[llvm] [MIPS] Implement llvm.fminimum and llvm.fmaximum with f32/f64 (PR #89907)

[Cinhi Young via llvm-commits]

Cyanoxygen wrote:

For `mipsel-linux-gnu`, it is neither soft float nor FP32, nor FP64, nor FPXX.
Updating code using `Subtarget.IsSingleFloat`.

https://github.com/llvm/llvm-project/pull/89907