Btrfs: don't skip writing out a empty block groups cache

I noticed a slight bug where we will not bother writing out the block group
cache's space cache if it's space tree is empty.  Since it could have a cluster
or pinned extents that need to be written out this is just not a valid test.
Thanks,

Signed-off-by: Josef Bacik <josef@redhat.com>

diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
index 0a8ccdbdd464c1e9d56d7f6f616b9050f87143e2..b81556ca75eab5e4b7d70a82aecf6fad60508e29 100644 (file)
--- a/fs/btrfs/free-space-cache.c
+++ b/fs/btrfs/free-space-cache.c
@@ -575,10 +575,6 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
 
        INIT_LIST_HEAD(&bitmap_list);
 
-       node = rb_first(&ctl->free_space_offset);
-       if (!node)
-               return -1;
-
        if (!i_size_read(inode))
                return -1;
 
@@ -639,6 +635,12 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
        if (block_group)
                start = block_group->key.objectid;
 
+       node = rb_first(&ctl->free_space_offset);
+       if (!node && cluster) {
+               node = rb_first(&cluster->root);
+               cluster = NULL;
+       }
+
        /* Write out the extent entries */
        do {
                struct btrfs_free_space_entry *entry;