[wasm-simd] Scalar lowering for load extends on BE machines

LoadMem already does the proper sign extension based on the
passed memtype, so all is left is splatting it to the desired
number of lanes.

Change-Id: I54317b6e2b4d3daf2e980315198577f61c3f556a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2068031Reviewed-by: Zhi An Ng <zhin@chromium.org>
Reviewed-by: Deepti Gandluri <gdeepti@chromium.org>
Commit-Queue: Deepti Gandluri <gdeepti@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66436}

src/compiler/wasm-compiler.cc

@@ -3754,14 +3754,20 @@ Node* WasmGraphBuilder::LoadTransformBigEndian(
       result = graph()->NewNode(mcgraph()->machine()->I8x16Splat(), value);
       break;
     }
+    case LoadTransformation::kI16x8Load8x8S:
+    case LoadTransformation::kI16x8Load8x8U:
     case LoadTransformation::kS16x8LoadSplat: {
       result = graph()->NewNode(mcgraph()->machine()->I16x8Splat(), value);
       break;
     }
+    case LoadTransformation::kI32x4Load16x4S:
+    case LoadTransformation::kI32x4Load16x4U:
     case LoadTransformation::kS32x4LoadSplat: {
       result = graph()->NewNode(mcgraph()->machine()->I32x4Splat(), value);
       break;
     }
+    case LoadTransformation::kI64x2Load32x2S:
+    case LoadTransformation::kI64x2Load32x2U:
     case LoadTransformation::kS64x2LoadSplat: {
       result = graph()->NewNode(mcgraph()->machine()->I64x2Splat(), value);
       break;