[llvm] Extend vector.reduce.add and splat transform to scalable vectors (PR #161101)

[Gábor Spaits via llvm-commits]

================
@@ -3785,13 +3785,19 @@ Instruction *InstCombinerImpl::visitCallInst(CallInst &CI) {
 
       // vector.reduce.add.vNiM(splat(%x)) -> mul(%x, N)
       if (Value *Splat = getSplatValue(Arg)) {
-        ElementCount VecToReduceCount =
-            cast<VectorType>(Arg->getType())->getElementCount();
+        VectorType *VecToReduceTy = cast<VectorType>(Arg->getType());
+        ElementCount VecToReduceCount = VecToReduceTy->getElementCount();
+        Value *RHS;
         if (VecToReduceCount.isFixed()) {
           unsigned VectorSize = VecToReduceCount.getFixedValue();
-          return BinaryOperator::CreateMul(
-              Splat, ConstantInt::get(Splat->getType(), VectorSize));
+          RHS = ConstantInt::get(Splat->getType(), VectorSize);
         }
+
+        RHS = Builder.CreateElementCount(Type::getInt64Ty(II->getContext()),
+                                         VecToReduceCount);
+        if (Splat->getType() != RHS->getType())
----------------
spaits wrote:

I will remove the check. I only did it because I see the same check a few lines above:
https://github.com/llvm/llvm-project/pull/161101/files#diff-bc7e40810dd612e1bea85874e9ee09c840e5d7ba1ddf2c164965aa9f366ead67R3777

Just there they do it with `Res` vector an `II` call. Is the check redundant there too? If so, may I remove that too?

https://github.com/llvm/llvm-project/pull/161101