Btrfs: fix sleep with spinlock held during unmount

The code to free block groups needs to drop the space info spin lock
before calling btrfs_remove_free_space_cache (which can schedule).

This is safe because at unmount time, nobody else is going to play
with the block groups.

Signed-off-by: Chris Mason <chris.mason@oracle.com>

diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 5258923d621f0b3030deba2cf9a5451d38711d3c..fe4e11b31a43898a64aec09db50667e488cb853e 100644 (file)
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -3157,9 +3157,13 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
                block_group = rb_entry(n, struct btrfs_block_group_cache,
                                       cache_node);
 
+               spin_unlock(&info->block_group_cache_lock);
                btrfs_remove_free_space_cache(block_group);
+               spin_lock(&info->block_group_cache_lock);
+
                rb_erase(&block_group->cache_node,
                         &info->block_group_cache_tree);
+
                spin_lock(&block_group->space_info->lock);
                list_del(&block_group->list);
                spin_unlock(&block_group->space_info->lock);