[clang] [AMDGPU] Add sema check for global_atomic_fadd_v2f16 builtin (PR #158145)

[Tim Gu via cfe-commits]

================
@@ -0,0 +1,46 @@
+// RUN: %clang_cc1 -triple amdgcn-amd-amdhsa -verify %s
+
+// Test semantic analysis for AMDGCN atomic fadd v2f16 builtins
+// These tests ensure proper type checking for the builtin arguments
+
+typedef _Float16 v2f16 __attribute__((ext_vector_type(2)));
+typedef float v2f32 __attribute__((ext_vector_type(2)));
+typedef _Float16 v4f16 __attribute__((ext_vector_type(4)));
+
----------------
tcgu-amd wrote:

I think the type was from hip though. How do I add a type from hip_fp16 to a clang test? Thanks! 

https://github.com/llvm/llvm-project/pull/158145