linux-kernel - Re: [PATCH] OOB read and write in mdp_prepare

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20230628063353.GA3625616@gnbcxd0016.gnb.st.com>
Date:   Wed, 28 Jun 2023 08:33:53 +0200
From:   Alain Volmat <alain.volmat@...s.st.com>
To:     sun yq <yqsun1997@...il.com>
CC:     <mchehab@...nel.org>, <matthias.bgg@...il.com>,
        <angelogioacchino.delregno@...labora.com>,
        <linux-media@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
        <linux-arm-kernel@...ts.infradead.org>,
        <linux-mediatek@...ts.infradead.org>, <499671216@...com>
Subject: Re: [PATCH] OOB read and write in mdp_prepare_buffer

Hi,

On Wed, Jun 28, 2023 at 07:28:54AM +0800, sun yq wrote:
> Hi,
> Because there are many functions using the plane, increasing the max
> number of the plane is to maximize the solution to all possible oob
> places.

I don't think it is the right approach then.  If the HW is only handling
3 planes, there should be no reason to have to allocate for 8 planes.  I
suspect that this 8 value is coming from the maximum allowed plane
number in V4L2 right ?
INHO driver should simply be fixed to ensure that num_plane won't go
higher than the real number of plane allocated in the structures.
It should be possible to get the num_plane value from the format
selected.

Alain

> 
> On Tue, Jun 27, 2023 at 10:06 PM Alain Volmat <alain.volmat@...s.st.com> wrote:
> >
> > Hi,
> >
> > On Tue, Jun 27, 2023 at 04:27:31PM +0800, yqsun1997@...il.com wrote:
> > > From: yqsun1997 <yqsun1997@...il.com>
> > >
> > > Because format in struct img_image_buffer max index is IMG_MAX_PLANES ==3,
> > > The num_planes max index is 8.so will be OOB like in mdp_prepare_buffer.
> >
> > Similarly as your other patch, could you describe why you need to
> > increase the IMG_MAX_PLANES while I suspect your driver only needs to
> > deal with 3 planes.  While the maximum num_planes value that can be
> > given by the user is 8, this has to be first compared to the configured
> > format prior to reaching this function.
> >
> > >
> > > static void mdp_prepare_buffer(struct img_image_buffer *b,
> > >                                struct mdp_frame *frame, struct vb2_buffer *vb)
> > > {
> > >         struct v4l2_pix_format_mplane *pix_mp = &frame->format.fmt.pix_mp;
> > >         unsigned int i;
> > >
> > >         b->format.colorformat = frame->mdp_fmt->mdp_color;
> > >         b->format.ycbcr_prof = frame->ycbcr_prof;
> > >         for (i = 0; i < pix_mp->num_planes; ++i) {
> > >                 u32 stride = mdp_fmt_get_stride(frame->mdp_fmt,
> > >                         pix_mp->plane_fmt[i].bytesperline, i);
> > >
> > >                 b->format.plane_fmt[i].stride = stride;  //oob
> > >                 ......
> > >
> > > Signed-off-by: yqsun1997 <yqsun1997@...il.com>
> > > ---
> > >  drivers/media/platform/mediatek/mdp3/mtk-mdp3-type.h | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-type.h b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-type.h
> > > index ae0396806..e2e991a34 100644
> > > --- a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-type.h
> > > +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-type.h
> > > @@ -11,7 +11,7 @@
> > >
> > >  #define IMG_MAX_HW_INPUTS    3
> > >  #define IMG_MAX_HW_OUTPUTS   4
> > > -#define IMG_MAX_PLANES               3
> > > +#define IMG_MAX_PLANES               8
> > >  #define IMG_MAX_COMPONENTS   20
> > >
> > >  struct img_crop {
> > > --
> > > 2.39.2
> > >
> >
> > Regards,
> > Alain