[PATCH 2/2] sched/fair: Honor asymmetric SMT priority in idle selection

Andrea Righi arighi at nvidia.com
Fri Sep 11 15:34:35 PDT 2026


On Fri, Sep 11, 2026 at 04:11:38PM +0200, Dietmar Eggemann wrote:
> On 09.09.26 08:26, Andrea Righi wrote:
> 
> [...]
> 
> > +static bool has_asym_smt_domain(int cpu)
> > +{
> > +	struct sched_domain *sd;
> > +
> > +	for_each_domain(cpu, sd) {
> > +		if (!(sd->flags & SD_SHARE_CPUCAPACITY))
> > +			break;
> > +
> > +		if (sd->flags & SD_ASYM_PACKING)
> > +			return true;
> > +	}
> > +
> > +	return false;
> > +}
> 
> Replace for_each_domain() w/ sd = rcu_dereference_all(cpu_rq(cpu)->sd)
> here as well?

Correct, I've changed it locally already to inspect the lowest scheduling domain
directly.

Thanks,
-Andrea



More information about the linux-arm-kernel mailing list