Re: [PATCH] cpufreq: intel_pstate: Fix compilation for !CONFIG_ACPI

From: Srinivas Pandruvada
Date: Tue Oct 23 2018 - 15:36:05 EST


On Tue, 2018-10-23 at 21:34 +0200, Dominik Brodowski wrote:
> On Tue, Oct 23, 2018 at 12:17:28PM -0700, Srinivas Pandruvada wrote:
> > On Tue, 2018-10-23 at 20:54 +0200, Dominik Brodowski wrote:
> > > Fixes: 86d333a8cc7f ("cpufreq: intel_pstate: Add base_frequency
> > > attribute")
> >
> > Thanks for the fix.
> >
> > > Cc: Srinivas Pandruvada <srinivas.pandruvada@xxxxxxxxxxxxxxx>
> > > Cc: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx>
> > > Cc: Len Brown <lenb@xxxxxxxxxx>
> > > Signed-off-by: Dominik Brodowski <linux@xxxxxxxxxxxxxxxxxxxx>
> > >
> > > diff --git a/drivers/cpufreq/intel_pstate.c
> > > b/drivers/cpufreq/intel_pstate.c
> > > index 49c0abf2d48f..50c5699970c5 100644
> > > --- a/drivers/cpufreq/intel_pstate.c
> > > +++ b/drivers/cpufreq/intel_pstate.c
> > > @@ -390,11 +390,6 @@ static int
> > > intel_pstate_get_cppc_guranteed(int
> > > cpu)
> > > static void intel_pstate_set_itmt_prio(int cpu)
> > > {
> > > }
> > > -
> > > -static int intel_pstate_get_cppc_guranteed(int cpu)
> > > -{
> > > - return -ENOTSUPP;
> > > -}
> >
> > What is ACPI is defined but SCHED_MC_PRIO is not defined?
> > Based on
> > "select ACPI_CPPC_LIB if X86_64 && ACPI && SCHED_MC_PRIO"
> >
> > So the above is still required. correct?
>
> Seems so, yes. Though that leads to either complicated #ifdefs or
> code
> duplications.
>
> In any case, I'd suggest marking at least nested #else and #endif
> lines
> with comments denoting which #ifdef they relate to, e.g.
>
> #else /* CONFIG_ACPI */
> #endif /* CONFIG_ACPI */
Will you submit a change?

Thanks,
Srinivas

>
> Thanks,
> Dominik