Message-ID: <7fa77e2239dd87004319a799dbea39988b718233.camel@linux.intel.com>
Date:   Tue, 07 Sep 2021 18:54:05 -0700
From:   Srinivas Pandruvada <srinivas.pandruvada@...ux.intel.com>
To:     xry111@...gyan1223.wang, rui.zhang@...el.com,
        daniel.lezcano@...aro.org, rjw@...ysocki.net,
        viresh.kumar@...aro.org, lenb@...nel.org
Cc:     linux-pm@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/2] cpufreq: intel_pstate: Process HWP Guaranteed
 change notification

Hi Wang,

I think this is Lenovo system. The FW sends some interrupts prematurely
before driver is ready,

There is a fix available but We have decided to revert the commit.

Please try the patch attached here.
https://bugzilla.kernel.org/show_bug.cgi?id=214329

Thanks,
Srinivas

On Wed, 2021-09-08 at 02:13 +0800, Xi Ruoyao wrote:
> Hi Srinivas,
> 
> Sorry for distrubing, but the mainline kernel panics on my system
> during
> boot. By reverting this commit the problem seems fixed.
> 
> I don't have kdump set up on the system, so I could only took a photo
> containing (a part of :( ) panic message.  The system is a laptop with
> Core i7-1065G7.  My kernel config is also attached.
> 
> Not sure if there is something wrong in this commit, or there is a
> firmware bug from the vendor of my laptop.  In the latter case, can we
> have something in kernel config or cmdline to disable HWP as a
> workaround?
> 
> On Thu, 2021-08-19 at 19:40 -0700, Srinivas Pandruvada wrote:
> > It is possible that HWP guaranteed ratio is changed in response to
> > change in power and thermal limits. For example when Intel Speed
> > Select
> > performance profile is changed or there is change in TDP, hardware
> > can
> > send notifications. It is possible that the guaranteed ratio is
> > increased. This creates an issue when turbo is disabled, as the old
> > limits set in MSR_HWP_REQUEST are still lower and hardware will clip
> > to older limits.
> > 
> > This change enables HWP interrupt and process HWP interrupts. When
> > guaranteed is changed, calls cpufreq_update_policy() so that driver
> > callbacks are called to update to new HWP limits. This callback
> > is called from a delayed workqueue of 10ms to avoid frequent updates.
> > 
> > Signed-off-by: Srinivas Pandruvada
> > <srinivas.pandruvada@...ux.intel.com>
> > ---
> >  drivers/cpufreq/intel_pstate.c | 39
> > ++++++++++++++++++++++++++++++++++
> >  1 file changed, 39 insertions(+)
> > 
> > diff --git a/drivers/cpufreq/intel_pstate.c
> > b/drivers/cpufreq/intel_pstate.c
> > index bb4549959b11..0fd2375c1f1e 100644
> > --- a/drivers/cpufreq/intel_pstate.c
> > +++ b/drivers/cpufreq/intel_pstate.c
> > @@ -32,6 +32,7 @@
> >  #include <asm/cpu_device_id.h>
> >  #include <asm/cpufeature.h>
> >  #include <asm/intel-family.h>
> > +#include "../drivers/thermal/intel/thermal_interrupt.h"
> >  
> >  #define INTEL_PSTATE_SAMPLING_INTERVAL (10 * NSEC_PER_MSEC)
> >  
> > @@ -219,6 +220,7 @@ struct global_params {
> >   * @sched_flags:       Store scheduler flags for possible cross CPU
> > update
> >   * @hwp_boost_min:     Last HWP boosted min performance
> >   * @suspended:         Whether or not the driver has been suspended.
> > + * @hwp_notify_work:   workqueue for HWP notifications.
> >   *
> >   * This structure stores per CPU instance data for all CPUs.
> >   */
> > @@ -257,6 +259,7 @@ struct cpudata {
> >         unsigned int sched_flags;
> >         u32 hwp_boost_min;
> >         bool suspended;
> > +       struct delayed_work hwp_notify_work;
> >  };
> >  
> >  static struct cpudata **all_cpu_data;
> > @@ -1625,6 +1628,40 @@ static void
> > intel_pstate_sysfs_hide_hwp_dynamic_boost(void)
> >  
> >  /************************** sysfs end ************************/
> >  
> > +static void intel_pstate_notify_work(struct work_struct *work)
> > +{
> > +       mutex_lock(&intel_pstate_driver_lock);
> > +       cpufreq_update_policy(smp_processor_id());
> > +       wrmsrl(MSR_HWP_STATUS, 0);
> > +       mutex_unlock(&intel_pstate_driver_lock);
> > +}
> > +
> > +void notify_hwp_interrupt(void)
> > +{
> > +       unsigned int this_cpu = smp_processor_id();
> > +       struct cpudata *cpudata;
> > +       u64 value;
> > +
> > +       if (!hwp_active || !boot_cpu_has(X86_FEATURE_HWP_NOTIFY))
> > +               return;
> > +
> > +       rdmsrl(MSR_HWP_STATUS, value);
> > +       if (!(value & 0x01))
> > +               return;
> > +
> > +       cpudata = all_cpu_data[this_cpu];
> > +       schedule_delayed_work_on(this_cpu, &cpudata->hwp_notify_work,
> > msecs_to_jiffies(10));
> > +}
> > +
> > +static void intel_pstate_enable_hwp_interrupt(struct cpudata
> > *cpudata)
> > +{
> > +       /* Enable HWP notification interrupt for guaranteed
> > performance change */
> > +       if (boot_cpu_has(X86_FEATURE_HWP_NOTIFY)) {
> > +               INIT_DELAYED_WORK(&cpudata->hwp_notify_work,
> > intel_pstate_notify_work);
> > +               wrmsrl_on_cpu(cpudata->cpu, MSR_HWP_INTERRUPT, 0x01);
> > +       }
> > +}
> > +
> >  static void intel_pstate_hwp_enable(struct cpudata *cpudata)
> >  {
> >         /* First disable HWP notification interrupt as we don't
> > process them */
> > @@ -1634,6 +1671,8 @@ static void intel_pstate_hwp_enable(struct
> > cpudata *cpudata)
> >         wrmsrl_on_cpu(cpudata->cpu, MSR_PM_ENABLE, 0x1);
> >         if (cpudata->epp_default == -EINVAL)
> >                 cpudata->epp_default = intel_pstate_get_epp(cpudata,
> > 0);
> > +
> > +       intel_pstate_enable_hwp_interrupt(cpudata);
> >  }
> >  
> >  static int atom_get_min_pstate(void)
> 

Powered by blists - more mailing lists