[MPlayer-dev-eng] [PATCH] Improved DGA vertical retrace sync

[Stephen Davies]

On Mon, 23 Sep 2002, Arpi wrote:

> most vga cards can trigger an IRQ at vsync, it should be possible to develop
> a kernel driver which can catch that. Zdenek Kabelac (avifile author) worked
> on such thing, and also made a draft of /dev/vbi for Linux, but afair it was
> refused. it could work similar way to /dev/rtc, you can read() and it will
> return only at vsync, without burning the cpu.
> afaik he implemented it in avifile's version of mga_vid, but only for matrox
> cards.

Hi A'rpi, others,

This is a topic that interests me.  I'd like to get vsync handling working
for me Voodoo3 using the vo_xv.  It seems like no-chance to get this
handled int the X driver itself.

Seeing that there are _so_ many ways to get video onto the screen it
seems to me that something separate from the drivers themselves might be a
good way.

That way to start with user apps like mplayer can use it to synchronize
page flips to the physical retrace interrupt whatever technique is used to
get frames on the screen.

Perhaps later the driver writers may want to use this interface to get
VSync info.

This topic gets discussed around the place off and on but in the end it
seems to end without consensus.

I guess the usual way out of that is for someone to write some code.  Even
if they do a bad job it may prompt someone more able to do a good job!

I looks like Zdenek already did some code last year?

(I can find this function in the avifile-0.6 CVS in mga_vid driver itself
but not as a separate driver?)

mplayer's version of mga_vid driver doesn't have this "read waits for
vsync" function and instead returns some config stuff on read().

So there is some drifting apart of the two versions?

Regards,
Steve