?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
nfsd4: typo logical vs bitwise negate for want_mask
[linux-btrfs-devel.git] / fs / btrfs / extent-tree.c
blob66bac226944e05ae82edb56a9fd43116f31a901a
1 /*
2 * Copyright (C) 2007 Oracle. All rights reserved.
3 *
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU General Public
6 * License v2 as published by the Free Software Foundation.
7 *
8 * This program is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
11 * General Public License for more details.
12 *
13 * You should have received a copy of the GNU General Public
14 * License along with this program; if not, write to the
15 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
16 * Boston, MA 021110-1307, USA.
17 */
18 #include <linux/sched.h>
19 #include <linux/pagemap.h>
20 #include <linux/writeback.h>
21 #include <linux/blkdev.h>
22 #include <linux/sort.h>
23 #include <linux/rcupdate.h>
24 #include <linux/kthread.h>
25 #include <linux/slab.h>
26 #include "compat.h"
27 #include "hash.h"
28 #include "ctree.h"
29 #include "disk-io.h"
30 #include "print-tree.h"
31 #include "transaction.h"
32 #include "volumes.h"
33 #include "locking.h"
34 #include "free-space-cache.h"
35
36 /* control flags for do_chunk_alloc's force field
37 * CHUNK_ALLOC_NO_FORCE means to only allocate a chunk
38 * if we really need one.
39 *
40 * CHUNK_ALLOC_FORCE means it must try to allocate one
41 *
42 * CHUNK_ALLOC_LIMITED means to only try and allocate one
43 * if we have very few chunks already allocated. This is
44 * used as part of the clustering code to help make sure
45 * we have a good pool of storage to cluster in, without
46 * filling the FS with empty chunks
47 *
48 */
49 enum {
50 CHUNK_ALLOC_NO_FORCE = 0,
51 CHUNK_ALLOC_FORCE = 1,
52 CHUNK_ALLOC_LIMITED = 2,
53 };
54
55 static int update_block_group(struct btrfs_trans_handle *trans,
56 struct btrfs_root *root,
57 u64 bytenr, u64 num_bytes, int alloc);
58 static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
59 struct btrfs_root *root,
60 u64 bytenr, u64 num_bytes, u64 parent,
61 u64 root_objectid, u64 owner_objectid,
62 u64 owner_offset, int refs_to_drop,
63 struct btrfs_delayed_extent_op *extra_op);
64 static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op,
65 struct extent_buffer *leaf,
66 struct btrfs_extent_item *ei);
67 static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
68 struct btrfs_root *root,
69 u64 parent, u64 root_objectid,
70 u64 flags, u64 owner, u64 offset,
71 struct btrfs_key *ins, int ref_mod);
72 static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
73 struct btrfs_root *root,
74 u64 parent, u64 root_objectid,
75 u64 flags, struct btrfs_disk_key *key,
76 int level, struct btrfs_key *ins);
77 static int do_chunk_alloc(struct btrfs_trans_handle *trans,
78 struct btrfs_root *extent_root, u64 alloc_bytes,
79 u64 flags, int force);
80 static int find_next_key(struct btrfs_path *path, int level,
81 struct btrfs_key *key);
82 static void dump_space_info(struct btrfs_space_info *info, u64 bytes,
83 int dump_block_groups);
84
85 static noinline int
86 block_group_cache_done(struct btrfs_block_group_cache *cache)
87 {
88 smp_mb();
89 return cache->cached == BTRFS_CACHE_FINISHED;
90 }
91
92 static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits)
93 {
94 return (cache->flags & bits) == bits;
95 }
96
97 static void btrfs_get_block_group(struct btrfs_block_group_cache *cache)
98 {
99 atomic_inc(&cache->count);
100 }
101
102 void btrfs_put_block_group(struct btrfs_block_group_cache *cache)
103 {
104 if (atomic_dec_and_test(&cache->count)) {
105 WARN_ON(cache->pinned > 0);
106 WARN_ON(cache->reserved > 0);
107 WARN_ON(cache->reserved_pinned > 0);
108 kfree(cache->free_space_ctl);
109 kfree(cache);
110 }
111 }
112
113 /*
114 * this adds the block group to the fs_info rb tree for the block group
115 * cache
116 */
117 static int btrfs_add_block_group_cache(struct btrfs_fs_info *info,
118 struct btrfs_block_group_cache *block_group)
119 {
120 struct rb_node **p;
121 struct rb_node *parent = NULL;
122 struct btrfs_block_group_cache *cache;
123
124 spin_lock(&info->block_group_cache_lock);
125 p = &info->block_group_cache_tree.rb_node;
126
127 while (*p) {
128 parent = *p;
129 cache = rb_entry(parent, struct btrfs_block_group_cache,
130 cache_node);
131 if (block_group->key.objectid < cache->key.objectid) {
132 p = &(*p)->rb_left;
133 } else if (block_group->key.objectid > cache->key.objectid) {
134 p = &(*p)->rb_right;
135 } else {
136 spin_unlock(&info->block_group_cache_lock);
137 return -EEXIST;
138 }
139 }
140
141 rb_link_node(&block_group->cache_node, parent, p);
142 rb_insert_color(&block_group->cache_node,
143 &info->block_group_cache_tree);
144 spin_unlock(&info->block_group_cache_lock);
145
146 return 0;
147 }
148
149 /*
150 * This will return the block group at or after bytenr if contains is 0, else
151 * it will return the block group that contains the bytenr
152 */
153 static struct btrfs_block_group_cache *
154 block_group_cache_tree_search(struct btrfs_fs_info *info, u64 bytenr,
155 int contains)
156 {
157 struct btrfs_block_group_cache *cache, *ret = NULL;
158 struct rb_node *n;
159 u64 end, start;
160
161 spin_lock(&info->block_group_cache_lock);
162 n = info->block_group_cache_tree.rb_node;
163
164 while (n) {
165 cache = rb_entry(n, struct btrfs_block_group_cache,
166 cache_node);
167 end = cache->key.objectid + cache->key.offset - 1;
168 start = cache->key.objectid;
169
170 if (bytenr < start) {
171 if (!contains && (!ret || start < ret->key.objectid))
172 ret = cache;
173 n = n->rb_left;
174 } else if (bytenr > start) {
175 if (contains && bytenr <= end) {
176 ret = cache;
177 break;
178 }
179 n = n->rb_right;
180 } else {
181 ret = cache;
182 break;
183 }
184 }
185 if (ret)
186 btrfs_get_block_group(ret);
187 spin_unlock(&info->block_group_cache_lock);
188
189 return ret;
190 }
191
192 static int add_excluded_extent(struct btrfs_root *root,
193 u64 start, u64 num_bytes)
194 {
195 u64 end = start + num_bytes - 1;
196 set_extent_bits(&root->fs_info->freed_extents[0],
197 start, end, EXTENT_UPTODATE, GFP_NOFS);
198 set_extent_bits(&root->fs_info->freed_extents[1],
199 start, end, EXTENT_UPTODATE, GFP_NOFS);
200 return 0;
201 }
202
203 static void free_excluded_extents(struct btrfs_root *root,
204 struct btrfs_block_group_cache *cache)
205 {
206 u64 start, end;
207
208 start = cache->key.objectid;
209 end = start + cache->key.offset - 1;
210
211 clear_extent_bits(&root->fs_info->freed_extents[0],
212 start, end, EXTENT_UPTODATE, GFP_NOFS);
213 clear_extent_bits(&root->fs_info->freed_extents[1],
214 start, end, EXTENT_UPTODATE, GFP_NOFS);
215 }
216
217 static int exclude_super_stripes(struct btrfs_root *root,
218 struct btrfs_block_group_cache *cache)
219 {
220 u64 bytenr;
221 u64 *logical;
222 int stripe_len;
223 int i, nr, ret;
224
225 if (cache->key.objectid < BTRFS_SUPER_INFO_OFFSET) {
226 stripe_len = BTRFS_SUPER_INFO_OFFSET - cache->key.objectid;
227 cache->bytes_super += stripe_len;
228 ret = add_excluded_extent(root, cache->key.objectid,
229 stripe_len);
230 BUG_ON(ret);
231 }
232
233 for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) {
234 bytenr = btrfs_sb_offset(i);
235 ret = btrfs_rmap_block(&root->fs_info->mapping_tree,
236 cache->key.objectid, bytenr,
237 0, &logical, &nr, &stripe_len);
238 BUG_ON(ret);
239
240 while (nr--) {
241 cache->bytes_super += stripe_len;
242 ret = add_excluded_extent(root, logical[nr],
243 stripe_len);
244 BUG_ON(ret);
245 }
246
247 kfree(logical);
248 }
249 return 0;
250 }
251
252 static struct btrfs_caching_control *
253 get_caching_control(struct btrfs_block_group_cache *cache)
254 {
255 struct btrfs_caching_control *ctl;
256
257 spin_lock(&cache->lock);
258 if (cache->cached != BTRFS_CACHE_STARTED) {
259 spin_unlock(&cache->lock);
260 return NULL;
261 }
262
263 /* We're loading it the fast way, so we don't have a caching_ctl. */
264 if (!cache->caching_ctl) {
265 spin_unlock(&cache->lock);
266 return NULL;
267 }
268
269 ctl = cache->caching_ctl;
270 atomic_inc(&ctl->count);
271 spin_unlock(&cache->lock);
272 return ctl;
273 }
274
275 static void put_caching_control(struct btrfs_caching_control *ctl)
276 {
277 if (atomic_dec_and_test(&ctl->count))
278 kfree(ctl);
279 }
280
281 /*
282 * this is only called by cache_block_group, since we could have freed extents
283 * we need to check the pinned_extents for any extents that can't be used yet
284 * since their free space will be released as soon as the transaction commits.
285 */
286 static u64 add_new_free_space(struct btrfs_block_group_cache *block_group,
287 struct btrfs_fs_info *info, u64 start, u64 end)
288 {
289 u64 extent_start, extent_end, size, total_added = 0;
290 int ret;
291
292 while (start < end) {
293 ret = find_first_extent_bit(info->pinned_extents, start,
294 &extent_start, &extent_end,
295 EXTENT_DIRTY | EXTENT_UPTODATE);
296 if (ret)
297 break;
298
299 if (extent_start <= start) {
300 start = extent_end + 1;
301 } else if (extent_start > start && extent_start < end) {
302 size = extent_start - start;
303 total_added += size;
304 ret = btrfs_add_free_space(block_group, start,
305 size);
306 BUG_ON(ret);
307 start = extent_end + 1;
308 } else {
309 break;
310 }
311 }
312
313 if (start < end) {
314 size = end - start;
315 total_added += size;
316 ret = btrfs_add_free_space(block_group, start, size);
317 BUG_ON(ret);
318 }
319
320 return total_added;
321 }
322
323 static noinline void caching_thread(struct btrfs_work *work)
324 {
325 struct btrfs_block_group_cache *block_group;
326 struct btrfs_fs_info *fs_info;
327 struct btrfs_caching_control *caching_ctl;
328 struct btrfs_root *extent_root;
329 struct btrfs_path *path;
330 struct extent_buffer *leaf;
331 struct btrfs_key key;
332 u64 total_found = 0;
333 u64 last = 0;
334 u32 nritems;
335 int ret = 0;
336
337 caching_ctl = container_of(work, struct btrfs_caching_control, work);
338 block_group = caching_ctl->block_group;
339 fs_info = block_group->fs_info;
340 extent_root = fs_info->extent_root;
341
342 path = btrfs_alloc_path();
343 if (!path)
344 goto out;
345
346 last = max_t(u64, block_group->key.objectid, BTRFS_SUPER_INFO_OFFSET);
347
348 /*
349 * We don't want to deadlock with somebody trying to allocate a new
350 * extent for the extent root while also trying to search the extent
351 * root to add free space. So we skip locking and search the commit
352 * root, since its read-only
353 */
354 path->skip_locking = 1;
355 path->search_commit_root = 1;
356 path->reada = 1;
357
358 key.objectid = last;
359 key.offset = 0;
360 key.type = BTRFS_EXTENT_ITEM_KEY;
361 again:
362 mutex_lock(&caching_ctl->mutex);
363 /* need to make sure the commit_root doesn't disappear */
364 down_read(&fs_info->extent_commit_sem);
365
366 ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
367 if (ret < 0)
368 goto err;
369
370 leaf = path->nodes[0];
371 nritems = btrfs_header_nritems(leaf);
372
373 while (1) {
374 if (btrfs_fs_closing(fs_info) > 1) {
375 last = (u64)-1;
376 break;
377 }
378
379 if (path->slots[0] < nritems) {
380 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
381 } else {
382 ret = find_next_key(path, 0, &key);
383 if (ret)
384 break;
385
386 if (need_resched() ||
387 btrfs_next_leaf(extent_root, path)) {
388 caching_ctl->progress = last;
389 btrfs_release_path(path);
390 up_read(&fs_info->extent_commit_sem);
391 mutex_unlock(&caching_ctl->mutex);
392 cond_resched();
393 goto again;
394 }
395 leaf = path->nodes[0];
396 nritems = btrfs_header_nritems(leaf);
397 continue;
398 }
399
400 if (key.objectid < block_group->key.objectid) {
401 path->slots[0]++;
402 continue;
403 }
404
405 if (key.objectid >= block_group->key.objectid +
406 block_group->key.offset)
407 break;
408
409 if (key.type == BTRFS_EXTENT_ITEM_KEY) {
410 total_found += add_new_free_space(block_group,
411 fs_info, last,
412 key.objectid);
413 last = key.objectid + key.offset;
414
415 if (total_found > (1024 * 1024 * 2)) {
416 total_found = 0;
417 wake_up(&caching_ctl->wait);
418 }
419 }
420 path->slots[0]++;
421 }
422 ret = 0;
423
424 total_found += add_new_free_space(block_group, fs_info, last,
425 block_group->key.objectid +
426 block_group->key.offset);
427 caching_ctl->progress = (u64)-1;
428
429 spin_lock(&block_group->lock);
430 block_group->caching_ctl = NULL;
431 block_group->cached = BTRFS_CACHE_FINISHED;
432 spin_unlock(&block_group->lock);
433
434 err:
435 btrfs_free_path(path);
436 up_read(&fs_info->extent_commit_sem);
437
438 free_excluded_extents(extent_root, block_group);
439
440 mutex_unlock(&caching_ctl->mutex);
441 out:
442 wake_up(&caching_ctl->wait);
443
444 put_caching_control(caching_ctl);
445 btrfs_put_block_group(block_group);
446 }
447
448 static int cache_block_group(struct btrfs_block_group_cache *cache,
449 struct btrfs_trans_handle *trans,
450 struct btrfs_root *root,
451 int load_cache_only)
452 {
453 struct btrfs_fs_info *fs_info = cache->fs_info;
454 struct btrfs_caching_control *caching_ctl;
455 int ret = 0;
456
457 smp_mb();
458 if (cache->cached != BTRFS_CACHE_NO)
459 return 0;
460
461 /*
462 * We can't do the read from on-disk cache during a commit since we need
463 * to have the normal tree locking. Also if we are currently trying to
464 * allocate blocks for the tree root we can't do the fast caching since
465 * we likely hold important locks.
466 */
467 if (trans && (!trans->transaction->in_commit) &&
468 (root && root != root->fs_info->tree_root)) {
469 spin_lock(&cache->lock);
470 if (cache->cached != BTRFS_CACHE_NO) {
471 spin_unlock(&cache->lock);
472 return 0;
473 }
474 cache->cached = BTRFS_CACHE_STARTED;
475 spin_unlock(&cache->lock);
476
477 ret = load_free_space_cache(fs_info, cache);
478
479 spin_lock(&cache->lock);
480 if (ret == 1) {
481 cache->cached = BTRFS_CACHE_FINISHED;
482 cache->last_byte_to_unpin = (u64)-1;
483 } else {
484 cache->cached = BTRFS_CACHE_NO;
485 }
486 spin_unlock(&cache->lock);
487 if (ret == 1) {
488 free_excluded_extents(fs_info->extent_root, cache);
489 return 0;
490 }
491 }
492
493 if (load_cache_only)
494 return 0;
495
496 caching_ctl = kzalloc(sizeof(*caching_ctl), GFP_NOFS);
497 BUG_ON(!caching_ctl);
498
499 INIT_LIST_HEAD(&caching_ctl->list);
500 mutex_init(&caching_ctl->mutex);
501 init_waitqueue_head(&caching_ctl->wait);
502 caching_ctl->block_group = cache;
503 caching_ctl->progress = cache->key.objectid;
504 /* one for caching kthread, one for caching block group list */
505 atomic_set(&caching_ctl->count, 2);
506 caching_ctl->work.func = caching_thread;
507
508 spin_lock(&cache->lock);
509 if (cache->cached != BTRFS_CACHE_NO) {
510 spin_unlock(&cache->lock);
511 kfree(caching_ctl);
512 return 0;
513 }
514 cache->caching_ctl = caching_ctl;
515 cache->cached = BTRFS_CACHE_STARTED;
516 spin_unlock(&cache->lock);
517
518 down_write(&fs_info->extent_commit_sem);
519 list_add_tail(&caching_ctl->list, &fs_info->caching_block_groups);
520 up_write(&fs_info->extent_commit_sem);
521
522 btrfs_get_block_group(cache);
523
524 btrfs_queue_worker(&fs_info->caching_workers, &caching_ctl->work);
525
526 return ret;
527 }
528
529 /*
530 * return the block group that starts at or after bytenr
531 */
532 static struct btrfs_block_group_cache *
533 btrfs_lookup_first_block_group(struct btrfs_fs_info *info, u64 bytenr)
534 {
535 struct btrfs_block_group_cache *cache;
536
537 cache = block_group_cache_tree_search(info, bytenr, 0);
538
539 return cache;
540 }
541
542 /*
543 * return the block group that contains the given bytenr
544 */
545 struct btrfs_block_group_cache *btrfs_lookup_block_group(
546 struct btrfs_fs_info *info,
547 u64 bytenr)
548 {
549 struct btrfs_block_group_cache *cache;
550
551 cache = block_group_cache_tree_search(info, bytenr, 1);
552
553 return cache;
554 }
555
556 static struct btrfs_space_info *__find_space_info(struct btrfs_fs_info *info,
557 u64 flags)
558 {
559 struct list_head *head = &info->space_info;
560 struct btrfs_space_info *found;
561
562 flags &= BTRFS_BLOCK_GROUP_DATA | BTRFS_BLOCK_GROUP_SYSTEM |
563 BTRFS_BLOCK_GROUP_METADATA;
564
565 rcu_read_lock();
566 list_for_each_entry_rcu(found, head, list) {
567 if (found->flags & flags) {
568 rcu_read_unlock();
569 return found;
570 }
571 }
572 rcu_read_unlock();
573 return NULL;
574 }
575
576 /*
577 * after adding space to the filesystem, we need to clear the full flags
578 * on all the space infos.
579 */
580 void btrfs_clear_space_info_full(struct btrfs_fs_info *info)
581 {
582 struct list_head *head = &info->space_info;
583 struct btrfs_space_info *found;
584
585 rcu_read_lock();
586 list_for_each_entry_rcu(found, head, list)
587 found->full = 0;
588 rcu_read_unlock();
589 }
590
591 static u64 div_factor(u64 num, int factor)
592 {
593 if (factor == 10)
594 return num;
595 num *= factor;
596 do_div(num, 10);
597 return num;
598 }
599
600 static u64 div_factor_fine(u64 num, int factor)
601 {
602 if (factor == 100)
603 return num;
604 num *= factor;
605 do_div(num, 100);
606 return num;
607 }
608
609 u64 btrfs_find_block_group(struct btrfs_root *root,
610 u64 search_start, u64 search_hint, int owner)
611 {
612 struct btrfs_block_group_cache *cache;
613 u64 used;
614 u64 last = max(search_hint, search_start);
615 u64 group_start = 0;
616 int full_search = 0;
617 int factor = 9;
618 int wrapped = 0;
619 again:
620 while (1) {
621 cache = btrfs_lookup_first_block_group(root->fs_info, last);
622 if (!cache)
623 break;
624
625 spin_lock(&cache->lock);
626 last = cache->key.objectid + cache->key.offset;
627 used = btrfs_block_group_used(&cache->item);
628
629 if ((full_search || !cache->ro) &&
630 block_group_bits(cache, BTRFS_BLOCK_GROUP_METADATA)) {
631 if (used + cache->pinned + cache->reserved <
632 div_factor(cache->key.offset, factor)) {
633 group_start = cache->key.objectid;
634 spin_unlock(&cache->lock);
635 btrfs_put_block_group(cache);
636 goto found;
637 }
638 }
639 spin_unlock(&cache->lock);
640 btrfs_put_block_group(cache);
641 cond_resched();
642 }
643 if (!wrapped) {
644 last = search_start;
645 wrapped = 1;
646 goto again;
647 }
648 if (!full_search && factor < 10) {
649 last = search_start;
650 full_search = 1;
651 factor = 10;
652 goto again;
653 }
654 found:
655 return group_start;
656 }
657
658 /* simple helper to search for an existing extent at a given offset */
659 int btrfs_lookup_extent(struct btrfs_root *root, u64 start, u64 len)
660 {
661 int ret;
662 struct btrfs_key key;
663 struct btrfs_path *path;
664
665 path = btrfs_alloc_path();
666 if (!path)
667 return -ENOMEM;
668
669 key.objectid = start;
670 key.offset = len;
671 btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
672 ret = btrfs_search_slot(NULL, root->fs_info->extent_root, &key, path,
673 0, 0);
674 btrfs_free_path(path);
675 return ret;
676 }
677
678 /*
679 * helper function to lookup reference count and flags of extent.
680 *
681 * the head node for delayed ref is used to store the sum of all the
682 * reference count modifications queued up in the rbtree. the head
683 * node may also store the extent flags to set. This way you can check
684 * to see what the reference count and extent flags would be if all of
685 * the delayed refs are not processed.
686 */
687 int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans,
688 struct btrfs_root *root, u64 bytenr,
689 u64 num_bytes, u64 *refs, u64 *flags)
690 {
691 struct btrfs_delayed_ref_head *head;
692 struct btrfs_delayed_ref_root *delayed_refs;
693 struct btrfs_path *path;
694 struct btrfs_extent_item *ei;
695 struct extent_buffer *leaf;
696 struct btrfs_key key;
697 u32 item_size;
698 u64 num_refs;
699 u64 extent_flags;
700 int ret;
701
702 path = btrfs_alloc_path();
703 if (!path)
704 return -ENOMEM;
705
706 key.objectid = bytenr;
707 key.type = BTRFS_EXTENT_ITEM_KEY;
708 key.offset = num_bytes;
709 if (!trans) {
710 path->skip_locking = 1;
711 path->search_commit_root = 1;
712 }
713 again:
714 ret = btrfs_search_slot(trans, root->fs_info->extent_root,
715 &key, path, 0, 0);
716 if (ret < 0)
717 goto out_free;
718
719 if (ret == 0) {
720 leaf = path->nodes[0];
721 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
722 if (item_size >= sizeof(*ei)) {
723 ei = btrfs_item_ptr(leaf, path->slots[0],
724 struct btrfs_extent_item);
725 num_refs = btrfs_extent_refs(leaf, ei);
726 extent_flags = btrfs_extent_flags(leaf, ei);
727 } else {
728 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
729 struct btrfs_extent_item_v0 *ei0;
730 BUG_ON(item_size != sizeof(*ei0));
731 ei0 = btrfs_item_ptr(leaf, path->slots[0],
732 struct btrfs_extent_item_v0);
733 num_refs = btrfs_extent_refs_v0(leaf, ei0);
734 /* FIXME: this isn't correct for data */
735 extent_flags = BTRFS_BLOCK_FLAG_FULL_BACKREF;
736 #else
737 BUG();
738 #endif
739 }
740 BUG_ON(num_refs == 0);
741 } else {
742 num_refs = 0;
743 extent_flags = 0;
744 ret = 0;
745 }
746
747 if (!trans)
748 goto out;
749
750 delayed_refs = &trans->transaction->delayed_refs;
751 spin_lock(&delayed_refs->lock);
752 head = btrfs_find_delayed_ref_head(trans, bytenr);
753 if (head) {
754 if (!mutex_trylock(&head->mutex)) {
755 atomic_inc(&head->node.refs);
756 spin_unlock(&delayed_refs->lock);
757
758 btrfs_release_path(path);
759
760 /*
761 * Mutex was contended, block until it's released and try
762 * again
763 */
764 mutex_lock(&head->mutex);
765 mutex_unlock(&head->mutex);
766 btrfs_put_delayed_ref(&head->node);
767 goto again;
768 }
769 if (head->extent_op && head->extent_op->update_flags)
770 extent_flags |= head->extent_op->flags_to_set;
771 else
772 BUG_ON(num_refs == 0);
773
774 num_refs += head->node.ref_mod;
775 mutex_unlock(&head->mutex);
776 }
777 spin_unlock(&delayed_refs->lock);
778 out:
779 WARN_ON(num_refs == 0);
780 if (refs)
781 *refs = num_refs;
782 if (flags)
783 *flags = extent_flags;
784 out_free:
785 btrfs_free_path(path);
786 return ret;
787 }
788
789 /*
790 * Back reference rules. Back refs have three main goals:
791 *
792 * 1) differentiate between all holders of references to an extent so that
793 * when a reference is dropped we can make sure it was a valid reference
794 * before freeing the extent.
795 *
796 * 2) Provide enough information to quickly find the holders of an extent
797 * if we notice a given block is corrupted or bad.
798 *
799 * 3) Make it easy to migrate blocks for FS shrinking or storage pool
800 * maintenance. This is actually the same as #2, but with a slightly
801 * different use case.
802 *
803 * There are two kinds of back refs. The implicit back refs is optimized
804 * for pointers in non-shared tree blocks. For a given pointer in a block,
805 * back refs of this kind provide information about the block's owner tree
806 * and the pointer's key. These information allow us to find the block by
807 * b-tree searching. The full back refs is for pointers in tree blocks not
808 * referenced by their owner trees. The location of tree block is recorded
809 * in the back refs. Actually the full back refs is generic, and can be
810 * used in all cases the implicit back refs is used. The major shortcoming
811 * of the full back refs is its overhead. Every time a tree block gets
812 * COWed, we have to update back refs entry for all pointers in it.
813 *
814 * For a newly allocated tree block, we use implicit back refs for
815 * pointers in it. This means most tree related operations only involve
816 * implicit back refs. For a tree block created in old transaction, the
817 * only way to drop a reference to it is COW it. So we can detect the
818 * event that tree block loses its owner tree's reference and do the
819 * back refs conversion.
820 *
821 * When a tree block is COW'd through a tree, there are four cases:
822 *
823 * The reference count of the block is one and the tree is the block's
824 * owner tree. Nothing to do in this case.
825 *
826 * The reference count of the block is one and the tree is not the
827 * block's owner tree. In this case, full back refs is used for pointers
828 * in the block. Remove these full back refs, add implicit back refs for
829 * every pointers in the new block.
830 *
831 * The reference count of the block is greater than one and the tree is
832 * the block's owner tree. In this case, implicit back refs is used for
833 * pointers in the block. Add full back refs for every pointers in the
834 * block, increase lower level extents' reference counts. The original
835 * implicit back refs are entailed to the new block.
836 *
837 * The reference count of the block is greater than one and the tree is
838 * not the block's owner tree. Add implicit back refs for every pointer in
839 * the new block, increase lower level extents' reference count.
840 *
841 * Back Reference Key composing:
842 *
843 * The key objectid corresponds to the first byte in the extent,
844 * The key type is used to differentiate between types of back refs.
845 * There are different meanings of the key offset for different types
846 * of back refs.
847 *
848 * File extents can be referenced by:
849 *
850 * - multiple snapshots, subvolumes, or different generations in one subvol
851 * - different files inside a single subvolume
852 * - different offsets inside a file (bookend extents in file.c)
853 *
854 * The extent ref structure for the implicit back refs has fields for:
855 *
856 * - Objectid of the subvolume root
857 * - objectid of the file holding the reference
858 * - original offset in the file
859 * - how many bookend extents
860 *
861 * The key offset for the implicit back refs is hash of the first
862 * three fields.
863 *
864 * The extent ref structure for the full back refs has field for:
865 *
866 * - number of pointers in the tree leaf
867 *
868 * The key offset for the implicit back refs is the first byte of
869 * the tree leaf
870 *
871 * When a file extent is allocated, The implicit back refs is used.
872 * the fields are filled in:
873 *
874 * (root_key.objectid, inode objectid, offset in file, 1)
875 *
876 * When a file extent is removed file truncation, we find the
877 * corresponding implicit back refs and check the following fields:
878 *
879 * (btrfs_header_owner(leaf), inode objectid, offset in file)
880 *
881 * Btree extents can be referenced by:
882 *
883 * - Different subvolumes
884 *
885 * Both the implicit back refs and the full back refs for tree blocks
886 * only consist of key. The key offset for the implicit back refs is
887 * objectid of block's owner tree. The key offset for the full back refs
888 * is the first byte of parent block.
889 *
890 * When implicit back refs is used, information about the lowest key and
891 * level of the tree block are required. These information are stored in
892 * tree block info structure.
