View previous topic :: View next topic |
Author |
Message |
andi_s Apprentice
Joined: 10 Feb 2005 Posts: 226 Location: Germany
|
Posted: Fri Nov 12, 2010 4:30 pm Post subject: arm filesystem boot problem (solved) |
|
|
hello,
i followed this
http://www.gentoo.org/doc/en/gentoo-x86-quickinstall.xml
to install gentoo on my armv5 pda - of course i adjusted the necessary steps for my device.
i used stage3-armv5tel-20101015.tar.bz2
i chrooted from an emdebian system that allready runs on this device, so i know my kernel is working (no modules - everything is build static). as i use haret to boot i have no /boot partition nor did i emerged a kernel (i use my custom kernel 2.6.29 that i use with debian system - this kernel is on vfat partition of the mmc) - as said - with the debian system i have no problems, so it must have something todo with the gentoo system.
my mmc (sd-card) is partitioned this way:
Code: |
mmcblk0p1 vfat (haret, zimage)
mmcblk0p2 ext2 (gentoo)
mmcblk0p3 swap
mmcblk0p4 ext2 (debian)
|
here is my gentoo-fstab
Code: |
/dev/mmcblk0p2 / ext2 noatime 0 1
/dev/mmcblk0p3 none swap sw 0 0
shm /dev/shm tmpfs nodev,nosuid,noexec 0 0
|
everything worked well so far - installing base system, chrooting, emerging syslog-ng etc.
but now if i try to boot into gentoo i get this error:
Code: |
mounting proc at /proc OK
mounting sysfs at /sys OK
mounting /dev ... OK
starting udevd ... OK
populating /dev ... OK
waiting for uevents ... OK
mounting devpts ... OK
remounting root filesystem read-only OK
checking root filesystem
ext2fs_check_if_mount: can't check if filesystem is mounted due to missing mtab file while determining wether /dev/mmcblk0p2 is mounted
fsck.ext2: no such file or directory while trying to open dev/mmcblk0p2
/dev/mmcblk0p2:
the superblock could not be read or does not describe a correct ext2 filesystem
some text... try e2fsck -b 8193 <device>
|
if i enter my root password for maintenance "fdisk -l" does not show me any devices, so i cannot run e2fsck from here
if i press ctrl-d my device simply reboots
i also checked mmcblk0p2 for errors (i run e2fsck from debian system) - but no errors found.
i hope anybody here can help - thanks!
Last edited by andi_s on Sat Nov 13, 2010 11:35 am; edited 1 time in total |
|
Back to top |
|
|
pianosaurus l33t
Joined: 19 Apr 2004 Posts: 944 Location: Bash$
|
Posted: Fri Nov 12, 2010 5:36 pm Post subject: |
|
|
Hm. The full installation documents has a cp /proc/mounts /etc/mtab step, which should stop boot from complaining. The boot scripts should keep the file updated after that.
Could you try that in maintenance mode and reboot? _________________ PKA Cuber
Please add [SOLVED] to the subject of your original post when you feel that your problem is resolved.
Adopt an unanswered post |
|
Back to top |
|
|
andi_s Apprentice
Joined: 10 Feb 2005 Posts: 226 Location: Germany
|
Posted: Fri Nov 12, 2010 6:31 pm Post subject: |
|
|
hi,
thanks for your reply!
in maintenance mode the filesystem was read-only, so i remounted it with rw and copied mtab to /etc
now the 'missing mtab' message is gone, however the rest of the error message is the same:
fsck.ext2: no such file... |
|
Back to top |
|
|
pianosaurus l33t
Joined: 19 Apr 2004 Posts: 944 Location: Bash$
|
Posted: Fri Nov 12, 2010 7:01 pm Post subject: |
|
|
Does that mean e2fsck is missing, or the device node? I assume the latter, but do an ls /dev just to check.
