Skip to content

Very late CMOD propagation and lift a restriction on conditional-discard optimization

Ian Romanick requested to merge idr/mesa:review/very-late-cmod-prop into master

While looking at something unrelated, I notice a case where CMOD propagation should have happened. I think spent almost an hour trying to figure out what the bug in CMOD propagation was. It turns out, the optimization pass was fine. The opportunity for it to work just didn't exist until after the scheduler moved some code around.

Hopefully this won't make too awful of a rebase for @mattst88.

Merge request reports