[llvm] [AMDGPU] Handle non-register operands for S_SUB/ADD_U64_PSEUDO (PR #86104)

[Matt Arsenault via llvm-commits]

================
@@ -4857,9 +4857,7 @@ MachineBasicBlock *SITargetLowering::EmitInstrWithCustomInserter(
     bool IsAdd = (MI.getOpcode() == AMDGPU::S_ADD_U64_PSEUDO);
     if (Subtarget->hasScalarAddSub64()) {
       unsigned Opc = IsAdd ? AMDGPU::S_ADD_U64 : AMDGPU::S_SUB_U64;
-      BuildMI(*BB, MI, DL, TII->get(Opc), Dest.getReg())
-          .add(Src0)
-          .add(Src1);
----------------
arsenm wrote:

Ignore clang-format, this is worse 

https://github.com/llvm/llvm-project/pull/86104