Re: Question about the 3COM 3C905

Bogdan Costescu <[email protected]> Tue, 30 Dec 2003 19:49:57 +0100 (CET)
Newsgroups gmane.linux.drivers.vortex.devel
Message-ID <Pine.LNX.4.44.0312301937560.24529-100000@kenzo.iwr.uni-heidelberg.de>
[[ CC-ed to the vortex list ]]

On Mon, 29 Dec 2003, Mark Hakbijl wrote:

> I have a question. I just got a 3COM 3C905 network card (3Com PCI 3c905 
> Boomerang 100baseTx, according to my Linux kernel). I use kernel 2.4.23 and 
> the 3c59x.o module included in this kernel version.

OK.

> If I boot my computer everything is alright: the network is working fine, 
> reaching speeds up to 11MB/s and there are no errors reported in 
> /var/log/kern.log.

Good.

> Now comes the problem. When I restart my network devices (in my case with 
> /etc/init.d/ethnet restart), the network doesn't work anymore (also not 
> ping-able, for example) and I keep getting kernel messages like:
> 
> Dec 29 12:27:25 henkieweb kernel: eth1: Transmit error, Tx status register 
> d0.

TX Underrun - the host was unable to supply the packet data fast enough to 
keep up with the network.
What kind of computer is this (CPU speed and mainboard) ?

> Dec 29 12:27:25 henkieweb kernel:   Flags; bus-master 1, dirty 1(1) current 
> 1(1)

This is a bit strange. "dirty" and "current" are usually a Tx-ring size 
apart, not equal.

> Dec 29 12:27:25 henkieweb kernel:   Transmit list 00000000 vs. ccb52240.
> Dec 29 12:27:25 henkieweb kernel:   0: @ccb52200  length 80000103 status 
> 80000103
> Dec 29 12:27:25 henkieweb kernel:   1: @ccb52240  length 00000000 status 
> 00000000
> Dec 29 12:27:25 henkieweb kernel:   2: @ccb52280  length 00000000 status 
> 00000000
> Dec 29 12:27:25 henkieweb kernel:   3: @ccb522c0  length 00000000 status 
> 00000000
> Dec 29 12:27:25 henkieweb kernel:   4: @ccb52300  length 00000000 status 
> 00000000
> Dec 29 12:27:25 henkieweb kernel:   5: @ccb52340  length 00000000 status 
> 00000000
> Dec 29 12:27:25 henkieweb kernel:   6: @ccb52380  length 00000000 status 
> 00000000
> Dec 29 12:27:25 henkieweb kernel:   7: @ccb523c0  length 00000000 status 
> 00000000
> Dec 29 12:27:25 henkieweb kernel:   8: @ccb52400  length 00000000 status 
> 00000000
> Dec 29 12:27:25 henkieweb kernel:   9: @ccb52440  length 00000000 status 
> 00000000
> Dec 29 12:27:25 henkieweb kernel:   10: @ccb52480  length 00000000 status 
> 00000000
> Dec 29 12:27:25 henkieweb kernel:   11: @ccb524c0  length 00000000 status 
> 00000000
> Dec 29 12:27:25 henkieweb kernel:   12: @ccb52500  length 00000000 status 
> 00000000
> Dec 29 12:27:25 henkieweb kernel:   13: @ccb52540  length 00000000 status 
> 00000000
> Dec 29 12:27:25 henkieweb kernel:   14: @ccb52580  length 00000000 status 
> 00000000
> Dec 29 12:27:25 henkieweb kernel:   15: @ccb525c0  length 00000000 status 
> 00000000

And the Tx ring really is empty...

> I searched on the internet for this problem and the only thing I could find 
> was that I could have something to do with duplex configuration or 
> something.

I don't think that your case has anything to do with duplex settings, 
especially as you say that the network works fine until you restart it. 
What happens if you repeatedly restart it ? Does it always log such 
messages ?

-- 
Bogdan Costescu

IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY
Telephone: +49 6221 54 8869, Telefax: +49 6221 54 8868
E-mail: [email protected]

_______________________________________________
vortex mailing list
[email protected]
http://www.scyld.com/mailman/listinfo/vortex