If the device isn't there, but it's there when you boot debian with the same kernel, I can only see two possibilities:
a) Gentoo's udev rules call the devices something else.
b) You need to probe some kernel modules first, for the mmc interface device. I imagine you'll get a list of candidates by running modprobe -ls | grep mmc. Make sure you copied the modules from the debian system too. They should be in /lib/modules/<kernel-version>/ _________________ PKA Cuber
Please add [SOLVED] to the subject of your original post when you feel that your problem is resolved.
Adopt an unanswered post |
|
Back to top |
|
|
andi_s Apprentice
Joined: 10 Feb 2005 Posts: 226 Location: Germany
|
Posted: Fri Nov 12, 2010 7:10 pm Post subject: |
|
|
e2fsck command is available, but "fdisk -l" does not show me any device
to b) : see above - no modules - everything is build static in the kernel (due to haret this is imho the best way)
btw: the kernel shows me during boot: mmc found p1 p2 p3 p4
to a) : i think this may be the case here, how can i check this? just with ls /dev ?
sorry, i must leave now, i will check this tomorrow |
|
Back to top |
|
|
pianosaurus l33t
Joined: 19 Apr 2004 Posts: 944 Location: Bash$
|
Posted: Fri Nov 12, 2010 7:30 pm Post subject: |
|
|
andi_s wrote: | to a) : i think this may be the case here, how can i check this? just with ls /dev ? |
Yes, that should do it. If udev recognises them as disks, you can also try ls -l /dev/disk/by-uuid/. That should list some unique identifiers, and show you what device they are really pointing to. _________________ PKA Cuber
Please add [SOLVED] to the subject of your original post when you feel that your problem is resolved.
Adopt an unanswered post |
|
Back to top |
|
|
chiefbag Guru
Joined: 01 Oct 2010 Posts: 542 Location: The Kingdom
|
Posted: Sat Nov 13, 2010 1:37 am Post subject: |
|
|
I have seen this issue on standard installs and it always relates to udev.
If you are seeing this error then the real root and the rest is correct.
I am not familiar with the ARM arch or if you are mounting a physical device or not.
However if you have the ability to mask udev please try the following.
vi /etc/portage/package.mask
>=sys-fs/udev-147 |
|
Back to top |
|
|
andi_s Apprentice
Joined: 10 Feb 2005 Posts: 226 Location: Germany
|
Posted: Sat Nov 13, 2010 11:34 am Post subject: |
|
|
it was indeed an udev-related problem!
thanks to both of you for your help! now i can boot into gentoo
however, during boot i get several messages:
Code: |
Unknown HZ value! (#) Assume 100
|
this seems to be a bug that is not fixed yet - but i think this is harmless...
but i recognized another error/warning line during boot:
Code: |
failed to execute /init. Attempting defaults...
INIT: version 2.88 booting
...
|
may this cause any problems? |
|
Back to top |
|
|
pianosaurus l33t
Joined: 19 Apr 2004 Posts: 944 Location: Bash$
|
Posted: Sat Nov 13, 2010 1:18 pm Post subject: |
|
|
I don't know why you are getting the Unknown HZ error, but it shouldn't be critical. It is written by certain tools, typically if they are too old for the kernel version.
The /init bug is simply because the init process is not located where the default is set up in the kernel. I don't know HaRET, but I assume you can set boot options somewhere? You should add init=/sbin/init to the kernel options.
If there is no such configuration, you can safely ignore it, as it is guessing correctly after the error message anyway. _________________ PKA Cuber
Please add [SOLVED] to the subject of your original post when you feel that your problem is resolved.
Adopt an unanswered post |
|
Back to top |
|
|
andi_s Apprentice
Joined: 10 Feb 2005 Posts: 226 Location: Germany
|
Posted: Sat Nov 13, 2010 3:58 pm Post subject: |
|
|
ok, thanks - yes haret has option to set init - i will try this |
|
Back to top |
|
|
|