Moi,
On Tue, Mar 21, 2023 at 11:25:04AM +0200, Tomi Valkeinen wrote:
> On 21/03/2023 10:43, Hans Verkuil wrote:
> > > @@ -250,6 +268,9 @@ struct v4l2_subdev_routing {
> > > #define VIDIOC_SUBDEV_S_SELECTION _IOWR('V', 62, struct v4l2_subdev_selection)
> > > #define VIDIOC_SUBDEV_G_ROUTING _IOWR('V', 38, struct v4l2_subdev_routing)
> > > #define VIDIOC_SUBDEV_S_ROUTING _IOWR('V', 39, struct v4l2_subdev_routing)
> > > +#define VIDIOC_SUBDEV_G_CLIENT_CAP _IOR('V', 101, struct v4l2_subdev_client_capability)
> > > +#define VIDIOC_SUBDEV_S_CLIENT_CAP _IOWR('V', 102, struct v4l2_subdev_client_capability)
> >
> > Can't we just use __u64 as the ioctl type? Do we need to wrap it inside a struct?
>
> We could, but then we need to pass a pointer to the u64 so that the kernel
> can modify it (for VIDIOC_SUBDEV_S_CLIENT_CAP). I personally find it more
> clear from the perspective of a user to use a struct here, instead of
> passing a generic u64 pointer. Also, all the other ioctls use a struct.
>
> So, I'm fine either way, but personally I like this one.
I prefer it the way it is.
--
Sakari Ailus
