On Wed, 2002-04-10 at 19:47, Trent Piepho wrote:
> > video_tuner->signal = (BT848_DSTATUS & BT848_DSTATUS_HLOC) ? 0xFFFF : 0;
> Not very useful. You can't query the tuner unless you have it selected as the
> input. So if your card doesn't have a tuner, or you are using the composite
> or s-video input, this won't work.
Although ugly, I'm planning on using that property for the signal
presence on the zoran cards too, although these cards have no tuner
whatsoever. We might just use the convention that everyone uses that
property for that goal, tuner or no tuner. If there would be a signal
property somewhere else in the v4l2 API, I might use that, but there
isn't any (yet). Maybe we need to change v4l2's int on the
VIDIOC_[SG]_INPUT and create a struct v4l2_input?
struct v4l2_input {
int index;
__u8 signal; // or maybe __u32 or so...
char reserved[some_number];
}
The problem is that this causes binary incompatibility, so I should
probably have come up with this suggestion a few years ago ;-). I guess
everyone would complain about this and they'd basically be right...
Ronald
--
- .-.
- /V\ | Ronald Bultje <rbultje@xxxxxxxxxxxxxxxxxxx>
- // \\ | Running: Linux 2.4.18-XFS and OpenBSD 3.0
- /( )\ | http://ronald.bitfreak.net/
- ^^-^^
