[igt-dev] [PATCH i-g-t 2/6] i915/i915_pm_rpm.c: create PC state subtest group

[Chris Wilson]

Quoting Anshuman Gupta (2020-12-09 16:25:02)
> On 2020-12-09 at 16:25:55 +0000, Chris Wilson wrote:
> > Quoting Anshuman Gupta (2020-12-09 16:06:38)
> > > Create a separate igt test group and move package C
> > > state in to this subgroup.
> > > Run powertop --auto-tune to tune SOC power configuration
> > > for package C state tests.
> > > 
> > > Signed-off-by: Anshuman Gupta <anshuman.gupta at intel.com>
> > > ---
> > >  tests/i915/i915_pm_rpm.c | 35 +++++++++++++++++++++++++++++++----
> > >  1 file changed, 31 insertions(+), 4 deletions(-)
> > > 
> > > diff --git a/tests/i915/i915_pm_rpm.c b/tests/i915/i915_pm_rpm.c
> > > index af55b569..42bc44d9 100644
> > > --- a/tests/i915/i915_pm_rpm.c
> > > +++ b/tests/i915/i915_pm_rpm.c
> > > @@ -832,6 +832,25 @@ static void basic_subtest(void)
> > >         /* XXX Also we can test wake up via exec nop */
> > >  }
> > >  
> > > +static bool setup_powertop(void)
> > > +{
> > > +       FILE *fp;
> > > +       char tmp[512];
> > > +
> > > +       fp = popen("powertop --auto-tune", "r");
> > 
> > Doesn't this defeat the point of having it work out of the box?
> > -Chris
> Yes, but there is no other way as PC state are being dependent on other
> non-gfx devices as well. These test kept disabled since gen9 platforms.

Could you be explicit in the steps required so that it is clear that
powertop is not working around our own bugs?
-Chris