[Dave asked me to repost this to the list -- not sure how useful this little lead is; haven't worked on this for more than half a decade.] I don't have a good platform to test this on today, but one way to detect a wireless bridge a couple of years ago was to send a SNAP packet (actually anything with an Ethernet length -- as opposed to ethertype -- in it) that is shorter than Ethernet minimum length (14 + 46), and then see whether the padding you put in on one end survives over to the other end. That of course only works with bridges that actually optimize this case (plus you can detect those that completely mess up this case -- there are quite a few of those). Hope that helps; would need to do experiments with real hardware to say anything more substantive. Grüße, Carsten PS.: In case you wonder what the work was that led to this observation: https://tools.ietf.org/html/draft-bormann-rohc-over-802-02