[llvm-branch-commits] [llvm] [AMDGPU] Dynamic VGPR support for llvm.amdgcn.cs.chain (PR #130094)

[Matt Arsenault via llvm-branch-commits]

================
@@ -116,14 +117,56 @@ static void splitBlock(MachineBasicBlock &MBB, MachineInstr &MI,
   MDT->applyUpdates(DTUpdates);
 }
 
-void SILateBranchLowering::expandChainCall(MachineInstr &MI) {
+static void addRegOrCopyOp(MachineInstrBuilder &MIB, MachineOperand &Op) {
+  if (Op.isReg())
+    MIB.addReg(Op.getReg());
+  else
+    MIB->addOperand(Op);
+}
+
+void SILateBranchLowering::expandChainCall(MachineInstr &MI,
+                                           const GCNSubtarget &ST,
+                                           bool DynamicVGPR) {
   // This is a tail call that needs to be expanded into at least
   // 2 instructions, one for setting EXEC and one for the actual tail call.
-  constexpr unsigned ExecIdx = 3;
+  unsigned ExecIdx =
----------------
arsenm wrote:

```suggestion
  int ExecIdx =
```

Implicit cast 

https://github.com/llvm/llvm-project/pull/130094