2 * Copyright (C) 2007 Oracle. All rights reserved.
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.
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.
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.
18 #include <linux/sched.h>
19 #include <linux/pagemap.h>
20 #include <linux/writeback.h>
25 #include "print-tree.h"
26 #include "transaction.h"
30 #define BLOCK_GROUP_DATA EXTENT_WRITEBACK
31 #define BLOCK_GROUP_METADATA EXTENT_UPTODATE
32 #define BLOCK_GROUP_SYSTEM EXTENT_NEW
34 #define BLOCK_GROUP_DIRTY EXTENT_DIRTY
36 static int finish_current_insert(struct btrfs_trans_handle *trans, struct
37 btrfs_root *extent_root);
38 static int del_pending_extents(struct btrfs_trans_handle *trans, struct
39 btrfs_root *extent_root);
40 static struct btrfs_block_group_cache *
41 __btrfs_find_block_group(struct btrfs_root *root,
42 struct btrfs_block_group_cache *hint,
43 u64 search_start, int data, int owner);
45 void maybe_lock_mutex(struct btrfs_root *root)
47 if (root != root->fs_info->extent_root &&
48 root != root->fs_info->chunk_root &&
49 root != root->fs_info->dev_root) {
50 mutex_lock(&root->fs_info->alloc_mutex);
54 void maybe_unlock_mutex(struct btrfs_root *root)
56 if (root != root->fs_info->extent_root &&
57 root != root->fs_info->chunk_root &&
58 root != root->fs_info->dev_root) {
59 mutex_unlock(&root->fs_info->alloc_mutex);
63 static int cache_block_group(struct btrfs_root *root,
64 struct btrfs_block_group_cache *block_group)
66 struct btrfs_path *path;
69 struct extent_buffer *leaf;
70 struct extent_io_tree *free_space_cache;
80 root = root->fs_info->extent_root;
81 free_space_cache = &root->fs_info->free_space_cache;
83 if (block_group->cached)
86 path = btrfs_alloc_path();
92 * we get into deadlocks with paths held by callers of this function.
93 * since the alloc_mutex is protecting things right now, just
94 * skip the locking here
96 path->skip_locking = 1;
97 first_free = block_group->key.objectid;
98 key.objectid = block_group->key.objectid;
100 btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
101 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
104 ret = btrfs_previous_item(root, path, 0, BTRFS_EXTENT_ITEM_KEY);
108 leaf = path->nodes[0];
109 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
110 if (key.objectid + key.offset > first_free)
111 first_free = key.objectid + key.offset;
114 leaf = path->nodes[0];
115 slot = path->slots[0];
116 if (slot >= btrfs_header_nritems(leaf)) {
117 ret = btrfs_next_leaf(root, path);
126 btrfs_item_key_to_cpu(leaf, &key, slot);
127 if (key.objectid < block_group->key.objectid) {
130 if (key.objectid >= block_group->key.objectid +
131 block_group->key.offset) {
135 if (btrfs_key_type(&key) == BTRFS_EXTENT_ITEM_KEY) {
140 if (key.objectid > last) {
141 hole_size = key.objectid - last;
142 set_extent_dirty(free_space_cache, last,
143 last + hole_size - 1,
146 last = key.objectid + key.offset;
154 if (block_group->key.objectid +
155 block_group->key.offset > last) {
156 hole_size = block_group->key.objectid +
157 block_group->key.offset - last;
158 set_extent_dirty(free_space_cache, last,
159 last + hole_size - 1, GFP_NOFS);
161 block_group->cached = 1;
163 btrfs_free_path(path);
167 struct btrfs_block_group_cache *btrfs_lookup_first_block_group(struct
171 struct extent_io_tree *block_group_cache;
172 struct btrfs_block_group_cache *block_group = NULL;
178 bytenr = max_t(u64, bytenr,
179 BTRFS_SUPER_INFO_OFFSET + BTRFS_SUPER_INFO_SIZE);
180 block_group_cache = &info->block_group_cache;
181 ret = find_first_extent_bit(block_group_cache,
182 bytenr, &start, &end,
183 BLOCK_GROUP_DATA | BLOCK_GROUP_METADATA |
188 ret = get_state_private(block_group_cache, start, &ptr);
192 block_group = (struct btrfs_block_group_cache *)(unsigned long)ptr;
196 struct btrfs_block_group_cache *btrfs_lookup_block_group(struct
200 struct extent_io_tree *block_group_cache;
201 struct btrfs_block_group_cache *block_group = NULL;
207 bytenr = max_t(u64, bytenr,
208 BTRFS_SUPER_INFO_OFFSET + BTRFS_SUPER_INFO_SIZE);
209 block_group_cache = &info->block_group_cache;
210 ret = find_first_extent_bit(block_group_cache,
211 bytenr, &start, &end,
212 BLOCK_GROUP_DATA | BLOCK_GROUP_METADATA |
217 ret = get_state_private(block_group_cache, start, &ptr);
221 block_group = (struct btrfs_block_group_cache *)(unsigned long)ptr;
222 if (block_group->key.objectid <= bytenr && bytenr <
223 block_group->key.objectid + block_group->key.offset)
228 static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits)
230 return (cache->flags & bits) == bits;
233 static int noinline find_search_start(struct btrfs_root *root,
234 struct btrfs_block_group_cache **cache_ret,
235 u64 *start_ret, u64 num, int data)
238 struct btrfs_block_group_cache *cache = *cache_ret;
239 struct extent_io_tree *free_space_cache;
240 struct extent_state *state;
245 u64 search_start = *start_ret;
248 WARN_ON(!mutex_is_locked(&root->fs_info->alloc_mutex));
249 total_fs_bytes = btrfs_super_total_bytes(&root->fs_info->super_copy);
250 free_space_cache = &root->fs_info->free_space_cache;
256 ret = cache_block_group(root, cache);
261 last = max(search_start, cache->key.objectid);
262 if (!block_group_bits(cache, data) || cache->ro)
265 spin_lock_irq(&free_space_cache->lock);
266 state = find_first_extent_bit_state(free_space_cache, last, EXTENT_DIRTY);
271 spin_unlock_irq(&free_space_cache->lock);
275 start = max(last, state->start);
276 last = state->end + 1;
277 if (last - start < num) {
279 state = extent_state_next(state);
280 } while(state && !(state->state & EXTENT_DIRTY));
283 spin_unlock_irq(&free_space_cache->lock);
287 if (start + num > cache->key.objectid + cache->key.offset)
289 if (!block_group_bits(cache, data)) {
290 printk("block group bits don't match %Lu %d\n", cache->flags, data);
296 cache = btrfs_lookup_block_group(root->fs_info, search_start);
298 printk("Unable to find block group for %Lu\n", search_start);
304 last = cache->key.objectid + cache->key.offset;
306 cache = btrfs_lookup_first_block_group(root->fs_info, last);
307 if (!cache || cache->key.objectid >= total_fs_bytes) {
316 if (cache_miss && !cache->cached) {
317 cache_block_group(root, cache);
319 cache = btrfs_lookup_first_block_group(root->fs_info, last);
322 cache = __btrfs_find_block_group(root, cache, last, data, 0);
329 static u64 div_factor(u64 num, int factor)
338 static int block_group_state_bits(u64 flags)
341 if (flags & BTRFS_BLOCK_GROUP_DATA)
342 bits |= BLOCK_GROUP_DATA;
343 if (flags & BTRFS_BLOCK_GROUP_METADATA)
344 bits |= BLOCK_GROUP_METADATA;
345 if (flags & BTRFS_BLOCK_GROUP_SYSTEM)
346 bits |= BLOCK_GROUP_SYSTEM;
350 static struct btrfs_block_group_cache *
351 __btrfs_find_block_group(struct btrfs_root *root,
352 struct btrfs_block_group_cache *hint,
353 u64 search_start, int data, int owner)
355 struct btrfs_block_group_cache *cache;
356 struct extent_io_tree *block_group_cache;
357 struct btrfs_block_group_cache *found_group = NULL;
358 struct btrfs_fs_info *info = root->fs_info;
371 block_group_cache = &info->block_group_cache;
373 if (data & BTRFS_BLOCK_GROUP_METADATA)
376 bit = block_group_state_bits(data);
379 struct btrfs_block_group_cache *shint;
380 shint = btrfs_lookup_first_block_group(info, search_start);
381 if (shint && block_group_bits(shint, data) && !shint->ro) {
382 used = btrfs_block_group_used(&shint->item);
383 if (used + shint->pinned <
384 div_factor(shint->key.offset, factor)) {
389 if (hint && !hint->ro && block_group_bits(hint, data)) {
390 used = btrfs_block_group_used(&hint->item);
391 if (used + hint->pinned <
392 div_factor(hint->key.offset, factor)) {
395 last = hint->key.objectid + hint->key.offset;
398 last = max(hint->key.objectid, search_start);
404 ret = find_first_extent_bit(block_group_cache, last,
409 ret = get_state_private(block_group_cache, start, &ptr);
415 cache = (struct btrfs_block_group_cache *)(unsigned long)ptr;
416 last = cache->key.objectid + cache->key.offset;
417 used = btrfs_block_group_used(&cache->item);
419 if (!cache->ro && block_group_bits(cache, data)) {
420 free_check = div_factor(cache->key.offset, factor);
421 if (used + cache->pinned < free_check) {
433 if (!full_search && factor < 10) {
443 struct btrfs_block_group_cache *btrfs_find_block_group(struct btrfs_root *root,
444 struct btrfs_block_group_cache
445 *hint, u64 search_start,
449 struct btrfs_block_group_cache *ret;
450 mutex_lock(&root->fs_info->alloc_mutex);
451 ret = __btrfs_find_block_group(root, hint, search_start, data, owner);
452 mutex_unlock(&root->fs_info->alloc_mutex);
455 static u64 hash_extent_ref(u64 root_objectid, u64 ref_generation,
456 u64 owner, u64 owner_offset)
458 u32 high_crc = ~(u32)0;
459 u32 low_crc = ~(u32)0;
461 lenum = cpu_to_le64(root_objectid);
462 high_crc = btrfs_crc32c(high_crc, &lenum, sizeof(lenum));
463 lenum = cpu_to_le64(ref_generation);
464 low_crc = btrfs_crc32c(low_crc, &lenum, sizeof(lenum));
465 if (owner >= BTRFS_FIRST_FREE_OBJECTID) {
466 lenum = cpu_to_le64(owner);
467 low_crc = btrfs_crc32c(low_crc, &lenum, sizeof(lenum));
468 lenum = cpu_to_le64(owner_offset);
469 low_crc = btrfs_crc32c(low_crc, &lenum, sizeof(lenum));
471 return ((u64)high_crc << 32) | (u64)low_crc;
474 static int match_extent_ref(struct extent_buffer *leaf,
475 struct btrfs_extent_ref *disk_ref,
476 struct btrfs_extent_ref *cpu_ref)
481 if (cpu_ref->objectid)
482 len = sizeof(*cpu_ref);
484 len = 2 * sizeof(u64);
485 ret = memcmp_extent_buffer(leaf, cpu_ref, (unsigned long)disk_ref,
490 static int noinline lookup_extent_backref(struct btrfs_trans_handle *trans,
491 struct btrfs_root *root,
492 struct btrfs_path *path, u64 bytenr,
494 u64 ref_generation, u64 owner,
495 u64 owner_offset, int del)
498 struct btrfs_key key;
499 struct btrfs_key found_key;
500 struct btrfs_extent_ref ref;
501 struct extent_buffer *leaf;
502 struct btrfs_extent_ref *disk_ref;
506 btrfs_set_stack_ref_root(&ref, root_objectid);
507 btrfs_set_stack_ref_generation(&ref, ref_generation);
508 btrfs_set_stack_ref_objectid(&ref, owner);
509 btrfs_set_stack_ref_offset(&ref, owner_offset);
511 hash = hash_extent_ref(root_objectid, ref_generation, owner,
514 key.objectid = bytenr;
515 key.type = BTRFS_EXTENT_REF_KEY;
518 ret = btrfs_search_slot(trans, root, &key, path,
522 leaf = path->nodes[0];
524 u32 nritems = btrfs_header_nritems(leaf);
525 if (path->slots[0] >= nritems) {
526 ret2 = btrfs_next_leaf(root, path);
529 leaf = path->nodes[0];
531 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
532 if (found_key.objectid != bytenr ||
533 found_key.type != BTRFS_EXTENT_REF_KEY)
535 key.offset = found_key.offset;
537 btrfs_release_path(root, path);
541 disk_ref = btrfs_item_ptr(path->nodes[0],
543 struct btrfs_extent_ref);
544 if (match_extent_ref(path->nodes[0], disk_ref, &ref)) {
548 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
549 key.offset = found_key.offset + 1;
550 btrfs_release_path(root, path);
557 * Back reference rules. Back refs have three main goals:
559 * 1) differentiate between all holders of references to an extent so that
560 * when a reference is dropped we can make sure it was a valid reference
561 * before freeing the extent.