893 */
894
895 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
896 static int convert_extent_item_v0(struct btrfs_trans_handle *trans,
897 struct btrfs_root *root,
898 struct btrfs_path *path,
899 u64 owner, u32 extra_size)
900 {
901 struct btrfs_extent_item *item;
902 struct btrfs_extent_item_v0 *ei0;
903 struct btrfs_extent_ref_v0 *ref0;
904 struct btrfs_tree_block_info *bi;
905 struct extent_buffer *leaf;
906 struct btrfs_key key;
907 struct btrfs_key found_key;
908 u32 new_size = sizeof(*item);
909 u64 refs;
910 int ret;
911
912 leaf = path->nodes[0];
913 BUG_ON(btrfs_item_size_nr(leaf, path->slots[0]) != sizeof(*ei0));
914
915 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
916 ei0 = btrfs_item_ptr(leaf, path->slots[0],
917 struct btrfs_extent_item_v0);
918 refs = btrfs_extent_refs_v0(leaf, ei0);
919
920 if (owner == (u64)-1) {
921 while (1) {
922 if (path->slots[0] >= btrfs_header_nritems(leaf)) {
923 ret = btrfs_next_leaf(root, path);
924 if (ret < 0)
925 return ret;
926 BUG_ON(ret > 0);
927 leaf = path->nodes[0];
928 }
929 btrfs_item_key_to_cpu(leaf, &found_key,
930 path->slots[0]);
931 BUG_ON(key.objectid != found_key.objectid);
932 if (found_key.type != BTRFS_EXTENT_REF_V0_KEY) {
933 path->slots[0]++;
934 continue;
935 }
936 ref0 = btrfs_item_ptr(leaf, path->slots[0],
937 struct btrfs_extent_ref_v0);
938 owner = btrfs_ref_objectid_v0(leaf, ref0);
939 break;
940 }
941 }
942 btrfs_release_path(path);
943
944 if (owner < BTRFS_FIRST_FREE_OBJECTID)
945 new_size += sizeof(*bi);
946
947 new_size -= sizeof(*ei0);
948 ret = btrfs_search_slot(trans, root, &key, path,
949 new_size + extra_size, 1);
950 if (ret < 0)
951 return ret;
952 BUG_ON(ret);
953
954 ret = btrfs_extend_item(trans, root, path, new_size);
955
956 leaf = path->nodes[0];
957 item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
958 btrfs_set_extent_refs(leaf, item, refs);
959 /* FIXME: get real generation */
960 btrfs_set_extent_generation(leaf, item, 0);
961 if (owner < BTRFS_FIRST_FREE_OBJECTID) {
962 btrfs_set_extent_flags(leaf, item,
963 BTRFS_EXTENT_FLAG_TREE_BLOCK |
964 BTRFS_BLOCK_FLAG_FULL_BACKREF);
965 bi = (struct btrfs_tree_block_info *)(item + 1);
966 /* FIXME: get first key of the block */
967 memset_extent_buffer(leaf, 0, (unsigned long)bi, sizeof(*bi));
968 btrfs_set_tree_block_level(leaf, bi, (int)owner);
969 } else {
970 btrfs_set_extent_flags(leaf, item, BTRFS_EXTENT_FLAG_DATA);
971 }
972 btrfs_mark_buffer_dirty(leaf);
973 return 0;
974 }
975 #endif
976
977 static u64 hash_extent_data_ref(u64 root_objectid, u64 owner, u64 offset)
978 {
979 u32 high_crc = ~(u32)0;
980 u32 low_crc = ~(u32)0;
981 __le64 lenum;
982
983 lenum = cpu_to_le64(root_objectid);
984 high_crc = crc32c(high_crc, &lenum, sizeof(lenum));
985 lenum = cpu_to_le64(owner);
986 low_crc = crc32c(low_crc, &lenum, sizeof(lenum));
987 lenum = cpu_to_le64(offset);
988 low_crc = crc32c(low_crc, &lenum, sizeof(lenum));
989
990 return ((u64)high_crc << 31) ^ (u64)low_crc;
991 }
992
993 static u64 hash_extent_data_ref_item(struct extent_buffer *leaf,
994 struct btrfs_extent_data_ref *ref)
995 {
996 return hash_extent_data_ref(btrfs_extent_data_ref_root(leaf, ref),
997 btrfs_extent_data_ref_objectid(leaf, ref),
998 btrfs_extent_data_ref_offset(leaf, ref));
999 }
1000
1001 static int match_extent_data_ref(struct extent_buffer *leaf,
1002 struct btrfs_extent_data_ref *ref,
1003 u64 root_objectid, u64 owner, u64 offset)
1004 {
1005 if (btrfs_extent_data_ref_root(leaf, ref) != root_objectid ||
1006 btrfs_extent_data_ref_objectid(leaf, ref) != owner ||
1007 btrfs_extent_data_ref_offset(leaf, ref) != offset)
1008 return 0;
1009 return 1;
1010 }
1011
1012 static noinline int lookup_extent_data_ref(struct btrfs_trans_handle *trans,
1013 struct btrfs_root *root,
1014 struct btrfs_path *path,
1015 u64 bytenr, u64 parent,
1016 u64 root_objectid,
1017 u64 owner, u64 offset)
1018 {
1019 struct btrfs_key key;
1020 struct btrfs_extent_data_ref *ref;
1021 struct extent_buffer *leaf;
1022 u32 nritems;
1023 int ret;
1024 int recow;
1025 int err = -ENOENT;
1026
1027 key.objectid = bytenr;
1028 if (parent) {
1029 key.type = BTRFS_SHARED_DATA_REF_KEY;
1030 key.offset = parent;
1031 } else {
1032 key.type = BTRFS_EXTENT_DATA_REF_KEY;
1033 key.offset = hash_extent_data_ref(root_objectid,
1034 owner, offset);
1035 }
1036 again:
1037 recow = 0;
1038 ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
1039 if (ret < 0) {
1040 err = ret;
1041 goto fail;
1042 }
1043
1044 if (parent) {
1045 if (!ret)
1046 return 0;
1047 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
1048 key.type = BTRFS_EXTENT_REF_V0_KEY;
1049 btrfs_release_path(path);
1050 ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
1051 if (ret < 0) {
1052 err = ret;
1053 goto fail;
1054 }
1055 if (!ret)
1056 return 0;
1057 #endif
1058 goto fail;
1059 }
1060
1061 leaf = path->nodes[0];
1062 nritems = btrfs_header_nritems(leaf);
1063 while (1) {
1064 if (path->slots[0] >= nritems) {
1065 ret = btrfs_next_leaf(root, path);
1066 if (ret < 0)
1067 err = ret;
1068 if (ret)
1069 goto fail;
1070
1071 leaf = path->nodes[0];
1072 nritems = btrfs_header_nritems(leaf);
1073 recow = 1;
1074 }
1075
1076 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
1077 if (key.objectid != bytenr ||
1078 key.type != BTRFS_EXTENT_DATA_REF_KEY)
1079 goto fail;
1080
1081 ref = btrfs_item_ptr(leaf, path->slots[0],
1082 struct btrfs_extent_data_ref);
1083
1084 if (match_extent_data_ref(leaf, ref, root_objectid,
1085 owner, offset)) {
1086 if (recow) {
1087 btrfs_release_path(path);
1088 goto again;
1089 }
1090 err = 0;
1091 break;
1092 }
1093 path->slots[0]++;
1094 }
1095 fail:
1096 return err;
1097 }
1098
1099 static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans,
1100 struct btrfs_root *root,
1101 struct btrfs_path *path,
1102 u64 bytenr, u64 parent,
1103 u64 root_objectid, u64 owner,
1104 u64 offset, int refs_to_add)
1105 {
1106 struct btrfs_key key;
1107 struct extent_buffer *leaf;
1108 u32 size;
1109 u32 num_refs;
1110 int ret;
1111
1112 key.objectid = bytenr;
1113 if (parent) {
1114 key.type = BTRFS_SHARED_DATA_REF_KEY;
1115 key.offset = parent;
1116 size = sizeof(struct btrfs_shared_data_ref);
1117 } else {
1118 key.type = BTRFS_EXTENT_DATA_REF_KEY;
1119 key.offset = hash_extent_data_ref(root_objectid,
1120 owner, offset);
1121 size = sizeof(struct btrfs_extent_data_ref);
1122 }
1123
1124 ret = btrfs_insert_empty_item(trans, root, path, &key, size);
1125 if (ret && ret != -EEXIST)
1126 goto fail;
1127
1128 leaf = path->nodes[0];
1129 if (parent) {
1130 struct btrfs_shared_data_ref *ref;
1131 ref = btrfs_item_ptr(leaf, path->slots[0],
1132 struct btrfs_shared_data_ref);
1133 if (ret == 0) {
1134 btrfs_set_shared_data_ref_count(leaf, ref, refs_to_add);
1135 } else {
1136 num_refs = btrfs_shared_data_ref_count(leaf, ref);
1137 num_refs += refs_to_add;
1138 btrfs_set_shared_data_ref_count(leaf, ref, num_refs);
1139 }
1140 } else {
1141 struct btrfs_extent_data_ref *ref;
1142 while (ret == -EEXIST) {
1143 ref = btrfs_item_ptr(leaf, path->slots[0],
1144 struct btrfs_extent_data_ref);
1145 if (match_extent_data_ref(leaf, ref, root_objectid,
1146 owner, offset))
1147 break;
1148 btrfs_release_path(path);
1149 key.offset++;
1150 ret = btrfs_insert_empty_item(trans, root, path, &key,
1151 size);
1152 if (ret && ret != -EEXIST)
1153 goto fail;
1154
1155 leaf = path->nodes[0];
1156 }
1157 ref = btrfs_item_ptr(leaf, path->slots[0],
1158 struct btrfs_extent_data_ref);
1159 if (ret == 0) {
1160 btrfs_set_extent_data_ref_root(leaf, ref,
1161 root_objectid);
1162 btrfs_set_extent_data_ref_objectid(leaf, ref, owner);
1163 btrfs_set_extent_data_ref_offset(leaf, ref, offset);
1164 btrfs_set_extent_data_ref_count(leaf, ref, refs_to_add);
1165 } else {
1166 num_refs = btrfs_extent_data_ref_count(leaf, ref);
1167 num_refs += refs_to_add;
1168 btrfs_set_extent_data_ref_count(leaf, ref, num_refs);
1169 }
1170 }
1171 btrfs_mark_buffer_dirty(leaf);
1172 ret = 0;
1173 fail:
1174 btrfs_release_path(path);
1175 return ret;
1176 }
1177
1178 static noinline int remove_extent_data_ref(struct btrfs_trans_handle *trans,
1179 struct btrfs_root *root,
1180 struct btrfs_path *path,
1181 int refs_to_drop)
1182 {
1183 struct btrfs_key key;
1184 struct btrfs_extent_data_ref *ref1 = NULL;
1185 struct btrfs_shared_data_ref *ref2 = NULL;
1186 struct extent_buffer *leaf;
1187 u32 num_refs = 0;
1188 int ret = 0;
1189
1190 leaf = path->nodes[0];
1191 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
1192
1193 if (key.type == BTRFS_EXTENT_DATA_REF_KEY) {
1194 ref1 = btrfs_item_ptr(leaf, path->slots[0],
1195 struct btrfs_extent_data_ref);
1196 num_refs = btrfs_extent_data_ref_count(leaf, ref1);
1197 } else if (key.type == BTRFS_SHARED_DATA_REF_KEY) {
1198 ref2 = btrfs_item_ptr(leaf, path->slots[0],
1199 struct btrfs_shared_data_ref);
1200 num_refs = btrfs_shared_data_ref_count(leaf, ref2);
1201 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
1202 } else if (key.type == BTRFS_EXTENT_REF_V0_KEY) {
1203 struct btrfs_extent_ref_v0 *ref0;
1204 ref0 = btrfs_item_ptr(leaf, path->slots[0],
1205 struct btrfs_extent_ref_v0);
1206 num_refs = btrfs_ref_count_v0(leaf, ref0);
1207 #endif
1208 } else {
1209 BUG();
1210 }
1211
1212 BUG_ON(num_refs < refs_to_drop);
1213 num_refs -= refs_to_drop;
1214
1215 if (num_refs == 0) {
1216 ret = btrfs_del_item(trans, root, path);
1217 } else {
1218 if (key.type == BTRFS_EXTENT_DATA_REF_KEY)
1219 btrfs_set_extent_data_ref_count(leaf, ref1, num_refs);
1220 else if (key.type == BTRFS_SHARED_DATA_REF_KEY)
1221 btrfs_set_shared_data_ref_count(leaf, ref2, num_refs);
1222 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
1223 else {
1224 struct btrfs_extent_ref_v0 *ref0;
1225 ref0 = btrfs_item_ptr(leaf, path->slots[0],
1226 struct btrfs_extent_ref_v0);
1227 btrfs_set_ref_count_v0(leaf, ref0, num_refs);
1228 }
1229 #endif
1230 btrfs_mark_buffer_dirty(leaf);
1231 }
1232 return ret;
1233 }
1234
1235 static noinline u32 extent_data_ref_count(struct btrfs_root *root,
1236 struct btrfs_path *path,
1237 struct btrfs_extent_inline_ref *iref)
1238 {
1239 struct btrfs_key key;
1240 struct extent_buffer *leaf;
1241 struct btrfs_extent_data_ref *ref1;
1242 struct btrfs_shared_data_ref *ref2;
1243 u32 num_refs = 0;
1244
1245 leaf = path->nodes[0];
1246 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
1247 if (iref) {
1248 if (btrfs_extent_inline_ref_type(leaf, iref) ==
1249 BTRFS_EXTENT_DATA_REF_KEY) {
1250 ref1 = (struct btrfs_extent_data_ref *)(&iref->offset);
1251 num_refs = btrfs_extent_data_ref_count(leaf, ref1);
1252 } else {
1253 ref2 = (struct btrfs_shared_data_ref *)(iref + 1);
1254 num_refs = btrfs_shared_data_ref_count(leaf, ref2);
1255 }
1256 } else if (key.type == BTRFS_EXTENT_DATA_REF_KEY) {
1257 ref1 = btrfs_item_ptr(leaf, path->slots[0],
1258 struct btrfs_extent_data_ref);
1259 num_refs = btrfs_extent_data_ref_count(leaf, ref1);
1260 } else if (key.type == BTRFS_SHARED_DATA_REF_KEY) {
1261 ref2 = btrfs_item_ptr(leaf, path->slots[0],
1262 struct btrfs_shared_data_ref);
1263 num_refs = btrfs_shared_data_ref_count(leaf, ref2);
1264 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
1265 } else if (key.type == BTRFS_EXTENT_REF_V0_KEY) {
1266 struct btrfs_extent_ref_v0 *ref0;
1267 ref0 = btrfs_item_ptr(leaf, path->slots[0],
1268 struct btrfs_extent_ref_v0);
1269 num_refs = btrfs_ref_count_v0(leaf, ref0);
1270 #endif
1271 } else {
1272 WARN_ON(1);
1273 }
1274 return num_refs;
1275 }
1276
1277 static noinline int lookup_tree_block_ref(struct btrfs_trans_handle *trans,
1278 struct btrfs_root *root,
1279 struct btrfs_path *path,
1280 u64 bytenr, u64 parent,
1281 u64 root_objectid)
1282 {
1283 struct btrfs_key key;
1284 int ret;
1285
1286 key.objectid = bytenr;
1287 if (parent) {
1288 key.type = BTRFS_SHARED_BLOCK_REF_KEY;
1289 key.offset = parent;
1290 } else {
1291 key.type = BTRFS_TREE_BLOCK_REF_KEY;
1292 key.offset = root_objectid;
1293 }
1294
1295 ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
1296 if (ret > 0)
1297 ret = -ENOENT;
1298 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
1299 if (ret == -ENOENT && parent) {
1300 btrfs_release_path(path);
1301 key.type = BTRFS_EXTENT_REF_V0_KEY;
1302 ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
1303 if (ret > 0)
1304 ret = -ENOENT;
1305 }
1306 #endif
1307 return ret;
1308 }
1309
1310 static noinline int insert_tree_block_ref(struct btrfs_trans_handle *trans,
1311 struct btrfs_root *root,
1312 struct btrfs_path *path,
1313 u64 bytenr, u64 parent,
1314 u64 root_objectid)
1315 {
1316 struct btrfs_key key;
1317 int ret;
1318
1319 key.objectid = bytenr;
1320 if (parent) {
1321 key.type = BTRFS_SHARED_BLOCK_REF_KEY;
1322 key.offset = parent;
1323 } else {
1324 key.type = BTRFS_TREE_BLOCK_REF_KEY;
1325 key.offset = root_objectid;
1326 }
1327
1328 ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
1329 btrfs_release_path(path);
1330 return ret;
1331 }
1332
1333 static inline int extent_ref_type(u64 parent, u64 owner)
1334 {
1335 int type;
1336 if (owner < BTRFS_FIRST_FREE_OBJECTID) {
1337 if (parent > 0)
1338 type = BTRFS_SHARED_BLOCK_REF_KEY;
1339 else
1340 type = BTRFS_TREE_BLOCK_REF_KEY;
1341 } else {
1342 if (parent > 0)
1343 type = BTRFS_SHARED_DATA_REF_KEY;
1344 else
1345 type = BTRFS_EXTENT_DATA_REF_KEY;
1346 }
1347 return type;
1348 }
1349
1350 static int find_next_key(struct btrfs_path *path, int level,
1351 struct btrfs_key *key)
1352
1353 {
1354 for (; level < BTRFS_MAX_LEVEL; level++) {
1355 if (!path->nodes[level])
1356 break;
1357 if (path->slots[level] + 1 >=
1358 btrfs_header_nritems(path->nodes[level]))
1359 continue;
1360 if (level == 0)
1361 btrfs_item_key_to_cpu(path->nodes[level], key,
1362 path->slots[level] + 1);
1363 else
1364 btrfs_node_key_to_cpu(path->nodes[level], key,
1365 path->slots[level] + 1);
1366 return 0;
1367 }
1368 return 1;
1369 }
1370
1371 /*
1372 * look for inline back ref. if back ref is found, *ref_ret is set
1373 * to the address of inline back ref, and 0 is returned.
1374 *
1375 * if back ref isn't found, *ref_ret is set to the address where it
1376 * should be inserted, and -ENOENT is returned.
1377 *
1378 * if insert is true and there are too many inline back refs, the path
1379 * points to the extent item, and -EAGAIN is returned.
1380 *
1381 * NOTE: inline back refs are ordered in the same way that back ref
1382 * items in the tree are ordered.
1383 */
1384 static noinline_for_stack
1385 int lookup_inline_extent_backref(struct btrfs_trans_handle *trans,
1386 struct btrfs_root *root,
1387 struct btrfs_path *path,
1388 struct btrfs_extent_inline_ref **ref_ret,
1389 u64 bytenr, u64 num_bytes,
1390 u64 parent, u64 root_objectid,
1391 u64 owner, u64 offset, int insert)
1392 {
1393 struct btrfs_key key;
1394 struct extent_buffer *leaf;
1395 struct btrfs_extent_item *ei;
1396 struct btrfs_extent_inline_ref *iref;
1397 u64 flags;
1398 u64 item_size;
1399 unsigned long ptr;
1400 unsigned long end;
1401 int extra_size;
1402 int type;
1403 int want;
1404 int ret;
1405 int err = 0;
1406
1407 key.objectid = bytenr;
1408 key.type = BTRFS_EXTENT_ITEM_KEY;
1409 key.offset = num_bytes;
1410
1411 want = extent_ref_type(parent, owner);
1412 if (insert) {
1413 extra_size = btrfs_extent_inline_ref_size(want);
1414 path->keep_locks = 1;
1415 } else
1416 extra_size = -1;
1417 ret = btrfs_search_slot(trans, root, &key, path, extra_size, 1);
1418 if (ret < 0) {
1419 err = ret;
1420 goto out;
1421 }
1422 BUG_ON(ret);
1423
1424 leaf = path->nodes[0];
1425 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1426 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
1427 if (item_size < sizeof(*ei)) {
1428 if (!insert) {
1429 err = -ENOENT;
1430 goto out;
1431 }
1432 ret = convert_extent_item_v0(trans, root, path, owner,
1433 extra_size);
1434 if (ret < 0) {
1435 err = ret;
1436 goto out;
1437 }
1438 leaf = path->nodes[0];
1439 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1440 }
1441 #endif
1442 BUG_ON(item_size < sizeof(*ei));
1443
1444 ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1445 flags = btrfs_extent_flags(leaf, ei);
1446
1447 ptr = (unsigned long)(ei + 1);
1448 end = (unsigned long)ei + item_size;
1449
1450 if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
1451 ptr += sizeof(struct btrfs_tree_block_info);
1452 BUG_ON(ptr > end);
1453 } else {
1454 BUG_ON(!(flags & BTRFS_EXTENT_FLAG_DATA));
1455 }
1456
1457 err = -ENOENT;
1458 while (1) {
1459 if (ptr >= end) {
1460 WARN_ON(ptr > end);
1461 break;
1462 }
1463 iref = (struct btrfs_extent_inline_ref *)ptr;
1464 type = btrfs_extent_inline_ref_type(leaf, iref);
1465 if (want < type)
1466 break;
1467 if (want > type) {
1468 ptr += btrfs_extent_inline_ref_size(type);
1469 continue;
1470 }
1471
1472 if (type == BTRFS_EXTENT_DATA_REF_KEY) {
1473 struct btrfs_extent_data_ref *dref;
1474 dref = (struct btrfs_extent_data_ref *)(&iref->offset);
1475 if (match_extent_data_ref(leaf, dref, root_objectid,
1476 owner, offset)) {
1477 err = 0;
1478 break;
1479 }
1480 if (hash_extent_data_ref_item(leaf, dref) <
1481 hash_extent_data_ref(root_objectid, owner, offset))
1482 break;
1483 } else {
1484 u64 ref_offset;
1485 ref_offset = btrfs_extent_inline_ref_offset(leaf, iref);
1486 if (parent > 0) {
1487 if (parent == ref_offset) {
1488 err = 0;
1489 break;
1490 }
1491 if (ref_offset < parent)
1492 break;
1493 } else {
1494 if (root_objectid == ref_offset) {
1495 err = 0;
1496 break;
1497 }
1498 if (ref_offset < root_objectid)
1499 break;
1500 }
1501 }
1502 ptr += btrfs_extent_inline_ref_size(type);
1503 }
1504 if (err == -ENOENT && insert) {
1505 if (item_size + extra_size >=
1506 BTRFS_MAX_EXTENT_ITEM_SIZE(root)) {
1507 err = -EAGAIN;
1508 goto out;
1509 }
1510 /*
1511 * To add new inline back ref, we have to make sure
1512 * there is no corresponding back ref item.
1513 * For simplicity, we just do not add new inline back
1514 * ref if there is any kind of item for this block
1515 */
1516 if (find_next_key(path, 0, &key) == 0 &&
1517 key.objectid == bytenr &&
1518 key.type < BTRFS_BLOCK_GROUP_ITEM_KEY) {
1519 err = -EAGAIN;
1520 goto out;
1521 }
1522 }
1523 *ref_ret = (struct btrfs_extent_inline_ref *)ptr;
1524 out:
1525 if (insert) {
1526 path->keep_locks = 0;
1527 btrfs_unlock_up_safe(path, 1);
1528 }
1529 return err;
1530 }
1531
1532 /*
1533 * helper to add new inline back ref
1534 */
1535 static noinline_for_stack
1536 int setup_inline_extent_backref(struct btrfs_trans_handle *trans,
1537 struct btrfs_root *root,
1538 struct btrfs_path *path,
1539 struct btrfs_extent_inline_ref *iref,
1540 u64 parent, u64 root_objectid,
1541 u64 owner, u64 offset, int refs_to_add,
1542 struct btrfs_delayed_extent_op *extent_op)
1543 {
1544 struct extent_buffer *leaf;
1545 struct btrfs_extent_item *ei;
1546 unsigned long ptr;
1547 unsigned long end;
1548 unsigned long item_offset;
1549 u64 refs;
1550 int size;
1551 int type;
1552 int ret;
1553
1554 leaf = path->nodes[0];
1555 ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1556 item_offset = (unsigned long)iref - (unsigned long)ei;
1557
1558 type = extent_ref_type(parent, owner);
1559 size = btrfs_extent_inline_ref_size(type);
1560
1561 ret = btrfs_extend_item(trans, root, path, size);
1562
1563 ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1564 refs = btrfs_extent_refs(leaf, ei);
1565 refs += refs_to_add;
1566 btrfs_set_extent_refs(leaf, ei, refs);
1567 if (extent_op)
1568 __run_delayed_extent_op(extent_op, leaf, ei);
1569
1570 ptr = (unsigned long)ei + item_offset;
1571 end = (unsigned long)ei + btrfs_item_size_nr(leaf, path->slots[0]);
1572 if (ptr < end - size)
1573 memmove_extent_buffer(leaf, ptr + size, ptr,
1574 end - size - ptr);
1575
1576 iref = (struct btrfs_extent_inline_ref *)ptr;
1577 btrfs_set_extent_inline_ref_type(leaf, iref, type);
1578 if (type == BTRFS_EXTENT_DATA_REF_KEY) {
1579 struct btrfs_extent_data_ref *dref;
1580 dref = (struct btrfs_extent_data_ref *)(&iref->offset);
1581 btrfs_set_extent_data_ref_root(leaf, dref, root_objectid);
1582 btrfs_set_extent_data_ref_objectid(leaf, dref, owner);
1583 btrfs_set_extent_data_ref_offset(leaf, dref, offset);
1584 btrfs_set_extent_data_ref_count(leaf, dref, refs_to_add);
1585 } else if (type == BTRFS_SHARED_DATA_REF_KEY) {
1586 struct btrfs_shared_data_ref *sref;
1587 sref = (struct btrfs_shared_data_ref *)(iref + 1);
1588 btrfs_set_shared_data_ref_count(leaf, sref, refs_to_add);
1589 btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
1590 } else if (type == BTRFS_SHARED_BLOCK_REF_KEY) {
1591 btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
1592 } else {
1593 btrfs_set_extent_inline_ref_offset(leaf, iref, root_objectid);
1594 }
1595 btrfs_mark_buffer_dirty(leaf);
1596 return 0;
1597 }
1598
1599 static int lookup_extent_backref(struct btrfs_trans_handle *trans,
1600 struct btrfs_root *root,
1601 struct btrfs_path *path,
1602 struct btrfs_extent_inline_ref **ref_ret,
1603 u64 bytenr, u64 num_bytes, u64 parent,
1604 u64 root_objectid, u64 owner, u64 offset)
1605 {
1606 int ret;
1607
1608 ret = lookup_inline_extent_backref(trans, root, path, ref_ret,
1609 bytenr, num_bytes, parent,
1610 root_objectid, owner, offset, 0);
1611 if (ret != -ENOENT)
1612 return ret;
1613
1614 btrfs_release_path(path);
1615 *ref_ret = NULL;
1616
1617 if (owner < BTRFS_FIRST_FREE_OBJECTID) {
1618 ret = lookup_tree_block_ref(trans, root, path, bytenr, parent,
1619 root_objectid);
1620 } else {
1621 ret = lookup_extent_data_ref(trans, root, path, bytenr, parent,
1622 root_objectid, owner, offset);
1623 }
1624 return ret;
1625 }
1626
1627 /*
1628 * helper to update/remove inline back ref
1629 */
1630 static noinline_for_stack
1631 int update_inline_extent_backref(struct btrfs_trans_handle *trans,
1632 struct btrfs_root *root,
1633 struct btrfs_path *path,
1634 struct btrfs_extent_inline_ref *iref,
1635 int refs_to_mod,
1636 struct btrfs_delayed_extent_op *extent_op)
1637 {
1638 struct extent_buffer *leaf;
1639 struct btrfs_extent_item *ei;
1640 struct btrfs_extent_data_ref *dref = NULL;
1641 struct btrfs_shared_data_ref *sref = NULL;
1642 unsigned long ptr;
1643 unsigned long end;
1644 u32 item_size;
1645 int size;
1646 int type;
1647 int ret;
1648 u64 refs;
1649
1650 leaf = path->nodes[0];
1651 ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1652 refs = btrfs_extent_refs(leaf, ei);
1653 WARN_ON(refs_to_mod < 0 && refs + refs_to_mod <= 0);
1654 refs += refs_to_mod;
1655 btrfs_set_extent_refs(leaf, ei, refs);
1656 if (extent_op)
1657 __run_delayed_extent_op(extent_op, leaf, ei);
1658
1659 type = btrfs_extent_inline_ref_type(leaf, iref);
1660
1661 if (type == BTRFS_EXTENT_DATA_REF_KEY) {
1662 dref = (struct btrfs_extent_data_ref *)(&iref->offset);
1663 refs = btrfs_extent_data_ref_count(leaf, dref);
1664 } else if (type == BTRFS_SHARED_DATA_REF_KEY) {
1665 sref = (struct btrfs_shared_data_ref *)(iref + 1);
1666 refs = btrfs_shared_data_ref_count(leaf, sref);
1667 } else {
1668 refs = 1;
1669 BUG_ON(refs_to_mod != -1);
1670 }
1671
1672 BUG_ON(refs_to_mod < 0 && refs < -refs_to_mod);
1673 refs += refs_to_mod;
1674
1675 if (refs > 0) {
1676 if (type == BTRFS_EXTENT_DATA_REF_KEY)
1677 btrfs_set_extent_data_ref_count(leaf, dref, refs);
1678 else
1679 btrfs_set_shared_data_ref_count(leaf, sref, refs);
1680 } else {
1681 size = btrfs_extent_inline_ref_size(type);
1682 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1683 ptr = (unsigned long)iref;
1684 end = (unsigned long)ei + item_size;
1685 if (ptr + size < end)
1686 memmove_extent_buffer(leaf, ptr, ptr + size,
1687 end - ptr - size);
1688 item_size -= size;
1689 ret = btrfs_truncate_item(trans, root, path, item_size, 1);
1690 }
1691 btrfs_mark_buffer_dirty(leaf);
1692 return 0;
1693 }
1694
1695 static noinline_for_stack
1696 int insert_inline_extent_backref(struct btrfs_trans_handle *trans,
1697 struct btrfs_root *root,
1698 struct btrfs_path *path,
1699 u64 bytenr, u64 num_bytes, u64 parent,
1700 u64 root_objectid, u64 owner,
1701 u64 offset, int refs_to_add,
1702 struct btrfs_delayed_extent_op *extent_op)
1703 {
1704 struct btrfs_extent_inline_ref *iref;
1705 int ret;
1706
1707 ret = lookup_inline_extent_backref(trans, root, path, &iref,
1708 bytenr, num_bytes, parent,
1709 root_objectid, owner, offset, 1);
1710 if (ret == 0) {
1711 BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID);
1712 ret = update_inline_extent_backref(trans, root, path, iref,
1713 refs_to_add, extent_op);
1714 } else if (ret == -ENOENT) {
1715 ret = setup_inline_extent_backref(trans, root, path, iref,
1716 parent, root_objectid,
1717 owner, offset, refs_to_add,
1718 extent_op);
1719 }
1720 return ret;
1721 }
1722
1723 static int insert_extent_backref(struct btrfs_trans_handle *trans,
1724 struct btrfs_root *root,
1725 struct btrfs_path *path,
1726 u64 bytenr, u64 parent, u64 root_objectid,
1727 u64 owner, u64 offset, int refs_to_add)
1728 {
1729 int ret;
1730 if (owner < BTRFS_FIRST_FREE_OBJECTID) {
1731 BUG_ON(refs_to_add != 1);
1732 ret = insert_tree_block_ref(trans, root, path, bytenr,
1733 parent, root_objectid);
1734 } else {
1735 ret = insert_extent_data_ref(trans, root, path, bytenr,
1736 parent, root_objectid,
1737 owner, offset, refs_to_add);
1738 }
1739 return ret;
1740 }
1741
1742 static int remove_extent_backref(struct btrfs_trans_handle *trans,
1743 struct btrfs_root *root,
1744 struct btrfs_path *path,
1745 struct btrfs_extent_inline_ref *iref,
1746 int refs_to_drop, int is_data)
1747 {
1748 int ret;
1749
1750 BUG_ON(!is_data && refs_to_drop != 1);
1751 if (iref) {
1752 ret = update_inline_extent_backref(trans, root, path, iref,
1753 -refs_to_drop, NULL);
1754 } else if (is_data) {
1755 ret = remove_extent_data_ref(trans, root, path, refs_to_drop);
1756 } else {
1757 ret = btrfs_del_item(trans, root, path);
1758 }
1759 return ret;
1760 }
1761
1762 static int btrfs_issue_discard(struct block_device *bdev,
1763 u64 start, u64 len)
1764 {
1765 return blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_NOFS, 0);
1766 }
1767
1768 static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr,
1769 u64 num_bytes, u64 *actual_bytes)
1770 {
1771 int ret;
1772 u64 discarded_bytes = 0;
1773 struct btrfs_multi_bio *multi = NULL;
1774
1775
1776 /* Tell the block device(s) that the sectors can be discarded */
1777 ret = btrfs_map_block(&root->fs_info->mapping_tree, REQ_DISCARD,
1778 bytenr, &num_bytes, &multi, 0);
1779 if (!ret) {
1780 struct btrfs_bio_stripe *stripe = multi->stripes;
1781 int i;
1782
1783
1784 for (i = 0; i < multi->num_stripes; i++, stripe++) {
1785 ret = btrfs_issue_discard(stripe->dev->bdev,
1786 stripe->physical,
1787 stripe->length);
1788 if (!ret)
1789 discarded_bytes += stripe->length;
1790 else if (ret != -EOPNOTSUPP)
1791 break;
1792 }
1793 kfree(multi);
1794 }
1795 if (discarded_bytes && ret == -EOPNOTSUPP)
1796 ret = 0;
1797
1798 if (actual_bytes)
1799 *actual_bytes = discarded_bytes;
1800
1801
1802 return ret;
1803 }
1804
1805 int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
1806 struct btrfs_root *root,
1807 u64 bytenr, u64 num_bytes, u64 parent,
1808 u64 root_objectid, u64 owner, u64 offset)
1809 {
1810 int ret;
1811 BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID &&
1812 root_objectid == BTRFS_TREE_LOG_OBJECTID);
1813
1814 if (owner < BTRFS_FIRST_FREE_OBJECTID) {
1815 ret = btrfs_add_delayed_tree_ref(trans, bytenr, num_bytes,
1816 parent, root_objectid, (int)owner,
1817 BTRFS_ADD_DELAYED_REF, NULL);
1818 } else {
1819 ret = btrfs_add_delayed_data_ref(trans, bytenr, num_bytes,
1820 parent, root_objectid, owner, offset,
1821 BTRFS_ADD_DELAYED_REF, NULL);
1822 }
1823 return ret;
1824 }
1825
1826 static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
1827 struct btrfs_root *root,
1828 u64 bytenr, u64 num_bytes,
1829 u64 parent, u64 root_objectid,
1830 u64 owner, u64 offset, int refs_to_add,
1831 struct btrfs_delayed_extent_op *extent_op)
1832 {
1833 struct btrfs_path *path;
1834 struct extent_buffer *leaf;
1835 struct btrfs_extent_item *item;
1836 u64 refs;
1837 int ret;
1838 int err = 0;
1839
1840 path = btrfs_alloc_path();
1841 if (!path)
1842 return -ENOMEM;
1843
1844 path->reada = 1;
1845 path->leave_spinning = 1;
1846 /* this will setup the path even if it fails to insert the back ref */
1847 ret = insert_inline_extent_backref(trans, root->fs_info->extent_root,
1848 path, bytenr, num_bytes, parent,
1849 root_objectid, owner, offset,
1850 refs_to_add, extent_op);
1851 if (ret == 0)
1852 goto out;
1853
1854 if (ret != -EAGAIN) {
1855 err = ret;
1856 goto out;
1857 }
1858
1859 leaf = path->nodes[0];
1860 item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1861 refs = btrfs_extent_refs(leaf, item);
1862 btrfs_set_extent_refs(leaf, item, refs + refs_to_add);
1863 if (extent_op)
1864 __run_delayed_extent_op(extent_op, leaf, item);
1865
1866 btrfs_mark_buffer_dirty(leaf);
1867 btrfs_release_path(path);
1868
1869 path->reada = 1;
1870 path->leave_spinning = 1;
1871
1872 /* now insert the actual backref */
1873 ret = insert_extent_backref(trans, root->fs_info->extent_root,
1874 path, bytenr, parent, root_objectid,
1875 owner, offset, refs_to_add);
1876 BUG_ON(ret);
1877 out:
1878 btrfs_free_path(path);
1879 return err;
1880 }
1881
1882 static int run_delayed_data_ref(struct btrfs_trans_handle *trans,
1883 struct btrfs_root *root,
1884 struct btrfs_delayed_ref_node *node,
1885 struct btrfs_delayed_extent_op *extent_op,
1886 int insert_reserved)
1887 {
1888 int ret = 0;
1889 struct btrfs_delayed_data_ref *ref;
1890 struct btrfs_key ins;
1891 u64 parent = 0;
1892 u64 ref_root = 0;
1893 u64 flags = 0;
1894
1895 ins.objectid = node->bytenr;
1896 ins.offset = node->num_bytes;
1897 ins.type = BTRFS_EXTENT_ITEM_KEY;
1898
1899 ref = btrfs_delayed_node_to_data_ref(node);
1900 if (node->type == BTRFS_SHARED_DATA_REF_KEY)
1901 parent = ref->parent;
1902 else
1903 ref_root = ref->root;
1904
1905 if (node->action == BTRFS_ADD_DELAYED_REF && insert_reserved) {
1906 if (extent_op) {
1907 BUG_ON(extent_op->update_key);
1908 flags |= extent_op->flags_to_set;
1909 }
1910 ret = alloc_reserved_file_extent(trans, root,
1911 parent, ref_root, flags,
1912 ref->objectid, ref->offset,
1913 &ins, node->ref_mod);
1914 } else if (node->action == BTRFS_ADD_DELAYED_REF) {
1915 ret = __btrfs_inc_extent_ref(trans, root, node->bytenr,
1916 node->num_bytes, parent,
1917 ref_root, ref->objectid,
1918 ref->offset, node->ref_mod,
1919 extent_op);
1920 } else if (node->action == BTRFS_DROP_DELAYED_REF) {
1921 ret = __btrfs_free_extent(trans, root, node->bytenr,
1922 node->num_bytes, parent,
1923 ref_root, ref->objectid,
1924 ref->offset, node->ref_mod,
1925 extent_op);
1926 } else {
1927 BUG();
1928 }
1929 return ret;
1930 }
1931
1932 static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op,
1933 struct extent_buffer *leaf,
1934 struct btrfs_extent_item *ei)
1935 {
1936 u64 flags = btrfs_extent_flags(leaf, ei);
1937 if (extent_op->update_flags) {
1938 flags |= extent_op->flags_to_set;
1939 btrfs_set_extent_flags(leaf, ei, flags);
1940 }
1941
1942 if (extent_op->update_key) {
1943 struct btrfs_tree_block_info *bi;
1944 BUG_ON(!(flags & BTRFS_EXTENT_FLAG_TREE_BLOCK));
1945 bi = (struct btrfs_tree_block_info *)(ei + 1);
1946 btrfs_set_tree_block_key(leaf, bi, &extent_op->key);
1947 }
1948 }
1949
1950 static int run_delayed_extent_op(struct btrfs_trans_handle *trans,
1951 struct btrfs_root *root,
1952 struct btrfs_delayed_ref_node *node,
1953 struct btrfs_delayed_extent_op *extent_op)
1954 {
1955 struct btrfs_key key;
1956 struct btrfs_path *path;
1957 struct btrfs_extent_item *ei;
1958 struct extent_buffer *leaf;
1959 u32 item_size;
1960 int ret;
1961 int err = 0;
1962
1963 path = btrfs_alloc_path();
1964 if (!path)
1965 return -ENOMEM;
1966
1967 key.objectid = node->bytenr;
1968 key.type = BTRFS_EXTENT_ITEM_KEY;
1969 key.offset = node->num_bytes;
1970
1971 path->reada = 1;
1972 path->leave_spinning = 1;
1973 ret = btrfs_search_slot(trans, root->fs_info->extent_root, &key,
1974 path, 0, 1);
1975 if (ret < 0) {
1976 err = ret;
1977 goto out;
1978 }
1979 if (ret > 0) {
1980 err = -EIO;
1981 goto out;
1982 }
1983
1984 leaf = path->nodes[0];
1985 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1986 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
1987 if (item_size < sizeof(*ei)) {
1988 ret = convert_extent_item_v0(trans, root->fs_info->extent_root,
1989 path, (u64)-1, 0);
1990 if (ret < 0) {
1991 err = ret;
1992 goto out;
1993 }
1994 leaf = path->nodes[0];
1995 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1996 }
1997 #endif
1998 BUG_ON(item_size < sizeof(*ei));
1999 ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
2000 __run_delayed_extent_op(extent_op, leaf, ei);
2001
2002 btrfs_mark_buffer_dirty(leaf);
2003 out:
2004 btrfs_free_path(path);
2005 return err;
2006 }
2007
2008 static int run_delayed_tree_ref(struct btrfs_trans_handle *trans,
2009 struct btrfs_root *root,
2010 struct btrfs_delayed_ref_node *node,
2011 struct btrfs_delayed_extent_op *extent_op,
2012 int insert_reserved)
2013 {
2014 int ret = 0;
2015 struct btrfs_delayed_tree_ref *ref;
2016 struct btrfs_key ins;
2017 u64 parent = 0;
2018 u64 ref_root = 0;
2019
2020 ins.objectid = node->bytenr;
2021 ins.offset = node->num_bytes;
2022 ins.type = BTRFS_EXTENT_ITEM_KEY;
2023
2024 ref = btrfs_delayed_node_to_tree_ref(node);
2025 if (node->type == BTRFS_SHARED_BLOCK_REF_KEY)
2026 parent = ref->parent;
2027 else
2028 ref_root = ref->root;
2029
2030 BUG_ON(node->ref_mod != 1);
2031 if (node->action == BTRFS_ADD_DELAYED_REF && insert_reserved) {
2032 BUG_ON(!extent_op || !extent_op->update_flags ||
2033 !extent_op->update_key);
2034 ret = alloc_reserved_tree_block(trans, root,
2035 parent, ref_root,
2036 extent_op->flags_to_set,
2037 &extent_op->key,
2038 ref->level, &ins);
2039 } else if (node->action == BTRFS_ADD_DELAYED_REF) {
2040 ret = __btrfs_inc_extent_ref(trans, root, node->bytenr,
2041 node->num_bytes, parent, ref_root,
2042 ref->level, 0, 1, extent_op);
2043 } else if (node->action == BTRFS_DROP_DELAYED_REF) {
2044 ret = __btrfs_free_extent(trans, root, node->bytenr,
2045 node->num_bytes, parent, ref_root,
2046 ref->level, 0, 1, extent_op);
2047 } else {
2048 BUG();
2049 }
2050 return ret;
2051 }
2052
2053 /* helper function to actually process a single delayed ref entry */
2054 static int run_one_delayed_ref(struct btrfs_trans_handle *trans,
2055 struct btrfs_root *root,
2056 struct btrfs_delayed_ref_node *node,
2057 struct btrfs_delayed_extent_op *extent_op,
2058 int insert_reserved)
2059 {
2060 int ret;
2061 if (btrfs_delayed_ref_is_head(node)) {
2062 struct btrfs_delayed_ref_head *head;
2063 /*
2064 * we've hit the end of the chain and we were supposed
2065 * to insert this extent into the tree. But, it got
2066 * deleted before we ever needed to insert it, so all
2067 * we have to do is clean up the accounting
2068 */
2069 BUG_ON(extent_op);
2070 head = btrfs_delayed_node_to_head(node);
2071 if (insert_reserved) {
2072 btrfs_pin_extent(root, node->bytenr,
2073 node->num_bytes, 1);
2074 if (head->is_data) {
2075 ret = btrfs_del_csums(trans, root,
2076 node->bytenr,
2077 node->num_bytes);
2078 BUG_ON(ret);
2079 }
2080 }
2081 mutex_unlock(&head->mutex);
2082 return 0;
2083 }
2084
2085 if (node->type == BTRFS_TREE_BLOCK_REF_KEY ||
2086 node->type == BTRFS_SHARED_BLOCK_REF_KEY)
2087 ret = run_delayed_tree_ref(trans, root, node, extent_op,
2088 insert_reserved);
2089 else if (node->type == BTRFS_EXTENT_DATA_REF_KEY ||
2090 node->type == BTRFS_SHARED_DATA_REF_KEY)
2091 ret = run_delayed_data_ref(trans, root, node, extent_op,
2092 insert_reserved);
2093 else
2094 BUG();
2095 return ret;
2096 }
2097
2098 static noinline struct btrfs_delayed_ref_node *
2099 select_delayed_ref(struct btrfs_delayed_ref_head *head)
2100 {
2101 struct rb_node *node;
2102 struct btrfs_delayed_ref_node *ref;
2103 int action = BTRFS_ADD_DELAYED_REF;
2104 again:
2105 /*
2106 * select delayed ref of type BTRFS_ADD_DELAYED_REF first.
