2 * Copyright (C) 2008 Red Hat. 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.
19 #include <linux/pagemap.h>
20 #include <linux/sched.h>
21 #include <linux/slab.h>
22 #include <linux/math64.h>
23 #include <linux/ratelimit.h>
25 #include "free-space-cache.h"
26 #include "transaction.h"
28 #include "extent_io.h"
29 #include "inode-map.h"
31 #define BITS_PER_BITMAP (PAGE_CACHE_SIZE * 8)
32 #define MAX_CACHE_BYTES_PER_GIG (32 * 1024)
34 static int link_free_space(struct btrfs_free_space_ctl *ctl,
35 struct btrfs_free_space *info);
37 static struct inode *__lookup_free_space_inode(struct btrfs_root *root,
38 struct btrfs_path *path,
42 struct btrfs_key location;
43 struct btrfs_disk_key disk_key;
44 struct btrfs_free_space_header *header;
45 struct extent_buffer *leaf;
46 struct inode *inode = NULL;
49 key.objectid = BTRFS_FREE_SPACE_OBJECTID;
53 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
57 btrfs_release_path(path);
58 return ERR_PTR(-ENOENT);
61 leaf = path->nodes[0];
62 header = btrfs_item_ptr(leaf, path->slots[0],
63 struct btrfs_free_space_header);
64 btrfs_free_space_key(leaf, header, &disk_key);
65 btrfs_disk_key_to_cpu(&location, &disk_key);
66 btrfs_release_path(path);
68 inode = btrfs_iget(root->fs_info->sb, &location, root, NULL);
70 return ERR_PTR(-ENOENT);
73 if (is_bad_inode(inode)) {
75 return ERR_PTR(-ENOENT);
78 inode->i_mapping->flags &= ~__GFP_FS;
83 struct inode *lookup_free_space_inode(struct btrfs_root *root,
84 struct btrfs_block_group_cache
85 *block_group, struct btrfs_path *path)
87 struct inode *inode = NULL;
88 u32 flags = BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW;
90 spin_lock(&block_group->lock);
91 if (block_group->inode)
92 inode = igrab(block_group->inode);
93 spin_unlock(&block_group->lock);
97 inode = __lookup_free_space_inode(root, path,
98 block_group->key.objectid);
102 spin_lock(&block_group->lock);
103 if (!((BTRFS_I(inode)->flags & flags) == flags)) {
104 printk(KERN_INFO "Old style space inode found, converting.\n");
105 BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM |
106 BTRFS_INODE_NODATACOW;
107 block_group->disk_cache_state = BTRFS_DC_CLEAR;
110 if (!block_group->iref) {
111 block_group->inode = igrab(inode);
112 block_group->iref = 1;
114 spin_unlock(&block_group->lock);
119 int __create_free_space_inode(struct btrfs_root *root,
120 struct btrfs_trans_handle *trans,
121 struct btrfs_path *path, u64 ino, u64 offset)
123 struct btrfs_key key;
124 struct btrfs_disk_key disk_key;
125 struct btrfs_free_space_header *header;
126 struct btrfs_inode_item *inode_item;
127 struct extent_buffer *leaf;
128 u64 flags = BTRFS_INODE_NOCOMPRESS | BTRFS_INODE_PREALLOC;
131 ret = btrfs_insert_empty_inode(trans, root, path, ino);
135 /* We inline crc's for the free disk space cache */
136 if (ino != BTRFS_FREE_INO_OBJECTID)
137 flags |= BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW;
139 leaf = path->nodes[0];
140 inode_item = btrfs_item_ptr(leaf, path->slots[0],
141 struct btrfs_inode_item);
142 btrfs_item_key(leaf, &disk_key, path->slots[0]);
143 memset_extent_buffer(leaf, 0, (unsigned long)inode_item,
144 sizeof(*inode_item));
145 btrfs_set_inode_generation(leaf, inode_item, trans->transid);
146 btrfs_set_inode_size(leaf, inode_item, 0);
147 btrfs_set_inode_nbytes(leaf, inode_item, 0);
148 btrfs_set_inode_uid(leaf, inode_item, 0);
149 btrfs_set_inode_gid(leaf, inode_item, 0);
150 btrfs_set_inode_mode(leaf, inode_item, S_IFREG | 0600);
151 btrfs_set_inode_flags(leaf, inode_item, flags);
152 btrfs_set_inode_nlink(leaf, inode_item, 1);
153 btrfs_set_inode_transid(leaf, inode_item, trans->transid);
154 btrfs_set_inode_block_group(leaf, inode_item, offset);
155 btrfs_mark_buffer_dirty(leaf);
156 btrfs_release_path(path);
158 key.objectid = BTRFS_FREE_SPACE_OBJECTID;
162 ret = btrfs_insert_empty_item(trans, root, path, &key,
163 sizeof(struct btrfs_free_space_header));
165 btrfs_release_path(path);
168 leaf = path->nodes[0];
169 header = btrfs_item_ptr(leaf, path->slots[0],
170 struct btrfs_free_space_header);
171 memset_extent_buffer(leaf, 0, (unsigned long)header, sizeof(*header));
172 btrfs_set_free_space_key(leaf, header, &disk_key);
173 btrfs_mark_buffer_dirty(leaf);
174 btrfs_release_path(path);
179 int create_free_space_inode(struct btrfs_root *root,
180 struct btrfs_trans_handle *trans,
181 struct btrfs_block_group_cache *block_group,
182 struct btrfs_path *path)
187 ret = btrfs_find_free_objectid(root, &ino);
191 return __create_free_space_inode(root, trans, path, ino,
192 block_group->key.objectid);
195 int btrfs_truncate_free_space_cache(struct btrfs_root *root,
196 struct btrfs_trans_handle *trans,
197 struct btrfs_path *path,
200 struct btrfs_block_rsv *rsv;
205 rsv = trans->block_rsv;
206 trans->block_rsv = &root->fs_info->global_block_rsv;
208 /* 1 for slack space, 1 for updating the inode */
209 needed_bytes = btrfs_calc_trunc_metadata_size(root, 1) +
210 btrfs_calc_trans_metadata_size(root, 1);
212 spin_lock(&trans->block_rsv->lock);
213 if (trans->block_rsv->reserved < needed_bytes) {
214 spin_unlock(&trans->block_rsv->lock);
215 trans->block_rsv = rsv;
218 spin_unlock(&trans->block_rsv->lock);
220 oldsize = i_size_read(inode);
221 btrfs_i_size_write(inode, 0);
222 truncate_pagecache(inode, oldsize, 0);
225 * We don't need an orphan item because truncating the free space cache
226 * will never be split across transactions.
228 ret = btrfs_truncate_inode_items(trans, root, inode,
229 0, BTRFS_EXTENT_DATA_KEY);
232 trans->block_rsv = rsv;
233 btrfs_abort_transaction(trans, root, ret);
237 ret = btrfs_update_inode(trans, root, inode);
239 btrfs_abort_transaction(trans, root, ret);
240 trans->block_rsv = rsv;
245 static int readahead_cache(struct inode *inode)
247 struct file_ra_state *ra;
248 unsigned long last_index;
250 ra = kzalloc(sizeof(*ra), GFP_NOFS);
254 file_ra_state_init(ra, inode->i_mapping);
255 last_index = (i_size_read(inode) - 1) >> PAGE_CACHE_SHIFT;
257 page_cache_sync_readahead(inode->i_mapping, ra, NULL, 0, last_index);
268 struct btrfs_root *root;
272 unsigned check_crcs:1;
275 static int io_ctl_init(struct io_ctl *io_ctl, struct inode *inode,
276 struct btrfs_root *root)
278 memset(io_ctl, 0, sizeof(struct io_ctl));
279 io_ctl->num_pages = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >>
281 io_ctl->pages = kzalloc(sizeof(struct page *) * io_ctl->num_pages,
286 if (btrfs_ino(inode) != BTRFS_FREE_INO_OBJECTID)
287 io_ctl->check_crcs = 1;
291 static void io_ctl_free(struct io_ctl *io_ctl)
293 kfree(io_ctl->pages);
296 static void io_ctl_unmap_page(struct io_ctl *io_ctl)
299 kunmap(io_ctl->page);
305 static void io_ctl_map_page(struct io_ctl *io_ctl, int clear)
307 WARN_ON(io_ctl->cur);
308 BUG_ON(io_ctl->index >= io_ctl->num_pages);
309 io_ctl->page = io_ctl->pages[io_ctl->index++];
310 io_ctl->cur = kmap(io_ctl->page);
311 io_ctl->orig = io_ctl->cur;
312 io_ctl->size = PAGE_CACHE_SIZE;
314 memset(io_ctl->cur, 0, PAGE_CACHE_SIZE);
317 static void io_ctl_drop_pages(struct io_ctl *io_ctl)
321 io_ctl_unmap_page(io_ctl);
323 for (i = 0; i < io_ctl->num_pages; i++) {
324 if (io_ctl->pages[i]) {
325 ClearPageChecked(io_ctl->pages[i]);
326 unlock_page(io_ctl->pages[i]);
327 page_cache_release(io_ctl->pages[i]);
332 static int io_ctl_prepare_pages(struct io_ctl *io_ctl, struct inode *inode,
336 gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping);
339 for (i = 0; i < io_ctl->num_pages; i++) {
340 page = find_or_create_page(inode->i_mapping, i, mask);
342 io_ctl_drop_pages(io_ctl);
345 io_ctl->pages[i] = page;
346 if (uptodate && !PageUptodate(page)) {
347 btrfs_readpage(NULL, page);
349 if (!PageUptodate(page)) {
350 printk(KERN_ERR "btrfs: error reading free "
352 io_ctl_drop_pages(io_ctl);
358 for (i = 0; i < io_ctl->num_pages; i++) {
359 clear_page_dirty_for_io(io_ctl->pages[i]);
360 set_page_extent_mapped(io_ctl->pages[i]);
366 static void io_ctl_set_generation(struct io_ctl *io_ctl, u64 generation)
370 io_ctl_map_page(io_ctl, 1);
373 * Skip the csum areas. If we don't check crcs then we just have a
374 * 64bit chunk at the front of the first page.