563 * 2) Provide enough information to quickly find the holders of an extent
564 * if we notice a given block is corrupted or bad.
566 * 3) Make it easy to migrate blocks for FS shrinking or storage pool
567 * maintenance. This is actually the same as #2, but with a slightly
568 * different use case.
570 * File extents can be referenced by:
572 * - multiple snapshots, subvolumes, or different generations in one subvol
573 * - different files inside a single subvolume (in theory, not implemented yet)
574 * - different offsets inside a file (bookend extents in file.c)
576 * The extent ref structure has fields for:
578 * - Objectid of the subvolume root
579 * - Generation number of the tree holding the reference
580 * - objectid of the file holding the reference
581 * - offset in the file corresponding to the key holding the reference
583 * When a file extent is allocated the fields are filled in:
584 * (root_key.objectid, trans->transid, inode objectid, offset in file)
586 * When a leaf is cow'd new references are added for every file extent found
587 * in the leaf. It looks the same as the create case, but trans->transid
588 * will be different when the block is cow'd.
590 * (root_key.objectid, trans->transid, inode objectid, offset in file)
592 * When a file extent is removed either during snapshot deletion or file
593 * truncation, the corresponding back reference is found
596 * (btrfs_header_owner(leaf), btrfs_header_generation(leaf),
597 * inode objectid, offset in file)
599 * Btree extents can be referenced by:
601 * - Different subvolumes
602 * - Different generations of the same subvolume
604 * Storing sufficient information for a full reverse mapping of a btree
605 * block would require storing the lowest key of the block in the backref,
606 * and it would require updating that lowest key either before write out or
607 * every time it changed. Instead, the objectid of the lowest key is stored
608 * along with the level of the tree block. This provides a hint
609 * about where in the btree the block can be found. Searches through the
610 * btree only need to look for a pointer to that block, so they stop one
611 * level higher than the level recorded in the backref.
613 * Some btrees do not do reference counting on their extents. These
614 * include the extent tree and the tree of tree roots. Backrefs for these
615 * trees always have a generation of zero.
617 * When a tree block is created, back references are inserted:
619 * (root->root_key.objectid, trans->transid or zero, level, lowest_key_objectid)
621 * When a tree block is cow'd in a reference counted root,
622 * new back references are added for all the blocks it points to.
623 * These are of the form (trans->transid will have increased since creation):
625 * (root->root_key.objectid, trans->transid, level, lowest_key_objectid)
627 * Because the lowest_key_objectid and the level are just hints
628 * they are not used when backrefs are deleted. When a backref is deleted:
630 * if backref was for a tree root:
631 * root_objectid = root->root_key.objectid
633 * root_objectid = btrfs_header_owner(parent)
635 * (root_objectid, btrfs_header_generation(parent) or zero, 0, 0)
637 * Back Reference Key hashing:
639 * Back references have four fields, each 64 bits long. Unfortunately,
640 * This is hashed into a single 64 bit number and placed into the key offset.
641 * The key objectid corresponds to the first byte in the extent, and the
642 * key type is set to BTRFS_EXTENT_REF_KEY
644 int btrfs_insert_extent_backref(struct btrfs_trans_handle *trans,
645 struct btrfs_root *root,
646 struct btrfs_path *path, u64 bytenr,
647 u64 root_objectid, u64 ref_generation,
648 u64 owner, u64 owner_offset)
651 struct btrfs_key key;
652 struct btrfs_extent_ref ref;
653 struct btrfs_extent_ref *disk_ref;
656 btrfs_set_stack_ref_root(&ref, root_objectid);
657 btrfs_set_stack_ref_generation(&ref, ref_generation);
658 btrfs_set_stack_ref_objectid(&ref, owner);
659 btrfs_set_stack_ref_offset(&ref, owner_offset);
661 hash = hash_extent_ref(root_objectid, ref_generation, owner,
664 key.objectid = bytenr;
665 key.type = BTRFS_EXTENT_REF_KEY;
667 ret = btrfs_insert_empty_item(trans, root, path, &key, sizeof(ref));
668 while (ret == -EEXIST) {
669 disk_ref = btrfs_item_ptr(path->nodes[0], path->slots[0],
670 struct btrfs_extent_ref);
671 if (match_extent_ref(path->nodes[0], disk_ref, &ref))
674 btrfs_release_path(root, path);
675 ret = btrfs_insert_empty_item(trans, root, path, &key,
680 disk_ref = btrfs_item_ptr(path->nodes[0], path->slots[0],
681 struct btrfs_extent_ref);
682 write_extent_buffer(path->nodes[0], &ref, (unsigned long)disk_ref,
684 btrfs_mark_buffer_dirty(path->nodes[0]);
686 btrfs_release_path(root, path);
690 static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
691 struct btrfs_root *root,
692 u64 bytenr, u64 num_bytes,
693 u64 root_objectid, u64 ref_generation,
694 u64 owner, u64 owner_offset)
696 struct btrfs_path *path;
698 struct btrfs_key key;
699 struct extent_buffer *l;
700 struct btrfs_extent_item *item;
703 WARN_ON(num_bytes < root->sectorsize);
704 path = btrfs_alloc_path();
709 key.objectid = bytenr;
710 btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
711 key.offset = num_bytes;
712 ret = btrfs_search_slot(trans, root->fs_info->extent_root, &key, path,
721 item = btrfs_item_ptr(l, path->slots[0], struct btrfs_extent_item);
722 refs = btrfs_extent_refs(l, item);
723 btrfs_set_extent_refs(l, item, refs + 1);
724 btrfs_mark_buffer_dirty(path->nodes[0]);
726 btrfs_release_path(root->fs_info->extent_root, path);
729 ret = btrfs_insert_extent_backref(trans, root->fs_info->extent_root,
730 path, bytenr, root_objectid,
731 ref_generation, owner, owner_offset);
733 finish_current_insert(trans, root->fs_info->extent_root);
734 del_pending_extents(trans, root->fs_info->extent_root);
736 btrfs_free_path(path);
740 int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
741 struct btrfs_root *root,
742 u64 bytenr, u64 num_bytes,
743 u64 root_objectid, u64 ref_generation,
744 u64 owner, u64 owner_offset)
748 mutex_lock(&root->fs_info->alloc_mutex);
749 ret = __btrfs_inc_extent_ref(trans, root, bytenr, num_bytes,
750 root_objectid, ref_generation,
751 owner, owner_offset);
752 mutex_unlock(&root->fs_info->alloc_mutex);
756 int btrfs_extent_post_op(struct btrfs_trans_handle *trans,
757 struct btrfs_root *root)
759 finish_current_insert(trans, root->fs_info->extent_root);
760 del_pending_extents(trans, root->fs_info->extent_root);
764 static int lookup_extent_ref(struct btrfs_trans_handle *trans,
765 struct btrfs_root *root, u64 bytenr,
766 u64 num_bytes, u32 *refs)
768 struct btrfs_path *path;
770 struct btrfs_key key;
771 struct extent_buffer *l;
772 struct btrfs_extent_item *item;
774 WARN_ON(num_bytes < root->sectorsize);
775 path = btrfs_alloc_path();
777 key.objectid = bytenr;
778 key.offset = num_bytes;
779 btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
780 ret = btrfs_search_slot(trans, root->fs_info->extent_root, &key, path,
785 btrfs_print_leaf(root, path->nodes[0]);
786 printk("failed to find block number %Lu\n", bytenr);
790 item = btrfs_item_ptr(l, path->slots[0], struct btrfs_extent_item);
791 *refs = btrfs_extent_refs(l, item);
793 btrfs_free_path(path);
797 u32 btrfs_count_snapshots_in_path(struct btrfs_root *root,
798 struct btrfs_path *count_path,
802 struct btrfs_root *extent_root = root->fs_info->extent_root;
803 struct btrfs_path *path;
807 u64 root_objectid = root->root_key.objectid;
813 struct btrfs_key key;
814 struct btrfs_key found_key;
815 struct extent_buffer *l;
816 struct btrfs_extent_item *item;
817 struct btrfs_extent_ref *ref_item;
820 /* FIXME, needs locking */
823 mutex_lock(&root->fs_info->alloc_mutex);
824 path = btrfs_alloc_path();
827 bytenr = first_extent;
829 bytenr = count_path->nodes[level]->start;
832 key.objectid = bytenr;
835 btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
836 ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
842 btrfs_item_key_to_cpu(l, &found_key, path->slots[0]);
844 if (found_key.objectid != bytenr ||
845 found_key.type != BTRFS_EXTENT_ITEM_KEY) {
849 item = btrfs_item_ptr(l, path->slots[0], struct btrfs_extent_item);
850 extent_refs = btrfs_extent_refs(l, item);
853 nritems = btrfs_header_nritems(l);
854 if (path->slots[0] >= nritems) {
855 ret = btrfs_next_leaf(extent_root, path);
860 btrfs_item_key_to_cpu(l, &found_key, path->slots[0]);
861 if (found_key.objectid != bytenr)
864 if (found_key.type != BTRFS_EXTENT_REF_KEY) {
870 ref_item = btrfs_item_ptr(l, path->slots[0],
871 struct btrfs_extent_ref);
872 found_objectid = btrfs_ref_root(l, ref_item);
874 if (found_objectid != root_objectid) {
879 found_owner = btrfs_ref_objectid(l, ref_item);
880 if (found_owner != expected_owner) {
885 * nasty. we don't count a reference held by
886 * the running transaction. This allows nodatacow
887 * to avoid cow most of the time
889 if (found_owner >= BTRFS_FIRST_FREE_OBJECTID &&
890 btrfs_ref_generation(l, ref_item) ==
