[clang] [AMDGPU] Use the AMDGPUToolChain when targeting C/C++ directly (PR #99687)

[Joseph Huber via cfe-commits]

================
@@ -6405,7 +6424,12 @@ const ToolChain &Driver::getToolChain(const ArgList &Args,
       TC = std::make_unique<toolchains::NVPTXToolChain>(*this, Target, Args);
       break;
     case llvm::Triple::AMDHSA:
-      TC = std::make_unique<toolchains::ROCMToolChain>(*this, Target, Args);
+      TC =
+          llvm::any_of(Inputs,
+                       [](auto &Input) { return types::isOpenCL(Input.first); })
+              ? std::make_unique<toolchains::ROCMToolChain>(*this, Target, Args)
----------------
jhuber6 wrote:

In the mixed case, all of them will be the `ROCMToolChain`, which I'm okay with. `ROCMToolChain` is what currently is invoked, the main difference is that it includes a lot of stuff in the ROCMToolChain while we want straight C/C++ to be freestanding.

https://github.com/llvm/llvm-project/pull/99687