached_state)
1024 : {
1025 0 : struct btrfs_fs_info *fs_info = inode->root->fs_info;
1026 0 : struct extent_changeset *data_reserved = NULL;
1027 0 : const u64 start = target->start;
1028 0 : const u64 len = target->len;
1029 0 : unsigned long last_index = (start + len - 1) >> PAGE_SHIFT;
1030 0 : unsigned long start_index = start >> PAGE_SHIFT;
1031 0 : unsigned long first_index = page_index(pages[0]);
1032 0 : int ret = 0;
1033 0 : int i;
1034 :
1035 0 : ASSERT(last_index - first_index + 1 <= nr_pages);
1036 :
1037 0 : ret = btrfs_delalloc_reserve_space(inode, &data_reserved, start, len);
1038 0 : if (ret < 0)
1039 : return ret;
1040 0 : clear_extent_bit(&inode->io_tree, start, start + len - 1,
1041 : EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING |
1042 : EXTENT_DEFRAG, cached_state);
1043 0 : set_extent_bit(&inode->io_tree, start, start + len - 1,
1044 : EXTENT_DELALLOC | EXTENT_DEFRAG, cached_state);
1045 :
1046 : /* Update the page status */
1047 0 : for (i = start_index - first_index; i <= last_index - first_index; i++) {
1048 0 : ClearPageChecked(pages[i]);
1049 0 : btrfs_page_clamp_set_dirty(fs_info, pages[i], start, len);
1050 : }
1051 0 : btrfs_delalloc_release_extents(inode, len);
1052 0 : extent_changeset_free(data_reserved);
1053 :
1054 0 : return ret;
1055 : }
1056 :
1057 0 : static int defrag_one_range(struct btrfs_inode *inode, u64 start, u32 len,
1058 : u32 extent_thresh, u64 newer_than, bool do_compress,
1059 : u64 *last_scanned_ret)
1060 : {
1061 0 : struct extent_state *cached_state = NULL;
1062 0 : struct defrag_target_range *entry;
1063 0 : struct defrag_target_range *tmp;
1064 0 : LIST_HEAD(target_list);
1065 0 : struct page **pages;
1066 0 : const u32 sectorsize = inode->root->fs_info->sectorsize;
1067 0 : u64 last_index = (start + len - 1) >> PAGE_SHIFT;
1068 0 : u64 start_index = start >> PAGE_SHIFT;
1069 0 : unsigned int nr_pages = last_index - start_index + 1;
1070 0 : int ret = 0;
1071 0 : int i;
1072 :
1073 0 : ASSERT(nr_pages <= CLUSTER_SIZE / PAGE_SIZE);
1074 0 : ASSERT(IS_ALIGNED(start, sectorsize) && IS_ALIGNED(len, sectorsize));
1075 :
1076 0 : pages = kcalloc(nr_pages, sizeof(struct page *), GFP_NOFS);
1077 0 : if (!pages)
1078 : return -ENOMEM;
1079 :
1080 : /* Prepare all pages */
1081 0 : for (i = 0; i < nr_pages; i++) {
1082 0 : pages[i] = defrag_prepare_one_page(inode, start_index + i);
1083 0 : if (IS_ERR(pages[i])) {
1084 0 : ret = PTR_ERR(pages[i]);
1085 0 : pages[i] = NULL;
1086 0 : goto free_pages;
1087 : }
1088 : }
1089 0 : for (i = 0; i < nr_pages; i++)
1090 0 : wait_on_page_writeback(pages[i]);
1091 :
1092 : /* Lock the pages range */
1093 0 : lock_extent(&inode->io_tree, start_index << PAGE_SHIFT,
1094 0 : (last_index << PAGE_SHIFT) + PAGE_SIZE - 1,
1095 : &cached_state);
1096 : /*
1097 : * Now we have a consistent view about the extent map, re-check
1098 : * which range really needs to be defragged.
1099 : *
1100 : * And this time we have extent locked already, pass @locked = true
1101 : * so that we won't relock the extent range and cause deadlock.
