f2fs: fix to skip shrinking extent nodes
authorChao Yu <chao2.yu@samsung.com>
Thu, 22 Oct 2015 10:24:12 +0000 (18:24 +0800)
committerJaegeuk Kim <jaegeuk@kernel.org>
Thu, 22 Oct 2015 16:39:35 +0000 (09:39 -0700)
In f2fs_shrink_extent_tree we should stop shrink flow if we have already
shrunk enough nodes in extent cache.

Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
fs/f2fs/extent_cache.c

index a38ee9b..7ddba81 100644 (file)
@@ -620,7 +620,7 @@ unsigned int f2fs_shrink_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
                        write_unlock(&et->lock);
 
                        if (node_cnt + tree_cnt >= nr_shrink)
-                               break;
+                               goto unlock_out;
                }
        }
 unlock_out: