[llvm] [CGP]: Optimize mul.overflow. (PR #148343)

[David Green via llvm-commits]

================
@@ -3470,6 +3470,13 @@ class LLVM_ABI TargetLoweringBase {
     return MathUsed && (VT.isSimple() || !isOperationExpand(Opcode, VT));
   }
 
+  // Return true if the target wants to optimize the mul overflow intrinsic
+  // for the given \p VT.
+  virtual bool shouldOptimizeMulOverflowIntrinsic(LLVMContext &Context,
----------------
davemgreen wrote:

Maybe we need a little more descriptive of a name - something like shouldOptimizeMulOverflowIntrinsicWithHighHalf  maybe.

https://github.com/llvm/llvm-project/pull/148343