[clang] Thread Safety Analysis: Support attributes on function pointers (PR #191187)

[Aaron Ballman via cfe-commits]

https://github.com/AaronBallman approved this pull request.

Thank you for the discussion @melver, I think I'm convinced this is a defensible way forward. I agree that converting these into type attributes will be a significant design shift that's likely to be disruptive in practice. LGTM, but please give a bit of time for @aaronpuchert to weigh in as well because he may have other ideas/opinions.

https://github.com/llvm/llvm-project/pull/191187