891 root->fs_info->generation) {
899 * if there is more than one reference against a data extent,
900 * we have to assume the other ref is another snapshot
902 if (level == -1 && extent_refs > 1) {
906 if (cur_count == 0) {
910 if (level >= 0 && root->node == count_path->nodes[level])
913 btrfs_release_path(root, path);
917 btrfs_free_path(path);
918 mutex_unlock(&root->fs_info->alloc_mutex);
922 int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
923 struct extent_buffer *buf)
927 struct btrfs_key key;
928 struct btrfs_file_extent_item *fi;
937 level = btrfs_header_level(buf);
938 nritems = btrfs_header_nritems(buf);
939 for (i = 0; i < nritems; i++) {
943 btrfs_item_key_to_cpu(buf, &key, i);
944 if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
946 fi = btrfs_item_ptr(buf, i,
947 struct btrfs_file_extent_item);
948 if (btrfs_file_extent_type(buf, fi) ==
949 BTRFS_FILE_EXTENT_INLINE)
951 disk_bytenr = btrfs_file_extent_disk_bytenr(buf, fi);
952 if (disk_bytenr == 0)
955 mutex_lock(&root->fs_info->alloc_mutex);
956 ret = __btrfs_inc_extent_ref(trans, root, disk_bytenr,
957 btrfs_file_extent_disk_num_bytes(buf, fi),
958 root->root_key.objectid, trans->transid,
959 key.objectid, key.offset);
960 mutex_unlock(&root->fs_info->alloc_mutex);
967 bytenr = btrfs_node_blockptr(buf, i);
968 btrfs_node_key_to_cpu(buf, &key, i);
970 mutex_lock(&root->fs_info->alloc_mutex);
971 ret = __btrfs_inc_extent_ref(trans, root, bytenr,
972 btrfs_level_size(root, level - 1),
973 root->root_key.objectid,
975 level - 1, key.objectid);
976 mutex_unlock(&root->fs_info->alloc_mutex);
988 for (i =0; i < faili; i++) {
991 btrfs_item_key_to_cpu(buf, &key, i);
992 if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
994 fi = btrfs_item_ptr(buf, i,
995 struct btrfs_file_extent_item);
996 if (btrfs_file_extent_type(buf, fi) ==
997 BTRFS_FILE_EXTENT_INLINE)
999 disk_bytenr = btrfs_file_extent_disk_bytenr(buf, fi);
1000 if (disk_bytenr == 0)
1002 err = btrfs_free_extent(trans, root, disk_bytenr,
1003 btrfs_file_extent_disk_num_bytes(buf,
1007 bytenr = btrfs_node_blockptr(buf, i);
1008 err = btrfs_free_extent(trans, root, bytenr,
1009 btrfs_level_size(root, level - 1), 0);
1017 static int write_one_cache_group(struct btrfs_trans_handle *trans,
1018 struct btrfs_root *root,
1019 struct btrfs_path *path,
1020 struct btrfs_block_group_cache *cache)
1024 struct btrfs_root *extent_root = root->fs_info->extent_root;
1026 struct extent_buffer *leaf;
1028 ret = btrfs_search_slot(trans, extent_root, &cache->key, path, 0, 1);
1033 leaf = path->nodes[0];
1034 bi = btrfs_item_ptr_offset(leaf, path->slots[0]);
1035 write_extent_buffer(leaf, &cache->item, bi, sizeof(cache->item));
1036 btrfs_mark_buffer_dirty(leaf);
1037 btrfs_release_path(extent_root, path);
1039 finish_current_insert(trans, extent_root);
1040 pending_ret = del_pending_extents(trans, extent_root);
1049 int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans,
1050 struct btrfs_root *root)
1052 struct extent_io_tree *block_group_cache;
1053 struct btrfs_block_group_cache *cache;
1057 struct btrfs_path *path;
1063 block_group_cache = &root->fs_info->block_group_cache;
1064 path = btrfs_alloc_path();
1068 mutex_lock(&root->fs_info->alloc_mutex);
1070 ret = find_first_extent_bit(block_group_cache, last,
1071 &start, &end, BLOCK_GROUP_DIRTY);
1076 ret = get_state_private(block_group_cache, start, &ptr);
1079 cache = (struct btrfs_block_group_cache *)(unsigned long)ptr;
1080 err = write_one_cache_group(trans, root,
1083 * if we fail to write the cache group, we want
1084 * to keep it marked dirty in hopes that a later
1091 clear_extent_bits(block_group_cache, start, end,
1092 BLOCK_GROUP_DIRTY, GFP_NOFS);
1094 btrfs_free_path(path);
1095 mutex_unlock(&root->fs_info->alloc_mutex);
1099 static struct btrfs_space_info *__find_space_info(struct btrfs_fs_info *info,
1102 struct list_head *head = &info->space_info;
1103 struct list_head *cur;
1104 struct btrfs_space_info *found;
1105 list_for_each(cur, head) {
1106 found = list_entry(cur, struct btrfs_space_info, list);
1107 if (found->flags == flags)
1114 static int update_space_info(struct btrfs_fs_info *info, u64 flags,
1115 u64 total_bytes, u64 bytes_used,
1116 struct btrfs_space_info **space_info)
1118 struct btrfs_space_info *found;
1120 found = __find_space_info(info, flags);
1122 found->total_bytes += total_bytes;
1123 found->bytes_used += bytes_used;
1125 WARN_ON(found->total_bytes < found->bytes_used);
1126 *space_info = found;
1129 found = kmalloc(sizeof(*found), GFP_NOFS);
1133 list_add(&found->list, &info->space_info);
1134 found->flags = flags;
1135 found->total_bytes = total_bytes;
1136 found->bytes_used = bytes_used;
1137 found->bytes_pinned = 0;
1139 found->force_alloc = 0;
1140 *space_info = found;
1144 static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
1146 u64 extra_flags = flags & (BTRFS_BLOCK_GROUP_RAID0 |
1147 BTRFS_BLOCK_GROUP_RAID1 |
1148 BTRFS_BLOCK_GROUP_RAID10 |
1149 BTRFS_BLOCK_GROUP_DUP);
1151 if (flags & BTRFS_BLOCK_GROUP_DATA)
1152 fs_info->avail_data_alloc_bits |= extra_flags;
1153 if (flags & BTRFS_BLOCK_GROUP_METADATA)
1154 fs_info->avail_metadata_alloc_bits |= extra_flags;
1155 if (flags & BTRFS_BLOCK_GROUP_SYSTEM)
1156 fs_info->avail_system_alloc_bits |= extra_flags;
1160 static u64 reduce_alloc_profile(struct btrfs_root *root, u64 flags)
1162 u64 num_devices = root->fs_info->fs_devices->num_devices;
1164 if (num_devices == 1)
1165 flags &= ~(BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID0);
1166 if (num_devices < 4)
1167 flags &= ~BTRFS_BLOCK_GROUP_RAID10;
1169 if ((flags & BTRFS_BLOCK_GROUP_DUP) &&
1170 (flags & (BTRFS_BLOCK_GROUP_RAID1 |
1171 BTRFS_BLOCK_GROUP_RAID10))) {
1172 flags &= ~BTRFS_BLOCK_GROUP_DUP;
1175 if ((flags & BTRFS_BLOCK_GROUP_RAID1) &&
1176 (flags & BTRFS_BLOCK_GROUP_RAID10)) {
1177 flags &= ~BTRFS_BLOCK_GROUP_RAID1;
1180 if ((flags & BTRFS_BLOCK_GROUP_RAID0) &&
1181 ((flags & BTRFS_BLOCK_GROUP_RAID1) |
1182 (flags & BTRFS_BLOCK_GROUP_RAID10) |
1183 (flags & BTRFS_BLOCK_GROUP_DUP)))
1184 flags &= ~BTRFS_BLOCK_GROUP_RAID0;
1188 static int do_chunk_alloc(struct btrfs_trans_handle *trans,
1189 struct btrfs_root *extent_root, u64 alloc_bytes,
1190 u64 flags, int force)
1192 struct btrfs_space_info *space_info;
1198 flags = reduce_alloc_profile(extent_root, flags);
1200 space_info = __find_space_info(extent_root->fs_info, flags);
1202 ret = update_space_info(extent_root->fs_info, flags,
1206 BUG_ON(!space_info);
1208 if (space_info->force_alloc) {
1210 space_info->force_alloc = 0;
1212 if (space_info->full)
1215 thresh = div_factor(space_info->total_bytes, 6);
1217 (space_info->bytes_used + space_info->bytes_pinned + alloc_bytes) <
1221 mutex_lock(&extent_root->fs_info->chunk_mutex);
1222 ret = btrfs_alloc_chunk(trans, extent_root, &start, &num_bytes, flags);
1223 if (ret == -ENOSPC) {
1224 printk("space info full %Lu\n", flags);
1225 space_info->full = 1;
1230 ret = btrfs_make_block_group(trans, extent_root, 0, flags,
1231 BTRFS_FIRST_CHUNK_TREE_OBJECTID, start, num_bytes);
1234 mutex_unlock(&extent_root->fs_info->chunk_mutex);
1239 static int update_block_group(struct btrfs_trans_handle *trans,
1240 struct btrfs_root *root,
1241 u64 bytenr, u64 num_bytes, int alloc,
1244 struct btrfs_block_group_cache *cache;
1245 struct btrfs_fs_info *info = root->fs_info;
1246 u64 total = num_bytes;
1252 WARN_ON(!mutex_is_locked(&root->fs_info->alloc_mutex));
1254 cache = btrfs_lookup_block_group(info, bytenr);
1258 byte_in_group = bytenr - cache->key.objectid;
1259 WARN_ON(byte_in_group > cache->key.offset);
1260 start = cache->key.objectid;
1261 end = start + cache->key.offset - 1;
1262 set_extent_bits(&info->block_group_cache, start, end,
1263 BLOCK_GROUP_DIRTY, GFP_NOFS);
1265 old_val = btrfs_block_group_used(&cache->item);
1266 num_bytes = min(total, cache->key.offset - byte_in_group);
1268 old_val += num_bytes;
1269 cache->space_info->bytes_used += num_bytes;
1271 old_val -= num_bytes;
1272 cache->space_info->bytes_used -= num_bytes;
1274 set_extent_dirty(&info->free_space_cache,
1275 bytenr, bytenr + num_bytes - 1,
1279 btrfs_set_block_group_used(&cache->item, old_val);
1281 bytenr += num_bytes;
1286 static u64 first_logical_byte(struct btrfs_root *root, u64 search_start)
1291 ret = find_first_extent_bit(&root->fs_info->block_group_cache,
1292 search_start, &start, &end,
1293 BLOCK_GROUP_DATA | BLOCK_GROUP_METADATA |
1294 BLOCK_GROUP_SYSTEM);
1301 static int update_pinned_extents(struct btrfs_root *root,
1302 u64 bytenr, u64 num, int pin)
1305 struct btrfs_block_group_cache *cache;
1306 struct btrfs_fs_info *fs_info = root->fs_info;
1308 WARN_ON(!mutex_is_locked(&root->fs_info->alloc_mutex));
1310 set_extent_dirty(&fs_info->pinned_extents,
1311 bytenr, bytenr + num - 1, GFP_NOFS);
1313 clear_extent_dirty(&fs_info->pinned_extents,
1314 bytenr, bytenr + num - 1, GFP_NOFS);
1317 cache = btrfs_lookup_block_group(fs_info, bytenr);
1319 u64 first = first_logical_byte(root, bytenr);
