[clang] [Clang][AMDGPU] Add builtins for some buffer resource atomics (PR #149216)

[Matt Arsenault via cfe-commits]

================
@@ -0,0 +1,24 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// RUN: %clang_cc1 -triple amdgcn-unknown-unknown -target-feature +atomic-fmin-fmax-global-f32 -target-feature +atomic-fmin-fmax-global-f64 -emit-llvm -o - %s | FileCheck %s
----------------
arsenm wrote:

We have this terrible system where we explicitly track features that are relevant to clang that is separate from the set of subtarget features added to the backend processor definition. In short the subtarget feature name strings used by clang are only mostly correlated with the feature in the backend. You need to [add these to the appropriate subtargets in TargetParser](https://github.com/llvm/llvm-project/blob/e1706763515dcd50edf67a377386bc6f1e33a6d0/llvm/lib/TargetParser/TargetParser.cpp#L367)

https://github.com/llvm/llvm-project/pull/149216