Argh,
>This is a part of a patch i sent to the list about a week ago. It also
>solved the problem of non-working sk(4) with jumbo frames enabled in my
>case.
I obviously didn't notice your posting...could have saved me some
valuable time! :) Anyway, I had jumbo frames enabled ever since, but
this caused me no trouble. It was just the wrong buffer size...so I
think there are probably other chips (beside the 88E8001) affected,
perhaps the xxx1000 family. As noticed before, I also did not see any
of these "watchdog timeout" messages.
But what the heck...it's just important sk? works!
Enjoy
--j
P.S. During my research, I checked the chip's ID on the motherboard,
and it reads 88E8001. Looking at the dmesg output, the system
identifies it as 88E1000* ("*" is "S" for a single port interface),
which is wrong. The manufacturer's site lists the 88E1000* as a member
of their "ALASCA" family, but according to dmesg, it's part of their
YUKON clan (which is correct). I'm not sure if we should patch
".../mii_layer/miidevs.h" to get this right, since after all, it
really doesn't matter...
------------------
cd /pub
more beer