Re: [RFC PATCH v5 4/7] sched: bias task wakeups to preferredsemi-idle packages

From: Balbir Singh
Date: Mon Dec 15 2008 - 03:46:58 EST


* Peter Zijlstra <peterz@xxxxxxxxxxxxx> [2008-12-15 09:33:04]:

> On Mon, 2008-12-15 at 09:25 +0100, Peter Zijlstra wrote:
> > On Mon, 2008-12-15 at 12:31 +0530, Balbir Singh wrote:
> >
> > > > kernel/sched_fair.c | 17 +++++++++++++++++
> > > > 1 files changed, 17 insertions(+), 0 deletions(-)
> > > >
> > > > diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c
> > > > index 98345e4..939f2a1 100644
> > > > --- a/kernel/sched_fair.c
> > > > +++ b/kernel/sched_fair.c
> > > > @@ -1027,6 +1027,23 @@ static int wake_idle(int cpu, struct task_struct *p)
> > > > cpumask_t tmp;
> > > > struct sched_domain *sd;
> > > > int i;
> > > > + unsigned int chosen_wakeup_cpu;
> > > > + int this_cpu;
> > > > +
> > > > + /*
> > > > + * At POWERSAVINGS_BALANCE_WAKEUP level, if both this_cpu and prev_cpu
> > > > + * are idle and this is not a kernel thread and this task's affinity
> > > > + * allows it to be moved to preferred cpu, then just move!
> > > > + */
> > > > +
> > > > + this_cpu = smp_processor_id();
> > > > + chosen_wakeup_cpu =
> > > > + cpu_rq(this_cpu)->rd->sched_mc_preferred_wakeup_cpu;
> > > > +
> > > > + if (sched_mc_power_savings >= POWERSAVINGS_BALANCE_WAKEUP &&
> > > > + idle_cpu(cpu) && idle_cpu(this_cpu) && p->mm &&
> > >
> > > The p->mm check is racy, it needs to be done under task_lock(). The
> > > best way to check for a kernel thread is get_task_mm(), followed by
> > > put_task_mm() is the mm is not NULL. We also need to check to see if
> > > the task is _hot_ on cpu. We should negate this optimization in case
> > > chosen_wakeup_cpu is idle, so check for that as well.
> >
> > Sure its racy, but so what?
> >
> > The worst I can see it that we exclude a dying task from this logic,
> > which isn't a problem at all, since its dying anyway.
>
> At which point I seriously doubt it'd still be on the rq anyway.
>

I forgot to mention that, the check should be (p->mm && !(p->flags & PF_KTHREAD))
--
Balbir
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/