Hi Kieran,

On Thu, Feb 13, 2020 at 10:27:11AM +0000, Kieran Bingham wrote:
> On 13/02/2020 10:20, Jacopo Mondi wrote:
> > Hi Kieran,
> >
> > On Thu, Feb 13, 2020 at 10:07:18AM +0000, Kieran Bingham wrote:
> >> Hi Jacopo,
> >>
> >> On 13/02/2020 09:46, Jacopo Mondi wrote:
> >>> Hi Kieran,
> >>>   very nice thanks for handling this
> >>
> >> :-)
> >>
> >>> Just a few minors
> >>
> >> :-s hehe
> >>
> >
> > Turned out to be lengthier than expected :)
> >
> >>>
> >>> On Wed, Feb 12, 2020 at 05:37:26PM +0000, Kieran Bingham wrote:
> >>>> Move all the V4L2 Subdev Async registration so that it can only happen once
> >>>> we know we will not need to -EPROBE_DEFER...
> >>>>
> >>>> Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
> >>>> ---
> >>>>  drivers/media/i2c/max9286.c | 88 +++++++++++++++++++++++--------------
> >>>>  1 file changed, 55 insertions(+), 33 deletions(-)
> >>>>
> >>>> diff --git a/drivers/media/i2c/max9286.c b/drivers/media/i2c/max9286.c
> >>>> index 1b4ff3533795..03c5fa232b6d 100644
> >>>> --- a/drivers/media/i2c/max9286.c
> >>>> +++ b/drivers/media/i2c/max9286.c
> >>>> @@ -503,6 +503,49 @@ static const struct v4l2_async_notifier_operations max9286_notify_ops = {
> >>>>  	.unbind = max9286_notify_unbind,
> >>>>  };
> >>>>
> >>>> +static int max9286_v4l2_async_register(struct max9286_priv *priv)
> >>>
> >>> Could you capture in the function name this actually deals with
> >>> notifiers ? Like max9286_notifier_register() or similar...
> >>
> >> I'd like to keep the 'v4l2' in there somewhere...
> >>
> >> 	max9286_v4l2_notifier_register() ?
> >>
> >> But then maybe even that could be confused with the notifiers/async
> >> handling for the max9286 itself.
> >>
> >> My aim was that max9286_v4l2_async_{un,}register() dealt with subdevices
> >> connected to the max9286 only ...
> >
> > To me async_register() calls for dealing with registering our own
> > subdev to the async framework, not collecting remote asds and adding
> > it our subnotifier. As you wish, it's really just a suggestion.
>
>
> So would you like to see the async registration between the max9286 and
> the connected CSI2 receiver handled in this pair of functions too?
>

No no, I was just bikeshedding on function names, no worries :)