Keith Whitwell wrote: > On Mon, 2009-01-26 at 10:10 -0800, Patrice Mandin wrote: >> Hello, >> >> [Excuse me if it's not the right place to talk about Gallium, is >> dri-devel a better place?] >> >> Last week I fixed a bug in nv30 gallium backend for nouveau, related to >> GL_EXT_blend_equation_separate (and not GL_EXT_blend_func_separate as I >> wrote in my commit), which seems to be a requirement for gallium >> drivers, as visible in gallium-0.2 branch, >> src/gallium/state_trackers/glx/dri/dri_extensions.c >> >> NV30 gpus (Geforce FX class) don't support it, so I would need to >> remove it from the extension list; and maybe make it a PIPE_CAP_??? >> like some other OpenGL stuff. >> >> My question are: is the current list of extensions for Gallium backend >> fixed (specific requirements from some shader engines?), and must all >> backend support said extensions? >> > > This is a bit of a missing chunk of functionality. > > There are some queries that the pipe_screen interface exposes, and these > could/should be used by the Mesa state tracker to figure out which > extensions to turn on and off for a particular driver... but currently, > we've just got some ugly bit of legacy code in each winsys which turns > on a fairly arbitary list. > > Until the state tracker starts using queries to build the extension > list, this type of bug will be common. > > If you're interested in figuring out how to do this, I think it'd be a > great contribution to Gallium. I think this ties into but is not > necessarily dependent on the code I've got in state_trackers/glx/dri > which is a start on hoisting out a bunch of duplicated code from all the > dri winsys's.
This++. ~ C. ------------------------------------------------------------------------------ This SF.net email is sponsored by: SourcForge Community SourceForge wants to tell your story. http://p.sf.net/sfu/sf-spreadtheword _______________________________________________ Mesa3d-dev mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/mesa3d-dev
