Btrfs: fix panic in balance due to EIO
authorLiu Bo <bo.li.liu@oracle.com>
Tue, 12 Jul 2016 17:29:37 +0000 (10:29 -0700)
committerDavid Sterba <dsterba@suse.com>
Tue, 26 Jul 2016 11:52:25 +0000 (13:52 +0200)
commit0fd8c3dae14fb64947842472940b807ca0781da9
treea45c5235f5267739564328a3ee0ba6ecf96e73fe
parentbaf863b9c29617cc9eaf24e039f58846e700db48
Btrfs: fix panic in balance due to EIO

During build_backref_tree(), if we fail to read a btree node,
we can eventually run into BUG_ON(cache->nr_nodes) that we put
in backref_cache_cleanup(), meaning we have at least one
memory leak.

This frees the backref_node that we's allocated at the very
beginning of build_backref_tree().

Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
fs/btrfs/relocation.c