linux-kernel - Re: [PATCH 01/13] media: v4l2-ctrls-api: fix error handling for v4l2_g

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: <dcb89d4c-72e0-42a3-8923-51cd18e05968@xs4all.nl>
Date: Wed, 16 Oct 2024 12:56:02 +0200
From: Hans Verkuil <hverkuil@...all.nl>
To: Mauro Carvalho Chehab <mchehab+huawei@...nel.org>
Cc: Ricardo Ribalda <ribalda@...omium.org>, linux-kernel@...r.kernel.org,
 linux-media@...r.kernel.org, stable@...r.kernel.org
Subject: Re: [PATCH 01/13] media: v4l2-ctrls-api: fix error handling for
 v4l2_g_ctrl()

On 16/10/2024 12:22, Mauro Carvalho Chehab wrote:
> The error check logic at get_ctrl() is broken: if ptr_to_user()
> fails to fill a control due to an error, no errors are returned
> and v4l2_g_ctrl() returns success on a failed operation, which
> may cause applications to fail.
> 
> Add an error check at get_ctrl() and ensure that it will
> be returned to userspace without filling the control value if
> get_ctrl() fails.
> 
> Fixes: 71c689dc2e73 ("media: v4l2-ctrls: split up into four source files")
> Cc: stable@...r.kernel.org
> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@...nel.org>
> ---
>  drivers/media/v4l2-core/v4l2-ctrls-api.c | 10 +++++++---
>  1 file changed, 7 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/media/v4l2-core/v4l2-ctrls-api.c b/drivers/media/v4l2-core/v4l2-ctrls-api.c
> index e5a364efd5e6..be32dccf9830 100644
> --- a/drivers/media/v4l2-core/v4l2-ctrls-api.c
> +++ b/drivers/media/v4l2-core/v4l2-ctrls-api.c
> @@ -753,9 +753,10 @@ static int get_ctrl(struct v4l2_ctrl *ctrl, struct v4l2_ext_control *c)
>  		for (i = 0; i < master->ncontrols; i++)
>  			cur_to_new(master->cluster[i]);
>  		ret = call_op(master, g_volatile_ctrl);
> -		new_to_user(c, ctrl);
> +		if (!ret)
> +			ret = new_to_user(c, ctrl);
>  	} else {
> -		cur_to_user(c, ctrl);
> +		ret = cur_to_user(c, ctrl);
>  	}
>  	v4l2_ctrl_unlock(master);
>  	return ret;
> @@ -770,7 +771,10 @@ int v4l2_g_ctrl(struct v4l2_ctrl_handler *hdl, struct v4l2_control *control)
>  	if (!ctrl || !ctrl->is_int)
>  		return -EINVAL;
>  	ret = get_ctrl(ctrl, &c);
> -	control->value = c.value;
> +
> +	if (!ret)
> +		control->value = c.value;
> +
>  	return ret;
>  }
>  EXPORT_SYMBOL(v4l2_g_ctrl);

Yeah, that's better.

There are also unchecked calls to cur_to_user() and user_to_new() in set_ctrl_lock().

Can you fix that as well?

Regards,

	Hans