linux-kernel - Re: [printk] b031a684bf: INFO:rcu_tasks_detected_stalls_on

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <YBLPhkHQ8cXFiY1X@alley>
Date:   Thu, 28 Jan 2021 15:51:50 +0100
From:   Petr Mladek <pmladek@...e.com>
To:     John Ogness <john.ogness@...utronix.de>
Cc:     Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
        kernel test robot <oliver.sang@...el.com>,
        LKML <linux-kernel@...r.kernel.org>, lkp@...ts.01.org,
        lkp@...el.com, zhengjun.xing@...ux.intel.com
Subject: Re: [printk]  b031a684bf: INFO:rcu_tasks_detected_stalls_on_tasks

On Thu 2021-01-28 12:32:46, John Ogness wrote:
> On 2021-01-28, Sergey Senozhatsky <sergey.senozhatsky@...il.com> wrote:
> >> # modprobe rcutorture onoff_interval=3 onoff_holdoff=30 torture_type=tasks
> >> 
> >> (Those are the same modprobe parameters used by the lkp job.)
> >> 
> >> After about a minute I see:
> >> 
> >> [   47.268292] tasks-torture: rcu_torture_read_exit: Start of episode
> >> [   51.273365] tasks-torture: rcu_torture_read_exit: End of episode
> >> [   55.823306] smpboot: do_boot_cpu failed(-1) to wakeup CPU#0
> >> [   55.824350] tasks-torture:torture_onoff task: online 0 failed: errno -5
> >> [   55.830661] tasks-torture:torture_onoff task: online 0 failed: errno -5
> >> [   55.848524] tasks-torture:torture_onoff task: online 0 failed: errno -5
> >
> > Just out of curious, this is seen only with the printk commit in
> > question applied?
> 
> No. The error messages are not related to the commit. But they are
> really the only thing printk'ing when the rcu stall happens. So you
> probably do need to see them in order to reproduce this. (Assuming this
> is somehow printk related.)

If you are able to reproduce this, it might be interesting to see
backtraces from all CPUs when RCU stall gets detected. Or even printing
all processes, even the sleeping ones.

I am still curious how this could be printk() related.

Best Regards,
Petr