Re: [PATCH v1 -next 0/3] RISC-V: ACPI: Enable CPPC based cpufreq support

From: Drew Fustini <drew@pdp7.com>
To: Sunil V L <sunilvl@ventanamicro.com>
Cc: linux-riscv@lists.infradead.org, linux-acpi@vger.kernel.org,
	linux-kernel@vger.kernel.org,
	"Rafael J . Wysocki" <rafael@kernel.org>,
	Len Brown <lenb@kernel.org>,
	Paul Walmsley <paul.walmsley@sifive.com>,
	Palmer Dabbelt <palmer@dabbelt.com>,
	Albert Ou <aou@eecs.berkeley.edu>,
	Viresh Kumar <viresh.kumar@linaro.org>,
	Conor Dooley <conor.dooley@microchip.com>,
	Andrew Jones <ajones@ventanamicro.com>,
	Atish Kumar Patra <atishp@rivosinc.com>,
	Anup Patel <apatel@ventanamicro.com>
Subject: Re: [PATCH v1 -next 0/3] RISC-V: ACPI: Enable CPPC based cpufreq support
Date: Tue, 19 Mar 2024 23:54:32 -0700	[thread overview]
Message-ID: <ZfqIKLnaAqOmXnWj@x1> (raw)
In-Reply-To: <Zflm5cje/+rnZ7HH@sunil-laptop>

On Tue, Mar 19, 2024 at 03:50:21PM +0530, Sunil V L wrote:
> On Mon, Mar 18, 2024 at 07:27:34PM -0700, Drew Fustini wrote:
> > On Mon, Mar 18, 2024 at 11:40:34AM -0700, Drew Fustini wrote:
> > > On Thu, Feb 08, 2024 at 09:14:11AM +0530, Sunil V L wrote:
> > > > This series enables the support for "Collaborative Processor Performance
> > > > Control (CPPC) on ACPI based RISC-V platforms. It depends on the
> > > > encoding of CPPC registers as defined in RISC-V FFH spec [2].
> > > > 
> > > > CPPC is described in the ACPI spec [1]. RISC-V FFH spec required to
> > > > enable this, is available at [2].
> > > > 
> > > > [1] - https://uefi.org/specs/ACPI/6.5/08_Processor_Configuration_and_Control.html#collaborative-processor-performance-control
> > > > [2] - https://github.com/riscv-non-isa/riscv-acpi-ffh/releases/download/v1.0.0/riscv-ffh.pdf
> > > > 
> > > > The series is based on the LPI support series.
> > > > Based-on: 20240118062930.245937-1-sunilvl@ventanamicro.com
> > > > (https://lore.kernel.org/lkml/20240118062930.245937-1-sunilvl@ventanamicro.com/)
> > > 
> > > Should the https://github.com/vlsunil/qemu/tree/lpi_exp branch also be
> > > used for this CPPC series too?
> > 
> > I noticed the ventanamicro qemu repo has a dev-upstream branch [1] which
> > contains 4bb6ba4d0fb9 ("riscv/virt: acpi: Enable CPPC - _CPC and _PSD").
> > I've built that but I still see 'SBI CPPC extension NOT detected!!' in
> > the Linux boot log.
> > 
> > I'm using upstream opensbi. It seems that sbi_cppc_probe() fails because
> > cppc_dev is not set. Nothing in the upstream opensbi repo seems to call
> > sbi_cppc_set_device(), so I am uncertain how it is possible for it to
> > work. Is there an opensbi branch I should be using?
> > 
> > Thanks,
> > Drew
> > 
> > [1] https://github.com/ventanamicro/qemu/tree/dev-upstream
> 
> Please use below branches for qemu and opensbi. These are just dummy
> objects/interfaces added to test the kernel change which are otherwise
> platform specific features.
> 
> https://github.com/vlsunil/qemu/tree/lpi_cppc_exp
> https://github.com/vlsunil/opensbi/tree/cppc_exp

Thank you for providing those branches.

The CPPC extension is now detected okay during boot but I see the
following in the boot log [1]:

[    1.094707] cpufreq: cpufreq_online: ->get() failed

Have you seen that too?

This appears to be from the following in drivers/cpufreq/cpufreq.c

1500         if (cpufreq_driver->get && has_target()) {
1501                 policy->cur = cpufreq_driver->get(policy->cpu);
1502                 if (!policy->cur) {
1503                         ret = -EIO;
1504                         pr_err("%s: ->get() failed\n", __func__);
1505                         goto out_destroy_policy;
1506                 }
1507         }

I'm not very familiar with the cpufreq code yet. My first thought is
that maybe the get() function pointer isn't set correctly?

thanks,
drew

[1] https://gist.github.com/pdp7/00ab047509d25871f87b3dc1b2445eaa