linux-kernel - Re: [PATCH v2] media: uapi: v4l: Change V4L2_TYPE_IS

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: Thu, 30 May 2024 12:31:59 +0200
From: Hans Verkuil <hverkuil@...all.nl>
To: Nas Chung <nas.chung@...psnmedia.com>, mchehab@...nel.org,
 linux-media@...r.kernel.org, sebastian.fricke@...labora.com,
 m.tretter@...gutronix.de
Cc: linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] media: uapi: v4l: Change V4L2_TYPE_IS_CAPTURE
 condition

On 28/05/2024 04:04, Nas Chung wrote:
> Explicitly compare a buffer type only with valid buffer types,
> to avoid matching the buffer type outside of valid buffer
> type set.
> 
> Signed-off-by: Nas Chung <nas.chung@...psnmedia.com>
> ---
>  include/uapi/linux/videodev2.h | 7 ++++++-
>  1 file changed, 6 insertions(+), 1 deletion(-)
> 
> diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h
> index fe6b67e83751..fa2b7086e480 100644
> --- a/include/uapi/linux/videodev2.h
> +++ b/include/uapi/linux/videodev2.h
> @@ -157,6 +157,10 @@ enum v4l2_buf_type {
>  	V4L2_BUF_TYPE_PRIVATE              = 0x80,
>  };
>  
> +#define V4L2_TYPE_IS_VALID(type)		\
> +	((type) >= V4L2_BUF_TYPE_VIDEO_CAPTURE	\
> +	 && (type) <= V4L2_BUF_TYPE_META_OUTPUT)
> +
>  #define V4L2_TYPE_IS_MULTIPLANAR(type)			\
>  	((type) == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE	\
>  	 || (type) == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE)
> @@ -171,7 +175,8 @@ enum v4l2_buf_type {
>  	 || (type) == V4L2_BUF_TYPE_SDR_OUTPUT			\
>  	 || (type) == V4L2_BUF_TYPE_META_OUTPUT)
>  
> -#define V4L2_TYPE_IS_CAPTURE(type) (!V4L2_TYPE_IS_OUTPUT(type))
> +#define V4L2_TYPE_IS_CAPTURE(type)	\
> +	(V4L2_TYPE_IS_VALID(type) && !V4L2_TYPE_IS_OUTPUT(type))
>  
>  enum v4l2_tuner_type {
>  	V4L2_TUNER_RADIO	     = 1,

This patch introduced this warning:

In function 'find_format_by_index',
    inlined from 'vdec_enum_fmt' at drivers/media/platform/qcom/venus/vdec.c:452:8:
drivers/media/platform/qcom/venus/vdec.c:172:32: warning: 'valid' may be used uninitialized [-Wmaybe-uninitialized]
  172 |                 if (k == index && valid)
      |                     ~~~~~~~~~~~^~~~~~~~
drivers/media/platform/qcom/venus/vdec.c: In function 'vdec_enum_fmt':
drivers/media/platform/qcom/venus/vdec.c:157:22: note: 'valid' was declared here
  157 |                 bool valid;
      |                      ^~~~~

This driver does this:

	bool valid;

        if (V4L2_TYPE_IS_OUTPUT(type)) {
                valid = venus_helper_check_codec(inst, fmt[i].pixfmt);
        } else if (V4L2_TYPE_IS_CAPTURE(type)) {
                valid = venus_helper_check_format(inst, fmt[i].pixfmt);

With your patch both V4L2_TYPE_IS_OUTPUT(type) and V4L2_TYPE_IS_CAPTURE(type)
can return false, something that wasn't possible without this patch.

In this driver the fix would just be to drop the second 'if' altogether, so just
'} else {'.

Since these defines are part of the public API, this change introduces a subtle
behavior difference that can affect applications.

That said, I do consider this an improvement.

I would like some changes, though:

1) Just after "V4L2_BUF_TYPE_META_OUTPUT          = 14," in enum v4l2_buf_type,
   add a comment saying that V4L2_TYPE_IS_VALID and (for output types)
   V4L2_TYPE_IS_OUTPUT must be updated if a new type is added. It's all
   too easy to forget that otherwise.
2) Add a patch fixing the venus/vdec.c code.
3) Something else I noticed, but I think this change should be in a separate patch:
   V4L2_TYPE_IS_OUTPUT() returns true for V4L2_BUF_TYPE_VIDEO_OVERLAY, but that
   definitely belongs to CAPTURE. Nobody really uses that type anymore, but still,
   it should be fixed.

Regards,

	Hans