Hello Hector and Neil,
I am nearly there to have a good emdebian system running.
emdebian is booting. According to Neil the emdebian has to be the root fs before executing the seondstage script.
So I bootet the board and chroot to emdebian and executed the script. BTW I got still the same error messages as before. But afterwards I set
the rootfs before booting to the emdebian tree and the initial scripts are started. The error is that /lib/modules/2.6.22.18-ELinOS-469/modules.dep could
not be found and afterwards process '/sbin/getty 38400 tty1' is restarted again and again.
VFS: Mounted root (nfs filesystem).
Freeing init memory: 72K
init started: BusyBox v1.11.1 (2008-08-08 18:56:00 ART)
starting pid 106, tty '': '/etc/init.d/rcS'
.udev/ already exists on the static /dev! (warning).
Starting the hotplug events dispatcher: udevd.
Synthesizing the initial hotplug events...done.
Waiting for /dev to be fully populated...done.
Setting hostname to 'emdebian-arm'...done.
Activating swap:.
FATAL: Could not load /lib/modules/2.6.22.18-ELinOS-469/modules.dep: No such file or directory
Can't open /dev/tty1: No such file or directory
Can't open /dev/tty2: No such file or directory
starting pid 364, tty '/dev/ttyS2': '/sbin/getty -L ttyS2 115200 vt100'
Can't open /dev/tty3: No such file or directory
process '/sbin/getty 38400 tty1' (pid 362) exited. Scheduling for restart.
process '/sbin/getty 38400 tty2' (pid 363) exited. Scheduling for restart.
process '/usr/bin/tail -f /var/log/messages' (pid 367) exited. Scheduling for restart.
Can't open /dev/tty1: No such file or directory
Can't open /dev/tty2: No such file or directory
Can't open /dev/tty3: No such file or directory
Debian GNU/Linux 5.0 emdebian-arm ttyS2
process '/sbin/getty 38400 tty1' (pid 368) exited. Scheduling for restart.
process '/sbin/getty 38400 tty2' (pid 369) exited. Scheduling for restart.
process '/usr/bin/tail -f /var/log/messages' (pid 370) exited. Scheduling for restart.
Can't open /dev/tty1: No such file or directory
Can't open /dev/tty2: No such file or directory
Can't open /dev/tty3: No such file or directory
Debian GNU/Linux 5.0 emdebian-arm ttyS2
process '/sbin/getty 38400 tty1' (pid 371) exited. Scheduling for restart.
.
.
I then copied the /lib/modules/2.6.22.18-ELinOS-469/modules.dep from the elinos fs and it looked better, but how is it intended to get that file?
After rebooting i get this:
VFS: Mounted root (nfs filesystem).
Freeing init memory: 72K
init started: BusyBox v1.11.1 (2008-08-08 18:56:00 ART)
starting pid 106, tty '': '/etc/init.d/rcS'
.udev/ already exists on the static /dev! (warning).
Starting the hotplug events dispatcher: udevd.
Synthesizing the initial hotplug events...done.
Waiting for /dev to be fully populated...done.
Setting hostname to 'emdebian-arm'...done.
Activating swap:.
Will now check all file systems.
fsck 1.41.3 (12-Oct-2008)
Checking all file systems.
Done checking file systems. A log is being saved in /var/log/fsck/checkfs if that location is writable..
Asking all remaining processes to terminate...done.
All processes ended within 1 seconds....done.
Will now mount local filesystems:.
The system is going down NOW!
Sending SIGTERM to all processes
Sending SIGKILL to all processes
Requesting system halt
System halted.
Rebooting results in the same behaviour. So how can I prevent the system to halt? The checkfs log contains:
Log of fsck -C -V -R -A -a
Thu Jan 1 00:00:40 1970
fsck 1.41.3 (12-Oct-2008)
Checking all file systems.
Thu Jan 1 00:00:40 1970
----------------
Marco
Hello,
>> What I did so far:
>> 1) build the emdebian root fs with emsandbox (architecture set to arm with emsetup)
>> 2) Set up a NFS server with the root file system of the original linux system running on the board. So the system is now booting from NFS instead of the small flash.
>> 3) Copied the emdebian-arm.tgz into a subfolder /emdebian
>
> I think that sounds wrong. The tarball is decompressed to *what will
> become the root directory of the machine* - but I've no idea about
> using NFS.
>
>> 4) On the device the .tgz archive was unpacked.
>> 5) Executing the secondstage script which fails.
>>
>> Apart from that I am wondering why the binaries (for example in the /bin dir of the emdebian rootfs) don't work when booted with the original linux system:
>
> Am I right in thinking that they work inside the chroot? Looks like it.
>
>> -/bin/sh-3.00# ls -l /emdebian/bin/busybox
>> -rwsr-xr-x 1 root root 736544 Aug 8 2008 /emdebian/bin/busybox
>> -/bin/sh-3.00# /emdebian/bin/busybox
>> -/bin/sh: /emdebian/bin/busybox: No such file or directory
>
> Why did the shell change?
>
>> The rootfs and the kernel where build with the arm cross compile tools from Elinos.
>
>> Here is the emsecondstage script log:
>>
>>
>> -/bin/sh-3.00# tar -xzpf emdebian-arm.tgz
>> nfs: server 192.168.99.237 not responding, still trying
>> nfs: server 192.168.99.237 not responding, still trying
>> nfs: server 192.168.99.237 OK
>> nfs: server 192.168.99.237 OK
>> INIT: Id "d9" respawning too fast: disabled for 5 minutes
All this looks a bit confusing. You need a kernel capable to run old
ARM ABI. You could unpack the tarball into a system and chroot to it,
then you should be in the emdebian system, just about to have a little
fight with apt and friends, but those are easily worked out.
If you try to replace a NFS rootfs with the one provided by emdebian,
you would experiment nfs errors, as the tarball is not properly setup
to support for NFS.
So, you need a NFS and old ARM ABI kernel, by soon there are plans to
have into emdebian the armel port, it is just a matter of time.
Currently, lots of testing and debugging is needed, but finding the
way arround, and proper configs you can have a nice working system.
>> Setting up debian-archive-keyring (2009.01.31em1) ...
>> gpg: symbol lookup error: /lib/libreadline.so.5: undefined symbol: PC
>> gpg: symbol lookup error: /lib/libreadline.so.5: undefined symbol: PC
>> gpg: symbol lookup error: /lib/libreadline.so.5: undefined symbol: PC
>> dpkg: error processing debian-archive-keyring (--configure):
>> subprocess post-installation script returned error exit status 127
You have undefined symbols, it most likely appears to me a problem
with your kernel, it is not exporting those symbols, probably because
you are mixing ABIs, if I am wrong please, correct me. There is
another mail in the list reporting this failure.
Regards
--
Héctor Orón