2107 * this prevents ref count from going down to zero when
2108 * there still are pending delayed ref.
2109 */
2110 node = rb_prev(&head->node.rb_node);
2111 while (1) {
2112 if (!node)
2113 break;
2114 ref = rb_entry(node, struct btrfs_delayed_ref_node,
2115 rb_node);
2116 if (ref->bytenr != head->node.bytenr)
2117 break;
2118 if (ref->action == action)
2119 return ref;
2120 node = rb_prev(node);
2121 }
2122 if (action == BTRFS_ADD_DELAYED_REF) {
2123 action = BTRFS_DROP_DELAYED_REF;
2124 goto again;
2125 }
2126 return NULL;
2127 }
2128
2129 static noinline int run_clustered_refs(struct btrfs_trans_handle *trans,
2130 struct btrfs_root *root,
2131 struct list_head *cluster)
2132 {
2133 struct btrfs_delayed_ref_root *delayed_refs;
2134 struct btrfs_delayed_ref_node *ref;
2135 struct btrfs_delayed_ref_head *locked_ref = NULL;
2136 struct btrfs_delayed_extent_op *extent_op;
2137 int ret;
2138 int count = 0;
2139 int must_insert_reserved = 0;
2140
2141 delayed_refs = &trans->transaction->delayed_refs;
2142 while (1) {
2143 if (!locked_ref) {
2144 /* pick a new head ref from the cluster list */
2145 if (list_empty(cluster))
2146 break;
2147
2148 locked_ref = list_entry(cluster->next,
2149 struct btrfs_delayed_ref_head, cluster);
2150
2151 /* grab the lock that says we are going to process
2152 * all the refs for this head */
2153 ret = btrfs_delayed_ref_lock(trans, locked_ref);
2154
2155 /*
2156 * we may have dropped the spin lock to get the head
2157 * mutex lock, and that might have given someone else
2158 * time to free the head. If that's true, it has been
2159 * removed from our list and we can move on.
2160 */
2161 if (ret == -EAGAIN) {
2162 locked_ref = NULL;
2163 count++;
2164 continue;
2165 }
2166 }
2167
2168 /*
2169 * record the must insert reserved flag before we
2170 * drop the spin lock.
2171 */
2172 must_insert_reserved = locked_ref->must_insert_reserved;
2173 locked_ref->must_insert_reserved = 0;
2174
2175 extent_op = locked_ref->extent_op;
2176 locked_ref->extent_op = NULL;
2177
2178 /*
2179 * locked_ref is the head node, so we have to go one
2180 * node back for any delayed ref updates
2181 */
2182 ref = select_delayed_ref(locked_ref);
2183 if (!ref) {
2184 /* All delayed refs have been processed, Go ahead
2185 * and send the head node to run_one_delayed_ref,
2186 * so that any accounting fixes can happen
2187 */
2188 ref = &locked_ref->node;
2189
2190 if (extent_op && must_insert_reserved) {
2191 kfree(extent_op);
2192 extent_op = NULL;
2193 }
2194
2195 if (extent_op) {
2196 spin_unlock(&delayed_refs->lock);
2197
2198 ret = run_delayed_extent_op(trans, root,
2199 ref, extent_op);
2200 BUG_ON(ret);
2201 kfree(extent_op);
2202
2203 cond_resched();
2204 spin_lock(&delayed_refs->lock);
2205 continue;
2206 }
2207
2208 list_del_init(&locked_ref->cluster);
2209 locked_ref = NULL;
2210 }
2211
2212 ref->in_tree = 0;
2213 rb_erase(&ref->rb_node, &delayed_refs->root);
2214 delayed_refs->num_entries--;
2215
2216 spin_unlock(&delayed_refs->lock);
2217
2218 ret = run_one_delayed_ref(trans, root, ref, extent_op,
2219 must_insert_reserved);
2220 BUG_ON(ret);
2221
2222 btrfs_put_delayed_ref(ref);
2223 kfree(extent_op);
2224 count++;
2225
2226 cond_resched();
2227 spin_lock(&delayed_refs->lock);
2228 }
2229 return count;
2230 }
2231
2232 /*
2233 * this starts processing the delayed reference count updates and
2234 * extent insertions we have queued up so far. count can be
2235 * 0, which means to process everything in the tree at the start
2236 * of the run (but not newly added entries), or it can be some target
2237 * number you'd like to process.
2238 */
2239 int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
2240 struct btrfs_root *root, unsigned long count)
2241 {
2242 struct rb_node *node;
2243 struct btrfs_delayed_ref_root *delayed_refs;
2244 struct btrfs_delayed_ref_node *ref;
2245 struct list_head cluster;
2246 int ret;
2247 int run_all = count == (unsigned long)-1;
2248 int run_most = 0;
2249
2250 if (root == root->fs_info->extent_root)
2251 root = root->fs_info->tree_root;
2252
2253 delayed_refs = &trans->transaction->delayed_refs;
2254 INIT_LIST_HEAD(&cluster);
2255 again:
2256 spin_lock(&delayed_refs->lock);
2257 if (count == 0) {
2258 count = delayed_refs->num_entries * 2;
2259 run_most = 1;
2260 }
2261 while (1) {
2262 if (!(run_all || run_most) &&
2263 delayed_refs->num_heads_ready < 64)
2264 break;
2265
2266 /*
2267 * go find something we can process in the rbtree. We start at
2268 * the beginning of the tree, and then build a cluster
2269 * of refs to process starting at the first one we are able to
2270 * lock
2271 */
2272 ret = btrfs_find_ref_cluster(trans, &cluster,
2273 delayed_refs->run_delayed_start);
2274 if (ret)
2275 break;
2276
2277 ret = run_clustered_refs(trans, root, &cluster);
2278 BUG_ON(ret < 0);
2279
2280 count -= min_t(unsigned long, ret, count);
2281
2282 if (count == 0)
2283 break;
2284 }
2285
2286 if (run_all) {
2287 node = rb_first(&delayed_refs->root);
2288 if (!node)
2289 goto out;
2290 count = (unsigned long)-1;
2291
2292 while (node) {
2293 ref = rb_entry(node, struct btrfs_delayed_ref_node,
2294 rb_node);
2295 if (btrfs_delayed_ref_is_head(ref)) {
2296 struct btrfs_delayed_ref_head *head;
2297
2298 head = btrfs_delayed_node_to_head(ref);
2299 atomic_inc(&ref->refs);
2300
2301 spin_unlock(&delayed_refs->lock);
2302 /*
2303 * Mutex was contended, block until it's
2304 * released and try again
2305 */
2306 mutex_lock(&head->mutex);
2307 mutex_unlock(&head->mutex);
2308
2309 btrfs_put_delayed_ref(ref);
2310 cond_resched();
2311 goto again;
2312 }
2313 node = rb_next(node);
2314 }
2315 spin_unlock(&delayed_refs->lock);
2316 schedule_timeout(1);
2317 goto again;
2318 }
2319 out:
2320 spin_unlock(&delayed_refs->lock);
2321 return 0;
2322 }
2323
2324 int btrfs_set_disk_extent_flags(struct btrfs_trans_handle *trans,
2325 struct btrfs_root *root,
2326 u64 bytenr, u64 num_bytes, u64 flags,
2327 int is_data)
2328 {
2329 struct btrfs_delayed_extent_op *extent_op;
2330 int ret;
2331
2332 extent_op = kmalloc(sizeof(*extent_op), GFP_NOFS);
2333 if (!extent_op)
2334 return -ENOMEM;
2335
2336 extent_op->flags_to_set = flags;
2337 extent_op->update_flags = 1;
2338 extent_op->update_key = 0;
2339 extent_op->is_data = is_data ? 1 : 0;
2340
2341 ret = btrfs_add_delayed_extent_op(trans, bytenr, num_bytes, extent_op);
2342 if (ret)
2343 kfree(extent_op);
2344 return ret;
2345 }
2346
2347 static noinline int check_delayed_ref(struct btrfs_trans_handle *trans,
2348 struct btrfs_root *root,
2349 struct btrfs_path *path,
2350 u64 objectid, u64 offset, u64 bytenr)
2351 {
2352 struct btrfs_delayed_ref_head *head;
2353 struct btrfs_delayed_ref_node *ref;
2354 struct btrfs_delayed_data_ref *data_ref;
2355 struct btrfs_delayed_ref_root *delayed_refs;
2356 struct rb_node *node;
2357 int ret = 0;
2358
2359 ret = -ENOENT;
2360 delayed_refs = &trans->transaction->delayed_refs;
2361 spin_lock(&delayed_refs->lock);
2362 head = btrfs_find_delayed_ref_head(trans, bytenr);
2363 if (!head)
2364 goto out;
2365
2366 if (!mutex_trylock(&head->mutex)) {
2367 atomic_inc(&head->node.refs);
2368 spin_unlock(&delayed_refs->lock);
2369
2370 btrfs_release_path(path);
2371
2372 /*
2373 * Mutex was contended, block until it's released and let
2374 * caller try again
2375 */
2376 mutex_lock(&head->mutex);
2377 mutex_unlock(&head->mutex);
2378 btrfs_put_delayed_ref(&head->node);
2379 return -EAGAIN;
2380 }
2381
2382 node = rb_prev(&head->node.rb_node);
2383 if (!node)
2384 goto out_unlock;
2385
2386 ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
2387
2388 if (ref->bytenr != bytenr)
2389 goto out_unlock;
2390
2391 ret = 1;
2392 if (ref->type != BTRFS_EXTENT_DATA_REF_KEY)
2393 goto out_unlock;
2394
2395 data_ref = btrfs_delayed_node_to_data_ref(ref);
2396
2397 node = rb_prev(node);
2398 if (node) {
2399 ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
2400 if (ref->bytenr == bytenr)
2401 goto out_unlock;
2402 }
2403
2404 if (data_ref->root != root->root_key.objectid ||
2405 data_ref->objectid != objectid || data_ref->offset != offset)
2406 goto out_unlock;
2407
2408 ret = 0;
2409 out_unlock:
2410 mutex_unlock(&head->mutex);
2411 out:
2412 spin_unlock(&delayed_refs->lock);
2413 return ret;
2414 }
2415
2416 static noinline int check_committed_ref(struct btrfs_trans_handle *trans,
2417 struct btrfs_root *root,
2418 struct btrfs_path *path,
2419 u64 objectid, u64 offset, u64 bytenr)
2420 {
2421 struct btrfs_root *extent_root = root->fs_info->extent_root;
2422 struct extent_buffer *leaf;
2423 struct btrfs_extent_data_ref *ref;
2424 struct btrfs_extent_inline_ref *iref;
2425 struct btrfs_extent_item *ei;
2426 struct btrfs_key key;
2427 u32 item_size;
2428 int ret;
2429
2430 key.objectid = bytenr;
2431 key.offset = (u64)-1;
2432 key.type = BTRFS_EXTENT_ITEM_KEY;
2433
2434 ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
2435 if (ret < 0)
2436 goto out;
2437 BUG_ON(ret == 0);
2438
2439 ret = -ENOENT;
2440 if (path->slots[0] == 0)
2441 goto out;
2442
2443 path->slots[0]--;
2444 leaf = path->nodes[0];
2445 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
2446
2447 if (key.objectid != bytenr || key.type != BTRFS_EXTENT_ITEM_KEY)
2448 goto out;
2449
2450 ret = 1;
2451 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
2452 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
2453 if (item_size < sizeof(*ei)) {
2454 WARN_ON(item_size != sizeof(struct btrfs_extent_item_v0));
2455 goto out;
2456 }
2457 #endif
2458 ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
2459
2460 if (item_size != sizeof(*ei) +
2461 btrfs_extent_inline_ref_size(BTRFS_EXTENT_DATA_REF_KEY))
2462 goto out;
2463
2464 if (btrfs_extent_generation(leaf, ei) <=
2465 btrfs_root_last_snapshot(&root->root_item))
2466 goto out;
2467
2468 iref = (struct btrfs_extent_inline_ref *)(ei + 1);
2469 if (btrfs_extent_inline_ref_type(leaf, iref) !=
2470 BTRFS_EXTENT_DATA_REF_KEY)
2471 goto out;
2472
2473 ref = (struct btrfs_extent_data_ref *)(&iref->offset);
2474 if (btrfs_extent_refs(leaf, ei) !=
2475 btrfs_extent_data_ref_count(leaf, ref) ||
2476 btrfs_extent_data_ref_root(leaf, ref) !=
2477 root->root_key.objectid ||
2478 btrfs_extent_data_ref_objectid(leaf, ref) != objectid ||
2479 btrfs_extent_data_ref_offset(leaf, ref) != offset)
2480 goto out;
2481
2482 ret = 0;
2483 out:
2484 return ret;
2485 }
2486
2487 int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans,
2488 struct btrfs_root *root,
2489 u64 objectid, u64 offset, u64 bytenr)
2490 {
2491 struct btrfs_path *path;
2492 int ret;
2493 int ret2;
2494
2495 path = btrfs_alloc_path();
2496 if (!path)
2497 return -ENOENT;
2498
2499 do {
2500 ret = check_committed_ref(trans, root, path, objectid,
2501 offset, bytenr);
2502 if (ret && ret != -ENOENT)
2503 goto out;
2504
2505 ret2 = check_delayed_ref(trans, root, path, objectid,
2506 offset, bytenr);
2507 } while (ret2 == -EAGAIN);
2508
2509 if (ret2 && ret2 != -ENOENT) {
2510 ret = ret2;
2511 goto out;
2512 }
2513
2514 if (ret != -ENOENT || ret2 != -ENOENT)
2515 ret = 0;
2516 out:
2517 btrfs_free_path(path);
2518 if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID)
2519 WARN_ON(ret > 0);
2520 return ret;
2521 }
2522
2523 static int __btrfs_mod_ref(struct btrfs_trans_handle *trans,
2524 struct btrfs_root *root,
2525 struct extent_buffer *buf,
2526 int full_backref, int inc)
2527 {
2528 u64 bytenr;
2529 u64 num_bytes;
2530 u64 parent;
2531 u64 ref_root;
2532 u32 nritems;
2533 struct btrfs_key key;
2534 struct btrfs_file_extent_item *fi;
2535 int i;
2536 int level;
2537 int ret = 0;
2538 int (*process_func)(struct btrfs_trans_handle *, struct btrfs_root *,
2539 u64, u64, u64, u64, u64, u64);
2540
2541 ref_root = btrfs_header_owner(buf);
2542 nritems = btrfs_header_nritems(buf);
2543 level = btrfs_header_level(buf);
2544
2545 if (!root->ref_cows && level == 0)
2546 return 0;
2547
2548 if (inc)
2549 process_func = btrfs_inc_extent_ref;
2550 else
2551 process_func = btrfs_free_extent;
2552
2553 if (full_backref)
2554 parent = buf->start;
2555 else
2556 parent = 0;
2557
2558 for (i = 0; i < nritems; i++) {
2559 if (level == 0) {
2560 btrfs_item_key_to_cpu(buf, &key, i);
2561 if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
2562 continue;
2563 fi = btrfs_item_ptr(buf, i,
2564 struct btrfs_file_extent_item);
2565 if (btrfs_file_extent_type(buf, fi) ==
2566 BTRFS_FILE_EXTENT_INLINE)
2567 continue;
2568 bytenr = btrfs_file_extent_disk_bytenr(buf, fi);
2569 if (bytenr == 0)
2570 continue;
2571
2572 num_bytes = btrfs_file_extent_disk_num_bytes(buf, fi);
2573 key.offset -= btrfs_file_extent_offset(buf, fi);
2574 ret = process_func(trans, root, bytenr, num_bytes,
2575 parent, ref_root, key.objectid,
2576 key.offset);
2577 if (ret)
2578 goto fail;
2579 } else {
2580 bytenr = btrfs_node_blockptr(buf, i);
2581 num_bytes = btrfs_level_size(root, level - 1);
2582 ret = process_func(trans, root, bytenr, num_bytes,
2583 parent, ref_root, level - 1, 0);
2584 if (ret)
2585 goto fail;
2586 }
2587 }
2588 return 0;
2589 fail:
2590 BUG();
2591 return ret;
2592 }
2593
2594 int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
2595 struct extent_buffer *buf, int full_backref)
2596 {
2597 return __btrfs_mod_ref(trans, root, buf, full_backref, 1);
2598 }
2599
2600 int btrfs_dec_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
2601 struct extent_buffer *buf, int full_backref)
2602 {
2603 return __btrfs_mod_ref(trans, root, buf, full_backref, 0);
2604 }
2605
2606 static int write_one_cache_group(struct btrfs_trans_handle *trans,
2607 struct btrfs_root *root,
2608 struct btrfs_path *path,
2609 struct btrfs_block_group_cache *cache)
2610 {
2611 int ret;
2612 struct btrfs_root *extent_root = root->fs_info->extent_root;
2613 unsigned long bi;
2614 struct extent_buffer *leaf;
2615
2616 ret = btrfs_search_slot(trans, extent_root, &cache->key, path, 0, 1);
2617 if (ret < 0)
2618 goto fail;
2619 BUG_ON(ret);
2620
2621 leaf = path->nodes[0];
2622 bi = btrfs_item_ptr_offset(leaf, path->slots[0]);
2623 write_extent_buffer(leaf, &cache->item, bi, sizeof(cache->item));
2624 btrfs_mark_buffer_dirty(leaf);
2625 btrfs_release_path(path);
2626 fail:
2627 if (ret)
2628 return ret;
2629 return 0;
2630
2631 }
2632
2633 static struct btrfs_block_group_cache *
2634 next_block_group(struct btrfs_root *root,
2635 struct btrfs_block_group_cache *cache)
2636 {
2637 struct rb_node *node;
2638 spin_lock(&root->fs_info->block_group_cache_lock);
2639 node = rb_next(&cache->cache_node);
2640 btrfs_put_block_group(cache);
2641 if (node) {
2642 cache = rb_entry(node, struct btrfs_block_group_cache,
2643 cache_node);
2644 btrfs_get_block_group(cache);
2645 } else
2646 cache = NULL;
2647 spin_unlock(&root->fs_info->block_group_cache_lock);
2648 return cache;
2649 }
2650
2651 static int cache_save_setup(struct btrfs_block_group_cache *block_group,
2652 struct btrfs_trans_handle *trans,
2653 struct btrfs_path *path)
2654 {
2655 struct btrfs_root *root = block_group->fs_info->tree_root;
2656 struct inode *inode = NULL;
2657 u64 alloc_hint = 0;
2658 int dcs = BTRFS_DC_ERROR;
2659 int num_pages = 0;
2660 int retries = 0;
2661 int ret = 0;
2662
2663 /*
2664 * If this block group is smaller than 100 megs don't bother caching the
2665 * block group.
2666 */
2667 if (block_group->key.offset < (100 * 1024 * 1024)) {
2668 spin_lock(&block_group->lock);
2669 block_group->disk_cache_state = BTRFS_DC_WRITTEN;
2670 spin_unlock(&block_group->lock);
2671 return 0;
2672 }
2673
2674 again:
2675 inode = lookup_free_space_inode(root, block_group, path);
2676 if (IS_ERR(inode) && PTR_ERR(inode) != -ENOENT) {
2677 ret = PTR_ERR(inode);
2678 btrfs_release_path(path);
2679 goto out;
2680 }
2681
2682 if (IS_ERR(inode)) {
2683 BUG_ON(retries);
2684 retries++;
2685
2686 if (block_group->ro)
2687 goto out_free;
2688
2689 ret = create_free_space_inode(root, trans, block_group, path);
2690 if (ret)
2691 goto out_free;
2692 goto again;
2693 }
2694
2695 /*
2696 * We want to set the generation to 0, that way if anything goes wrong
2697 * from here on out we know not to trust this cache when we load up next
2698 * time.
2699 */
2700 BTRFS_I(inode)->generation = 0;
2701 ret = btrfs_update_inode(trans, root, inode);
2702 WARN_ON(ret);
2703
2704 if (i_size_read(inode) > 0) {
2705 ret = btrfs_truncate_free_space_cache(root, trans, path,
2706 inode);
2707 if (ret)
2708 goto out_put;
2709 }
2710
2711 spin_lock(&block_group->lock);
2712 if (block_group->cached != BTRFS_CACHE_FINISHED) {
2713 /* We're not cached, don't bother trying to write stuff out */
2714 dcs = BTRFS_DC_WRITTEN;
2715 spin_unlock(&block_group->lock);
2716 goto out_put;
2717 }
2718 spin_unlock(&block_group->lock);
2719
2720 num_pages = (int)div64_u64(block_group->key.offset, 1024 * 1024 * 1024);
2721 if (!num_pages)
2722 num_pages = 1;
2723
2724 /*
2725 * Just to make absolutely sure we have enough space, we're going to
2726 * preallocate 12 pages worth of space for each block group. In
2727 * practice we ought to use at most 8, but we need extra space so we can
2728 * add our header and have a terminator between the extents and the
2729 * bitmaps.
2730 */
2731 num_pages *= 16;
2732 num_pages *= PAGE_CACHE_SIZE;
2733
2734 ret = btrfs_check_data_free_space(inode, num_pages);
2735 if (ret)
2736 goto out_put;
2737
2738 ret = btrfs_prealloc_file_range_trans(inode, trans, 0, 0, num_pages,
2739 num_pages, num_pages,
2740 &alloc_hint);
2741 if (!ret)
2742 dcs = BTRFS_DC_SETUP;
2743 btrfs_free_reserved_data_space(inode, num_pages);
2744 out_put:
2745 iput(inode);
2746 out_free:
2747 btrfs_release_path(path);
2748 out:
2749 spin_lock(&block_group->lock);
2750 block_group->disk_cache_state = dcs;
2751 spin_unlock(&block_group->lock);
2752
2753 return ret;
2754 }
2755
2756 int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans,
2757 struct btrfs_root *root)
2758 {
2759 struct btrfs_block_group_cache *cache;
2760 int err = 0;
2761 struct btrfs_path *path;
2762 u64 last = 0;
2763
2764 path = btrfs_alloc_path();
2765 if (!path)
2766 return -ENOMEM;
2767
2768 again:
2769 while (1) {
2770 cache = btrfs_lookup_first_block_group(root->fs_info, last);
2771 while (cache) {
2772 if (cache->disk_cache_state == BTRFS_DC_CLEAR)
2773 break;
2774 cache = next_block_group(root, cache);
2775 }
2776 if (!cache) {
2777 if (last == 0)
2778 break;
2779 last = 0;
2780 continue;
2781 }
2782 err = cache_save_setup(cache, trans, path);
2783 last = cache->key.objectid + cache->key.offset;
2784 btrfs_put_block_group(cache);
2785 }
2786
2787 while (1) {
2788 if (last == 0) {
2789 err = btrfs_run_delayed_refs(trans, root,
2790 (unsigned long)-1);
2791 BUG_ON(err);
2792 }
2793
2794 cache = btrfs_lookup_first_block_group(root->fs_info, last);
2795 while (cache) {
2796 if (cache->disk_cache_state == BTRFS_DC_CLEAR) {
2797 btrfs_put_block_group(cache);
2798 goto again;
2799 }
2800
2801 if (cache->dirty)
2802 break;
2803 cache = next_block_group(root, cache);
2804 }
2805 if (!cache) {
2806 if (last == 0)
2807 break;
2808 last = 0;
2809 continue;
2810 }
2811
2812 if (cache->disk_cache_state == BTRFS_DC_SETUP)
2813 cache->disk_cache_state = BTRFS_DC_NEED_WRITE;
2814 cache->dirty = 0;
2815 last = cache->key.objectid + cache->key.offset;
2816
2817 err = write_one_cache_group(trans, root, path, cache);
2818 BUG_ON(err);
2819 btrfs_put_block_group(cache);
2820 }
2821
2822 while (1) {
2823 /*
2824 * I don't think this is needed since we're just marking our
2825 * preallocated extent as written, but just in case it can't
2826 * hurt.
2827 */
2828 if (last == 0) {
2829 err = btrfs_run_delayed_refs(trans, root,
2830 (unsigned long)-1);
2831 BUG_ON(err);
2832 }
2833
2834 cache = btrfs_lookup_first_block_group(root->fs_info, last);
2835 while (cache) {
2836 /*
2837 * Really this shouldn't happen, but it could if we
2838 * couldn't write the entire preallocated extent and
2839 * splitting the extent resulted in a new block.
2840 */
2841 if (cache->dirty) {
2842 btrfs_put_block_group(cache);
2843 goto again;
2844 }
2845 if (cache->disk_cache_state == BTRFS_DC_NEED_WRITE)
2846 break;
2847 cache = next_block_group(root, cache);
2848 }
2849 if (!cache) {
2850 if (last == 0)
2851 break;
2852 last = 0;
2853 continue;
2854 }
2855
2856 btrfs_write_out_cache(root, trans, cache, path);
2857
2858 /*
2859 * If we didn't have an error then the cache state is still
2860 * NEED_WRITE, so we can set it to WRITTEN.
