How is the Ethernet MAC address for usb0 defined?

[Dr. H. Nikolaus Schaller]

Am 07.08.2007 um 16:35 schrieb Mikko Rauhala:

> ti, 2007-08-07 kello 11:47 -0400, Dr. H. Nikolaus Schaller kirjoitti:
>> If it is so, I would suggest to derive the MAC address from the
>> device serial number - or if that is not available - from the
>> Bluetooth MAC (but not using the same).
>
> Incidentally, using the Bluetooth MAC directly wouldn't be such a  
> silly
> idea. Different interfaces on the same host can share the same MAC  
> just

Hm. I remember that the only unique identifier in the world is the 48  
bit Ethernet address...

> fine. For instance on (at least older) SPARC hardware the MAC address
> was actually the property of the computer; all network interfaces used
> the same one. 'course, you couldn't hook two interfaces with the same
> advertised MAC onto the same LAN segment, but that's not really a
> relevant limitation here.

Are you sure? The same computer could be used to connect through
USB and BT in parallel.

You might ask why someone would want to do that? E.g. debugging the
BT interface - which needs a parallel ssh session into the OpenMoko.
>
> So, if one were to use a static MAC on the usbnet interface, I'd  
> say go
> for the BT MAC; it's easy, and there's one associated with each Neo.
> Personally I don't have a problem with a randomized one, but this  
> would
> seem like a logical thing to do.

IMHO a better solution would be to rehash (MD5, CRC48 or something) the
Bluetooth MAC - so that probability of conflicts is approx. 2^-48