1320 WARN_ON(first < bytenr);
1321 len = min(first - bytenr, num);
1323 len = min(num, cache->key.offset -
1324 (bytenr - cache->key.objectid));
1328 cache->pinned += len;
1329 cache->space_info->bytes_pinned += len;
1331 fs_info->total_pinned += len;
1334 cache->pinned -= len;
1335 cache->space_info->bytes_pinned -= len;
1337 fs_info->total_pinned -= len;
1345 int btrfs_copy_pinned(struct btrfs_root *root, struct extent_io_tree *copy)
1350 struct extent_io_tree *pinned_extents = &root->fs_info->pinned_extents;
1354 ret = find_first_extent_bit(pinned_extents, last,
1355 &start, &end, EXTENT_DIRTY);
1358 set_extent_dirty(copy, start, end, GFP_NOFS);
1364 int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
1365 struct btrfs_root *root,
1366 struct extent_io_tree *unpin)
1371 struct extent_io_tree *free_space_cache;
1372 free_space_cache = &root->fs_info->free_space_cache;
1374 mutex_lock(&root->fs_info->alloc_mutex);
1376 ret = find_first_extent_bit(unpin, 0, &start, &end,
1380 update_pinned_extents(root, start, end + 1 - start, 0);
1381 clear_extent_dirty(unpin, start, end, GFP_NOFS);
1382 set_extent_dirty(free_space_cache, start, end, GFP_NOFS);
1384 mutex_unlock(&root->fs_info->alloc_mutex);
1388 static int finish_current_insert(struct btrfs_trans_handle *trans,
1389 struct btrfs_root *extent_root)
1393 struct btrfs_fs_info *info = extent_root->fs_info;
1394 struct extent_buffer *eb;
1395 struct btrfs_path *path;
1396 struct btrfs_key ins;
1397 struct btrfs_disk_key first;
1398 struct btrfs_extent_item extent_item;
1403 WARN_ON(!mutex_is_locked(&extent_root->fs_info->alloc_mutex));
1404 btrfs_set_stack_extent_refs(&extent_item, 1);
1405 btrfs_set_key_type(&ins, BTRFS_EXTENT_ITEM_KEY);
1406 path = btrfs_alloc_path();
1409 ret = find_first_extent_bit(&info->extent_ins, 0, &start,
1410 &end, EXTENT_LOCKED);
1414 ins.objectid = start;
1415 ins.offset = end + 1 - start;
1416 err = btrfs_insert_item(trans, extent_root, &ins,
1417 &extent_item, sizeof(extent_item));
1418 clear_extent_bits(&info->extent_ins, start, end, EXTENT_LOCKED,
1420 eb = read_tree_block(extent_root, ins.objectid, ins.offset,
1422 btrfs_tree_lock(eb);
1423 level = btrfs_header_level(eb);
1425 btrfs_item_key(eb, &first, 0);
1427 btrfs_node_key(eb, &first, 0);
1429 btrfs_tree_unlock(eb);
1430 free_extent_buffer(eb);
1432 * the first key is just a hint, so the race we've created
1433 * against reading it is fine
1435 err = btrfs_insert_extent_backref(trans, extent_root, path,
1436 start, extent_root->root_key.objectid,
1438 btrfs_disk_key_objectid(&first));
1441 btrfs_free_path(path);
1445 static int pin_down_bytes(struct btrfs_root *root, u64 bytenr, u32 num_bytes,
1450 WARN_ON(!mutex_is_locked(&root->fs_info->alloc_mutex));
1452 struct extent_buffer *buf;
1453 buf = btrfs_find_tree_block(root, bytenr, num_bytes);
1455 if (btrfs_try_tree_lock(buf) &&
1456 btrfs_buffer_uptodate(buf, 0)) {
1458 root->fs_info->running_transaction->transid;
1459 u64 header_transid =
1460 btrfs_header_generation(buf);
1461 if (header_transid == transid &&
1462 !btrfs_header_flag(buf,
1463 BTRFS_HEADER_FLAG_WRITTEN)) {
1464 clean_tree_block(NULL, root, buf);
1465 btrfs_tree_unlock(buf);
1466 free_extent_buffer(buf);
1469 btrfs_tree_unlock(buf);
1471 free_extent_buffer(buf);
1473 update_pinned_extents(root, bytenr, num_bytes, 1);
1475 set_extent_bits(&root->fs_info->pending_del,
1476 bytenr, bytenr + num_bytes - 1,
1477 EXTENT_LOCKED, GFP_NOFS);
1484 * remove an extent from the root, returns 0 on success
1486 static int __free_extent(struct btrfs_trans_handle *trans, struct btrfs_root
1487 *root, u64 bytenr, u64 num_bytes,
1488 u64 root_objectid, u64 ref_generation,
1489 u64 owner_objectid, u64 owner_offset, int pin,
1492 struct btrfs_path *path;
1493 struct btrfs_key key;
1494 struct btrfs_fs_info *info = root->fs_info;
1495 struct btrfs_root *extent_root = info->extent_root;
1496 struct extent_buffer *leaf;
1498 int extent_slot = 0;
1499 int found_extent = 0;
1501 struct btrfs_extent_item *ei;
1504 WARN_ON(!mutex_is_locked(&root->fs_info->alloc_mutex));
1505 key.objectid = bytenr;
1506 btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
1507 key.offset = num_bytes;
1508 path = btrfs_alloc_path();
1513 ret = lookup_extent_backref(trans, extent_root, path,
1514 bytenr, root_objectid,
1516 owner_objectid, owner_offset, 1);
1518 struct btrfs_key found_key;
1519 extent_slot = path->slots[0];
1520 while(extent_slot > 0) {
1522 btrfs_item_key_to_cpu(path->nodes[0], &found_key,
1524 if (found_key.objectid != bytenr)
1526 if (found_key.type == BTRFS_EXTENT_ITEM_KEY &&
1527 found_key.offset == num_bytes) {
1531 if (path->slots[0] - extent_slot > 5)
1535 ret = btrfs_del_item(trans, extent_root, path);
1537 btrfs_print_leaf(extent_root, path->nodes[0]);
1539 printk("Unable to find ref byte nr %Lu root %Lu "
1540 " gen %Lu owner %Lu offset %Lu\n", bytenr,
1541 root_objectid, ref_generation, owner_objectid,
1544 if (!found_extent) {
1545 btrfs_release_path(extent_root, path);
1546 ret = btrfs_search_slot(trans, extent_root, &key, path, -1, 1);
1550 extent_slot = path->slots[0];
1553 leaf = path->nodes[0];
1554 ei = btrfs_item_ptr(leaf, extent_slot,
1555 struct btrfs_extent_item);
1556 refs = btrfs_extent_refs(leaf, ei);
1559 btrfs_set_extent_refs(leaf, ei, refs);
1561 btrfs_mark_buffer_dirty(leaf);
1563 if (refs == 0 && found_extent && path->slots[0] == extent_slot + 1) {
1564 /* if the back ref and the extent are next to each other
1565 * they get deleted below in one shot
1567 path->slots[0] = extent_slot;
1569 } else if (found_extent) {
1570 /* otherwise delete the extent back ref */
1571 ret = btrfs_del_item(trans, extent_root, path);
1573 /* if refs are 0, we need to setup the path for deletion */
1575 btrfs_release_path(extent_root, path);
1576 ret = btrfs_search_slot(trans, extent_root, &key, path,
1589 ret = pin_down_bytes(root, bytenr, num_bytes, 0);
1595 /* block accounting for super block */
1596 spin_lock_irq(&info->delalloc_lock);
1597 super_used = btrfs_super_bytes_used(&info->super_copy);
1598 btrfs_set_super_bytes_used(&info->super_copy,
1599 super_used - num_bytes);
1600 spin_unlock_irq(&info->delalloc_lock);
1602 /* block accounting for root item */
1603 root_used = btrfs_root_used(&root->root_item);
1604 btrfs_set_root_used(&root->root_item,
1605 root_used - num_bytes);
1606 ret = btrfs_del_items(trans, extent_root, path, path->slots[0],
1611 ret = update_block_group(trans, root, bytenr, num_bytes, 0,
1615 btrfs_free_path(path);
1616 finish_current_insert(trans, extent_root);
1621 * find all the blocks marked as pending in the radix tree and remove
1622 * them from the extent map
1624 static int del_pending_extents(struct btrfs_trans_handle *trans, struct
1625 btrfs_root *extent_root)
1631 struct extent_io_tree *pending_del;
1632 struct extent_io_tree *pinned_extents;
1634 WARN_ON(!mutex_is_locked(&extent_root->fs_info->alloc_mutex));
1635 pending_del = &extent_root->fs_info->pending_del;
1636 pinned_extents = &extent_root->fs_info->pinned_extents;
1639 ret = find_first_extent_bit(pending_del, 0, &start, &end,
1643 update_pinned_extents(extent_root, start, end + 1 - start, 1);
1644 clear_extent_bits(pending_del, start, end, EXTENT_LOCKED,
1646 ret = __free_extent(trans, extent_root,
1647 start, end + 1 - start,
1648 extent_root->root_key.objectid,
1657 * remove an extent from the root, returns 0 on success
1659 static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
1660 struct btrfs_root *root, u64 bytenr,
1661 u64 num_bytes, u64 root_objectid,
1662 u64 ref_generation, u64 owner_objectid,
1663 u64 owner_offset, int pin)
1665 struct btrfs_root *extent_root = root->fs_info->extent_root;
1669 WARN_ON(num_bytes < root->sectorsize);
1670 if (!root->ref_cows)
1673 if (root == extent_root) {
1674 pin_down_bytes(root, bytenr, num_bytes, 1);
1677 ret = __free_extent(trans, root, bytenr, num_bytes, root_objectid,
1678 ref_generation, owner_objectid, owner_offset,
1681 finish_current_insert(trans, root->fs_info->extent_root);
1682 pending_ret = del_pending_extents(trans, root->fs_info->extent_root);
1683 return ret ? ret : pending_ret;
1686 int btrfs_free_extent(struct btrfs_trans_handle *trans,
1687 struct btrfs_root *root, u64 bytenr,
1688 u64 num_bytes, u64 root_objectid,
1689 u64 ref_generation, u64 owner_objectid,
1690 u64 owner_offset, int pin)
1694 maybe_lock_mutex(root);
1695 ret = __btrfs_free_extent(trans, root, bytenr, num_bytes,
1696 root_objectid, ref_generation,
1697 owner_objectid, owner_offset, pin);
1698 maybe_unlock_mutex(root);
1702 static u64 stripe_align(struct btrfs_root *root, u64 val)
1704 u64 mask = ((u64)root->stripesize - 1);
1705 u64 ret = (val + mask) & ~mask;
1710 * walks the btree of allocated extents and find a hole of a given size.
1711 * The key ins is changed to record the hole:
1712 * ins->objectid == block start
1713 * ins->flags = BTRFS_EXTENT_ITEM_KEY
1714 * ins->offset == number of blocks
1715 * Any available blocks before search_start are skipped.
