[turbofan] Also optimize unsigned division by constant.

src/compiler/arm64/instruction-selector-arm64.cc

Index: src/compiler/arm64/instruction-selector-arm64.cc
diff --git a/src/compiler/arm64/instruction-selector-arm64.cc b/src/compiler/arm64/instruction-selector-arm64.cc
index 39cd6ca77e0cc31b1a50c4d4a4ed53cf0ebd672c..bf768c91bde37c1b94033126874007c9b98b96d6 100644
--- a/src/compiler/arm64/instruction-selector-arm64.cc
+++ b/src/compiler/arm64/instruction-selector-arm64.cc
@@ -778,6 +778,16 @@ void InstructionSelector::VisitInt32MulHigh(Node* node) {
 }
 
 
+void InstructionSelector::VisitUint32MulHigh(Node* node) {
+  // TODO(arm64): Can we do better here?
+  Arm64OperandGenerator g(this);
+  InstructionOperand* const smull_operand = g.TempRegister();
+  Emit(kArm64Umull, smull_operand, g.UseRegister(node->InputAt(0)),
+       g.UseRegister(node->InputAt(1)));
+  Emit(kArm64Lsr, g.DefineAsRegister(node), smull_operand, g.TempImmediate(32));
+}
+
+
 void InstructionSelector::VisitInt32Div(Node* node) {
   VisitRRR(this, kArm64Idiv32, node);
 }