btrfs: fix deadlock when writing out space cache

If we fail to prepare our pages for whatever reason (out of memory in
our case) we need to make sure to drop the block_group->data_rwsem,
otherwise hilarity ensues.

Signed-off-by: Josef Bacik <jbacik@fb.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Liu Bo <bo.li.liu@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ add label and use existing unlocking code ]
Signed-off-by: David Sterba <dsterba@suse.com>
This commit is contained in:
Josef Bacik 2017-11-15 16:20:52 -05:00 committed by David Sterba
parent 8e138e0d92
commit b77000ed55

View file

@ -1264,7 +1264,7 @@ static int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
/* Lock all pages first so we can lock the extent safely. */
ret = io_ctl_prepare_pages(io_ctl, inode, 0);
if (ret)
goto out;
goto out_unlock;
lock_extent_bits(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1,
&cached_state);
@ -1358,6 +1358,7 @@ static int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
out_nospc:
cleanup_write_cache_enospc(inode, io_ctl, &cached_state);
out_unlock:
if (block_group && (block_group->flags & BTRFS_BLOCK_GROUP_DATA))
up_write(&block_group->data_rwsem);