f2fs: don't do checkpoint if error is occurred

If we met an error during the dentry recovery, we should not conduct checkpoint.
Otherwise, some errorneous dentry blocks overwrites the existing blocks that
contain the remaining recovery information.

Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c
index 3a4b51c..5148d90 100644
--- a/fs/f2fs/recovery.c
+++ b/fs/f2fs/recovery.c
@@ -403,6 +403,7 @@
 	destroy_fsync_dnodes(sbi, &inode_list);
 	kmem_cache_destroy(fsync_entry_slab);
 	sbi->por_doing = 0;
-	write_checkpoint(sbi, false);
+	if (!err)
+		write_checkpoint(sbi, false);
 	return err;
 }