1717 static int noinline find_free_extent(struct btrfs_trans_handle *trans,
1718 struct btrfs_root *orig_root,
1719 u64 num_bytes, u64 empty_size,
1720 u64 search_start, u64 search_end,
1721 u64 hint_byte, struct btrfs_key *ins,
1722 u64 exclude_start, u64 exclude_nr,
1726 u64 orig_search_start;
1727 struct btrfs_root * root = orig_root->fs_info->extent_root;
1728 struct btrfs_fs_info *info = root->fs_info;
1729 u64 total_needed = num_bytes;
1730 u64 *last_ptr = NULL;
1731 struct btrfs_block_group_cache *block_group;
1734 int chunk_alloc_done = 0;
1735 int empty_cluster = 2 * 1024 * 1024;
1736 int allowed_chunk_alloc = 0;
1738 WARN_ON(num_bytes < root->sectorsize);
1739 btrfs_set_key_type(ins, BTRFS_EXTENT_ITEM_KEY);
1741 if (orig_root->ref_cows || empty_size)
1742 allowed_chunk_alloc = 1;
1744 if (data & BTRFS_BLOCK_GROUP_METADATA) {
1745 last_ptr = &root->fs_info->last_alloc;
1746 empty_cluster = 256 * 1024;
1749 if ((data & BTRFS_BLOCK_GROUP_DATA) && btrfs_test_opt(root, SSD)) {
1750 last_ptr = &root->fs_info->last_data_alloc;
1755 hint_byte = *last_ptr;
1757 empty_size += empty_cluster;
1761 search_start = max(search_start, first_logical_byte(root, 0));
1762 orig_search_start = search_start;
1764 if (search_end == (u64)-1)
1765 search_end = btrfs_super_total_bytes(&info->super_copy);
1768 block_group = btrfs_lookup_first_block_group(info, hint_byte);
1770 hint_byte = search_start;
1771 block_group = __btrfs_find_block_group(root, block_group,
1772 hint_byte, data, 1);
1773 if (last_ptr && *last_ptr == 0 && block_group)
1774 hint_byte = block_group->key.objectid;
1776 block_group = __btrfs_find_block_group(root,
1778 search_start, data, 1);
1780 search_start = max(search_start, hint_byte);
1782 total_needed += empty_size;
1786 block_group = btrfs_lookup_first_block_group(info,
1789 block_group = btrfs_lookup_first_block_group(info,
1792 if (full_scan && !chunk_alloc_done) {
1793 if (allowed_chunk_alloc) {
1794 do_chunk_alloc(trans, root,
1795 num_bytes + 2 * 1024 * 1024, data, 1);
1796 allowed_chunk_alloc = 0;
1797 } else if (block_group && block_group_bits(block_group, data)) {
1798 block_group->space_info->force_alloc = 1;
1800 chunk_alloc_done = 1;
1802 ret = find_search_start(root, &block_group, &search_start,
1803 total_needed, data);
1804 if (ret == -ENOSPC && last_ptr && *last_ptr) {
1806 block_group = btrfs_lookup_first_block_group(info,
1808 search_start = orig_search_start;
1809 ret = find_search_start(root, &block_group, &search_start,
1810 total_needed, data);
1817 if (last_ptr && *last_ptr && search_start != *last_ptr) {
1820 empty_size += empty_cluster;
1821 total_needed += empty_size;
1823 block_group = btrfs_lookup_first_block_group(info,
1825 search_start = orig_search_start;
1826 ret = find_search_start(root, &block_group,
1827 &search_start, total_needed, data);
1834 search_start = stripe_align(root, search_start);
1835 ins->objectid = search_start;
1836 ins->offset = num_bytes;
1838 if (ins->objectid + num_bytes >= search_end)
1841 if (ins->objectid + num_bytes >
1842 block_group->key.objectid + block_group->key.offset) {
1843 search_start = block_group->key.objectid +
1844 block_group->key.offset;
1848 if (test_range_bit(&info->extent_ins, ins->objectid,
1849 ins->objectid + num_bytes -1, EXTENT_LOCKED, 0)) {
1850 search_start = ins->objectid + num_bytes;
1854 if (test_range_bit(&info->pinned_extents, ins->objectid,
1855 ins->objectid + num_bytes -1, EXTENT_DIRTY, 0)) {
1856 search_start = ins->objectid + num_bytes;
1860 if (exclude_nr > 0 && (ins->objectid + num_bytes > exclude_start &&
1861 ins->objectid < exclude_start + exclude_nr)) {
1862 search_start = exclude_start + exclude_nr;
1866 if (!(data & BTRFS_BLOCK_GROUP_DATA)) {
1867 block_group = btrfs_lookup_block_group(info, ins->objectid);
1869 trans->block_group = block_group;
1871 ins->offset = num_bytes;
1873 *last_ptr = ins->objectid + ins->offset;
1875 btrfs_super_total_bytes(&root->fs_info->super_copy)) {
1882 if (search_start + num_bytes >= search_end) {
1884 search_start = orig_search_start;
1891 total_needed -= empty_size;
1896 block_group = btrfs_lookup_first_block_group(info, search_start);
1898 block_group = __btrfs_find_block_group(root, block_group,
1899 search_start, data, 0);
1906 static int __btrfs_reserve_extent(struct btrfs_trans_handle *trans,
1907 struct btrfs_root *root,
1908 u64 num_bytes, u64 min_alloc_size,
1909 u64 empty_size, u64 hint_byte,
1910 u64 search_end, struct btrfs_key *ins,
1914 u64 search_start = 0;
1916 struct btrfs_fs_info *info = root->fs_info;
1919 alloc_profile = info->avail_data_alloc_bits &
1920 info->data_alloc_profile;
1921 data = BTRFS_BLOCK_GROUP_DATA | alloc_profile;
1922 } else if (root == root->fs_info->chunk_root) {
1923 alloc_profile = info->avail_system_alloc_bits &
1924 info->system_alloc_profile;
1925 data = BTRFS_BLOCK_GROUP_SYSTEM | alloc_profile;
1927 alloc_profile = info->avail_metadata_alloc_bits &
1928 info->metadata_alloc_profile;
1929 data = BTRFS_BLOCK_GROUP_METADATA | alloc_profile;
1932 data = reduce_alloc_profile(root, data);
1934 * the only place that sets empty_size is btrfs_realloc_node, which
1935 * is not called recursively on allocations
1937 if (empty_size || root->ref_cows) {
1938 if (!(data & BTRFS_BLOCK_GROUP_METADATA)) {
1939 ret = do_chunk_alloc(trans, root->fs_info->extent_root,
1941 BTRFS_BLOCK_GROUP_METADATA |
1942 (info->metadata_alloc_profile &
1943 info->avail_metadata_alloc_bits), 0);
1946 ret = do_chunk_alloc(trans, root->fs_info->extent_root,
1947 num_bytes + 2 * 1024 * 1024, data, 0);
1951 WARN_ON(num_bytes < root->sectorsize);
1952 ret = find_free_extent(trans, root, num_bytes, empty_size,
1953 search_start, search_end, hint_byte, ins,
1954 trans->alloc_exclude_start,
1955 trans->alloc_exclude_nr, data);
1957 if (ret == -ENOSPC && num_bytes > min_alloc_size) {
1958 num_bytes = num_bytes >> 1;
1959 num_bytes = max(num_bytes, min_alloc_size);
1960 do_chunk_alloc(trans, root->fs_info->extent_root,
1961 num_bytes, data, 1);
1965 printk("allocation failed flags %Lu\n", data);
1968 clear_extent_dirty(&root->fs_info->free_space_cache,
1969 ins->objectid, ins->objectid + ins->offset - 1,
1974 int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
1975 struct btrfs_root *root,
1976 u64 num_bytes, u64 min_alloc_size,
1977 u64 empty_size, u64 hint_byte,
1978 u64 search_end, struct btrfs_key *ins,
1982 maybe_lock_mutex(root);
1983 ret = __btrfs_reserve_extent(trans, root, num_bytes, min_alloc_size,
1984 empty_size, hint_byte, search_end, ins,
1986 maybe_unlock_mutex(root);
1990 static int __btrfs_alloc_reserved_extent(struct btrfs_trans_handle *trans,
1991 struct btrfs_root *root,
1992 u64 root_objectid, u64 ref_generation,
1993 u64 owner, u64 owner_offset,
1994 struct btrfs_key *ins)
2000 u64 num_bytes = ins->offset;
2002 struct btrfs_fs_info *info = root->fs_info;
2003 struct btrfs_root *extent_root = info->extent_root;
2004 struct btrfs_extent_item *extent_item;
2005 struct btrfs_extent_ref *ref;
2006 struct btrfs_path *path;
2007 struct btrfs_key keys[2];
2009 /* block accounting for super block */
2010 spin_lock_irq(&info->delalloc_lock);
2011 super_used = btrfs_super_bytes_used(&info->super_copy);
2012 btrfs_set_super_bytes_used(&info->super_copy, super_used + num_bytes);
2013 spin_unlock_irq(&info->delalloc_lock);
2015 /* block accounting for root item */
2016 root_used = btrfs_root_used(&root->root_item);
2017 btrfs_set_root_used(&root->root_item, root_used + num_bytes);
2019 if (root == extent_root) {
2020 set_extent_bits(&root->fs_info->extent_ins, ins->objectid,
2021 ins->objectid + ins->offset - 1,
2022 EXTENT_LOCKED, GFP_NOFS);
2026 memcpy(&keys[0], ins, sizeof(*ins));
2027 keys[1].offset = hash_extent_ref(root_objectid, ref_generation,
2028 owner, owner_offset);
2029 keys[1].objectid = ins->objectid;
2030 keys[1].type = BTRFS_EXTENT_REF_KEY;
2031 sizes[0] = sizeof(*extent_item);
2032 sizes[1] = sizeof(*ref);
2034 path = btrfs_alloc_path();
2037 ret = btrfs_insert_empty_items(trans, extent_root, path, keys,
2041 extent_item = btrfs_item_ptr(path->nodes[0], path->slots[0],
2042 struct btrfs_extent_item);
2043 btrfs_set_extent_refs(path->nodes[0], extent_item, 1);
2044 ref = btrfs_item_ptr(path->nodes[0], path->slots[0] + 1,
2045 struct btrfs_extent_ref);
2047 btrfs_set_ref_root(path->nodes[0], ref, root_objectid);
2048 btrfs_set_ref_generation(path->nodes[0], ref, ref_generation);
2049 btrfs_set_ref_objectid(path->nodes[0], ref, owner);
2050 btrfs_set_ref_offset(path->nodes[0], ref, owner_offset);
2052 btrfs_mark_buffer_dirty(path->nodes[0]);
2054 trans->alloc_exclude_start = 0;
2055 trans->alloc_exclude_nr = 0;
2056 btrfs_free_path(path);
2057 finish_current_insert(trans, extent_root);
2058 pending_ret = del_pending_extents(trans, extent_root);
2068 ret = update_block_group(trans, root, ins->objectid, ins->offset, 1, 0);
2070 printk("update block group failed for %Lu %Lu\n",
2071 ins->objectid, ins->offset);
2078 int btrfs_alloc_reserved_extent(struct btrfs_trans_handle *trans,
2079 struct btrfs_root *root,
2080 u64 root_objectid, u64 ref_generation,
2081 u64 owner, u64 owner_offset,
2082 struct btrfs_key *ins)
2085 maybe_lock_mutex(root);
2086 ret = __btrfs_alloc_reserved_extent(trans, root, root_objectid,
2087 ref_generation, owner,
2089 maybe_unlock_mutex(root);
2093 * finds a free extent and does all the dirty work required for allocation
2094 * returns the key for the extent through ins, and a tree buffer for
2095 * the first block of the extent through buf.
2097 * returns 0 if everything worked, non-zero otherwise.
2099 int btrfs_alloc_extent(struct btrfs_trans_handle *trans,
2100 struct btrfs_root *root,
2101 u64 num_bytes, u64 min_alloc_size,
2102 u64 root_objectid, u64 ref_generation,
2103 u64 owner, u64 owner_offset,
2104 u64 empty_size, u64 hint_byte,
2105 u64 search_end, struct btrfs_key *ins, u64 data)
2109 maybe_lock_mutex(root);
2111 ret = __btrfs_reserve_extent(trans, root, num_bytes,
2112 min_alloc_size, empty_size, hint_byte,
2113 search_end, ins, data);
2115 ret = __btrfs_alloc_reserved_extent(trans, root, root_objectid,
2116 ref_generation, owner,
2120 maybe_unlock_mutex(root);
2124 * helper function to allocate a block for a given tree
2125 * returns the tree buffer or NULL.