1102 : */
1103 0 : ret = defrag_collect_targets(inode, start, len, extent_thresh,
1104 : newer_than, do_compress, true,
1105 : &target_list, last_scanned_ret);
1106 0 : if (ret < 0)
1107 0 : goto unlock_extent;
1108 :
1109 0 : list_for_each_entry(entry, &target_list, list) {
1110 0 : ret = defrag_one_locked_target(inode, entry, pages, nr_pages,
1111 : &cached_state);
1112 0 : if (ret < 0)
1113 : break;
1114 : }
1115 :
1116 0 : list_for_each_entry_safe(entry, tmp, &target_list, list) {
1117 0 : list_del_init(&entry->list);
1118 0 : kfree(entry);
1119 : }
1120 0 : unlock_extent:
1121 0 : unlock_extent(&inode->io_tree, start_index << PAGE_SHIFT,
1122 : (last_index << PAGE_SHIFT) + PAGE_SIZE - 1,
1123 : &cached_state);
1124 0 : free_pages:
1125 0 : for (i = 0; i < nr_pages; i++) {
1126 0 : if (pages[i]) {
1127 0 : unlock_page(pages[i]);
1128 0 : put_page(pages[i]);
1129 : }
1130 : }
1131 0 : kfree(pages);
1132 0 : return ret;
1133 : }
1134 :
1135 0 : static int defrag_one_cluster(struct btrfs_inode *inode,
1136 : struct file_ra_state *ra,
1137 : u64 start, u32 len, u32 extent_thresh,
1138 : u64 newer_than, bool do_compress,
1139 : unsigned long *sectors_defragged,
1140 : unsigned long max_sectors,
1141 : u64 *last_scanned_ret)
1142 : {
1143 0 : const u32 sectorsize = inode->root->fs_info->sectorsize;
1144 0 : struct defrag_target_range *entry;
1145 0 : struct defrag_target_range *tmp;
1146 0 : LIST_HEAD(target_list);
1147 0 : int ret;
1148 :
1149 0 : ret = defrag_collect_targets(inode, start, len, extent_thresh,
1150 : newer_than, do_compress, false,
1151 : &target_list, NULL);
1152 0 : if (ret < 0)
1153 0 : goto out;
1154 :
1155 0 : list_for_each_entry(entry, &target_list, list) {
1156 0 : u32 range_len = entry->len;
1157 :
1158 : /* Reached or beyond the limit */
1159 0 : if (max_sectors && *sectors_defragged >= max_sectors) {
1160 : ret = 1;
1161 : break;
1162 : }
1163 :
1164 0 : if (max_sectors)
1165 0 : range_len = min_t(u32, range_len,
1166 : (max_sectors - *sectors_defragged) * sectorsize);
1167 :
1168 : /*
1169 : * If defrag_one_range() has updated last_scanned_ret,
1170 : * our range may already be invalid (e.g. hole punched).
1171 : * Skip if our range is before last_scanned_ret, as there is
1172 : * no need to defrag the range anymore.
1173 : */
1174 0 : if (entry->start + range_len <= *last_scanned_ret)
1175 0 : continue;
1176 :
1177 0 : if (ra)
1178 0 : page_cache_sync_readahead(inode->vfs_inode.i_mapping,
1179 0 : ra, NULL, entry->start >> PAGE_SHIFT,
1180 0 : ((entry->start + range_len - 1) >> PAGE_SHIFT) -
1181 0 : (entry->start >> PAGE_SHIFT) + 1);
1182 : /*
1183 : * Here we may not defrag any range if holes are punched before
1184 : * we locked the pages.
1185 : * But that's fine, it only affects the @sectors_defragged
1186 : * accounting.
1187 : */
1188 0 : ret = defrag_one_range(inode, entry->start, range_len,
1189 : extent_thresh, newer_than, do_compress,
1190 : last_scanned_ret);
1191 0 : if (ret < 0)
1192 : break;
1193 0 : *sectors_defragged += range_len >>
1194 0 : inode->root->fs_info->sectorsize_bits;
1195 : }
1196 0 : out:
1197 0 : list_for_each_entry_safe(entry, tmp, &target_list, list) {
1198 0 : list_del_init(&entry->list);
1199 0 : kfree(entry);
1200 : }
1201 0 : if (ret >= 0)
1202 0 : *last_scanned_ret = max(*last_scanned_ret, start + len);
1203 0 : return ret;
1204 : }
1205 :
1206 : /*
1207 : * Entry point to file defragmentation.
