On 06/06/16 09:00, Dave Airlie wrote: > On 6 June 2016 at 16:30, Alejandro Piñeiro <apinhe...@igalia.com> wrote: >> On 03/06/16 23:04, Dave Airlie wrote: >>> On 4 June 2016 at 03:39, Alejandro Piñeiro <apinhe...@igalia.com> wrote: >>>> On 03/06/16 02:46, Dave Airlie wrote: >>>>> From: Dave Airlie <airl...@redhat.com> >>>>> >>>>> "all geometry shader output vertex count declarations in a >>>>> program must declare the same count." >>>> This spec quote lacks context. As far as I see it comes from GLSL 4.5 >>>> spec, "4.4.2.3 Geometry Outputs". >>>> >>>>> Fixes: >>>>> GL45-CTS.geometry_shader.output.conflicted_output_vertices_max >>>> The patch looks good. But testing on Skylake and Broadwell, current >>>> master is passing this test without this patch. Did you mean a different >>>> test, or you made the test with a different hw? >>> It might be a regression since, >>> aaa69c79cd584db4d9c6ea7794e93d29f3d54572 >>> >>> Can you confirm you are running on master, >> Yes. And just in case, I updated today, and also tried at that commit >> point, and just before. In all cases the test is passing on Skylake. >> > Can you send the results from TestResults.qpa for just this test, and > maybe MESA_GLSL=dump. > > I suppose it my be failing to link for some unrelated reason.
No, the problem was on my side. Checking qpa and MESA_GLSL=dump, I realized that in order to properly execute this test, I need to override the GL version to at least 4.4. The test is expecting the program to fail to link. Without the override, it fails because GLSL 440 is not supported. Overriding the GL version gets the program linking, so your patch is needed to get it properly failing. So sorry fo the noise. That patch is: Reviewed-by: Alejandro Piñeiro <apinhe...@igalia.com> BR _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