376 if (io_ctl->check_crcs) {
377 io_ctl->cur += (sizeof(u32) * io_ctl->num_pages);
378 io_ctl->size -= sizeof(u64) + (sizeof(u32) * io_ctl->num_pages);
380 io_ctl->cur += sizeof(u64);
381 io_ctl->size -= sizeof(u64) * 2;
385 *val = cpu_to_le64(generation);
386 io_ctl->cur += sizeof(u64);
389 static int io_ctl_check_generation(struct io_ctl *io_ctl, u64 generation)
394 * Skip the crc area. If we don't check crcs then we just have a 64bit
395 * chunk at the front of the first page.
397 if (io_ctl->check_crcs) {
398 io_ctl->cur += sizeof(u32) * io_ctl->num_pages;
399 io_ctl->size -= sizeof(u64) +
400 (sizeof(u32) * io_ctl->num_pages);
402 io_ctl->cur += sizeof(u64);
403 io_ctl->size -= sizeof(u64) * 2;
407 if (le64_to_cpu(*gen) != generation) {
408 printk_ratelimited(KERN_ERR "btrfs: space cache generation "
409 "(%Lu) does not match inode (%Lu)\n", *gen,
411 io_ctl_unmap_page(io_ctl);
414 io_ctl->cur += sizeof(u64);
418 static void io_ctl_set_crc(struct io_ctl *io_ctl, int index)
424 if (!io_ctl->check_crcs) {
425 io_ctl_unmap_page(io_ctl);
430 offset = sizeof(u32) * io_ctl->num_pages;
432 crc = btrfs_csum_data(io_ctl->root, io_ctl->orig + offset, crc,
433 PAGE_CACHE_SIZE - offset);
434 btrfs_csum_final(crc, (char *)&crc);
435 io_ctl_unmap_page(io_ctl);
436 tmp = kmap(io_ctl->pages[0]);
439 kunmap(io_ctl->pages[0]);
442 static int io_ctl_check_crc(struct io_ctl *io_ctl, int index)
448 if (!io_ctl->check_crcs) {
449 io_ctl_map_page(io_ctl, 0);
454 offset = sizeof(u32) * io_ctl->num_pages;
456 tmp = kmap(io_ctl->pages[0]);
459 kunmap(io_ctl->pages[0]);
461 io_ctl_map_page(io_ctl, 0);
462 crc = btrfs_csum_data(io_ctl->root, io_ctl->orig + offset, crc,
463 PAGE_CACHE_SIZE - offset);
464 btrfs_csum_final(crc, (char *)&crc);
466 printk_ratelimited(KERN_ERR "btrfs: csum mismatch on free "
468 io_ctl_unmap_page(io_ctl);
475 static int io_ctl_add_entry(struct io_ctl *io_ctl, u64 offset, u64 bytes,
478 struct btrfs_free_space_entry *entry;
484 entry->offset = cpu_to_le64(offset);
485 entry->bytes = cpu_to_le64(bytes);
486 entry->type = (bitmap) ? BTRFS_FREE_SPACE_BITMAP :
487 BTRFS_FREE_SPACE_EXTENT;
488 io_ctl->cur += sizeof(struct btrfs_free_space_entry);
489 io_ctl->size -= sizeof(struct btrfs_free_space_entry);
491 if (io_ctl->size >= sizeof(struct btrfs_free_space_entry))
494 io_ctl_set_crc(io_ctl, io_ctl->index - 1);
496 /* No more pages to map */
497 if (io_ctl->index >= io_ctl->num_pages)
500 /* map the next page */
501 io_ctl_map_page(io_ctl, 1);
505 static int io_ctl_add_bitmap(struct io_ctl *io_ctl, void *bitmap)
511 * If we aren't at the start of the current page, unmap this one and
512 * map the next one if there is any left.
514 if (io_ctl->cur != io_ctl->orig) {
515 io_ctl_set_crc(io_ctl, io_ctl->index - 1);
516 if (io_ctl->index >= io_ctl->num_pages)
518 io_ctl_map_page(io_ctl, 0);
521 memcpy(io_ctl->cur, bitmap, PAGE_CACHE_SIZE);
522 io_ctl_set_crc(io_ctl, io_ctl->index - 1);
523 if (io_ctl->index < io_ctl->num_pages)
524 io_ctl_map_page(io_ctl, 0);
528 static void io_ctl_zero_remaining_pages(struct io_ctl *io_ctl)
531 * If we're not on the boundary we know we've modified the page and we
532 * need to crc the page.
534 if (io_ctl->cur != io_ctl->orig)
535 io_ctl_set_crc(io_ctl, io_ctl->index - 1);
537 io_ctl_unmap_page(io_ctl);
539 while (io_ctl->index < io_ctl->num_pages) {
540 io_ctl_map_page(io_ctl, 1);
541 io_ctl_set_crc(io_ctl, io_ctl->index - 1);
545 static int io_ctl_read_entry(struct io_ctl *io_ctl,
546 struct btrfs_free_space *entry, u8 *type)
548 struct btrfs_free_space_entry *e;
552 ret = io_ctl_check_crc(io_ctl, io_ctl->index);
558 entry->offset = le64_to_cpu(e->offset);
559 entry->bytes = le64_to_cpu(e->bytes);
561 io_ctl->cur += sizeof(struct btrfs_free_space_entry);
562 io_ctl->size -= sizeof(struct btrfs_free_space_entry);
564 if (io_ctl->size >= sizeof(struct btrfs_free_space_entry))
567 io_ctl_unmap_page(io_ctl);
572 static int io_ctl_read_bitmap(struct io_ctl *io_ctl,
573 struct btrfs_free_space *entry)
577 ret = io_ctl_check_crc(io_ctl, io_ctl->index);
581 memcpy(entry->bitmap, io_ctl->cur, PAGE_CACHE_SIZE);
582 io_ctl_unmap_page(io_ctl);
587 int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
588 struct btrfs_free_space_ctl *ctl,
589 struct btrfs_path *path, u64 offset)
591 struct btrfs_free_space_header *header;
592 struct extent_buffer *leaf;
593 struct io_ctl io_ctl;
594 struct btrfs_key key;
595 struct btrfs_free_space *e, *n;
596 struct list_head bitmaps;
603 INIT_LIST_HEAD(&bitmaps);
605 /* Nothing in the space cache, goodbye */
606 if (!i_size_read(inode))
609 key.objectid = BTRFS_FREE_SPACE_OBJECTID;
613 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
617 btrfs_release_path(path);
623 leaf = path->nodes[0];
624 header = btrfs_item_ptr(leaf, path->slots[0],
625 struct btrfs_free_space_header);
626 num_entries = btrfs_free_space_entries(leaf, header);
627 num_bitmaps = btrfs_free_space_bitmaps(leaf, header);
628 generation = btrfs_free_space_generation(leaf, header);
629 btrfs_release_path(path);
631 if (BTRFS_I(inode)->generation != generation) {
632 printk(KERN_ERR "btrfs: free space inode generation (%llu) did"
633 " not match free space cache generation (%llu)\n",
634 (unsigned long long)BTRFS_I(inode)->generation,
635 (unsigned long long)generation);
642 ret = io_ctl_init(&io_ctl, inode, root);
646 ret = readahead_cache(inode);
650 ret = io_ctl_prepare_pages(&io_ctl, inode, 1);
654 ret = io_ctl_check_crc(&io_ctl, 0);
658 ret = io_ctl_check_generation(&io_ctl, generation);
662 while (num_entries) {
663 e = kmem_cache_zalloc(btrfs_free_space_cachep,
668 ret = io_ctl_read_entry(&io_ctl, e, &type);
670 kmem_cache_free(btrfs_free_space_cachep, e);
675 kmem_cache_free(btrfs_free_space_cachep, e);
679 if (type == BTRFS_FREE_SPACE_EXTENT) {
680 spin_lock(&ctl->tree_lock);
681 ret = link_free_space(ctl, e);
682 spin_unlock(&ctl->tree_lock);
684 printk(KERN_ERR "Duplicate entries in "
685 "free space cache, dumping\n");
686 kmem_cache_free(btrfs_free_space_cachep, e);
690 BUG_ON(!num_bitmaps);
692 e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
695 btrfs_free_space_cachep, e);
698 spin_lock(&ctl->tree_lock);
699 ret = link_free_space(ctl, e);
700 ctl->total_bitmaps++;
701 ctl->op->recalc_thresholds(ctl);
702 spin_unlock(&ctl->tree_lock);
704 printk(KERN_ERR "Duplicate entries in "
705 "free space cache, dumping\n");
706 kmem_cache_free(btrfs_free_space_cachep, e);
709 list_add_tail(&e->list, &bitmaps);
715 io_ctl_unmap_page(&io_ctl);
718 * We add the bitmaps at the end of the entries in order that
719 * the bitmap entries are added to the cache.
