Re: [PATCH v3] cpuset: restore sanity to cpuset_cpus_allowed_fallback()

From: Tejun Heo
Date: Wed Jun 12 2019 - 11:43:39 EST


Hello, Joel.

On Wed, Jun 12, 2019 at 11:13:15AM -0400, Joel Savitz wrote:
> In the case that a process is constrained by taskset(1) (i.e.
> sched_setaffinity(2)) to a subset of available cpus, and all of those are
> subsequently offlined, the scheduler will set tsk->cpus_allowed to
> the current value of task_cs(tsk)->effective_cpus.
>
> This is done via a call to do_set_cpus_allowed() in the context of
> cpuset_cpus_allowed_fallback() made by the scheduler when this case is
> detected. This is the only call made to cpuset_cpus_allowed_fallback()
> in the latest mainline kernel.
>
> However, this is not sane behavior.

So, I fully agree the whole thing is insane. It has always been so
and the root cause of the problem is that we aren't distinguishing
restrictions put on by the cpuset and individual tasks' configured
cpumask. Ultimately, for this to make actual sense, we need to
separate out per-task effective and configured and cpuset-imposed
masks.

That said, the suggested behavior seems way better than the current
one, so no objection from me, but can you please repost w/ scheduler
ppl cc'd?

Thanks.

--
tejun