[Mesa-dev] [PATCH 10/16] i965/fs: Eliminate null-dst instructions without side-effects.

[Connor Abbott]

> That makes sense.  However, in the vec4 backend, we do emit scratch reads
> and writes prior to optimization - in move_grf_array_access_to_scratch().
> I've been meaning to port that over to the scalar backend.  At which
> point, SHADER_OPCODE_GEN4_SCRATCH_WRITE would be a problem.

Sort of off-topic, but it would be really nice if you could do this. A
lot of the code Jason wrote for NIR is about dealing with indirect
array accesses, which currently get lowered to if-ladders in GLSL
which makes all of it dead code. It would be nice to be able to bugfix
+ test it.