On Thursday, October 01, 2015 06:50:26 PM Jason Ekstrand wrote: > This series does a bunch of code-shuffling with the end objective of > getting the gl_* data structures out of the backend compiler. Now that we > have NIR, we would like the compiler to be a NIR -> binary translator and > not be loaded full of mesa and GL-specific stuff. This series gets us 95% > of the way there. Unfortunately, we still have to pass gl_program into > geometry shaders for transform-fedback and, for fragment shaders, we need > to pass in gl_shader_program so that we can fix up TEXTURE_RECTANGLE on > older hardware (pre-Broadwell). > > Incidentally, according to our CI system, this series fixes 2 piglit tests > on Iron Lake and GM45. I have no idea why. > > Jason Ekstrand (13): > i965/shader: Pull assign_common_binding_table_offsets out of > backend_shader > i965: Move binding table setup to codegen time. > i965: Move prog_data uniform setup to the codegen level > nir/glsl: Take a gl_shader_program and a stage rather than a gl_shader > nir: Add a a nir_shader_info struct > nir: Move GS data to nir_shader_info > i965/backend_shader: Add a field to store the NIR shader > i965/cs: Remove the prog argument from local_id_payload_dwords > i965/fs: Move sampler unit lookup into rescale_texcoord > i965/fs: Use the nir info instead of pulling things out of > [shader_]prog > i965/vec4: Use nir info instead of pulling things out of [shader_]prog > i965/fs,vec4: Get rid of the sanity_param_count > i965/shader: Get rid of the shader, prog, and shader_prog fields
Series is: Reviewed-by: Kenneth Graunke <kenn...@whitecape.org> I used this to read patch 13, as it never seemed to hit the mailing list for some reason or other: http://cgit.freedesktop.org/~jekstrand/mesa/log/?h=wip/compiler-separation
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