2861 */
2862 if (cache->disk_cache_state == BTRFS_DC_NEED_WRITE)
2863 cache->disk_cache_state = BTRFS_DC_WRITTEN;
2864 last = cache->key.objectid + cache->key.offset;
2865 btrfs_put_block_group(cache);
2866 }
2867
2868 btrfs_free_path(path);
2869 return 0;
2870 }
2871
2872 int btrfs_extent_readonly(struct btrfs_root *root, u64 bytenr)
2873 {
2874 struct btrfs_block_group_cache *block_group;
2875 int readonly = 0;
2876
2877 block_group = btrfs_lookup_block_group(root->fs_info, bytenr);
2878 if (!block_group || block_group->ro)
2879 readonly = 1;
2880 if (block_group)
2881 btrfs_put_block_group(block_group);
2882 return readonly;
2883 }
2884
2885 static int update_space_info(struct btrfs_fs_info *info, u64 flags,
2886 u64 total_bytes, u64 bytes_used,
2887 struct btrfs_space_info **space_info)
2888 {
2889 struct btrfs_space_info *found;
2890 int i;
2891 int factor;
2892
2893 if (flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 |
2894 BTRFS_BLOCK_GROUP_RAID10))
2895 factor = 2;
2896 else
2897 factor = 1;
2898
2899 found = __find_space_info(info, flags);
2900 if (found) {
2901 spin_lock(&found->lock);
2902 found->total_bytes += total_bytes;
2903 found->disk_total += total_bytes * factor;
2904 found->bytes_used += bytes_used;
2905 found->disk_used += bytes_used * factor;
2906 found->full = 0;
2907 spin_unlock(&found->lock);
2908 *space_info = found;
2909 return 0;
2910 }
2911 found = kzalloc(sizeof(*found), GFP_NOFS);
2912 if (!found)
2913 return -ENOMEM;
2914
2915 for (i = 0; i < BTRFS_NR_RAID_TYPES; i++)
2916 INIT_LIST_HEAD(&found->block_groups[i]);
2917 init_rwsem(&found->groups_sem);
2918 spin_lock_init(&found->lock);
2919 found->flags = flags & (BTRFS_BLOCK_GROUP_DATA |
2920 BTRFS_BLOCK_GROUP_SYSTEM |
2921 BTRFS_BLOCK_GROUP_METADATA);
2922 found->total_bytes = total_bytes;
2923 found->disk_total = total_bytes * factor;
2924 found->bytes_used = bytes_used;
2925 found->disk_used = bytes_used * factor;
2926 found->bytes_pinned = 0;
2927 found->bytes_reserved = 0;
2928 found->bytes_readonly = 0;
2929 found->bytes_may_use = 0;
2930 found->full = 0;
2931 found->force_alloc = CHUNK_ALLOC_NO_FORCE;
2932 found->chunk_alloc = 0;
2933 found->flush = 0;
2934 init_waitqueue_head(&found->wait);
2935 *space_info = found;
2936 list_add_rcu(&found->list, &info->space_info);
2937 return 0;
2938 }
2939
2940 static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
2941 {
2942 u64 extra_flags = flags & (BTRFS_BLOCK_GROUP_RAID0 |
2943 BTRFS_BLOCK_GROUP_RAID1 |
2944 BTRFS_BLOCK_GROUP_RAID10 |
2945 BTRFS_BLOCK_GROUP_DUP);
2946 if (extra_flags) {
2947 if (flags & BTRFS_BLOCK_GROUP_DATA)
2948 fs_info->avail_data_alloc_bits |= extra_flags;
2949 if (flags & BTRFS_BLOCK_GROUP_METADATA)
2950 fs_info->avail_metadata_alloc_bits |= extra_flags;
2951 if (flags & BTRFS_BLOCK_GROUP_SYSTEM)
2952 fs_info->avail_system_alloc_bits |= extra_flags;
2953 }
2954 }
2955
2956 u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags)
2957 {
2958 /*
2959 * we add in the count of missing devices because we want
2960 * to make sure that any RAID levels on a degraded FS
2961 * continue to be honored.
2962 */
2963 u64 num_devices = root->fs_info->fs_devices->rw_devices +
2964 root->fs_info->fs_devices->missing_devices;
2965
2966 if (num_devices == 1)
2967 flags &= ~(BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID0);
2968 if (num_devices < 4)
2969 flags &= ~BTRFS_BLOCK_GROUP_RAID10;
2970
2971 if ((flags & BTRFS_BLOCK_GROUP_DUP) &&
2972 (flags & (BTRFS_BLOCK_GROUP_RAID1 |
2973 BTRFS_BLOCK_GROUP_RAID10))) {
2974 flags &= ~BTRFS_BLOCK_GROUP_DUP;
2975 }
2976
2977 if ((flags & BTRFS_BLOCK_GROUP_RAID1) &&
2978 (flags & BTRFS_BLOCK_GROUP_RAID10)) {
2979 flags &= ~BTRFS_BLOCK_GROUP_RAID1;
2980 }
2981
2982 if ((flags & BTRFS_BLOCK_GROUP_RAID0) &&
2983 ((flags & BTRFS_BLOCK_GROUP_RAID1) |
2984 (flags & BTRFS_BLOCK_GROUP_RAID10) |
2985 (flags & BTRFS_BLOCK_GROUP_DUP)))
2986 flags &= ~BTRFS_BLOCK_GROUP_RAID0;
2987 return flags;
2988 }
2989
2990 static u64 get_alloc_profile(struct btrfs_root *root, u64 flags)
2991 {
2992 if (flags & BTRFS_BLOCK_GROUP_DATA)
2993 flags |= root->fs_info->avail_data_alloc_bits &
2994 root->fs_info->data_alloc_profile;
2995 else if (flags & BTRFS_BLOCK_GROUP_SYSTEM)
2996 flags |= root->fs_info->avail_system_alloc_bits &
2997 root->fs_info->system_alloc_profile;
2998 else if (flags & BTRFS_BLOCK_GROUP_METADATA)
2999 flags |= root->fs_info->avail_metadata_alloc_bits &
3000 root->fs_info->metadata_alloc_profile;
3001 return btrfs_reduce_alloc_profile(root, flags);
3002 }
3003
3004 u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data)
3005 {
3006 u64 flags;
3007
3008 if (data)
3009 flags = BTRFS_BLOCK_GROUP_DATA;
3010 else if (root == root->fs_info->chunk_root)
3011 flags = BTRFS_BLOCK_GROUP_SYSTEM;
3012 else
3013 flags = BTRFS_BLOCK_GROUP_METADATA;
3014
3015 return get_alloc_profile(root, flags);
3016 }
3017
3018 void btrfs_set_inode_space_info(struct btrfs_root *root, struct inode *inode)
3019 {
3020 BTRFS_I(inode)->space_info = __find_space_info(root->fs_info,
3021 BTRFS_BLOCK_GROUP_DATA);
3022 }
3023
3024 /*
3025 * This will check the space that the inode allocates from to make sure we have
3026 * enough space for bytes.
3027 */
3028 int btrfs_check_data_free_space(struct inode *inode, u64 bytes)
3029 {
3030 struct btrfs_space_info *data_sinfo;
3031 struct btrfs_root *root = BTRFS_I(inode)->root;
3032 u64 used;
3033 int ret = 0, committed = 0, alloc_chunk = 1;
3034
3035 /* make sure bytes are sectorsize aligned */
3036 bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1);
3037
3038 if (root == root->fs_info->tree_root ||
3039 BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID) {
3040 alloc_chunk = 0;
3041 committed = 1;
3042 }
3043
3044 data_sinfo = BTRFS_I(inode)->space_info;
3045 if (!data_sinfo)
3046 goto alloc;
3047
3048 again:
3049 /* make sure we have enough space to handle the data first */
3050 spin_lock(&data_sinfo->lock);
3051 used = data_sinfo->bytes_used + data_sinfo->bytes_reserved +
3052 data_sinfo->bytes_pinned + data_sinfo->bytes_readonly +
3053 data_sinfo->bytes_may_use;
3054
3055 if (used + bytes > data_sinfo->total_bytes) {
3056 struct btrfs_trans_handle *trans;
3057
3058 /*
3059 * if we don't have enough free bytes in this space then we need
3060 * to alloc a new chunk.
3061 */
3062 if (!data_sinfo->full && alloc_chunk) {
3063 u64 alloc_target;
3064
3065 data_sinfo->force_alloc = CHUNK_ALLOC_FORCE;
3066 spin_unlock(&data_sinfo->lock);
3067 alloc:
3068 alloc_target = btrfs_get_alloc_profile(root, 1);
3069 trans = btrfs_join_transaction(root);
3070 if (IS_ERR(trans))
3071 return PTR_ERR(trans);
3072
3073 ret = do_chunk_alloc(trans, root->fs_info->extent_root,
3074 bytes + 2 * 1024 * 1024,
3075 alloc_target,
3076 CHUNK_ALLOC_NO_FORCE);
3077 btrfs_end_transaction(trans, root);
3078 if (ret < 0) {
3079 if (ret != -ENOSPC)
3080 return ret;
3081 else
3082 goto commit_trans;
3083 }
3084
3085 if (!data_sinfo) {
3086 btrfs_set_inode_space_info(root, inode);
3087 data_sinfo = BTRFS_I(inode)->space_info;
3088 }
3089 goto again;
3090 }
3091
3092 /*
3093 * If we have less pinned bytes than we want to allocate then
3094 * don't bother committing the transaction, it won't help us.
3095 */
3096 if (data_sinfo->bytes_pinned < bytes)
3097 committed = 1;
3098 spin_unlock(&data_sinfo->lock);
3099
3100 /* commit the current transaction and try again */
3101 commit_trans:
3102 if (!committed &&
3103 !atomic_read(&root->fs_info->open_ioctl_trans)) {
3104 committed = 1;
3105 trans = btrfs_join_transaction(root);
3106 if (IS_ERR(trans))
3107 return PTR_ERR(trans);
3108 ret = btrfs_commit_transaction(trans, root);
3109 if (ret)
3110 return ret;
3111 goto again;
3112 }
3113
3114 return -ENOSPC;
3115 }
3116 data_sinfo->bytes_may_use += bytes;
3117 BTRFS_I(inode)->reserved_bytes += bytes;
3118 spin_unlock(&data_sinfo->lock);
3119
3120 return 0;
3121 }
3122
3123 /*
3124 * called when we are clearing an delalloc extent from the
3125 * inode's io_tree or there was an error for whatever reason
3126 * after calling btrfs_check_data_free_space
3127 */
3128 void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes)
3129 {
3130 struct btrfs_root *root = BTRFS_I(inode)->root;
3131 struct btrfs_space_info *data_sinfo;
3132
3133 /* make sure bytes are sectorsize aligned */
3134 bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1);
3135
3136 data_sinfo = BTRFS_I(inode)->space_info;
3137 spin_lock(&data_sinfo->lock);
3138 data_sinfo->bytes_may_use -= bytes;
3139 BTRFS_I(inode)->reserved_bytes -= bytes;
3140 spin_unlock(&data_sinfo->lock);
3141 }
3142
3143 static void force_metadata_allocation(struct btrfs_fs_info *info)
3144 {
3145 struct list_head *head = &info->space_info;
3146 struct btrfs_space_info *found;
3147
3148 rcu_read_lock();
3149 list_for_each_entry_rcu(found, head, list) {
3150 if (found->flags & BTRFS_BLOCK_GROUP_METADATA)
3151 found->force_alloc = CHUNK_ALLOC_FORCE;
3152 }
3153 rcu_read_unlock();
3154 }
3155
3156 static int should_alloc_chunk(struct btrfs_root *root,
3157 struct btrfs_space_info *sinfo, u64 alloc_bytes,
3158 int force)
3159 {
3160 u64 num_bytes = sinfo->total_bytes - sinfo->bytes_readonly;
3161 u64 num_allocated = sinfo->bytes_used + sinfo->bytes_reserved;
3162 u64 thresh;
3163
3164 if (force == CHUNK_ALLOC_FORCE)
3165 return 1;
3166
3167 /*
3168 * in limited mode, we want to have some free space up to
3169 * about 1% of the FS size.
3170 */
3171 if (force == CHUNK_ALLOC_LIMITED) {
3172 thresh = btrfs_super_total_bytes(&root->fs_info->super_copy);
3173 thresh = max_t(u64, 64 * 1024 * 1024,
3174 div_factor_fine(thresh, 1));
3175
3176 if (num_bytes - num_allocated < thresh)
3177 return 1;
3178 }
3179
3180 /*
3181 * we have two similar checks here, one based on percentage
3182 * and once based on a hard number of 256MB. The idea
3183 * is that if we have a good amount of free
3184 * room, don't allocate a chunk. A good mount is
3185 * less than 80% utilized of the chunks we have allocated,
3186 * or more than 256MB free
3187 */
3188 if (num_allocated + alloc_bytes + 256 * 1024 * 1024 < num_bytes)
3189 return 0;
3190
3191 if (num_allocated + alloc_bytes < div_factor(num_bytes, 8))
3192 return 0;
3193
3194 thresh = btrfs_super_total_bytes(&root->fs_info->super_copy);
3195
3196 /* 256MB or 5% of the FS */
3197 thresh = max_t(u64, 256 * 1024 * 1024, div_factor_fine(thresh, 5));
3198
3199 if (num_bytes > thresh && sinfo->bytes_used < div_factor(num_bytes, 3))
3200 return 0;
3201 return 1;
3202 }
3203
3204 static int do_chunk_alloc(struct btrfs_trans_handle *trans,
3205 struct btrfs_root *extent_root, u64 alloc_bytes,
3206 u64 flags, int force)
3207 {
3208 struct btrfs_space_info *space_info;
3209 struct btrfs_fs_info *fs_info = extent_root->fs_info;
3210 int wait_for_alloc = 0;
3211 int ret = 0;
3212
3213 flags = btrfs_reduce_alloc_profile(extent_root, flags);
3214
3215 space_info = __find_space_info(extent_root->fs_info, flags);
3216 if (!space_info) {
3217 ret = update_space_info(extent_root->fs_info, flags,
3218 0, 0, &space_info);
3219 BUG_ON(ret);
3220 }
3221 BUG_ON(!space_info);
3222
3223 again:
3224 spin_lock(&space_info->lock);
3225 if (space_info->force_alloc)
3226 force = space_info->force_alloc;
3227 if (space_info->full) {
3228 spin_unlock(&space_info->lock);
3229 return 0;
3230 }
3231
3232 if (!should_alloc_chunk(extent_root, space_info, alloc_bytes, force)) {
3233 spin_unlock(&space_info->lock);
3234 return 0;
3235 } else if (space_info->chunk_alloc) {
3236 wait_for_alloc = 1;
3237 } else {
3238 space_info->chunk_alloc = 1;
3239 }
3240
3241 spin_unlock(&space_info->lock);
3242
3243 mutex_lock(&fs_info->chunk_mutex);
3244
3245 /*
3246 * The chunk_mutex is held throughout the entirety of a chunk
3247 * allocation, so once we've acquired the chunk_mutex we know that the
3248 * other guy is done and we need to recheck and see if we should
3249 * allocate.
3250 */
3251 if (wait_for_alloc) {
3252 mutex_unlock(&fs_info->chunk_mutex);
3253 wait_for_alloc = 0;
3254 goto again;
3255 }
3256
3257 /*
3258 * If we have mixed data/metadata chunks we want to make sure we keep
3259 * allocating mixed chunks instead of individual chunks.
3260 */
3261 if (btrfs_mixed_space_info(space_info))
3262 flags |= (BTRFS_BLOCK_GROUP_DATA | BTRFS_BLOCK_GROUP_METADATA);
3263
3264 /*
3265 * if we're doing a data chunk, go ahead and make sure that
3266 * we keep a reasonable number of metadata chunks allocated in the
3267 * FS as well.
3268 */
3269 if (flags & BTRFS_BLOCK_GROUP_DATA && fs_info->metadata_ratio) {
3270 fs_info->data_chunk_allocations++;
3271 if (!(fs_info->data_chunk_allocations %
3272 fs_info->metadata_ratio))
3273 force_metadata_allocation(fs_info);
3274 }
3275
3276 ret = btrfs_alloc_chunk(trans, extent_root, flags);
3277 if (ret < 0 && ret != -ENOSPC)
3278 goto out;
3279
3280 spin_lock(&space_info->lock);
3281 if (ret)
3282 space_info->full = 1;
3283 else
3284 ret = 1;
3285
3286 space_info->force_alloc = CHUNK_ALLOC_NO_FORCE;
3287 space_info->chunk_alloc = 0;
3288 spin_unlock(&space_info->lock);
3289 out:
3290 mutex_unlock(&extent_root->fs_info->chunk_mutex);
3291 return ret;
3292 }
3293
3294 /*
3295 * shrink metadata reservation for delalloc
3296 */
3297 static int shrink_delalloc(struct btrfs_trans_handle *trans,
3298 struct btrfs_root *root, u64 to_reclaim, int sync)
3299 {
3300 struct btrfs_block_rsv *block_rsv;
3301 struct btrfs_space_info *space_info;
3302 u64 reserved;
3303 u64 max_reclaim;
3304 u64 reclaimed = 0;
3305 long time_left;
3306 int nr_pages = (2 * 1024 * 1024) >> PAGE_CACHE_SHIFT;
3307 int loops = 0;
3308 unsigned long progress;
3309
3310 block_rsv = &root->fs_info->delalloc_block_rsv;
3311 space_info = block_rsv->space_info;
3312
3313 smp_mb();
3314 reserved = space_info->bytes_reserved;
3315 progress = space_info->reservation_progress;
3316
3317 if (reserved == 0)
3318 return 0;
3319
3320 smp_mb();
3321 if (root->fs_info->delalloc_bytes == 0) {
3322 if (trans)
3323 return 0;
3324 btrfs_wait_ordered_extents(root, 0, 0);
3325 return 0;
3326 }
3327
3328 max_reclaim = min(reserved, to_reclaim);
3329
3330 while (loops < 1024) {
3331 /* have the flusher threads jump in and do some IO */
3332 smp_mb();
3333 nr_pages = min_t(unsigned long, nr_pages,
3334 root->fs_info->delalloc_bytes >> PAGE_CACHE_SHIFT);
3335 writeback_inodes_sb_nr_if_idle(root->fs_info->sb, nr_pages);
3336
3337 spin_lock(&space_info->lock);
3338 if (reserved > space_info->bytes_reserved)
3339 reclaimed += reserved - space_info->bytes_reserved;
3340 reserved = space_info->bytes_reserved;
3341 spin_unlock(&space_info->lock);
3342
3343 loops++;
3344
3345 if (reserved == 0 || reclaimed >= max_reclaim)
3346 break;
3347
3348 if (trans && trans->transaction->blocked)
3349 return -EAGAIN;
3350
3351 time_left = schedule_timeout_interruptible(1);
3352
3353 /* We were interrupted, exit */
3354 if (time_left)
3355 break;
3356
3357 /* we've kicked the IO a few times, if anything has been freed,
3358 * exit. There is no sense in looping here for a long time
3359 * when we really need to commit the transaction, or there are
3360 * just too many writers without enough free space
3361 */
3362
3363 if (loops > 3) {
3364 smp_mb();
3365 if (progress != space_info->reservation_progress)
3366 break;
3367 }
3368
3369 }
3370 if (reclaimed >= to_reclaim && !trans)
3371 btrfs_wait_ordered_extents(root, 0, 0);
3372 return reclaimed >= to_reclaim;
3373 }
3374
3375 /*
3376 * Retries tells us how many times we've called reserve_metadata_bytes. The
3377 * idea is if this is the first call (retries == 0) then we will add to our
3378 * reserved count if we can't make the allocation in order to hold our place
3379 * while we go and try and free up space. That way for retries > 1 we don't try
3380 * and add space, we just check to see if the amount of unused space is >= the
3381 * total space, meaning that our reservation is valid.
3382 *
3383 * However if we don't intend to retry this reservation, pass -1 as retries so
3384 * that it short circuits this logic.
3385 */
3386 static int reserve_metadata_bytes(struct btrfs_trans_handle *trans,
3387 struct btrfs_root *root,
3388 struct btrfs_block_rsv *block_rsv,
3389 u64 orig_bytes, int flush)
3390 {
3391 struct btrfs_space_info *space_info = block_rsv->space_info;
3392 u64 unused;
3393 u64 num_bytes = orig_bytes;
3394 int retries = 0;
3395 int ret = 0;
3396 bool committed = false;
3397 bool flushing = false;
3398
3399 again:
3400 ret = 0;
3401 spin_lock(&space_info->lock);
3402 /*
3403 * We only want to wait if somebody other than us is flushing and we are
3404 * actually alloed to flush.
3405 */
3406 while (flush && !flushing && space_info->flush) {
3407 spin_unlock(&space_info->lock);
3408 /*
3409 * If we have a trans handle we can't wait because the flusher
3410 * may have to commit the transaction, which would mean we would
3411 * deadlock since we are waiting for the flusher to finish, but
3412 * hold the current transaction open.
3413 */
3414 if (trans)
3415 return -EAGAIN;
3416 ret = wait_event_interruptible(space_info->wait,
3417 !space_info->flush);
3418 /* Must have been interrupted, return */
3419 if (ret)
3420 return -EINTR;
3421
3422 spin_lock(&space_info->lock);
3423 }
3424
3425 ret = -ENOSPC;
3426 unused = space_info->bytes_used + space_info->bytes_reserved +
3427 space_info->bytes_pinned + space_info->bytes_readonly +
3428 space_info->bytes_may_use;
3429
3430 /*
3431 * The idea here is that we've not already over-reserved the block group
3432 * then we can go ahead and save our reservation first and then start
3433 * flushing if we need to. Otherwise if we've already overcommitted
3434 * lets start flushing stuff first and then come back and try to make
3435 * our reservation.
3436 */
3437 if (unused <= space_info->total_bytes) {
3438 unused = space_info->total_bytes - unused;
3439 if (unused >= num_bytes) {
3440 space_info->bytes_reserved += orig_bytes;
3441 ret = 0;
3442 } else {
3443 /*
3444 * Ok set num_bytes to orig_bytes since we aren't
3445 * overocmmitted, this way we only try and reclaim what
3446 * we need.
3447 */
3448 num_bytes = orig_bytes;
3449 }
3450 } else {
3451 /*
3452 * Ok we're over committed, set num_bytes to the overcommitted
3453 * amount plus the amount of bytes that we need for this
3454 * reservation.
3455 */
3456 num_bytes = unused - space_info->total_bytes +
3457 (orig_bytes * (retries + 1));
3458 }
3459
3460 /*
3461 * Couldn't make our reservation, save our place so while we're trying
3462 * to reclaim space we can actually use it instead of somebody else
3463 * stealing it from us.
3464 */
3465 if (ret && flush) {
3466 flushing = true;
3467 space_info->flush = 1;
3468 }
3469
3470 spin_unlock(&space_info->lock);
3471
3472 if (!ret || !flush)
3473 goto out;
3474
3475 /*
3476 * We do synchronous shrinking since we don't actually unreserve
3477 * metadata until after the IO is completed.
3478 */
3479 ret = shrink_delalloc(trans, root, num_bytes, 1);
3480 if (ret < 0)
3481 goto out;
3482
3483 ret = 0;
3484
3485 /*
3486 * So if we were overcommitted it's possible that somebody else flushed
3487 * out enough space and we simply didn't have enough space to reclaim,
3488 * so go back around and try again.
3489 */
3490 if (retries < 2) {
3491 retries++;
3492 goto again;
3493 }
3494
3495 /*
3496 * Not enough space to be reclaimed, don't bother committing the
3497 * transaction.
3498 */
3499 spin_lock(&space_info->lock);
3500 if (space_info->bytes_pinned < orig_bytes)
3501 ret = -ENOSPC;
3502 spin_unlock(&space_info->lock);
3503 if (ret)
3504 goto out;
3505
3506 ret = -EAGAIN;
3507 if (trans)
3508 goto out;
3509
3510 ret = -ENOSPC;
3511 if (committed)
3512 goto out;
3513
3514 trans = btrfs_join_transaction(root);
3515 if (IS_ERR(trans))
3516 goto out;
3517 ret = btrfs_commit_transaction(trans, root);
3518 if (!ret) {
3519 trans = NULL;
3520 committed = true;
3521 goto again;
3522 }
3523
3524 out:
3525 if (flushing) {
3526 spin_lock(&space_info->lock);
3527 space_info->flush = 0;
3528 wake_up_all(&space_info->wait);
3529 spin_unlock(&space_info->lock);
3530 }
3531 return ret;
3532 }
3533
3534 static struct btrfs_block_rsv *get_block_rsv(struct btrfs_trans_handle *trans,
3535 struct btrfs_root *root)
3536 {
3537 struct btrfs_block_rsv *block_rsv;
3538 if (root->ref_cows)
3539 block_rsv = trans->block_rsv;
3540 else
3541 block_rsv = root->block_rsv;
3542
3543 if (!block_rsv)
3544 block_rsv = &root->fs_info->empty_block_rsv;
3545
3546 return block_rsv;
3547 }
3548
3549 static int block_rsv_use_bytes(struct btrfs_block_rsv *block_rsv,
3550 u64 num_bytes)
3551 {
3552 int ret = -ENOSPC;
3553 spin_lock(&block_rsv->lock);
3554 if (block_rsv->reserved >= num_bytes) {
3555 block_rsv->reserved -= num_bytes;
3556 if (block_rsv->reserved < block_rsv->size)
3557 block_rsv->full = 0;
3558 ret = 0;
3559 }
3560 spin_unlock(&block_rsv->lock);
3561 return ret;
3562 }
3563
3564 static void block_rsv_add_bytes(struct btrfs_block_rsv *block_rsv,
3565 u64 num_bytes, int update_size)
3566 {
3567 spin_lock(&block_rsv->lock);
3568 block_rsv->reserved += num_bytes;
3569 if (update_size)
3570 block_rsv->size += num_bytes;
3571 else if (block_rsv->reserved >= block_rsv->size)
3572 block_rsv->full = 1;
3573 spin_unlock(&block_rsv->lock);
3574 }
3575
3576 static void block_rsv_release_bytes(struct btrfs_block_rsv *block_rsv,
3577 struct btrfs_block_rsv *dest, u64 num_bytes)
3578 {
3579 struct btrfs_space_info *space_info = block_rsv->space_info;
3580
3581 spin_lock(&block_rsv->lock);
3582 if (num_bytes == (u64)-1)
3583 num_bytes = block_rsv->size;
3584 block_rsv->size -= num_bytes;
3585 if (block_rsv->reserved >= block_rsv->size) {
3586 num_bytes = block_rsv->reserved - block_rsv->size;
3587 block_rsv->reserved = block_rsv->size;
3588 block_rsv->full = 1;
3589 } else {
3590 num_bytes = 0;
3591 }
3592 spin_unlock(&block_rsv->lock);
3593
3594 if (num_bytes > 0) {
3595 if (dest) {
3596 spin_lock(&dest->lock);
3597 if (!dest->full) {
3598 u64 bytes_to_add;
3599
3600 bytes_to_add = dest->size - dest->reserved;
3601 bytes_to_add = min(num_bytes, bytes_to_add);
3602 dest->reserved += bytes_to_add;
3603 if (dest->reserved >= dest->size)
3604 dest->full = 1;
3605 num_bytes -= bytes_to_add;
3606 }
3607 spin_unlock(&dest->lock);
3608 }
3609 if (num_bytes) {
3610 spin_lock(&space_info->lock);
3611 space_info->bytes_reserved -= num_bytes;
3612 space_info->reservation_progress++;
3613 spin_unlock(&space_info->lock);
3614 }
3615 }
3616 }
3617
3618 static int block_rsv_migrate_bytes(struct btrfs_block_rsv *src,
3619 struct btrfs_block_rsv *dst, u64 num_bytes)
3620 {
3621 int ret;
3622
3623 ret = block_rsv_use_bytes(src, num_bytes);
3624 if (ret)
3625 return ret;
3626
3627 block_rsv_add_bytes(dst, num_bytes, 1);
3628 return 0;
3629 }
3630
3631 void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv)
3632 {
3633 memset(rsv, 0, sizeof(*rsv));
3634 spin_lock_init(&rsv->lock);
3635 atomic_set(&rsv->usage, 1);
3636 rsv->priority = 6;
3637 INIT_LIST_HEAD(&rsv->list);
3638 }
3639
3640 struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root)
3641 {
3642 struct btrfs_block_rsv *block_rsv;
3643 struct btrfs_fs_info *fs_info = root->fs_info;
3644
3645 block_rsv = kmalloc(sizeof(*block_rsv), GFP_NOFS);
3646 if (!block_rsv)
3647 return NULL;
3648
3649 btrfs_init_block_rsv(block_rsv);
3650 block_rsv->space_info = __find_space_info(fs_info,
3651 BTRFS_BLOCK_GROUP_METADATA);
3652 return block_rsv;
3653 }
3654
3655 void btrfs_free_block_rsv(struct btrfs_root *root,
3656 struct btrfs_block_rsv *rsv)
3657 {
3658 if (rsv && atomic_dec_and_test(&rsv->usage)) {
3659 btrfs_block_rsv_release(root, rsv, (u64)-1);
3660 if (!rsv->durable)
3661 kfree(rsv);
3662 }
3663 }
3664
3665 /*
3666 * make the block_rsv struct be able to capture freed space.
3667 * the captured space will re-add to the the block_rsv struct
3668 * after transaction commit
3669 */
3670 void btrfs_add_durable_block_rsv(struct btrfs_fs_info *fs_info,
3671 struct btrfs_block_rsv *block_rsv)
3672 {
3673 block_rsv->durable = 1;
3674 mutex_lock(&fs_info->durable_block_rsv_mutex);
3675 list_add_tail(&block_rsv->list, &fs_info->durable_block_rsv_list);
3676 mutex_unlock(&fs_info->durable_block_rsv_mutex);
3677 }
3678
3679 int btrfs_block_rsv_add(struct btrfs_trans_handle *trans,
3680 struct btrfs_root *root,
3681 struct btrfs_block_rsv *block_rsv,
3682 u64 num_bytes)
3683 {
3684 int ret;
3685
3686 if (num_bytes == 0)
3687 return 0;
3688
3689 ret = reserve_metadata_bytes(trans, root, block_rsv, num_bytes, 1);
3690 if (!ret) {
3691 block_rsv_add_bytes(block_rsv, num_bytes, 1);
3692 return 0;
3693 }
3694
3695 return ret;
3696 }
3697
3698 int btrfs_block_rsv_check(struct btrfs_trans_handle *trans,
3699 struct btrfs_root *root,
3700 struct btrfs_block_rsv *block_rsv,
3701 u64 min_reserved, int min_factor)
3702 {
3703 u64 num_bytes = 0;
3704 int commit_trans = 0;
3705 int ret = -ENOSPC;
3706
3707 if (!block_rsv)
3708 return 0;
3709
3710 spin_lock(&block_rsv->lock);
3711 if (min_factor > 0)
3712 num_bytes = div_factor(block_rsv->size, min_factor);
3713 if (min_reserved > num_bytes)
3714 num_bytes = min_reserved;
3715
3716 if (block_rsv->reserved >= num_bytes) {
3717 ret = 0;
3718 } else {
3719 num_bytes -= block_rsv->reserved;
3720 if (block_rsv->durable &&
3721 block_rsv->freed[0] + block_rsv->freed[1] >= num_bytes)
3722 commit_trans = 1;
3723 }
3724 spin_unlock(&block_rsv->lock);
3725 if (!ret)
3726 return 0;
3727
3728 if (block_rsv->refill_used) {
3729 ret = reserve_metadata_bytes(trans, root, block_rsv,
3730 num_bytes, 0);
3731 if (!ret) {
3732 block_rsv_add_bytes(block_rsv, num_bytes, 0);
3733 return 0;
3734 }
3735 }
3736
3737 if (commit_trans) {
3738 if (trans)
3739 return -EAGAIN;
3740 trans = btrfs_join_transaction(root);
3741 BUG_ON(IS_ERR(trans));
3742 ret = btrfs_commit_transaction(trans, root);
3743 return 0;
3744 }
3745
3746 return -ENOSPC;
3747 }
3748
3749 int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv,
3750 struct btrfs_block_rsv *dst_rsv,
3751 u64 num_bytes)
3752 {
3753 return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes);
3754 }
3755
3756 void btrfs_block_rsv_release(struct btrfs_root *root,
3757 struct btrfs_block_rsv *block_rsv,
3758 u64 num_bytes)
3759 {
3760 struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv;
3761 if (global_rsv->full || global_rsv == block_rsv ||
3762 block_rsv->space_info != global_rsv->space_info)
3763 global_rsv = NULL;
3764 block_rsv_release_bytes(block_rsv, global_rsv, num_bytes);
3765 }
3766
3767 /*
3768 * helper to calculate size of global block reservation.
