[llvm] [AMDGPU] Implement Waitcnt Expansion for Profiling (PR #169345)

[Pierre van Houtryve via llvm-commits]

================
@@ -297,6 +302,30 @@ class WaitcntGenerator {
   // optimization.
   bool isOptNone() const { return OptNone; }
 
+  // Get the maximum wait count value for a given counter type
+  unsigned getWaitCountMax(InstCounterType T) const {
----------------
Pierre-vh wrote:

This does not take into account `hasExtendedWaitCounts`, check how `HardwareLimits` is used below

https://github.com/llvm/llvm-project/pull/169345