[igt-dev] [PATCH i-g-t] i915/i915_pm_rpm: Split the planes into dynamic subtests

[Chris Wilson]

Quoting Petri Latvala (2020-04-22 09:33:49)
> On Wed, Apr 22, 2020 at 09:12:58AM +0100, Chris Wilson wrote:
> > Use the dynamic subtests to allow the user to individually run the
> > per-plane rpm tests.
> > 
> > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > Cc: Petri Latvala <petri.latvala at intel.com>
> > ---
> >  tests/i915/i915_pm_rpm.c | 18 +++++++-----------
> >  1 file changed, 7 insertions(+), 11 deletions(-)
> > 
> > diff --git a/tests/i915/i915_pm_rpm.c b/tests/i915/i915_pm_rpm.c
> > index 4f8124dc4..a34e78b6b 100644
> > --- a/tests/i915/i915_pm_rpm.c
> > +++ b/tests/i915/i915_pm_rpm.c
> > @@ -1778,7 +1778,7 @@ static void test_one_plane(bool dpms, uint32_t plane_id,
> >  /* This one also triggered WARNs on our driver at some point in time. */
> >  static void planes_subtest(bool universal, bool dpms)
> >  {
> > -     int i, rc, planes_tested = 0, crtc_idx;
> > +     int i, rc, crtc_idx;
> >       drmModePlaneResPtr planes;
> >  
> >       igt_require(default_mode_params);
> > @@ -1803,8 +1803,8 @@ static void planes_subtest(bool universal, bool dpms)
> >  
> >                       type = universal ? get_plane_type(plane->plane_id) :
> >                                          PLANE_OVERLAY;
> > -                     test_one_plane(dpms, plane->plane_id, type);
> > -                     planes_tested++;
> > +                     igt_dynamic_f("plane-%d\n", plane->plane_id)
> > +                             test_one_plane(dpms, plane->plane_id, type);
> >               }
> >               drmModeFreePlane(plane);
> >       }
> > @@ -1813,10 +1813,6 @@ static void planes_subtest(bool universal, bool dpms)
> >       if (universal) {
> >               rc = drmSetClientCap(drm_fd, DRM_CLIENT_CAP_UNIVERSAL_PLANES, 0);
> >               igt_assert_eq(rc, 0);
> > -
> > -             igt_assert_lte(3, planes_tested);
> > -     } else {
> > -             igt_assert_lte(1, planes_tested);
> 
> Sure, we cannot have these asserts anymore when people can run a
> limited set. Do we need to check them somewhere else?
> 
> Hang on, what was the point of these asserts? Asserting that we have
> at least 3 universal planes? And at least 1 of non-universal? Wtf?

I'm glad I wasn't the only one asking that :)

This isn't even a test for the universal/legacy API, it's just checking
that we don't get runtime errors when poking them.
-Chris