who take the responsibility to write pid of NodeManager to hadoop-yarn/tasks when using cgroup

[upo]

If we mount the yarn.nodemanager.linux-container-executor.cgroups.mount-path=/sys/fs/cgroup manually, must we  echo "pid" > /sys/fs/cgroup/cpu/hadoop-yarn/tasks manually?

And if NodeManager auto mount the cgroup subsystem, then who take the responsibility to write pid of NodeManager to hadoop-yarn/tasks?