721 list_for_each_entry_safe(e, n, &bitmaps, list) {
722 list_del_init(&e->list);
723 ret = io_ctl_read_bitmap(&io_ctl, e);
728 io_ctl_drop_pages(&io_ctl);
731 io_ctl_free(&io_ctl);
734 io_ctl_drop_pages(&io_ctl);
735 __btrfs_remove_free_space_cache(ctl);
739 int load_free_space_cache(struct btrfs_fs_info *fs_info,
740 struct btrfs_block_group_cache *block_group)
742 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
743 struct btrfs_root *root = fs_info->tree_root;
745 struct btrfs_path *path;
748 u64 used = btrfs_block_group_used(&block_group->item);
751 * If this block group has been marked to be cleared for one reason or
752 * another then we can't trust the on disk cache, so just return.
754 spin_lock(&block_group->lock);
755 if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) {
756 spin_unlock(&block_group->lock);
759 spin_unlock(&block_group->lock);
761 path = btrfs_alloc_path();
764 path->search_commit_root = 1;
765 path->skip_locking = 1;
767 inode = lookup_free_space_inode(root, block_group, path);
769 btrfs_free_path(path);
773 /* We may have converted the inode and made the cache invalid. */
774 spin_lock(&block_group->lock);
775 if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) {
776 spin_unlock(&block_group->lock);
777 btrfs_free_path(path);
780 spin_unlock(&block_group->lock);
782 ret = __load_free_space_cache(fs_info->tree_root, inode, ctl,
783 path, block_group->key.objectid);
784 btrfs_free_path(path);
788 spin_lock(&ctl->tree_lock);
789 matched = (ctl->free_space == (block_group->key.offset - used -
790 block_group->bytes_super));
791 spin_unlock(&ctl->tree_lock);
794 __btrfs_remove_free_space_cache(ctl);
795 printk(KERN_ERR "block group %llu has an wrong amount of free "
796 "space\n", block_group->key.objectid);
801 /* This cache is bogus, make sure it gets cleared */
802 spin_lock(&block_group->lock);
803 block_group->disk_cache_state = BTRFS_DC_CLEAR;
804 spin_unlock(&block_group->lock);
807 printk(KERN_ERR "btrfs: failed to load free space cache "
808 "for block group %llu\n", block_group->key.objectid);
816 * __btrfs_write_out_cache - write out cached info to an inode
817 * @root - the root the inode belongs to
818 * @ctl - the free space cache we are going to write out
819 * @block_group - the block_group for this cache if it belongs to a block_group
820 * @trans - the trans handle
821 * @path - the path to use
822 * @offset - the offset for the key we'll insert
824 * This function writes out a free space cache struct to disk for quick recovery
825 * on mount. This will return 0 if it was successfull in writing the cache out,
826 * and -1 if it was not.
828 int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
829 struct btrfs_free_space_ctl *ctl,
830 struct btrfs_block_group_cache *block_group,
831 struct btrfs_trans_handle *trans,
832 struct btrfs_path *path, u64 offset)
834 struct btrfs_free_space_header *header;
835 struct extent_buffer *leaf;
836 struct rb_node *node;
837 struct list_head *pos, *n;
838 struct extent_state *cached_state = NULL;
839 struct btrfs_free_cluster *cluster = NULL;
840 struct extent_io_tree *unpin = NULL;
841 struct io_ctl io_ctl;
842 struct list_head bitmap_list;
843 struct btrfs_key key;
844 u64 start, extent_start, extent_end, len;
850 INIT_LIST_HEAD(&bitmap_list);
852 if (!i_size_read(inode))
855 ret = io_ctl_init(&io_ctl, inode, root);
859 /* Get the cluster for this block_group if it exists */
860 if (block_group && !list_empty(&block_group->cluster_list))
861 cluster = list_entry(block_group->cluster_list.next,
862 struct btrfs_free_cluster,
865 /* Lock all pages first so we can lock the extent safely. */
866 io_ctl_prepare_pages(&io_ctl, inode, 0);
868 lock_extent_bits(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1,
871 node = rb_first(&ctl->free_space_offset);
872 if (!node && cluster) {
873 node = rb_first(&cluster->root);
877 /* Make sure we can fit our crcs into the first page */
878 if (io_ctl.check_crcs &&
879 (io_ctl.num_pages * sizeof(u32)) >= PAGE_CACHE_SIZE) {
884 io_ctl_set_generation(&io_ctl, trans->transid);
886 /* Write out the extent entries */
888 struct btrfs_free_space *e;
890 e = rb_entry(node, struct btrfs_free_space, offset_index);
893 ret = io_ctl_add_entry(&io_ctl, e->offset, e->bytes,
899 list_add_tail(&e->list, &bitmap_list);
902 node = rb_next(node);
903 if (!node && cluster) {
904 node = rb_first(&cluster->root);
910 * We want to add any pinned extents to our free space cache
911 * so we don't leak the space
915 * We shouldn't have switched the pinned extents yet so this is the
918 unpin = root->fs_info->pinned_extents;
921 start = block_group->key.objectid;
923 while (block_group && (start < block_group->key.objectid +
924 block_group->key.offset)) {
925 ret = find_first_extent_bit(unpin, start,
926 &extent_start, &extent_end,
933 /* This pinned extent is out of our range */
934 if (extent_start >= block_group->key.objectid +
935 block_group->key.offset)
938 extent_start = max(extent_start, start);
939 extent_end = min(block_group->key.objectid +
940 block_group->key.offset, extent_end + 1);
941 len = extent_end - extent_start;
944 ret = io_ctl_add_entry(&io_ctl, extent_start, len, NULL);
951 /* Write out the bitmaps */
952 list_for_each_safe(pos, n, &bitmap_list) {
953 struct btrfs_free_space *entry =
954 list_entry(pos, struct btrfs_free_space, list);
956 ret = io_ctl_add_bitmap(&io_ctl, entry->bitmap);
959 list_del_init(&entry->list);
962 /* Zero out the rest of the pages just to make sure */
963 io_ctl_zero_remaining_pages(&io_ctl);
965 ret = btrfs_dirty_pages(root, inode, io_ctl.pages, io_ctl.num_pages,
966 0, i_size_read(inode), &cached_state);
967 io_ctl_drop_pages(&io_ctl);
968 unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
969 i_size_read(inode) - 1, &cached_state, GFP_NOFS);
975 ret = filemap_write_and_wait(inode->i_mapping);
979 key.objectid = BTRFS_FREE_SPACE_OBJECTID;
983 ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
985 clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, inode->i_size - 1,
986 EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, NULL,
990 leaf = path->nodes[0];
992 struct btrfs_key found_key;
993 BUG_ON(!path->slots[0]);
995 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
996 if (found_key.objectid != BTRFS_FREE_SPACE_OBJECTID ||
997 found_key.offset != offset) {
998 clear_extent_bit(&BTRFS_I(inode)->io_tree, 0,
1000 EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0,
1002 btrfs_release_path(path);
1007 BTRFS_I(inode)->generation = trans->transid;
1008 header = btrfs_item_ptr(leaf, path->slots[0],
1009 struct btrfs_free_space_header);
1010 btrfs_set_free_space_entries(leaf, header, entries);
1011 btrfs_set_free_space_bitmaps(leaf, header, bitmaps);
1012 btrfs_set_free_space_generation(leaf, header, trans->transid);
1013 btrfs_mark_buffer_dirty(leaf);
1014 btrfs_release_path(path);
1018 io_ctl_free(&io_ctl);
1020 invalidate_inode_pages2(inode->i_mapping);
1021 BTRFS_I(inode)->generation = 0;
1023 btrfs_update_inode(trans, root, inode);
1027 list_for_each_safe(pos, n, &bitmap_list) {
1028 struct btrfs_free_space *entry =
1029 list_entry(pos, struct btrfs_free_space, list);
1030 list_del_init(&entry->list);
1032 io_ctl_drop_pages(&io_ctl);
1033 unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
1034 i_size_read(inode) - 1, &cached_state, GFP_NOFS);
1038 int btrfs_write_out_cache(struct btrfs_root *root,
1039 struct btrfs_trans_handle *trans,
1040 struct btrfs_block_group_cache *block_group,
1041 struct btrfs_path *path)
1043 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
1044 struct inode *inode;
1047 root = root->fs_info->tree_root;
1049 spin_lock(&block_group->lock);
1050 if (block_group->disk_cache_state < BTRFS_DC_SETUP) {
1051 spin_unlock(&block_group->lock);
1054 spin_unlock(&block_group->lock);
1056 inode = lookup_free_space_inode(root, block_group, path);
1060 ret = __btrfs_write_out_cache(root, inode, ctl, block_group, trans,
1061 path, block_group->key.objectid);
1063 spin_lock(&block_group->lock);
1064 block_group->disk_cache_state = BTRFS_DC_ERROR;
1065 spin_unlock(&block_group->lock);
1068 printk(KERN_ERR "btrfs: failed to write free space cache "
1069 "for block group %llu\n", block_group->key.objectid);
1077 static inline unsigned long offset_to_bit(u64 bitmap_start, u32 unit,
1080 BUG_ON(offset < bitmap_start);
1081 offset -= bitmap_start;
1082 return (unsigned long)(div_u64(offset, unit));
1085 static inline unsigned long bytes_to_bits(u64 bytes, u32 unit)
1087 return (unsigned long)(div_u64(bytes, unit));
1090 static inline u64 offset_to_bitmap(struct btrfs_free_space_ctl *ctl,
1094 u64 bytes_per_bitmap;
1096 bytes_per_bitmap = BITS_PER_BITMAP * ctl->unit;
1097 bitmap_start = offset - ctl->start;
1098 bitmap_start = div64_u64(bitmap_start, bytes_per_bitmap);
1099 bitmap_start *= bytes_per_bitmap;
1100 bitmap_start += ctl->start;
1102 return bitmap_start;
1105 static int tree_insert_offset(struct rb_root *root, u64 offset,
1106 struct rb_node *node, int bitmap)
1108 struct rb_node **p = &root->rb_node;
1109 struct rb_node *parent = NULL;
1110 struct btrfs_free_space *info;
1114 info = rb_entry(parent, struct btrfs_free_space, offset_index);
1116 if (offset < info->offset) {
1118 } else if (offset > info->offset) {
1119 p = &(*p)->rb_right;
1122 * we could have a bitmap entry and an extent entry
1123 * share the same offset. If this is the case, we want
1124 * the extent entry to always be found first if we do a
1125 * linear search through the tree, since we want to have
1126 * the quickest allocation time, and allocating from an
1127 * extent is faster than allocating from a bitmap. So
1128 * if we're inserting a bitmap and we find an entry at
1129 * this offset, we want to go right, or after this entry
1130 * logically. If we are inserting an extent and we've
1131 * found a bitmap, we want to go left, or before
1139 p = &(*p)->rb_right;
1141 if (!info->bitmap) {
1150 rb_link_node(node, parent, p);
1151 rb_insert_color(node, root);
1157 * searches the tree for the given offset.
