Btrfs: extent_map and data=ordered fixes for space balancing
* Add an EXTENT_BOUNDARY state bit to keep the writepage code
from merging data extents that are in the process of being
relocated. This allows us to do accounting for them properly.
* The balancing code relocates data extents indepdent of the underlying
inode. The extent_map code was modified to properly account for
things moving around (invalidating extent_map caches in the inode).
* Don't take the drop_mutex in the create_subvol ioctl. It isn't
required.
* Fix walking of the ordered extent list to avoid races with sys_unlink
* Change the lock ordering rules. Transaction start goes outside
the drop_mutex. This allows btrfs_commit_transaction to directly
drop the relocation trees.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 48a3dc0..4516fbf 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -135,7 +135,7 @@
BUG_ON(num_bytes > btrfs_super_total_bytes(&root->fs_info->super_copy));
mutex_lock(&BTRFS_I(inode)->extent_mutex);
- btrfs_drop_extent_cache(inode, start, start + num_bytes - 1);
+ btrfs_drop_extent_cache(inode, start, start + num_bytes - 1, 0);
mutex_unlock(&BTRFS_I(inode)->extent_mutex);
while(num_bytes > 0) {
@@ -163,7 +163,7 @@
break;
}
btrfs_drop_extent_cache(inode, start,
- start + ins.offset - 1);
+ start + ins.offset - 1, 0);
}
mutex_unlock(&BTRFS_I(inode)->extent_mutex);
@@ -587,7 +587,7 @@
btrfs_drop_extent_cache(inode, ordered_extent->file_offset,
ordered_extent->file_offset +
- ordered_extent->len - 1);
+ ordered_extent->len - 1, 0);
mutex_unlock(&BTRFS_I(inode)->extent_mutex);
ins.objectid = ordered_extent->start;
@@ -880,7 +880,7 @@
int ret = 0, nr_unlink = 0, nr_truncate = 0;
/* don't do orphan cleanup if the fs is readonly. */
- if (root->inode->i_sb->s_flags & MS_RDONLY)
+ if (root->fs_info->sb->s_flags & MS_RDONLY)
return;
path = btrfs_alloc_path();
@@ -892,8 +892,6 @@
btrfs_set_key_type(&key, BTRFS_ORPHAN_ITEM_KEY);
key.offset = (u64)-1;
- trans = btrfs_start_transaction(root, 1);
- btrfs_set_trans_block_group(trans, root->inode);
while (1) {
ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
@@ -933,7 +931,7 @@
* crossing root thing. we store the inode number in the
* offset of the orphan item.
*/
- inode = btrfs_iget_locked(root->inode->i_sb,
+ inode = btrfs_iget_locked(root->fs_info->sb,
found_key.offset, root);
if (!inode)
break;
@@ -965,7 +963,9 @@
* do a destroy_inode
*/
if (is_bad_inode(inode)) {
+ trans = btrfs_start_transaction(root, 1);
btrfs_orphan_del(trans, inode);
+ btrfs_end_transaction(trans, root);
iput(inode);
continue;
}
@@ -988,7 +988,6 @@
printk(KERN_INFO "btrfs: truncated %d orphans\n", nr_truncate);
btrfs_free_path(path);
- btrfs_end_transaction(trans, root);
}
void btrfs_read_locked_inode(struct inode *inode)
@@ -1343,8 +1342,7 @@
u64 mask = root->sectorsize - 1;
if (root->ref_cows)
- btrfs_drop_extent_cache(inode,
- new_size & (~mask), (u64)-1);
+ btrfs_drop_extent_cache(inode, new_size & (~mask), (u64)-1, 0);
path = btrfs_alloc_path();
path->reada = -1;
BUG_ON(!path);
@@ -1677,7 +1675,7 @@
hole_start, 0, 0,
hole_size, 0);
btrfs_drop_extent_cache(inode, hole_start,
- (u64)-1);
+ (u64)-1, 0);
btrfs_check_file(root, inode);
}
mutex_unlock(&BTRFS_I(inode)->extent_mutex);
@@ -1843,6 +1841,24 @@
args->root == BTRFS_I(inode)->root);
}
+struct inode *btrfs_ilookup(struct super_block *s, u64 objectid,
+ struct btrfs_root *root, int wait)
+{
+ struct inode *inode;
+ struct btrfs_iget_args args;
+ args.ino = objectid;
+ args.root = root;
+
+ if (wait) {
+ inode = ilookup5(s, objectid, btrfs_find_actor,
+ (void *)&args);
+ } else {
+ inode = ilookup5_nowait(s, objectid, btrfs_find_actor,
+ (void *)&args);
+ }
+ return inode;
+}
+
struct inode *btrfs_iget_locked(struct super_block *s, u64 objectid,
struct btrfs_root *root)
{
@@ -3266,7 +3282,7 @@
btrfs_put_ordered_extent(ordered);
}
}
- btrfs_drop_extent_cache(inode, 0, (u64)-1);
+ btrfs_drop_extent_cache(inode, 0, (u64)-1, 0);
kmem_cache_free(btrfs_inode_cachep, BTRFS_I(inode));
}
@@ -3412,16 +3428,22 @@
{
struct list_head *head = &root->fs_info->delalloc_inodes;
struct btrfs_inode *binode;
+ struct inode *inode;
unsigned long flags;
spin_lock_irqsave(&root->fs_info->delalloc_lock, flags);
while(!list_empty(head)) {
binode = list_entry(head->next, struct btrfs_inode,
delalloc_inodes);
- atomic_inc(&binode->vfs_inode.i_count);
+ inode = igrab(&binode->vfs_inode);
+ if (!inode)
+ list_del_init(&binode->delalloc_inodes);
spin_unlock_irqrestore(&root->fs_info->delalloc_lock, flags);
- filemap_write_and_wait(binode->vfs_inode.i_mapping);
- iput(&binode->vfs_inode);
+ if (inode) {
+ filemap_write_and_wait(inode->i_mapping);
+ iput(inode);
+ }
+ cond_resched();
spin_lock_irqsave(&root->fs_info->delalloc_lock, flags);
}
spin_unlock_irqrestore(&root->fs_info->delalloc_lock, flags);