Re: [PATCH 7/7 v3] sched: fix wrong utilization accounting when switching to fair class

From	Peter Zijlstra <peterz@infradead.org>
Newsgroups	linux.kernel
Subject	Re: [PATCH 7/7 v3] sched: fix wrong utilization accounting when switching to fair class
Date	2016-09-26 17:00 +0200
Message-ID	<slFxT-567-1@gated-at.bofh.it> (permalink)
References	(4 earlier) <si0hz-4BJ-9@gated-at.bofh.it> <si2jo-5SZ-35@gated-at.bofh.it> <sjrSp-2rY-11@gated-at.bofh.it> <sjsYa-3nb-41@gated-at.bofh.it> <skbiy-6i0-17@gated-at.bofh.it>
Organization	linux.* mail to news gateway

Show all headers | View raw

On Thu, Sep 22, 2016 at 02:25:18PM +0200, Peter Zijlstra wrote:
> On Tue, Sep 20, 2016 at 03:06:04PM +0200, Vincent Guittot wrote:
> > > Also, even it things magically worked out, its still very icky to mix
> > > the normalized vruntime into things.
> > 
> > I agree
> 
> In any case, I pushed out a bunch of patches to:
> 
>   git://git.kernel.org/pub/scm/linux/kernel/git/peterz/queue.git sched/tmp
> 
> they appear to be able to build a kernel, but that's not much testing.

Hurmph, its buggered for cgroups. When we get the newidle lockbreak
between dequeue and put_prev_task, the load-balance pull can add a task
to the 'current' cgroup.

Which suggests we cannot do that enqueue_entity() optimization, unless
we fix that dequeue+put atomicy muck.

Back to linux.kernel | Previous | Next | Find similar | Unroll thread

Thread

Re: [PATCH 7/7 v3] sched: fix wrong utilization accounting when  switching to fair class Peter Zijlstra <peterz@infradead.org> - 2016-09-26 17:00 +0200

csiph-web