1159 * fuzzy - If this is set, then we are trying to make an allocation, and we just
1160 * want a section that has at least bytes size and comes at or after the given
1163 static struct btrfs_free_space *
1164 tree_search_offset(struct btrfs_free_space_ctl *ctl,
1165 u64 offset, int bitmap_only, int fuzzy)
1167 struct rb_node *n = ctl->free_space_offset.rb_node;
1168 struct btrfs_free_space *entry, *prev = NULL;
1170 /* find entry that is closest to the 'offset' */
1177 entry = rb_entry(n, struct btrfs_free_space, offset_index);
1180 if (offset < entry->offset)
1182 else if (offset > entry->offset)
1195 * bitmap entry and extent entry may share same offset,
1196 * in that case, bitmap entry comes after extent entry.
1201 entry = rb_entry(n, struct btrfs_free_space, offset_index);
1202 if (entry->offset != offset)
1205 WARN_ON(!entry->bitmap);
1208 if (entry->bitmap) {
1210 * if previous extent entry covers the offset,
1211 * we should return it instead of the bitmap entry
1213 n = &entry->offset_index;
1218 prev = rb_entry(n, struct btrfs_free_space,
1220 if (!prev->bitmap) {
1221 if (prev->offset + prev->bytes > offset)
1233 /* find last entry before the 'offset' */
1235 if (entry->offset > offset) {
1236 n = rb_prev(&entry->offset_index);
1238 entry = rb_entry(n, struct btrfs_free_space,
1240 BUG_ON(entry->offset > offset);
1249 if (entry->bitmap) {
1250 n = &entry->offset_index;
1255 prev = rb_entry(n, struct btrfs_free_space,
1257 if (!prev->bitmap) {
1258 if (prev->offset + prev->bytes > offset)
1263 if (entry->offset + BITS_PER_BITMAP * ctl->unit > offset)
1265 } else if (entry->offset + entry->bytes > offset)
1272 if (entry->bitmap) {
1273 if (entry->offset + BITS_PER_BITMAP *
1277 if (entry->offset + entry->bytes > offset)
1281 n = rb_next(&entry->offset_index);
1284 entry = rb_entry(n, struct btrfs_free_space, offset_index);
1290 __unlink_free_space(struct btrfs_free_space_ctl *ctl,
1291 struct btrfs_free_space *info)
1293 rb_erase(&info->offset_index, &ctl->free_space_offset);
1294 ctl->free_extents--;
1297 static void unlink_free_space(struct btrfs_free_space_ctl *ctl,
1298 struct btrfs_free_space *info)
1300 __unlink_free_space(ctl, info);
1301 ctl->free_space -= info->bytes;
1304 static int link_free_space(struct btrfs_free_space_ctl *ctl,
1305 struct btrfs_free_space *info)
1309 BUG_ON(!info->bitmap && !info->bytes);
1310 ret = tree_insert_offset(&ctl->free_space_offset, info->offset,
1311 &info->offset_index, (info->bitmap != NULL));
1315 ctl->free_space += info->bytes;
1316 ctl->free_extents++;
1320 static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl)
1322 struct btrfs_block_group_cache *block_group = ctl->private;
1326 u64 size = block_group->key.offset;
1327 u64 bytes_per_bg = BITS_PER_BITMAP * block_group->sectorsize;
1328 int max_bitmaps = div64_u64(size + bytes_per_bg - 1, bytes_per_bg);
1330 BUG_ON(ctl->total_bitmaps > max_bitmaps);
1333 * The goal is to keep the total amount of memory used per 1gb of space
1334 * at or below 32k, so we need to adjust how much memory we allow to be
1335 * used by extent based free space tracking
1337 if (size < 1024 * 1024 * 1024)
1338 max_bytes = MAX_CACHE_BYTES_PER_GIG;
1340 max_bytes = MAX_CACHE_BYTES_PER_GIG *
1341 div64_u64(size, 1024 * 1024 * 1024);
1344 * we want to account for 1 more bitmap than what we have so we can make
1345 * sure we don't go over our overall goal of MAX_CACHE_BYTES_PER_GIG as
1346 * we add more bitmaps.
1348 bitmap_bytes = (ctl->total_bitmaps + 1) * PAGE_CACHE_SIZE;
1350 if (bitmap_bytes >= max_bytes) {
1351 ctl->extents_thresh = 0;
1356 * we want the extent entry threshold to always be at most 1/2 the maxw
1357 * bytes we can have, or whatever is less than that.
1359 extent_bytes = max_bytes - bitmap_bytes;
1360 extent_bytes = min_t(u64, extent_bytes, div64_u64(max_bytes, 2));
1362 ctl->extents_thresh =
1363 div64_u64(extent_bytes, (sizeof(struct btrfs_free_space)));
1366 static inline void __bitmap_clear_bits(struct btrfs_free_space_ctl *ctl,
1367 struct btrfs_free_space *info,
1368 u64 offset, u64 bytes)
1370 unsigned long start, count;
1372 start = offset_to_bit(info->offset, ctl->unit, offset);
1373 count = bytes_to_bits(bytes, ctl->unit);
1374 BUG_ON(start + count > BITS_PER_BITMAP);
1376 bitmap_clear(info->bitmap, start, count);
1378 info->bytes -= bytes;
1381 static void bitmap_clear_bits(struct btrfs_free_space_ctl *ctl,
1382 struct btrfs_free_space *info, u64 offset,
1385 __bitmap_clear_bits(ctl, info, offset, bytes);
1386 ctl->free_space -= bytes;
1389 static void bitmap_set_bits(struct btrfs_free_space_ctl *ctl,
1390 struct btrfs_free_space *info, u64 offset,
1393 unsigned long start, count;
1395 start = offset_to_bit(info->offset, ctl->unit, offset);
1396 count = bytes_to_bits(bytes, ctl->unit);
1397 BUG_ON(start + count > BITS_PER_BITMAP);
1399 bitmap_set(info->bitmap, start, count);
1401 info->bytes += bytes;
1402 ctl->free_space += bytes;
1405 static int search_bitmap(struct btrfs_free_space_ctl *ctl,
1406 struct btrfs_free_space *bitmap_info, u64 *offset,
1409 unsigned long found_bits = 0;
1410 unsigned long bits, i;
1411 unsigned long next_zero;
1413 i = offset_to_bit(bitmap_info->offset, ctl->unit,
1414 max_t(u64, *offset, bitmap_info->offset));
1415 bits = bytes_to_bits(*bytes, ctl->unit);
1417 for (i = find_next_bit(bitmap_info->bitmap, BITS_PER_BITMAP, i);
1418 i < BITS_PER_BITMAP;
1419 i = find_next_bit(bitmap_info->bitmap, BITS_PER_BITMAP, i + 1)) {
1420 next_zero = find_next_zero_bit(bitmap_info->bitmap,
1421 BITS_PER_BITMAP, i);
1422 if ((next_zero - i) >= bits) {
1423 found_bits = next_zero - i;
1430 *offset = (u64)(i * ctl->unit) + bitmap_info->offset;
1431 *bytes = (u64)(found_bits) * ctl->unit;
1438 static struct btrfs_free_space *
1439 find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes)
1441 struct btrfs_free_space *entry;
1442 struct rb_node *node;
1445 if (!ctl->free_space_offset.rb_node)
1448 entry = tree_search_offset(ctl, offset_to_bitmap(ctl, *offset), 0, 1);
1452 for (node = &entry->offset_index; node; node = rb_next(node)) {
1453 entry = rb_entry(node, struct btrfs_free_space, offset_index);
1454 if (entry->bytes < *bytes)
1457 if (entry->bitmap) {
1458 ret = search_bitmap(ctl, entry, offset, bytes);
1464 *offset = entry->offset;
1465 *bytes = entry->bytes;
1472 static void add_new_bitmap(struct btrfs_free_space_ctl *ctl,
1473 struct btrfs_free_space *info, u64 offset)
1475 info->offset = offset_to_bitmap(ctl, offset);
1477 INIT_LIST_HEAD(&info->list);
1478 link_free_space(ctl, info);
1479 ctl->total_bitmaps++;
1481 ctl->op->recalc_thresholds(ctl);
1484 static void free_bitmap(struct btrfs_free_space_ctl *ctl,
1485 struct btrfs_free_space *bitmap_info)
1487 unlink_free_space(ctl, bitmap_info);
1488 kfree(bitmap_info->bitmap);
1489 kmem_cache_free(btrfs_free_space_cachep, bitmap_info);
1490 ctl->total_bitmaps--;
1491 ctl->op->recalc_thresholds(ctl);
1494 static noinline int remove_from_bitmap(struct btrfs_free_space_ctl *ctl,
1495 struct btrfs_free_space *bitmap_info,
1496 u64 *offset, u64 *bytes)
1499 u64 search_start, search_bytes;
1503 end = bitmap_info->offset + (u64)(BITS_PER_BITMAP * ctl->unit) - 1;
1506 * XXX - this can go away after a few releases.
