On Thu, Jun 5, 2025 at 6:14 PM Dave Hansen <dave.hansen@xxxxxxxxx> wrote:
>
> On 6/5/25 08:07, Rafael J. Wysocki wrote:
> > #ifdef CONFIG_ACPI_PROCESSOR_CSTATE
> > +void acpi_idle_rescan_dead_smt_siblings(void)
> > +{
> > + if (cpuidle_get_driver() == &acpi_idle_driver)
> > + arch_cpu_rescan_dead_smt_siblings();
> > +}
>
> My only thought in reading this is that maybe cpuidle_register_driver()
> would be a better spot to force the arch_cpu_rescan_dead_smt_siblings().
> That way, each driver would not have to do the rescan.
Unfortunately, this wouldn't work in the current arrangement of things
because cpuidle_register_driver() can be called in a CPU online path.
It should be possible to make this work in the future, but first things first.
> But that's just a little nit at worst, otherwise the series looks good
> to me. Thanks for chasing this down.
>
> For the x86 bits:
>
> Acked-by: Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx>
Thank you!
