[wasm] simd scalar lowering F32x4Add and I32x4Add

test/cctest/cctest.gyp

Index: test/cctest/cctest.gyp
diff --git a/test/cctest/cctest.gyp b/test/cctest/cctest.gyp
index dc848ad94699d466ca61c4bb1d61732b43963990..59d75a0cc17b2ca5dea352c00d2e5922e84b9c7b 100644
--- a/test/cctest/cctest.gyp
+++ b/test/cctest/cctest.gyp
@@ -192,6 +192,7 @@
       'wasm/test-run-wasm-interpreter.cc',
       'wasm/test-run-wasm-js.cc',
       'wasm/test-run-wasm-module.cc',
+      'wasm/test-run-wasm-module-helper.h',
       'wasm/test-run-wasm-relocation.cc',
       'wasm/test-run-wasm-try-catch.cc',
       'wasm/test-signatures.h',
@@ -217,7 +218,8 @@
       'test-macro-assembler-x64.cc',
       'test-log-stack-tracer.cc',
       'test-run-wasm-relocation-x64.cc',
-      'wasm/test-run-wasm-simd.cc'
+      'wasm/test-run-wasm-simd.cc',
+      'wasm/test-run-wasm-simd-lowering.cc',

[gdeepti, 2016/08/30 17:16:06]

Is there a reason we run the tests only on x64? Ignore if you're using x64 as a
starting point I'm just wondering if you hit anything that was architecture
specific.

[aseemgarg, 2016/10/10 17:35:17]

I set it up to run on architectures where the other simd tests are running.

     ],
     'cctest_sources_arm': [  ### gcmole(arch:arm) ###
       'test-assembler-arm.cc',