Sudden Timeout Error of RTL-8139 Card

Donald Becker becker@cesdis1.gsfc.nasa.gov
Tue Sep 15 16:08:33 1998


On Tue, 15 Sep 1998, Janos Farkas wrote:

> Just FYI, since then, there was not a single occurence of the 'transmit
> timeout' problem in normal circumstances, on neither host I'm using the
> modified 1.02 driver (i.e. only the stressed ping floods caused transmit
> timeouts, but it still can do that; it probably is just too extreme?).

Thanks for the report.

> But, debug=2 still showed two more messages on the first day:
> 
> eth2: Transmit error, Tx status b10aa059.
> eth2: Transmit error, Tx status b10aa24a.

These are both out-of-window collisions.  My guess is that you are on a
repeater and some other host is trying to use full-duplex.
Fix: Change that other host's setting to the proper half-duplex.

> This status is different than the ones I saw earlier, and indeed, these
> were not followed by a transmit timeout message.

The driver should be able to restart the transmitter after encountering this
error.

There is only one expected Tx error: Tx FIFO underflow.  The driver starts
with the minimum threshold to minimize transmit latency, and increases the
threshold on each FIFO underrun.  If I were trying to sell a product I would
say "Intelligent hardware analyzes you machine and adapts to provide the
highest possible performance", but it's still just increasing the
threshold.. 

Donald Becker					  becker@cesdis.gsfc.nasa.gov
USRA-CESDIS, Center of Excellence in Space Data and Information Sciences.
Code 930.5, Goddard Space Flight Center,  Greenbelt, MD.  20771
301-286-0882	     http://cesdis.gsfc.nasa.gov/people/becker/whoiam.html

 | 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.