Le vendredi 31 juillet 2026 à 14:07 +0000, Tharit Tangkijwanichakul a écrit :
> The MPEG-2 sequence and picture controls are validated
> before a job is queued, so hantro_get_ctrl() is not expected to return
> NULL in hantro_g1_mpeg2_dec_run(). The controls are nonetheless
> dereferenced unconditionally.
> 
> Harden the invariant by checking the sequence and picture controls with
> WARN_ON().
> 
> Found by code inspection.
> 
> Fixes: f329e21e9dad ("media: uapi: mpeg2: Split sequence and picture parameters")
> Signed-off-by: Tharit Tangkijwanichakul <tharitt97@gmail.com>

Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>

p.s. this is a bit cosmetic, since only a programming error could lead to that,
once a control is registered, it has a value. But that makes it consistent with
all other use of hantro_get_ctrl() in this driver, which I like.

Nicolas

> ---
> Tested on a Rockchip RK3588 (Rock 5B) board with Fluster:
>   MPEG-2 (MPEG2_VIDEO-MAIN): 23/43, unchanged
> 
>  drivers/media/platform/verisilicon/hantro_g1_mpeg2_dec.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/drivers/media/platform/verisilicon/hantro_g1_mpeg2_dec.c b/drivers/media/platform/verisilicon/hantro_g1_mpeg2_dec.c
> index e0d6bd0a6e44..6a942f4b8a2d 100644
> --- a/drivers/media/platform/verisilicon/hantro_g1_mpeg2_dec.c
> +++ b/drivers/media/platform/verisilicon/hantro_g1_mpeg2_dec.c
> @@ -161,8 +161,13 @@ int hantro_g1_mpeg2_dec_run(struct hantro_ctx *ctx)
>  
>  	seq = hantro_get_ctrl(ctx,
>  			      V4L2_CID_STATELESS_MPEG2_SEQUENCE);
> +	if (WARN_ON(!seq))
> +		return -EINVAL;
> +
>  	pic = hantro_get_ctrl(ctx,
>  			      V4L2_CID_STATELESS_MPEG2_PICTURE);
> +	if (WARN_ON(!pic))
> +		return -EINVAL;
>  
>  	reg = G1_REG_DEC_AXI_RD_ID(0) |
>  	      G1_REG_DEC_TIMEOUT_E(1) |