nir: crocus: iris: Implement finalize_nir callback
This MR includes a couple small changes to NIR and crocus as part of the initial clean up / refactoring.
Improves performance of SynMark OglDrvShComp by +241.879%±1.01366% (n=5) on a random KBL desktop that I have. That seems to put it at about the same performance as i965, but I did not test that in a statistically sound way.
@kwg had a similar commit about a year ago (kwg/mesa@35d0a59c). In that commit message he said:
Improves performance in SynMark OglDrvShComp (a compiler microbenchmark) by 2.1x on my i7-7700HQ. This is still much slower than i965, though.
I'm not sure if that "2.1x" mean +210% or +110%. My result seems different in that the delta with i965 seems pretty small now. I was using a i7-7700K, so the CPUs were pretty similar. Both are Kaby Lake, but mine is a much higher clocked desktop part.