Re: [Patch 07/16] media: ti-vpe: vpe: fix a v4l2-compliance failure causing a kernel panic

[Benoit Parrot]

Hans Verkuil <hverkuil@xxxxxxxxx> wrote on Mon [2019-Sep-30 10:35:05 +0200]:
> On 9/27/19 8:36 PM, Benoit Parrot wrote:
> > v4l2-compliance fails with this message:
> > 
> >    warn: v4l2-test-formats.cpp(717): \
> >    	TRY_FMT cannot handle an invalid pixelformat.
> >    test VIDIOC_TRY_FMT: FAIL
> > 
> > This causes the following kernel panic:
> > 
> > Unable to handle kernel paging request at virtual address 56595561
> > pgd = ecd80e00
> > *pgd=00000000
> > Internal error: Oops: 205 [#1] PREEMPT SMP ARM
> > ...
> > CPU: 0 PID: 930 Comm: v4l2-compliance Not tainted \
> > 	4.14.62-01715-gc8cd67f49a19 #1
> > Hardware name: Generic DRA72X (Flattened Device Tree)
> > task: ece44d80 task.stack: ecc6e000
> > PC is at __vpe_try_fmt+0x18c/0x2a8 [ti_vpe]
> > LR is at 0x8
> > 
> > Because the driver fails to properly check the 'num_planes' values for
> > proper ranges it ends up accessing out of bound data causing the kernel
> > panic.
> > 
> > Since this driver only handle single or dual plane pixel format, make
> > sure the provided value does not exceed 2 planes.
> > 
> > Signed-off-by: Benoit Parrot <bparrot@xxxxxx>
> > Reviewed-by: Tomi Valkeinen <tomi.valkeinen@xxxxxx>
> > ---
> >  drivers/media/platform/ti-vpe/vpe.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/drivers/media/platform/ti-vpe/vpe.c b/drivers/media/platform/ti-vpe/vpe.c
> > index bbbf11174e16..1278d457f753 100644
> > --- a/drivers/media/platform/ti-vpe/vpe.c
> > +++ b/drivers/media/platform/ti-vpe/vpe.c
> > @@ -1650,7 +1650,7 @@ static int __vpe_try_fmt(struct vpe_ctx *ctx, struct v4l2_format *f,
> >  			      &pix->height, MIN_H, MAX_H, H_ALIGN,
> >  			      S_ALIGN);
> >  
> > -	if (!pix->num_planes)
> > +	if (!pix->num_planes || pix->num_planes > 2)
> >  		pix->num_planes = fmt->coplanar ? 2 : 1;
> >  	else if (pix->num_planes > 1 && !fmt->coplanar)
> >  		pix->num_planes = 1;
> > 
> 
> This looks weird.
> 
> Why not just unconditionally do:
> 
> 	pix->num_planes = fmt->coplanar ? 2 : 1;

In order to not change the behavior, VPE would assume that NV12 format for
instance were always sent as 2 separate planes buffers. So for backward
compatibility this is order to handle both cases where NV12 could be
handled as both a single plane and a dual plane buffers based on what the
user space application intent on passing in/out.

Benoit

> 
> Regards,
> 
> 	Hans