[llvm] [LV] Prefer FLT_MIN/MAX for fmin/fmax reductions with ninf (PR #107141)

[Nikita Popov via llvm-commits]

https://github.com/nikic approved this pull request.

LGTM. This looks like a bug fix to me, using inf in an ninf instruction would result in poison.

https://github.com/llvm/llvm-project/pull/107141