[OpenGL] Issue when passing gl_FrontColor through Tessellation & Geometry shader stages

[vsarthou]

Hello,

I am using OpenGL 4.6 in Compatibility Profile mode on an HD Graphics 630, and I am encountering a rendering issue when using the gl_FrontColor built-in shader variable through Tessellation and Geometry shader stages.

It seems that the value of gl_FrontColor incorrectly becomes (0,0,0,0) at some point through the stages.

I am not experiencing this issue on Nvidia or AMD GPUs.

I made a minimal reproducer that I attached in 3 versions: the executable, the source code, and an OpenGL trace generated using apitrace (https://github.com/apitrace/apitrace).

The reproducer renders a full-window triangle, and sets a different gl_FrontColor value for each vertex.

The expected rendering can be seen in OK.png.

The rendering I get is just an all-black frame buffer (KO.png).

Other info:

• The issue is still reproducible if I only use a Geometry shader (Vert => Geom => Frag)
• The issue is still reproducible if I only use Tessallation shaders (Vert => TessCtrl => TessEval => Frag)
• There is no issue if I don't use Tessellation or Geometry shader stages (only Vertex + Fragment).

More detailed info about my setup:

System Setup Information:
-----------------------------------------

CPU: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
GPU: Intel® UHD Graphics 630
Graphics Driver Version: 31.0.101.2115
Operating System: Windows 10 Enterprise
OS Version: 21H2 19044.2364
API: OpenGL
Occurs on non-Intel GPUs?: No

Steps to Reproduce:
-------------------------------
- Launch attached reproducer executable OR replay attached OpenGL trace using qapitrace (https://github.com/apitrace/apitrace)

Expected Result:
-------------------------------
- Rendering contains colors as seen in OK.png

Actual Result:
-------------------------------
- Rendering is all black as seen in KO.png

Additional Information:
-------------------------------
- The source code of the reproducer is attached (FrontColorIssue.c)

Thanks in advance

[Hugo_Intel]

Hello vsarthou

Thank you for posting on the Intel Communities.

Please allow us more time to check the information that you have provided so far. I will post back once I have an update for you or in case a need more information from you.

Best Regards,

Hugo O.

Intel Customer Support Technician.

[Hugo_Intel]

Hello vsarthou

Thank you for your patience. Please let us know the following information:

• Are you contacting us on behalf of a company for a specific type of project? Or, Are you an independent developer?
• Are you able to test this on and 11th Generation processor or newer?
• Please run Intel® System Support Utility and attach the log file to get more information about your system.

Best Regards,

Hugo O.

Intel Customer Support Technician.

[vsarthou]

Hello Hugo,

Thank you for your reply.

• Yes I am contacting you on behalf of my company for a specific project.
• I was able to test my reproducer on this CPU: 12th Gen Intel(R) Core(TM) i7-12700T
  And the bug is still reproducible.
• You will find attached the report generated by Intel System Support Utility (sys_info.txt)

Regards,

[Hugo_Intel]

Hello vsarthou

Thank you for sharing these details with us. We will continue researching the information you have provided so far and I will post back as soon as I have an update for you.

Best Regards,

Hugo O.

Intel Customer Support Technician.