[PATCH] D25482: [PPC] Allow two post RA schedulers to be in the pipeline and select one depending on the Machine Function's subtarget

[Matthias Braun via llvm-commits]

MatzeB added a comment.

I think I should be more specific in my critique and I noticed that I missed part of the problem while reading this the first time:

- I'm against adding infarstructure/special cases at a basic level like MachineFunctionPass. It is already more complicated than it should be today.
- The current way of replacing the PostRAScheduler with the MachinePostRAScheduler is messy indeed. How about always scheduling both passes and letting enablePostRAScheduler() return an enum value indicating whether we want the machine or the list scheduler to run?


https://reviews.llvm.org/D25482