Re: [PATCH v3 3/3] [RFC] CPUFreq: Add support for cpu-perf-dependencies

[Rafael J. Wysocki]

On Tue, Nov 17, 2020 at 11:53 AM Viresh Kumar <viresh.kumar@xxxxxxxxxx> wrote:
>
> On 17-11-20, 10:47, Nicola Mazzucato wrote:
> > Freq-invariance has been mentioned. I suppose the fix will depend on which
> > strategy we prefer to solve this.
>
> I am not sure how FIE will use this information, as I thought the
> problem is about knowing the current frequency on a CPU instead of
> which CPUs are related. Anyway, EM is good enough to get this stuff
> going.
>
> > As a reminder, two solutions:
> > 1) dependent_cpus cpumask in cpufreq and involved entities pick this info
>
> Yeah, this one. And it will be called freqdomain_cpus. Add support for
> freqdomain_cpus in core, update acpi-cpufreq to reuse it instead of
> adding its own and you can have it in other drivers then.

Is this really a cpufreq thing, though, or is it arch stuff?  I think
the latter, because it is not necessary for anything in cpufreq.

Yes, acpi-cpufreq happens to know this information, because it uses
processor_perflib, but the latter may as well be used by the arch
enumeration of CPUs and the freqdomain_cpus mask may be populated from
there.

As far as cpufreq is concerned, if the interface to the hardware is
per-CPU, there is one CPU per policy and cpufreq has no business
knowing anything about the underlying hardware coordination.