AMDGPU: Add compile time hack for hasCFUser

[arsenm]

Summary

Assume the control flow intrinsic results are never casted, and early
exit based on the type.

[hliao]

look reasonable for me. any test cases?

[arsenm]

In D74140#1861985, @hliao wrote:

look reasonable for me. any test cases?

It's compile time. The broken case would probably successfully compile anyway, since we aren't really fully dependent on the DAG divergence

[hliao]

LGTM

[arsenm]

03a2d0045d2c11f4601bb2cbf5cdd5389ed2d72a

llvm/lib/Target/AMDGPU/SIISelLowering.cpp

SITargetLowering::getRegClassFor(MVT VT, bool isDivergent) const {
 
   return RC;
 }
 
 // FIXME: This is a workaround for DivergenceAnalysis not understanding always
 // uniform values (as produced by the mask results of control flow intrinsics)
 // used outside of divergent blocks. The phi users need to also be treated as
 // always uniform.
-static bool hasCFUser(const Value *V, SmallPtrSet<const Value *, 16> &Visited) {
+static bool hasCFUser(const Value *V, SmallPtrSet<const Value *, 16> &Visited,
+                      unsigned WaveSize) {
+  // FIXME: We asssume we never cast the mask results of a control flow
+  // intrinsic.
+  // Early exit if the type won't be consistent as a compile time hack.
+  IntegerType *IT = dyn_cast<IntegerType>(V->getType());
+  if (!IT || IT->getBitWidth() != WaveSize)
+    return false;
+
   if (!isa<Instruction>(V))
     return false;
   if (!Visited.insert(V).second)
     return false;
   bool Result = false;
   for (auto U : V->users()) {
     if (const IntrinsicInst *Intrinsic = dyn_cast<IntrinsicInst>(U)) {
       if (V == U->getOperand(1)) {
if (const IntrinsicInst *Intrinsic = dyn_cast<IntrinsicInst>(U)) {
           break;
         case Intrinsic::amdgcn_end_cf:
         case Intrinsic::amdgcn_loop:
           Result = true;
           break;
         }
       }
     } else {
-      Result = hasCFUser(U, Visited);
+      Result = hasCFUser(U, Visited, WaveSize);
     }
     if (Result)
       break;
   }
   return Result;
 }
 
 bool SITargetLowering::requiresUniformRegister(MachineFunction &MF,
if (isa<InlineAsm>(CI->getCalledValue())) {
             else if (SIRI->isSGPRClass(RC))
               return true;
           }
         }
       }
     }
   }
   SmallPtrSet<const Value *, 16> Visited;
-  return hasCFUser(V, Visited);
+  return hasCFUser(V, Visited, Subtarget->getWavefrontSize());
 }