Re: Debian images on a Risc PC
Philip Blundell <pb@labs.futuretv.com> writes:
> > I have tried using the image in a chroot environment and it does
> > not have this problem (although it does go wrong in other ways).
> > But if I get Linux to run a small C program on boot to chroot to
> > the image, init gives the same respawning errors.
>
> What kernel are you using? I wonder if you might not have NWFPE
> built in or something.
Yes, you're right. I'm using a 2.0 kernel at the moment. So, I've
changed my chroot program to insmod the fpemulator module before
starting init, and booting now gets as far as the login prompt.
(I had to comment out the ``kill -USR1 1'' line in
/etc/init.d/mountall.sh because it was causing the system to hang.)
I get:
Debian GNU/Linux potato (none) tty1
(none) login: root
System bootup in progress - please wait
[Disconnect bypassed -- root login allowed.]
Unable to change tty /dev/tty1: No such device
I get the same error if I chroot into the Debian image and run login.
It's puzzling because I've looked at the login code, and I can do the
chown/chmod that it performs manually without getting an error.
On the other hand, if I run the Debian login binary from the old
RedHat system, after symlinking login.defs to /etc, it does work.
Both /dev/tty1 files seem to be the same. (Probably isn't a fair test
though, since running the Debian login from RedHat wouldn't use the
Debian-built libc.)
(The `bootup in progress' message also goes away if I delete
/etc/nologin{,.boot} manually. Shouldn't files like this be in /var,
anyway?)
Any ideas what's happening now? I'll try a 2.2 kernel some time if
that's likely to still be the problem.
Thanks for the help so far!
--
Mark Seaborn
- mseaborn@bigfoot.com - http://members.xoom.com/mseaborn/ -
``Water boils at a lower temperature at high altitude, which partly
accounts for the nasty taste of coffee on the summit of Mauna Kea''
Reply to: