HD4770: Screen corruption and artefacts with the from-git graphical stack

[MùPùF]

Hi,

I have had corruption problems with the open source driver for a few days now and it doesn't seem to go away.

Here is a video of the current state on my computer:
http://mupuf.org/blog/article/10/

Do you guys have any idea of what's going on ?

Thanks a lot

[bridgman]

Are you running openarena with LIBGL_ALWAYS_INDIRECT=yes ? It will look pretty bad if you don't right now...

[agd5f]

Also, EXA DFS seems to have issues on some rv740s.
Option "EXANoDownloadFromScreen"
should fix any issues with the cursor or small pixmaps.

[MùPùF]

Thanks to both of you! It works quite well.
There are still some minor corruption problems when typing text for example but I can live with it

[Perry3D]

Thanks to both of you! It works quite well.
There are still some minor corruption problems when typing text for example but I can live with it

I have the same issues with compiz enabled.


I hope this is a known bug and not Arch related, since I am using the same packages as MùPùF.

[MùPùF]

Perry3D: Did you activate Option "EXANoDownloadFromScreen" ?
It solved my problems and I no longer have this kind of corruption.

[Perry3D]

Yes, i already tried this option: no difference.
The corruptions appears when typing text, opening the gnome menu and some progress bars:

[legume]

I have the same issues with compiz enabled.

I hope this is a known bug and not Arch related, since I am using the same packages as MùPùF.

Yes, this is a known issue.

[pingufunkybeat]

Speaking of corruption, after upgrading to KDE 4.3, all the XRender-related artefacts disappeared and compositing in XRender mode is working great (though seemingly a bit slower than before).

OpenGL compositing starts, but hangs after a while.

[Veerappan]

I'd just like to chime in and say that EXANoDownloadFromScreen fixed the pointer corruption issues I was having in Gnome as well (no compositing, Ubuntu 9.04 w/ git drm/mesa/xorg-driver).

My card is a Radeon HD 4770 on a Gigabyte 785G-based mATX motherboard.

I'm still getting glxgears corruption with both normal GL, and with LIBGB_ALWAYS_INDIRECT=1, but I won't be able to do much troubleshooting with that until after work probably.