Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Message-ID: <aa2a937379d85a93358f3d63f92e2eb23b4006d7.camel@bootlin.com>
Subject: Re: [PATCH v5 3/4] drm/vc4: Check for the binner bo before handling
 OOM interrupt
From:   Paul Kocialkowski <paul.kocialkowski@bootlin.com>
To:     Eric Anholt <eric@anholt.net>, dri-devel@lists.freedesktop.org,
        linux-kernel@vger.kernel.org
Cc:     David Airlie <airlied@linux.ie>, Daniel Vetter <daniel@ffwll.ch>,
        Thomas Petazzoni <thomas.petazzoni@bootlin.com>,
        Maxime Ripard <maxime.ripard@bootlin.com>,
        Eben Upton <eben@raspberrypi.org>
Date:   Tue, 23 Apr 2019 10:50:59 +0200
In-Reply-To: <d9ccaaf4db46abf0098fbeb178a42c71a7860b9d.camel@bootlin.com>
References: <20190415123908.28986-1-paul.kocialkowski@bootlin.com>
         <20190415123908.28986-4-paul.kocialkowski@bootlin.com>
         <87v9zf806m.fsf@anholt.net>
         <d9ccaaf4db46abf0098fbeb178a42c71a7860b9d.camel@bootlin.com>
Organization: Bootlin
Content-Type: text/plain; charset="UTF-8"
User-Agent: Evolution 3.32.1 
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, 2019-04-23 at 10:30 +0200, Paul Kocialkowski wrote:
> Hi,
> 
> On Mon, 2019-04-15 at 13:48 -0700, Eric Anholt wrote:
> > Paul Kocialkowski <paul.kocialkowski@bootlin.com> writes:
> > 
> > > Since the OOM interrupt directly deals with the binner bo, it doesn't
> > > make sense to try and handle it without a binner buffer registered.
> > > 
> > > Signed-off-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
> > > ---
> > >  drivers/gpu/drm/vc4/vc4_irq.c | 3 +++
> > >  1 file changed, 3 insertions(+)
> > > 
> > > diff --git a/drivers/gpu/drm/vc4/vc4_irq.c b/drivers/gpu/drm/vc4/vc4_irq.c
> > > index ffd0a4388752..723dc86b4511 100644
> > > --- a/drivers/gpu/drm/vc4/vc4_irq.c
> > > +++ b/drivers/gpu/drm/vc4/vc4_irq.c
> > > @@ -64,6 +64,9 @@ vc4_overflow_mem_work(struct work_struct *work)
> > >  	struct vc4_exec_info *exec;
> > >  	unsigned long irqflags;
> > >  
> > > +	if (!bo)
> > > +		return;
> > > +
> > >  	bin_bo_slot = vc4_v3d_get_bin_slot(vc4);
> > >  	if (bin_bo_slot < 0) {
> > >  		DRM_ERROR("Couldn't allocate binner overflow mem\n");
> > > -- 
> > > 2.21.0
> > 
> > I don't think this is going to be race-free.  You're checking outside of
> > a lock, then proceeding to use it even if (in patch 4) the bin BO was in
> > the process of being freed during the file close path.  Can we put all
> > of the overflow process here under the same lock as freeing?
> 
> Definitely, sorry I missed that.

That being said, I think it should be fixed in 4/4 since this patch is
not problematic on its own (we haven't introduced the mutex/kref yet). 

Cheers,

Paul

-- 
Paul Kocialkowski, Bootlin
Embedded Linux and kernel engineering
https://bootlin.com