1508 * since the only user of btrfs_remove_free_space is the tree logging
1509 * stuff, and the only way to test that is under crash conditions, we
1510 * want to have this debug stuff here just in case somethings not
1511 * working. Search the bitmap for the space we are trying to use to
1512 * make sure its actually there. If its not there then we need to stop
1513 * because something has gone wrong.
1515 search_start = *offset;
1516 search_bytes = *bytes;
1517 search_bytes = min(search_bytes, end - search_start + 1);
1518 ret = search_bitmap(ctl, bitmap_info, &search_start, &search_bytes);
1519 BUG_ON(ret < 0 || search_start != *offset);
1521 if (*offset > bitmap_info->offset && *offset + *bytes > end) {
1522 bitmap_clear_bits(ctl, bitmap_info, *offset, end - *offset + 1);
1523 *bytes -= end - *offset + 1;
1525 } else if (*offset >= bitmap_info->offset && *offset + *bytes <= end) {
1526 bitmap_clear_bits(ctl, bitmap_info, *offset, *bytes);
1531 struct rb_node *next = rb_next(&bitmap_info->offset_index);
1532 if (!bitmap_info->bytes)
1533 free_bitmap(ctl, bitmap_info);
1536 * no entry after this bitmap, but we still have bytes to
1537 * remove, so something has gone wrong.
1542 bitmap_info = rb_entry(next, struct btrfs_free_space,
1546 * if the next entry isn't a bitmap we need to return to let the
1547 * extent stuff do its work.
1549 if (!bitmap_info->bitmap)
1553 * Ok the next item is a bitmap, but it may not actually hold
1554 * the information for the rest of this free space stuff, so
1555 * look for it, and if we don't find it return so we can try
1556 * everything over again.
1558 search_start = *offset;
1559 search_bytes = *bytes;
1560 ret = search_bitmap(ctl, bitmap_info, &search_start,
1562 if (ret < 0 || search_start != *offset)
1566 } else if (!bitmap_info->bytes)
1567 free_bitmap(ctl, bitmap_info);
1572 static u64 add_bytes_to_bitmap(struct btrfs_free_space_ctl *ctl,
1573 struct btrfs_free_space *info, u64 offset,
1576 u64 bytes_to_set = 0;
1579 end = info->offset + (u64)(BITS_PER_BITMAP * ctl->unit);
1581 bytes_to_set = min(end - offset, bytes);
1583 bitmap_set_bits(ctl, info, offset, bytes_to_set);
1585 return bytes_to_set;
1589 static bool use_bitmap(struct btrfs_free_space_ctl *ctl,
1590 struct btrfs_free_space *info)
1592 struct btrfs_block_group_cache *block_group = ctl->private;
1595 * If we are below the extents threshold then we can add this as an
1596 * extent, and don't have to deal with the bitmap
1598 if (ctl->free_extents < ctl->extents_thresh) {
1600 * If this block group has some small extents we don't want to
1601 * use up all of our free slots in the cache with them, we want
1602 * to reserve them to larger extents, however if we have plent
1603 * of cache left then go ahead an dadd them, no sense in adding
1604 * the overhead of a bitmap if we don't have to.
1606 if (info->bytes <= block_group->sectorsize * 4) {
1607 if (ctl->free_extents * 2 <= ctl->extents_thresh)
1615 * some block groups are so tiny they can't be enveloped by a bitmap, so
1616 * don't even bother to create a bitmap for this
1618 if (BITS_PER_BITMAP * block_group->sectorsize >
1619 block_group->key.offset)
1625 static struct btrfs_free_space_op free_space_op = {
1626 .recalc_thresholds = recalculate_thresholds,
1627 .use_bitmap = use_bitmap,
1630 static int insert_into_bitmap(struct btrfs_free_space_ctl *ctl,
1631 struct btrfs_free_space *info)
1633 struct btrfs_free_space *bitmap_info;
1634 struct btrfs_block_group_cache *block_group = NULL;
1636 u64 bytes, offset, bytes_added;
1639 bytes = info->bytes;
1640 offset = info->offset;
1642 if (!ctl->op->use_bitmap(ctl, info))
1645 if (ctl->op == &free_space_op)
1646 block_group = ctl->private;
1649 * Since we link bitmaps right into the cluster we need to see if we
1650 * have a cluster here, and if so and it has our bitmap we need to add
1651 * the free space to that bitmap.
1653 if (block_group && !list_empty(&block_group->cluster_list)) {
1654 struct btrfs_free_cluster *cluster;
1655 struct rb_node *node;
1656 struct btrfs_free_space *entry;
1658 cluster = list_entry(block_group->cluster_list.next,
1659 struct btrfs_free_cluster,
1661 spin_lock(&cluster->lock);
1662 node = rb_first(&cluster->root);
1664 spin_unlock(&cluster->lock);
1665 goto no_cluster_bitmap;
1668 entry = rb_entry(node, struct btrfs_free_space, offset_index);
1669 if (!entry->bitmap) {
1670 spin_unlock(&cluster->lock);
1671 goto no_cluster_bitmap;
1674 if (entry->offset == offset_to_bitmap(ctl, offset)) {
1675 bytes_added = add_bytes_to_bitmap(ctl, entry,
1677 bytes -= bytes_added;
1678 offset += bytes_added;
1680 spin_unlock(&cluster->lock);
1688 bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
1695 bytes_added = add_bytes_to_bitmap(ctl, bitmap_info, offset, bytes);
1696 bytes -= bytes_added;
1697 offset += bytes_added;
1707 if (info && info->bitmap) {
1708 add_new_bitmap(ctl, info, offset);
1713 spin_unlock(&ctl->tree_lock);
1715 /* no pre-allocated info, allocate a new one */
1717 info = kmem_cache_zalloc(btrfs_free_space_cachep,
1720 spin_lock(&ctl->tree_lock);
1726 /* allocate the bitmap */
1727 info->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
1728 spin_lock(&ctl->tree_lock);
1729 if (!info->bitmap) {
1739 kfree(info->bitmap);
1740 kmem_cache_free(btrfs_free_space_cachep, info);
1746 static bool try_merge_free_space(struct btrfs_free_space_ctl *ctl,
1747 struct btrfs_free_space *info, bool update_stat)
1749 struct btrfs_free_space *left_info;
1750 struct btrfs_free_space *right_info;
1751 bool merged = false;
1752 u64 offset = info->offset;
1753 u64 bytes = info->bytes;
1756 * first we want to see if there is free space adjacent to the range we
1757 * are adding, if there is remove that struct and add a new one to
1758 * cover the entire range
1760 right_info = tree_search_offset(ctl, offset + bytes, 0, 0);
1761 if (right_info && rb_prev(&right_info->offset_index))
1762 left_info = rb_entry(rb_prev(&right_info->offset_index),
1763 struct btrfs_free_space, offset_index);
1765 left_info = tree_search_offset(ctl, offset - 1, 0, 0);
1767 if (right_info && !right_info->bitmap) {
1769 unlink_free_space(ctl, right_info);
1771 __unlink_free_space(ctl, right_info);
1772 info->bytes += right_info->bytes;
1773 kmem_cache_free(btrfs_free_space_cachep, right_info);
1777 if (left_info && !left_info->bitmap &&
1778 left_info->offset + left_info->bytes == offset) {
1780 unlink_free_space(ctl, left_info);
1782 __unlink_free_space(ctl, left_info);
1783 info->offset = left_info->offset;
1784 info->bytes += left_info->bytes;
1785 kmem_cache_free(btrfs_free_space_cachep, left_info);
1792 int __btrfs_add_free_space(struct btrfs_free_space_ctl *ctl,
1793 u64 offset, u64 bytes)
1795 struct btrfs_free_space *info;
1798 info = kmem_cache_zalloc(btrfs_free_space_cachep, GFP_NOFS);
1802 info->offset = offset;
1803 info->bytes = bytes;
1805 spin_lock(&ctl->tree_lock);
1807 if (try_merge_free_space(ctl, info, true))
1811 * There was no extent directly to the left or right of this new
1812 * extent then we know we're going to have to allocate a new extent, so
1813 * before we do that see if we need to drop this into a bitmap
1815 ret = insert_into_bitmap(ctl, info);
1823 ret = link_free_space(ctl, info);
1825 kmem_cache_free(btrfs_free_space_cachep, info);
1827 spin_unlock(&ctl->tree_lock);
1830 printk(KERN_CRIT "btrfs: unable to add free space :%d\n", ret);
1831 BUG_ON(ret == -EEXIST);
1837 int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
1838 u64 offset, u64 bytes)
1840 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
1841 struct btrfs_free_space *info;
1842 struct btrfs_free_space *next_info = NULL;
1845 spin_lock(&ctl->tree_lock);
1848 info = tree_search_offset(ctl, offset, 0, 0);
1851 * oops didn't find an extent that matched the space we wanted
1852 * to remove, look for a bitmap instead
1854 info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
1857 /* the tree logging code might be calling us before we
1858 * have fully loaded the free space rbtree for this
1859 * block group. So it is possible the entry won't
1860 * be in the rbtree yet at all. The caching code
1861 * will make sure not to put it in the rbtree if
1862 * the logging code has pinned it.