2127 struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
2128 struct btrfs_root *root,
2137 struct btrfs_key ins;
2139 struct extent_buffer *buf;
2141 ret = btrfs_alloc_extent(trans, root, blocksize, blocksize,
2142 root_objectid, ref_generation,
2143 level, first_objectid, empty_size, hint,
2147 return ERR_PTR(ret);
2149 buf = btrfs_find_create_tree_block(root, ins.objectid, blocksize);
2151 btrfs_free_extent(trans, root, ins.objectid, blocksize,
2152 root->root_key.objectid, ref_generation,
2154 return ERR_PTR(-ENOMEM);
2156 btrfs_set_header_generation(buf, trans->transid);
2157 btrfs_tree_lock(buf);
2158 clean_tree_block(trans, root, buf);
2159 btrfs_set_buffer_uptodate(buf);
2161 if (PageDirty(buf->first_page)) {
2162 printk("page %lu dirty\n", buf->first_page->index);
2166 set_extent_dirty(&trans->transaction->dirty_pages, buf->start,
2167 buf->start + buf->len - 1, GFP_NOFS);
2168 trans->blocks_used++;
2172 static int noinline drop_leaf_ref(struct btrfs_trans_handle *trans,
2173 struct btrfs_root *root,
2174 struct extent_buffer *leaf)
2177 u64 leaf_generation;
2178 struct btrfs_key key;
2179 struct btrfs_file_extent_item *fi;
2184 BUG_ON(!btrfs_is_leaf(leaf));
2185 nritems = btrfs_header_nritems(leaf);
2186 leaf_owner = btrfs_header_owner(leaf);
2187 leaf_generation = btrfs_header_generation(leaf);
2189 mutex_unlock(&root->fs_info->alloc_mutex);
2191 for (i = 0; i < nritems; i++) {
2195 btrfs_item_key_to_cpu(leaf, &key, i);
2196 if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
2198 fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item);
2199 if (btrfs_file_extent_type(leaf, fi) ==
2200 BTRFS_FILE_EXTENT_INLINE)
2203 * FIXME make sure to insert a trans record that
2204 * repeats the snapshot del on crash
2206 disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
2207 if (disk_bytenr == 0)
2210 mutex_lock(&root->fs_info->alloc_mutex);
2211 ret = __btrfs_free_extent(trans, root, disk_bytenr,
2212 btrfs_file_extent_disk_num_bytes(leaf, fi),
2213 leaf_owner, leaf_generation,
2214 key.objectid, key.offset, 0);
2215 mutex_unlock(&root->fs_info->alloc_mutex);
2219 mutex_lock(&root->fs_info->alloc_mutex);
2223 static void noinline reada_walk_down(struct btrfs_root *root,
2224 struct extent_buffer *node,
2237 nritems = btrfs_header_nritems(node);
2238 level = btrfs_header_level(node);
2242 for (i = slot; i < nritems && skipped < 32; i++) {
2243 bytenr = btrfs_node_blockptr(node, i);
2244 if (last && ((bytenr > last && bytenr - last > 32 * 1024) ||
2245 (last > bytenr && last - bytenr > 32 * 1024))) {
2249 blocksize = btrfs_level_size(root, level - 1);
2251 ret = lookup_extent_ref(NULL, root, bytenr,
2259 ret = readahead_tree_block(root, bytenr, blocksize,
2260 btrfs_node_ptr_generation(node, i));
2261 last = bytenr + blocksize;
2269 * we want to avoid as much random IO as we can with the alloc mutex
2270 * held, so drop the lock and do the lookup, then do it again with the
2273 int drop_snap_lookup_refcount(struct btrfs_root *root, u64 start, u64 len,
2276 mutex_unlock(&root->fs_info->alloc_mutex);
2277 lookup_extent_ref(NULL, root, start, len, refs);
2279 mutex_lock(&root->fs_info->alloc_mutex);
2280 return lookup_extent_ref(NULL, root, start, len, refs);
2284 * helper function for drop_snapshot, this walks down the tree dropping ref
2285 * counts as it goes.
2287 static int noinline walk_down_tree(struct btrfs_trans_handle *trans,
2288 struct btrfs_root *root,
2289 struct btrfs_path *path, int *level)
2295 struct extent_buffer *next;
2296 struct extent_buffer *cur;
2297 struct extent_buffer *parent;
2302 mutex_lock(&root->fs_info->alloc_mutex);
2304 WARN_ON(*level < 0);
2305 WARN_ON(*level >= BTRFS_MAX_LEVEL);
2306 ret = drop_snap_lookup_refcount(root, path->nodes[*level]->start,
2307 path->nodes[*level]->len, &refs);
2313 * walk down to the last node level and free all the leaves
2315 while(*level >= 0) {
2316 WARN_ON(*level < 0);
2317 WARN_ON(*level >= BTRFS_MAX_LEVEL);
2318 cur = path->nodes[*level];
2320 if (btrfs_header_level(cur) != *level)
2323 if (path->slots[*level] >=
2324 btrfs_header_nritems(cur))
2327 ret = drop_leaf_ref(trans, root, cur);
2331 bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
2332 ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
2333 blocksize = btrfs_level_size(root, *level - 1);
2335 ret = drop_snap_lookup_refcount(root, bytenr, blocksize, &refs);
2338 parent = path->nodes[*level];
2339 root_owner = btrfs_header_owner(parent);
2340 root_gen = btrfs_header_generation(parent);
2341 path->slots[*level]++;
2342 ret = __btrfs_free_extent(trans, root, bytenr,
2343 blocksize, root_owner,
2348 next = btrfs_find_tree_block(root, bytenr, blocksize);
2349 if (!next || !btrfs_buffer_uptodate(next, ptr_gen)) {
2350 free_extent_buffer(next);
2351 mutex_unlock(&root->fs_info->alloc_mutex);
2353 if (path->slots[*level] == 0)
2354 reada_walk_down(root, cur, path->slots[*level]);
2356 next = read_tree_block(root, bytenr, blocksize,
2359 mutex_lock(&root->fs_info->alloc_mutex);
2361 /* we've dropped the lock, double check */
2362 ret = lookup_extent_ref(NULL, root, bytenr, blocksize,
2366 parent = path->nodes[*level];
2367 root_owner = btrfs_header_owner(parent);
2368 root_gen = btrfs_header_generation(parent);
2370 path->slots[*level]++;
2371 free_extent_buffer(next);
2372 ret = __btrfs_free_extent(trans, root, bytenr,
2380 WARN_ON(*level <= 0);
2381 if (path->nodes[*level-1])
2382 free_extent_buffer(path->nodes[*level-1]);
2383 path->nodes[*level-1] = next;
2384 *level = btrfs_header_level(next);
2385 path->slots[*level] = 0;
2388 WARN_ON(*level < 0);
2389 WARN_ON(*level >= BTRFS_MAX_LEVEL);
2391 if (path->nodes[*level] == root->node) {
2392 root_owner = root->root_key.objectid;
2393 parent = path->nodes[*level];
2395 parent = path->nodes[*level + 1];
2396 root_owner = btrfs_header_owner(parent);
2399 root_gen = btrfs_header_generation(parent);
2400 ret = __btrfs_free_extent(trans, root, path->nodes[*level]->start,
2401 path->nodes[*level]->len,
2402 root_owner, root_gen, 0, 0, 1);
2403 free_extent_buffer(path->nodes[*level]);
2404 path->nodes[*level] = NULL;
2407 mutex_unlock(&root->fs_info->alloc_mutex);
2413 * helper for dropping snapshots. This walks back up the tree in the path
2414 * to find the first node higher up where we haven't yet gone through
2417 static int noinline walk_up_tree(struct btrfs_trans_handle *trans,
2418 struct btrfs_root *root,
2419 struct btrfs_path *path, int *level)
2423 struct btrfs_root_item *root_item = &root->root_item;
2428 for(i = *level; i < BTRFS_MAX_LEVEL - 1 && path->nodes[i]; i++) {
2429 slot = path->slots[i];
2430 if (slot < btrfs_header_nritems(path->nodes[i]) - 1) {
2431 struct extent_buffer *node;
2432 struct btrfs_disk_key disk_key;
2433 node = path->nodes[i];
2436 WARN_ON(*level == 0);
2437 btrfs_node_key(node, &disk_key, path->slots[i]);
2438 memcpy(&root_item->drop_progress,
2439 &disk_key, sizeof(disk_key));
2440 root_item->drop_level = i;
2443 if (path->nodes[*level] == root->node) {
2444 root_owner = root->root_key.objectid;
2446 btrfs_header_generation(path->nodes[*level]);
2448 struct extent_buffer *node;
2449 node = path->nodes[*level + 1];
2450 root_owner = btrfs_header_owner(node);
2451 root_gen = btrfs_header_generation(node);
2453 ret = btrfs_free_extent(trans, root,
2454 path->nodes[*level]->start,
2455 path->nodes[*level]->len,
2456 root_owner, root_gen, 0, 0, 1);
2458 free_extent_buffer(path->nodes[*level]);
2459 path->nodes[*level] = NULL;
2467 * drop the reference count on the tree rooted at 'snap'. This traverses
2468 * the tree freeing any blocks that have a ref count of zero after being
2471 int btrfs_drop_snapshot(struct btrfs_trans_handle *trans, struct btrfs_root
2477 struct btrfs_path *path;
2480 struct btrfs_root_item *root_item = &root->root_item;
2482 WARN_ON(!mutex_is_locked(&root->fs_info->drop_mutex));
2483 path = btrfs_alloc_path();
2486 level = btrfs_header_level(root->node);
2488 if (btrfs_disk_key_objectid(&root_item->drop_progress) == 0) {
2489 path->nodes[level] = root->node;
2490 extent_buffer_get(root->node);
2491 path->slots[level] = 0;
2493 struct btrfs_key key;
2494 struct btrfs_disk_key found_key;
2495 struct extent_buffer *node;
2497 btrfs_disk_key_to_cpu(&key, &root_item->drop_progress);
2498 level = root_item->drop_level;
2499 path->lowest_level = level;
2500 wret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
2505 node = path->nodes[level];
2506 btrfs_node_key(node, &found_key, path->slots[level]);
2507 WARN_ON(memcmp(&found_key, &root_item->drop_progress,
2508 sizeof(found_key)));
2510 * unlock our path, this is safe because only this
2511 * function is allowed to delete this snapshot
2513 for (i = 0; i < BTRFS_MAX_LEVEL; i++) {
2514 if (path->nodes[i] && path->locks[i]) {
2516 btrfs_tree_unlock(path->nodes[i]);
2521 wret = walk_down_tree(trans, root, path, &level);
2527 wret = walk_up_tree(trans, root, path, &level);
2532 if (trans->transaction->in_commit) {
2537 for (i = 0; i <= orig_level; i++) {
2538 if (path->nodes[i]) {
2539 free_extent_buffer(path->nodes[i]);
2540 path->nodes[i] = NULL;
2544 btrfs_free_path(path);
2548 int btrfs_free_block_groups(struct btrfs_fs_info *info)
2555 mutex_lock(&info->alloc_mutex);
2557 ret = find_first_extent_bit(&info->block_group_cache, 0,
2558 &start, &end, (unsigned int)-1);
2561 ret = get_state_private(&info->block_group_cache, start, &ptr);
2563 kfree((void *)(unsigned long)ptr);
2564 clear_extent_bits(&info->block_group_cache, start,
2565 end, (unsigned int)-1, GFP_NOFS);
2568 ret = find_first_extent_bit(&info->free_space_cache, 0,
2569 &start, &end, EXTENT_DIRTY);
2572 clear_extent_dirty(&info->free_space_cache, start,
2575 mutex_unlock(&info->alloc_mutex);
2579 static unsigned long calc_ra(unsigned long start, unsigned long last,
2582 return min(last, start + nr - 1);
2585 static int noinline relocate_inode_pages(struct inode *inode, u64 start,
2590 unsigned long last_index;
2593 struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
2594 struct file_ra_state *ra;
2595 unsigned long total_read = 0;
2596 unsigned long ra_pages;
2597 struct btrfs_trans_handle *trans;
2599 ra = kzalloc(sizeof(*ra), GFP_NOFS);
2601 mutex_lock(&inode->i_mutex);
2602 i = start >> PAGE_CACHE_SHIFT;
2603 last_index = (start + len - 1) >> PAGE_CACHE_SHIFT;
2605 ra_pages = BTRFS_I(inode)->root->fs_info->bdi.ra_pages;
2607 file_ra_state_init(ra, inode->i_mapping);
2609 for (; i <= last_index; i++) {
2610 if (total_read % ra_pages == 0) {
2611 btrfs_force_ra(inode->i_mapping, ra, NULL, i,
2612 calc_ra(i, last_index, ra_pages));
2615 if (((u64)i << PAGE_CACHE_SHIFT) > inode->i_size)
2616 goto truncate_racing;
2618 page = grab_cache_page(inode->i_mapping, i);
2622 if (!PageUptodate(page)) {
2623 btrfs_readpage(NULL, page);
2625 if (!PageUptodate(page)) {
2627 page_cache_release(page);
2631 #if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,18)
2632 ClearPageDirty(page);
2634 cancel_dirty_page(page, PAGE_CACHE_SIZE);
2636 wait_on_page_writeback(page);
2637 set_page_extent_mapped(page);
2638 page_start = (u64)page->index << PAGE_CACHE_SHIFT;
2639 page_end = page_start + PAGE_CACHE_SIZE - 1;
2641 lock_extent(io_tree, page_start, page_end, GFP_NOFS);
2643 set_extent_delalloc(io_tree, page_start,
2644 page_end, GFP_NOFS);
2645 set_page_dirty(page);
2647 unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
2649 page_cache_release(page);
2651 balance_dirty_pages_ratelimited_nr(inode->i_mapping,
2656 trans = btrfs_start_transaction(BTRFS_I(inode)->root, 1);
2658 btrfs_end_transaction(trans, BTRFS_I(inode)->root);
2659 mark_inode_dirty(inode);
2661 mutex_unlock(&inode->i_mutex);
2665 vmtruncate(inode, inode->i_size);
2666 balance_dirty_pages_ratelimited_nr(inode->i_mapping,
2672 * The back references tell us which tree holds a ref on a block,
2673 * but it is possible for the tree root field in the reference to
2674 * reflect the original root before a snapshot was made. In this
2675 * case we should search through all the children of a given root
2676 * to find potential holders of references on a block.