1208 : *
1209 : * @inode: inode to be defragged
1210 : * @ra: readahead state (can be NUL)
1211 : * @range: defrag options including range and flags
1212 : * @newer_than: minimum transid to defrag
1213 : * @max_to_defrag: max number of sectors to be defragged, if 0, the whole inode
1214 : * will be defragged.
1215 : *
1216 : * Return <0 for error.
1217 : * Return >=0 for the number of sectors defragged, and range->start will be updated
1218 : * to indicate the file offset where next defrag should be started at.
1219 : * (Mostly for autodefrag, which sets @max_to_defrag thus we may exit early without
1220 : * defragging all the range).
1221 : */
1222 0 : int btrfs_defrag_file(struct inode *inode, struct file_ra_state *ra,
1223 : struct btrfs_ioctl_defrag_range_args *range,
1224 : u64 newer_than, unsigned long max_to_defrag)
1225 : {
1226 0 : struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
1227 0 : unsigned long sectors_defragged = 0;
1228 0 : u64 isize = i_size_read(inode);
1229 0 : u64 cur;
1230 0 : u64 last_byte;
1231 0 : bool do_compress = (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS);
1232 0 : bool ra_allocated = false;
1233 0 : int compress_type = BTRFS_COMPRESS_ZLIB;
1234 0 : int ret = 0;
1235 0 : u32 extent_thresh = range->extent_thresh;
1236 0 : pgoff_t start_index;
1237 :
1238 0 : if (isize == 0)
1239 : return 0;
1240 :
1241 0 : if (range->start >= isize)
1242 : return -EINVAL;
1243 :
1244 0 : if (do_compress) {
1245 0 : if (range->compress_type >= BTRFS_NR_COMPRESS_TYPES)
1246 : return -EINVAL;
1247 0 : if (range->compress_type)
1248 0 : compress_type = range->compress_type;
1249 : }
1250 :
1251 0 : if (extent_thresh == 0)
1252 0 : extent_thresh = SZ_256K;
1253 :
1254 0 : if (range->start + range->len > range->start) {
1255 : /* Got a specific range */
1256 0 : last_byte = min(isize, range->start + range->len);
1257 : } else {
1258 : /* Defrag until file end */
1259 : last_byte = isize;
1260 : }
1261 :
1262 : /* Align the range */
1263 0 : cur = round_down(range->start, fs_info->sectorsize);
1264 0 : last_byte = round_up(last_byte, fs_info->sectorsize) - 1;
1265 :
1266 : /*
1267 : * If we were not given a ra, allocate a readahead context. As
1268 : * readahead is just an optimization, defrag will work without it so
1269 : * we don't error out.
1270 : */
1271 0 : if (!ra) {
1272 0 : ra_allocated = true;
1273 0 : ra = kzalloc(sizeof(*ra), GFP_KERNEL);
1274 0 : if (ra)
1275 0 : file_ra_state_init(ra, inode->i_mapping);
1276 : }
1277 :
1278 : /*
1279 : * Make writeback start from the beginning of the range, so that the
1280 : * defrag range can be written sequentially.
