[SER] HitObject_MakeNop|Miss DXIL opcodes and verification tests

[simoll]

• DXIL opcodes for HitObject_MakeNop and HitObject_MakeMiss
• DXV validation test

[bob80905]

LGTM. Any validation rules could be added separately.

[damyanp]

LGTM. Any validation rules could be added separately.

@bob80905 - are there validation rules that need adding? If so, can we make sure that there are issues tracking adding them please, @simoll?

[bob80905]

LGTM. Any validation rules could be added separately.

@bob80905 - are there validation rules that need adding? If so, can we make sure that there are issues tracking adding them please, @simoll?

From looking at the spec, I noticed this:

HitObject cannot be stored in memory buffers, in groupshared memory, in ray payloads, or in intersection attributes. 

I think this should be a validation rule, though I could see it being much more efficient to add all the hit object variants first, then one PR to add the validation rule and test them all. So that doesn't have to be added here.

[simoll]

LGTM. Any validation rules could be added separately.

@bob80905 - are there validation rules that need adding? If so, can we make sure that there are issues tracking adding them please, @simoll?

There is a checkbox specifically for validation tests for these opcodes in the SER umbrella bug #7214 now.

[simoll]

LGTM. Any validation rules could be added separately.

@bob80905 - are there validation rules that need adding? If so, can we make sure that there are issues tracking adding them please, @simoll?

From looking at the spec, I noticed this:

HitObject cannot be stored in memory buffers, in groupshared memory, in ray payloads, or in intersection attributes. 

I think this should be a validation rule, though I could see it being much more efficient to add all the hit object variants first, then one PR to add the validation rule and test them all. So that doesn't have to be added here.

Added an explicit issue to track HitObject validation / diagnostics for unsupported uses: #7234

[simoll]

@tex3d @bob80905 - is this one good to merge? No planned changes from my side at this point.

[bogner]

LGTM