Hi Laurent,
On 18/05/17 15:01, Laurent Pinchart wrote:
> Hi Kieran,
>
> Thank you for the patch.
>
> On Wednesday 17 May 2017 16:03:39 Kieran Bingham wrote:
>> From: Kieran Bingham <kieran.bingham+renesas@xxxxxxxxxxxxxxxx>
>>
>> Devices supporting multiple endpoints on a single device node must set
>> their subdevice fwnode to the endpoint to allow distinct comparisons.
>>
>> Adapt the match_fwnode call to compare against the provided fwnodes
>> first, but also to search for a comparison against the parent fwnode.
>>
>> This allows notifiers to pass the endpoint for comparison and still
>> support existing subdevices which store their default parent device
>> node.
>>
>> Signed-off-by: Kieran Bingham <kieran.bingham+renesas@xxxxxxxxxxxxxxxx>
>> ---
>> drivers/media/v4l2-core/v4l2-async.c | 20 ++++++++++++++++----
>> 1 file changed, 16 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/media/v4l2-core/v4l2-async.c
>> b/drivers/media/v4l2-core/v4l2-async.c index e1e181db90f7..65735a5c4350
>> 100644
>> --- a/drivers/media/v4l2-core/v4l2-async.c
>> +++ b/drivers/media/v4l2-core/v4l2-async.c
>> @@ -41,14 +41,26 @@ static bool match_devname(struct v4l2_subdev *sd,
>> return !strcmp(asd->match.device_name.name, dev_name(sd->dev));
>> }
>>
> /*
> * Check whether the two device_node pointers refer to the same OF node. We
> * can't compare pointers directly as they can differ if overlays have been
> * applied.
> */
Thanks - that's a good addition - I've put it in.
>
>> +static bool match_of(struct device_node *a, struct device_node *b)
>> +{
>> + return !of_node_cmp(of_node_full_name(a), of_node_full_name(b));
>> +}
>> +
>> static bool match_fwnode(struct v4l2_subdev *sd, struct v4l2_async_subdev
>> *asd)
>> {
>> + struct device_node *sdnode;
>> + struct fwnode_handle *async_device;
>
> I would name this asd_fwnode, and to be consistent rename sdnode to sd_ofnode.
Actually, now that I agree with Sakari, and the parent of both the SD and the
ASD should be cross-referenced, I have used:
sd_parent = fwnode_graph_get_port_parent(sd->fwnode);
asd_parent = fwnode_graph_get_port_parent(asd_fwnode);
>
>> +
>> + async_device = fwnode_graph_get_port_parent(asd->match.fwnode.fwnode);
>> +
>> if (!is_of_node(sd->fwnode) || !is_of_node(asd->match.fwnode.fwnode))
>> - return sd->fwnode == asd->match.fwnode.fwnode;
>> + return sd->fwnode == asd->match.fwnode.fwnode ||
>> + sd->fwnode == async_device;
>
> I wonder whether we could simplify this by changing the
> fwnode_graph_get_port_parent() API. At the moment the function walks two or
> three levels up depending on whether there's a ports name or not. If we turned
> in into a function that accepts an endpoint, port or device node, and returns
> the device node unconditionally (basically, returning the argument if its name
> is not "port(@[0-9]+)?" or "endpoint(@[0-9]+)?", and walking up until it
> reaches the device node otherwise), you could write the above
>
> asd_fwnode = fwnode_graph_get_port_parent(asd->match.fwnode.fwnode);
>
> if (!is_of_node(sd->fwnode) || !is_of_node(asd_fwnode))
> sd->fwnode == asd_fwnode;
>
> sdnode = to_of_node(sd->fwnode);
>
> return match_of(sdnode, to_of_node(asd_node));
I don't think that would help here. I want the function to do comparisons on the
endpoint when provided - I don't want helpers to suddenly bring the comparison
up to the device level.
>
>> +
>> + sdnode = to_of_node(sd->fwnode);
>>
>> - return !of_node_cmp(of_node_full_name(to_of_node(sd->fwnode)),
>> - of_node_full_name(
>> - to_of_node(asd->match.fwnode.fwnode)));
>> + return match_of(sdnode, to_of_node(asd->match.fwnode.fwnode)) ||
>> + match_of(sdnode, to_of_node(async_device));
>
> This is getting a bit complex, could you document the function ?
I've added comments, and improved helpers - I think it's looking a lot better now :)
>
>> }
>>
>> static bool match_custom(struct v4l2_subdev *sd, struct v4l2_async_subdev
>> *asd)
>
--
Kieran
