Re: Kernel panic - not syncing: Attempted to kill init! exitcode=0x00007f00

[Valerio Vanni]

Il 25/02/2023 13:28, Valerio Vanni ha scritto:
> Il 24/02/2023 19:06, Slade Watkins ha scritto:
>
> > > https://bugzilla.kernel.org/show_bug.cgi?id=217083
> > >
> > > I got this error after upgrading a Debian machine (x86-64) from Stretch
> > > to Buster. Upgrade is successful, but the next boot it crashes.
> >
> > Thanks for the details. What was the last known good kernel version
> > that did not exhibit this issue on your system?
>
> There isn't a last known good version. As I said, the issue appeared 
> after an OS upgrade, not after a kernel one.
I updated bugzilla and I update also here.

I use serial console to collect diagnostic data, but now I've found that 
it misses lines.
In kernel boot line I put "console=ttyS0 console=tty0" so I get both 
local and serial console.

But the issue is: tty0 shows more messages. Is there a way to get 
everything also on ttyS0? It would be useful, since you cant' scroll 
tty0 after a kernel panic.

When kernel loads, tty0 scrolls very fast. I could see some more line 
but wasn't able to read them.
The only way I found was adding a (slow) speed to ttyS0, 
"console=ttyS0,1200" on kernel boot line.
This slowed down ttyS0, and tty0 was forced to go at the same pace.

So I could read a line (only on tty0), between "Run /sbin/init as init 
process" and start of kernel panic.
That line complained that it could not find something in /usr/lib/.... 
At that stage, /usr was not mounted yet.

Then I restored full disk image on the machine, an before upgrading from 
stretch to buster I merged / and /usr, so that /usr/lib/... was 
available from / mount.

After that, kernel panic did not show anymore.