3769 * the desired value is sum of space used by extent tree,
3770 * checksum tree and root tree
3771 */
3772 static u64 calc_global_metadata_size(struct btrfs_fs_info *fs_info)
3773 {
3774 struct btrfs_space_info *sinfo;
3775 u64 num_bytes;
3776 u64 meta_used;
3777 u64 data_used;
3778 int csum_size = btrfs_super_csum_size(&fs_info->super_copy);
3779
3780 sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA);
3781 spin_lock(&sinfo->lock);
3782 data_used = sinfo->bytes_used;
3783 spin_unlock(&sinfo->lock);
3784
3785 sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA);
3786 spin_lock(&sinfo->lock);
3787 if (sinfo->flags & BTRFS_BLOCK_GROUP_DATA)
3788 data_used = 0;
3789 meta_used = sinfo->bytes_used;
3790 spin_unlock(&sinfo->lock);
3791
3792 num_bytes = (data_used >> fs_info->sb->s_blocksize_bits) *
3793 csum_size * 2;
3794 num_bytes += div64_u64(data_used + meta_used, 50);
3795
3796 if (num_bytes * 3 > meta_used)
3797 num_bytes = div64_u64(meta_used, 3);
3798
3799 return ALIGN(num_bytes, fs_info->extent_root->leafsize << 10);
3800 }
3801
3802 static void update_global_block_rsv(struct btrfs_fs_info *fs_info)
3803 {
3804 struct btrfs_block_rsv *block_rsv = &fs_info->global_block_rsv;
3805 struct btrfs_space_info *sinfo = block_rsv->space_info;
3806 u64 num_bytes;
3807
3808 num_bytes = calc_global_metadata_size(fs_info);
3809
3810 spin_lock(&block_rsv->lock);
3811 spin_lock(&sinfo->lock);
3812
3813 block_rsv->size = num_bytes;
3814
3815 num_bytes = sinfo->bytes_used + sinfo->bytes_pinned +
3816 sinfo->bytes_reserved + sinfo->bytes_readonly +
3817 sinfo->bytes_may_use;
3818
3819 if (sinfo->total_bytes > num_bytes) {
3820 num_bytes = sinfo->total_bytes - num_bytes;
3821 block_rsv->reserved += num_bytes;
3822 sinfo->bytes_reserved += num_bytes;
3823 }
3824
3825 if (block_rsv->reserved >= block_rsv->size) {
3826 num_bytes = block_rsv->reserved - block_rsv->size;
3827 sinfo->bytes_reserved -= num_bytes;
3828 sinfo->reservation_progress++;
3829 block_rsv->reserved = block_rsv->size;
3830 block_rsv->full = 1;
3831 }
3832
3833 spin_unlock(&sinfo->lock);
3834 spin_unlock(&block_rsv->lock);
3835 }
3836
3837 static void init_global_block_rsv(struct btrfs_fs_info *fs_info)
3838 {
3839 struct btrfs_space_info *space_info;
3840
3841 space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_SYSTEM);
3842 fs_info->chunk_block_rsv.space_info = space_info;
3843 fs_info->chunk_block_rsv.priority = 10;
3844
3845 space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA);
3846 fs_info->global_block_rsv.space_info = space_info;
3847 fs_info->global_block_rsv.priority = 10;
3848 fs_info->global_block_rsv.refill_used = 1;
3849 fs_info->delalloc_block_rsv.space_info = space_info;
3850 fs_info->trans_block_rsv.space_info = space_info;
3851 fs_info->empty_block_rsv.space_info = space_info;
3852 fs_info->empty_block_rsv.priority = 10;
3853
3854 fs_info->extent_root->block_rsv = &fs_info->global_block_rsv;
3855 fs_info->csum_root->block_rsv = &fs_info->global_block_rsv;
3856 fs_info->dev_root->block_rsv = &fs_info->global_block_rsv;
3857 fs_info->tree_root->block_rsv = &fs_info->global_block_rsv;
3858 fs_info->chunk_root->block_rsv = &fs_info->chunk_block_rsv;
3859
3860 btrfs_add_durable_block_rsv(fs_info, &fs_info->global_block_rsv);
3861
3862 btrfs_add_durable_block_rsv(fs_info, &fs_info->delalloc_block_rsv);
3863
3864 update_global_block_rsv(fs_info);
3865 }
3866
3867 static void release_global_block_rsv(struct btrfs_fs_info *fs_info)
3868 {
3869 block_rsv_release_bytes(&fs_info->global_block_rsv, NULL, (u64)-1);
3870 WARN_ON(fs_info->delalloc_block_rsv.size > 0);
3871 WARN_ON(fs_info->delalloc_block_rsv.reserved > 0);
3872 WARN_ON(fs_info->trans_block_rsv.size > 0);
3873 WARN_ON(fs_info->trans_block_rsv.reserved > 0);
3874 WARN_ON(fs_info->chunk_block_rsv.size > 0);
3875 WARN_ON(fs_info->chunk_block_rsv.reserved > 0);
3876 }
3877
3878 int btrfs_truncate_reserve_metadata(struct btrfs_trans_handle *trans,
3879 struct btrfs_root *root,
3880 struct btrfs_block_rsv *rsv)
3881 {
3882 struct btrfs_block_rsv *trans_rsv = &root->fs_info->trans_block_rsv;
3883 u64 num_bytes;
3884 int ret;
3885
3886 /*
3887 * Truncate should be freeing data, but give us 2 items just in case it
3888 * needs to use some space. We may want to be smarter about this in the
3889 * future.
3890 */
3891 num_bytes = btrfs_calc_trans_metadata_size(root, 2);
3892
3893 /* We already have enough bytes, just return */
3894 if (rsv->reserved >= num_bytes)
3895 return 0;
3896
3897 num_bytes -= rsv->reserved;
3898
3899 /*
3900 * You should have reserved enough space before hand to do this, so this
3901 * should not fail.
3902 */
3903 ret = block_rsv_migrate_bytes(trans_rsv, rsv, num_bytes);
3904 BUG_ON(ret);
3905
3906 return 0;
3907 }
3908
3909 void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans,
3910 struct btrfs_root *root)
3911 {
3912 if (!trans->bytes_reserved)
3913 return;
3914
3915 BUG_ON(trans->block_rsv != &root->fs_info->trans_block_rsv);
3916 btrfs_block_rsv_release(root, trans->block_rsv,
3917 trans->bytes_reserved);
3918 trans->bytes_reserved = 0;
3919 }
3920
3921 int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans,
3922 struct inode *inode)
3923 {
3924 struct btrfs_root *root = BTRFS_I(inode)->root;
3925 struct btrfs_block_rsv *src_rsv = get_block_rsv(trans, root);
3926 struct btrfs_block_rsv *dst_rsv = root->orphan_block_rsv;
3927
3928 /*
3929 * We need to hold space in order to delete our orphan item once we've
3930 * added it, so this takes the reservation so we can release it later
3931 * when we are truly done with the orphan item.
3932 */
3933 u64 num_bytes = btrfs_calc_trans_metadata_size(root, 1);
3934 return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes);
3935 }
3936
3937 void btrfs_orphan_release_metadata(struct inode *inode)
3938 {
3939 struct btrfs_root *root = BTRFS_I(inode)->root;
3940 u64 num_bytes = btrfs_calc_trans_metadata_size(root, 1);
3941 btrfs_block_rsv_release(root, root->orphan_block_rsv, num_bytes);
3942 }
3943
3944 int btrfs_snap_reserve_metadata(struct btrfs_trans_handle *trans,
3945 struct btrfs_pending_snapshot *pending)
3946 {
3947 struct btrfs_root *root = pending->root;
3948 struct btrfs_block_rsv *src_rsv = get_block_rsv(trans, root);
3949 struct btrfs_block_rsv *dst_rsv = &pending->block_rsv;
3950 /*
3951 * two for root back/forward refs, two for directory entries
3952 * and one for root of the snapshot.
3953 */
3954 u64 num_bytes = btrfs_calc_trans_metadata_size(root, 5);
3955 dst_rsv->space_info = src_rsv->space_info;
3956 return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes);
3957 }
3958
3959 static unsigned drop_outstanding_extent(struct inode *inode)
3960 {
3961 unsigned dropped_extents = 0;
3962
3963 spin_lock(&BTRFS_I(inode)->lock);
3964 BUG_ON(!BTRFS_I(inode)->outstanding_extents);
3965 BTRFS_I(inode)->outstanding_extents--;
3966
3967 /*
3968 * If we have more or the same amount of outsanding extents than we have
3969 * reserved then we need to leave the reserved extents count alone.
3970 */
3971 if (BTRFS_I(inode)->outstanding_extents >=
3972 BTRFS_I(inode)->reserved_extents)
3973 goto out;
3974
3975 dropped_extents = BTRFS_I(inode)->reserved_extents -
3976 BTRFS_I(inode)->outstanding_extents;
3977 BTRFS_I(inode)->reserved_extents -= dropped_extents;
3978 out:
3979 spin_unlock(&BTRFS_I(inode)->lock);
3980 return dropped_extents;
3981 }
3982
3983 static u64 calc_csum_metadata_size(struct inode *inode, u64 num_bytes)
3984 {
3985 return num_bytes >>= 3;
3986 }
3987
3988 int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes)
3989 {
3990 struct btrfs_root *root = BTRFS_I(inode)->root;
3991 struct btrfs_block_rsv *block_rsv = &root->fs_info->delalloc_block_rsv;
3992 u64 to_reserve = 0;
3993 unsigned nr_extents = 0;
3994 int ret;
3995
3996 if (btrfs_transaction_in_commit(root->fs_info))
3997 schedule_timeout(1);
3998
3999 num_bytes = ALIGN(num_bytes, root->sectorsize);
4000
4001 spin_lock(&BTRFS_I(inode)->lock);
4002 BTRFS_I(inode)->outstanding_extents++;
4003
4004 if (BTRFS_I(inode)->outstanding_extents >
4005 BTRFS_I(inode)->reserved_extents) {
4006 nr_extents = BTRFS_I(inode)->outstanding_extents -
4007 BTRFS_I(inode)->reserved_extents;
4008 BTRFS_I(inode)->reserved_extents += nr_extents;
4009
4010 to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents);
4011 }
4012 spin_unlock(&BTRFS_I(inode)->lock);
4013
4014 to_reserve += calc_csum_metadata_size(inode, num_bytes);
4015 ret = reserve_metadata_bytes(NULL, root, block_rsv, to_reserve, 1);
4016 if (ret) {
4017 unsigned dropped;
4018 /*
4019 * We don't need the return value since our reservation failed,
4020 * we just need to clean up our counter.
4021 */
4022 dropped = drop_outstanding_extent(inode);
4023 WARN_ON(dropped > 1);
4024 return ret;
4025 }
4026
4027 block_rsv_add_bytes(block_rsv, to_reserve, 1);
4028
4029 return 0;
4030 }
4031
4032 void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes)
4033 {
4034 struct btrfs_root *root = BTRFS_I(inode)->root;
4035 u64 to_free = 0;
4036 unsigned dropped;
4037
4038 num_bytes = ALIGN(num_bytes, root->sectorsize);
4039 dropped = drop_outstanding_extent(inode);
4040
4041 to_free = calc_csum_metadata_size(inode, num_bytes);
4042 if (dropped > 0)
4043 to_free += btrfs_calc_trans_metadata_size(root, dropped);
4044
4045 btrfs_block_rsv_release(root, &root->fs_info->delalloc_block_rsv,
4046 to_free);
4047 }
4048
4049 int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes)
4050 {
4051 int ret;
4052
4053 ret = btrfs_check_data_free_space(inode, num_bytes);
4054 if (ret)
4055 return ret;
4056
4057 ret = btrfs_delalloc_reserve_metadata(inode, num_bytes);
4058 if (ret) {
4059 btrfs_free_reserved_data_space(inode, num_bytes);
4060 return ret;
4061 }
4062
4063 return 0;
4064 }
4065
4066 void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes)
4067 {
4068 btrfs_delalloc_release_metadata(inode, num_bytes);
4069 btrfs_free_reserved_data_space(inode, num_bytes);
4070 }
4071
4072 static int update_block_group(struct btrfs_trans_handle *trans,
4073 struct btrfs_root *root,
4074 u64 bytenr, u64 num_bytes, int alloc)
4075 {
4076 struct btrfs_block_group_cache *cache = NULL;
4077 struct btrfs_fs_info *info = root->fs_info;
4078 u64 total = num_bytes;
4079 u64 old_val;
4080 u64 byte_in_group;
4081 int factor;
4082
4083 /* block accounting for super block */
4084 spin_lock(&info->delalloc_lock);
4085 old_val = btrfs_super_bytes_used(&info->super_copy);
4086 if (alloc)
4087 old_val += num_bytes;
4088 else
4089 old_val -= num_bytes;
4090 btrfs_set_super_bytes_used(&info->super_copy, old_val);
4091 spin_unlock(&info->delalloc_lock);
4092
4093 while (total) {
4094 cache = btrfs_lookup_block_group(info, bytenr);
4095 if (!cache)
4096 return -1;
4097 if (cache->flags & (BTRFS_BLOCK_GROUP_DUP |
4098 BTRFS_BLOCK_GROUP_RAID1 |
4099 BTRFS_BLOCK_GROUP_RAID10))
4100 factor = 2;
4101 else
4102 factor = 1;
4103 /*
4104 * If this block group has free space cache written out, we
4105 * need to make sure to load it if we are removing space. This
4106 * is because we need the unpinning stage to actually add the
4107 * space back to the block group, otherwise we will leak space.
4108 */
4109 if (!alloc && cache->cached == BTRFS_CACHE_NO)
4110 cache_block_group(cache, trans, NULL, 1);
4111
4112 byte_in_group = bytenr - cache->key.objectid;
4113 WARN_ON(byte_in_group > cache->key.offset);
4114
4115 spin_lock(&cache->space_info->lock);
4116 spin_lock(&cache->lock);
4117
4118 if (btrfs_super_cache_generation(&info->super_copy) != 0 &&
4119 cache->disk_cache_state < BTRFS_DC_CLEAR)
4120 cache->disk_cache_state = BTRFS_DC_CLEAR;
4121
4122 cache->dirty = 1;
4123 old_val = btrfs_block_group_used(&cache->item);
4124 num_bytes = min(total, cache->key.offset - byte_in_group);
4125 if (alloc) {
4126 old_val += num_bytes;
4127 btrfs_set_block_group_used(&cache->item, old_val);
4128 cache->reserved -= num_bytes;
4129 cache->space_info->bytes_reserved -= num_bytes;
4130 cache->space_info->reservation_progress++;
4131 cache->space_info->bytes_used += num_bytes;
4132 cache->space_info->disk_used += num_bytes * factor;
4133 spin_unlock(&cache->lock);
4134 spin_unlock(&cache->space_info->lock);
4135 } else {
4136 old_val -= num_bytes;
4137 btrfs_set_block_group_used(&cache->item, old_val);
4138 cache->pinned += num_bytes;
4139 cache->space_info->bytes_pinned += num_bytes;
4140 cache->space_info->bytes_used -= num_bytes;
4141 cache->space_info->disk_used -= num_bytes * factor;
4142 spin_unlock(&cache->lock);
4143 spin_unlock(&cache->space_info->lock);
4144
4145 set_extent_dirty(info->pinned_extents,
4146 bytenr, bytenr + num_bytes - 1,
4147 GFP_NOFS | __GFP_NOFAIL);
4148 }
4149 btrfs_put_block_group(cache);
4150 total -= num_bytes;
4151 bytenr += num_bytes;
4152 }
4153 return 0;
4154 }
4155
4156 static u64 first_logical_byte(struct btrfs_root *root, u64 search_start)
4157 {
4158 struct btrfs_block_group_cache *cache;
4159 u64 bytenr;
4160
4161 cache = btrfs_lookup_first_block_group(root->fs_info, search_start);
4162 if (!cache)
4163 return 0;
4164
4165 bytenr = cache->key.objectid;
4166 btrfs_put_block_group(cache);
4167
4168 return bytenr;
4169 }
4170
4171 static int pin_down_extent(struct btrfs_root *root,
4172 struct btrfs_block_group_cache *cache,
4173 u64 bytenr, u64 num_bytes, int reserved)
4174 {
4175 spin_lock(&cache->space_info->lock);
4176 spin_lock(&cache->lock);
4177 cache->pinned += num_bytes;
4178 cache->space_info->bytes_pinned += num_bytes;
4179 if (reserved) {
4180 cache->reserved -= num_bytes;
4181 cache->space_info->bytes_reserved -= num_bytes;
4182 cache->space_info->reservation_progress++;
4183 }
4184 spin_unlock(&cache->lock);
4185 spin_unlock(&cache->space_info->lock);
4186
4187 set_extent_dirty(root->fs_info->pinned_extents, bytenr,
4188 bytenr + num_bytes - 1, GFP_NOFS | __GFP_NOFAIL);
4189 return 0;
4190 }
4191
4192 /*
4193 * this function must be called within transaction
4194 */
4195 int btrfs_pin_extent(struct btrfs_root *root,
4196 u64 bytenr, u64 num_bytes, int reserved)
4197 {
4198 struct btrfs_block_group_cache *cache;
4199
4200 cache = btrfs_lookup_block_group(root->fs_info, bytenr);
4201 BUG_ON(!cache);
4202
4203 pin_down_extent(root, cache, bytenr, num_bytes, reserved);
4204
4205 btrfs_put_block_group(cache);
4206 return 0;
4207 }
4208
4209 /*
4210 * update size of reserved extents. this function may return -EAGAIN
4211 * if 'reserve' is true or 'sinfo' is false.
4212 */
4213 int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache,
4214 u64 num_bytes, int reserve, int sinfo)
4215 {
4216 int ret = 0;
4217 if (sinfo) {
4218 struct btrfs_space_info *space_info = cache->space_info;
4219 spin_lock(&space_info->lock);
4220 spin_lock(&cache->lock);
4221 if (reserve) {
4222 if (cache->ro) {
4223 ret = -EAGAIN;
4224 } else {
4225 cache->reserved += num_bytes;
4226 space_info->bytes_reserved += num_bytes;
4227 }
4228 } else {
4229 if (cache->ro)
4230 space_info->bytes_readonly += num_bytes;
4231 cache->reserved -= num_bytes;
4232 space_info->bytes_reserved -= num_bytes;
4233 space_info->reservation_progress++;
4234 }
4235 spin_unlock(&cache->lock);
4236 spin_unlock(&space_info->lock);
4237 } else {
4238 spin_lock(&cache->lock);
4239 if (cache->ro) {
4240 ret = -EAGAIN;
4241 } else {
4242 if (reserve)
4243 cache->reserved += num_bytes;
4244 else
4245 cache->reserved -= num_bytes;
4246 }
4247 spin_unlock(&cache->lock);
4248 }
4249 return ret;
4250 }
4251
4252 int btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans,
4253 struct btrfs_root *root)
4254 {
4255 struct btrfs_fs_info *fs_info = root->fs_info;
4256 struct btrfs_caching_control *next;
4257 struct btrfs_caching_control *caching_ctl;
4258 struct btrfs_block_group_cache *cache;
4259
4260 down_write(&fs_info->extent_commit_sem);
4261
4262 list_for_each_entry_safe(caching_ctl, next,
4263 &fs_info->caching_block_groups, list) {
4264 cache = caching_ctl->block_group;
4265 if (block_group_cache_done(cache)) {
4266 cache->last_byte_to_unpin = (u64)-1;
4267 list_del_init(&caching_ctl->list);
4268 put_caching_control(caching_ctl);
4269 } else {
4270 cache->last_byte_to_unpin = caching_ctl->progress;
4271 }
4272 }
4273
4274 if (fs_info->pinned_extents == &fs_info->freed_extents[0])
4275 fs_info->pinned_extents = &fs_info->freed_extents[1];
4276 else
4277 fs_info->pinned_extents = &fs_info->freed_extents[0];
4278
4279 up_write(&fs_info->extent_commit_sem);
4280
4281 update_global_block_rsv(fs_info);
4282 return 0;
4283 }
4284
4285 static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end)
4286 {
4287 struct btrfs_fs_info *fs_info = root->fs_info;
4288 struct btrfs_block_group_cache *cache = NULL;
4289 u64 len;
4290
4291 while (start <= end) {
4292 if (!cache ||
4293 start >= cache->key.objectid + cache->key.offset) {
4294 if (cache)
4295 btrfs_put_block_group(cache);
4296 cache = btrfs_lookup_block_group(fs_info, start);
4297 BUG_ON(!cache);
4298 }
4299
4300 len = cache->key.objectid + cache->key.offset - start;
4301 len = min(len, end + 1 - start);
4302
4303 if (start < cache->last_byte_to_unpin) {
4304 len = min(len, cache->last_byte_to_unpin - start);
4305 btrfs_add_free_space(cache, start, len);
4306 }
4307
4308 start += len;
4309
4310 spin_lock(&cache->space_info->lock);
4311 spin_lock(&cache->lock);
4312 cache->pinned -= len;
4313 cache->space_info->bytes_pinned -= len;
4314 if (cache->ro) {
4315 cache->space_info->bytes_readonly += len;
4316 } else if (cache->reserved_pinned > 0) {
4317 len = min(len, cache->reserved_pinned);
4318 cache->reserved_pinned -= len;
4319 cache->space_info->bytes_reserved += len;
4320 }
4321 spin_unlock(&cache->lock);
4322 spin_unlock(&cache->space_info->lock);
4323 }
4324
4325 if (cache)
4326 btrfs_put_block_group(cache);
4327 return 0;
4328 }
4329
4330 int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
4331 struct btrfs_root *root)
4332 {
4333 struct btrfs_fs_info *fs_info = root->fs_info;
4334 struct extent_io_tree *unpin;
4335 struct btrfs_block_rsv *block_rsv;
4336 struct btrfs_block_rsv *next_rsv;
4337 u64 start;
4338 u64 end;
4339 int idx;
4340 int ret;
4341
4342 if (fs_info->pinned_extents == &fs_info->freed_extents[0])
4343 unpin = &fs_info->freed_extents[1];
4344 else
4345 unpin = &fs_info->freed_extents[0];
4346
4347 while (1) {
4348 ret = find_first_extent_bit(unpin, 0, &start, &end,
4349 EXTENT_DIRTY);
4350 if (ret)
4351 break;
4352
4353 if (btrfs_test_opt(root, DISCARD))
4354 ret = btrfs_discard_extent(root, start,
4355 end + 1 - start, NULL);
4356
4357 clear_extent_dirty(unpin, start, end, GFP_NOFS);
4358 unpin_extent_range(root, start, end);
4359 cond_resched();
4360 }
4361
4362 mutex_lock(&fs_info->durable_block_rsv_mutex);
4363 list_for_each_entry_safe(block_rsv, next_rsv,
4364 &fs_info->durable_block_rsv_list, list) {
4365
4366 idx = trans->transid & 0x1;
4367 if (block_rsv->freed[idx] > 0) {
4368 block_rsv_add_bytes(block_rsv,
4369 block_rsv->freed[idx], 0);
4370 block_rsv->freed[idx] = 0;
4371 }
4372 if (atomic_read(&block_rsv->usage) == 0) {
4373 btrfs_block_rsv_release(root, block_rsv, (u64)-1);
4374
4375 if (block_rsv->freed[0] == 0 &&
4376 block_rsv->freed[1] == 0) {
4377 list_del_init(&block_rsv->list);
4378 kfree(block_rsv);
4379 }
4380 } else {
4381 btrfs_block_rsv_release(root, block_rsv, 0);
4382 }
4383 }
4384 mutex_unlock(&fs_info->durable_block_rsv_mutex);
4385
4386 return 0;
4387 }
4388
4389 static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
4390 struct btrfs_root *root,
4391 u64 bytenr, u64 num_bytes, u64 parent,
4392 u64 root_objectid, u64 owner_objectid,
4393 u64 owner_offset, int refs_to_drop,
4394 struct btrfs_delayed_extent_op *extent_op)
4395 {
4396 struct btrfs_key key;
4397 struct btrfs_path *path;
4398 struct btrfs_fs_info *info = root->fs_info;
4399 struct btrfs_root *extent_root = info->extent_root;
4400 struct extent_buffer *leaf;
4401 struct btrfs_extent_item *ei;
4402 struct btrfs_extent_inline_ref *iref;
4403 int ret;
4404 int is_data;
4405 int extent_slot = 0;
4406 int found_extent = 0;
4407 int num_to_del = 1;
4408 u32 item_size;
4409 u64 refs;
4410
4411 path = btrfs_alloc_path();
4412 if (!path)
4413 return -ENOMEM;
4414
4415 path->reada = 1;
4416 path->leave_spinning = 1;
4417
4418 is_data = owner_objectid >= BTRFS_FIRST_FREE_OBJECTID;
4419 BUG_ON(!is_data && refs_to_drop != 1);
4420
4421 ret = lookup_extent_backref(trans, extent_root, path, &iref,
4422 bytenr, num_bytes, parent,
4423 root_objectid, owner_objectid,
4424 owner_offset);
4425 if (ret == 0) {
4426 extent_slot = path->slots[0];
4427 while (extent_slot >= 0) {
4428 btrfs_item_key_to_cpu(path->nodes[0], &key,
4429 extent_slot);
4430 if (key.objectid != bytenr)
4431 break;
4432 if (key.type == BTRFS_EXTENT_ITEM_KEY &&
4433 key.offset == num_bytes) {
4434 found_extent = 1;
4435 break;
4436 }
4437 if (path->slots[0] - extent_slot > 5)
4438 break;
4439 extent_slot--;
4440 }
4441 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
4442 item_size = btrfs_item_size_nr(path->nodes[0], extent_slot);
4443 if (found_extent && item_size < sizeof(*ei))
4444 found_extent = 0;
4445 #endif
4446 if (!found_extent) {
4447 BUG_ON(iref);
4448 ret = remove_extent_backref(trans, extent_root, path,
4449 NULL, refs_to_drop,
4450 is_data);
4451 BUG_ON(ret);
4452 btrfs_release_path(path);
4453 path->leave_spinning = 1;
4454
4455 key.objectid = bytenr;
4456 key.type = BTRFS_EXTENT_ITEM_KEY;
4457 key.offset = num_bytes;
4458
4459 ret = btrfs_search_slot(trans, extent_root,
4460 &key, path, -1, 1);
4461 if (ret) {
4462 printk(KERN_ERR "umm, got %d back from search"
4463 ", was looking for %llu\n", ret,
4464 (unsigned long long)bytenr);
4465 if (ret > 0)
4466 btrfs_print_leaf(extent_root,
4467 path->nodes[0]);
4468 }
4469 BUG_ON(ret);
4470 extent_slot = path->slots[0];
4471 }
4472 } else {
4473 btrfs_print_leaf(extent_root, path->nodes[0]);
4474 WARN_ON(1);
4475 printk(KERN_ERR "btrfs unable to find ref byte nr %llu "
4476 "parent %llu root %llu owner %llu offset %llu\n",
4477 (unsigned long long)bytenr,
4478 (unsigned long long)parent,
4479 (unsigned long long)root_objectid,
4480 (unsigned long long)owner_objectid,
4481 (unsigned long long)owner_offset);
4482 }
4483
4484 leaf = path->nodes[0];
4485 item_size = btrfs_item_size_nr(leaf, extent_slot);
4486 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
4487 if (item_size < sizeof(*ei)) {
4488 BUG_ON(found_extent || extent_slot != path->slots[0]);
4489 ret = convert_extent_item_v0(trans, extent_root, path,
4490 owner_objectid, 0);
4491 BUG_ON(ret < 0);
4492
4493 btrfs_release_path(path);
4494 path->leave_spinning = 1;
4495
4496 key.objectid = bytenr;
4497 key.type = BTRFS_EXTENT_ITEM_KEY;
4498 key.offset = num_bytes;
4499
4500 ret = btrfs_search_slot(trans, extent_root, &key, path,
4501 -1, 1);
4502 if (ret) {
4503 printk(KERN_ERR "umm, got %d back from search"
4504 ", was looking for %llu\n", ret,
4505 (unsigned long long)bytenr);
4506 btrfs_print_leaf(extent_root, path->nodes[0]);
4507 }
4508 BUG_ON(ret);
4509 extent_slot = path->slots[0];
4510 leaf = path->nodes[0];
4511 item_size = btrfs_item_size_nr(leaf, extent_slot);
4512 }
4513 #endif
4514 BUG_ON(item_size < sizeof(*ei));
4515 ei = btrfs_item_ptr(leaf, extent_slot,
4516 struct btrfs_extent_item);
4517 if (owner_objectid < BTRFS_FIRST_FREE_OBJECTID) {
4518 struct btrfs_tree_block_info *bi;
4519 BUG_ON(item_size < sizeof(*ei) + sizeof(*bi));
4520 bi = (struct btrfs_tree_block_info *)(ei + 1);
4521 WARN_ON(owner_objectid != btrfs_tree_block_level(leaf, bi));
4522 }
4523
4524 refs = btrfs_extent_refs(leaf, ei);
4525 BUG_ON(refs < refs_to_drop);
4526 refs -= refs_to_drop;
4527
4528 if (refs > 0) {
4529 if (extent_op)
4530 __run_delayed_extent_op(extent_op, leaf, ei);
4531 /*
4532 * In the case of inline back ref, reference count will
4533 * be updated by remove_extent_backref
4534 */
4535 if (iref) {
4536 BUG_ON(!found_extent);
4537 } else {
4538 btrfs_set_extent_refs(leaf, ei, refs);
4539 btrfs_mark_buffer_dirty(leaf);
4540 }
4541 if (found_extent) {
4542 ret = remove_extent_backref(trans, extent_root, path,
4543 iref, refs_to_drop,
4544 is_data);
4545 BUG_ON(ret);
4546 }
4547 } else {
4548 if (found_extent) {
4549 BUG_ON(is_data && refs_to_drop !=
4550 extent_data_ref_count(root, path, iref));
4551 if (iref) {
4552 BUG_ON(path->slots[0] != extent_slot);
4553 } else {
4554 BUG_ON(path->slots[0] != extent_slot + 1);
4555 path->slots[0] = extent_slot;
4556 num_to_del = 2;
4557 }
4558 }
4559
4560 ret = btrfs_del_items(trans, extent_root, path, path->slots[0],
4561 num_to_del);
4562 BUG_ON(ret);
4563 btrfs_release_path(path);
4564
4565 if (is_data) {
4566 ret = btrfs_del_csums(trans, root, bytenr, num_bytes);
4567 BUG_ON(ret);
4568 } else {
4569 invalidate_mapping_pages(info->btree_inode->i_mapping,
4570 bytenr >> PAGE_CACHE_SHIFT,
4571 (bytenr + num_bytes - 1) >> PAGE_CACHE_SHIFT);
4572 }
4573
4574 ret = update_block_group(trans, root, bytenr, num_bytes, 0);
4575 BUG_ON(ret);
4576 }
4577 btrfs_free_path(path);
4578 return ret;
4579 }
4580
4581 /*
4582 * when we free an block, it is possible (and likely) that we free the last
4583 * delayed ref for that extent as well. This searches the delayed ref tree for
4584 * a given extent, and if there are no other delayed refs to be processed, it
4585 * removes it from the tree.
4586 */
4587 static noinline int check_ref_cleanup(struct btrfs_trans_handle *trans,
4588 struct btrfs_root *root, u64 bytenr)
4589 {
4590 struct btrfs_delayed_ref_head *head;
4591 struct btrfs_delayed_ref_root *delayed_refs;
4592 struct btrfs_delayed_ref_node *ref;
4593 struct rb_node *node;
4594 int ret = 0;
4595
4596 delayed_refs = &trans->transaction->delayed_refs;
4597 spin_lock(&delayed_refs->lock);
4598 head = btrfs_find_delayed_ref_head(trans, bytenr);
4599 if (!head)
4600 goto out;
4601
4602 node = rb_prev(&head->node.rb_node);
4603 if (!node)
4604 goto out;
4605
4606 ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
4607
4608 /* there are still entries for this ref, we can't drop it */
4609 if (ref->bytenr == bytenr)
4610 goto out;
4611
4612 if (head->extent_op) {
4613 if (!head->must_insert_reserved)
4614 goto out;
4615 kfree(head->extent_op);
4616 head->extent_op = NULL;
4617 }
4618
4619 /*
4620 * waiting for the lock here would deadlock. If someone else has it
4621 * locked they are already in the process of dropping it anyway
4622 */
4623 if (!mutex_trylock(&head->mutex))
4624 goto out;
4625
4626 /*
4627 * at this point we have a head with no other entries. Go
4628 * ahead and process it.
4629 */
4630 head->node.in_tree = 0;
4631 rb_erase(&head->node.rb_node, &delayed_refs->root);
4632
4633 delayed_refs->num_entries--;
4634
4635 /*
4636 * we don't take a ref on the node because we're removing it from the
4637 * tree, so we just steal the ref the tree was holding.
