f2fs: lost matching-pair of trace in f2fs_truncate_inode_blocks
authorYubo Feng <fengyubo3@huawei.com>
Sat, 20 Jun 2020 08:39:43 +0000 (16:39 +0800)
committerJaegeuk Kim <jaegeuk@kernel.org>
Wed, 8 Jul 2020 04:51:47 +0000 (21:51 -0700)
if get_node_path() return -E2BIG and trace of
f2fs_truncate_inode_blocks_enter/exit enabled
then the matching-pair of trace_exit will lost
in log.

Signed-off-by: Yubo Feng <fengyubo3@huawei.com>
Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
fs/f2fs/node.c

index b4f4b0d..6551d5e 100644 (file)
@@ -1041,8 +1041,10 @@ int f2fs_truncate_inode_blocks(struct inode *inode, pgoff_t from)
        trace_f2fs_truncate_inode_blocks_enter(inode, from);
 
        level = get_node_path(inode, from, offset, noffset);
-       if (level < 0)
+       if (level < 0) {
+               trace_f2fs_truncate_inode_blocks_exit(inode, level);
                return level;
+       }
 
        page = f2fs_get_node_page(sbi, inode->i_ino);
        if (IS_ERR(page)) {