1868 if (info->bytes < bytes && rb_next(&info->offset_index)) {
1870 next_info = rb_entry(rb_next(&info->offset_index),
1871 struct btrfs_free_space,
1874 if (next_info->bitmap)
1875 end = next_info->offset +
1876 BITS_PER_BITMAP * ctl->unit - 1;
1878 end = next_info->offset + next_info->bytes;
1880 if (next_info->bytes < bytes ||
1881 next_info->offset > offset || offset > end) {
1882 printk(KERN_CRIT "Found free space at %llu, size %llu,"
1883 " trying to use %llu\n",
1884 (unsigned long long)info->offset,
1885 (unsigned long long)info->bytes,
1886 (unsigned long long)bytes);
1895 if (info->bytes == bytes) {
1896 unlink_free_space(ctl, info);
1898 kfree(info->bitmap);
1899 ctl->total_bitmaps--;
1901 kmem_cache_free(btrfs_free_space_cachep, info);
1906 if (!info->bitmap && info->offset == offset) {
1907 unlink_free_space(ctl, info);
1908 info->offset += bytes;
1909 info->bytes -= bytes;
1910 ret = link_free_space(ctl, info);
1915 if (!info->bitmap && info->offset <= offset &&
1916 info->offset + info->bytes >= offset + bytes) {
1917 u64 old_start = info->offset;
1919 * we're freeing space in the middle of the info,
1920 * this can happen during tree log replay
1922 * first unlink the old info and then
1923 * insert it again after the hole we're creating
1925 unlink_free_space(ctl, info);
1926 if (offset + bytes < info->offset + info->bytes) {
1927 u64 old_end = info->offset + info->bytes;
1929 info->offset = offset + bytes;
1930 info->bytes = old_end - info->offset;
1931 ret = link_free_space(ctl, info);
1936 /* the hole we're creating ends at the end
1937 * of the info struct, just free the info
1939 kmem_cache_free(btrfs_free_space_cachep, info);
1941 spin_unlock(&ctl->tree_lock);
1943 /* step two, insert a new info struct to cover
1944 * anything before the hole
1946 ret = btrfs_add_free_space(block_group, old_start,
1947 offset - old_start);
1948 WARN_ON(ret); /* -ENOMEM */
1952 ret = remove_from_bitmap(ctl, info, &offset, &bytes);
1955 BUG_ON(ret); /* logic error */
1957 spin_unlock(&ctl->tree_lock);
1962 void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
1965 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
1966 struct btrfs_free_space *info;
1970 for (n = rb_first(&ctl->free_space_offset); n; n = rb_next(n)) {
1971 info = rb_entry(n, struct btrfs_free_space, offset_index);
1972 if (info->bytes >= bytes)
1974 printk(KERN_CRIT "entry offset %llu, bytes %llu, bitmap %s\n",
1975 (unsigned long long)info->offset,
1976 (unsigned long long)info->bytes,
1977 (info->bitmap) ? "yes" : "no");
1979 printk(KERN_INFO "block group has cluster?: %s\n",
1980 list_empty(&block_group->cluster_list) ? "no" : "yes");
1981 printk(KERN_INFO "%d blocks of free space at or bigger than bytes is"
1985 void btrfs_init_free_space_ctl(struct btrfs_block_group_cache *block_group)
1987 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
1989 spin_lock_init(&ctl->tree_lock);
1990 ctl->unit = block_group->sectorsize;
1991 ctl->start = block_group->key.objectid;
1992 ctl->private = block_group;
1993 ctl->op = &free_space_op;
1996 * we only want to have 32k of ram per block group for keeping
1997 * track of free space, and if we pass 1/2 of that we want to
1998 * start converting things over to using bitmaps
2000 ctl->extents_thresh = ((1024 * 32) / 2) /
2001 sizeof(struct btrfs_free_space);
2005 * for a given cluster, put all of its extents back into the free
2006 * space cache. If the block group passed doesn't match the block group
2007 * pointed to by the cluster, someone else raced in and freed the
2008 * cluster already. In that case, we just return without changing anything
2011 __btrfs_return_cluster_to_free_space(
2012 struct btrfs_block_group_cache *block_group,
2013 struct btrfs_free_cluster *cluster)
2015 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2016 struct btrfs_free_space *entry;
2017 struct rb_node *node;
2019 spin_lock(&cluster->lock);
2020 if (cluster->block_group != block_group)
2023 cluster->block_group = NULL;
2024 cluster->window_start = 0;
2025 list_del_init(&cluster->block_group_list);
2027 node = rb_first(&cluster->root);
2031 entry = rb_entry(node, struct btrfs_free_space, offset_index);
2032 node = rb_next(&entry->offset_index);
2033 rb_erase(&entry->offset_index, &cluster->root);
2035 bitmap = (entry->bitmap != NULL);
2037 try_merge_free_space(ctl, entry, false);
2038 tree_insert_offset(&ctl->free_space_offset,
2039 entry->offset, &entry->offset_index, bitmap);
2041 cluster->root = RB_ROOT;
2044 spin_unlock(&cluster->lock);
2045 btrfs_put_block_group(block_group);
2049 void __btrfs_remove_free_space_cache_locked(struct btrfs_free_space_ctl *ctl)
2051 struct btrfs_free_space *info;
2052 struct rb_node *node;
2054 while ((node = rb_last(&ctl->free_space_offset)) != NULL) {
2055 info = rb_entry(node, struct btrfs_free_space, offset_index);
2056 if (!info->bitmap) {
2057 unlink_free_space(ctl, info);
2058 kmem_cache_free(btrfs_free_space_cachep, info);
2060 free_bitmap(ctl, info);
2062 if (need_resched()) {
2063 spin_unlock(&ctl->tree_lock);
2065 spin_lock(&ctl->tree_lock);
2070 void __btrfs_remove_free_space_cache(struct btrfs_free_space_ctl *ctl)
2072 spin_lock(&ctl->tree_lock);
2073 __btrfs_remove_free_space_cache_locked(ctl);
2074 spin_unlock(&ctl->tree_lock);
2077 void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group)
2079 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2080 struct btrfs_free_cluster *cluster;
2081 struct list_head *head;
2083 spin_lock(&ctl->tree_lock);
2084 while ((head = block_group->cluster_list.next) !=
2085 &block_group->cluster_list) {
2086 cluster = list_entry(head, struct btrfs_free_cluster,
2089 WARN_ON(cluster->block_group != block_group);
2090 __btrfs_return_cluster_to_free_space(block_group, cluster);
2091 if (need_resched()) {
2092 spin_unlock(&ctl->tree_lock);
2094 spin_lock(&ctl->tree_lock);
2097 __btrfs_remove_free_space_cache_locked(ctl);
2098 spin_unlock(&ctl->tree_lock);
2102 u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
2103 u64 offset, u64 bytes, u64 empty_size)
2105 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2106 struct btrfs_free_space *entry = NULL;
2107 u64 bytes_search = bytes + empty_size;
2110 spin_lock(&ctl->tree_lock);
2111 entry = find_free_space(ctl, &offset, &bytes_search);
2116 if (entry->bitmap) {
2117 bitmap_clear_bits(ctl, entry, offset, bytes);
2119 free_bitmap(ctl, entry);
2121 unlink_free_space(ctl, entry);
2122 entry->offset += bytes;
2123 entry->bytes -= bytes;
2125 kmem_cache_free(btrfs_free_space_cachep, entry);
2127 link_free_space(ctl, entry);
2131 spin_unlock(&ctl->tree_lock);
2137 * given a cluster, put all of its extents back into the free space
2138 * cache. If a block group is passed, this function will only free
2139 * a cluster that belongs to the passed block group.
2141 * Otherwise, it'll get a reference on the block group pointed to by the
2142 * cluster and remove the cluster from it.
