Re: urgent help about NIC question: e1000_clean_tx_irq: Detected Tx Unit Hang
On Fri, Jul 06, 2007 at 09:12:43AM -0400, chloe K wrote:
> Hi all
>
> I am running debian etch as router. one optical fibre cardis is using
> PCI-X and other 4 copper ports card is using PCI 64 bits
>
> Four ports copper ethernet cards are testing fine. but when I plug
> the optical fibre card to our upstream, this card is up and I can ping
> to my upstream. but at the same time, one of the ports in 4 ports card
> is showing error message in the console.
>
> ls it the irq issue? How can I check it? thank you very much for your
> help
>
> one of the ports in my 4 ports is showing error message
>
> e1000: eth2: e1000_clean_tx_irq: Detected Tx Unit Hang
> Tx Queue <0>
> TDH <3a>
> TDT <25>
> next_to_use <25>
> next_to_clean <38>
> buffer_info [next_to_clean>
> time_stamp <ffff2062>
> next_to_watch <3c>
> juffies <ffff2761>
> next_to_watch_status <0>
> NETDEV WATCHDOG: eth2: transit timed out
> e1000: eth2: e1000_watchdog: NIC Link is up 1000 Mpbs Full Duplex
> e1000: eth2: e1000_watchdog: Link is Down
>
I think that you are saying that in both cases, both cards are installed
in the box. The problem is that one of the copper ports stops working
when the fiber port is connected to something. Correct?
Give us the output of
#dmsg | grep -i eth
#lspci | grep -i eth
And tell us which port is the fibre and which ports are on the 4-port
card. Also, does you MB have its own ports?
Doug.
Reply to: