[llvm] [AMDGPU] Allocate i1 argument to SGPRs (PR #72461)

[Matt Arsenault via llvm-commits]

================
@@ -73,7 +79,7 @@ struct AMDGPUOutgoingValueHandler : public CallLowering::OutgoingValueHandler {
     if (TRI->isSGPRReg(MRI, PhysReg)) {
       LLT Ty = MRI.getType(ExtReg);
       LLT S32 = LLT::scalar(32);
-      if (Ty != S32) {
+      if (Ty != S32 && Ty != LLT::scalar(64)) {
----------------
arsenm wrote:

Shouldn't need to special case the types here 

https://github.com/llvm/llvm-project/pull/72461