[llvm] [clang] [AMDGPU] Add code model (#70760) test for amdgpu target. (PR #71019)

[Matt Arsenault via cfe-commits]

================
@@ -13,6 +13,11 @@
 // RUN: not %clang -### -c --target=aarch64 -mcmodel=medium %s 2>&1 | FileCheck --check-prefix=ERR-MEDIUM %s
 // RUN: not %clang -### -c --target=aarch64 -mcmodel=kernel %s 2>&1 | FileCheck --check-prefix=ERR-KERNEL %s
 // RUN: not %clang --target=aarch64_32-linux -### -S -mcmodel=small %s 2>&1 | FileCheck --check-prefix=ERR-AARCH64_32 %s
+// RUN: %clang --target=amdgcn-amd-amdhsa -nogpulib -### -c -mcmodel=tiny %s 2>&1 | FileCheck --check-prefix=AMDGPU-MCMODEL-TINY %s
----------------
arsenm wrote:

There's 2 cases. The case where we're directly codegening to the target, in which case it behaves like a regular CPU target. That's what you're testing here, but it's somewhat unusual. We could error here as before.

The more realistic case is treating it as an offload target from OpenMP or HIP, so I think it's more important to test that case. We definitely don't want an error here.

https://github.com/llvm/llvm-project/pull/71019