Received-SPF: pass (google.com: domain of linux-kernel+bounces-117282-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
Date: Mon, 25 Mar 2024 15:44:12 +0100
From: Luca Ceresoli <luca.ceresoli@bootlin.com>
To: Ricardo Ribalda <ribalda@chromium.org>
Cc: Thierry Reding <thierry.reding@gmail.com>, Jonathan Hunter
 <jonathanh@nvidia.com>, Sowjanya Komatineni <skomatineni@nvidia.com>, Mauro
 Carvalho Chehab <mchehab@kernel.org>, Greg Kroah-Hartman
 <gregkh@linuxfoundation.org>, Hans Verkuil <hverkuil@xs4all.nl>,
 linux-media@vger.kernel.org, linux-tegra@vger.kernel.org,
 linux-staging@lists.linux.dev, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 3/3] media: dvbdev: Initialize sbuf
Message-ID: <20240325154412.0b768835@booty>
In-Reply-To: <20240325-gcc-arm-warnings-v2-3-47523cf5c8ca@chromium.org>
References: <20240325-gcc-arm-warnings-v2-0-47523cf5c8ca@chromium.org>
	<20240325-gcc-arm-warnings-v2-3-47523cf5c8ca@chromium.org>
Organization: Bootlin
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Hello Ricardo,

On Mon, 25 Mar 2024 14:19:55 +0000
Ricardo Ribalda <ribalda@chromium.org> wrote:

> Because the size passed to copy_from_user() cannot be known beforehand,
> it needs to be checked during runtime with check_object_size. That makes
> gcc believe that the content of sbuf can be used before init.
>=20
> Fix:
> ./include/linux/thread_info.h:215:17: warning: =E2=80=98sbuf=E2=80=99 may=
 be used uninitialized [-Wmaybe-uninitialized]
>=20
> Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> ---
>  drivers/media/dvb-core/dvbdev.c             | 2 +-
>  drivers/staging/media/tegra-video/tegra20.c | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
>=20
> diff --git a/drivers/media/dvb-core/dvbdev.c b/drivers/media/dvb-core/dvb=
dev.c
> index 733d0bc4b4cc3..b43695bc51e75 100644
> --- a/drivers/media/dvb-core/dvbdev.c
> +++ b/drivers/media/dvb-core/dvbdev.c
> @@ -956,7 +956,7 @@ int dvb_usercopy(struct file *file,
>  		 int (*func)(struct file *file,
>  			     unsigned int cmd, void *arg))
>  {
> -	char    sbuf[128];
> +	char    sbuf[128] =3D {};
>  	void    *mbuf =3D NULL;
>  	void    *parg =3D NULL;
>  	int     err  =3D -EINVAL;
> diff --git a/drivers/staging/media/tegra-video/tegra20.c b/drivers/stagin=
g/media/tegra-video/tegra20.c
> index c39b52d0e4447..630e2ff987a37 100644
> --- a/drivers/staging/media/tegra-video/tegra20.c
> +++ b/drivers/staging/media/tegra-video/tegra20.c
> @@ -164,6 +164,7 @@ static void tegra20_vi_get_input_formats(struct tegra=
_vi_channel *chan,
>  	unsigned int input_mbus_code =3D chan->fmtinfo->code;
> =20
>  	(*main_input_format) =3D VI_INPUT_INPUT_FORMAT_YUV422;
> +	(*yuv_input_format) =3D VI_INPUT_YUV_INPUT_FORMAT_UYVY;

I agree with Thierry this seems to be the best fix, making this
function similar to tegra20_vi_get_output_formats().

However in this v2 your change ended up in patch 3 while it should be
in patch 1. I'm glad to add my Reviewed-by tag on a new version with
this fixed.

Luca

--=20
Luca Ceresoli, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com