[igt-dev] [PATCH i-g-t] Fix IGT Warnings due to dynamic subtest inclusion

[Zbigniew Kempczyński]

On Mon, Nov 09, 2020 at 09:16:01PM -0800, Dixit, Ashutosh wrote:
> On Mon, 09 Nov 2020 20:55:10 -0800, <nidhi1.gupta at intel.com> wrote:
> >
> > From: Nidhi Gupta <nidhi1.gupta at intel.com>
> >
> > Signed-off-by: Nidhi Gupta <nidhi1.gupta at intel.com>
> > ---
> >  tests/kms_atomic_interruptible.c | 14 ++++++++++++++
> >  1 file changed, 14 insertions(+)
> >
> > diff --git a/tests/kms_atomic_interruptible.c b/tests/kms_atomic_interruptible.c
> > index 27bb6a2b..f98023ca 100644
> > --- a/tests/kms_atomic_interruptible.c
> > +++ b/tests/kms_atomic_interruptible.c
> > @@ -288,7 +288,9 @@ igt_main
> >		for_each_pipe_with_valid_output(&display, pipe, output) {
> >			igt_dynamic_f("%s-pipe-%s", igt_output_name(output), kmstest_pipe_name(pipe))
> >				run_plane_test(&display, pipe, output, test_legacy_modeset, DRM_PLANE_TYPE_PRIMARY);
> > +			{
> >				break;
> > +			}
> >		}
> 
> Are we trying to break after the first call to run_plane_test() here? In
> that case why are we using a for loop (for_each_pipe_with_valid_output)? I
> am not sure if igt_subtest_with_dynamic/igt_dynamic_f combinations can take
> a break statement, this code pattern is certainly not used elsewhere.
> 
> So the patch does fix the warnings (which is a relief) but otherwise I
> don't know if it makes sense. If not, I'd say revert the previous patch and
> do it a different way.

IMO patch doesn't makes sense. @Nidhi - please decide which "for" loop you 
want to break and resend.

--
Zbigneiw