On Saturday 29 June 2002 14:38, Ville Syrj�l� wrote:
> On Sat, Jun 29, 2002 at 09:43:14AM +0200, Mike Pieper wrote:
> > On Thursday 27 June 2002 22:05, Denis Oliver Kropp wrote:
> > > ...
> > > Someone could add the missing bits by using Mike's code.
> >
> > I'll make a new patch. The existing patch with an extra file which is
> > mostly a copy of the existing v4l-provider wasn't such a good idea,
> > because it's difficult to maintain. The new one will patch directly the
> > existing file.
>
> Did you see my latest patch (v4l_fix.patch)? It fixed the field sync
> stuff and cleaned up the code. Did anyone see the patch? I haven't
> received any comments on it nor has it appeared in cvs.
I didn't had the time to look at it, but I will ...

> > I also have an idea how to tell the provider what it should do
> > (grabbing/overlay) at runtime. The idea is, to add a string at the end of
> > the filename to give the driver some hints. Like this:
> > /dev/video0:overlay  or /dev/video0:grab
> > What do you think about that?
>
> I'm not sure what benefit that would have. When AGP surfaces become
> supported it might be useful depending on how the surface allocation is
> done.
The benefit is that it gives you the chance to control what the driver will 
use: grabbing or overlay (dma). The overlay solution needs less CPU. The grab 
solution uses the CPU much more, but delivers real sync to the input, what is 
important if you need a high quality picture.

The best would be of course to have overlay combined with synchronisation, but 
v4l doesn't offer that.

Mike


--
Info:  To unsubscribe send a mail to [EMAIL PROTECTED] with
"unsubscribe directfb-dev" as subject.

Reply via email to