[libc] [llvm] [libc] Make ldexpf128/scalbnf128/scalblnf128 use emulated Float128 (PR #220295)

[Parth Patel via llvm-commits]

================

----------------
rigalis wrote:

I used a small wrapper for the scalblnf128 MPFR test to convert the int arg to long, as LdExpFunc in LdExpTest.h doesn't have a long-exponent variant (couldn't find precedent for that anywhere). Is this alright or would smoke-only be enough here?

https://github.com/llvm/llvm-project/pull/220295