linux-kernel - Re: [PATCH v2] media: mediatek: encoder: Fix uninitialized scalar variable issue

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: <c751e015c0a9fb2ab6514a45952e01424cfbb0cb.camel@collabora.com>
Date: Fri, 29 Aug 2025 15:10:02 -0400
From: Nicolas Dufresne <nicolas.dufresne@...labora.com>
To: Irui Wang <irui.wang@...iatek.com>, Hans Verkuil
 <hverkuil-cisco@...all.nl>,  Mauro Carvalho Chehab	 <mchehab@...nel.org>,
 Matthias Brugger <matthias.bgg@...il.com>, 
	angelogioacchino.delregno@...labora.com, Qianfeng Rong
 <rongqianfeng@...o.com>
Cc: Project_Global_Chrome_Upstream_Group@...iatek.com, 
	linux-media@...r.kernel.org, linux-kernel@...r.kernel.org, 
	linux-arm-kernel@...ts.infradead.org, linux-mediatek@...ts.infradead.org, 
 Yunfei Dong <yunfei.dong@...iatek.com>, Longfei Wang
 <longfei.wang@...iatek.com>
Subject: Re: [PATCH v2] media: mediatek: encoder: Fix uninitialized scalar
 variable issue

Le mercredi 16 juillet 2025 à 15:14 +0800, Irui Wang a écrit :
> UNINIT checker finds some instances of variables that are used
> without being initialized, for example using the uninitialized
> value enc_result.is_key_frm can result in unpredictable behavior,
> so initialize these variables after declaring.
> 
> Fixes: 4e855a6efa54 ("[media] vcodec: mediatek: Add Mediatek V4L2 Video
> Encoder Driver")
> 
> Signed-off-by: Irui Wang <irui.wang@...iatek.com>
> ---
> v2:
>  - Add Fixes tag, update commit message
>  - Remove unnecessary memset
>  - Move memset to before the first usage
> ---
>  .../media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc.c   | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc.c
> b/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc.c
> index a01dc25a7699..3065f3e66336 100644
> --- a/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc.c
> +++ b/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc.c
> @@ -865,7 +865,7 @@ static void vb2ops_venc_buf_queue(struct vb2_buffer *vb)
>  static int vb2ops_venc_start_streaming(struct vb2_queue *q, unsigned int
> count)
>  {
>  	struct mtk_vcodec_enc_ctx *ctx = vb2_get_drv_priv(q);
> -	struct venc_enc_param param;
> +	struct venc_enc_param param = { 0 };
>  	int ret;
>  	int i;
>  
> @@ -1036,6 +1036,7 @@ static int mtk_venc_encode_header(void *priv)
>  			  ctx->id, dst_buf->vb2_buf.index, bs_buf.va,
>  			  (u64)bs_buf.dma_addr, bs_buf.size);
>  
> +	memset(&enc_result, 0, sizeof(enc_result));

Please, apply review comment to all occurrence, so same here.

>  	ret = venc_if_encode(ctx,
>  			VENC_START_OPT_ENCODE_SEQUENCE_HEADER,
>  			NULL, &bs_buf, &enc_result);
> @@ -1185,6 +1186,7 @@ static void mtk_venc_worker(struct work_struct *work)
>  			  (u64)frm_buf.fb_addr[1].dma_addr,
> frm_buf.fb_addr[1].size,
>  			  (u64)frm_buf.fb_addr[2].dma_addr,
> frm_buf.fb_addr[2].size);
>  
> +	memset(&enc_result, 0, sizeof(enc_result));

Same here.

>  	ret = venc_if_encode(ctx, VENC_START_OPT_ENCODE_FRAME,
>  			     &frm_buf, &bs_buf, &enc_result);
>  
> 


Would be nice to coordinate with Qianfeng Rong <rongqianfeng@...o.com> [0], he
ported the entire driver to this initialization method, which is clearly the way
to go.

- Patch 1 will port the driver to {} stack init
- Patch 2 will add missing initializes

Consistency is key for this type of things since developer usually follow the
surrounding style.

regards
Nicolas

[0] https://patchwork.linuxtv.org/project/linux-media/patch/20250803135514.118892-1-rongqianfeng@vivo.com/

Download attachment "signature.asc" of type "application/pgp-signature" (229 bytes)