sample mask input to fs shouldn't force persample execution

[airlied]

From my reading of the spec, using gl_SampleId or gl_SamplePosition should force per sample execution, however the spec doesn't seem to mention the gl_SampleMask input causing per sample.

AMDVLK sets m_pResUsage->builtInUsage.fs.runAtSampleRate for gl_SampleMask.

[cainguo]

I didn't see AMDVLK sets m_pResUsage->builtInUsage.fs.runAtSampleRate for gl_SampleMask, but only for gl_SampleMaskIn.

[Kai-AMD]

Hi Airlied,

We are now working on this issue.

Thanks for your help!

Best Regards,
Kai

[jinjianrong]

Fix is merged to master branch