Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   "Hans Verkuil (hansverk)" <hansverk@cisco.com>
To:     Wen Yang <yellowriver2010@hotmail.com>,
        Hans Verkuil <hans.verkuil@cisco.com>,
        Mauro Carvalho Chehab <mchehab@kernel.org>
CC:     Linux Media Mailing List <linux-media@vger.kernel.org>,
        LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 1/4] media: cec-notifier: fix possible object reference
 leak
Thread-Topic: [PATCH 1/4] media: cec-notifier: fix possible object reference
 leak
Thread-Index: AQHUwCH951ryS7AxYU2IkMLJINPqbQ==
Date:   Mon, 11 Feb 2019 10:57:31 +0000
Message-ID: <6b7fc837599e4069b876b32473b31746@XCH-ALN-012.cisco.com>
References: <HK0PR02MB363461179B3A702DB9CAEC55B26A0@HK0PR02MB3634.apcprd02.prod.outlook.com>
 <3ed515b78a404ec4a22b5f69ed9d6e28@XCH-ALN-012.cisco.com>
Accept-Language: en-US
Content-Language: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 11/02/2019 11:38, Hans Verkuil (hansverk) wrote:=0A=
> On 09/02/2019 03:48, Wen Yang wrote:=0A=
>> put_device() should be called in cec_notifier_release(),=0A=
>> since the dev is being passed down to cec_notifier_get_conn(),=0A=
>> which holds reference. On cec_notifier destruction, it=0A=
>> should drop the reference to the device.=0A=
>>=0A=
>> Fixes: 6917a7b77413 ("[media] media: add CEC notifier support")=0A=
>> Signed-off-by: Wen Yang <yellowriver2010@hotmail.com>=0A=
>> ---=0A=
>>  drivers/media/cec/cec-notifier.c | 1 +=0A=
>>  1 file changed, 1 insertion(+)=0A=
>>=0A=
>> diff --git a/drivers/media/cec/cec-notifier.c b/drivers/media/cec/cec-no=
tifier.c=0A=
>> index dd2078b..621d4ae 100644=0A=
>> --- a/drivers/media/cec/cec-notifier.c=0A=
>> +++ b/drivers/media/cec/cec-notifier.c=0A=
>> @@ -66,6 +66,7 @@ static void cec_notifier_release(struct kref *kref)=0A=
>>  		container_of(kref, struct cec_notifier, kref);=0A=
>>  =0A=
>>  	list_del(&n->head);=0A=
>> +	put_device(n->dev);=0A=
>>  	kfree(n->conn);=0A=
>>  	kfree(n);=0A=
>>  }=0A=
>>=0A=
> =0A=
> Sorry, no. The dev pointer is just a search key that the notifier code lo=
oks=0A=
> for. It is not the notifier's responsibility to take a reference, that wo=
uld=0A=
> be the responsibility of the hdmi and cec drivers.=0A=
=0A=
Correction: the cec driver should never take a reference of the hdmi device=
.=0A=
It never accesses the HDMI device, it only needs the HDMI device pointer as=
=0A=
a key in the notifier list.=0A=
=0A=
The real problem is that several CEC drivers take a reference of the HDMI d=
evice=0A=
and never release it. So those drivers need to be fixed.=0A=
=0A=
Regards,=0A=
=0A=
	Hans=0A=
=0A=
> =0A=
> If you can demonstrate that there is an object reference leak, then pleas=
e=0A=
> provide the details: it is likely a bug elsewhere and not in the notifier=
=0A=
> code.=0A=
> =0A=
> BTW, your patch series didn't arrive on the linux-media mailinglist for=
=0A=
> some reason.=0A=
> =0A=
> Regards,=0A=
> =0A=
> 	Hans=0A=
> =0A=
=0A=