RE: [PATCH v1 00/11] amd-pstate CPU Performance Scaling Driver

["Penny, Zheng" <penny.zheng@xxxxxxx>]

[AMD Official Use Only - AMD Internal Distribution Only]

Hi,

> -----Original Message-----
> From: Jan Beulich <jbeulich@xxxxxxxx>
> Sent: Tuesday, December 3, 2024 5:01 PM
> To: Penny, Zheng <penny.zheng@xxxxxxx>
> Cc: Stabellini, Stefano <stefano.stabellini@xxxxxxx>; Huang, Ray
> <Ray.Huang@xxxxxxx>; Ragiadakou, Xenia <Xenia.Ragiadakou@xxxxxxx>;
> Andryuk, Jason <Jason.Andryuk@xxxxxxx>; Andrew Cooper
> <andrew.cooper3@xxxxxxxxxx>; Roger Pau Monné <roger.pau@xxxxxxxxxx>; Julien
> Grall <julien@xxxxxxx>; Stefano Stabellini <sstabellini@xxxxxxxxxx>; Anthony
> PERARD <anthony.perard@xxxxxxxxxx>; xen-devel@xxxxxxxxxxxxxxxxxxxx
> Subject: Re: [PATCH v1 00/11] amd-pstate CPU Performance Scaling Driver
>
> On 03.12.2024 09:11, Penny Zheng wrote:
> > amd-pstate is the AMD CPU performance scaling driver that introduces a
> > new CPU frequency control mechanism on modern AMD APU and CPU series
> > in Xen. The new mechanism is based on Collaborative Processor
> > Performance Control (CPPC) which provides finer grain frequency
> > management than legacy ACPI hardware P-States. Current AMD CPU/APU
> > platforms are using the ACPI P-states driver to manage CPU frequency
> > and clocks with switching only in 3 P-states. CPPC replaces the ACPI
> > P-states controls and allows a flexible, low-latency interface for Xen
> > to directly communicate the performance hints to hardware.
> >
> > amd_pstate CPPC has 2 operation modes: autonomous (active) mode, and
> > non-autonomous (passive) mode. We register different CPUFreq driver
> > for different modes, "amd-pstate" for passive mode and "amd-pstate-epp"
> > for active mode.
> >
> > The passive mode leverages common governors such as *ondemand*,
> > *performance*, etc, to manage the performance hints. And the active
> > mode uses epp to provides a hint to the hardware if software wants to
> > bias toward performance (0x0) or energy efficiency (0xff). CPPC power
> > algorithm will calculate the runtime workload and adjust the realtime
> > cpu cores frequency according to the power supply and thermal, core
> > voltage and some other hardware conditions.
> >
> > amd-pstate is enabled with a top-level cpufreq=amd-pstate option. It
> > will fallback to cpufreq=xen if amd-pstate is unavailable.
> >
> > With `cpufreq=amd-pstate,active`, We did a 60s sampling test to see
> > the CPU frequency change, through tweaking the energy_perf preference
> > from `xenpm set-cpufreq-cppc powersave` to `xenpm set-cpufreq-cppc
> performance`.
> > The outputs are as follows:
> > ```
> > Setting CPU in powersave mode
> > Sampling and Outputs:
> >   Avg freq      2000000 KHz
> >   Avg freq      2000000 KHz
> >   Avg freq      2000000 KHz
> > Setting CPU in performance mode
> > Sampling and Outputs:
> >   Avg freq      4640000 KHz
> >   Avg freq      4220000 KHz
> >   Avg freq      4640000 KHz
> > ```
> >
> > Penny Zheng (11):
> >   xen/x86: add CPPC feature flag for AMD processors
> >   xen/x86: introduce new sub-hypercall to get CPPC data
> >   xen/x86: introduce "cpufreq=amd-pstate" xen cmdline
> >   xen/x86: get processor max speed from DMI table
> >   xen/x86: introduce a new amd pstate driver for cpufreq scaling
> >   xen/cpufreq: introduce policy type when cpufreq_driver->setpolicy
> >     exists
> >   xen/cpufreq: only set gov NULL when cpufreq_driver.target() exists
> >   x86/cpufreq: add "cpufreq=amd-pstate,active" para
> >   xen/x86: implement EPP support for the AMD processors
> >   tools/xenpm: Print CPPC parameters for amd-pstate driver
> >   xen/cpufreq: Adapt SET/GET_CPUFREQ_CPPC xen_sysctl_pm_op for
> >     amd-pstate driver
>
> Just to clarify: While certainly fine, it is a little surprising to see such 
> a submission
> just after the submission deadline for 4.20 was passed.

So sorry.... Forgot we're in feature frozen window

> This is intended for 4.21 then, I expect? Or else have you talked to the 
> release
> manager?

Yes, it is for 4.21

>
> Jan

Many thanks,
Penny,