[realtek-bug] problem: rtl8139 V1.17 and Linux kernel 2.2.19
Wed Jul 24 17:45:00 2002
Oh, that was it!! My co-worker had switched cables to try to solve a perceived comm problem, and we get the first message when we disconnect, and the second when we re-connect!! We just confirmed it by connecting and disconnecting the cable again... Great...
Thanks for your quick help!!
> -----Original Message-----
> From: Donald Becker [mailto:firstname.lastname@example.org]
> Sent: Wednesday, July 24, 2002 14:36
> To: Dan Miller
> Cc: email@example.com
> Subject: Re: [realtek-bug] problem: rtl8139 V1.17 and Linux kernel
> On Wed, 24 Jul 2002, Dan Miller wrote:
> > We're trying to use a Realtek Ethernet chip (built into the
> > motherboard) on a Linux 2.2.19 system.
> > We originally using rtl8139.c version 1.07, but after a few hours
> > under heavy loads (WebStone with 500 connections) it would start to
> > report errors of various sorts. Today, we tried upgrading
> to version
> > 1.17, from the scyld web site. We also obtained pci-scan.c and the
> > required headers from this same page. When we load this
> driver pair,
> > it loads successfully, but when we try to ping through the
> driver, in
> > reports:
> > eth0: Abnormal interrupt, status 00000020.
> > eth0: Abnormal interrupt, status 00002020.
> Hmmm, this could be either a link change indication (overloaded on the
> RxUnderrun indication), or a RxUnderrun. Which chip do you have, and
> what is the detection message? If it's a rtl8139c, the second status
> indicates that the cable has changed length -- perhaps due to the link
> Donald Becker firstname.lastname@example.org
> Scyld Computing Corporation http://www.scyld.com
> 410 Severn Ave. Suite 210 Second Generation
> Beowulf Clusters
> Annapolis MD 21403 410-990-9993