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

[Ralf Jung via cfe-commits]

RalfJung wrote:

> possibly the frontend should be marking all atomicrmw ops as volatile if -fms-kernel is used?

If the intent of the kernel is "this memory access must be lowered to an actual asm access" then yeah it should be marked volatile.

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