reset: Duplicated checks for Arm architecture, lacking explanation

[afaerber]

As visible in #5, both wh_reset.py and bh_reset.py have open-coded checks for Arm architecture.
This should probably be refactored into a common function called by both, so that they don't start deviating.

Also, both sites do not offer further explanatory comments as to why PCIe rescan would not work on Arm. Is this really an architectural issue or just related to specific PCIe host controller drivers or to some quirk on a specific vendor chip? If architectural, how would we know whether any other architecture may need to be added to that check? (for example, have RISC-V host systems like HiFive Unmatched been verified?)