Date:   Wed, 4 Mar 2020 21:47:15 +0100
From:   Niklas <niklas.soderlund@...natech.se>
To:     Lad Prabhakar <prabhakar.csengg@...il.com>
Cc:     Mauro Carvalho Chehab <mchehab@...nel.org>,
        linux-media@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-renesas-soc@...r.kernel.org,
        Lad Prabhakar <prabhakar.mahadev-lad.rj@...renesas.com>
Subject: Re: [PATCH 1/2] media: rcar-csi2: Fix registering camera endpoint to
 VIN

Hi Lad,

Thanks for your work.

On 2020-02-28 16:50:10 +0000, Lad Prabhakar wrote:
> CSI2 registers camera/sensor as v4l2 async sub device with fwnode is
> remote endpoint and the camera/sensor register itself as v4l2 sub device
> with fwnode is remote device as a result the match.fwnode should be
> fwnode_graph_get_remote_port_parent and not
> fwnode_graph_get_remote_endpoint.
> 
> This patch makes use of v4l2 helper function
> v4l2_async_notifier_add_fwnode_remote_subdev() which uses
> fwnode_graph_get_remote_port_parent as match.fwnode fixing the issue
> of registering camera endpoint to the driver.
> 
> Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@...renesas.com>

I'm afraid this is not the right solution. The rcar-csi2 driver uses the 
v4l2-async framework to do endpoint matching instead of node matching.  
This is needed as it needs to work with the adv748x driver which 
register it self in v4l2-async using endpoints instead of nodes. The 
reason for this is that from a single DT node it creates multiple 
subdevices, one for each endpoint IIRC.

Without this patch the two CSI-2 receivers on R-Car M3-n registers the 
two following 'paths' in v4l2 to be able to find the two subdevice CSI-2 
transmitters created by the ADV748x.

rcar-csi2 fea80000.csi2: '/soc/i2c@...d8000/video-receiver@...port@...ndpoint'
rcar-csi2 feaa0000.csi2: '/soc/i2c@...d8000/video-receiver@...port@...ndpoint'

With this patch applied it registers the following which can't be found 
as they are not present in the v4l2-async list of subdevices (as they 
are registerd as above).

rcar-csi2 fea80000.csi2: '/soc/i2c@...d8000/video-receiver@70'
rcar-csi2 feaa0000.csi2: '/soc/i2c@...d8000/video-receiver@70'
rcar-csi2: probe of feaa0000.csi2 failed with error -17

This patch may unlock your use-case as it's a known problem that 
endpoint and node matching do not mix. But it will break the already 
upstream use-case and for that reason, I'm really sorry about this.

Nacked-by: Niklas Söderlund <niklas.soderlund+renesas@...natech.se>

The real solution to this problem IMHO is to make all of v4l2-async 
operate using endpoint matching or possibly some kind of fallback to 
node matching if no endpoint can be found. Never the less some work is 
required in the v4l2-async core to sort out node and endpoint matching 
coexistence.

> ---
>  drivers/media/platform/rcar-vin/rcar-csi2.c | 14 ++++++--------
>  1 file changed, 6 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/media/platform/rcar-vin/rcar-csi2.c b/drivers/media/platform/rcar-vin/rcar-csi2.c
> index faa9fb23a2e9..5b04e4768eb1 100644
> --- a/drivers/media/platform/rcar-vin/rcar-csi2.c
> +++ b/drivers/media/platform/rcar-vin/rcar-csi2.c
> @@ -833,20 +833,18 @@ static int rcsi2_parse_dt(struct rcar_csi2 *priv)
>  		return ret;
>  	}
>  
> -	priv->asd.match.fwnode =
> -		fwnode_graph_get_remote_endpoint(of_fwnode_handle(ep));
> -	priv->asd.match_type = V4L2_ASYNC_MATCH_FWNODE;
> -
> -	of_node_put(ep);
> -
>  	v4l2_async_notifier_init(&priv->notifier);
>  
> -	ret = v4l2_async_notifier_add_subdev(&priv->notifier, &priv->asd);
> +	ret = v4l2_async_notifier_add_fwnode_remote_subdev(&priv->notifier,
> +							   of_fwnode_handle(ep),
> +							   &priv->asd);
>  	if (ret) {
> -		fwnode_handle_put(priv->asd.match.fwnode);
> +		of_node_put(ep);
>  		return ret;
>  	}
>  
> +	of_node_put(ep);
> +
>  	priv->notifier.ops = &rcar_csi2_notify_ops;
>  
>  	dev_dbg(priv->dev, "Found '%pOF'\n",
> -- 
> 2.20.1
> 

-- 
Regards,
Niklas Söderlund

Powered by blists - more mailing lists