Re: [SLUG] Quick question - 3 NICs

From: R P Herrold (herrold@owlriver.com)
Date: Sat Nov 10 2001 - 12:29:48 EST


On Sat, 10 Nov 2001 Doumbeck1@aol.com wrote:

> I am not sure that is correct.

I spent a day of my life at a customer's site with a mix of 4
PCI NICs learning this lesson -- I could not believe it
either. It was true. And as I think back, it was true with
ISA Intel EtherExpress with 2.0 kernel as well; I had a
response from Don Becker, to the effect that the 'lower
numbered MAC' address NIC would assign first ... most of the
time.

> basically if something is changing, then your hardware is not reseting
> consistently. I can always count on my hardware resetting the same...

Even through warm boots? I really doubt that -- it is too
dependant on the given manufacturer's ROM code and hardware
RC detection (analog) circuitry -- it will even vary as
component tolerances vary.

You are fortunate, is about as far as I would go in 'always
count[ing]' on this ... I have some HP vectra's and Compaq
units which 'wander' -- the older Phoenix Bios' are markedly
easy to demonstrate this with, in the Bios' inability to be
'nailed' to a given detection order.

-- Russ



This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 17:31:40 EDT