[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#1111353: linux-image-6.1.0-38-cloud-amd64: system crashed with rcu_preempt kthread starved for 5252 jiffies error



Hi Salvatore,

You can mark this bug as a duplicate of #1111017 (linux 6.1.147 hangs when loading audit rules / booting).

I tested the following and that temporarily fixed the issue:

sudo systemctl stop auditd
sudo apt install linux-image-6.1.0-36-cloud-amd64

Without stopping auditd first, the second command hangs the system with a 100% certainty.

PS: do you have an indication for when 6.1.148 will be released?

-- 
Mvg, Sander Devrieze.


Op zo 17 aug 2025 om 13:20 schreef Salvatore Bonaccorso <carnil@debian.org>:
Control: tags -1 + moreinfo

Hi

On Sun, Aug 17, 2025 at 11:36:05AM +0200, sander wrote:
> Package: src:linux
> Version: 6.1.147-1
> Severity: important
> X-Debbugs-Cc: s.devrieze@gmail.com
>
> Dear Maintainer,
>
>    * What led up to the situation?
>
> Executing 'composer update' to update a website's software stack or installing a new package with apt, make the system crash with latest kernel version.
>
>    * What exactly did you do (or not do) that was effective (or
>      ineffective)?
>
> Rebooting and running the previous kernel version.
>
>    * What was the outcome of this action?
>
> rcu: INFO: rcu-preempt detected stalls on CPUs/tasks:
> rcu: All QSes seen, last rcu_preempt kthread activity 5252 (4294928323-4294923071), jiffies_till_next_fqs=1, root ->qsmask 0x0
> rcu: rcu_preempt kthread starved for 5252 jiffies! g41193 f0x2 RCU_GP_WAIT_FQS(5) ->state=0x0 ->cpu=0
> rcu: #Unless rcu_preempt kthread gets sufficient CPU time, OOM is now expected behavior.
> rcu: RCU grace-period kthread stack dump:
> rcu: Stack dump where RCU GP kthread last ran:
>
>    * What outcome did you expect instead?

Can you capture a full kernel log when this happens and provide it
here?

Do you have some easy reproducing steps for it? In case yes would you
be able to do some bisection to find the offending commit between
6.1.140 and 6.1.147?

If you can test on that infrastrcture: Does the version in trixie
(6.12.41-1) or experimental (6.16.1-1~exp1) exhibit the problem as
well?

Most ideal is if we have an simple reproducer so that bisection can be
performed and isolate the problem.

Regards,
Salvatore

Reply to: