[llvm] [AMDGPU][CodeGen][True16] Track waitcnt of vgpr32 instead of vgpr16 for 16bit reg in GFX11 (PR #157795)

[Joe Nash via llvm-commits]

================
@@ -2570,6 +2578,12 @@ def UseFakeTrue16Insts : True16PredicateClass<"Subtarget->hasTrue16BitInsts() &&
   // FIXME When we default to RealTrue16 instead of Fake, change the line as follows.
   // AssemblerPredicate<(all_of FeatureTrue16BitInsts, (not FeatureRealTrue16Insts))>;
 
+def Has16bitD16HWBug: Predicate<"Subtarget->has16bitD16HWBug()">,
+  AssemblerPredicate<(all_of FeatureTrue16BitInsts, FeatureRealTrue16Insts, Feature16bitD16HWBug)>;
+def NotHas16bitD16HWBug: Predicate<"Subtarget->useRealTrue16Insts() && "
----------------
Sisyph wrote:

```suggestion
def NotHas16bitD16HWBug: Predicate<"!Subtarget->has16bitD16HWBug()"
```
? Not sure what the point of that field is.

https://github.com/llvm/llvm-project/pull/157795