1281 : */
1282 0 : start_index = cur >> PAGE_SHIFT;
1283 0 : if (start_index < inode->i_mapping->writeback_index)
1284 0 : inode->i_mapping->writeback_index = start_index;
1285 :
1286 0 : while (cur < last_byte) {
1287 0 : const unsigned long prev_sectors_defragged = sectors_defragged;
1288 0 : u64 last_scanned = cur;
1289 0 : u64 cluster_end;
1290 :
1291 0 : if (btrfs_defrag_cancelled(fs_info)) {
1292 : ret = -EAGAIN;
1293 0 : break;
1294 : }
1295 :
1296 : /* We want the cluster end at page boundary when possible */
1297 0 : cluster_end = (((cur >> PAGE_SHIFT) +
1298 0 : (SZ_256K >> PAGE_SHIFT)) << PAGE_SHIFT) - 1;
1299 0 : cluster_end = min(cluster_end, last_byte);
1300 :
1301 0 : btrfs_inode_lock(BTRFS_I(inode), 0);
1302 0 : if (IS_SWAPFILE(inode)) {
1303 0 : ret = -ETXTBSY;
1304 0 : btrfs_inode_unlock(BTRFS_I(inode), 0);
1305 0 : break;
1306 : }
1307 0 : if (!(inode->i_sb->s_flags & SB_ACTIVE)) {
1308 0 : btrfs_inode_unlock(BTRFS_I(inode), 0);
1309 0 : break;
1310 : }
1311 0 : if (do_compress)
1312 0 : BTRFS_I(inode)->defrag_compress = compress_type;
1313 0 : ret = defrag_one_cluster(BTRFS_I(inode), ra, cur,
1314 0 : cluster_end + 1 - cur, extent_thresh,
1315 : newer_than, do_compress, §ors_defragged,
1316 : max_to_defrag, &last_scanned);
1317 :
1318 0 : if (sectors_defragged > prev_sectors_defragged)
1319 0 : balance_dirty_pages_ratelimited(inode->i_mapping);
1320 :
1321 0 : btrfs_inode_unlock(BTRFS_I(inode), 0);
1322 0 : if (ret < 0)
1323 : break;
1324 0 : cur = max(cluster_end + 1, last_scanned);
1325 0 : if (ret > 0) {
1326 : ret = 0;
1327 : break;
1328 : }
1329 0 : cond_resched();
1330 : }
1331 :
1332 0 : if (ra_allocated)
1333 0 : kfree(ra);
1334 : /*
1335 : * Update range.start for autodefrag, this will indicate where to start
1336 : * in next run.
1337 : */
1338 0 : range->start = cur;
1339 0 : if (sectors_defragged) {
1340 : /*
1341 : * We have defragged some sectors, for compression case they
1342 : * need to be written back immediately.
1343 : */
1344 0 : if (range->flags & BTRFS_DEFRAG_RANGE_START_IO) {
1345 0 : filemap_flush(inode->i_mapping);
1346 0 : if (test_bit(BTRFS_INODE_HAS_ASYNC_EXTENT,
1347 : &BTRFS_I(inode)->runtime_flags))
1348 0 : filemap_flush(inode->i_mapping);
1349 : }
1350 0 : if (range->compress_type == BTRFS_COMPRESS_LZO)
1351 0 : btrfs_set_fs_incompat(fs_info, COMPRESS_LZO);
1352 0 : else if (range->compress_type == BTRFS_COMPRESS_ZSTD)
1353 0 : btrfs_set_fs_incompat(fs_info, COMPRESS_ZSTD);
1354 0 : ret = sectors_defragged;
1355 : }
1356 0 : if (do_compress) {
1357 0 : btrfs_inode_lock(BTRFS_I(inode), 0);
1358 0 : BTRFS_I(inode)->defrag_compress = BTRFS_COMPRESS_NONE;
1359 0 : btrfs_inode_unlock(BTRFS_I(inode), 0);
1360 : }
1361 : return ret;
1362 : }
1363 :
1364 0 : void __cold btrfs_auto_defrag_exit(void)
1365 : {
1366 0 : kmem_cache_destroy(btrfs_inode_defrag_cachep);
1367 0 : }
1368 :
1369 2 : int __init btrfs_auto_defrag_init(void)
1370 : {
1371 2 : btrfs_inode_defrag_cachep = kmem_cache_create("btrfs_inode_defrag",
1372 : sizeof(struct inode_defrag), 0,
1373 : SLAB_MEM_SPREAD,
1374 : NULL);
1375 2 : if (!btrfs_inode_defrag_cachep)
1376 0 : return -ENOMEM;
1377 :
1378 : return 0;
1379 : }
|