2678 * Instead, we do something a little less fancy and just search
2679 * all the roots for a given key/block combination.
2681 static int find_root_for_ref(struct btrfs_root *root,
2682 struct btrfs_path *path,
2683 struct btrfs_key *key0,
2686 struct btrfs_root **found_root,
2689 struct btrfs_key root_location;
2690 struct btrfs_root *cur_root = *found_root;
2691 struct btrfs_file_extent_item *file_extent;
2692 u64 root_search_start = BTRFS_FS_TREE_OBJECTID;
2696 root_location.offset = (u64)-1;
2697 root_location.type = BTRFS_ROOT_ITEM_KEY;
2698 path->lowest_level = level;
2701 ret = btrfs_search_slot(NULL, cur_root, key0, path, 0, 0);
2703 if (ret == 0 && file_key) {
2704 struct extent_buffer *leaf = path->nodes[0];
2705 file_extent = btrfs_item_ptr(leaf, path->slots[0],
2706 struct btrfs_file_extent_item);
2707 if (btrfs_file_extent_type(leaf, file_extent) ==
2708 BTRFS_FILE_EXTENT_REG) {
2710 btrfs_file_extent_disk_bytenr(leaf,
2713 } else if (!file_key) {
2714 if (path->nodes[level])
2715 found_bytenr = path->nodes[level]->start;
2718 btrfs_release_path(cur_root, path);
2720 if (found_bytenr == bytenr) {
2721 *found_root = cur_root;
2725 ret = btrfs_search_root(root->fs_info->tree_root,
2726 root_search_start, &root_search_start);
2730 root_location.objectid = root_search_start;
2731 cur_root = btrfs_read_fs_root_no_name(root->fs_info,
2739 path->lowest_level = 0;
2744 * note, this releases the path
2746 static int noinline relocate_one_reference(struct btrfs_root *extent_root,
2747 struct btrfs_path *path,
2748 struct btrfs_key *extent_key,
2749 u64 *last_file_objectid,
2750 u64 *last_file_offset,
2751 u64 *last_file_root,
2754 struct inode *inode;
2755 struct btrfs_root *found_root;
2756 struct btrfs_key root_location;
2757 struct btrfs_key found_key;
2758 struct btrfs_extent_ref *ref;
2766 WARN_ON(!mutex_is_locked(&extent_root->fs_info->alloc_mutex));
2768 ref = btrfs_item_ptr(path->nodes[0], path->slots[0],
2769 struct btrfs_extent_ref);
2770 ref_root = btrfs_ref_root(path->nodes[0], ref);
2771 ref_gen = btrfs_ref_generation(path->nodes[0], ref);
2772 ref_objectid = btrfs_ref_objectid(path->nodes[0], ref);
2773 ref_offset = btrfs_ref_offset(path->nodes[0], ref);
2774 btrfs_release_path(extent_root, path);
2776 root_location.objectid = ref_root;
2778 root_location.offset = 0;
2780 root_location.offset = (u64)-1;
2781 root_location.type = BTRFS_ROOT_ITEM_KEY;
2783 found_root = btrfs_read_fs_root_no_name(extent_root->fs_info,
2785 BUG_ON(!found_root);
2786 mutex_unlock(&extent_root->fs_info->alloc_mutex);
2788 if (ref_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
2789 found_key.objectid = ref_objectid;
2790 found_key.type = BTRFS_EXTENT_DATA_KEY;
2791 found_key.offset = ref_offset;
2794 if (last_extent == extent_key->objectid &&
2795 *last_file_objectid == ref_objectid &&
2796 *last_file_offset == ref_offset &&
2797 *last_file_root == ref_root)
2800 ret = find_root_for_ref(extent_root, path, &found_key,
2801 level, 1, &found_root,
2802 extent_key->objectid);
2807 if (last_extent == extent_key->objectid &&
2808 *last_file_objectid == ref_objectid &&
2809 *last_file_offset == ref_offset &&
2810 *last_file_root == ref_root)
2813 inode = btrfs_iget_locked(extent_root->fs_info->sb,
2814 ref_objectid, found_root);
2815 if (inode->i_state & I_NEW) {
2816 /* the inode and parent dir are two different roots */
2817 BTRFS_I(inode)->root = found_root;
2818 BTRFS_I(inode)->location.objectid = ref_objectid;
2819 BTRFS_I(inode)->location.type = BTRFS_INODE_ITEM_KEY;
2820 BTRFS_I(inode)->location.offset = 0;
2821 btrfs_read_locked_inode(inode);
2822 unlock_new_inode(inode);
2825 /* this can happen if the reference is not against
2826 * the latest version of the tree root
2828 if (is_bad_inode(inode))
2831 *last_file_objectid = inode->i_ino;
2832 *last_file_root = found_root->root_key.objectid;
2833 *last_file_offset = ref_offset;
2835 relocate_inode_pages(inode, ref_offset, extent_key->offset);
2838 struct btrfs_trans_handle *trans;
2839 struct extent_buffer *eb;
2842 eb = read_tree_block(found_root, extent_key->objectid,
2843 extent_key->offset, 0);
2844 btrfs_tree_lock(eb);
2845 level = btrfs_header_level(eb);
2848 btrfs_item_key_to_cpu(eb, &found_key, 0);
2850 btrfs_node_key_to_cpu(eb, &found_key, 0);
2852 btrfs_tree_unlock(eb);
2853 free_extent_buffer(eb);
2855 ret = find_root_for_ref(extent_root, path, &found_key,
2856 level, 0, &found_root,
2857 extent_key->objectid);
2863 * right here almost anything could happen to our key,
2864 * but that's ok. The cow below will either relocate it
2865 * or someone else will have relocated it. Either way,
2866 * it is in a different spot than it was before and
2870 trans = btrfs_start_transaction(found_root, 1);
2872 if (found_root == extent_root->fs_info->extent_root ||
2873 found_root == extent_root->fs_info->chunk_root ||
2874 found_root == extent_root->fs_info->dev_root) {
2876 mutex_lock(&extent_root->fs_info->alloc_mutex);
2879 path->lowest_level = level;
2881 ret = btrfs_search_slot(trans, found_root, &found_key, path,
2883 path->lowest_level = 0;
2884 btrfs_release_path(found_root, path);
2886 if (found_root == found_root->fs_info->extent_root)
2887 btrfs_extent_post_op(trans, found_root);
2889 mutex_unlock(&extent_root->fs_info->alloc_mutex);
2891 btrfs_end_transaction(trans, found_root);
2895 mutex_lock(&extent_root->fs_info->alloc_mutex);
2899 static int noinline del_extent_zero(struct btrfs_root *extent_root,
2900 struct btrfs_path *path,
2901 struct btrfs_key *extent_key)
2904 struct btrfs_trans_handle *trans;
2906 trans = btrfs_start_transaction(extent_root, 1);
2907 ret = btrfs_search_slot(trans, extent_root, extent_key, path, -1, 1);
2914 ret = btrfs_del_item(trans, extent_root, path);
2916 btrfs_end_transaction(trans, extent_root);
2920 static int noinline relocate_one_extent(struct btrfs_root *extent_root,
2921 struct btrfs_path *path,
2922 struct btrfs_key *extent_key)
2924 struct btrfs_key key;
2925 struct btrfs_key found_key;
2926 struct extent_buffer *leaf;
2927 u64 last_file_objectid = 0;
2928 u64 last_file_root = 0;
2929 u64 last_file_offset = (u64)-1;
2930 u64 last_extent = 0;
2935 if (extent_key->objectid == 0) {
2936 ret = del_extent_zero(extent_root, path, extent_key);
2939 key.objectid = extent_key->objectid;
2940 key.type = BTRFS_EXTENT_REF_KEY;
2944 ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
2950 leaf = path->nodes[0];
2951 nritems = btrfs_header_nritems(leaf);
2952 if (path->slots[0] == nritems) {
2953 ret = btrfs_next_leaf(extent_root, path);
2960 leaf = path->nodes[0];
2963 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
2964 if (found_key.objectid != extent_key->objectid) {
2968 if (found_key.type != BTRFS_EXTENT_REF_KEY) {
2972 key.offset = found_key.offset + 1;
2973 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
2975 ret = relocate_one_reference(extent_root, path, extent_key,
2976 &last_file_objectid,
2978 &last_file_root, last_extent);
2981 last_extent = extent_key->objectid;
2985 btrfs_release_path(extent_root, path);
2989 static u64 update_block_group_flags(struct btrfs_root *root, u64 flags)
2992 u64 stripped = BTRFS_BLOCK_GROUP_RAID0 |
2993 BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10;
2995 num_devices = root->fs_info->fs_devices->num_devices;
2996 if (num_devices == 1) {
2997 stripped |= BTRFS_BLOCK_GROUP_DUP;
2998 stripped = flags & ~stripped;
3000 /* turn raid0 into single device chunks */
3001 if (flags & BTRFS_BLOCK_GROUP_RAID0)
3004 /* turn mirroring into duplication */
3005 if (flags & (BTRFS_BLOCK_GROUP_RAID1 |
3006 BTRFS_BLOCK_GROUP_RAID10))
3007 return stripped | BTRFS_BLOCK_GROUP_DUP;
3010 /* they already had raid on here, just return */
3011 if (flags & stripped)
3014 stripped |= BTRFS_BLOCK_GROUP_DUP;
3015 stripped = flags & ~stripped;
3017 /* switch duplicated blocks with raid1 */
3018 if (flags & BTRFS_BLOCK_GROUP_DUP)
3019 return stripped | BTRFS_BLOCK_GROUP_RAID1;
3021 /* turn single device chunks into raid0 */
3022 return stripped | BTRFS_BLOCK_GROUP_RAID0;
3027 int __alloc_chunk_for_shrink(struct btrfs_root *root,
3028 struct btrfs_block_group_cache *shrink_block_group,
3031 struct btrfs_trans_handle *trans;
3032 u64 new_alloc_flags;
3035 if (btrfs_block_group_used(&shrink_block_group->item) > 0) {
3037 mutex_unlock(&root->fs_info->alloc_mutex);
3038 trans = btrfs_start_transaction(root, 1);
3039 mutex_lock(&root->fs_info->alloc_mutex);
3041 new_alloc_flags = update_block_group_flags(root,
3042 shrink_block_group->flags);
3043 if (new_alloc_flags != shrink_block_group->flags) {
3045 btrfs_block_group_used(&shrink_block_group->item);