4638 */
4639 delayed_refs->num_heads--;
4640 if (list_empty(&head->cluster))
4641 delayed_refs->num_heads_ready--;
4642
4643 list_del_init(&head->cluster);
4644 spin_unlock(&delayed_refs->lock);
4645
4646 BUG_ON(head->extent_op);
4647 if (head->must_insert_reserved)
4648 ret = 1;
4649
4650 mutex_unlock(&head->mutex);
4651 btrfs_put_delayed_ref(&head->node);
4652 return ret;
4653 out:
4654 spin_unlock(&delayed_refs->lock);
4655 return 0;
4656 }
4657
4658 void btrfs_free_tree_block(struct btrfs_trans_handle *trans,
4659 struct btrfs_root *root,
4660 struct extent_buffer *buf,
4661 u64 parent, int last_ref)
4662 {
4663 struct btrfs_block_rsv *block_rsv;
4664 struct btrfs_block_group_cache *cache = NULL;
4665 int ret;
4666
4667 if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
4668 ret = btrfs_add_delayed_tree_ref(trans, buf->start, buf->len,
4669 parent, root->root_key.objectid,
4670 btrfs_header_level(buf),
4671 BTRFS_DROP_DELAYED_REF, NULL);
4672 BUG_ON(ret);
4673 }
4674
4675 if (!last_ref)
4676 return;
4677
4678 block_rsv = get_block_rsv(trans, root);
4679 cache = btrfs_lookup_block_group(root->fs_info, buf->start);
4680 if (block_rsv->space_info != cache->space_info)
4681 goto out;
4682
4683 if (btrfs_header_generation(buf) == trans->transid) {
4684 if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
4685 ret = check_ref_cleanup(trans, root, buf->start);
4686 if (!ret)
4687 goto pin;
4688 }
4689
4690 if (btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) {
4691 pin_down_extent(root, cache, buf->start, buf->len, 1);
4692 goto pin;
4693 }
4694
4695 WARN_ON(test_bit(EXTENT_BUFFER_DIRTY, &buf->bflags));
4696
4697 btrfs_add_free_space(cache, buf->start, buf->len);
4698 ret = btrfs_update_reserved_bytes(cache, buf->len, 0, 0);
4699 if (ret == -EAGAIN) {
4700 /* block group became read-only */
4701 btrfs_update_reserved_bytes(cache, buf->len, 0, 1);
4702 goto out;
4703 }
4704
4705 ret = 1;
4706 spin_lock(&block_rsv->lock);
4707 if (block_rsv->reserved < block_rsv->size) {
4708 block_rsv->reserved += buf->len;
4709 ret = 0;
4710 }
4711 spin_unlock(&block_rsv->lock);
4712
4713 if (ret) {
4714 spin_lock(&cache->space_info->lock);
4715 cache->space_info->bytes_reserved -= buf->len;
4716 cache->space_info->reservation_progress++;
4717 spin_unlock(&cache->space_info->lock);
4718 }
4719 goto out;
4720 }
4721 pin:
4722 if (block_rsv->durable && !cache->ro) {
4723 ret = 0;
4724 spin_lock(&cache->lock);
4725 if (!cache->ro) {
4726 cache->reserved_pinned += buf->len;
4727 ret = 1;
4728 }
4729 spin_unlock(&cache->lock);
4730
4731 if (ret) {
4732 spin_lock(&block_rsv->lock);
4733 block_rsv->freed[trans->transid & 0x1] += buf->len;
4734 spin_unlock(&block_rsv->lock);
4735 }
4736 }
4737 out:
4738 /*
4739 * Deleting the buffer, clear the corrupt flag since it doesn't matter
4740 * anymore.
4741 */
4742 clear_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags);
4743 btrfs_put_block_group(cache);
4744 }
4745
4746 int btrfs_free_extent(struct btrfs_trans_handle *trans,
4747 struct btrfs_root *root,
4748 u64 bytenr, u64 num_bytes, u64 parent,
4749 u64 root_objectid, u64 owner, u64 offset)
4750 {
4751 int ret;
4752
4753 /*
4754 * tree log blocks never actually go into the extent allocation
4755 * tree, just update pinning info and exit early.
4756 */
4757 if (root_objectid == BTRFS_TREE_LOG_OBJECTID) {
4758 WARN_ON(owner >= BTRFS_FIRST_FREE_OBJECTID);
4759 /* unlocks the pinned mutex */
4760 btrfs_pin_extent(root, bytenr, num_bytes, 1);
4761 ret = 0;
4762 } else if (owner < BTRFS_FIRST_FREE_OBJECTID) {
4763 ret = btrfs_add_delayed_tree_ref(trans, bytenr, num_bytes,
4764 parent, root_objectid, (int)owner,
4765 BTRFS_DROP_DELAYED_REF, NULL);
4766 BUG_ON(ret);
4767 } else {
4768 ret = btrfs_add_delayed_data_ref(trans, bytenr, num_bytes,
4769 parent, root_objectid, owner,
4770 offset, BTRFS_DROP_DELAYED_REF, NULL);
4771 BUG_ON(ret);
4772 }
4773 return ret;
4774 }
4775
4776 static u64 stripe_align(struct btrfs_root *root, u64 val)
4777 {
4778 u64 mask = ((u64)root->stripesize - 1);
4779 u64 ret = (val + mask) & ~mask;
4780 return ret;
4781 }
4782
4783 /*
4784 * when we wait for progress in the block group caching, its because
4785 * our allocation attempt failed at least once. So, we must sleep
4786 * and let some progress happen before we try again.
4787 *
4788 * This function will sleep at least once waiting for new free space to
4789 * show up, and then it will check the block group free space numbers
4790 * for our min num_bytes. Another option is to have it go ahead
4791 * and look in the rbtree for a free extent of a given size, but this
4792 * is a good start.
4793 */
4794 static noinline int
4795 wait_block_group_cache_progress(struct btrfs_block_group_cache *cache,
4796 u64 num_bytes)
4797 {
4798 struct btrfs_caching_control *caching_ctl;
4799 DEFINE_WAIT(wait);
4800
4801 caching_ctl = get_caching_control(cache);
4802 if (!caching_ctl)
4803 return 0;
4804
4805 wait_event(caching_ctl->wait, block_group_cache_done(cache) ||
4806 (cache->free_space_ctl->free_space >= num_bytes));
4807
4808 put_caching_control(caching_ctl);
4809 return 0;
4810 }
4811
4812 static noinline int
4813 wait_block_group_cache_done(struct btrfs_block_group_cache *cache)
4814 {
4815 struct btrfs_caching_control *caching_ctl;
4816 DEFINE_WAIT(wait);
4817
4818 caching_ctl = get_caching_control(cache);
4819 if (!caching_ctl)
4820 return 0;
4821
4822 wait_event(caching_ctl->wait, block_group_cache_done(cache));
4823
4824 put_caching_control(caching_ctl);
4825 return 0;
4826 }
4827
4828 static int get_block_group_index(struct btrfs_block_group_cache *cache)
4829 {
4830 int index;
4831 if (cache->flags & BTRFS_BLOCK_GROUP_RAID10)
4832 index = 0;
4833 else if (cache->flags & BTRFS_BLOCK_GROUP_RAID1)
4834 index = 1;
4835 else if (cache->flags & BTRFS_BLOCK_GROUP_DUP)
4836 index = 2;
4837 else if (cache->flags & BTRFS_BLOCK_GROUP_RAID0)
4838 index = 3;
4839 else
4840 index = 4;
4841 return index;
4842 }
4843
4844 enum btrfs_loop_type {
4845 LOOP_FIND_IDEAL = 0,
4846 LOOP_CACHING_NOWAIT = 1,
4847 LOOP_CACHING_WAIT = 2,
4848 LOOP_ALLOC_CHUNK = 3,
4849 LOOP_NO_EMPTY_SIZE = 4,
4850 };
4851
4852 /*
4853 * walks the btree of allocated extents and find a hole of a given size.
4854 * The key ins is changed to record the hole:
4855 * ins->objectid == block start
4856 * ins->flags = BTRFS_EXTENT_ITEM_KEY
4857 * ins->offset == number of blocks
4858 * Any available blocks before search_start are skipped.
4859 */
4860 static noinline int find_free_extent(struct btrfs_trans_handle *trans,
4861 struct btrfs_root *orig_root,
4862 u64 num_bytes, u64 empty_size,
4863 u64 search_start, u64 search_end,
4864 u64 hint_byte, struct btrfs_key *ins,
4865 u64 data)
4866 {
4867 int ret = 0;
4868 struct btrfs_root *root = orig_root->fs_info->extent_root;
4869 struct btrfs_free_cluster *last_ptr = NULL;
4870 struct btrfs_block_group_cache *block_group = NULL;
4871 int empty_cluster = 2 * 1024 * 1024;
4872 int allowed_chunk_alloc = 0;
4873 int done_chunk_alloc = 0;
4874 struct btrfs_space_info *space_info;
4875 int last_ptr_loop = 0;
4876 int loop = 0;
4877 int index = 0;
4878 bool found_uncached_bg = false;
4879 bool failed_cluster_refill = false;
4880 bool failed_alloc = false;
4881 bool use_cluster = true;
4882 u64 ideal_cache_percent = 0;
4883 u64 ideal_cache_offset = 0;
4884
4885 WARN_ON(num_bytes < root->sectorsize);
4886 btrfs_set_key_type(ins, BTRFS_EXTENT_ITEM_KEY);
4887 ins->objectid = 0;
4888 ins->offset = 0;
4889
4890 space_info = __find_space_info(root->fs_info, data);
4891 if (!space_info) {
4892 printk(KERN_ERR "No space info for %llu\n", data);
4893 return -ENOSPC;
4894 }
4895
4896 /*
4897 * If the space info is for both data and metadata it means we have a
4898 * small filesystem and we can't use the clustering stuff.
4899 */
4900 if (btrfs_mixed_space_info(space_info))
4901 use_cluster = false;
4902
4903 if (orig_root->ref_cows || empty_size)
4904 allowed_chunk_alloc = 1;
4905
4906 if (data & BTRFS_BLOCK_GROUP_METADATA && use_cluster) {
4907 last_ptr = &root->fs_info->meta_alloc_cluster;
4908 if (!btrfs_test_opt(root, SSD))
4909 empty_cluster = 64 * 1024;
4910 }
4911
4912 if ((data & BTRFS_BLOCK_GROUP_DATA) && use_cluster &&
4913 btrfs_test_opt(root, SSD)) {
4914 last_ptr = &root->fs_info->data_alloc_cluster;
4915 }
4916
4917 if (last_ptr) {
4918 spin_lock(&last_ptr->lock);
4919 if (last_ptr->block_group)
4920 hint_byte = last_ptr->window_start;
4921 spin_unlock(&last_ptr->lock);
4922 }
4923
4924 search_start = max(search_start, first_logical_byte(root, 0));
4925 search_start = max(search_start, hint_byte);
4926
4927 if (!last_ptr)
4928 empty_cluster = 0;
4929
4930 if (search_start == hint_byte) {
4931 ideal_cache:
4932 block_group = btrfs_lookup_block_group(root->fs_info,
4933 search_start);
4934 /*
4935 * we don't want to use the block group if it doesn't match our
4936 * allocation bits, or if its not cached.
4937 *
4938 * However if we are re-searching with an ideal block group
4939 * picked out then we don't care that the block group is cached.
4940 */
4941 if (block_group && block_group_bits(block_group, data) &&
4942 (block_group->cached != BTRFS_CACHE_NO ||
4943 search_start == ideal_cache_offset)) {
4944 down_read(&space_info->groups_sem);
4945 if (list_empty(&block_group->list) ||
4946 block_group->ro) {
4947 /*
4948 * someone is removing this block group,
4949 * we can't jump into the have_block_group
4950 * target because our list pointers are not
4951 * valid
4952 */
4953 btrfs_put_block_group(block_group);
4954 up_read(&space_info->groups_sem);
4955 } else {
4956 index = get_block_group_index(block_group);
4957 goto have_block_group;
4958 }
4959 } else if (block_group) {
4960 btrfs_put_block_group(block_group);
4961 }
4962 }
4963 search:
4964 down_read(&space_info->groups_sem);
4965 list_for_each_entry(block_group, &space_info->block_groups[index],
4966 list) {
4967 u64 offset;
4968 int cached;
4969
4970 btrfs_get_block_group(block_group);
4971 search_start = block_group->key.objectid;
4972
4973 /*
4974 * this can happen if we end up cycling through all the
4975 * raid types, but we want to make sure we only allocate
4976 * for the proper type.
4977 */
4978 if (!block_group_bits(block_group, data)) {
4979 u64 extra = BTRFS_BLOCK_GROUP_DUP |
4980 BTRFS_BLOCK_GROUP_RAID1 |
4981 BTRFS_BLOCK_GROUP_RAID10;
4982
4983 /*
4984 * if they asked for extra copies and this block group
4985 * doesn't provide them, bail. This does allow us to
4986 * fill raid0 from raid1.
4987 */
4988 if ((data & extra) && !(block_group->flags & extra))
4989 goto loop;
4990 }
4991
4992 have_block_group:
4993 if (unlikely(block_group->cached == BTRFS_CACHE_NO)) {
4994 u64 free_percent;
4995
4996 ret = cache_block_group(block_group, trans,
4997 orig_root, 1);
4998 if (block_group->cached == BTRFS_CACHE_FINISHED)
4999 goto have_block_group;
5000
5001 free_percent = btrfs_block_group_used(&block_group->item);
5002 free_percent *= 100;
5003 free_percent = div64_u64(free_percent,
5004 block_group->key.offset);
5005 free_percent = 100 - free_percent;
5006 if (free_percent > ideal_cache_percent &&
5007 likely(!block_group->ro)) {
5008 ideal_cache_offset = block_group->key.objectid;
5009 ideal_cache_percent = free_percent;
5010 }
5011
5012 /*
5013 * The caching workers are limited to 2 threads, so we
5014 * can queue as much work as we care to.
5015 */
5016 if (loop > LOOP_FIND_IDEAL) {
5017 ret = cache_block_group(block_group, trans,
5018 orig_root, 0);
5019 BUG_ON(ret);
5020 }
5021 found_uncached_bg = true;
5022
5023 /*
5024 * If loop is set for cached only, try the next block
5025 * group.
5026 */
5027 if (loop == LOOP_FIND_IDEAL)
5028 goto loop;
5029 }
5030
5031 cached = block_group_cache_done(block_group);
5032 if (unlikely(!cached))
5033 found_uncached_bg = true;
5034
5035 if (unlikely(block_group->ro))
5036 goto loop;
5037
5038 spin_lock(&block_group->free_space_ctl->tree_lock);
5039 if (cached &&
5040 block_group->free_space_ctl->free_space <
5041 num_bytes + empty_size) {
5042 spin_unlock(&block_group->free_space_ctl->tree_lock);
5043 goto loop;
5044 }
5045 spin_unlock(&block_group->free_space_ctl->tree_lock);
5046
5047 /*
5048 * Ok we want to try and use the cluster allocator, so lets look
5049 * there, unless we are on LOOP_NO_EMPTY_SIZE, since we will
5050 * have tried the cluster allocator plenty of times at this
5051 * point and not have found anything, so we are likely way too
5052 * fragmented for the clustering stuff to find anything, so lets
5053 * just skip it and let the allocator find whatever block it can
5054 * find
5055 */
5056 if (last_ptr && loop < LOOP_NO_EMPTY_SIZE) {
5057 /*
5058 * the refill lock keeps out other
5059 * people trying to start a new cluster
5060 */
5061 spin_lock(&last_ptr->refill_lock);
5062 if (last_ptr->block_group &&
5063 (last_ptr->block_group->ro ||
5064 !block_group_bits(last_ptr->block_group, data))) {
5065 offset = 0;
5066 goto refill_cluster;
5067 }
5068
5069 offset = btrfs_alloc_from_cluster(block_group, last_ptr,
5070 num_bytes, search_start);
5071 if (offset) {
5072 /* we have a block, we're done */
5073 spin_unlock(&last_ptr->refill_lock);
5074 goto checks;
5075 }
5076
5077 spin_lock(&last_ptr->lock);
5078 /*
5079 * whoops, this cluster doesn't actually point to
5080 * this block group. Get a ref on the block
5081 * group is does point to and try again
5082 */
5083 if (!last_ptr_loop && last_ptr->block_group &&
5084 last_ptr->block_group != block_group &&
5085 index <=
5086 get_block_group_index(last_ptr->block_group)) {
5087
5088 btrfs_put_block_group(block_group);
5089 block_group = last_ptr->block_group;
5090 btrfs_get_block_group(block_group);
5091 spin_unlock(&last_ptr->lock);
5092 spin_unlock(&last_ptr->refill_lock);
5093
5094 last_ptr_loop = 1;
5095 search_start = block_group->key.objectid;
5096 /*
5097 * we know this block group is properly
5098 * in the list because
5099 * btrfs_remove_block_group, drops the
5100 * cluster before it removes the block
5101 * group from the list
5102 */
5103 goto have_block_group;
5104 }
5105 spin_unlock(&last_ptr->lock);
5106 refill_cluster:
5107 /*
5108 * this cluster didn't work out, free it and
5109 * start over
5110 */
5111 btrfs_return_cluster_to_free_space(NULL, last_ptr);
5112
5113 last_ptr_loop = 0;
5114
5115 /* allocate a cluster in this block group */
5116 ret = btrfs_find_space_cluster(trans, root,
5117 block_group, last_ptr,
5118 offset, num_bytes,
5119 empty_cluster + empty_size);
5120 if (ret == 0) {
5121 /*
5122 * now pull our allocation out of this
5123 * cluster
5124 */
5125 offset = btrfs_alloc_from_cluster(block_group,
5126 last_ptr, num_bytes,
5127 search_start);
5128 if (offset) {
5129 /* we found one, proceed */
5130 spin_unlock(&last_ptr->refill_lock);
5131 goto checks;
5132 }
5133 } else if (!cached && loop > LOOP_CACHING_NOWAIT
5134 && !failed_cluster_refill) {
5135 spin_unlock(&last_ptr->refill_lock);
5136
5137 failed_cluster_refill = true;
5138 wait_block_group_cache_progress(block_group,
5139 num_bytes + empty_cluster + empty_size);
5140 goto have_block_group;
5141 }
5142
5143 /*
5144 * at this point we either didn't find a cluster
5145 * or we weren't able to allocate a block from our
5146 * cluster. Free the cluster we've been trying
5147 * to use, and go to the next block group
5148 */
5149 btrfs_return_cluster_to_free_space(NULL, last_ptr);
5150 spin_unlock(&last_ptr->refill_lock);
5151 goto loop;
5152 }
5153
5154 offset = btrfs_find_space_for_alloc(block_group, search_start,
5155 num_bytes, empty_size);
5156 /*
5157 * If we didn't find a chunk, and we haven't failed on this
5158 * block group before, and this block group is in the middle of
5159 * caching and we are ok with waiting, then go ahead and wait
5160 * for progress to be made, and set failed_alloc to true.
5161 *
5162 * If failed_alloc is true then we've already waited on this
5163 * block group once and should move on to the next block group.
5164 */
5165 if (!offset && !failed_alloc && !cached &&
5166 loop > LOOP_CACHING_NOWAIT) {
5167 wait_block_group_cache_progress(block_group,
5168 num_bytes + empty_size);
5169 failed_alloc = true;
5170 goto have_block_group;
5171 } else if (!offset) {
5172 goto loop;
5173 }
5174 checks:
5175 search_start = stripe_align(root, offset);
5176 /* move on to the next group */
5177 if (search_start + num_bytes >= search_end) {
5178 btrfs_add_free_space(block_group, offset, num_bytes);
5179 goto loop;
5180 }
5181
5182 /* move on to the next group */
5183 if (search_start + num_bytes >
5184 block_group->key.objectid + block_group->key.offset) {
5185 btrfs_add_free_space(block_group, offset, num_bytes);
5186 goto loop;
5187 }
5188
5189 ins->objectid = search_start;
5190 ins->offset = num_bytes;
5191
5192 if (offset < search_start)
5193 btrfs_add_free_space(block_group, offset,
5194 search_start - offset);
5195 BUG_ON(offset > search_start);
5196
5197 ret = btrfs_update_reserved_bytes(block_group, num_bytes, 1,
5198 (data & BTRFS_BLOCK_GROUP_DATA));
5199 if (ret == -EAGAIN) {
5200 btrfs_add_free_space(block_group, offset, num_bytes);
5201 goto loop;
5202 }
5203
5204 /* we are all good, lets return */
5205 ins->objectid = search_start;
5206 ins->offset = num_bytes;
5207
5208 if (offset < search_start)
5209 btrfs_add_free_space(block_group, offset,
5210 search_start - offset);
5211 BUG_ON(offset > search_start);
5212 btrfs_put_block_group(block_group);
5213 break;
5214 loop:
5215 failed_cluster_refill = false;
5216 failed_alloc = false;
5217 BUG_ON(index != get_block_group_index(block_group));
5218 btrfs_put_block_group(block_group);
5219 }
5220 up_read(&space_info->groups_sem);
5221
5222 if (!ins->objectid && ++index < BTRFS_NR_RAID_TYPES)
5223 goto search;
5224
5225 /* LOOP_FIND_IDEAL, only search caching/cached bg's, and don't wait for
5226 * for them to make caching progress. Also
5227 * determine the best possible bg to cache
5228 * LOOP_CACHING_NOWAIT, search partially cached block groups, kicking
5229 * caching kthreads as we move along
5230 * LOOP_CACHING_WAIT, search everything, and wait if our bg is caching
5231 * LOOP_ALLOC_CHUNK, force a chunk allocation and try again
5232 * LOOP_NO_EMPTY_SIZE, set empty_size and empty_cluster to 0 and try
5233 * again
5234 */
5235 if (!ins->objectid && loop < LOOP_NO_EMPTY_SIZE) {
5236 index = 0;
5237 if (loop == LOOP_FIND_IDEAL && found_uncached_bg) {
5238 found_uncached_bg = false;
5239 loop++;
5240 if (!ideal_cache_percent)
5241 goto search;
5242
5243 /*
5244 * 1 of the following 2 things have happened so far
5245 *
5246 * 1) We found an ideal block group for caching that
5247 * is mostly full and will cache quickly, so we might
5248 * as well wait for it.
5249 *
5250 * 2) We searched for cached only and we didn't find
5251 * anything, and we didn't start any caching kthreads
5252 * either, so chances are we will loop through and
5253 * start a couple caching kthreads, and then come back
5254 * around and just wait for them. This will be slower
5255 * because we will have 2 caching kthreads reading at
5256 * the same time when we could have just started one
5257 * and waited for it to get far enough to give us an
5258 * allocation, so go ahead and go to the wait caching
5259 * loop.
5260 */
5261 loop = LOOP_CACHING_WAIT;
5262 search_start = ideal_cache_offset;
5263 ideal_cache_percent = 0;
5264 goto ideal_cache;
5265 } else if (loop == LOOP_FIND_IDEAL) {
5266 /*
5267 * Didn't find a uncached bg, wait on anything we find
5268 * next.
5269 */
5270 loop = LOOP_CACHING_WAIT;
5271 goto search;
5272 }
5273
5274 loop++;
5275
5276 if (loop == LOOP_ALLOC_CHUNK) {
5277 if (allowed_chunk_alloc) {
5278 ret = do_chunk_alloc(trans, root, num_bytes +
5279 2 * 1024 * 1024, data,
5280 CHUNK_ALLOC_LIMITED);
5281 allowed_chunk_alloc = 0;
5282 if (ret == 1)
5283 done_chunk_alloc = 1;
5284 } else if (!done_chunk_alloc &&
5285 space_info->force_alloc ==
5286 CHUNK_ALLOC_NO_FORCE) {
5287 space_info->force_alloc = CHUNK_ALLOC_LIMITED;
5288 }
5289
5290 /*
5291 * We didn't allocate a chunk, go ahead and drop the
5292 * empty size and loop again.
5293 */
5294 if (!done_chunk_alloc)
5295 loop = LOOP_NO_EMPTY_SIZE;
5296 }
5297
5298 if (loop == LOOP_NO_EMPTY_SIZE) {
5299 empty_size = 0;
5300 empty_cluster = 0;
5301 }
5302
5303 goto search;
5304 } else if (!ins->objectid) {
5305 ret = -ENOSPC;
5306 } else if (ins->objectid) {
5307 ret = 0;
5308 }
5309
5310 return ret;
5311 }
5312
5313 static void dump_space_info(struct btrfs_space_info *info, u64 bytes,
5314 int dump_block_groups)
5315 {
5316 struct btrfs_block_group_cache *cache;
5317 int index = 0;
5318
5319 spin_lock(&info->lock);
5320 printk(KERN_INFO "space_info has %llu free, is %sfull\n",
5321 (unsigned long long)(info->total_bytes - info->bytes_used -
5322 info->bytes_pinned - info->bytes_reserved -
5323 info->bytes_readonly),
5324 (info->full) ? "" : "not ");
5325 printk(KERN_INFO "space_info total=%llu, used=%llu, pinned=%llu, "
5326 "reserved=%llu, may_use=%llu, readonly=%llu\n",
5327 (unsigned long long)info->total_bytes,
5328 (unsigned long long)info->bytes_used,
5329 (unsigned long long)info->bytes_pinned,
5330 (unsigned long long)info->bytes_reserved,
5331 (unsigned long long)info->bytes_may_use,
5332 (unsigned long long)info->bytes_readonly);
5333 spin_unlock(&info->lock);
5334
5335 if (!dump_block_groups)
5336 return;
5337
5338 down_read(&info->groups_sem);
5339 again:
5340 list_for_each_entry(cache, &info->block_groups[index], list) {
5341 spin_lock(&cache->lock);
5342 printk(KERN_INFO "block group %llu has %llu bytes, %llu used "
5343 "%llu pinned %llu reserved\n",
5344 (unsigned long long)cache->key.objectid,
5345 (unsigned long long)cache->key.offset,
5346 (unsigned long long)btrfs_block_group_used(&cache->item),
5347 (unsigned long long)cache->pinned,
5348 (unsigned long long)cache->reserved);
5349 btrfs_dump_free_space(cache, bytes);
5350 spin_unlock(&cache->lock);
5351 }
5352 if (++index < BTRFS_NR_RAID_TYPES)
5353 goto again;
5354 up_read(&info->groups_sem);
5355 }
5356
5357 int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
5358 struct btrfs_root *root,
5359 u64 num_bytes, u64 min_alloc_size,
5360 u64 empty_size, u64 hint_byte,
5361 u64 search_end, struct btrfs_key *ins,
5362 u64 data)
5363 {
5364 int ret;
5365 u64 search_start = 0;
5366
5367 data = btrfs_get_alloc_profile(root, data);
5368 again:
5369 /*
5370 * the only place that sets empty_size is btrfs_realloc_node, which
5371 * is not called recursively on allocations
5372 */
5373 if (empty_size || root->ref_cows)
5374 ret = do_chunk_alloc(trans, root->fs_info->extent_root,
5375 num_bytes + 2 * 1024 * 1024, data,
5376 CHUNK_ALLOC_NO_FORCE);
5377
5378 WARN_ON(num_bytes < root->sectorsize);
5379 ret = find_free_extent(trans, root, num_bytes, empty_size,
5380 search_start, search_end, hint_byte,
5381 ins, data);
5382
5383 if (ret == -ENOSPC && num_bytes > min_alloc_size) {
5384 num_bytes = num_bytes >> 1;
5385 num_bytes = num_bytes & ~(root->sectorsize - 1);
5386 num_bytes = max(num_bytes, min_alloc_size);
5387 do_chunk_alloc(trans, root->fs_info->extent_root,
5388 num_bytes, data, CHUNK_ALLOC_FORCE);
5389 goto again;
5390 }
5391 if (ret == -ENOSPC && btrfs_test_opt(root, ENOSPC_DEBUG)) {
5392 struct btrfs_space_info *sinfo;
5393
5394 sinfo = __find_space_info(root->fs_info, data);
5395 printk(KERN_ERR "btrfs allocation failed flags %llu, "
5396 "wanted %llu\n", (unsigned long long)data,
5397 (unsigned long long)num_bytes);
5398 dump_space_info(sinfo, num_bytes, 1);
5399 }
5400
5401 trace_btrfs_reserved_extent_alloc(root, ins->objectid, ins->offset);
5402
5403 return ret;
5404 }
5405
5406 int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len)
5407 {
5408 struct btrfs_block_group_cache *cache;
5409 int ret = 0;
5410
5411 cache = btrfs_lookup_block_group(root->fs_info, start);
5412 if (!cache) {
5413 printk(KERN_ERR "Unable to find block group for %llu\n",
5414 (unsigned long long)start);
5415 return -ENOSPC;
5416 }
5417
5418 if (btrfs_test_opt(root, DISCARD))
5419 ret = btrfs_discard_extent(root, start, len, NULL);
5420
5421 btrfs_add_free_space(cache, start, len);
5422 btrfs_update_reserved_bytes(cache, len, 0, 1);
5423 btrfs_put_block_group(cache);
5424
5425 trace_btrfs_reserved_extent_free(root, start, len);
5426
5427 return ret;
5428 }
5429
5430 static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
5431 struct btrfs_root *root,
5432 u64 parent, u64 root_objectid,
5433 u64 flags, u64 owner, u64 offset,
5434 struct btrfs_key *ins, int ref_mod)
5435 {
5436 int ret;
5437 struct btrfs_fs_info *fs_info = root->fs_info;
5438 struct btrfs_extent_item *extent_item;
5439 struct btrfs_extent_inline_ref *iref;
5440 struct btrfs_path *path;
5441 struct extent_buffer *leaf;
5442 int type;
5443 u32 size;
5444
5445 if (parent > 0)
5446 type = BTRFS_SHARED_DATA_REF_KEY;
5447 else
5448 type = BTRFS_EXTENT_DATA_REF_KEY;
5449
5450 size = sizeof(*extent_item) + btrfs_extent_inline_ref_size(type);
5451
5452 path = btrfs_alloc_path();
5453 if (!path)
5454 return -ENOMEM;
5455
5456 path->leave_spinning = 1;
5457 ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path,
5458 ins, size);
5459 BUG_ON(ret);
5460
5461 leaf = path->nodes[0];
5462 extent_item = btrfs_item_ptr(leaf, path->slots[0],
5463 struct btrfs_extent_item);
5464 btrfs_set_extent_refs(leaf, extent_item, ref_mod);
5465 btrfs_set_extent_generation(leaf, extent_item, trans->transid);
5466 btrfs_set_extent_flags(leaf, extent_item,
5467 flags | BTRFS_EXTENT_FLAG_DATA);
5468
5469 iref = (struct btrfs_extent_inline_ref *)(extent_item + 1);
5470 btrfs_set_extent_inline_ref_type(leaf, iref, type);
5471 if (parent > 0) {
5472 struct btrfs_shared_data_ref *ref;
5473 ref = (struct btrfs_shared_data_ref *)(iref + 1);
5474 btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
5475 btrfs_set_shared_data_ref_count(leaf, ref, ref_mod);
5476 } else {
5477 struct btrfs_extent_data_ref *ref;
5478 ref = (struct btrfs_extent_data_ref *)(&iref->offset);
5479 btrfs_set_extent_data_ref_root(leaf, ref, root_objectid);
5480 btrfs_set_extent_data_ref_objectid(leaf, ref, owner);
5481 btrfs_set_extent_data_ref_offset(leaf, ref, offset);
5482 btrfs_set_extent_data_ref_count(leaf, ref, ref_mod);
5483 }
5484
5485 btrfs_mark_buffer_dirty(path->nodes[0]);
5486 btrfs_free_path(path);
5487
5488 ret = update_block_group(trans, root, ins->objectid, ins->offset, 1);
5489 if (ret) {
5490 printk(KERN_ERR "btrfs update block group failed for %llu "
5491 "%llu\n", (unsigned long long)ins->objectid,
5492 (unsigned long long)ins->offset);
5493 BUG();
5494 }
5495 return ret;
5496 }
5497
5498 static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
5499 struct btrfs_root *root,
5500 u64 parent, u64 root_objectid,
5501 u64 flags, struct btrfs_disk_key *key,
5502 int level, struct btrfs_key *ins)
5503 {
5504 int ret;
5505 struct btrfs_fs_info *fs_info = root->fs_info;
5506 struct btrfs_extent_item *extent_item;
5507 struct btrfs_tree_block_info *block_info;
5508 struct btrfs_extent_inline_ref *iref;
5509 struct btrfs_path *path;
5510 struct extent_buffer *leaf;
5511 u32 size = sizeof(*extent_item) + sizeof(*block_info) + sizeof(*iref);
5512
5513 path = btrfs_alloc_path();
5514 if (!path)
5515 return -ENOMEM;
5516
5517 path->leave_spinning = 1;
5518 ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path,
5519 ins, size);
5520 BUG_ON(ret);
5521
5522 leaf = path->nodes[0];
5523 extent_item = btrfs_item_ptr(leaf, path->slots[0],
5524 struct btrfs_extent_item);
5525 btrfs_set_extent_refs(leaf, extent_item, 1);
5526 btrfs_set_extent_generation(leaf, extent_item, trans->transid);
5527 btrfs_set_extent_flags(leaf, extent_item,
5528 flags | BTRFS_EXTENT_FLAG_TREE_BLOCK);
5529 block_info = (struct btrfs_tree_block_info *)(extent_item + 1);
5530
5531 btrfs_set_tree_block_key(leaf, block_info, key);
5532 btrfs_set_tree_block_level(leaf, block_info, level);
5533
5534 iref = (struct btrfs_extent_inline_ref *)(block_info + 1);
5535 if (parent > 0) {
5536 BUG_ON(!(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF));
5537 btrfs_set_extent_inline_ref_type(leaf, iref,
5538 BTRFS_SHARED_BLOCK_REF_KEY);
5539 btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
5540 } else {
5541 btrfs_set_extent_inline_ref_type(leaf, iref,
5542 BTRFS_TREE_BLOCK_REF_KEY);
5543 btrfs_set_extent_inline_ref_offset(leaf, iref, root_objectid);
5544 }
5545
5546 btrfs_mark_buffer_dirty(leaf);
5547 btrfs_free_path(path);
5548
5549 ret = update_block_group(trans, root, ins->objectid, ins->offset, 1);
5550 if (ret) {
5551 printk(KERN_ERR "btrfs update block group failed for %llu "
5552 "%llu\n", (unsigned long long)ins->objectid,
5553 (unsigned long long)ins->offset);
5554 BUG();
5555 }
5556 return ret;
5557 }
5558
5559 int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
5560 struct btrfs_root *root,
5561 u64 root_objectid, u64 owner,
5562 u64 offset, struct btrfs_key *ins)
5563 {
5564 int ret;
5565
5566 BUG_ON(root_objectid == BTRFS_TREE_LOG_OBJECTID);
5567
5568 ret = btrfs_add_delayed_data_ref(trans, ins->objectid, ins->offset,
5569 0, root_objectid, owner, offset,
5570 BTRFS_ADD_DELAYED_EXTENT, NULL);
5571 return ret;
5572 }
5573
5574 /*
5575 * this is used by the tree logging recovery code. It records that
5576 * an extent has been allocated and makes sure to clear the free
5577 * space cache bits as well
5578 */
5579 int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans,
5580 struct btrfs_root *root,
5581 u64 root_objectid, u64 owner, u64 offset,
5582 struct btrfs_key *ins)
5583 {
5584 int ret;
5585 struct btrfs_block_group_cache *block_group;
5586 struct btrfs_caching_control *caching_ctl;
5587 u64 start = ins->objectid;
5588 u64 num_bytes = ins->offset;
5589
5590 block_group = btrfs_lookup_block_group(root->fs_info, ins->objectid);
5591 cache_block_group(block_group, trans, NULL, 0);
5592 caching_ctl = get_caching_control(block_group);
5593
5594 if (!caching_ctl) {
5595 BUG_ON(!block_group_cache_done(block_group));
5596 ret = btrfs_remove_free_space(block_group, start, num_bytes);
5597 BUG_ON(ret);
5598 } else {
5599 mutex_lock(&caching_ctl->mutex);
5600
5601 if (start >= caching_ctl->progress) {
5602 ret = add_excluded_extent(root, start, num_bytes);
5603 BUG_ON(ret);
5604 } else if (start + num_bytes <= caching_ctl->progress) {
5605 ret = btrfs_remove_free_space(block_group,
5606 start, num_bytes);
5607 BUG_ON(ret);
5608 } else {
5609 num_bytes = caching_ctl->progress - start;
5610 ret = btrfs_remove_free_space(block_group,
5611 start, num_bytes);
5612 BUG_ON(ret);
5613
5614 start = caching_ctl->progress;
5615 num_bytes = ins->objectid + ins->offset -
5616 caching_ctl->progress;
5617 ret = add_excluded_extent(root, start, num_bytes);
5618 BUG_ON(ret);
5619 }
5620
5621 mutex_unlock(&caching_ctl->mutex);
5622 put_caching_control(caching_ctl);
5623 }
5624
5625 ret = btrfs_update_reserved_bytes(block_group, ins->offset, 1, 1);
5626 BUG_ON(ret);
5627 btrfs_put_block_group(block_group);
5628 ret = alloc_reserved_file_extent(trans, root, 0, root_objectid,
5629 0, owner, offset, ins, 1);
5630 return ret;
5631 }
5632
5633 struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans,
5634 struct btrfs_root *root,
5635 u64 bytenr, u32 blocksize,
5636 int level)
5637 {
5638 struct extent_buffer *buf;
5639
5640 buf = btrfs_find_create_tree_block(root, bytenr, blocksize);
5641 if (!buf)
5642 return ERR_PTR(-ENOMEM);
5643 btrfs_set_header_generation(buf, trans->transid);
5644 btrfs_set_buffer_lockdep_class(root->root_key.objectid, buf, level);
5645 btrfs_tree_lock(buf);
5646 clean_tree_block(trans, root, buf);
5647
5648 btrfs_set_lock_blocking(buf);
5649 btrfs_set_buffer_uptodate(buf);
5650
5651 if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) {
5652 /*
5653 * we allow two log transactions at a time, use different
5654 * EXENT bit to differentiate dirty pages.
