[SOLVED] Kernel Panic when reboot after change in elilo.conf to hibernate

tdmsoares · 07-04-2020, 10:46 AM

Everything was fine untill I wanted to hibernate my laptop.
Following the instructions from https://docs.slackware.com/howtos:sl...in:hibernation, since I have elilo, I added the following line in elilo.conf:

Code:

append="resume=/dev/sda5"

Where sda5 is my swap partition

then I typed

Code:

# pm-hibernate

and everything worked fine, it loaded where I was working.

The problem was when I did

Code:

# reboot

Now my laptop is in Kernel Panic.
I forced a reset but the problem continues

What I can do now?

Notes: I cannot access my laptop files right now since it is in kernel panic, but I remember in elilo.conf a line before I added (to load at swap partiton) like this:

Code:

append="root=/dev/sda4"

hazel · 07-04-2020, 11:04 AM

You don't provide the actual panic message. Is it something like "can't find root partition"? That seems the likeliest cause of a panic after your modifications.

You need to boot from another device such as a dvd or memory stick and check your elilo.conf file. If you have accidentally deleted or damaged the line that specifies the root partition, that could be the cause of your problem.

Also, are you using the generic kernel with an initrd or the huge kernel booting directly to the hard drive?

tdmsoares · 07-04-2020, 11:17 AM

Unfortunately I can't post the image of the messages, but I will try to write a few of it:

Code:

Call Trace:

<IRQ> [ffffffff81627b33>] dump_stack+0x57/0x74
...
warn_slowpath_common ...
warn_slowpath_null +0x1a/0x20
native_smp_send_reschedule...
trigger_load_balance
scheduler_tick
? tick_sched_handle.isra.14 ...
update_process_times
tick_sched_handle.isra.14...
tick_sched_timer+0x3d...
__hrtimer_run_queues+0xf1...
hrtimer_interrupt+0x...
local_apic_timer_interrupt+0x...
smp_apic_timer_interrupt...
apic_timer_interrupt
<EOI> [fff...] ? panic+0x1c3/0x202
? panic+0x...
? printk+0x4b...
mount_block_root+0x...
mount_root+0xad...
prepare_namespace+0x16a
kernel_init_freeable+0x1bf...
? rest_init +0x80...
kernel_init+0xe...
ret_from_fork+0x55
? rest_init+0x80
----[end trace a26d...]

I have a usb stick with puppy ubuntu linux, I can mount the root partition /dev/sda4, but /boot/efi is empty as well as /proc folders...

tdmsoares · 07-04-2020, 11:19 AM

Quote:

Originally Posted by hazel

Also, are you using the generic kernel with an initrd or the huge kernel booting directly to the hard drive?

AS far as I know I am using the hard drive, I think with huge kernel

When I upgraded the kernel I just use it eliloconfig

hazel · 07-04-2020, 12:35 PM

If you are not using an initrd, the kernel is entirely dependent on elilo for passing it the root device name. I suspect that you have accidentally deleted that line from elilo.conf. If so, it's not a big problem. You just have to edit it back in again.

/boot/efi is empty because it's a mount point. The EFI system partition can be mounted there when required, but it isn't mounted by default. You have to mount it by hand (as root). You can do that from Ubuntu. You don't need to mount your Slackware partition at all; just mount the ESP on Ubuntu's /boot/efi. Then you can take a good look at elilo.conf and see if it passes the root partition correctly. Incidently, the reason Slackware's /proc directory was empty is because you weren't running Slackware when you looked. All the proc stuff from the kernel was in Ubuntu's /proc.

In your panic output, I think the actual error message comes just before the backtrace that you printed.

tdmsoares · 07-04-2020, 01:49 PM

I thought /boot was inside the root partition and not a mount point... More one thing learned!

I mounted the ESP and now I found my elilo.conf, here is the content of the file:

Code:

chooser=simple
delay=1
timeout=1
#
image=vmlinuz
        label=vmlinuz
        initrd=initrd.gz
        read-only
        append="root=/dev/sda4 vga=normal ro"
        append="resume=/dev/sda5"

For the hibernation I added the last line, for the hibernation it worked, but when trying to turn on the laptop I get kernel panic

Petri Kaukasoina · 07-04-2020, 02:10 PM

Quote:

Originally Posted by tdmsoares

Code:

        append="root=/dev/sda4 vga=normal ro"
        append="resume=/dev/sda5"

I don't know elilo, but in lilo you can't have two append lines. Why not put everything in one line?

baumei · 07-04-2020, 02:50 PM

With LILO, the kernel configuration parameter for additional append lines goes like this:

Code:

addappend="resume=/dev/sda5"

I imagine that your second append-line obliterated the first, and so the kernel does not know where the root partition is.

Petri Kaukasoina · 07-04-2020, 03:10 PM

I'm not sure addappend works in elilo. But there were already three kernel commandline parameters in the append command, so why not put the fourth there, too, like this:

Code:

chooser=simple
delay=1
timeout=1
#
image=vmlinuz
        label=vmlinuz
        initrd=initrd.gz
        read-only
        append="root=/dev/sda4 vga=normal ro resume=/dev/sda5"

tdmsoares · 07-04-2020, 04:04 PM

Quote:

Originally Posted by baumei

With LILO, the kernel configuration parameter for additional append lines goes like this:

Code:

addappend="resume=/dev/sda5"

I imagine that your second append-line obliterated the first, and so the kernel does not know where the root partition is.

I add

Code:

addappend

and it doesn't work with elilo, not a Kernel Panic but I cannot boot

Quote:

Originally Posted by Petri Kaukasoina

I'm not sure addappend works in elilo. But there were already three kernel commandline parameters in the append command, so why not put the fourth there, too, like this:

Code:

chooser=simple
delay=1
timeout=1
#
image=vmlinuz
        label=vmlinuz
        initrd=initrd.gz
        read-only
        append="root=/dev/sda4 vga=normal ro resume=/dev/sda5"

So this is the cause: only one append line is allowed, more than one, and with elilo resuming from hibernating works, but a simple reboot it will get a kernel panic.

Now my slackware is doing fine! Thanks!