2144 int btrfs_return_cluster_to_free_space(
2145 struct btrfs_block_group_cache *block_group,
2146 struct btrfs_free_cluster *cluster)
2148 struct btrfs_free_space_ctl *ctl;
2151 /* first, get a safe pointer to the block group */
2152 spin_lock(&cluster->lock);
2154 block_group = cluster->block_group;
2156 spin_unlock(&cluster->lock);
2159 } else if (cluster->block_group != block_group) {
2160 /* someone else has already freed it don't redo their work */
2161 spin_unlock(&cluster->lock);
2164 atomic_inc(&block_group->count);
2165 spin_unlock(&cluster->lock);
2167 ctl = block_group->free_space_ctl;
2169 /* now return any extents the cluster had on it */
2170 spin_lock(&ctl->tree_lock);
2171 ret = __btrfs_return_cluster_to_free_space(block_group, cluster);
2172 spin_unlock(&ctl->tree_lock);
2174 /* finally drop our ref */
2175 btrfs_put_block_group(block_group);
2179 static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group,
2180 struct btrfs_free_cluster *cluster,
2181 struct btrfs_free_space *entry,
2182 u64 bytes, u64 min_start)
2184 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2186 u64 search_start = cluster->window_start;
2187 u64 search_bytes = bytes;
2190 search_start = min_start;
2191 search_bytes = bytes;
2193 err = search_bitmap(ctl, entry, &search_start, &search_bytes);
2198 __bitmap_clear_bits(ctl, entry, ret, bytes);
2204 * given a cluster, try to allocate 'bytes' from it, returns 0
2205 * if it couldn't find anything suitably large, or a logical disk offset
2206 * if things worked out
2208 u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group,
2209 struct btrfs_free_cluster *cluster, u64 bytes,
2212 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2213 struct btrfs_free_space *entry = NULL;
2214 struct rb_node *node;
2217 spin_lock(&cluster->lock);
2218 if (bytes > cluster->max_size)
2221 if (cluster->block_group != block_group)
2224 node = rb_first(&cluster->root);
2228 entry = rb_entry(node, struct btrfs_free_space, offset_index);
2230 if (entry->bytes < bytes ||
2231 (!entry->bitmap && entry->offset < min_start)) {
2232 node = rb_next(&entry->offset_index);
2235 entry = rb_entry(node, struct btrfs_free_space,
2240 if (entry->bitmap) {
2241 ret = btrfs_alloc_from_bitmap(block_group,
2242 cluster, entry, bytes,
2243 cluster->window_start);
2245 node = rb_next(&entry->offset_index);
2248 entry = rb_entry(node, struct btrfs_free_space,
2252 cluster->window_start += bytes;
2254 ret = entry->offset;
2256 entry->offset += bytes;
2257 entry->bytes -= bytes;
2260 if (entry->bytes == 0)
2261 rb_erase(&entry->offset_index, &cluster->root);
2265 spin_unlock(&cluster->lock);
2270 spin_lock(&ctl->tree_lock);
2272 ctl->free_space -= bytes;
2273 if (entry->bytes == 0) {
2274 ctl->free_extents--;
2275 if (entry->bitmap) {
2276 kfree(entry->bitmap);
2277 ctl->total_bitmaps--;
2278 ctl->op->recalc_thresholds(ctl);
2280 kmem_cache_free(btrfs_free_space_cachep, entry);
2283 spin_unlock(&ctl->tree_lock);
2288 static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group,
2289 struct btrfs_free_space *entry,
2290 struct btrfs_free_cluster *cluster,
2291 u64 offset, u64 bytes,
2292 u64 cont1_bytes, u64 min_bytes)
2294 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2295 unsigned long next_zero;
2297 unsigned long want_bits;
2298 unsigned long min_bits;
2299 unsigned long found_bits;
2300 unsigned long start = 0;
2301 unsigned long total_found = 0;
2304 i = offset_to_bit(entry->offset, block_group->sectorsize,
2305 max_t(u64, offset, entry->offset));
2306 want_bits = bytes_to_bits(bytes, block_group->sectorsize);
2307 min_bits = bytes_to_bits(min_bytes, block_group->sectorsize);
2311 for (i = find_next_bit(entry->bitmap, BITS_PER_BITMAP, i);
2312 i < BITS_PER_BITMAP;
2313 i = find_next_bit(entry->bitmap, BITS_PER_BITMAP, i + 1)) {
2314 next_zero = find_next_zero_bit(entry->bitmap,
2315 BITS_PER_BITMAP, i);
2316 if (next_zero - i >= min_bits) {
2317 found_bits = next_zero - i;
2328 cluster->max_size = 0;
2331 total_found += found_bits;
2333 if (cluster->max_size < found_bits * block_group->sectorsize)
2334 cluster->max_size = found_bits * block_group->sectorsize;
2336 if (total_found < want_bits || cluster->max_size < cont1_bytes) {
2341 cluster->window_start = start * block_group->sectorsize +
2343 rb_erase(&entry->offset_index, &ctl->free_space_offset);
2344 ret = tree_insert_offset(&cluster->root, entry->offset,
2345 &entry->offset_index, 1);
2346 BUG_ON(ret); /* -EEXIST; Logic error */
2348 trace_btrfs_setup_cluster(block_group, cluster,
2349 total_found * block_group->sectorsize, 1);
2354 * This searches the block group for just extents to fill the cluster with.
2355 * Try to find a cluster with at least bytes total bytes, at least one
2356 * extent of cont1_bytes, and other clusters of at least min_bytes.
2359 setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group,
2360 struct btrfs_free_cluster *cluster,
2361 struct list_head *bitmaps, u64 offset, u64 bytes,
2362 u64 cont1_bytes, u64 min_bytes)
2364 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2365 struct btrfs_free_space *first = NULL;
2366 struct btrfs_free_space *entry = NULL;
2367 struct btrfs_free_space *last;
2368 struct rb_node *node;
2374 entry = tree_search_offset(ctl, offset, 0, 1);
2379 * We don't want bitmaps, so just move along until we find a normal
2382 while (entry->bitmap || entry->bytes < min_bytes) {
2383 if (entry->bitmap && list_empty(&entry->list))
2384 list_add_tail(&entry->list, bitmaps);
2385 node = rb_next(&entry->offset_index);
2388 entry = rb_entry(node, struct btrfs_free_space, offset_index);
2391 window_start = entry->offset;
2392 window_free = entry->bytes;
2393 max_extent = entry->bytes;
2397 for (node = rb_next(&entry->offset_index); node;
2398 node = rb_next(&entry->offset_index)) {
2399 entry = rb_entry(node, struct btrfs_free_space, offset_index);
2401 if (entry->bitmap) {
2402 if (list_empty(&entry->list))
2403 list_add_tail(&entry->list, bitmaps);
2407 if (entry->bytes < min_bytes)
2411 window_free += entry->bytes;
2412 if (entry->bytes > max_extent)
2413 max_extent = entry->bytes;
2416 if (window_free < bytes || max_extent < cont1_bytes)
2419 cluster->window_start = first->offset;
2421 node = &first->offset_index;
2424 * now we've found our entries, pull them out of the free space
2425 * cache and put them into the cluster rbtree
2430 entry = rb_entry(node, struct btrfs_free_space, offset_index);
2431 node = rb_next(&entry->offset_index);
2432 if (entry->bitmap || entry->bytes < min_bytes)
2435 rb_erase(&entry->offset_index, &ctl->free_space_offset);
2436 ret = tree_insert_offset(&cluster->root, entry->offset,
2437 &entry->offset_index, 0);
2438 total_size += entry->bytes;
2439 BUG_ON(ret); /* -EEXIST; Logic error */
2440 } while (node && entry != last);
2442 cluster->max_size = max_extent;
2443 trace_btrfs_setup_cluster(block_group, cluster, total_size, 0);
2448 * This specifically looks for bitmaps that may work in the cluster, we assume
2449 * that we have already failed to find extents that will work.
2452 setup_cluster_bitmap(struct btrfs_block_group_cache *block_group,
2453 struct btrfs_free_cluster *cluster,
2454 struct list_head *bitmaps, u64 offset, u64 bytes,
2455 u64 cont1_bytes, u64 min_bytes)
2457 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2458 struct btrfs_free_space *entry;
2460 u64 bitmap_offset = offset_to_bitmap(ctl, offset);
2462 if (ctl->total_bitmaps == 0)
2466 * The bitmap that covers offset won't be in the list unless offset
2467 * is just its start offset.
2469 entry = list_first_entry(bitmaps, struct btrfs_free_space, list);
2470 if (entry->offset != bitmap_offset) {
2471 entry = tree_search_offset(ctl, bitmap_offset, 1, 0);
2472 if (entry && list_empty(&entry->list))
2473 list_add(&entry->list, bitmaps);
2476 list_for_each_entry(entry, bitmaps, list) {
2477 if (entry->bytes < bytes)
2479 ret = btrfs_bitmap_cluster(block_group, entry, cluster, offset,
2480 bytes, cont1_bytes, min_bytes);
2486 * The bitmaps list has all the bitmaps that record free space
2487 * starting after offset, so no more search is required.
2493 * here we try to find a cluster of blocks in a block group. The goal
2494 * is to find at least bytes+empty_size.
2495 * We might not find them all in one contiguous area.
2497 * returns zero and sets up cluster if things worked out, otherwise
2498 * it returns -enospc
2500 int btrfs_find_space_cluster(struct btrfs_trans_handle *trans,
2501 struct btrfs_root *root,
2502 struct btrfs_block_group_cache *block_group,
2503 struct btrfs_free_cluster *cluster,
2504 u64 offset, u64 bytes, u64 empty_size)
2506 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2507 struct btrfs_free_space *entry, *tmp;
2514 * Choose the minimum extent size we'll require for this
2515 * cluster. For SSD_SPREAD, don't allow any fragmentation.
2516 * For metadata, allow allocates with smaller extents. For
2517 * data, keep it dense.
2519 if (btrfs_test_opt(root, SSD_SPREAD)) {
2520 cont1_bytes = min_bytes = bytes + empty_size;
2521 } else if (block_group->flags & BTRFS_BLOCK_GROUP_METADATA) {
2522 cont1_bytes = bytes;
2523 min_bytes = block_group->sectorsize;
2525 cont1_bytes = max(bytes, (bytes + empty_size) >> 2);
2526 min_bytes = block_group->sectorsize;
2529 spin_lock(&ctl->tree_lock);
2532 * If we know we don't have enough space to make a cluster don't even
2533 * bother doing all the work to try and find one.