5655 */
5656 if (root->log_transid % 2 == 0)
5657 set_extent_dirty(&root->dirty_log_pages, buf->start,
5658 buf->start + buf->len - 1, GFP_NOFS);
5659 else
5660 set_extent_new(&root->dirty_log_pages, buf->start,
5661 buf->start + buf->len - 1, GFP_NOFS);
5662 } else {
5663 set_extent_dirty(&trans->transaction->dirty_pages, buf->start,
5664 buf->start + buf->len - 1, GFP_NOFS);
5665 }
5666 trans->blocks_used++;
5667 /* this returns a buffer locked for blocking */
5668 return buf;
5669 }
5670
5671 static struct btrfs_block_rsv *
5672 use_block_rsv(struct btrfs_trans_handle *trans,
5673 struct btrfs_root *root, u32 blocksize)
5674 {
5675 struct btrfs_block_rsv *block_rsv;
5676 struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv;
5677 int ret;
5678
5679 block_rsv = get_block_rsv(trans, root);
5680
5681 if (block_rsv->size == 0) {
5682 ret = reserve_metadata_bytes(trans, root, block_rsv,
5683 blocksize, 0);
5684 /*
5685 * If we couldn't reserve metadata bytes try and use some from
5686 * the global reserve.
5687 */
5688 if (ret && block_rsv != global_rsv) {
5689 ret = block_rsv_use_bytes(global_rsv, blocksize);
5690 if (!ret)
5691 return global_rsv;
5692 return ERR_PTR(ret);
5693 } else if (ret) {
5694 return ERR_PTR(ret);
5695 }
5696 return block_rsv;
5697 }
5698
5699 ret = block_rsv_use_bytes(block_rsv, blocksize);
5700 if (!ret)
5701 return block_rsv;
5702 if (ret) {
5703 WARN_ON(1);
5704 ret = reserve_metadata_bytes(trans, root, block_rsv, blocksize,
5705 0);
5706 if (!ret) {
5707 spin_lock(&block_rsv->lock);
5708 block_rsv->size += blocksize;
5709 spin_unlock(&block_rsv->lock);
5710 return block_rsv;
5711 } else if (ret && block_rsv != global_rsv) {
5712 ret = block_rsv_use_bytes(global_rsv, blocksize);
5713 if (!ret)
5714 return global_rsv;
5715 }
5716 }
5717
5718 return ERR_PTR(-ENOSPC);
5719 }
5720
5721 static void unuse_block_rsv(struct btrfs_block_rsv *block_rsv, u32 blocksize)
5722 {
5723 block_rsv_add_bytes(block_rsv, blocksize, 0);
5724 block_rsv_release_bytes(block_rsv, NULL, 0);
5725 }
5726
5727 /*
5728 * finds a free extent and does all the dirty work required for allocation
5729 * returns the key for the extent through ins, and a tree buffer for
5730 * the first block of the extent through buf.
5731 *
5732 * returns the tree buffer or NULL.
5733 */
5734 struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
5735 struct btrfs_root *root, u32 blocksize,
5736 u64 parent, u64 root_objectid,
5737 struct btrfs_disk_key *key, int level,
5738 u64 hint, u64 empty_size)
5739 {
5740 struct btrfs_key ins;
5741 struct btrfs_block_rsv *block_rsv;
5742 struct extent_buffer *buf;
5743 u64 flags = 0;
5744 int ret;
5745
5746
5747 block_rsv = use_block_rsv(trans, root, blocksize);
5748 if (IS_ERR(block_rsv))
5749 return ERR_CAST(block_rsv);
5750
5751 ret = btrfs_reserve_extent(trans, root, blocksize, blocksize,
5752 empty_size, hint, (u64)-1, &ins, 0);
5753 if (ret) {
5754 unuse_block_rsv(block_rsv, blocksize);
5755 return ERR_PTR(ret);
5756 }
5757
5758 buf = btrfs_init_new_buffer(trans, root, ins.objectid,
5759 blocksize, level);
5760 BUG_ON(IS_ERR(buf));
5761
5762 if (root_objectid == BTRFS_TREE_RELOC_OBJECTID) {
5763 if (parent == 0)
5764 parent = ins.objectid;
5765 flags |= BTRFS_BLOCK_FLAG_FULL_BACKREF;
5766 } else
5767 BUG_ON(parent > 0);
5768
5769 if (root_objectid != BTRFS_TREE_LOG_OBJECTID) {
5770 struct btrfs_delayed_extent_op *extent_op;
5771 extent_op = kmalloc(sizeof(*extent_op), GFP_NOFS);
5772 BUG_ON(!extent_op);
5773 if (key)
5774 memcpy(&extent_op->key, key, sizeof(extent_op->key));
5775 else
5776 memset(&extent_op->key, 0, sizeof(extent_op->key));
5777 extent_op->flags_to_set = flags;
5778 extent_op->update_key = 1;
5779 extent_op->update_flags = 1;
5780 extent_op->is_data = 0;
5781
5782 ret = btrfs_add_delayed_tree_ref(trans, ins.objectid,
5783 ins.offset, parent, root_objectid,
5784 level, BTRFS_ADD_DELAYED_EXTENT,
5785 extent_op);
5786 BUG_ON(ret);
5787 }
5788 return buf;
5789 }
5790
5791 struct walk_control {
5792 u64 refs[BTRFS_MAX_LEVEL];
5793 u64 flags[BTRFS_MAX_LEVEL];
5794 struct btrfs_key update_progress;
5795 int stage;
5796 int level;
5797 int shared_level;
5798 int update_ref;
5799 int keep_locks;
5800 int reada_slot;
5801 int reada_count;
5802 };
5803
5804 #define DROP_REFERENCE 1
5805 #define UPDATE_BACKREF 2
5806
5807 static noinline void reada_walk_down(struct btrfs_trans_handle *trans,
5808 struct btrfs_root *root,
5809 struct walk_control *wc,
5810 struct btrfs_path *path)
5811 {
5812 u64 bytenr;
5813 u64 generation;
5814 u64 refs;
5815 u64 flags;
5816 u32 nritems;
5817 u32 blocksize;
5818 struct btrfs_key key;
5819 struct extent_buffer *eb;
5820 int ret;
5821 int slot;
5822 int nread = 0;
5823
5824 if (path->slots[wc->level] < wc->reada_slot) {
5825 wc->reada_count = wc->reada_count * 2 / 3;
5826 wc->reada_count = max(wc->reada_count, 2);
5827 } else {
5828 wc->reada_count = wc->reada_count * 3 / 2;
5829 wc->reada_count = min_t(int, wc->reada_count,
5830 BTRFS_NODEPTRS_PER_BLOCK(root));
5831 }
5832
5833 eb = path->nodes[wc->level];
5834 nritems = btrfs_header_nritems(eb);
5835 blocksize = btrfs_level_size(root, wc->level - 1);
5836
5837 for (slot = path->slots[wc->level]; slot < nritems; slot++) {
5838 if (nread >= wc->reada_count)
5839 break;
5840
5841 cond_resched();
5842 bytenr = btrfs_node_blockptr(eb, slot);
5843 generation = btrfs_node_ptr_generation(eb, slot);
5844
5845 if (slot == path->slots[wc->level])
5846 goto reada;
5847
5848 if (wc->stage == UPDATE_BACKREF &&
5849 generation <= root->root_key.offset)
5850 continue;
5851
5852 /* We don't lock the tree block, it's OK to be racy here */
5853 ret = btrfs_lookup_extent_info(trans, root, bytenr, blocksize,
5854 &refs, &flags);
5855 BUG_ON(ret);
5856 BUG_ON(refs == 0);
5857
5858 if (wc->stage == DROP_REFERENCE) {
5859 if (refs == 1)
5860 goto reada;
5861
5862 if (wc->level == 1 &&
5863 (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF))
5864 continue;
5865 if (!wc->update_ref ||
5866 generation <= root->root_key.offset)
5867 continue;
5868 btrfs_node_key_to_cpu(eb, &key, slot);
5869 ret = btrfs_comp_cpu_keys(&key,
5870 &wc->update_progress);
5871 if (ret < 0)
5872 continue;
5873 } else {
5874 if (wc->level == 1 &&
5875 (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF))
5876 continue;
5877 }
5878 reada:
5879 ret = readahead_tree_block(root, bytenr, blocksize,
5880 generation);
5881 if (ret)
5882 break;
5883 nread++;
5884 }
5885 wc->reada_slot = slot;
5886 }
5887
5888 /*
5889 * hepler to process tree block while walking down the tree.
5890 *
5891 * when wc->stage == UPDATE_BACKREF, this function updates
5892 * back refs for pointers in the block.
5893 *
5894 * NOTE: return value 1 means we should stop walking down.
5895 */
5896 static noinline int walk_down_proc(struct btrfs_trans_handle *trans,
5897 struct btrfs_root *root,
5898 struct btrfs_path *path,
5899 struct walk_control *wc, int lookup_info)
5900 {
5901 int level = wc->level;
5902 struct extent_buffer *eb = path->nodes[level];
5903 u64 flag = BTRFS_BLOCK_FLAG_FULL_BACKREF;
5904 int ret;
5905
5906 if (wc->stage == UPDATE_BACKREF &&
5907 btrfs_header_owner(eb) != root->root_key.objectid)
5908 return 1;
5909
5910 /*
5911 * when reference count of tree block is 1, it won't increase
5912 * again. once full backref flag is set, we never clear it.
5913 */
5914 if (lookup_info &&
5915 ((wc->stage == DROP_REFERENCE && wc->refs[level] != 1) ||
5916 (wc->stage == UPDATE_BACKREF && !(wc->flags[level] & flag)))) {
5917 BUG_ON(!path->locks[level]);
5918 ret = btrfs_lookup_extent_info(trans, root,
5919 eb->start, eb->len,
5920 &wc->refs[level],
5921 &wc->flags[level]);
5922 BUG_ON(ret);
5923 BUG_ON(wc->refs[level] == 0);
5924 }
5925
5926 if (wc->stage == DROP_REFERENCE) {
5927 if (wc->refs[level] > 1)
5928 return 1;
5929
5930 if (path->locks[level] && !wc->keep_locks) {
5931 btrfs_tree_unlock_rw(eb, path->locks[level]);
5932 path->locks[level] = 0;
5933 }
5934 return 0;
5935 }
5936
5937 /* wc->stage == UPDATE_BACKREF */
5938 if (!(wc->flags[level] & flag)) {
5939 BUG_ON(!path->locks[level]);
5940 ret = btrfs_inc_ref(trans, root, eb, 1);
5941 BUG_ON(ret);
5942 ret = btrfs_dec_ref(trans, root, eb, 0);
5943 BUG_ON(ret);
5944 ret = btrfs_set_disk_extent_flags(trans, root, eb->start,
5945 eb->len, flag, 0);
5946 BUG_ON(ret);
5947 wc->flags[level] |= flag;
5948 }
5949
5950 /*
5951 * the block is shared by multiple trees, so it's not good to
5952 * keep the tree lock
5953 */
5954 if (path->locks[level] && level > 0) {
5955 btrfs_tree_unlock_rw(eb, path->locks[level]);
5956 path->locks[level] = 0;
5957 }
5958 return 0;
5959 }
5960
5961 /*
5962 * hepler to process tree block pointer.
5963 *
5964 * when wc->stage == DROP_REFERENCE, this function checks
5965 * reference count of the block pointed to. if the block
5966 * is shared and we need update back refs for the subtree
5967 * rooted at the block, this function changes wc->stage to
5968 * UPDATE_BACKREF. if the block is shared and there is no
5969 * need to update back, this function drops the reference
5970 * to the block.
5971 *
5972 * NOTE: return value 1 means we should stop walking down.
5973 */
5974 static noinline int do_walk_down(struct btrfs_trans_handle *trans,
5975 struct btrfs_root *root,
5976 struct btrfs_path *path,
5977 struct walk_control *wc, int *lookup_info)
5978 {
5979 u64 bytenr;
5980 u64 generation;
5981 u64 parent;
5982 u32 blocksize;
5983 struct btrfs_key key;
5984 struct extent_buffer *next;
5985 int level = wc->level;
5986 int reada = 0;
5987 int ret = 0;
5988
5989 generation = btrfs_node_ptr_generation(path->nodes[level],
5990 path->slots[level]);
5991 /*
5992 * if the lower level block was created before the snapshot
5993 * was created, we know there is no need to update back refs
5994 * for the subtree
5995 */
5996 if (wc->stage == UPDATE_BACKREF &&
5997 generation <= root->root_key.offset) {
5998 *lookup_info = 1;
5999 return 1;
6000 }
6001
6002 bytenr = btrfs_node_blockptr(path->nodes[level], path->slots[level]);
6003 blocksize = btrfs_level_size(root, level - 1);
6004
6005 next = btrfs_find_tree_block(root, bytenr, blocksize);
6006 if (!next) {
6007 next = btrfs_find_create_tree_block(root, bytenr, blocksize);
6008 if (!next)
6009 return -ENOMEM;
6010 reada = 1;
6011 }
6012 btrfs_tree_lock(next);
6013 btrfs_set_lock_blocking(next);
6014
6015 ret = btrfs_lookup_extent_info(trans, root, bytenr, blocksize,
6016 &wc->refs[level - 1],
6017 &wc->flags[level - 1]);
6018 BUG_ON(ret);
6019 BUG_ON(wc->refs[level - 1] == 0);
6020 *lookup_info = 0;
6021
6022 if (wc->stage == DROP_REFERENCE) {
6023 if (wc->refs[level - 1] > 1) {
6024 if (level == 1 &&
6025 (wc->flags[0] & BTRFS_BLOCK_FLAG_FULL_BACKREF))
6026 goto skip;
6027
6028 if (!wc->update_ref ||
6029 generation <= root->root_key.offset)
6030 goto skip;
6031
6032 btrfs_node_key_to_cpu(path->nodes[level], &key,
6033 path->slots[level]);
6034 ret = btrfs_comp_cpu_keys(&key, &wc->update_progress);
6035 if (ret < 0)
6036 goto skip;
6037
6038 wc->stage = UPDATE_BACKREF;
6039 wc->shared_level = level - 1;
6040 }
6041 } else {
6042 if (level == 1 &&
6043 (wc->flags[0] & BTRFS_BLOCK_FLAG_FULL_BACKREF))
6044 goto skip;
6045 }
6046
6047 if (!btrfs_buffer_uptodate(next, generation)) {
6048 btrfs_tree_unlock(next);
6049 free_extent_buffer(next);
6050 next = NULL;
6051 *lookup_info = 1;
6052 }
6053
6054 if (!next) {
6055 if (reada && level == 1)
6056 reada_walk_down(trans, root, wc, path);
6057 next = read_tree_block(root, bytenr, blocksize, generation);
6058 if (!next)
6059 return -EIO;
6060 btrfs_tree_lock(next);
6061 btrfs_set_lock_blocking(next);
6062 }
6063
6064 level--;
6065 BUG_ON(level != btrfs_header_level(next));
6066 path->nodes[level] = next;
6067 path->slots[level] = 0;
6068 path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING;
6069 wc->level = level;
6070 if (wc->level == 1)
6071 wc->reada_slot = 0;
6072 return 0;
6073 skip:
6074 wc->refs[level - 1] = 0;
6075 wc->flags[level - 1] = 0;
6076 if (wc->stage == DROP_REFERENCE) {
6077 if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF) {
6078 parent = path->nodes[level]->start;
6079 } else {
6080 BUG_ON(root->root_key.objectid !=
6081 btrfs_header_owner(path->nodes[level]));
6082 parent = 0;
6083 }
6084
6085 ret = btrfs_free_extent(trans, root, bytenr, blocksize, parent,
6086 root->root_key.objectid, level - 1, 0);
6087 BUG_ON(ret);
6088 }
6089 btrfs_tree_unlock(next);
6090 free_extent_buffer(next);
6091 *lookup_info = 1;
6092 return 1;
6093 }
6094
6095 /*
6096 * hepler to process tree block while walking up the tree.
6097 *
6098 * when wc->stage == DROP_REFERENCE, this function drops
6099 * reference count on the block.
6100 *
6101 * when wc->stage == UPDATE_BACKREF, this function changes
6102 * wc->stage back to DROP_REFERENCE if we changed wc->stage
6103 * to UPDATE_BACKREF previously while processing the block.
6104 *
6105 * NOTE: return value 1 means we should stop walking up.
6106 */
6107 static noinline int walk_up_proc(struct btrfs_trans_handle *trans,
6108 struct btrfs_root *root,
6109 struct btrfs_path *path,
6110 struct walk_control *wc)
6111 {
6112 int ret;
6113 int level = wc->level;
6114 struct extent_buffer *eb = path->nodes[level];
6115 u64 parent = 0;
6116
6117 if (wc->stage == UPDATE_BACKREF) {
6118 BUG_ON(wc->shared_level < level);
6119 if (level < wc->shared_level)
6120 goto out;
6121
6122 ret = find_next_key(path, level + 1, &wc->update_progress);
6123 if (ret > 0)
6124 wc->update_ref = 0;
6125
6126 wc->stage = DROP_REFERENCE;
6127 wc->shared_level = -1;
6128 path->slots[level] = 0;
6129
6130 /*
6131 * check reference count again if the block isn't locked.
6132 * we should start walking down the tree again if reference
6133 * count is one.
6134 */
6135 if (!path->locks[level]) {
6136 BUG_ON(level == 0);
6137 btrfs_tree_lock(eb);
6138 btrfs_set_lock_blocking(eb);
6139 path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING;
6140
6141 ret = btrfs_lookup_extent_info(trans, root,
6142 eb->start, eb->len,
6143 &wc->refs[level],
6144 &wc->flags[level]);
6145 BUG_ON(ret);
6146 BUG_ON(wc->refs[level] == 0);
6147 if (wc->refs[level] == 1) {
6148 btrfs_tree_unlock_rw(eb, path->locks[level]);
6149 return 1;
6150 }
6151 }
6152 }
6153
6154 /* wc->stage == DROP_REFERENCE */
6155 BUG_ON(wc->refs[level] > 1 && !path->locks[level]);
6156
6157 if (wc->refs[level] == 1) {
6158 if (level == 0) {
6159 if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF)
6160 ret = btrfs_dec_ref(trans, root, eb, 1);
6161 else
6162 ret = btrfs_dec_ref(trans, root, eb, 0);
6163 BUG_ON(ret);
6164 }
6165 /* make block locked assertion in clean_tree_block happy */
6166 if (!path->locks[level] &&
6167 btrfs_header_generation(eb) == trans->transid) {
6168 btrfs_tree_lock(eb);
6169 btrfs_set_lock_blocking(eb);
6170 path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING;
6171 }
6172 clean_tree_block(trans, root, eb);
6173 }
6174
6175 if (eb == root->node) {
6176 if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF)
6177 parent = eb->start;
6178 else
6179 BUG_ON(root->root_key.objectid !=
6180 btrfs_header_owner(eb));
6181 } else {
6182 if (wc->flags[level + 1] & BTRFS_BLOCK_FLAG_FULL_BACKREF)
6183 parent = path->nodes[level + 1]->start;
6184 else
6185 BUG_ON(root->root_key.objectid !=
6186 btrfs_header_owner(path->nodes[level + 1]));
6187 }
6188
6189 btrfs_free_tree_block(trans, root, eb, parent, wc->refs[level] == 1);
6190 out:
6191 wc->refs[level] = 0;
6192 wc->flags[level] = 0;
6193 return 0;
6194 }
6195
6196 static noinline int walk_down_tree(struct btrfs_trans_handle *trans,
6197 struct btrfs_root *root,
6198 struct btrfs_path *path,
6199 struct walk_control *wc)
6200 {
6201 int level = wc->level;
6202 int lookup_info = 1;
6203 int ret;
6204
6205 while (level >= 0) {
6206 ret = walk_down_proc(trans, root, path, wc, lookup_info);
6207 if (ret > 0)
6208 break;
6209
6210 if (level == 0)
6211 break;
6212
6213 if (path->slots[level] >=
6214 btrfs_header_nritems(path->nodes[level]))
6215 break;
6216
6217 ret = do_walk_down(trans, root, path, wc, &lookup_info);
6218 if (ret > 0) {
6219 path->slots[level]++;
6220 continue;
6221 } else if (ret < 0)
6222 return ret;
6223 level = wc->level;
6224 }
6225 return 0;
6226 }
6227
6228 static noinline int walk_up_tree(struct btrfs_trans_handle *trans,
6229 struct btrfs_root *root,
6230 struct btrfs_path *path,
6231 struct walk_control *wc, int max_level)
6232 {
6233 int level = wc->level;
6234 int ret;
6235
6236 path->slots[level] = btrfs_header_nritems(path->nodes[level]);
6237 while (level < max_level && path->nodes[level]) {
6238 wc->level = level;
6239 if (path->slots[level] + 1 <
6240 btrfs_header_nritems(path->nodes[level])) {
6241 path->slots[level]++;
6242 return 0;
6243 } else {
6244 ret = walk_up_proc(trans, root, path, wc);
6245 if (ret > 0)
6246 return 0;
6247
6248 if (path->locks[level]) {
6249 btrfs_tree_unlock_rw(path->nodes[level],
6250 path->locks[level]);
6251 path->locks[level] = 0;
6252 }
6253 free_extent_buffer(path->nodes[level]);
6254 path->nodes[level] = NULL;
6255 level++;
6256 }
6257 }
6258 return 1;
6259 }
6260
6261 /*
6262 * drop a subvolume tree.
6263 *
6264 * this function traverses the tree freeing any blocks that only
6265 * referenced by the tree.
6266 *
6267 * when a shared tree block is found. this function decreases its
6268 * reference count by one. if update_ref is true, this function
6269 * also make sure backrefs for the shared block and all lower level
6270 * blocks are properly updated.
6271 */
6272 int btrfs_drop_snapshot(struct btrfs_root *root,
6273 struct btrfs_block_rsv *block_rsv, int update_ref)
6274 {
6275 struct btrfs_path *path;
6276 struct btrfs_trans_handle *trans;
6277 struct btrfs_root *tree_root = root->fs_info->tree_root;
6278 struct btrfs_root_item *root_item = &root->root_item;
6279 struct walk_control *wc;
6280 struct btrfs_key key;
6281 int err = 0;
6282 int ret;
6283 int level;
6284
6285 path = btrfs_alloc_path();
6286 if (!path)
6287 return -ENOMEM;
6288
6289 wc = kzalloc(sizeof(*wc), GFP_NOFS);
6290 if (!wc) {
6291 btrfs_free_path(path);
6292 return -ENOMEM;
6293 }
6294
6295 trans = btrfs_start_transaction(tree_root, 0);
6296 BUG_ON(IS_ERR(trans));
6297
6298 if (block_rsv)
6299 trans->block_rsv = block_rsv;
6300
6301 if (btrfs_disk_key_objectid(&root_item->drop_progress) == 0) {
6302 level = btrfs_header_level(root->node);
6303 path->nodes[level] = btrfs_lock_root_node(root);
6304 btrfs_set_lock_blocking(path->nodes[level]);
6305 path->slots[level] = 0;
6306 path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING;
6307 memset(&wc->update_progress, 0,
6308 sizeof(wc->update_progress));
6309 } else {
6310 btrfs_disk_key_to_cpu(&key, &root_item->drop_progress);
6311 memcpy(&wc->update_progress, &key,
6312 sizeof(wc->update_progress));
6313
6314 level = root_item->drop_level;
6315 BUG_ON(level == 0);
6316 path->lowest_level = level;
6317 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
6318 path->lowest_level = 0;
6319 if (ret < 0) {
6320 err = ret;
6321 goto out;
6322 }
6323 WARN_ON(ret > 0);
6324
6325 /*
6326 * unlock our path, this is safe because only this
6327 * function is allowed to delete this snapshot
6328 */
6329 btrfs_unlock_up_safe(path, 0);
6330
6331 level = btrfs_header_level(root->node);
6332 while (1) {
6333 btrfs_tree_lock(path->nodes[level]);
6334 btrfs_set_lock_blocking(path->nodes[level]);
6335
6336 ret = btrfs_lookup_extent_info(trans, root,
6337 path->nodes[level]->start,
6338 path->nodes[level]->len,
6339 &wc->refs[level],
6340 &wc->flags[level]);
6341 BUG_ON(ret);
6342 BUG_ON(wc->refs[level] == 0);
6343
6344 if (level == root_item->drop_level)
6345 break;
6346
6347 btrfs_tree_unlock(path->nodes[level]);
6348 WARN_ON(wc->refs[level] != 1);
6349 level--;
6350 }
6351 }
6352
6353 wc->level = level;
6354 wc->shared_level = -1;
6355 wc->stage = DROP_REFERENCE;
6356 wc->update_ref = update_ref;
6357 wc->keep_locks = 0;
6358 wc->reada_count = BTRFS_NODEPTRS_PER_BLOCK(root);
6359
6360 while (1) {
6361 ret = walk_down_tree(trans, root, path, wc);
6362 if (ret < 0) {
6363 err = ret;
6364 break;
6365 }
6366
6367 ret = walk_up_tree(trans, root, path, wc, BTRFS_MAX_LEVEL);
6368 if (ret < 0) {
6369 err = ret;
6370 break;
6371 }
6372
6373 if (ret > 0) {
6374 BUG_ON(wc->stage != DROP_REFERENCE);
6375 break;
6376 }
6377
6378 if (wc->stage == DROP_REFERENCE) {
6379 level = wc->level;
6380 btrfs_node_key(path->nodes[level],
6381 &root_item->drop_progress,
6382 path->slots[level]);
6383 root_item->drop_level = level;
6384 }
6385
6386 BUG_ON(wc->level == 0);
6387 if (btrfs_should_end_transaction(trans, tree_root)) {
6388 ret = btrfs_update_root(trans, tree_root,
6389 &root->root_key,
6390 root_item);
6391 BUG_ON(ret);
6392
6393 btrfs_end_transaction_throttle(trans, tree_root);
6394 trans = btrfs_start_transaction(tree_root, 0);
6395 BUG_ON(IS_ERR(trans));
6396 if (block_rsv)
6397 trans->block_rsv = block_rsv;
6398 }
6399 }
6400 btrfs_release_path(path);
6401 BUG_ON(err);
6402
6403 ret = btrfs_del_root(trans, tree_root, &root->root_key);
6404 BUG_ON(ret);
6405
6406 if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) {
6407 ret = btrfs_find_last_root(tree_root, root->root_key.objectid,
6408 NULL, NULL);
6409 BUG_ON(ret < 0);
6410 if (ret > 0) {
6411 /* if we fail to delete the orphan item this time
6412 * around, it'll get picked up the next time.
6413 *
6414 * The most common failure here is just -ENOENT.
6415 */
6416 btrfs_del_orphan_item(trans, tree_root,
6417 root->root_key.objectid);
6418 }
6419 }
6420
6421 if (root->in_radix) {
6422 btrfs_free_fs_root(tree_root->fs_info, root);
6423 } else {
6424 free_extent_buffer(root->node);
6425 free_extent_buffer(root->commit_root);
6426 kfree(root);
6427 }
6428 out:
6429 btrfs_end_transaction_throttle(trans, tree_root);
6430 kfree(wc);
6431 btrfs_free_path(path);
6432 return err;
6433 }
6434
6435 /*
6436 * drop subtree rooted at tree block 'node'.
