RE: [PATCH v4 3/9] cpufreq: amd-pstate: change amd-pstate driver to be built-in type

[Yuan, Perry]

[AMD Official Use Only - General]

Hi Mario.

> -----Original Message-----
> From: Limonciello, Mario <Mario.Limonciello@xxxxxxx>
> Sent: Tuesday, November 15, 2022 6:46 AM
> To: Yuan, Perry <Perry.Yuan@xxxxxxx>; rafael.j.wysocki@xxxxxxxxx; Huang,
> Ray <Ray.Huang@xxxxxxx>; viresh.kumar@xxxxxxxxxx
> Cc: Sharma, Deepak <Deepak.Sharma@xxxxxxx>; Fontenot, Nathan
> <Nathan.Fontenot@xxxxxxx>; Deucher, Alexander
> <Alexander.Deucher@xxxxxxx>; Huang, Shimmer
> <Shimmer.Huang@xxxxxxx>; Du, Xiaojian <Xiaojian.Du@xxxxxxx>; Meng,
> Li (Jassmine) <Li.Meng@xxxxxxx>; Karny, Wyes <Wyes.Karny@xxxxxxx>;
> linux-pm@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx
> Subject: Re: [PATCH v4 3/9] cpufreq: amd-pstate: change amd-pstate driver
> to be built-in type
> 
> On 11/10/2022 11:58, Perry Yuan wrote:
> > Change the `amd-pstate` driver as the built-in type which can help to
> > load the driver before the acpi_cpufreq driver as the default pstate
> > driver for the AMD processors.
> >
> > for the processors do not have the dedicated MSR functions, add
> > `amd-pstate=legacy_cppc` to grub which enable shared memory interface
> > to communicate with cppc_acpi module to control pstate hints.
> 
> Did you sync with Wyes already as I had suggested?  Was this the outcome?
> 
> I was a bit surprised to see this still as legacy_cppc when reviewing v4.

The discussion we made will be in V5 soon. 
I am working on it.


> 
> >
> > Signed-off-by: Perry Yuan <Perry.Yuan@xxxxxxx>
> > ---
> >   drivers/cpufreq/Kconfig.x86  |  2 +-
> >   drivers/cpufreq/amd-pstate.c | 25 +++++++++++++++----------
> >   2 files changed, 16 insertions(+), 11 deletions(-)
> >
> > diff --git a/drivers/cpufreq/Kconfig.x86 b/drivers/cpufreq/Kconfig.x86
> > index 310779b07daf..00476e94db90 100644
> > --- a/drivers/cpufreq/Kconfig.x86
> > +++ b/drivers/cpufreq/Kconfig.x86
> > @@ -35,7 +35,7 @@ config X86_PCC_CPUFREQ
> >   	  If in doubt, say N.
> >
> >   config X86_AMD_PSTATE
> > -	tristate "AMD Processor P-State driver"
> > +	bool "AMD Processor P-State driver"
> >   	depends on X86 && ACPI
> >   	select ACPI_PROCESSOR
> >   	select ACPI_CPPC_LIB if X86_64
> > diff --git a/drivers/cpufreq/amd-pstate.c
> > b/drivers/cpufreq/amd-pstate.c index ace7d50cf2ac..85a0b3fb56c2 100644
> > --- a/drivers/cpufreq/amd-pstate.c
> > +++ b/drivers/cpufreq/amd-pstate.c
> > @@ -59,10 +59,7 @@
> >    * we disable it by default to go acpi-cpufreq on these processors and add
> a
> >    * module parameter to be able to enable it manually for debugging.
> >    */
> > -static bool shared_mem = false;
> > -module_param(shared_mem, bool, 0444); -
> MODULE_PARM_DESC(shared_mem,
> > -		 "enable amd-pstate on processors with shared memory
> solution (false = disabled (default), true = enabled)");
> > +static bool shared_mem __read_mostly;
> >
> >   static struct cpufreq_driver amd_pstate_driver;
> >
> > @@ -653,16 +650,24 @@ static int __init amd_pstate_init(void)
> >
> >   	return ret;
> >   }
> > +device_initcall(amd_pstate_init);
> >
> > -static void __exit amd_pstate_exit(void)
> > +static int __init amd_pstate_param(char *str)
> >   {
> > -	cpufreq_unregister_driver(&amd_pstate_driver);
> > +	if (!str)
> > +		return -EINVAL;
> >
> > -	amd_pstate_enable(false);
> > -}
> > +	/*
> > +	 * support shared memory type CPPC which has no MSR function.
> > +	 * enable amd-pstate on processors with shared memory solution
> > +	 * (amd-pstate=legacy_cppc enabled), it is disabled by default.
> > +	 */
> > +	if (!strcmp(str, "legacy_cppc"))
> > +		shared_mem = true;
> >
> > -module_init(amd_pstate_init);
> > -module_exit(amd_pstate_exit);
> > +	return 0;
> > +}
> > +early_param("amd-pstate", amd_pstate_param);
> 
> Documentation/kernel-parameters.txt needs to be updated for this early
> parameter support.

Yeah, that will also be in V5 together.

> 
> >
> >   MODULE_AUTHOR("Huang Rui <ray.huang@xxxxxxx>");
> >   MODULE_DESCRIPTION("AMD Processor P-state Frequency Driver");