[llvm] [clang] [AMDGPU] Emit a waitcnt instruction after each memory instruction (PR #68932)

[Jay Foad via cfe-commits]

================
@@ -1708,6 +1710,19 @@ bool SIInsertWaitcnts::insertWaitcntInBlock(MachineFunction &MF,
     }
 
     ++Iter;
+    if (ST->isPreciseMemoryEnabled() && Inst.mayLoadOrStore()) {
+      auto Builder =
+          BuildMI(Block, Iter, DebugLoc(), TII->get(AMDGPU::S_WAITCNT))
+              .addImm(0);
+      if (IsGFX10Plus) {
----------------
jayfoad wrote:

I guess this works but it seems a bit wasteful to insert S_WAITCNT after stores and S_WAITCNT_VSCNT after loads. Does anyone care?

Stepping back a bit, I think you can probably implement this by calling generateWaitcnt instead of building the instructions yourself.

https://github.com/llvm/llvm-project/pull/68932