latest feisty kernel does not boot

Thanks for your bug report. If you modify the kernel boot options in grub for the new kernel and remove 'splash' and 'quiet' is there a more informative error returned?

If is possible that you are being hit by a critical issue we found & resolved between yesterday and today.

A must for this error is you have multiple filesystems -- and, specifically, /usr/lib is on a different filesystem than /. If this is the case... then:

If you installed libx86-1 yesterday to this morning, this *may* be it (together with other updates that probably went along)

Check is there is a /usr/lib/libx86.so.1. If it is there, copy it over to /lib/libx86.so.1

You will probably be able to boot in single user mode, then log in as root on the console.

This *might* be it. For reference, this was bug # 83920.

If you are using grub, you can also stop usplash by ESCaping at the beginning of the grub process, then:

1. highlight the kernel you want to boot
2. hit the 'e' key (without quotes); this will edit the boot line
3. look for a line that has the keyword 'splash' on it
4. position yourself on this line, and hit the 'e' key again, to edit the line
5. remove the 'splash' keyword from the line
6. hit ENTER
7. hit the 'b' key to boot

This should allow you to look at the boot messages, and find (or, at least to have an idea) of what is going wrong.

Note: this is on powerpc so there is no grub, only yaboot - using 'nosplash' on the boot prompt doesn't make any difference. And I'm using only 1 filesystem, /dev/hda3 is / and /dev/hda4 is swap (hda1/hda2 are used by yaboot - I think).

Ah, found /etc/yaboot.conf via the rescue cd - its a kernel crash...:

Faulting instruction address: 0xc0084a0c
Oops: kernel access of bad area: sig: 11 [#1]

And a lot of other numbers....

Is this a kernel bug or a bad disk?

I might be able to make a picture of the dump if that is really needed.

Yes. It's really needed. Especially the traceback portion.

Cheers,
Kyle

Sorry for the bad quality, I only have a mobile laying around here. Tomorrow I will type
the backtrace of the screen.

Need to know exactly what kernel version you are using. First off, you should be able to boot the old kernel, and yaboot prompt type "old" (hitting TAB should show you the available options).

Run this command:

dpkg -l | grep linux-source

Send the output of that command.

Backtrace:

init-bottom
Done
Unable to handle kernel paging request for data at address 0x0000004
Faulting instructing address: 0xc0084a0c
Oops: kernel access of bad area: sig: 11 [#1]
....
....
Call Trace:
DFFC3E40 C0084AAB free_block+0x160/0x16c (unreliable)
DFFC3E60 C00845B8 cache_flusharray+0x5c/0x100
DFFC3E80 C00847D4 kmem_cache_free+0xb4/0xd0
DFFC3EA0 C0080280 destroy_inode+0x78/0x8c
DFFC3EC0 C009F4D4 iput+0x94/0xb4
DFFC3ED0 C0095280 do_inlinkat+0x130/0x194
DFFC3ED0 C0013510 ret_from_syscal+0x0/0x38
    --- Exception: c01 at 0xf810a18
        LR: 0x1000002cc
 Instruction dump:
 5499c .....

The hex numbers on the left are not correct; I had limited time this morning and after the crash I get only 180 seconds before the system reboots :P

On booting the old kernel: I can select 'old' in the yaboot prompt, but this kernel isn't configured... No vmlinuz.old nor an initramfs.old is there in the /boot directory.

I can only boot from a rescue CD and then chroot to the root partition. I should be able to retrieve
the boot.log that way.

The crashing kernel does not write any dmesg/boot logging, so I cannot give that in this bugreport :(

It is kernel version 2.6.20-6.11.

I've did a apt-get install --reinstall of the kernel and this helped. I can now boot the laptop... I'm still suspecting a bad block on the disk or something.

Can I run some tests?

Did a badblocks /dev/hda3 which didn't report any errors. Kernel crashes again...with the same backtrace as above.

The latest kernel upgrade, 2.6.20-8.x seems to fix to issue.

Hi Miek

Can you confirm that you have not experienced this bug since upgrading to 2.6.20-8? If so, we can close this bug.

Thanks

[On 14 Mar, @19:25, Phillip Lougher wrote in "[Bug 84059] Re: latest feisty ..."]
> Hi Miek
>
> Can you confirm that you have not experienced this bug since upgrading
> to 2.6.20-8? If so, we can close this bug.

I think this is a duplicate of: 75681

regards,