[Mesa-dev] [PATCH] st/mesa: expose OES_geometry_shader and OES_texture_cube_map_array

[Marek Olšák]

On Mon, Aug 29, 2016 at 5:41 PM, Ilia Mirkin <imirkin at alum.mit.edu> wrote:
> On Mon, Aug 29, 2016 at 11:35 AM, Marek Olšák <maraeo at gmail.com> wrote:
>> On Sat, Aug 27, 2016 at 11:53 PM, Ilia Mirkin <imirkin at alum.mit.edu> wrote:
>>> Signed-off-by: Ilia Mirkin <imirkin at alum.mit.edu>
>>> ---
>>>  docs/features.txt                      |  4 ++--
>>>  docs/relnotes/12.1.0.html              |  4 ++--
>>>  src/mesa/state_tracker/st_extensions.c | 14 ++++++++++++++
>>>  3 files changed, 18 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/docs/features.txt b/docs/features.txt
>>> index 26e8ff7..4c755c6 100644
>>> --- a/docs/features.txt
>>> +++ b/docs/features.txt
>>> @@ -260,7 +260,7 @@ GLES3.2, GLSL ES 3.2:
>>>    GL_OES_copy_image                                     DONE (all drivers)
>>>    GL_OES_draw_buffers_indexed                           DONE (all drivers that support GL_ARB_draw_buffers_blend)
>>>    GL_OES_draw_elements_base_vertex                      DONE (all drivers)
>>> -  GL_OES_geometry_shader                                DONE (i965/gen8+)
>>> +  GL_OES_geometry_shader                                DONE (i965/gen8+, nvc0, radeonsi)
>>>    GL_OES_gpu_shader5                                    DONE (all drivers that support GL_ARB_gpu_shader5)
>>>    GL_OES_primitive_bounding_box                         not started
>>>    GL_OES_sample_shading                                 DONE (i965, nvc0, r600, radeonsi)
>>> @@ -271,7 +271,7 @@ GLES3.2, GLSL ES 3.2:
>>>    GL_OES_tessellation_shader                            started (Ken)
>>>    GL_OES_texture_border_clamp                           DONE (all drivers)
>>>    GL_OES_texture_buffer                                 DONE (i965, nvc0, radeonsi)
>>> -  GL_OES_texture_cube_map_array                         DONE (i965/gen8+)
>>> +  GL_OES_texture_cube_map_array                         DONE (i965/gen8+, nvc0, radeonsi)
>>>    GL_OES_texture_stencil8                               DONE (all drivers that support GL_ARB_texture_stencil8)
>>>    GL_OES_texture_storage_multisample_2d_array           DONE (all drivers that support GL_ARB_texture_multisample)
>>>
>>> diff --git a/docs/relnotes/12.1.0.html b/docs/relnotes/12.1.0.html
>>> index d22d14b..f77ef91 100644
>>> --- a/docs/relnotes/12.1.0.html
>>> +++ b/docs/relnotes/12.1.0.html
>>> @@ -57,8 +57,8 @@ Note: some of the new features are only available with certain drivers.
>>>  <li>GL_KHR_blend_equation_advanced on i965</li>
>>>  <li>GL_KHR_texture_compression_astc_sliced_3d on i965</li>
>>>  <li>GL_OES_copy_image on nv50, nvc0, r600, radeonsi, softpipe, llvmpipe</li>
>>> -<li>GL_OES_geometry_shader on i965/gen8+</li>
>>> -<li>GL_OES_texture_cube_map_array on i965/gen8+</li>
>>> +<li>GL_OES_geometry_shader on i965/gen8+, nvc0, radeonsi</li>
>>> +<li>GL_OES_texture_cube_map_array on i965/gen8+, nvc0, radeonsi</li>
>>>  </ul>
>>>
>>>  <h2>Bug fixes</h2>
>>> diff --git a/src/mesa/state_tracker/st_extensions.c b/src/mesa/state_tracker/st_extensions.c
>>> index f86a5a3..c7ec10c 100644
>>> --- a/src/mesa/state_tracker/st_extensions.c
>>> +++ b/src/mesa/state_tracker/st_extensions.c
>>> @@ -946,6 +946,15 @@ void st_init_extensions(struct pipe_screen *screen,
>>>        extensions->ARB_tessellation_shader = GL_TRUE;
>>>     }
>>>
>>> +   /* Ideally this should also check that invocations are supported. In
>>> +    * practice, all of the hw that supports ES 3.1 also supports multiple
>>> +    * invocations.
>>> +    */
>>> +   if (screen->get_shader_param(screen, PIPE_SHADER_GEOMETRY,
>>> +                                PIPE_SHADER_CAP_MAX_INSTRUCTIONS) > 0) {
>>> +      extensions->OES_geometry_shader = GL_TRUE;
>>> +   }
>>
>> This should also check ARB_gpu_shader5, which adds support for
>> multiple GS invocations.
>
> Yeah, I just thought it was a little heavy to check for gs5 when all
> we wanted were multiple invocations. And I didn't really want to go
> through and add a new cap. I guess the safe thing is to add a gs5 dep
> and then when some hw comes along that wants OES_geom but not ARB_gs5,
> we can worry about it then.

Yeah that sounds good.

Marek