In order to avoid the removal/creation of the slurm root cgroups
of the managed subsystems, disable the notify_on_release to prevent
the release agent from removing them automatically.

In order to be more deterministic in the removal of the different layers
of memcg involved when a step is running, do it manually by first acquiring
a lock on the root memcg to protect concurrent creations of step memcg from
invalid intermediate memcg directories in their own hierarchies.

Now relies on the task_pre_launch_priv to add tasks to the
step memcg and delete it at the end of the step.
slurmstepd no longer has to be added to any memcg for the
logic to work.

Prepare the use of task_pre_launch_priv to attach tasks to
the various cgroup subsystem in order to no longer have to
add the slurmstepd to the cgroup and let him belong to the
default cg of each subsystem.

This privileged call is executed just after the fork of each
forked task by user root before becoming the user. It enables
the task plugin to perform actions as a privileged user in
every task context.

Conflicts:
	doc/html/slurm.shtml

The workk for associations is still not complete

for passthrough gets removed on a dynamic system.

in error for passthrough.

user's pending allocation was started with srun and then for some reason
the slurmctld was brought down and while it was down the srun was removed.

previously it overwrote the poll_thread id

This is needed for when a free request is added to a block but there are
jobs finishing up so we don't start new jobs on the block since they will
fail on start.

Previously the node count would change from c-node count to midplane count
(but still be interpreted as a c-node count).

Missing spaces in dodump function of sacct

No real changes to logic other than some additional error checking.

Preempt jobs only when insufficient idle resources exist to start job,
regardless of the node weight.