6437 *
6438 * NOTE: this function will unlock and release tree block 'node'
6439 */
6440 int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
6441 struct btrfs_root *root,
6442 struct extent_buffer *node,
6443 struct extent_buffer *parent)
6444 {
6445 struct btrfs_path *path;
6446 struct walk_control *wc;
6447 int level;
6448 int parent_level;
6449 int ret = 0;
6450 int wret;
6451
6452 BUG_ON(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID);
6453
6454 path = btrfs_alloc_path();
6455 if (!path)
6456 return -ENOMEM;
6457
6458 wc = kzalloc(sizeof(*wc), GFP_NOFS);
6459 if (!wc) {
6460 btrfs_free_path(path);
6461 return -ENOMEM;
6462 }
6463
6464 btrfs_assert_tree_locked(parent);
6465 parent_level = btrfs_header_level(parent);
6466 extent_buffer_get(parent);
6467 path->nodes[parent_level] = parent;
6468 path->slots[parent_level] = btrfs_header_nritems(parent);
6469
6470 btrfs_assert_tree_locked(node);
6471 level = btrfs_header_level(node);
6472 path->nodes[level] = node;
6473 path->slots[level] = 0;
6474 path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING;
6475
6476 wc->refs[parent_level] = 1;
6477 wc->flags[parent_level] = BTRFS_BLOCK_FLAG_FULL_BACKREF;
6478 wc->level = level;
6479 wc->shared_level = -1;
6480 wc->stage = DROP_REFERENCE;
6481 wc->update_ref = 0;
6482 wc->keep_locks = 1;
6483 wc->reada_count = BTRFS_NODEPTRS_PER_BLOCK(root);
6484
6485 while (1) {
6486 wret = walk_down_tree(trans, root, path, wc);
6487 if (wret < 0) {
6488 ret = wret;
6489 break;
6490 }
6491
6492 wret = walk_up_tree(trans, root, path, wc, parent_level);
6493 if (wret < 0)
6494 ret = wret;
6495 if (wret != 0)
6496 break;
6497 }
6498
6499 kfree(wc);
6500 btrfs_free_path(path);
6501 return ret;
6502 }
6503
6504 static u64 update_block_group_flags(struct btrfs_root *root, u64 flags)
6505 {
6506 u64 num_devices;
6507 u64 stripped = BTRFS_BLOCK_GROUP_RAID0 |
6508 BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10;
6509
6510 /*
6511 * we add in the count of missing devices because we want
6512 * to make sure that any RAID levels on a degraded FS
6513 * continue to be honored.
6514 */
6515 num_devices = root->fs_info->fs_devices->rw_devices +
6516 root->fs_info->fs_devices->missing_devices;
6517
6518 if (num_devices == 1) {
6519 stripped |= BTRFS_BLOCK_GROUP_DUP;
6520 stripped = flags & ~stripped;
6521
6522 /* turn raid0 into single device chunks */
6523 if (flags & BTRFS_BLOCK_GROUP_RAID0)
6524 return stripped;
6525
6526 /* turn mirroring into duplication */
6527 if (flags & (BTRFS_BLOCK_GROUP_RAID1 |
6528 BTRFS_BLOCK_GROUP_RAID10))
6529 return stripped | BTRFS_BLOCK_GROUP_DUP;
6530 return flags;
6531 } else {
6532 /* they already had raid on here, just return */
6533 if (flags & stripped)
6534 return flags;
6535
6536 stripped |= BTRFS_BLOCK_GROUP_DUP;
6537 stripped = flags & ~stripped;
6538
6539 /* switch duplicated blocks with raid1 */
6540 if (flags & BTRFS_BLOCK_GROUP_DUP)
6541 return stripped | BTRFS_BLOCK_GROUP_RAID1;
6542
6543 /* turn single device chunks into raid0 */
6544 return stripped | BTRFS_BLOCK_GROUP_RAID0;
6545 }
6546 return flags;
6547 }
6548
6549 static int set_block_group_ro(struct btrfs_block_group_cache *cache, int force)
6550 {
6551 struct btrfs_space_info *sinfo = cache->space_info;
6552 u64 num_bytes;
6553 u64 min_allocable_bytes;
6554 int ret = -ENOSPC;
6555
6556
6557 /*
6558 * We need some metadata space and system metadata space for
6559 * allocating chunks in some corner cases until we force to set
6560 * it to be readonly.
6561 */
6562 if ((sinfo->flags &
6563 (BTRFS_BLOCK_GROUP_SYSTEM | BTRFS_BLOCK_GROUP_METADATA)) &&
6564 !force)
6565 min_allocable_bytes = 1 * 1024 * 1024;
6566 else
6567 min_allocable_bytes = 0;
6568
6569 spin_lock(&sinfo->lock);
6570 spin_lock(&cache->lock);
6571
6572 if (cache->ro) {
6573 ret = 0;
6574 goto out;
6575 }
6576
6577 num_bytes = cache->key.offset - cache->reserved - cache->pinned -
6578 cache->bytes_super - btrfs_block_group_used(&cache->item);
6579
6580 if (sinfo->bytes_used + sinfo->bytes_reserved + sinfo->bytes_pinned +
6581 sinfo->bytes_may_use + sinfo->bytes_readonly +
6582 cache->reserved_pinned + num_bytes + min_allocable_bytes <=
6583 sinfo->total_bytes) {
6584 sinfo->bytes_readonly += num_bytes;
6585 sinfo->bytes_reserved += cache->reserved_pinned;
6586 cache->reserved_pinned = 0;
6587 cache->ro = 1;
6588 ret = 0;
6589 }
6590 out:
6591 spin_unlock(&cache->lock);
6592 spin_unlock(&sinfo->lock);
6593 return ret;
6594 }
6595
6596 int btrfs_set_block_group_ro(struct btrfs_root *root,
6597 struct btrfs_block_group_cache *cache)
6598
6599 {
6600 struct btrfs_trans_handle *trans;
6601 u64 alloc_flags;
6602 int ret;
6603
6604 BUG_ON(cache->ro);
6605
6606 trans = btrfs_join_transaction(root);
6607 BUG_ON(IS_ERR(trans));
6608
6609 alloc_flags = update_block_group_flags(root, cache->flags);
6610 if (alloc_flags != cache->flags)
6611 do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags,
6612 CHUNK_ALLOC_FORCE);
6613
6614 ret = set_block_group_ro(cache, 0);
6615 if (!ret)
6616 goto out;
6617 alloc_flags = get_alloc_profile(root, cache->space_info->flags);
6618 ret = do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags,
6619 CHUNK_ALLOC_FORCE);
6620 if (ret < 0)
6621 goto out;
6622 ret = set_block_group_ro(cache, 0);
6623 out:
6624 btrfs_end_transaction(trans, root);
6625 return ret;
6626 }
6627
6628 int btrfs_force_chunk_alloc(struct btrfs_trans_handle *trans,
6629 struct btrfs_root *root, u64 type)
6630 {
6631 u64 alloc_flags = get_alloc_profile(root, type);
6632 return do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags,
6633 CHUNK_ALLOC_FORCE);
6634 }
6635
6636 /*
6637 * helper to account the unused space of all the readonly block group in the
6638 * list. takes mirrors into account.
6639 */
6640 static u64 __btrfs_get_ro_block_group_free_space(struct list_head *groups_list)
6641 {
6642 struct btrfs_block_group_cache *block_group;
6643 u64 free_bytes = 0;
6644 int factor;
6645
6646 list_for_each_entry(block_group, groups_list, list) {
6647 spin_lock(&block_group->lock);
6648
6649 if (!block_group->ro) {
6650 spin_unlock(&block_group->lock);
6651 continue;
6652 }
6653
6654 if (block_group->flags & (BTRFS_BLOCK_GROUP_RAID1 |
6655 BTRFS_BLOCK_GROUP_RAID10 |
6656 BTRFS_BLOCK_GROUP_DUP))
6657 factor = 2;
6658 else
6659 factor = 1;
6660
6661 free_bytes += (block_group->key.offset -
6662 btrfs_block_group_used(&block_group->item)) *
6663 factor;
6664
6665 spin_unlock(&block_group->lock);
6666 }
6667
6668 return free_bytes;
6669 }
6670
6671 /*
6672 * helper to account the unused space of all the readonly block group in the
6673 * space_info. takes mirrors into account.
6674 */
6675 u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo)
6676 {
6677 int i;
6678 u64 free_bytes = 0;
6679
6680 spin_lock(&sinfo->lock);
6681
6682 for(i = 0; i < BTRFS_NR_RAID_TYPES; i++)
6683 if (!list_empty(&sinfo->block_groups[i]))
6684 free_bytes += __btrfs_get_ro_block_group_free_space(
6685 &sinfo->block_groups[i]);
6686
6687 spin_unlock(&sinfo->lock);
6688
6689 return free_bytes;
6690 }
6691
6692 int btrfs_set_block_group_rw(struct btrfs_root *root,
6693 struct btrfs_block_group_cache *cache)
6694 {
6695 struct btrfs_space_info *sinfo = cache->space_info;
6696 u64 num_bytes;
6697
6698 BUG_ON(!cache->ro);
6699
6700 spin_lock(&sinfo->lock);
6701 spin_lock(&cache->lock);
6702 num_bytes = cache->key.offset - cache->reserved - cache->pinned -
6703 cache->bytes_super - btrfs_block_group_used(&cache->item);
6704 sinfo->bytes_readonly -= num_bytes;
6705 cache->ro = 0;
6706 spin_unlock(&cache->lock);
6707 spin_unlock(&sinfo->lock);
6708 return 0;
6709 }
6710
6711 /*
6712 * checks to see if its even possible to relocate this block group.
6713 *
6714 * @return - -1 if it's not a good idea to relocate this block group, 0 if its
6715 * ok to go ahead and try.
6716 */
6717 int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr)
6718 {
6719 struct btrfs_block_group_cache *block_group;
6720 struct btrfs_space_info *space_info;
6721 struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
6722 struct btrfs_device *device;
6723 int full = 0;
6724 int ret = 0;
6725
6726 block_group = btrfs_lookup_block_group(root->fs_info, bytenr);
6727
6728 /* odd, couldn't find the block group, leave it alone */
6729 if (!block_group)
6730 return -1;
6731
6732 /* no bytes used, we're good */
6733 if (!btrfs_block_group_used(&block_group->item))
6734 goto out;
6735
6736 space_info = block_group->space_info;
6737 spin_lock(&space_info->lock);
6738
6739 full = space_info->full;
6740
6741 /*
6742 * if this is the last block group we have in this space, we can't
6743 * relocate it unless we're able to allocate a new chunk below.
6744 *
6745 * Otherwise, we need to make sure we have room in the space to handle
6746 * all of the extents from this block group. If we can, we're good
6747 */
6748 if ((space_info->total_bytes != block_group->key.offset) &&
6749 (space_info->bytes_used + space_info->bytes_reserved +
6750 space_info->bytes_pinned + space_info->bytes_readonly +
6751 btrfs_block_group_used(&block_group->item) <
6752 space_info->total_bytes)) {
6753 spin_unlock(&space_info->lock);
6754 goto out;
6755 }
6756 spin_unlock(&space_info->lock);
6757
6758 /*
6759 * ok we don't have enough space, but maybe we have free space on our
6760 * devices to allocate new chunks for relocation, so loop through our
6761 * alloc devices and guess if we have enough space. However, if we
6762 * were marked as full, then we know there aren't enough chunks, and we
6763 * can just return.
6764 */
6765 ret = -1;
6766 if (full)
6767 goto out;
6768
6769 mutex_lock(&root->fs_info->chunk_mutex);
6770 list_for_each_entry(device, &fs_devices->alloc_list, dev_alloc_list) {
6771 u64 min_free = btrfs_block_group_used(&block_group->item);
6772 u64 dev_offset;
6773
6774 /*
6775 * check to make sure we can actually find a chunk with enough
6776 * space to fit our block group in.
6777 */
6778 if (device->total_bytes > device->bytes_used + min_free) {
6779 ret = find_free_dev_extent(NULL, device, min_free,
6780 &dev_offset, NULL);
6781 if (!ret)
6782 break;
6783 ret = -1;
6784 }
6785 }
6786 mutex_unlock(&root->fs_info->chunk_mutex);
6787 out:
6788 btrfs_put_block_group(block_group);
6789 return ret;
6790 }
6791
6792 static int find_first_block_group(struct btrfs_root *root,
6793 struct btrfs_path *path, struct btrfs_key *key)
6794 {
6795 int ret = 0;
6796 struct btrfs_key found_key;
6797 struct extent_buffer *leaf;
6798 int slot;
6799
6800 ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
6801 if (ret < 0)
6802 goto out;
6803
6804 while (1) {
6805 slot = path->slots[0];
6806 leaf = path->nodes[0];
6807 if (slot >= btrfs_header_nritems(leaf)) {
6808 ret = btrfs_next_leaf(root, path);
6809 if (ret == 0)
6810 continue;
6811 if (ret < 0)
6812 goto out;
6813 break;
6814 }
6815 btrfs_item_key_to_cpu(leaf, &found_key, slot);
6816
6817 if (found_key.objectid >= key->objectid &&
6818 found_key.type == BTRFS_BLOCK_GROUP_ITEM_KEY) {
6819 ret = 0;
6820 goto out;
6821 }
6822 path->slots[0]++;
6823 }
6824 out:
6825 return ret;
6826 }
6827
6828 void btrfs_put_block_group_cache(struct btrfs_fs_info *info)
6829 {
6830 struct btrfs_block_group_cache *block_group;
6831 u64 last = 0;
6832
6833 while (1) {
6834 struct inode *inode;
6835
6836 block_group = btrfs_lookup_first_block_group(info, last);
6837 while (block_group) {
6838 spin_lock(&block_group->lock);
6839 if (block_group->iref)
6840 break;
6841 spin_unlock(&block_group->lock);
6842 block_group = next_block_group(info->tree_root,
6843 block_group);
6844 }
6845 if (!block_group) {
6846 if (last == 0)
6847 break;
6848 last = 0;
6849 continue;
6850 }
6851
6852 inode = block_group->inode;
6853 block_group->iref = 0;
6854 block_group->inode = NULL;
6855 spin_unlock(&block_group->lock);
6856 iput(inode);
6857 last = block_group->key.objectid + block_group->key.offset;
6858 btrfs_put_block_group(block_group);
6859 }
6860 }
6861
6862 int btrfs_free_block_groups(struct btrfs_fs_info *info)
6863 {
6864 struct btrfs_block_group_cache *block_group;
6865 struct btrfs_space_info *space_info;
6866 struct btrfs_caching_control *caching_ctl;
6867 struct rb_node *n;
6868
6869 down_write(&info->extent_commit_sem);
6870 while (!list_empty(&info->caching_block_groups)) {
6871 caching_ctl = list_entry(info->caching_block_groups.next,
6872 struct btrfs_caching_control, list);
6873 list_del(&caching_ctl->list);
6874 put_caching_control(caching_ctl);
6875 }
6876 up_write(&info->extent_commit_sem);
6877
6878 spin_lock(&info->block_group_cache_lock);
6879 while ((n = rb_last(&info->block_group_cache_tree)) != NULL) {
6880 block_group = rb_entry(n, struct btrfs_block_group_cache,
6881 cache_node);
6882 rb_erase(&block_group->cache_node,
6883 &info->block_group_cache_tree);
6884 spin_unlock(&info->block_group_cache_lock);
6885
6886 down_write(&block_group->space_info->groups_sem);
6887 list_del(&block_group->list);
6888 up_write(&block_group->space_info->groups_sem);
6889
6890 if (block_group->cached == BTRFS_CACHE_STARTED)
6891 wait_block_group_cache_done(block_group);
6892
6893 /*
6894 * We haven't cached this block group, which means we could
6895 * possibly have excluded extents on this block group.
6896 */
6897 if (block_group->cached == BTRFS_CACHE_NO)
6898 free_excluded_extents(info->extent_root, block_group);
6899
6900 btrfs_remove_free_space_cache(block_group);
6901 btrfs_put_block_group(block_group);
6902
6903 spin_lock(&info->block_group_cache_lock);
6904 }
6905 spin_unlock(&info->block_group_cache_lock);
6906
6907 /* now that all the block groups are freed, go through and
6908 * free all the space_info structs. This is only called during
6909 * the final stages of unmount, and so we know nobody is
6910 * using them. We call synchronize_rcu() once before we start,
6911 * just to be on the safe side.
6912 */
6913 synchronize_rcu();
6914
6915 release_global_block_rsv(info);
6916
6917 while(!list_empty(&info->space_info)) {
6918 space_info = list_entry(info->space_info.next,
6919 struct btrfs_space_info,
6920 list);
6921 if (space_info->bytes_pinned > 0 ||
6922 space_info->bytes_reserved > 0) {
6923 WARN_ON(1);
6924 dump_space_info(space_info, 0, 0);
6925 }
6926 list_del(&space_info->list);
6927 kfree(space_info);
6928 }
6929 return 0;
6930 }
6931
6932 static void __link_block_group(struct btrfs_space_info *space_info,
6933 struct btrfs_block_group_cache *cache)
6934 {
6935 int index = get_block_group_index(cache);
6936
6937 down_write(&space_info->groups_sem);
6938 list_add_tail(&cache->list, &space_info->block_groups[index]);
6939 up_write(&space_info->groups_sem);
6940 }
6941
6942 int btrfs_read_block_groups(struct btrfs_root *root)
6943 {
6944 struct btrfs_path *path;
6945 int ret;
6946 struct btrfs_block_group_cache *cache;
6947 struct btrfs_fs_info *info = root->fs_info;
6948 struct btrfs_space_info *space_info;
6949 struct btrfs_key key;
6950 struct btrfs_key found_key;
6951 struct extent_buffer *leaf;
6952 int need_clear = 0;
6953 u64 cache_gen;
6954
6955 root = info->extent_root;
6956 key.objectid = 0;
6957 key.offset = 0;
6958 btrfs_set_key_type(&key, BTRFS_BLOCK_GROUP_ITEM_KEY);
6959 path = btrfs_alloc_path();
6960 if (!path)
6961 return -ENOMEM;
6962 path->reada = 1;
6963
6964 cache_gen = btrfs_super_cache_generation(&root->fs_info->super_copy);
6965 if (cache_gen != 0 &&
6966 btrfs_super_generation(&root->fs_info->super_copy) != cache_gen)
6967 need_clear = 1;
6968 if (btrfs_test_opt(root, CLEAR_CACHE))
6969 need_clear = 1;
6970 if (!btrfs_test_opt(root, SPACE_CACHE) && cache_gen)
6971 printk(KERN_INFO "btrfs: disk space caching is enabled\n");
6972
6973 while (1) {
6974 ret = find_first_block_group(root, path, &key);
6975 if (ret > 0)
6976 break;
6977 if (ret != 0)
6978 goto error;
6979 leaf = path->nodes[0];
6980 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
6981 cache = kzalloc(sizeof(*cache), GFP_NOFS);
6982 if (!cache) {
6983 ret = -ENOMEM;
6984 goto error;
6985 }
6986 cache->free_space_ctl = kzalloc(sizeof(*cache->free_space_ctl),
6987 GFP_NOFS);
6988 if (!cache->free_space_ctl) {
6989 kfree(cache);
6990 ret = -ENOMEM;
6991 goto error;
6992 }
6993
6994 atomic_set(&cache->count, 1);
6995 spin_lock_init(&cache->lock);
6996 cache->fs_info = info;
6997 INIT_LIST_HEAD(&cache->list);
6998 INIT_LIST_HEAD(&cache->cluster_list);
6999
7000 if (need_clear)
7001 cache->disk_cache_state = BTRFS_DC_CLEAR;
7002
7003 read_extent_buffer(leaf, &cache->item,
7004 btrfs_item_ptr_offset(leaf, path->slots[0]),
7005 sizeof(cache->item));
7006 memcpy(&cache->key, &found_key, sizeof(found_key));
7007
7008 key.objectid = found_key.objectid + found_key.offset;
7009 btrfs_release_path(path);
7010 cache->flags = btrfs_block_group_flags(&cache->item);
7011 cache->sectorsize = root->sectorsize;
7012
7013 btrfs_init_free_space_ctl(cache);
7014
7015 /*
7016 * We need to exclude the super stripes now so that the space
7017 * info has super bytes accounted for, otherwise we'll think
7018 * we have more space than we actually do.
7019 */
7020 exclude_super_stripes(root, cache);
7021
7022 /*
7023 * check for two cases, either we are full, and therefore
7024 * don't need to bother with the caching work since we won't
7025 * find any space, or we are empty, and we can just add all
7026 * the space in and be done with it. This saves us _alot_ of
7027 * time, particularly in the full case.
7028 */
7029 if (found_key.offset == btrfs_block_group_used(&cache->item)) {
7030 cache->last_byte_to_unpin = (u64)-1;
7031 cache->cached = BTRFS_CACHE_FINISHED;
7032 free_excluded_extents(root, cache);
7033 } else if (btrfs_block_group_used(&cache->item) == 0) {
7034 cache->last_byte_to_unpin = (u64)-1;
7035 cache->cached = BTRFS_CACHE_FINISHED;
7036 add_new_free_space(cache, root->fs_info,
7037 found_key.objectid,
7038 found_key.objectid +
7039 found_key.offset);
7040 free_excluded_extents(root, cache);
7041 }
7042
7043 ret = update_space_info(info, cache->flags, found_key.offset,
7044 btrfs_block_group_used(&cache->item),
7045 &space_info);
7046 BUG_ON(ret);
7047 cache->space_info = space_info;
7048 spin_lock(&cache->space_info->lock);
7049 cache->space_info->bytes_readonly += cache->bytes_super;
7050 spin_unlock(&cache->space_info->lock);
7051
7052 __link_block_group(space_info, cache);
7053
7054 ret = btrfs_add_block_group_cache(root->fs_info, cache);
7055 BUG_ON(ret);
7056
7057 set_avail_alloc_bits(root->fs_info, cache->flags);
7058 if (btrfs_chunk_readonly(root, cache->key.objectid))
7059 set_block_group_ro(cache, 1);
7060 }
7061
7062 list_for_each_entry_rcu(space_info, &root->fs_info->space_info, list) {
7063 if (!(get_alloc_profile(root, space_info->flags) &
7064 (BTRFS_BLOCK_GROUP_RAID10 |
7065 BTRFS_BLOCK_GROUP_RAID1 |
7066 BTRFS_BLOCK_GROUP_DUP)))
7067 continue;
7068 /*
7069 * avoid allocating from un-mirrored block group if there are
7070 * mirrored block groups.
7071 */
7072 list_for_each_entry(cache, &space_info->block_groups[3], list)
7073 set_block_group_ro(cache, 1);
7074 list_for_each_entry(cache, &space_info->block_groups[4], list)
7075 set_block_group_ro(cache, 1);
7076 }
7077
7078 init_global_block_rsv(info);
7079 ret = 0;
7080 error:
7081 btrfs_free_path(path);
7082 return ret;
7083 }
7084
7085 int btrfs_make_block_group(struct btrfs_trans_handle *trans,
7086 struct btrfs_root *root, u64 bytes_used,
7087 u64 type, u64 chunk_objectid, u64 chunk_offset,
7088 u64 size)
7089 {
7090 int ret;
7091 struct btrfs_root *extent_root;
7092 struct btrfs_block_group_cache *cache;
7093
7094 extent_root = root->fs_info->extent_root;
7095
7096 root->fs_info->last_trans_log_full_commit = trans->transid;
7097
7098 cache = kzalloc(sizeof(*cache), GFP_NOFS);
7099 if (!cache)
7100 return -ENOMEM;
7101 cache->free_space_ctl = kzalloc(sizeof(*cache->free_space_ctl),
7102 GFP_NOFS);
7103 if (!cache->free_space_ctl) {
7104 kfree(cache);
7105 return -ENOMEM;
7106 }
7107
7108 cache->key.objectid = chunk_offset;
7109 cache->key.offset = size;
7110 cache->key.type = BTRFS_BLOCK_GROUP_ITEM_KEY;
7111 cache->sectorsize = root->sectorsize;
7112 cache->fs_info = root->fs_info;
7113
7114 atomic_set(&cache->count, 1);
7115 spin_lock_init(&cache->lock);
7116 INIT_LIST_HEAD(&cache->list);
7117 INIT_LIST_HEAD(&cache->cluster_list);
7118
7119 btrfs_init_free_space_ctl(cache);
7120
7121 btrfs_set_block_group_used(&cache->item, bytes_used);
7122 btrfs_set_block_group_chunk_objectid(&cache->item, chunk_objectid);
7123 cache->flags = type;
7124 btrfs_set_block_group_flags(&cache->item, type);
7125
7126 cache->last_byte_to_unpin = (u64)-1;
7127 cache->cached = BTRFS_CACHE_FINISHED;
7128 exclude_super_stripes(root, cache);
7129
7130 add_new_free_space(cache, root->fs_info, chunk_offset,
7131 chunk_offset + size);
7132
7133 free_excluded_extents(root, cache);
7134
7135 ret = update_space_info(root->fs_info, cache->flags, size, bytes_used,
7136 &cache->space_info);
7137 BUG_ON(ret);
7138
7139 spin_lock(&cache->space_info->lock);
7140 cache->space_info->bytes_readonly += cache->bytes_super;
7141 spin_unlock(&cache->space_info->lock);
7142
7143 __link_block_group(cache->space_info, cache);
7144
7145 ret = btrfs_add_block_group_cache(root->fs_info, cache);
7146 BUG_ON(ret);
7147
7148 ret = btrfs_insert_item(trans, extent_root, &cache->key, &cache->item,
7149 sizeof(cache->item));
7150 BUG_ON(ret);
7151
7152 set_avail_alloc_bits(extent_root->fs_info, type);
7153
7154 return 0;
7155 }
7156
7157 int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
7158 struct btrfs_root *root, u64 group_start)
7159 {
7160 struct btrfs_path *path;
7161 struct btrfs_block_group_cache *block_group;
7162 struct btrfs_free_cluster *cluster;
7163 struct btrfs_root *tree_root = root->fs_info->tree_root;
7164 struct btrfs_key key;
7165 struct inode *inode;
7166 int ret;
7167 int factor;
7168
7169 root = root->fs_info->extent_root;
7170
7171 block_group = btrfs_lookup_block_group(root->fs_info, group_start);
7172 BUG_ON(!block_group);
7173 BUG_ON(!block_group->ro);
7174
7175 /*
7176 * Free the reserved super bytes from this block group before
7177 * remove it.
7178 */
7179 free_excluded_extents(root, block_group);
7180
7181 memcpy(&key, &block_group->key, sizeof(key));
7182 if (block_group->flags & (BTRFS_BLOCK_GROUP_DUP |
7183 BTRFS_BLOCK_GROUP_RAID1 |
7184 BTRFS_BLOCK_GROUP_RAID10))
7185 factor = 2;
7186 else
7187 factor = 1;
7188
7189 /* make sure this block group isn't part of an allocation cluster */
7190 cluster = &root->fs_info->data_alloc_cluster;
7191 spin_lock(&cluster->refill_lock);
7192 btrfs_return_cluster_to_free_space(block_group, cluster);
7193 spin_unlock(&cluster->refill_lock);
7194
7195 /*
7196 * make sure this block group isn't part of a metadata
7197 * allocation cluster
7198 */
7199 cluster = &root->fs_info->meta_alloc_cluster;
7200 spin_lock(&cluster->refill_lock);
7201 btrfs_return_cluster_to_free_space(block_group, cluster);
7202 spin_unlock(&cluster->refill_lock);
7203
7204 path = btrfs_alloc_path();
7205 if (!path) {
7206 ret = -ENOMEM;
7207 goto out;
7208 }
7209
7210 inode = lookup_free_space_inode(root, block_group, path);
7211 if (!IS_ERR(inode)) {
7212 ret = btrfs_orphan_add(trans, inode);
7213 BUG_ON(ret);
7214 clear_nlink(inode);
7215 /* One for the block groups ref */
7216 spin_lock(&block_group->lock);
7217 if (block_group->iref) {
7218 block_group->iref = 0;
7219 block_group->inode = NULL;
7220 spin_unlock(&block_group->lock);
7221 iput(inode);
7222 } else {
7223 spin_unlock(&block_group->lock);
7224 }
7225 /* One for our lookup ref */
7226 iput(inode);
7227 }
7228
7229 key.objectid = BTRFS_FREE_SPACE_OBJECTID;
7230 key.offset = block_group->key.objectid;
7231 key.type = 0;
7232
7233 ret = btrfs_search_slot(trans, tree_root, &key, path, -1, 1);
7234 if (ret < 0)
7235 goto out;
7236 if (ret > 0)
7237 btrfs_release_path(path);
7238 if (ret == 0) {
7239 ret = btrfs_del_item(trans, tree_root, path);
7240 if (ret)
7241 goto out;
7242 btrfs_release_path(path);
7243 }
7244
7245 spin_lock(&root->fs_info->block_group_cache_lock);
7246 rb_erase(&block_group->cache_node,
7247 &root->fs_info->block_group_cache_tree);
7248 spin_unlock(&root->fs_info->block_group_cache_lock);
7249
7250 down_write(&block_group->space_info->groups_sem);
7251 /*
7252 * we must use list_del_init so people can check to see if they
7253 * are still on the list after taking the semaphore
7254 */
7255 list_del_init(&block_group->list);
7256 up_write(&block_group->space_info->groups_sem);
7257
7258 if (block_group->cached == BTRFS_CACHE_STARTED)
7259 wait_block_group_cache_done(block_group);
7260
7261 btrfs_remove_free_space_cache(block_group);
7262
7263 spin_lock(&block_group->space_info->lock);
7264 block_group->space_info->total_bytes -= block_group->key.offset;
7265 block_group->space_info->bytes_readonly -= block_group->key.offset;
7266 block_group->space_info->disk_total -= block_group->key.offset * factor;
7267 spin_unlock(&block_group->space_info->lock);
7268
7269 memcpy(&key, &block_group->key, sizeof(key));
7270
7271 btrfs_clear_space_info_full(root->fs_info);
7272
7273 btrfs_put_block_group(block_group);
7274 btrfs_put_block_group(block_group);
7275
7276 ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
7277 if (ret > 0)
7278 ret = -EIO;
7279 if (ret < 0)
7280 goto out;
7281
7282 ret = btrfs_del_item(trans, root, path);
7283 out:
7284 btrfs_free_path(path);
7285 return ret;
7286 }
7287
7288 int btrfs_init_space_info(struct btrfs_fs_info *fs_info)
7289 {
7290 struct btrfs_space_info *space_info;
7291 struct btrfs_super_block *disk_super;
7292 u64 features;
7293 u64 flags;
7294 int mixed = 0;
7295 int ret;
7296
7297 disk_super = &fs_info->super_copy;
7298 if (!btrfs_super_root(disk_super))
7299 return 1;
7300
7301 features = btrfs_super_incompat_flags(disk_super);
7302 if (features & BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS)
7303 mixed = 1;
7304
7305 flags = BTRFS_BLOCK_GROUP_SYSTEM;
7306 ret = update_space_info(fs_info, flags, 0, 0, &space_info);
7307 if (ret)
7308 goto out;
7309
7310 if (mixed) {
7311 flags = BTRFS_BLOCK_GROUP_METADATA | BTRFS_BLOCK_GROUP_DATA;
7312 ret = update_space_info(fs_info, flags, 0, 0, &space_info);
7313 } else {
7314 flags = BTRFS_BLOCK_GROUP_METADATA;
7315 ret = update_space_info(fs_info, flags, 0, 0, &space_info);
7316 if (ret)
7317 goto out;
7318
7319 flags = BTRFS_BLOCK_GROUP_DATA;
7320 ret = update_space_info(fs_info, flags, 0, 0, &space_info);
7321 }
7322 out:
7323 return ret;
7324 }
7325
7326 int btrfs_error_unpin_extent_range(struct btrfs_root *root, u64 start, u64 end)
7327 {
7328 return unpin_extent_range(root, start, end);
7329 }
7330
7331 int btrfs_error_discard_extent(struct btrfs_root *root, u64 bytenr,
7332 u64 num_bytes, u64 *actual_bytes)
7333 {
7334 return btrfs_discard_extent(root, bytenr, num_bytes, actual_bytes);
7335 }
7336
7337 int btrfs_trim_fs(struct btrfs_root *root, struct fstrim_range *range)
7338 {
7339 struct btrfs_fs_info *fs_info = root->fs_info;
7340 struct btrfs_block_group_cache *cache = NULL;
7341 u64 group_trimmed;
7342 u64 start;
7343 u64 end;
7344 u64 trimmed = 0;
7345 int ret = 0;
7346
7347 cache = btrfs_lookup_block_group(fs_info, range->start);
7348
7349 while (cache) {
7350 if (cache->key.objectid >= (range->start + range->len)) {
7351 btrfs_put_block_group(cache);
7352 break;
7353 }
7354
7355 start = max(range->start, cache->key.objectid);
7356 end = min(range->start + range->len,
7357 cache->key.objectid + cache->key.offset);
7358
7359 if (end - start >= range->minlen) {
7360 if (!block_group_cache_done(cache)) {
7361 ret = cache_block_group(cache, NULL, root, 0);
7362 if (!ret)
7363 wait_block_group_cache_done(cache);
7364 }
7365 ret = btrfs_trim_block_group(cache,
7366 &group_trimmed,
7367 start,
7368 end,
7369 range->minlen);
7370
7371 trimmed += group_trimmed;
7372 if (ret) {
7373 btrfs_put_block_group(cache);
7374 break;
7375 }
7376 }
7377
7378 cache = next_block_group(fs_info->tree_root, cache);
7379 }
7380
7381 range->len = trimmed;
7382 return ret;
7383 }
Linux Btrfs Development