Problems with Tulip on second PCI Bus

[Andrew McMillan]

Donald Becker wrote:

> On Mon, 15 Mar 1999, Andrew McMillan wrote:
>
> > I can't seem to get my Tulip card working under Linux, and I was
> > wondering if there is perhaps a bug somewhere in either the driver, or
> > the code to handle the second PCI bus.  The only error message I can
> > seem to get out of the tulip driver (and only with debug on) is:
> >
> >     eth0: Too much work during an interrupt, csr5=0xfc90a000
>
> This is reporting a fatal PCI bus error!
> Normally I would recommend running 'tulip-diag' to report the exact error
> type, but if it's locking up the machine...
>
> 0xfc90....  means "PCI master abort".
>
> This occurred while fetching the transmit descriptor, with the receive side
> still stopped.  That likely means that it occured while setting up the chip
> -- loading the station address using a special configuration command that
> looks like a transmit packet.
>
> The error is cleared, but probably immediately recurs.

Well that would certainly account for it's scrolling off screen very quickly..
:-)

Does this make you think that it is an error in the Linux kernel PCI code?


> > I have a DEC21140-based network card (Kingston 10/100) in the
> > docking station for my laptop.  We use these network cards all
> > around our office, including several other linux boxes. We use
> > the 'tulip' driver which works everywhere but on my laptop.
>
> This is likely an error in configuring the bus bridge for the docking
> station.  Do the PCI slot support bus-masters?

I'm not sure.  How do I tell?  The card works fine with Windows 9X.

Thanks,
                                    Andrew McMillan.
_____________________________________________________________________
            Andrew McMillan, e-mail: Andrew@cat-it.co.nz
Catalyst IT Ltd, PO Box 10-225, Level 21, 105 The Terrace, Wellington
Me: +64 (21) 635 694, Fax: +64 (4) 499 5596, Office: +64 (4) 499 2267