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

[David Green via llvm-commits]

================
@@ -6389,6 +6395,235 @@ bool CodeGenPrepare::optimizeGatherScatterInst(Instruction *MemoryInst,
   return true;
 }
 
+// Rewrite the mul_with_overflow intrinsic by checking if both of the
+// operands' value range is within the legal type. If so, we can optimize the
+// multiplication algorithm. This code is supposed to be written during the step
+// of type legalization, but given that we need to reconstruct the IR which is
+// not doable there, we do it here.
+bool CodeGenPrepare::optimizeMulWithOverflow(Instruction *I, bool IsSigned,
+                                             ModifyDT &ModifiedDT) {
+  if (!TLI->shouldOptimizeMulOverflowIntrinsic())
----------------
davemgreen wrote:

I would pass the type into the shouldOptimizeMulOverflowIntrinsic, and have the target return whether it should expand for the given type. Some of the legality checks bellow could then be removed from here.

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