Network problems with VIA chipset and Athlon XP
becker at scyld.com
Wed Oct 23 05:47:02 PDT 2002
On Wed, 23 Oct 2002, Claudio Martins wrote:
> Subject: Network problems with VIA chipset and Athlon XP
> We have a set of systems that become network unreachable, presenting the
> following messages in dmesg output, repeated at aprox. 5 sec. intervals:
What driver and version?
> NETDEV WATCHDOG: eth0: transmit timed out
> eth0: Transmit timed out, status fc664010, CSR12 00000000, resetting...
Hmmm, this looks like a modified tulip driver.
> eth0: Transmit timed out, status fc684010, CSR12 00000000, resetting...
> eth0: Out-of-sync dirty pointer, 34089 vs. 34106.
With a very serious problem.
> switch. The nodes are all identical and use an Asus A7V266-EX motherboard
> (VIA KT266), Athlon XP 1800+ CPU, 1.5GB of PC2100 DDR RAM, a 40GB Seagate IDE
> disk, Accton EN-1216 10/100 NIC (Tulip) and ATI Rage XL AGP graphics card.
Yup, you are using a Tulip card with the modified driver distributed
with the kernel.
The modified Tulip driver in the kernel is broken with certain cards,
and has been for a long time. Even with the example of my working
driver, the people modifying it prefer to keep patching their patches.
Someday they will might it right by chance, but that approach takes much
longer than understanding, design and testing.
[[ Actually, the situation is worse than this: there are _two_ modified
tulip drivers in the kernel, tulip and old_tulip. Both were modified by
the same person, and one abandoned. ]]
> I googled for a while and found some mailing list messages of people
> reporting errors that resembled this one, and some replies by Donald Becker.
And in almost every reply I advise "Use the unmodified driver set from
Scyld". People get tired of hearing that, but it really does fix many
tar xfvz netdrivers.tgz
Donald Becker becker at scyld.com
Scyld Computing Corporation http://www.scyld.com
410 Severn Ave. Suite 210 Scyld Beowulf cluster system
Annapolis MD 21403 410-990-9993
More information about the Beowulf