[V8] Implement SIMD Boolean vector types to allow mask registers.

src/compiler/representation-change.cc

Index: src/compiler/representation-change.cc
diff --git a/src/compiler/representation-change.cc b/src/compiler/representation-change.cc
index 9b834d84179bdb92908bb2f2e2cb4b26e1c79b9e..b38ddf34680d8a08fb2196be646d33691ae9ee2c 100644
--- a/src/compiler/representation-change.cc
+++ b/src/compiler/representation-change.cc
@@ -169,9 +169,10 @@ Node* RepresentationChanger::GetRepresentationFor(
     case MachineRepresentation::kWord64:
       DCHECK(use_info.type_check() == TypeCheckKind::kNone);
       return GetWord64RepresentationFor(node, output_rep, output_type);
-    case MachineRepresentation::kSimd128:  // Fall through.
-      // TODO(bbudge) Handle conversions between tagged and untagged.
-      break;
+    case MachineRepresentation::kSimd128:
+    case MachineRepresentation::kSimd1x4:
+    case MachineRepresentation::kSimd1x8:
+    case MachineRepresentation::kSimd1x16:
     case MachineRepresentation::kNone:
       return node;
   }