[eepro100] Help please

Donald Becker becker@scyld.com
Fri Mar 14 11:41:05 2003


On 11 Mar 2003, Guillermo Gomez wrote:

> Basic Symptom: network is down (eth0). eth1 is up and running ok.
> Question: is my lan adapter gone ? See the logs i've got.
> System: Red Hat Linux 8.0
> The system is not local (remote access so far untill i travel there).
> What can i do ?
> 
> # more /var/log/messages
> ... [relevant info only]
> Mar 11 09:23:39 sunix kernel: eepro100.c:v1.09j-t 9/29/99 Donald Becker
> http://www.scyld.com/network/eepro100.html
> Mar 11 09:23:39 sunix kernel: eepro100.c: $Revision: 1.36 $ 2000/11/17
> Modified by Andrey V. Savochkin <saw@saw.sw.com.sg> and others
> Mar 11 09:23:39 sunix kernel: PCI: Enabling device 01:00.0 (0000 ->
> 0003)

Hmmm, very curious... not activated.

My driver release always bring the adater out of ACPI power saving
state.  I suspect that the driver you are using does not do this.
You can check if this is the problem by running 'lspci -v' and verifying
that the eepro100 is in D0 state, not D1, D2 or D3.

> Mar 11 09:23:39 sunix kernel: eth0: Invalid EEPROM checksum 0xff00,
> check settings before activating this device!
> Mar 11 09:23:39 sunix kernel: eth0: OEM i82557/i82558 10/100 Ethernet,
> FF:FF:FF:FF:FF:FF, IRQ 9.
...
> Mar 11 09:23:40 sunix kernel:  Failure to initialize the i82557.

Yup, not power up, not plugged in, or just broken.

> 01:00.0 Ethernet controller: Intel Corp. 82557/8/9 [Ethernet Pro 100]
> (rev 08)
...
> 		Status: D0 PME-Enable- DSel=0 DScale=2 PME-

Ohhh, OK, it must be broken.  This should be impossible, but might
happen with power glitches or internal firmware errors.

With a local machine, a full power cycle (unplug or hard switch, not
soft-power-off) will likely restore operation.
With a remote machine, setting the card to D3 state and then back to D0
state might reset operation.  I use pci-config to do this, but it does
require some expert knowledge to run.

> IRQ seems not to be assigned in the scanpci listing (INT_LINE 0x00) ?

That might be a problem, but the device is too broken to tell.


-- 
Donald Becker				becker@scyld.com
Scyld Computing Corporation		http://www.scyld.com
410 Severn Ave. Suite 210		Scyld Beowulf cluster system
Annapolis MD 21403			410-990-9993