[llvm] Enable logf128 constant folding for AArch64 (PR #96287)

[Matthew Devereau via llvm-commits]

MDevereau wrote:

> I'd still prefer to use some sort of detection. If you can't figure out a way to do it in the header itself, please use CMake tests.

I've added detection of long double size. Confusingly some x86 targets defined long double as 16 bytes but are actually fp80 types, so I've checked the size of the mantissa too. I've also removed the CMake opt-in and detect logf128 by default now.

https://github.com/llvm/llvm-project/pull/96287