[clang] [llvm] [clang][RISCV] Support the Swift calling convention (PR #213448)

[Aaron Ballman via cfe-commits]

AaronBallman wrote:

> @AaronBallman I don't think we've ever said the policy parts of that anywhere, but I'd be happy to write up a quick RFC for the forums if you'd like.

Thank you for the offer! I think that makes sense; we should also update the attribute documentation for this:

> ABI stability is the most interesting question here. From the Swift project perspective, the only Swift ABIs that are officially stable are the ones on Apple platforms. I would ask that the Clang honor and document that for these attributes. That is, people are welcome to use these attributes in their own code if they like, and of course that covers enabling them in order to bring up new Swift ports. However, the attributes should be considered ABI-unstable on any given platform (and so should not be used in ABI-stable C interfaces) until such a point as the Swift project declares a stable ABI for that platform.

But one thing that would be helpful for us is if we can find someone to be the maintainer for swift within the Clang community. I don't expect they'd have a lot of review work to do related to the role, but that gives contributors an idea of who to reach out for on PRs like this (or the 32-bit PPC swift calling convention PR that came up recently). Do you know of anyone with enough crossover between the projects who might be interested in stepping up for it?

https://github.com/llvm/llvm-project/pull/213448