[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sun, 22 May 2016 10:04:34 +0200
From: Mike Galbraith <mgalbraith@...e.de>
To: Wanpeng Li <kernellwp@...il.com>
Cc: Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...nel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
Pavan Kondeti <pkondeti@...eaurora.org>,
Ben Segall <bsegall@...gle.com>,
Matt Fleming <matt@...eblueprint.co.uk>,
Morten Rasmussen <morten.rasmussen@....com>,
Paul Turner <pjt@...gle.com>,
Thomas Gleixner <tglx@...utronix.de>,
Byungchul Park <byungchul.park@....com>,
Andrew Hunter <ahh@...gle.com>
Subject: Re: [PATCH 2/3] sched,fair: Fix local starvation
On Sun, 2016-05-22 at 15:42 +0800, Wanpeng Li wrote:
> 2016-05-22 15:32 GMT+08:00 Mike Galbraith <mgalbraith@...e.de>:
> > On Sun, 2016-05-22 at 15:27 +0800, Wanpeng Li wrote:
> >
> > > What's the meaning of 'x-cpu wakeup'? ;-)
> >
> > Generally, cross CPU, as in waker/wakee reside on different CPUs,
> > but
> > in this case, it's cross socket wakeup.
>
> Do you mean wakeup wakees on remote socket don't imply
> migration/normalized, why?
Because the wakee is NOT necessarily migrated simply because it lives
in some remote cache domain. It was a simple but nasty booboo.
ttwu():
cpu = select_task_rq(p, p->wake_cpu, SD_BALANCE_WAKE, wake_flags);
if (task_cpu(p) != cpu) {
wake_flags |= WF_MIGRATED;
set_task_cpu(p, cpu);
set_task_cpu():
if (task_cpu(p) != new_cpu) {
if (p->sched_class->migrate_task_rq)
p->sched_class->migrate_task_rq(p);
^^^^^^^^^^^^^^^^^^
migrate_task_rq_fair() normalizes wakee, those wakees that did not
migrate have NOT been normalized, leaving two flavors of wakee on the
wake_list, with no discriminator. Store class information internally,
and the x-socket information disconnect evaporates.
-Mike
Powered by blists - more mailing lists