[PATCH] D28489: [CodeGen] Move MacroFusion to the target

[Matthias Braun via Phabricator via llvm-commits]

MatzeB added a comment.

In https://reviews.llvm.org/D28489#659179, @MatzeB wrote:

> The targets add the DAG Mutators anyway, so you should be able to remove the whole `shouldScheduleAdjacent()` callback let the targets define their own `class MacroFusion : public ScheduleDAGMutation { ... }` class which they then add as a mutator.


That would also mean to get rid of the `createMacroFusionDAGMutation(const TargetInstrInfo *TII)` function and thereby the EnableMacroFusion flag. This is fine IMO as on AArch64 you can just as well enable/disable it with the FeatureArithmeticBccFusion/FeatureArithmeticCbzFusion.


Repository:
  rL LLVM

https://reviews.llvm.org/D28489