Re: rcu_prempt stalls / lockup

[Paul E. McKenney]

On Tue, Apr 01, 2014 at 06:16:16PM -0400, Dave Jones wrote:
> On Tue, Apr 01, 2014 at 11:32:45AM -0700, Paul E. McKenney wrote:
> 
>  > >  > Given that you have CONFIG_RCU_TRACE=y, could you please enable the
>  > >  > following trace events and dump the trace before things hang?
>  > >  > 
>  > >  > 	trace_event=rcu:rcu_grace_period,rcu:rcu_grace_period_init
>  > >  > 
>  > >  > If it is not feasible to dump the trace before things hang, let me
>  > >  > know, and I will work out some other diagnostic regime.
>  > > 
>  > > I'll give that a shot when I get back in a few hours.
> 
> ok, this is tricky, and I'm not sure how helpful the below is.
> because I don't know when the hang is going to happen, in parallel to my usual workload
> I did this..
> while [ 1 ];
> do
>   cat trace
>   echo > trace
> done
> 
> and got this...
> 
> 
> 
> # entries-in-buffer/entries-written: 7/7   #P:4
> #
> #                              _-----=> irqs-off
> #                             / _----=> need-resched
> #                            | / _---=> hardirq/softirq
> #                            || / _--=> preempt-depth
> #                            ||| /     delay
> #           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
> #              | |       |   ||||       |         |
>           <idle>-0     [000] dNs3   851.748475: rcu_grace_period: rcu_preempt 21342 cpuend
>           <idle>-0     [000] dNs3   851.748477: rcu_grace_period: rcu_preempt 21343 cpustart
>           <idle>-0     [000] dNs3   851.748477: rcu_grace_period: rcu_sched 20080 cpuend
>           <idle>-0     [000] dN.2   851.748480: rcu_grace_period: rcu_preempt 21343 cpuqs
>      rcu_preempt-9     [000] ...1   851.748485: rcu_grace_period: rcu_preempt 21343 fqsstart
>      rcu_preempt-9     [000] ...1   851.748487: rcu_grace_period: rcu_preempt 21343 fqsend
>      rcu_preempt-9     [000] ...1   851.748487: rcu_grace_period: rcu_preempt 21343 fqswait
> # tracer: nop
> #
> # entries-in-buffer/entries-written: 13/13   #P:4
> #
> #                              _-----=> irqs-off
> #                             / _----=> need-resched
> #                            | / _---=> hardirq/softirq
> #                            || / _--=> preempt-depth
> #                            ||| /     delay
> #           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
> #              | |       |   ||||       |         |
>      rcu_preempt-9     [003] d..2   851.878457: rcu_grace_period: rcu_preempt 21349 end
>      rcu_preempt-9     [003] d..2   851.878459: rcu_grace_period: rcu_preempt 21349 newreq
>      rcu_preempt-9     [003] ...1   851.878459: rcu_grace_period: rcu_preempt 21349 reqwait
>      rcu_preempt-9     [003] d..2   851.878460: rcu_grace_period: rcu_preempt 21350 start
>      rcu_preempt-9     [003] d..2   851.878461: rcu_grace_period: rcu_preempt 21350 cpustart
>      rcu_preempt-9     [003] d..2   851.878462: rcu_grace_period_init: rcu_preempt 21350 0 0 7 f
>      rcu_preempt-9     [003] ...1   851.878463: rcu_grace_period: rcu_preempt 21350 fqswait
>      rcu_preempt-9     [003] d..2   851.878464: rcu_grace_period: rcu_preempt 21350 cpuqs
>           <idle>-0     [000] dNs3   851.878482: rcu_grace_period: rcu_preempt 21349 cpuend
>           <idle>-0     [000] dNs3   851.878483: rcu_grace_period: rcu_preempt 21350 cpustart
>           <idle>-0     [000] dNs3   851.878484: rcu_grace_period: rcu_sched 20086 cpuend
>           <idle>-0     [000] .N.2   851.878486: rcu_grace_period: rcu_sched 20086 cpuqs
>           <idle>-0     [000] dN.2   851.878487: rcu_grace_period: rcu_preempt 21350 cpuqs
> 
> 
> followed by a half dozen 'empty' traces before it totally locked up.

Cool, thank you!

Could you please also send the RCU CPU stall warning messages?

							Thanx, Paul

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/