Hi :)

after using etherreal i could see that the write-access is cause the =
qube
want's to write a file named 'init.core' after mounting the nfs-volume =
and
running the init-process.

i think, it could be a kind of core-dump. Nothing needfull found via =
google
:(

maybee it could be a problem of cross-compiling ? i build a toolset for
cobalt at my i386-pc; but as i can see booting & switching to the 3com =
does
work at all.=20

any tips ?

regard,

axel



-----Urspr=FCngliche Nachricht-----
Von: port-cobalt-owner@NetBSD.org [mailto:port-cobalt-owner@NetBSD.org] =
Im
Auftrag von Bryan Olmstead
Gesendet: Mittwoch, 28. Januar 2004 14:50
An: port-cobalt@NetBSD.org
Betreff: Re: Qube2700, Restore-CD with custom Kernel & 2nd NIC



> couse of the tulip-problem at the qube2700 i try'd to build a modifyed
> restore-cd with support for a 3c905b/c-Nic. I have used the cvs-source =
to
> build the kernel.

this is a great idea!  i talked with a friend last week about the
possibility of doing something like this as the tulip card in the 2700
seems to be possessed by the devil.  i've just been to lazy to put
together the cross complier on my main box.

> but then the install-process failed; via tcpdump i can read messages
> between the qube an the nfs/boot-pc that the qube wants to get
> write-access to the NFS-Volume. I guess, that it could be couse of =
some
> modifications in the cvs-kernel (atabus <-> idebus).

first let me say that i am also quite new to netbsd and nfs but,
none-the-less, i have a few ideas as i would really like to see this =
work
out as it would be a great way to get these guys up and running until =
the
tulip driver can be fixed.

if it's a read-write problem on the nfs volume, you might make sure that
something along the lines of:

	/nfsroot 10.0.0.0/255.255.255.0(rw)

is in the /etc/exports file.  the (rw) option at the end should force
read-write, if i understand correctly.

how long have waited for the process?  i copied an original 2700 drive =
to
an nfsroot partition to boot from one time, just to make sure things =
were
working fine on the nfs side, and it took something like 20+ minutes for
it to boot.  it might just be that it's taking a long, long time.  :)

if you get this working and make the kernel availible to all, i'll have =
to
buy you a beer. <g>

thanks,

-bryan

----------------------------
Bryan Olmstead
reply-to: olmsteab@acm.org
http://www.tr1tium.com