[llvm] [AMDGPU] Lazily emit waitcnts on function entry (PR #73122)

[Jay Foad via llvm-commits]

================
@@ -44,7 +46,6 @@ entry:
 define ptr addrspace(1) @tail_call_assert_align() {
 ; CHECK-LABEL: tail_call_assert_align:
 ; CHECK:       ; %bb.0: ; %entry
-; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
----------------
jayfoad wrote:

This function ends with a tail call, not a return. The ABI says that counters are undefined on function entry but must be 0 on return.

https://github.com/llvm/llvm-project/pull/73122