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;
Subject: Re: [PATCH v3 7/7] drm/i2c: tda998x: register as a drm bridge
To:     Russell King - ARM Linux <linux@armlinux.org.uk>,
        Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc:     dri-devel@lists.freedesktop.org,
        Mark Rutland <mark.rutland@arm.com>,
        Boris Brezillon <boris.brezillon@free-electrons.com>,
        Alexandre Belloni <alexandre.belloni@bootlin.com>,
        devicetree@vger.kernel.org, David Airlie <airlied@linux.ie>,
        linux-kernel@vger.kernel.org, Rob Herring <robh+dt@kernel.org>,
        Jacopo Mondi <jacopo+renesas@jmondi.org>,
        Daniel Vetter <daniel.vetter@intel.com>,
        linux-arm-kernel@lists.infradead.org
References: <20180419162751.25223-1-peda@axentia.se>
 <20180419162751.25223-8-peda@axentia.se> <2556566.3fxMPoIyx0@avalon>
 <20180420102408.GW16141@n2100.armlinux.org.uk>
From:   Peter Rosin <peda@axentia.se>
Organization: Axentia Technologies AB
Message-ID: <790d7a89-ae13-098f-3c0c-a8dfbc84e4a7@axentia.se>
Date:   Fri, 20 Apr 2018 15:28:41 +0200
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:52.0) Gecko/20100101
 Thunderbird/52.7.0
MIME-Version: 1.0
In-Reply-To: <20180420102408.GW16141@n2100.armlinux.org.uk>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit
SpamDiagnosticOutput: 1:99
SpamDiagnosticMetadata: NSPM
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 20 Apr 2018 13:28:46.2015 (UTC)
X-MS-Exchange-CrossTenant-Network-Message-Id: 143b8563-5799-449a-a2bd-08d5a6c2a59e
X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted
X-MS-Exchange-CrossTenant-Id: 4ee68585-03e1-4785-942a-df9c1871a234
X-MS-Exchange-Transport-CrossTenantHeadersStamped: VI1PR0202MB2783
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On 2018-04-20 12:24, Russell King - ARM Linux wrote:
> On Fri, Apr 20, 2018 at 01:06:49PM +0300, Laurent Pinchart wrote:
>> Hi Peter,
>>
>> Thank you for the patch.
>>
>> On Thursday, 19 April 2018 19:27:51 EEST Peter Rosin wrote:
>>> This makes this driver work with all(?) drivers that are not
>>> componentized and instead expect to connect to a panel/bridge. That
>>> said, the only one tested is atmel_hlcdc.
>>>
>>> This hooks the relevant work function previously called by the encoder
>>> and the component also to the bridge, since the encoder goes away when
>>> connecting to the bridge interface of the driver and the equivalent of
>>> bind/unbind of the component is handled by bridge attach/detach.
>>>
>>> The lifetime requirements of a bridge and a component are slightly
>>> different, which is the reason for struct tda998x_bridge.
>>
>> Couldn't you move the allocation and initialization (tda998x_create) of the 
>> tda998x_priv structure to probe time ? I think you wouldn't need a separate 
>> structure in that case. Unless I'm mistaken there would be an added benefit of 
>> separating component and bridge initialization, resulting in the encoder not 
>> being initialized at all if the component isn't used. You wouldn't need to add 
>> a local_encoder parameter to the tda998x_init() function.
> 
> No, I don't like that idea one bit, as I've stated in the past about the
> component API.  The same (probably) goes for the bridge stuff too.
> 
> Consider the following:
> 
> Your DRM system is initialised.  You then remove a module, which results
> in the DRM system being torn down.  You re-insert the module (eg, having
> made a change to it).  The DRM system is then re-initialised.
> 
> At this point, what is the state of variables such as priv->is_on if
> you allocate the structure at probe time?
> 
> What about all the other variables in the driver private structure - are
> you sure that the driver can cope with random values from the previous
> "usage" remaining there?
> 
> At the moment, this isn't a concern for the driver because we
> dev_kzalloc() the structure in the bind callback.  Move that to the
> probe function, and the structure is no longer re-initialised each
> time, and so it retains the previous state.  The driver is not setup
> to cope with that.
> 
> So, to work around that, you would need to reinitialise _everything_
> in the structure that the driver requires, which IMHO is a very
> open to bugs (eg, if a member is missed, or added without the
> necessary re-initialisation), _especially_ when this is not a path
> that will get regular testing.
> 
> If you want to do this for a subset of data, it would be much better
> to separate them into independent structures (maybe one embedded into
> the other) so that this problem can not occur.  That way, a subset
> of the data can be memset() when bound to the rest of the DRM system
> ensuring a consistent driver state and still achieve what you're
> suggesting.

This was the exact reason I added struct tda998x_bridge. It seemed
very risky to move the tda998x_create call (or some of its meat) to
the probe function. Even if that could be done I think it should
definitely be a separate patch.

Cheers,
Peter