3047 calc = shrink_block_group->key.offset;
3049 do_chunk_alloc(trans, root->fs_info->extent_root,
3050 calc + 2 * 1024 * 1024, new_alloc_flags, force);
3052 mutex_unlock(&root->fs_info->alloc_mutex);
3053 btrfs_end_transaction(trans, root);
3054 mutex_lock(&root->fs_info->alloc_mutex);
3059 int btrfs_shrink_extent_tree(struct btrfs_root *root, u64 shrink_start)
3061 struct btrfs_trans_handle *trans;
3062 struct btrfs_root *tree_root = root->fs_info->tree_root;
3063 struct btrfs_path *path;
3066 u64 shrink_last_byte;
3067 struct btrfs_block_group_cache *shrink_block_group;
3068 struct btrfs_fs_info *info = root->fs_info;
3069 struct btrfs_key key;
3070 struct btrfs_key found_key;
3071 struct extent_buffer *leaf;
3076 mutex_lock(&root->fs_info->alloc_mutex);
3077 shrink_block_group = btrfs_lookup_block_group(root->fs_info,
3079 BUG_ON(!shrink_block_group);
3081 shrink_last_byte = shrink_block_group->key.objectid +
3082 shrink_block_group->key.offset;
3084 shrink_block_group->space_info->total_bytes -=
3085 shrink_block_group->key.offset;
3086 path = btrfs_alloc_path();
3087 root = root->fs_info->extent_root;
3090 printk("btrfs relocating block group %llu flags %llu\n",
3091 (unsigned long long)shrink_start,
3092 (unsigned long long)shrink_block_group->flags);
3094 __alloc_chunk_for_shrink(root, shrink_block_group, 1);
3098 shrink_block_group->ro = 1;
3102 key.objectid = shrink_start;
3105 cur_byte = key.objectid;
3107 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
3111 ret = btrfs_previous_item(root, path, 0, BTRFS_EXTENT_ITEM_KEY);
3116 leaf = path->nodes[0];
3117 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
3118 if (found_key.objectid + found_key.offset > shrink_start &&
3119 found_key.objectid < shrink_last_byte) {
3120 cur_byte = found_key.objectid;
3121 key.objectid = cur_byte;
3124 btrfs_release_path(root, path);
3127 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
3132 leaf = path->nodes[0];
3133 nritems = btrfs_header_nritems(leaf);
3134 if (path->slots[0] >= nritems) {
3135 ret = btrfs_next_leaf(root, path);
3142 leaf = path->nodes[0];
3143 nritems = btrfs_header_nritems(leaf);
3146 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
3148 if (found_key.objectid >= shrink_last_byte)
3151 if (progress && need_resched()) {
3152 memcpy(&key, &found_key, sizeof(key));
3154 btrfs_release_path(root, path);
3155 btrfs_search_slot(NULL, root, &key, path, 0, 0);
3161 if (btrfs_key_type(&found_key) != BTRFS_EXTENT_ITEM_KEY ||
3162 found_key.objectid + found_key.offset <= cur_byte) {
3163 memcpy(&key, &found_key, sizeof(key));
3170 cur_byte = found_key.objectid + found_key.offset;
3171 key.objectid = cur_byte;
3172 btrfs_release_path(root, path);
3173 ret = relocate_one_extent(root, path, &found_key);
3174 __alloc_chunk_for_shrink(root, shrink_block_group, 0);
3177 btrfs_release_path(root, path);
3179 if (total_found > 0) {
3180 printk("btrfs relocate found %llu last extent was %llu\n",
3181 (unsigned long long)total_found,
3182 (unsigned long long)found_key.objectid);
3183 mutex_unlock(&root->fs_info->alloc_mutex);
3184 trans = btrfs_start_transaction(tree_root, 1);
3185 btrfs_commit_transaction(trans, tree_root);
3187 btrfs_clean_old_snapshots(tree_root);
3189 trans = btrfs_start_transaction(tree_root, 1);
3190 btrfs_commit_transaction(trans, tree_root);
3191 mutex_lock(&root->fs_info->alloc_mutex);
3196 * we've freed all the extents, now remove the block
3197 * group item from the tree
3199 mutex_unlock(&root->fs_info->alloc_mutex);
3201 trans = btrfs_start_transaction(root, 1);
3202 mutex_lock(&root->fs_info->alloc_mutex);
3203 memcpy(&key, &shrink_block_group->key, sizeof(key));
3205 ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
3211 clear_extent_bits(&info->block_group_cache, key.objectid,
3212 key.objectid + key.offset - 1,
3213 (unsigned int)-1, GFP_NOFS);
3216 clear_extent_bits(&info->free_space_cache,
3217 key.objectid, key.objectid + key.offset - 1,
3218 (unsigned int)-1, GFP_NOFS);
3220 memset(shrink_block_group, 0, sizeof(*shrink_block_group));
3221 kfree(shrink_block_group);
3223 btrfs_del_item(trans, root, path);
3224 btrfs_release_path(root, path);
3225 mutex_unlock(&root->fs_info->alloc_mutex);
3226 btrfs_commit_transaction(trans, root);
3228 mutex_lock(&root->fs_info->alloc_mutex);
3230 /* the code to unpin extents might set a few bits in the free
3231 * space cache for this range again
3233 clear_extent_bits(&info->free_space_cache,
3234 key.objectid, key.objectid + key.offset - 1,
3235 (unsigned int)-1, GFP_NOFS);
3237 btrfs_free_path(path);
3238 mutex_unlock(&root->fs_info->alloc_mutex);
3242 int find_first_block_group(struct btrfs_root *root, struct btrfs_path *path,
3243 struct btrfs_key *key)
3246 struct btrfs_key found_key;
3247 struct extent_buffer *leaf;
3250 ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
3255 slot = path->slots[0];
3256 leaf = path->nodes[0];
3257 if (slot >= btrfs_header_nritems(leaf)) {
3258 ret = btrfs_next_leaf(root, path);
3265 btrfs_item_key_to_cpu(leaf, &found_key, slot);
3267 if (found_key.objectid >= key->objectid &&
3268 found_key.type == BTRFS_BLOCK_GROUP_ITEM_KEY) {
3279 int btrfs_read_block_groups(struct btrfs_root *root)
3281 struct btrfs_path *path;
3284 struct btrfs_block_group_cache *cache;
3285 struct btrfs_fs_info *info = root->fs_info;
3286 struct btrfs_space_info *space_info;
3287 struct extent_io_tree *block_group_cache;
3288 struct btrfs_key key;
3289 struct btrfs_key found_key;
3290 struct extent_buffer *leaf;
3292 block_group_cache = &info->block_group_cache;
3293 root = info->extent_root;
3296 btrfs_set_key_type(&key, BTRFS_BLOCK_GROUP_ITEM_KEY);
3297 path = btrfs_alloc_path();
3301 mutex_lock(&root->fs_info->alloc_mutex);
3303 ret = find_first_block_group(root, path, &key);
3311 leaf = path->nodes[0];
3312 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
3313 cache = kzalloc(sizeof(*cache), GFP_NOFS);
3319 read_extent_buffer(leaf, &cache->item,
3320 btrfs_item_ptr_offset(leaf, path->slots[0]),
3321 sizeof(cache->item));
3322 memcpy(&cache->key, &found_key, sizeof(found_key));
3324 key.objectid = found_key.objectid + found_key.offset;
3325 btrfs_release_path(root, path);
3326 cache->flags = btrfs_block_group_flags(&cache->item);
3328 if (cache->flags & BTRFS_BLOCK_GROUP_DATA) {
3329 bit = BLOCK_GROUP_DATA;
3330 } else if (cache->flags & BTRFS_BLOCK_GROUP_SYSTEM) {
3331 bit = BLOCK_GROUP_SYSTEM;
3332 } else if (cache->flags & BTRFS_BLOCK_GROUP_METADATA) {
3333 bit = BLOCK_GROUP_METADATA;
3335 set_avail_alloc_bits(info, cache->flags);
3337 ret = update_space_info(info, cache->flags, found_key.offset,
3338 btrfs_block_group_used(&cache->item),
3341 cache->space_info = space_info;
3343 /* use EXTENT_LOCKED to prevent merging */
3344 set_extent_bits(block_group_cache, found_key.objectid,
3345 found_key.objectid + found_key.offset - 1,
3346 bit | EXTENT_LOCKED, GFP_NOFS);
3347 set_state_private(block_group_cache, found_key.objectid,
3348 (unsigned long)cache);
3351 btrfs_super_total_bytes(&info->super_copy))
3356 btrfs_free_path(path);
3357 mutex_unlock(&root->fs_info->alloc_mutex);
3361 int btrfs_make_block_group(struct btrfs_trans_handle *trans,
3362 struct btrfs_root *root, u64 bytes_used,
3363 u64 type, u64 chunk_objectid, u64 chunk_offset,
3368 struct btrfs_root *extent_root;
3369 struct btrfs_block_group_cache *cache;
3370 struct extent_io_tree *block_group_cache;
3372 WARN_ON(!mutex_is_locked(&root->fs_info->alloc_mutex));
3373 extent_root = root->fs_info->extent_root;
3374 block_group_cache = &root->fs_info->block_group_cache;
3376 cache = kzalloc(sizeof(*cache), GFP_NOFS);
3378 cache->key.objectid = chunk_offset;
3379 cache->key.offset = size;
3380 btrfs_set_key_type(&cache->key, BTRFS_BLOCK_GROUP_ITEM_KEY);
3382 btrfs_set_block_group_used(&cache->item, bytes_used);
3383 btrfs_set_block_group_chunk_objectid(&cache->item, chunk_objectid);
3384 cache->flags = type;
3385 btrfs_set_block_group_flags(&cache->item, type);
3387 ret = update_space_info(root->fs_info, cache->flags, size, bytes_used,
3388 &cache->space_info);
3391 bit = block_group_state_bits(type);
3392 set_extent_bits(block_group_cache, chunk_offset,
3393 chunk_offset + size - 1,
3394 bit | EXTENT_LOCKED, GFP_NOFS);
3396 set_state_private(block_group_cache, chunk_offset,
3397 (unsigned long)cache);
3398 ret = btrfs_insert_item(trans, extent_root, &cache->key, &cache->item,
3399 sizeof(cache->item));
3402 finish_current_insert(trans, extent_root);
3403 ret = del_pending_extents(trans, extent_root);
3405 set_avail_alloc_bits(extent_root->fs_info, type);