[dm-devel] dm raid: avoid mddev->suspended access

[Mike Snitzer]

On Tue, Jul 25 2017 at  2:17pm -0400,
Mike Snitzer <snitzer at redhat.com> wrote:

> On Thu, Jul 13 2017 at 11:34am -0400,
> heinzm at redhat.com <heinzm at redhat.com> wrote:
> 
> > From: Heinz Mauelshagen <heinzm at redhat.com>
> > 
> > Use runtime flag to ensure that an mddev gets suspended/resumed just once.
> > 
> > Signed-off-by: Heinz Mauelshagen <heinzm at redhat.com>
> > ---
> >  drivers/md/dm-raid.c | 12 +++++++-----
> >  1 file changed, 7 insertions(+), 5 deletions(-)
> > 
> > diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c
> > index b409015..60c524b 100644
> > --- a/drivers/md/dm-raid.c
> > +++ b/drivers/md/dm-raid.c
> > @@ -3760,7 +3762,7 @@ static int rs_start_reshape(struct raid_set *rs)
> >  		return r;
> >  
> >  	/* Need to be resumed to be able to start reshape, recovery is frozen until raid_resume() though */
> > -	if (mddev->suspended)
> > +	if (test_and_clear_bit(RT_FLAG_RS_SUSPENDED, &rs->runtime_flags))
> >  		mddev_resume(mddev);
> >  
> >  	/*
> > @@ -3787,8 +3789,8 @@ static int rs_start_reshape(struct raid_set *rs)
> >  	}
> >  
> >  	/* Suspend because a resume will happen in raid_resume() */
> > -	if (!mddev->suspended)
> > -		mddev_suspend(mddev);
> > +	set_bit(RT_FLAG_RS_SUSPENDED, &rs->runtime_flags);
> > +	mddev_suspend(mddev);
> >  
> >  	/*
> >  	 * Now reshape got set up, update superblocks to
> 
> Shouldn't this be the following?
> 
> if (!test_and_set_bit(RT_FLAG_RS_SUSPENDED, &rs->runtime_flags))
>    mddev_suspend(mddev);

Looking closer, the preceding test_and_clear_bit() should ensure that
the device is always resumed by the time it gets to the code I called
into question.