Re: [PATCH v2 2/2] nohz: make nohz_full imply isolcpus

[Chris Metcalf]

On 04/08/2015 05:41 AM, Peter Zijlstra wrote:
> On Mon, Apr 06, 2015 at 02:16:45PM -0400, cmetcalf@xxxxxxxxxx wrote:
> > From: Chris Metcalf <cmetcalf@xxxxxxxxxx>
> >
> > It's not clear that nohz_full is useful without isolcpus also
> > set, since otherwise the scheduler has to run periodically to
> > try to determine whether to steal work from other cores.
> So the Changelog and the patch don't seem to agree with one another.
>
> The Changelog states that nohz_full should depend on isolcpus.

The git commit message says "make nohz_full imply isolcpus".
That's consistent with the code.

> The patch implies nohz_full for isolcpus.
>
> These are not the same; and I don't see the argument for the former make
> sense for the latter.
>
> In specific isolcpus without nohz_full does make sense.
>
> > Signed-off-by: Chris Metcalf <cmetcalf@xxxxxxxxxx>
> > Acked-by: Mike Galbraith <umgwanakikbuti@xxxxxxxxx> ["thumbs up!"]
> > Acked-by: Rik van Riel <riel@xxxxxxxxxx>
> > ---
> >   kernel/sched/core.c | 1 +
> >   1 file changed, 1 insertion(+)
> >
> > diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> > index f0f831e8a345..275f12c608f2 100644
> > --- a/kernel/sched/core.c
> > +++ b/kernel/sched/core.c
> > @@ -6836,6 +6836,7 @@ static int init_sched_domains(const struct cpumask *cpu_map)
> >   	doms_cur = alloc_sched_domains(ndoms_cur);
> >   	if (!doms_cur)
> >   		doms_cur = &fallback_doms;
> > +	tick_nohz_full_set_cpus(cpu_isolated_map);
> >   	cpumask_andnot(doms_cur[0], cpu_map, cpu_isolated_map);
> >   	err = build_sched_domains(doms_cur[0], NULL);
> >   	register_sched_domain_sysctl();

--
Chris Metcalf, EZChip Semiconductor
http://www.ezchip.com

--
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/