Re: nonexistant device, eth2 loaded? and another net related lock-up

Chris Black (
Fri, 31 Jul 1998 16:57:40 +0200

Brian Craft wrote:
[stuff about crashes I have no clue about deleted]
idunno what could be causing this, but I do have something to say on the
nonexistant eth2...

> Anyone else having problems like this?
[previous boot messages deleted]
> Jul 30 13:40:10 bcboy-linux kernel: eth1: NE2000 found at 0x240, using IRQ 3.
> Jul 30 13:40:10 bcboy-linux kernel: loading device 'eth2'...
> ^^^^^^^^^^^^
> there is no eth2 device! What's up with this? I don't see this
> in the other 2.0.34 machines I have running. There's no eth2
> defined in conf.modules or on the boot line.
I have seen stuff like "loading device ethn" (where n = number of nics,
ie, it will be the first eth device that isn't actually in your machine)
on a machine with multiple ethernet cards.
I don't know if it is probing for another card, but I would think that
after it doesn't find it, there isn't an issue. Check cat /proc/net/dev,
there isn't an entry for the 'ghost' eth device. I don't know exactly
how the ethernet device probing for multiple cards works though.
Are you loading these eth drivers via entries in conf.modules and using
modules or by stuff compiled into the kernel and passing arguments thru
I have done it both ways but don't have access to the boot logs for
doing it the non-module way.
I also don't have actual messages for the current configuration using
modules because the command dmesg just shows lots of other messages
about SYN stuff (I seem to have written some network software somewhat
incorrectly, bad me. It isn't the hardware anyways).
A suggestion/question: why aren't kernel messages dealing with modules
loaded at boot time thru conf.modules or whatever recorded in the syslog
dmesg file? They don't seem to be for me anyways (2.0.35 + Jumbo-6).


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at