Fix/enable inlined min/maxFD

[matthewhall2]

On Z, fmin/fmax/dmin/dmax intrinsics were changed to support the Java spec (+0.0 compares as strictly greater than -0.0, and NaNs are returned unchanged) on the OpenJ9 side, and to support the IEEE-754 spec in OMR (quiet bit is set for when NaNs are involved). Previously +0 and -0 compared as equal.

This should be investigated for other platforms that support the fmin/fmax/dmin/dmax opcodes (namely aarch64 and Power) as discussed in https://github.com/eclipse-openj9/openj9/pull/20185#discussion_r1797047516 so that recognized calls for Math.max and Math.min can be enabled (assuming it does not cause a performance degradation on these platforms).

@zl-wang @IBMJimmyk @r30shah