[clang] [llvm] Don't optimize out no-op atomics in kernel mode (PR #193562)

[Eli Friedman via llvm-commits]

efriedma-quic wrote:

> The question is why MakeBinaryAtomicValue still emits non-volatile atomicrmw even though intrinsics do have volatile arguments?

The intrinsics have a fixed signature; from the perspective of semantic analysis, the pointer is always technically volatile-qualified.

We could mess with it to "overload" the signature so there are volatile and non-volatile variants.  But I'm not sure that really helps.

https://github.com/llvm/llvm-project/pull/193562