Re: [tip:core/urgent] rcu: avoid pointless blocked-task warnings

Re: [tip:core/urgent] rcu: avoid pointless blocked-task warnings

[Ingo Molnar]

* tip-bot for Paul E. McKenney <paul.mckenney@linaro.org> wrote:

> Commit-ID:  b24efdfdf679cf9b05947c531971905fc727dd40
> Gitweb:     http://git.kernel.org/tip/b24efdfdf679cf9b05947c531971905fc727dd40
> Author:     Paul E. McKenney <paul.mckenney@linaro.org>
> AuthorDate: Wed, 12 Jan 2011 14:18:11 -0800
> Committer:  Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> CommitDate: Fri, 14 Jan 2011 04:58:08 -0800
> 
> rcu: avoid pointless blocked-task warnings
> 
> If the RCU callback-processing kthread has nothing to do, it parks in
> a wait_event().  If RCU remains idle for more than two minutes, the
> kernel complains about this.  This commit changes from wait_event()
> to wait_event_interruptible() to prevent the kernel from complaining
> just because RCU is idle.

Btw., a sidenote, it's not just about the blocked-tasks lockup detector warning, 
uninterruptible sleeps will also artificially inflate the load average of the box, 
+1.0 per such task. So it's definitely a bug to sleep for a very long time 
uninterruptible.

Thanks,

	Ingo

Re: [tip:core/urgent] rcu: avoid pointless blocked-task warnings

[Paul E. McKenney]

On Sat, Jan 15, 2011 at 03:28:57PM +0100, Ingo Molnar wrote:
> 
> * tip-bot for Paul E. McKenney <paul.mckenney@linaro.org> wrote:
> 
> > Commit-ID:  b24efdfdf679cf9b05947c531971905fc727dd40
> > Gitweb:     http://git.kernel.org/tip/b24efdfdf679cf9b05947c531971905fc727dd40
> > Author:     Paul E. McKenney <paul.mckenney@linaro.org>
> > AuthorDate: Wed, 12 Jan 2011 14:18:11 -0800
> > Committer:  Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> > CommitDate: Fri, 14 Jan 2011 04:58:08 -0800
> > 
> > rcu: avoid pointless blocked-task warnings
> > 
> > If the RCU callback-processing kthread has nothing to do, it parks in
> > a wait_event().  If RCU remains idle for more than two minutes, the
> > kernel complains about this.  This commit changes from wait_event()
> > to wait_event_interruptible() to prevent the kernel from complaining
> > just because RCU is idle.
> 
> Btw., a sidenote, it's not just about the blocked-tasks lockup detector warning, 
> uninterruptible sleeps will also artificially inflate the load average of the box, 
> +1.0 per such task. So it's definitely a bug to sleep for a very long time 
> uninterruptible.

Good point, and I have made the corresponding change in the in-progress
TREE_PREEMPT_RCU variant of RCU priority boosting.  Some thought will be
required when merging SRCU into the Tiny and Tree RCU implementations.  ;-)

								Thanx, Paul