Re: [PATCH] cpufreq: intel_pstate: Override parameters if HWP forced by BIOS

From: Srinivas Pandruvada
Date: Thu Sep 09 2021 - 10:53:14 EST


On Thu, 2021-09-09 at 06:30 -0700, Doug Smythies wrote:
> On Wed, Sep 8, 2021 at 11:33 PM Srinivas Pandruvada
> <srinivas.pandruvada@xxxxxxxxxxxxxxx> wrote:
> >
> > On Wed, 2021-09-08 at 20:48 -0700, Doug Smythies wrote:
> > > If HWP has been already been enabled by BIOS, it may be
> > > necessary to override some kernel command line parameters.
> > > Once it has been enabled it requires a reset to be disabled.
> > >
> > > Signed-off-by: Doug Smythies <dsmythies@xxxxxxxxx>
> > > ---
> > >  drivers/cpufreq/intel_pstate.c | 22 ++++++++++++++++------
> > >  1 file changed, 16 insertions(+), 6 deletions(-)
> > >
> > > diff --git a/drivers/cpufreq/intel_pstate.c
> > > b/drivers/cpufreq/intel_pstate.c
> > > index bb4549959b11..073bae5d4498 100644
> > > --- a/drivers/cpufreq/intel_pstate.c
> > > +++ b/drivers/cpufreq/intel_pstate.c
> > > @@ -3267,7 +3267,7 @@ static int __init intel_pstate_init(void)
> > >                  */
> > >                 if ((!no_hwp &&
> > > boot_cpu_has(X86_FEATURE_HWP_EPP)) ||
> > >                     intel_pstate_hwp_is_enabled()) {
> > > -                       hwp_active++;
> > > +                       hwp_active = 1;
> > Why this change?
>
> It was just to keep it at 1, but I agree not absolutely needed.
>
> >
> > >                         hwp_mode_bdw = id->driver_data;
> > >                         intel_pstate.attr = hwp_cpufreq_attrs;
> > >                         intel_cpufreq.attr = hwp_cpufreq_attrs;
> > > @@ -3347,17 +3347,27 @@ device_initcall(intel_pstate_init);
> > >
> > >  static int __init intel_pstate_setup(char *str)
> > >  {
> > > +       /*
> > > +        * If BIOS is forcing HWP, then parameter
> > > +        * overrides might be needed. Only print
> > > +        * the message once, and regardless of
> > > +        * any overrides.
> > > +        */
> > > +       if(!hwp_active
> > This part of code is from early_param, Is it possible that
> > hwp_active is not 0?
>
> Not at this point, in any testing I did.
> But I do not know the authoritative answer
> to your question.
>
But as you explained you want to prevent repeated print of
"HWP enabled by BIOS". So you need this.

> >
> > > && boot_cpu_has(X86_FEATURE_HWP))
> > > +               if(intel_pstate_hwp_is_enabled()){
> > > +                       pr_info("HWP enabled by BIOS\n");
> > > +                       hwp_active = 1;
> > > +               }
> > >         if (!str)
> > >                 return -EINVAL;
> > >
> > > -       if (!strcmp(str, "disable"))
> > > +       if (!strcmp(str, "disable") && !hwp_active)
> > >                 no_load = 1;
> > > -       else if (!strcmp(str, "active"))
> > > +       if (!strcmp(str, "active"))
> > >                 default_driver = &intel_pstate;
> > > -       else if (!strcmp(str, "passive"))
> > > +       if (!strcmp(str, "passive"))
> > >                 default_driver = &intel_cpufreq;
> >
> > Why "else if" changed to "if" ?
>
> Because it doesn't matter anyway and I would
> have had to figure out another qualifier.
> This way, and given that this executes once per
> intel_pstate command line parameter, the code
> executes the way it used to, overall.
If someone specified intel_pstate=active, it will also compare with
"passive" with this change.

Thanks,
Srinivas

>
> >
> >
> > Thanks,
> > Srinivas
> >
> > > -
> > > -       if (!strcmp(str, "no_hwp")) {
> > > +       if (!strcmp(str, "no_hwp") && !hwp_active) {
> > >                 pr_info("HWP disabled\n");
> > >                 no_hwp = 1;
> > >         }
> >
> >