hverkuil: new request API RFC for you to look at. It should address your concerns, so I hope we can find some common ground on this one
I know there is still a lot missing, but since the previous attempts were dead on arrival I prefer to validate this first
we need to move forward with this on the Chrome OS side, and would like to do so with an upstream(able) base
anyway, looking forward to seeing your comments!
hverkuil: Huomenta!
sailus: hi!
hverkuil: How do you do?
busy with the cve crap
I'll try to review the set today.
I'd have a quick question on lens controls.
I've proposed a class for voice coil lenses; these are simple devices that still have a variety of ways how they're controled.
I'd add a control to tell whether a lens is still moving; apprently some controllers support that.
Would you add that control to a new class or use an existing class?
Bear in mind that there will presumably be more controls related to voice coil lenses.
Such as enabling or disabling ringing compensation; I haven't proposed one as no driver has supported it yet. The hardware support exists though.
I have no problem with a new control class for this. Would it be specific for voice coils or just 'lens control' in general? I think I prefer it to be just 'lens control'.
The voice coil lenses are way more simple than other devices.
The control is different, too.
And the V4L2 controls could be also mostly orthogonal between the two.
I'm leaning towards voice coil control class but I'm fine with just generic lens control class, too.
Perhaps actually some controls could be actually the same. Such as the one telling whether a lens is still moving.
I guess that answers the question actually.
:-)
I think a lens control class makes sense.
In the case of dw9807, you'll still need to ask the controller whether the lens is moving, so it'll be volatile read-only control.
For a decent step motor based controller, I'd presume the controller tells you when it's done.
Through an interrupt.
yes
Or perhaps I'm expecting too much sophistication.
I believe that's what happens. We have such hardware.
I'll go with a lens control class the.
then.
Thanks.
sailus: ping
neg: Pong.
sailus: when you find the time could you rebase your vc branch? There are currently some small conflicts with the media-tree. I'm preparing next version of the multiplexed streams for R-Car and I want to be able to describe your branch as a dependecy :-) Nothing urgent
neg: Ack.
I'll see if I could have that done over the week-end.
sailus: no enjoy your weekend no rush
Otherwise I'll forget about it and you'll have to ping me again about them. :-)