Kernel panic on new machine

Kernel panic on new machine

[phodina via]

[-- Attachment #1: Type: text/plain, Size: 27 bytes --]

Kernel panic on new machine

[-- Attachment #2: IMG_20220411_180003.jpg --]
[-- Type: image/jpeg, Size: 6246183 bytes --]

Re: Kernel panic on new machine

[phodina via]

Hi,

I've installed Guix to a new machine and after rebooting the live install I get kernel panic.

The machine is Dell x86_64 uefi machine with FAT32 boot partition and Luks partition with Btrfs filesystem with rootfs.

Could you please help me?

Petr

Re: Kernel panic on new machine

[phodina via]

[-- Attachment #1: Type: text/plain, Size: 21 bytes --]

Here's the screenshot

[-- Attachment #2: IMG_20220411_180003.jpg --]
[-- Type: image/jpeg, Size: 6246183 bytes --]

Re: Kernel panic on new machine

[phodina via]

[-- Attachment #1: Type: text/plain, Size: 17 bytes --]

Here's screenshot

[-- Attachment #2: IMG_20220411_190703.jpg --]
[-- Type: image/jpeg, Size: 827432 bytes --]

Re: Kernel panic on new machine

[raingloom]

On Mon, 11 Apr 2022 16:51:30 +0000
phodina via <help-guix@gnu.org> wrote:

> Hi,
> 
> I've installed Guix to a new machine and after rebooting the live
> install I get kernel panic.
> 
> The machine is Dell x86_64 uefi machine with FAT32 boot partition and
> Luks partition with Btrfs filesystem with rootfs.
> 
> Could you please help me?
> 
> Petr

This is a Guix init issue, not "really" a kernel panic. As you can see,
there is some kind of error in the Scheme code being run at boot time,
but instead of being dropped into a debug shell, it exits, which causes
the kernel to panic.
This is because Guile is running as PID 1, that is, it is the first
process the kernel starts. PID 1 is kinda special and AFAIK it is not
allowed to exit, ever. Otherwise you get that panic.

Re: Kernel panic on new machine

[phodina]

Hi,

> This is a Guix init issue, not "really" a kernel panic. As you can see,
> there is some kind of error in the Scheme code being run at boot time,
> but instead of being dropped into a debug shell, it exits, which causes
> the kernel to panic.
> This is because Guile is running as PID 1, that is, it is the first
> process the kernel starts. PID 1 is kinda special and AFAIK it is not
> allowed to exit, ever. Otherwise you get that panic.

yes, I know Guix init process exiting and due to being PID 1 kernel panics.

In the meantime I reinstalled Guix again and it booted correctly, no idea what caused it.


However, my question was/still is how to fix this? Also is there a way to determine what caused this? Is there a way to launch the debug shell instead of exiting?

PS: Sorry for the image spam but it appeared that the image was not posted the first time so I tried also sending a smaller image

Petr

Re: Kernel panic on new machine

[raingloom]

On Thu, 14 Apr 2022 07:47:47 +0000
phodina <phodina@protonmail.com> wrote:

> Hi,
> 
> > This is a Guix init issue, not "really" a kernel panic. As you can
> > see, there is some kind of error in the Scheme code being run at
> > boot time, but instead of being dropped into a debug shell, it
> > exits, which causes the kernel to panic.
> > This is because Guile is running as PID 1, that is, it is the first
> > process the kernel starts. PID 1 is kinda special and AFAIK it is
> > not allowed to exit, ever. Otherwise you get that panic.  
> 
> yes, I know Guix init process exiting and due to being PID 1 kernel
> panics.
> 
> In the meantime I reinstalled Guix again and it booted correctly, no
> idea what caused it.
> 
> 
> However, my question was/still is how to fix this? Also is there a
> way to determine what caused this? Is there a way to launch the debug
> shell instead of exiting?
> 
> PS: Sorry for the image spam but it appeared that the image was not
> posted the first time so I tried also sending a smaller image
> 
> Petr
> 

Maybe you could boot that system profile in QEMU.
It looks like it is trying to open a debug shell but failing, so... not
sure what you could do in that case if you are on bare metal. At least
on QEMU you might be able to generate some kind of core dump or attach
gdb to it or something.