linux-kernel - Re: [PATCH] media: i2c: ov5648: Plug runtime pm counter leak

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] [day] [month] [year] [list]

Message-ID: <20210521140037.GA3@valkosipuli.retiisi.eu>
Date:   Fri, 21 May 2021 17:00:38 +0300
From:   Sakari Ailus <sakari.ailus@....fi>
To:     Pratyush Yadav <p.yadav@...com>
Cc:     Mauro Carvalho Chehab <mchehab@...nel.org>,
        Sakari Ailus <sakari.ailus@...ux.intel.com>,
        Paul Kocialkowski <paul.kocialkowski@...tlin.com>,
        Dan Carpenter <dan.carpenter@...cle.com>,
        linux-media@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] media: i2c: ov5648: Plug runtime pm counter leak

Hi Pratyush,

Thanks for the patch.

On Tue, May 11, 2021 at 06:14:37PM +0530, Pratyush Yadav wrote:
> When the stream is being enabled, the runtime pm usage counter is
> incremented. Then if ov5648_sw_standby() fails, the function returns
> error without decrementing the counter, leaking it.
> 
> Signed-off-by: Pratyush Yadav <p.yadav@...com>
> ---
> 
> Hi,
> 
> I spotted this when converting OV5640 driver to use runtime PM using
> this driver as reference. I only have a very surface level understanding
> of runtime PM system as of now so please review with that in mind.
> 
> This patch is only compile-tested since I don't have the hardware with
> me.
> 
>  drivers/media/i2c/ov5648.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/media/i2c/ov5648.c b/drivers/media/i2c/ov5648.c
> index 3ecb4a3e8773..6aa2c950f505 100644
> --- a/drivers/media/i2c/ov5648.c
> +++ b/drivers/media/i2c/ov5648.c
> @@ -2143,8 +2143,12 @@ static int ov5648_s_stream(struct v4l2_subdev *subdev, int enable)
>  	ret = ov5648_sw_standby(sensor, !enable);
>  	mutex_unlock(&sensor->mutex);
> 
> -	if (ret)
> +	if (ret) {
> +		if (enable)
> +			pm_runtime_put(sensor->dev);

The function powers the sensor off at the end if streaming was to be
disabled. You could do:

	if (ret || !enable)
		pm_runtime_put(sensor->dev);

> +
>  		return ret;
> +	}
> 
>  	state->streaming = !!enable;

state->streaming is assigned when not holding the lock. Could you address
that at the same time?

-- 
Kind regards,

Sakari Ailus