>
> On Wed, Aug 15, 2007 at 10:48:19AM +0200, Ulf Lamping wrote:
> > In the meantime I've checked in the "two pass" approach into addr_resolv.c:
> >
> > First try to find a "perfect match" -> if found all is well
> > Second strip off both flags and try to find a match then -> if found
> > return that one
> > No match found -> Give up
>
> OK, as it is with opensource: he who does it get's to say how it's done
> in the end :-)
> Still, is there a specific reason for you to mask out both bits? Do you
> have examples where masking out the localally administered bit is
> useful?
This wasn't really backed up by knowledge or deep thought ;-)
I thought it just might be better to check in the stuff that I had, and "iron out" the remaining flag details.
> Otherwise, I'd like to change the mask to 0xfe
>
After a short thought now: As it's not unlikely that a locally administered card is not from the manufacturer OUI of that address, masking out the LA flag might even be misleading.
So feel free to change the mask to 0xfe, but please leave a comment explaining why the LA flag should not be masked out - so we won't get into this discussion later again and again ;-)))
Regards, ULFL
_______________________________________________________________________
Jetzt neu! Schützen Sie Ihren PC mit McAfee und WEB.DE. 3 Monate
kostenlos testen. http://www.pc-sicherheit.web.de/startseite/?mc=022220