Re: INFO: task rcu_gp:3 blocked for more than 122 seconds.

[Paul E. McKenney]

On Fri, Sep 30, 2022 at 12:19:34PM +0200, Borislav Petkov wrote:
> On Thu, Sep 29, 2022 at 08:37:22AM -0700, Paul E. McKenney wrote:
> > As discussed on IRC, I got nuttin' on this one.  Looks like an RCU
> > expedited grace period is stuck waiting for its workqueue kthread, and
> > that this workqueue kthread is stuck for some reason.
> 
> Yeah, it is not that easy to repro - I need to wait a couple of minutes
> after it boots. Look at the timestamps:
> 
> [  249.149378] INFO: task rcu_gp:3 blocked for more than 122 seconds.
> ^^^^^^^^^^^^^^

That can be annoying!  I well recall one bisection session where each step
took several hours, and where each step required manual intervention,
and where the bug was urgent.  I got the bisection overnight, but was
pretty useless then next day.  ;-)

> But it looks like sched/core is the culprit because merging it into
> Linus master triggers it.

It is possible that Peter would want proof of this.  ;-)

> Lemme bisect that.

Please, and thank you in advance!

							Thanx, Paul

> Stay tuned.
> 
> -- 
> Regards/Gruss,
>     Boris.
> 
> https://people.kernel.org/tglx/notes-about-netiquette