[PATCH] media: v4l: Fix dereference before NULL check

Sakari Ailus sakari.ailus at iki.fi
Thu Aug 21 10:25:48 PDT 2025


Hi Chandra,

Please choose an appropriate subject on the next time, this is missing what
the patch is for -- see git log if in doubt.

On Mon, Aug 18, 2025 at 03:01:57PM +0530, Chandra Mohan Sundar wrote:
> In 'stm32_csi_start', 'csidev->s_subdev' is dereferenced directly while
> assigning a value to the 'src_pad'. However the same value is being
> checked against NULL at a later point of time indicating that there
> are chances that the value can be NULL.
> 
> Move the dereference after the NULL check.
> 
> Fixes: e7bad98c205d1 ("media: v4l: Convert the users of v4l2_get_link_freq to call it on a pad")

Fixes: requires Cc: stable. I've fixed these this time. See
Documentation/process/submitting-patches.rst .

> Signed-off-by: Chandra Mohan Sundar <chandramohan.explore at gmail.com>
> ---
>  drivers/media/platform/st/stm32/stm32-csi.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/media/platform/st/stm32/stm32-csi.c b/drivers/media/platform/st/stm32/stm32-csi.c
> index b69048144cc1..fd2b6dfbd44c 100644
> --- a/drivers/media/platform/st/stm32/stm32-csi.c
> +++ b/drivers/media/platform/st/stm32/stm32-csi.c
> @@ -443,8 +443,7 @@ static void stm32_csi_phy_reg_write(struct stm32_csi_dev *csidev,
>  static int stm32_csi_start(struct stm32_csi_dev *csidev,
>  			   struct v4l2_subdev_state *state)
>  {
> -	struct media_pad *src_pad =
> -		&csidev->s_subdev->entity.pads[csidev->s_subdev_pad_nb];
> +	struct media_pad *src_pad;
>  	const struct stm32_csi_mbps_phy_reg *phy_regs = NULL;
>  	struct v4l2_mbus_framefmt *sink_fmt;
>  	const struct stm32_csi_fmts *fmt;
> @@ -466,6 +465,7 @@ static int stm32_csi_start(struct stm32_csi_dev *csidev,
>  	if (!csidev->s_subdev)
>  		return -EIO;
>  
> +	src_pad = &csidev->s_subdev->entity.pads[csidev->s_subdev_pad_nb];
>  	link_freq = v4l2_get_link_freq(src_pad,
>  				       fmt->bpp, 2 * csidev->num_lanes);
>  	if (link_freq < 0)

-- 
Regards,

Sakari Ailus



More information about the linux-arm-kernel mailing list