[clang] [llvm] [aarch64] Add more MSVC intrinsics (PR #193221)

[Daniel Paoliello via llvm-commits]

dpaoliello wrote:

> It would expect intrisincs to be used along code that want to have some form of runtime detection to use different path depending of the underlying hardware. It should not matter much for Win11, which sets ARMv8.1/LSE as the baseline; but some code might want to optimize some atomics for RCPC.

Correct, this is how these intrinsics tend to be used: in some function that will guard their use based on runtime feature detection rather than enabling the target feature for the entire compilation. This also means that we have to be **very** careful with how the optimizer treats these intrinsics, they must NOT be lifted into other blocks.

https://github.com/llvm/llvm-project/pull/193221