Re: [PATCH] zram: fix possible race when checking idle_strm

[Sergey Senozhatsky]

On (08/07/15 18:37), Sergey Senozhatsky wrote:
[..]
> > Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@xxxxxxx>
> 
> Acked-by: Sergey Senozhatsky <sergey.senozhatsky@xxxxxxxxx>

Andrew, my apologies, can we hold on a bit with my Acked-by. I
need more time to think about the issue -- do we actually have
one or we don't.

It seems that to be on a safe side this check is better be done
under the lock, but at the same time I sort of fail to convenience
myself that doing this check outside the lock introduces any danger.
Need some rest. May be Joonsoo and Minchan can add some bits
(that I currently miss) or thoughts.

	-ss

> > ---
> >  drivers/block/zram/zcomp.c | 15 +++++++++++++--
> >  1 file changed, 13 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/block/zram/zcomp.c b/drivers/block/zram/zcomp.c
> > index 80a62e7..837e9c3 100644
> > --- a/drivers/block/zram/zcomp.c
> > +++ b/drivers/block/zram/zcomp.c
> > @@ -76,6 +76,17 @@ static struct zcomp_strm *zcomp_strm_alloc(struct zcomp *comp)
> >  	return zstrm;
> >  }
> >  
> > +static bool avail_idle_strm(struct zcomp_strm_multi *zs)
> > +{
> > +	int avail;
> > +
> > +	spin_lock(&zs->strm_lock);
> > +	avail = !list_empty(&zs->idle_strm);
> > +	spin_unlock(&zs->strm_lock);
> > +
> > +	return avail;
> > +}
> > +
> >  /*
> >   * get idle zcomp_strm or wait until other process release
> >   * (zcomp_strm_release()) one for us
> > @@ -97,7 +108,7 @@ static struct zcomp_strm *zcomp_strm_multi_find(struct zcomp *comp)
> >  		/* zstrm streams limit reached, wait for idle stream */
> >  		if (zs->avail_strm >= zs->max_strm) {
> >  			spin_unlock(&zs->strm_lock);
> > -			wait_event(zs->strm_wait, !list_empty(&zs->idle_strm));
> > +			wait_event(zs->strm_wait, avail_idle_strm(zs));
> >  			continue;
> >  		}
> >  		/* allocate new zstrm stream */
> > @@ -109,7 +120,7 @@ static struct zcomp_strm *zcomp_strm_multi_find(struct zcomp *comp)
> >  			spin_lock(&zs->strm_lock);
> >  			zs->avail_strm--;
> >  			spin_unlock(&zs->strm_lock);
> > -			wait_event(zs->strm_wait, !list_empty(&zs->idle_strm));
> > +			wait_event(zs->strm_wait, avail_idle_strm(zs));
> >  			continue;
> >  		}
> >  		break;
> > -- 
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/