[RISCV] Remove getPostRAMutations

[wangpc-pp]

We are using PostMachineScheduler instead of PostRAScheduler
since #68696.

The hook getPostRAMutations is only used in PostRAScheduler so
it is actually dead code for RISC-V now.

[llvmbot]

@llvm/pr-subscribers-backend-risc-v

Author: Pengcheng Wang (wangpc-pp)

Changes

We are using PostMachineScheduler instead of PostRAScheduler
since #68696.

The hook getPostRAMutations is only used in PostRAScheduler so
it is actually dead code for RISC-V now.

---

Full diff: https://github.com/llvm/llvm-project/pull/117527.diff

2 Files Affected:

• (modified) llvm/lib/Target/RISCV/RISCVSubtarget.cpp (-5)
• (modified) llvm/lib/Target/RISCV/RISCVSubtarget.h (-3)

diff --git a/llvm/lib/Target/RISCV/RISCVSubtarget.cpp b/llvm/lib/Target/RISCV/RISCVSubtarget.cpp
index 03397e1e0d89ee..426d3682049046 100644
--- a/llvm/lib/Target/RISCV/RISCVSubtarget.cpp
+++ b/llvm/lib/Target/RISCV/RISCVSubtarget.cpp
@@ -186,11 +186,6 @@ bool RISCVSubtarget::useRVVForFixedLengthVectors() const {
 
 bool RISCVSubtarget::enableSubRegLiveness() const { return true; }
 
-void RISCVSubtarget::getPostRAMutations(
-    std::vector<std::unique_ptr<ScheduleDAGMutation>> &Mutations) const {
-  Mutations.push_back(createMacroFusionDAGMutation(getMacroFusions()));
-}
-
   /// Enable use of alias analysis during code generation (during MI
   /// scheduling, DAGCombine, etc.).
 bool RISCVSubtarget::useAA() const { return UseAA; }
diff --git a/llvm/lib/Target/RISCV/RISCVSubtarget.h b/llvm/lib/Target/RISCV/RISCVSubtarget.h
index f2c0a3d85c998a..043838e13b964d 100644
--- a/llvm/lib/Target/RISCV/RISCVSubtarget.h
+++ b/llvm/lib/Target/RISCV/RISCVSubtarget.h
@@ -301,9 +301,6 @@ class RISCVSubtarget : public RISCVGenSubtargetInfo {
 
   bool enableSubRegLiveness() const override;
 
-  void getPostRAMutations(std::vector<std::unique_ptr<ScheduleDAGMutation>>
-                              &Mutations) const override;
-
   bool useAA() const override;
 
   unsigned getCacheLineSize() const override {

[preames]

LGTM

[michaelmaitland]

LGTM.

On another note, have we looked at using the MacroFusionDAGMutation for PostMachineSched?

[topperc]

LGTM.

On another note, have we looked at using the MacroFusionDAGMutation for PostMachineSched?

It's handled in llvm::createGenericSchedPostRA already

[topperc]

LGTM