LLVMpipe FP64 Support Knocks Off Some GL4 Extensions

While LLVMpipe tends to be an afterthought in supporting new OpenGL extensions within Mesa/Gallium3D and is in need of some help, David Airlie managed to land some improvements for it today in Mesa by adding support for double-precision floating-points.

David crossed off the LLVMpipe FP64 support to now support the ARB_gpu_shader_fp64 extension needed by OpenGL 4.0 and the ARB_vertex_attrib_64bit extension needed by OpenGL 4.1. Up to now only the Nouveau NVC0 and Softpipe drivers have supported the said extensions. The support landed with this Git commit.

ARB_gpu_shader_fp64 allows for shaders to use double-precision floating-point data types. ARB_vertex_attrib_64bit extends the 64-bit floating-point support by adding the support to vertex shader inputs. This work will be found with the upcoming Mesa 10.7 release.