Geometry Shaders For Mesa Not Yet Ready

While there's been early code available for several months, Mesa support for OpenGL Geometry Shaders still isn't ready for merging into mainline Mesa.

While Mesa 9.1 has picked up many features over the past six months, OpenGL Geometry Shaders won't be found in the release due out by late February.

Geometry Shaders are GLSL programs that deal with primitives and can yield more primitives. Geometry Shaders have been part of the OpenGL Core specification since version 3.2 with the ARB_geometry_shader4 extension.

There's been early OpenGL GS Mesa code going back to last summer thanks to a Mesa branch that was worked on by Bryan Cain. While it ended up progressing and back in September it looked like Geometry Shaders would merge soon for Mesa, that didn't end up happening. Geometry Shaders are one of the last holdouts for Mesa being able to advertise itself as OpenGL 3.2 friendly.

Earlier this month when writing about R600 Gallium3D Getting Close On OpenGL 3.3 Support, David Airlie wrote about the state of geometry shaders in the Phoronix Forums. "There is a tree on github with unproven geometry shader code for r600g, intel would need to add 965 support, it just needs a lot of piglit tests and a lot of review...but writing the piglit tests for geometry shaders is probably the biggest problem." (Also see The State Of Open-Source Radeon Driver Features.)

Brian Paul, Mesa founder and current VMware employee, asked today on the Mesa developers' mailing list about the state of Geometry Shaders for this open-source graphics stack. Intel's Paul Berry responded with the current open-source state of affairs for OpenGL Geometry Shaders:

It looks like Bryan made a lot of headway before he set the branch aside in October, so I'm planning to use that as a starting point. My impression from talking to Bryan is that he doesn't have too much time/interest to put into geometry shaders these days, so for the moment I'm planning to take over responsibility for the branch myself.

My first order of business is to write some piglit tests for geometry shaders, so that I don't have to push anything to Mesa master that isn't well-tested. I have an nVidia system that I can use as a reference platform to make sure my tests are correct. I also want to spend a little bit of time prototyping some i965 back-end code just to increase my familiarity with the problem domain and to find out if i965 hardware has any sharp corners I need to watch out for. Once I've done those two things, my plan is to start rebasing Bryan's patch series onto master and sending it out for review piece by piece.

Geometry shaders haven't risen to the top of Intel's priority list quite yet, but they're letting me work on them part-time for now, so progress is likely to be slow for a while. I'll keep the mailing list updated as to my progress. And I'll try to make sure that major design decisions get discussed here on the list so that no one gets left out of the loop.

So getting the GS support cleaned-up and merged isn't at the top of Intel's TODO list, but at least they are working on it. This won't happen for Mesa 9.1 obviously, but hopefully this OpenGL 3.2 feature will be baked and ready for tasting by the next Mesa release six months down the road. The follow-up release will be Mesa 9.2 or more than likely it will be dubbed Mesa 10.0 assuming a major new OpenGL revision is reached (likely OpenGL 3.2/3.3 compliance in one go).