[eepro100] eepro100 driver frequently dying in Linux 2.4.4
Kallol Biswas
kallol@bugula.fpk.hp.com
Mon, 23 Jul 2001 23:52:14 EDT
Donald,
They will eventually accept, they never denied any problem
I found, but they did not come forward with existing problems.
I never encountered command skipping problem, my programming model is
a little different, the TX and non TX command are not mixed in the
same ring.
Kallol
>
> On Mon, 23 Jul 2001, Kallol Biswas wrote:
>
> > "Thanks" me too:)
>
> Yup.
>
> > I am having hard time to workaround other two problems with 82559
> > card, I hope intel will come up with solutions again.
>
> Does Intel admit the problems?
>
> > Cards eats outgoing packets and does not generate interrupt
> > for incoming packets:
> ...
> > packet. However with kernel debugger I have seen that the card posts
> > transmit complete status and interrupt. No error counter is update.
> > That means the card really devours the frames. This problem also happens
> > only with high workload.
> ...
> > 2)
> > On a highly stressed machine repeated call to selftest may cause it
> > to fail.
> > ... This also happens with high network
> > load.
>
> The '559 seems to have a whole set of problems related to high PCI or
> network loads. Once they occur, they quick recur. My interpretation is
> that the chip seems to corrupt its firmware or state machines.
>
> Is there any Intel report on the long-standing problem of the command
> unit "skipping" -- not marking commands as having completed?
>
>
>
> Donald Becker becker@scyld.com
> Scyld Computing Corporation http://www.scyld.com
> 410 Severn Ave. Suite 210 Second Generation Beowulf Clusters
> Annapolis MD 21403 410-990-9993
>
>
> _______________________________________________
> eepro100 mailing list
> eepro100@scyld.com
> http://www.scyld.com/mailman/listinfo/eepro100
>
--
Phone: 973-443-7469 | Design Engineer
Telnet: 1-443-7469 | EIAL UNIX OS LAB
www.kallolbiswas.com | Hewlett Packard Company
kallol_biswas@hp.com | Florham Park, NJ