Re: [PATCH] media: mtk-vcodec: Fix reference count leak in mtk_vcodec_fw_scp_init

[Nicolas Dufresne]

Hi,

Le mardi 18 mars 2025 à 19:05 +0800, Miaoqian Lin a écrit :
> scp_get() returns a reference that needs to be released with scp_put().
> Add missing scp_put() before returning error in mtk_vcodec_fw_scp_init().
> 
> Fixes: 53dbe0850444 ("media: mtk-vcodec: potential null pointer deference in SCP")
> Signed-off-by: Miaoqian Lin <linmq006@xxxxxxxxx>
> ---
>  .../platform/mediatek/vcodec/common/mtk_vcodec_fw_scp.c      | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw_scp.c b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw_scp.c
> index ff23b225db70..11ab3bc60217 100644
> --- a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw_scp.c
> +++ b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw_scp.c
> @@ -71,7 +71,6 @@ struct mtk_vcodec_fw *mtk_vcodec_fw_scp_init(void *priv, enum mtk_vcodec_fw_use
>  		pr_err("Invalid fw_use %d (use a reasonable fw id here)\n", fw_use);
>  		return ERR_PTR(-EINVAL);
>  	}
> -

Might sound nit-picky, but don't do style fixes in patches intended for
backports. It increases the chance of conflicts.

>  	scp = scp_get(plat_dev);
>  	if (!scp) {
>  		dev_err(&plat_dev->dev, "could not get vdec scp handle");
> @@ -79,8 +78,10 @@ struct mtk_vcodec_fw *mtk_vcodec_fw_scp_init(void *priv, enum mtk_vcodec_fw_use
>  	}
>  
>  	fw = devm_kzalloc(&plat_dev->dev, sizeof(*fw), GFP_KERNEL);
> -	if (!fw)
> +	if (!fw) {
> +		scp_put(scp);
>  		return ERR_PTR(-ENOMEM);
> +	}

With the above style change removed:

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

>  	fw->type = SCP;
>  	fw->ops = &mtk_vcodec_rproc_msg;
>  	fw->scp = scp;