[eepro100] sleep mode persistence

Kallol Biswas Kallol.Biswas@efi.com
Tue Oct 1 02:52:01 2002

In some cases the code modification was done
just to workaround  problems instead of fixing them.
For example in the routine speedo_found1() the eeprom
words are read using I/O space, which is not really required.

We were getting all eeprom words as  Fs from an 82559 card that was behind
a PCI-PCI bridge,  the problem went away with  Donald's version of the driver , i.e.
accessing  the eeprom in memory space fixed the problem. The bridge's settings were
fine. We are yet to find out why IO access returns all Fs.

We had an issue though, his driver could not be compiled statically in 2.4.17 kernel.

BTW, does any  one know about the fix for the 82559's bug that on very very
high traffic the card consumes all incoming and out going packets? However
it generates Tx complete interrupts.


Donald Becker wrote:

> On Mon, 30 Sep 2002, Frank Lenaerts wrote:
> > > > eepro100: wait_for_cmd_done timeout!
> > >
> > > The chip isn't repsonding.
> >
> > Would it be responding with the newer driver? Actually, the chip is
> > responding after such a reboot, so I wonder what|if something has
> > changed the driver initialisation or so (between v1.09 and v1.24 of
> > eepro100.c).
> The driver in the kernel split from my release.  Part of the reason for
> the split was that I insisted on a correctly designed API that
>   - didn't need to change with every kernel release,
>   - was capable of providing backwards and forward compatibility
>   - supported hot-swap (CardBus and PCI) and power management
>   - handled card activation consistently
> Instead the device interface was repeated changed to incrementally add
> the features, without a overall design.  The result is that the drivers
> shipped with the later 2.4 kernel are not usable even with earlier 2.4
> kernel versions, and would need major changes to use with earlier kernels.
> In this case I suspect a problem a problem with card activation that may
> be solved by using my driver set and driver interface.
> > # dmesg
> > eepro100.c:v1.09j-t 9/29/99 Donald Becker http://www.scyld.com/network/eepro100.html
> > eepro100.c: $Revision: 1.36 $ 2000/11/17 Modified by Andrey V. Savochkin <saw@saw.sw.com.sg> and others
> And what happened to the people that modified the driver?  They aren't
> around providing fixes and updates...
> --
> 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