2535 if (ctl->free_space < bytes) {
2536 spin_unlock(&ctl->tree_lock);
2540 spin_lock(&cluster->lock);
2542 /* someone already found a cluster, hooray */
2543 if (cluster->block_group) {
2548 trace_btrfs_find_cluster(block_group, offset, bytes, empty_size,
2551 INIT_LIST_HEAD(&bitmaps);
2552 ret = setup_cluster_no_bitmap(block_group, cluster, &bitmaps, offset,
2554 cont1_bytes, min_bytes);
2556 ret = setup_cluster_bitmap(block_group, cluster, &bitmaps,
2557 offset, bytes + empty_size,
2558 cont1_bytes, min_bytes);
2560 /* Clear our temporary list */
2561 list_for_each_entry_safe(entry, tmp, &bitmaps, list)
2562 list_del_init(&entry->list);
2565 atomic_inc(&block_group->count);
2566 list_add_tail(&cluster->block_group_list,
2567 &block_group->cluster_list);
2568 cluster->block_group = block_group;
2570 trace_btrfs_failed_cluster_setup(block_group);
2573 spin_unlock(&cluster->lock);
2574 spin_unlock(&ctl->tree_lock);
2580 * simple code to zero out a cluster
2582 void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster)
2584 spin_lock_init(&cluster->lock);
2585 spin_lock_init(&cluster->refill_lock);
2586 cluster->root = RB_ROOT;
2587 cluster->max_size = 0;
2588 INIT_LIST_HEAD(&cluster->block_group_list);
2589 cluster->block_group = NULL;
2592 static int do_trimming(struct btrfs_block_group_cache *block_group,
2593 u64 *total_trimmed, u64 start, u64 bytes,
2594 u64 reserved_start, u64 reserved_bytes)
2596 struct btrfs_space_info *space_info = block_group->space_info;
2597 struct btrfs_fs_info *fs_info = block_group->fs_info;
2602 spin_lock(&space_info->lock);
2603 spin_lock(&block_group->lock);
2604 if (!block_group->ro) {
2605 block_group->reserved += reserved_bytes;
2606 space_info->bytes_reserved += reserved_bytes;
2609 spin_unlock(&block_group->lock);
2610 spin_unlock(&space_info->lock);
2612 ret = btrfs_error_discard_extent(fs_info->extent_root,
2613 start, bytes, &trimmed);
2615 *total_trimmed += trimmed;
2617 btrfs_add_free_space(block_group, reserved_start, reserved_bytes);
2620 spin_lock(&space_info->lock);
2621 spin_lock(&block_group->lock);
2622 if (block_group->ro)
2623 space_info->bytes_readonly += reserved_bytes;
2624 block_group->reserved -= reserved_bytes;
2625 space_info->bytes_reserved -= reserved_bytes;
2626 spin_unlock(&space_info->lock);
2627 spin_unlock(&block_group->lock);
2633 static int trim_no_bitmap(struct btrfs_block_group_cache *block_group,
2634 u64 *total_trimmed, u64 start, u64 end, u64 minlen)
2636 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2637 struct btrfs_free_space *entry;
2638 struct rb_node *node;
2644 while (start < end) {
2645 spin_lock(&ctl->tree_lock);
2647 if (ctl->free_space < minlen) {
2648 spin_unlock(&ctl->tree_lock);
2652 entry = tree_search_offset(ctl, start, 0, 1);
2654 spin_unlock(&ctl->tree_lock);
2659 while (entry->bitmap) {
2660 node = rb_next(&entry->offset_index);
2662 spin_unlock(&ctl->tree_lock);
2665 entry = rb_entry(node, struct btrfs_free_space,
2669 if (entry->offset >= end) {
2670 spin_unlock(&ctl->tree_lock);
2674 extent_start = entry->offset;
2675 extent_bytes = entry->bytes;
2676 start = max(start, extent_start);
2677 bytes = min(extent_start + extent_bytes, end) - start;
2678 if (bytes < minlen) {
2679 spin_unlock(&ctl->tree_lock);
2683 unlink_free_space(ctl, entry);
2684 kmem_cache_free(btrfs_free_space_cachep, entry);
2686 spin_unlock(&ctl->tree_lock);
2688 ret = do_trimming(block_group, total_trimmed, start, bytes,
2689 extent_start, extent_bytes);
2695 if (fatal_signal_pending(current)) {
2706 static int trim_bitmaps(struct btrfs_block_group_cache *block_group,
2707 u64 *total_trimmed, u64 start, u64 end, u64 minlen)
2709 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2710 struct btrfs_free_space *entry;
2714 u64 offset = offset_to_bitmap(ctl, start);
2716 while (offset < end) {
2717 bool next_bitmap = false;
2719 spin_lock(&ctl->tree_lock);
2721 if (ctl->free_space < minlen) {
2722 spin_unlock(&ctl->tree_lock);
2726 entry = tree_search_offset(ctl, offset, 1, 0);
2728 spin_unlock(&ctl->tree_lock);
2734 ret2 = search_bitmap(ctl, entry, &start, &bytes);
2735 if (ret2 || start >= end) {
2736 spin_unlock(&ctl->tree_lock);
2741 bytes = min(bytes, end - start);
2742 if (bytes < minlen) {
2743 spin_unlock(&ctl->tree_lock);
2747 bitmap_clear_bits(ctl, entry, start, bytes);
2748 if (entry->bytes == 0)
2749 free_bitmap(ctl, entry);
2751 spin_unlock(&ctl->tree_lock);
2753 ret = do_trimming(block_group, total_trimmed, start, bytes,
2759 offset += BITS_PER_BITMAP * ctl->unit;
2762 if (start >= offset + BITS_PER_BITMAP * ctl->unit)
2763 offset += BITS_PER_BITMAP * ctl->unit;
2766 if (fatal_signal_pending(current)) {
2777 int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group,
2778 u64 *trimmed, u64 start, u64 end, u64 minlen)
2784 ret = trim_no_bitmap(block_group, trimmed, start, end, minlen);
2788 ret = trim_bitmaps(block_group, trimmed, start, end, minlen);
2794 * Find the left-most item in the cache tree, and then return the
2795 * smallest inode number in the item.
2797 * Note: the returned inode number may not be the smallest one in
2798 * the tree, if the left-most item is a bitmap.
2800 u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root)
2802 struct btrfs_free_space_ctl *ctl = fs_root->free_ino_ctl;
2803 struct btrfs_free_space *entry = NULL;
2806 spin_lock(&ctl->tree_lock);
2808 if (RB_EMPTY_ROOT(&ctl->free_space_offset))
2811 entry = rb_entry(rb_first(&ctl->free_space_offset),
2812 struct btrfs_free_space, offset_index);
2814 if (!entry->bitmap) {
2815 ino = entry->offset;
2817 unlink_free_space(ctl, entry);
2821 kmem_cache_free(btrfs_free_space_cachep, entry);
2823 link_free_space(ctl, entry);
2829 ret = search_bitmap(ctl, entry, &offset, &count);
2830 /* Logic error; Should be empty if it can't find anything */
2834 bitmap_clear_bits(ctl, entry, offset, 1);
2835 if (entry->bytes == 0)
2836 free_bitmap(ctl, entry);
2839 spin_unlock(&ctl->tree_lock);
2844 struct inode *lookup_free_ino_inode(struct btrfs_root *root,
2845 struct btrfs_path *path)
2847 struct inode *inode = NULL;
2849 spin_lock(&root->cache_lock);
2850 if (root->cache_inode)
2851 inode = igrab(root->cache_inode);
2852 spin_unlock(&root->cache_lock);
2856 inode = __lookup_free_space_inode(root, path, 0);
2860 spin_lock(&root->cache_lock);
2861 if (!btrfs_fs_closing(root->fs_info))
2862 root->cache_inode = igrab(inode);
2863 spin_unlock(&root->cache_lock);
2868 int create_free_ino_inode(struct btrfs_root *root,
2869 struct btrfs_trans_handle *trans,
2870 struct btrfs_path *path)
2872 return __create_free_space_inode(root, trans, path,
2873 BTRFS_FREE_INO_OBJECTID, 0);
2876 int load_free_ino_cache(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
2878 struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
2879 struct btrfs_path *path;
2880 struct inode *inode;
2882 u64 root_gen = btrfs_root_generation(&root->root_item);
2884 if (!btrfs_test_opt(root, INODE_MAP_CACHE))
2888 * If we're unmounting then just return, since this does a search on the
2889 * normal root and not the commit root and we could deadlock.
2891 if (btrfs_fs_closing(fs_info))
2894 path = btrfs_alloc_path();
2898 inode = lookup_free_ino_inode(root, path);
2902 if (root_gen != BTRFS_I(inode)->generation)
2905 ret = __load_free_space_cache(root, inode, ctl, path, 0);
2908 printk(KERN_ERR "btrfs: failed to load free ino cache for "
2909 "root %llu\n", root->root_key.objectid);
2913 btrfs_free_path(path);
2917 int btrfs_write_out_ino_cache(struct btrfs_root *root,
2918 struct btrfs_trans_handle *trans,
2919 struct btrfs_path *path)
2921 struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
2922 struct inode *inode;
2925 if (!btrfs_test_opt(root, INODE_MAP_CACHE))
2928 inode = lookup_free_ino_inode(root, path);
2932 ret = __btrfs_write_out_cache(root, inode, ctl, NULL, trans, path, 0);
2934 btrfs_delalloc_release_metadata(inode, inode->i_size);
2936 printk(KERN_ERR "btrfs: failed to write free ino cache "
2937 "for root %llu\n", root->root_key.objectid);