This is my third attempt to post this - I think maybe the mail
list was rejecting it because my envelope from address was wrong?
Apologies for any multiple copies that eventually appear.
Ok, so what actually is supposed to be the state of the latest kernels
on a CATS board? Wasn't PhilB working on that some time ago?
In some spare moments, with lots of spare cpu time, We have
successfully cross compiled a toolchain and the kernel, made a boot
header with a pretty logo that even manages to accept command line
arguments and pass them to linux, and it boots ok from a UFS
partition, and successfully nfs mounts the root.
We initally tried the JimPick debian boot image in a tar file, and
that worked fine once we got past the point of everything non-trivial
dying due to lack of floating point. Configure verbose user aborts,
was the solution to finding the problem! However, the nfs root stuff
appears to horribly disagree with dselect or something, and we were
unable to get additional packages to install.
Also working reasonably fine, if one is prepared to create an initial
image by slight amounts of cross architecture fighting with rpm is the
redhat rawhide stuff. So far it doesn't quite allow login, but booting
single user works, and rpm happily installs extra packages. Various
things are missed during the bootup, eg where is fsck.nfs (presumably
a null operation) suppose to come from?
Anyway, what doesn't work is the IDE. :-( On a plain 2.2.1-rmk4 it
detected it ok, and even seemed to work, but was very unstable and
would crash seemingly if one pressed a key while it was running. On
the 2.2.2-rmk4-philb990301 we just tried, it doesn't even detect it,
seemingly whatever configuration we try.
So, is there some secret set of configure options to choose to make
this work? What are they? Or is debugging the ide something that we
would have to do ourselves if we really want it?
unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]
