[clang] [clang][X86] Support __attribute__((model("small"/"large"))) (PR #124834)

[Yaxun Liu via cfe-commits]

================
@@ -1,64 +1,40 @@
-// RUN: %clang_cc1 -triple aarch64 -verify=expected,aarch64 -fsyntax-only %s
+// RUN: %clang_cc1 -triple aarch64 -verify=expected,unsupported -fsyntax-only %s
 // RUN: %clang_cc1 -triple loongarch64 -verify=expected,loongarch64 -fsyntax-only %s
-// RUN: %clang_cc1 -triple mips64 -verify=expected,mips64 -fsyntax-only %s
-// RUN: %clang_cc1 -triple powerpc64 -verify=expected,powerpc64 -fsyntax-only %s
-// RUN: %clang_cc1 -triple riscv64 -verify=expected,riscv64 -fsyntax-only %s
+// RUN: %clang_cc1 -triple mips64 -verify=expected,unsupported -fsyntax-only %s
+// RUN: %clang_cc1 -triple powerpc64 -verify=expected,unsupported -fsyntax-only %s
+// RUN: %clang_cc1 -triple riscv64 -verify=expected,unsupported -fsyntax-only %s
 // RUN: %clang_cc1 -triple x86_64 -verify=expected,x86_64 -fsyntax-only %s
+// RUN: %clang_cc1 -triple nvptx64-unknown-cuda -fcuda-is-device -x cuda -verify=expected,unsupported -fsyntax-only %s
----------------
yxsamliu wrote:

> The attribute _is_ technically is ignored, with a warning, but the warning is likely non-actionable as it would be issues for a perfectly correct host code which just happens to be seen by a GPU compilation.
> 
> I think the correct way to handle it would be to issue a deferred diagnostic which will fire only if we end up code-gen-ing something that uses the attribute.
> 
> @yxsamliu Sam, does that make sense?

My understanding is that deferred diagnostics only works in a function, but this diagnostic happens outside functions.

https://github.com/llvm/llvm-project/pull/124834