Subject: Re: The impossible has happened.
To: David Gilbert <dgilbert@jaywon.pci.on.ca>
From: Jason Thorpe <thorpej@nas.nasa.gov>
List: port-sparc
Date: 03/12/1996 18:17:46
On Tue, 12 Mar 1996 18:19:10 -0500 (EST) 
 David Gilbert <dgilbert@jaywon.pci.on.ca> wrote:

 > 	While testing ethernet things, I discovered that the
 > impossible has happened.  I have two ethernet cards with the same
 > ethernet (raw) address.  Actually, it's highly unlikely that this is
 > true, but I get the following from my messages file (and on boot):

This isn't impossible!  On a Sun, this is a *feature*!  :-)

 > Mar 11 23:49:05 repeat /netbsd: mainbus0 (root)
 > Mar 11 23:49:05 repeat /netbsd: cpu0 at mainbus0: SUN-4/200 series (L64802 or ACT8847 FPU)
 > Mar 11 23:49:05 repeat /netbsd: cpu0: 131072 byte write-back, 16 bytes/line, sw flush cache enabled
 > Mar 11 23:49:05 repeat /netbsd: obio0 at mainbus0
 > Mar 11 23:49:06 repeat /netbsd: ie0 at obio0 addr 0xf6000000 pri 6 address 08:00:20:06:ac:a6, type onboard
 > Mar 11 23:49:06 repeat /netbsd: vmel0 at mainbus0
 > Mar 11 23:49:06 repeat /netbsd: vmes0 at mainbus0
 > Mar 11 23:49:06 repeat /netbsd: ie1 at vmes0 addr 0xffe88000 vec 0x75 pri 5 address 08:00:20:06:ac:a6, type multibus/vme
 > 
 > 	Now... I don't know if this is a causing a problem (my
 > previously posted problems are reproducable without the two cards
 > installed), but this certainly says something interesting about our
 > driver.

Not at all.  This is expected on a Sun system.  Note that on a Sun, the 
MAC address isn't stored on a PROM on the card, but rather on a PROM on 
the *motherboard*.  (This comes from the idprom.)

Sun's approach was more like what the original Xerox way was ... A host 
identifier, then an interface identifier.  Unfortunately, the ethernet 
standard didn't really provide for the former, but Sun didn't change :-)

You can push a new MAC address into one of your interfaces using 
ifconfig(8).  Just pick one that's unique to your particular physical 
subnet :-)

For the record, this *really* breaks a dually-homed Sun that has both 
interfaces on the same subnet.  Is this what's happening in your case?

--------------------------------------------------------------------------
Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                               Home: 408.866.1912
NAS: M/S 258-6                                          Work: 415.604.0935
Moffett Field, CA 94035                                Pager: 415.428.6939