Add HLSL memory barrier intrinsics, fix dst, add lit & EvaluateAttributeSnapped

Test/hlsl.intrinsics.lit.frag

@@ -0,0 +1,4 @@
+void PixelShaderFunction(float n_dot_l, float n_dot_h, float m)
+{
+    float4 r0 = lit(n_dot_l, n_dot_h, m);
+}