Stefan Lucke wrote:
Hi,
after getting vdr-softdevice running with my radeon I switched back to my matrox-G550 and neither OSD nor video works with directfb. Startup messages showed nothing unusual.
I also have problems with softdevice 0.0.6 / directfb CVS / viafb / 2.4.26-epia2.
I am currently documenting these problems, along with full testing of softmepg.
Problems are in the area of :
* video is only seen "through" the OSD (when no OSD, no video, when OSD is open, I can see the vedio behind the text / drawing areas
* OSD is *monocolor*, as if it consisted only in transparency showing the video
* crashes (I still have to investigate where they occur exactly)
To exit I had to type blind "menu -> 6 -> left -> OK -> OK"
I type ^C on the SSH shell I run vdr from... Is it particularly bad ?
Xv-out and vidix-out work without recompile.
About Xv : it seems that there is a lot of Xv-specific code in the generic code, ie. a lot of #ifdef in softdevice.c
That leads to the fact that if I only use DFB output method, the code is not the same if I compile with ou without Xv support. Is the impact protentieally bad, ot these Xv improvements are in fact really generic ?
I currently compile with Xv support, but never tried to run it.
It would be nice if someone could give a hint.
The only hint I could give you is that you're not alone with this kind of problems...
Current code base is available via: http://www.lucke.in-berlin.de
file: softdevice_av_sync_test.tar.bz2
kernel: is 2.4.27-pre5, matroxfb_base loaded directfb: cvs from 2004-06-16 19:00
Stefan Lucke
-- NH
