On Thu, 26 Jan 2012 12:46:03 -0800, Chad Versace <chad.vers...@linux.intel.com> 
wrote:
> On i965, _mesa_ir_link_shader is never called. As a consequence, the
> current fragment program (ctx->FragmentProgram->_Current) exists but does
> not differ from the fixed function fragment program
> (ctx->FragmentProgram->_TexEnvProgram). This confuses swrast.

I'd say "As a consequence, the current gl_fragment_program exists but is
invalid because its instructions were never filled in.  Avoid using the
empty program when possible by checking that the fragment shader is not
a translation of the current fixed function state.  This should also
result in some performance improvement by taking advantage of custom
fixed function paths where previously the Mesa IR interpreter was used."

I think it makes sense to me how the _swrast_update_fragment_program
hunk is needed.  So:

Reviewed-by: Eric Anholt <e...@anholt.net>

Attachment: pgpPA7rZNo07b.pgp
Description: PGP signature

_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev

Reply via email to