Re: [PATCH] workqueue: warn if memory reclaim tries to flush !WQ_MEM_RECLAIM workqueue

From: Tejun Heo <tj@kernel.org>
To: Adrian Hunter <adrian.hunter@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>,
	Ulrich Obergfell <uobergfe@redhat.com>,
	Ingo Molnar <mingo@redhat.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	linux-kernel@vger.kernel.org, kernel-team@fb.com
Subject: Re: [PATCH] workqueue: warn if memory reclaim tries to flush !WQ_MEM_RECLAIM workqueue
Date: Fri, 11 Mar 2016 12:52:16 -0500	[thread overview]
Message-ID: <20160311175216.GK24046@htj.duckdns.org> (raw)
In-Reply-To: <56E18EF6.1010006@intel.com>

Hello,

On Thu, Mar 10, 2016 at 05:12:54PM +0200, Adrian Hunter wrote:
> > @@ -2711,6 +2744,8 @@ static bool start_flush_work(struct work
> >  		pwq = worker->current_pwq;
> >  	}
> >  
> > +	check_flush_dependency(pwq->wq, work);
> > +
> >  	insert_wq_barrier(pwq, barr, work, worker);
> >  	spin_unlock_irq(&pool->lock);
> 
> I am hitting the warnings when using cancel_delayed_work_sync().  I would
> have thought that forward progress would still be guaranteed in that case.
> Is it true that it is not?

I'm not sure I understand what you're trying to say.  If you're trying
to do cancel_delayed_work_sync() from a memreclaim wq on a work item
which is executing on !memreclaim wq, that'd be an incorrect thing to
do as that can deadlock the memreclaim wq under memory pressure.

Thanks.

-- 
tejun