AMDGPU: Stop forcing RequiresCodeGenSCCOrder

[arsenm]

This hasn't been strictly necessary since c897c13.
Practically this makes little difference; we still enable IPRA
by default which implies this option. By removing this explicit
force, -enable-ipra=0 has the expected change in the pass pipeline
to remove the DummyCGSCC runs.

[arsenm]

• AMDGPU: Stop forcing RequiresCodeGenSCCOrder #169522 👈 (View in Graphite)
• AMDGPU: Remove DummyCGSCC use after buffer lowering passes #169519
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[llvmbot]

@llvm/pr-subscribers-backend-amdgpu

Author: Matt Arsenault (arsenm)

Changes

This hasn't been strictly necessary since c897c13.
Practically this makes little difference; we still enable IPRA
by default which implies this option. By removing this explicit
force, -enable-ipra=0 has the expected change in the pass pipeline
to remove the DummyCGSCC runs.

---

Full diff: https://github.com/llvm/llvm-project/pull/169522.diff

1 Files Affected:

• (modified) llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp (-4)

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
index 23ca4dcc009c8..e5a35abe6da6b 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
@@ -1226,10 +1226,6 @@ class GCNPassConfig final : public AMDGPUPassConfig {
 public:
   GCNPassConfig(TargetMachine &TM, PassManagerBase &PM)
       : AMDGPUPassConfig(TM, PM) {
-    // It is necessary to know the register usage of the entire call graph.  We
-    // allow calls without EnableAMDGPUFunctionCalls if they are marked
-    // noinline, so this is always required.
-    setRequiresCodeGenSCCOrder(true);
     substitutePass(&PostRASchedulerID, &PostMachineSchedulerID);
   }