linux-kernel - Re: [PATCH v7 2/4] drm/mediatek: Fix using wrong drm private data to bind mediatek-drm

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]

Date:   Mon, 31 Jul 2023 08:21:28 +0000
From:   Jason-JH Lin (林睿祥) 
        <Jason-JH.Lin@...iatek.com>
To:     "chunkuang.hu@...nel.org" <chunkuang.hu@...nel.org>,
        "angelogioacchino.delregno@...labora.com" 
        <angelogioacchino.delregno@...labora.com>,
        "eugen.hristev@...labora.com" <eugen.hristev@...labora.com>
CC:     "linux-mediatek@...ts.infradead.org" 
        <linux-mediatek@...ts.infradead.org>,
        Singo Chang (張興國) 
        <Singo.Chang@...iatek.com>,
        Johnson Wang (王聖鑫) 
        <Johnson.Wang@...iatek.com>,
        Jason-ch Chen (陳建豪) 
        <Jason-ch.Chen@...iatek.com>,
        Shawn Sung (宋孝謙) 
        <Shawn.Sung@...iatek.com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Nancy Lin (林欣螢) <Nancy.Lin@...iatek.com>,
        "dri-devel@...ts.freedesktop.org" <dri-devel@...ts.freedesktop.org>,
        Project_Global_Chrome_Upstream_Group 
        <Project_Global_Chrome_Upstream_Group@...iatek.com>,
        "linux-arm-kernel@...ts.infradead.org" 
        <linux-arm-kernel@...ts.infradead.org>,
        "matthias.bgg@...il.com" <matthias.bgg@...il.com>
Subject: Re: [PATCH v7 2/4] drm/mediatek: Fix using wrong drm private data to
 bind mediatek-drm

Hi Eugen,

Thanks for the reviews.

On Fri, 2023-07-28 at 11:47 +0300, Eugen Hristev wrote:
> Hi,
> 
> On 7/27/23 19:41, Jason-JH.Lin wrote:
> > Add checking the length of each data path before assigning drm
> > private
> > data into all_drm_priv array.
> > 
> > Fixes: 1ef7ed48356c ("drm/mediatek: Modify mediatek-drm for mt8195
> > multi mmsys support")
> > Signed-off-by: Jason-JH.Lin <jason-jh.lin@...iatek.com>
> > ---
> >   drivers/gpu/drm/mediatek/mtk_drm_drv.c | 16 +++++++++++++---
> >   1 file changed, 13 insertions(+), 3 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/mediatek/mtk_drm_drv.c
> > b/drivers/gpu/drm/mediatek/mtk_drm_drv.c
> > index 249c9fd6347e..d2fb1fb4e682 100644
> > --- a/drivers/gpu/drm/mediatek/mtk_drm_drv.c
> > +++ b/drivers/gpu/drm/mediatek/mtk_drm_drv.c
> > @@ -351,6 +351,7 @@ static bool mtk_drm_get_all_drm_priv(struct
> > device *dev)
> >   {
> >   	struct mtk_drm_private *drm_priv = dev_get_drvdata(dev);
> >   	struct mtk_drm_private *all_drm_priv[MAX_CRTC];
> > +	struct mtk_drm_private *temp_drm_priv;
> >   	struct device_node *phandle = dev->parent->of_node;
> >   	const struct of_device_id *of_id;
> >   	struct device_node *node;
> > @@ -373,9 +374,18 @@ static bool mtk_drm_get_all_drm_priv(struct
> > device *dev)
> >   		if (!drm_dev || !dev_get_drvdata(drm_dev))
> >   			continue;
> >   
> > -		all_drm_priv[cnt] = dev_get_drvdata(drm_dev);
> > -		if (all_drm_priv[cnt] && all_drm_priv[cnt]-
> > >mtk_drm_bound)
> > -			cnt++;
> > +		temp_drm_priv = dev_get_drvdata(drm_dev);
> > +		if (temp_drm_priv) {
> > +			if (temp_drm_priv->mtk_drm_bound)
> > +				cnt++;
> > +
> > +			if (temp_drm_priv->data->main_len)
> > +				all_drm_priv[0] = temp_drm_priv;
> > +			else if (temp_drm_priv->data->ext_len)
> > +				all_drm_priv[1] = temp_drm_priv;
> > +			else if (temp_drm_priv->data->third_len)
> > +				all_drm_priv[2] = temp_drm_priv;
> > +		}
> 
> Previously the code was assigning stuff into all_drm_priv[cnt] and 
> incrementing it.
> With your change, it assigns to all_drm_priv[0], [1], [2]. Is this
> what 
> you intended ?

Because dev_get_drvdata(drm_dev) will get the driver data by drm_dev.
Each drm_dev represents a display path.
e,g.
drm_dev of "mediatek,mt8195-vdosys0" represents main path.
drm_dev of "mediatek,mt8195-vdosys1" represents ext path.

So we want to make sure all_drm_priv[] store the private data in
the order of display path, such as:
all_drm_priv[0] = the private data of main display
all_drm_priv[1] = the private data of ext display
all_drm_priv[2] = the private data of third display

> If this loop has second run, you will reassign to all_drm_priv again
> ?

Because the previous code will store all_drm_priv[] in the order of
mtk_drm_bind() was called.

If drm_dev of ext path bound earlier than drm_dev of main path,
all_drm_priv[] in mtk_drm_get_all_drm_priv() may be re-assigned like
this:
all_drm_priv[0]->all_drm_priv[0] = private data of ext path
all_drm_priv[1]->all_drm_priv[0] = private data of ext path
all_drm_priv[0]->all_drm_priv[1] = private data of main path
all_drm_priv[1]->all_drm_priv[1] = private data of main path

But we expect all_drm_priv[] be re-assigned like this:
all_drm_priv[0]->all_drm_priv[0] = private data of main path
all_drm_priv[1]->all_drm_priv[0] = private data of main path
all_drm_priv[0]->all_drm_priv[1] = private data of ext path
all_drm_priv[1]->all_drm_priv[1] = private data of ext path

> I would expect you to take `cnt` into account.
> Also, is it expected that all_drm_priv has holes in the array ?

Each drm_dev will only called mtk_drm_bind() once, so all holes
will be filled after all drm_dev has called mtk_drm_bind().

Do you agree with this statement? :)

Regards,
Jason-JH.Lin

> 
> Eugen
> 
> 
> 
> >   	}
> >   
> >   	if (drm_priv->data->mmsys_dev_num == cnt) {
> 
>