> Just a quick note to tell that the 1.3.96 kernel (+ Linus hint about adding
> <asm/types.h> to skbuff.h, fixed in .97) boots and runs fine on my
> Olivetti. I never got interrupts working on 1.3.62, so this is progress.
> (with 1.3.62 I could send, but not receive, with NE2000 as well as 3c503)
> I don't have any docs on the m700-10 except for the inst manual, so I'm in
> a difficult situtation for trying to fix such things myself. :-(
Until 1.2.63 or so I still had the EISA interrupts disabled just because
this part of the kernel was *completly* untested on Magnum/M700. Probably
this was your problem.
Anyway, networking is still broken; the problem is a weired combination
between a bug in my checksum code and erroneous code generated by GCC.
> My 3c503 ethernet board works fine now, in PIO mode. There's still something
> wrong with the shared mem detection, I'll come back to that when I get time
> again. The nfsmounts works fine, things look good, and the console is really
> bad, as Ralf indicated.. :-)
Question: I had a M700 that shows a line of trash at the top of the screen
after some time - even when just idleing around. Did you see this problem,
too? What CPU version are you using? Milo prints this and you can do a
cat /proc/cpuinfo to get this information.
All that now is still missing is an updated version of libc which I've built
yesterday for big (untested) and little endian. The versions that I've
published can no longer be used with current kernel header files.