Re: [PATCH 3/5] MTD: blktrans: fix a race vs kthread_stop.

[Maxim Levitsky <maximlevitsky@gmail.com>]

On Mon, 2010-10-25 at 01:31 +0100, David Woodhouse wrote:
> On Fri, 2010-10-15 at 17:20 +0200, Maxim Levitsky wrote:
> > There is small race window that could make kthread_stop hang forever.
> > I found that while hacking the IR subsystem.
> > 
> > Signed-off-by: Maxim Levitsky <maximlevisky@gmail.com>
> > ---
> >  drivers/mtd/mtd_blkdevs.c |    4 ++++
> >  1 files changed, 4 insertions(+), 0 deletions(-)
> > 
> > diff --git a/drivers/mtd/mtd_blkdevs.c b/drivers/mtd/mtd_blkdevs.c
> > index 040c2d9..a919587 100644
> > --- a/drivers/mtd/mtd_blkdevs.c
> > +++ b/drivers/mtd/mtd_blkdevs.c
> > @@ -133,6 +133,10 @@ static int mtd_blktrans_thread(void *arg)
> >  
> >  		if (!req && !(req = blk_fetch_request(rq))) {
> >  			set_current_state(TASK_INTERRUPTIBLE);
> > +
> > +			if (kthread_should_stop())
> > +				set_current_state(TASK_RUNNING);
> > +
> >  			spin_unlock_irq(rq->queue_lock);
> >  			schedule();
> >  			spin_lock_irq(rq->queue_lock);
> 
> 
> Shouldn't we add a break in there too, so it immediately breaks out of
> the loop (with the lock held)?
It doesn't matter here, as schedule() won't make task sleep, so the
thread will break at next iteration.


Best regards
	Maxim Levitsky