[PATCH] D114652: [AMDGPU] Move call clobbered return address registers s[30:31] to callee saved range

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/SIInstrInfo.cpp:2081
+        BuildMI(MBB, MI, DL, get(AMDGPU::S_SETPC_B64_return))
+            .addReg(TRI->getReturnAddressReg(*MF), RegState::Undef);
+
----------------
If we're going to just hardcode this to s[30:31], probably should do away with getReturnAddressReg


================
Comment at: llvm/lib/Target/AMDGPU/SIInstrInfo.cpp:2084-2089
+    for (unsigned i = Desc.getNumOperands(), e = MI.getNumOperands(); i != e;
+         ++i) {
+      const MachineOperand &MO = MI.getOperand(i);
+      assert(MO.isReg() && MO.getReg());
+      MIB.add(MO);
+    }
----------------
There's a copyImplicitOperands helper for this


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D114652/new/

https://reviews.llvm.org/D114652