[Net] Kernel renams eth0 -> eth7

[Net] Kernel renams eth0 -> eth7

[Nico Schottelius]

[-- Attachment #1: Type: text/plain, Size: 2458 bytes --]

Hello!

I am runnig 2.6.17.13 on a Geode, with 3 natsemi ports onboard
and 4 ports in a pci card (soekris net4801).

The strange thing is, that it currently renames eth0 to eth7
and eth1 to eth8 (or somehow different).

Have a look at [0] for dmesg output. You'll see this:

--------------------------------------------------------------------------------
[   64.643627] natsemi eth0: NatSemi DP8381[56] at 0xa0000000
(0000:00:06.0), 00:00:24:c5:69:5c, IRQ 10, port TP.
[   64.676833] natsemi eth1: NatSemi DP8381[56] at 0xa0001000
(0000:00:07.0), 00:00:24:c5:69:5d, IRQ 10, port TP.
[   64.710006] natsemi eth2: NatSemi DP8381[56] at 0xa0002000
(0000:00:08.0), 00:00:24:c5:69:5e, IRQ 10, port TP.
[   64.743402] natsemi eth3: NatSemi DP8381[56] at 0xa4000000
(0000:01:00.0), 00:00:24:c4:fd:14, IRQ 5, port TP.
[   64.776463] natsemi eth4: NatSemi DP8381[56] at 0xa4001000
(0000:01:01.0), 00:00:24:c4:fd:15, IRQ 11, port TP.
[   64.809891] natsemi eth5: NatSemi DP8381[56] at 0xa4002000
(0000:01:02.0), 00:00:24:c4:fd:16, IRQ 5, port TP.
[   64.842982] natsemi eth6: NatSemi DP8381[56] at 0xa4003000
(0000:01:03.0), 00:00:24:c4:fd:17, IRQ 11, port TP.
--------------------------------------------------------------------------------

which is correct. Later you see this:
--------------------------------------------------------------------------------
[  122.142761] eth8: DSPCFG accepted after 0 usec.
[  122.156411] eth8: link up.
[  122.164584] eth8: Setting full-duplex based on negotiated link
capability.
[  122.719529] eth8: remaining active for wake-on-lan
[  122.927613] eth8: DSPCFG accepted after 0 usec.
[  122.941269] eth8: link up.
--------------------------------------------------------------------------------
which feels really strange, because eth0 and eth1 simply
'disappeared'.

More information about the device can be found at [1]

My questions are:

   - why is it getting renamed and why did it not happen some boots
     (before I moved) before (like in [2])?
   
   - how can I tell the kernel NOT to rename it anymore / what
     must I fix?

Any hints appreciated.

Sincerly

Nico

[0]: http://unix.schottelius.org/zwerg/dmesg.renamed.eth0%2c1
[1]: http://unix.schottelius.org/zwerg/
[2]: http://unix.schottelius.org/zwerg/dmesg

-- 
``...if there's one thing about Linux users, they're do-ers, not whiners.''
(A quotation of Andy Patrizio I completely agree with)

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [Net] Kernel renams eth0 -> eth7

[Stephen Hemminger]

On Mon, 9 Oct 2006 22:18:06 +0200
Nico Schottelius <nico-kernel20061009@schottelius.org> wrote:

> Hello!
> 
> I am runnig 2.6.17.13 on a Geode, with 3 natsemi ports onboard
> and 4 ports in a pci card (soekris net4801).
> 
> The strange thing is, that it currently renames eth0 to eth7
> and eth1 to eth8 (or somehow different).
> 
> Have a look at [0] for dmesg output. You'll see this:
> 
> --------------------------------------------------------------------------------
> [   64.643627] natsemi eth0: NatSemi DP8381[56] at 0xa0000000
> (0000:00:06.0), 00:00:24:c5:69:5c, IRQ 10, port TP.
> [   64.676833] natsemi eth1: NatSemi DP8381[56] at 0xa0001000
> (0000:00:07.0), 00:00:24:c5:69:5d, IRQ 10, port TP.
> [   64.710006] natsemi eth2: NatSemi DP8381[56] at 0xa0002000
> (0000:00:08.0), 00:00:24:c5:69:5e, IRQ 10, port TP.
> [   64.743402] natsemi eth3: NatSemi DP8381[56] at 0xa4000000
> (0000:01:00.0), 00:00:24:c4:fd:14, IRQ 5, port TP.
> [   64.776463] natsemi eth4: NatSemi DP8381[56] at 0xa4001000
> (0000:01:01.0), 00:00:24:c4:fd:15, IRQ 11, port TP.
> [   64.809891] natsemi eth5: NatSemi DP8381[56] at 0xa4002000
> (0000:01:02.0), 00:00:24:c4:fd:16, IRQ 5, port TP.
> [   64.842982] natsemi eth6: NatSemi DP8381[56] at 0xa4003000
> (0000:01:03.0), 00:00:24:c4:fd:17, IRQ 11, port TP.
> --------------------------------------------------------------------------------
> 
> which is correct. Later you see this:
> --------------------------------------------------------------------------------
> [  122.142761] eth8: DSPCFG accepted after 0 usec.
> [  122.156411] eth8: link up.
> [  122.164584] eth8: Setting full-duplex based on negotiated link
> capability.
> [  122.719529] eth8: remaining active for wake-on-lan
> [  122.927613] eth8: DSPCFG accepted after 0 usec.
> [  122.941269] eth8: link up.
> --------------------------------------------------------------------------------
> which feels really strange, because eth0 and eth1 simply
> 'disappeared'.
> 
> More information about the device can be found at [1]
> 
> My questions are:
> 
>    - why is it getting renamed and why did it not happen some boots
>      (before I moved) before (like in [2])?
>    
>    - how can I tell the kernel NOT to rename it anymore / what
>      must I fix?
> 
> Any hints appreciated.

Look at the hotplug scripts, that is probably what is doing it
in user space.  Or you could add a hook to dump the caller to
dev_change_name() in net/core/dev.c. 

-- 
Stephen Hemminger <shemminger@osdl.org>