[Mesa-dev] Request for information on Radeon

[Ilia Mirkin]

On Sun, Jun 8, 2014 at 6:16 PM, Matt Turner <mattst88 at gmail.com> wrote:
> On Sun, Jun 8, 2014 at 3:12 PM, Ilia Mirkin <imirkin at alum.mit.edu> wrote:
>> On Sun, Jun 8, 2014 at 6:02 PM, Matt Turner <mattst88 at gmail.com> wrote:
>>> On Sun, Jun 8, 2014 at 2:40 PM, Aditya Avinash <adityaavinash1 at gmail.com> wrote:
>>>> Can I join them? I have good knowledge on Tessellation using proprietary
>>>> drivers.
>>>
>>> I don't think Summer of Code allows people to work together on the
>>> same project. Then again, I don't know that anyone has heard from
>>> Fabian (Cc'd) since he sent some piglit tests at the end of March. His
>>> github activity seems to have stopped around the same time. Since the
>>> coding start date was May 19th, I'm a bit concerned.
>>>
>>> GSoC's midterm date is June 27th, less than three weeks away. It's an
>>> awkward situation in which we want to let Fabian do the project he was
>>> accepted for, but if we can't get in contact with him we don't want to
>>> prevent others from making progress. Let's see if we hear something
>>> from him soon.
>>>
>>>> I don't know where to start. Can you provide me some documentation on it?
>>>
>>> Presumably you want to implement tessellation for Radeon hardware?
>>> There's a lot of common infrastructure needed first (the GLSL compiler
>>> lives in src/glsl).
>>
>> What was the scope of the GSoC? Even if GLSL + mesa/core supports it,
>> a bunch of stuff would need to be piped through to gallium, for
>> nouveau and radeon to support it. (And the nvc0 driver already has a
>> bunch of provisional support for it, waiting for the gallium
>> interfaces to materialize.)
>
> The short description is "Implementation of the
> ARB_tessellation_shader extension in Mesa This includes * Piglit tests
> * Mesa front-end implementation * GLSL compiler support for two new
> shader targets * driver back-end implementation"
>
> The backend implementation is for IvyBridge. I'm not sure if I'm
> allowed or should post more of the proposal.
>
> But, I'm not sure what point you're making?

I was more asking than making a point :) The implicit suggestion was
that if one's goal was to add support to radeon, and the GSoC didn't
cover gallium (as Chris just mentioned it doesn't), then one could
still make a meaningful contribution by adding the proper interfaces
to Gallium without stepping on the toes of someone doing the glsl/mesa
core work.

  -ilia