[clang] [CIR] Lowering to LLVM for global pointers (PR #125619)

[Andy Kaylor via cfe-commits]

andykaylor wrote:

This patch includes an experimental attempt at a CIR attribute visitor along the lines that @erichkeane suggested in a previous PR review. Unfortunately, I wasn't able to use the new visitor in the place that he suggested it without an ugly shoehorn because different handlers there would need different return behavior. I thought it would be worth posting it for discussion of the one place I was able to use it to see if we think it's worth pursuing to implement with proper TableGen support and some way of handling the init value case.

https://github.com/llvm/llvm-project/pull/125619