Re: net/bluetooth: workqueue destruction WARNING in hci_unregister_dev

From: Tejun Heo
Date: Fri Mar 11 2016 - 12:12:41 EST


Hello,

Sorry about the delay.

On Thu, Mar 03, 2016 at 10:12:01AM +0100, Jiri Slaby wrote:
> On 03/02/2016, 04:45 PM, Tejun Heo wrote:
> > On Fri, Feb 19, 2016 at 01:10:00PM +0100, Jiri Slaby wrote:
> >>> 1. didn't help, the problem persists. So I haven't applied the patch from 2.
> >>
> >> FWIW I dumped more info about the wq:
> >> wq->name='hci0' pwq=ffff8800390d7600 wq->dfl_pwq=ffff8800390d5200
> >> pwq->refcnt=2 pwq->nr_active=0 delayed_works: <nothing>
> >
> > Can you please print out the same info for all pwq's during shutdown?
> > It looks like we're leaking pwq refcnt but I can't spot a place where
> > that could happen on an empty pwq.
>
> I have not done that yet, but today, I see:
> destroy_workqueue: name='req_hci0' pwq=ffff88002f590300
> wq->dfl_pwq=ffff88002f591e00 pwq->refcnt=2 pwq->nr_active=0 delayed_works:
> pwq 12: cpus=0-1 node=0 flags=0x4 nice=-20 active=0/1
> in-flight: 18568:wq_barrier_func

So, this means that there's flush_work() racing against workqueue
destruction, which can't be safe. :(

Thanks.

--
tejun