rtl8139 problems

[Ben Shakal]

I seem to be having the same problem as you two (and probably 
others), and I came up with a band-aid until the problem is 
resolved.  Note that I am using it on my personal machine, so 
100% uptime is not mandatory.  I've got a little network monitor 
applet in X that you can bind commands to the different mouse 
clicks.  So I have it set up to shut down the interface when I right 
click on the applet, and bring the interface back up when I left click 
on it.  (Using commands similar to what Lenart Gabor is using.)  
Also, so that I can (semi-)reliably telnet into the machine without 
permanently locking myself out, I have the same thing done every 
20 minutes by cron... And it doesn't seem to adversely affect 
transfers, as long as they don't time out before the interface is 
reset.  Anyway, I suppsoe everyone else with the same problem 
has automated it by now, but I just thought I'd share my temporary 
fix with anyone who cares... :)

Oh, and thank you James for the note about the Intel EEPro... I 
was just about to permanently fix my problem by going out and 
investing in a better card.... But I guess I won't be considering that 
one, either... :)

'Course, when I am booted into M$ Windoze, the card works fine... 
:(



James Stevens wrote:

> Lenart Gabor wrote:
> > 
> > Hi !
> > 
> > I'm using "rtl8139.c:v1.00F 8/12/98" version of rtl8139 driver. After a few
> > transferred megabytes kernel says :
> > Sep  2 20:08:03 world kernel: eth0: Transmit timeout, status 0d 0000 media 08.
> > Sep  2 20:08:03 world kernel: eth0: Tx queue start entry 4  dirty entry 0.
> > Sep  2 20:08:03 world kernel: eth0:  Tx descriptor 0 is 00002000. (queue head)
> > Sep  2 20:08:03 world kernel: eth0:  Tx descriptor 1 is 00002000.
> > Sep  2 20:08:03 world kernel: eth0:  Tx descriptor 2 is 00002000.
> > Sep  2 20:08:03 world kernel: eth0:  Tx descriptor 3 is 00002000.
> > Sep  2 20:08:03 world kernel: eth0: MII #32 registers are: 1000 782d 0000 0000 05e1 0000 0000 0000.
> > After this point I can't use the eth interface  ping does not work etc.
> > If I "ifconfig eth0 down" than rerun /etc/init.d/network script
> > (debian) it will work again .... but after a few megabytes everything can
> > be started again :(
> > 
> > Sorry for my Bad English and other misstakes (I'm typing on about
> > 120bps line at the moment !!)
> > 
> > - Gabor
> 
> I get nearly the same thing which I posted to this mailing list
> yesterday. I think it is becuase I am saturating 10Mhz ethernet and the
> driver can not cope correctly with the transmit buffer filling up.
> 
> Interstingly, I am seeing a VERY similar looking problem with the Intel
> EEPro 10/100 driver.
> 
> James


--
Ben Shakal
neb@uclink4.berkeley.edu
ben@sixg.com

"Linux: Because reboots are for upgrades."
 | To unsubscribe, send mail to Majordomo@cesdis.gsfc.nasa.gov, and within the
 |  body of the mail, include only the text:
 |   unsubscribe this-list-name youraddress@wherever.org
 | You will be unsubscribed as speedily as possible.