[clang] [clang] Pass fp128 indirectly and return in xmm0 on Windows (PR #115052)

[Trevor Gross via cfe-commits]

tgross35 wrote:

Thanks for answering that question, I wouldn't have had a good answer outside of consistency. Does mid-level optimizations refer to optimizations done in Clang rather than in LLVM?

Somebody will need to land this for me, the two commits should come separate (first is NFC).

> Otherwise, I think this is good to go. I suppose Rust folks are watching this PR, otherwise, I'd say loop them in.

For future reference that's mostly me for f16/f128, I'll update our frontend after this lands.

https://github.com/llvm/llvm-project/pull/115052