[llvm] [AMDGPU][NewInsertWaitcnts] Boilerplate for the new wait insertion pass (PR #226335)

[Pierre van Houtryve via llvm-commits]

================
@@ -0,0 +1,47 @@
+//===--- AMDGPUNewInsertWaitcnts.h ------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+
+#ifndef LLVM_LIB_TARGET_AMDGPU_AMDGPUNEWINSERTWAITCNTS_AMDGPUNEWINSERTWAITCNTS_H
+#define LLVM_LIB_TARGET_AMDGPU_AMDGPUNEWINSERTWAITCNTS_AMDGPUNEWINSERTWAITCNTS_H
+
+#include "llvm/CodeGen/MachineFunction.h"
+#include "llvm/CodeGen/MachineFunctionAnalysisManager.h"
+#include "llvm/IR/PassManager.h"
+
+namespace llvm {
+
+namespace AMDGPU {
+
+/// The main class of the wait insertion pass.
+class InsertWaitcnts {
+  /// The bulk of the wait insertion logic is in this function. Returns true if
----------------
Pierre-vh wrote:

I guess this is fine for now. We'll want to flesh this out more once we get there.

I'd vote for actually removing this and just using `run`, then add this back in when there is some actual functionality

https://github.com/llvm/llvm-project/pull/226335