Created attachment 24293 [details] Desktop background corruption after moving a window Hi, Using the radeonhd driver with EXA acceleration enabled on my Radeon HD 3870 (RV670) card, I have desktop background corruption when I move windows in KDE 4.2 with desktop effects disabled. The corruption does not always appear immediately when starting moving a window, but shaking the window will trigger the corruption for sure. Once the background is corrupted, a refresh (for example by clicking on it) will restore it. With EXA acceleration disabled, the problem disappears. The radeon driver does not have this issue with EXA acceleration enabled. I have tested latest radeon and radeonhd drivers from their git master branch, with the drm modules from the git r6xx-r7xx-support branch.
Created attachment 24294 [details] Same desktop background area without corruption
I can second this bug and reported it already on ML and @freenode in #radeonhd channel. I made a video showing the corruption: http://www.nightwulf.org/radeonhd.avi The corruptions aren't that massive, when using xrender with kwin but they don't disappear completly. Greets, Torsten
Do you have similar problems with radeon (xf86-video-ati) as well? If not that might help isolate the problem.
Created attachment 24687 [details] Xorg log file
(In reply to comment #4) > Created an attachment (id=24687) [details] > Xorg log file > sorry, rest of the text was lost while saving. As you can see in the attached log file, EXA works with xf86-video-ati and it doesn't produce artifacts. Video looks good and xvinfo seconds, that XV works. In opposite to the original poster, I have a R680 (=HD3870X2 with 2xRV670).
Happens to me on a RHD 4670 (RV730XT) on KDE 4.2.2 desktop. Don't need to do any thing specific - appears randomly. Have not tested radeon enough to comment on whether it happens there - will try radeon shortly.
Can you narrow down the problem to a particular accel hook? try the following options to selectively disable hooks: Option "EXANoComposite" Option "EXANoUploadToScreen" Option "EXANoDownloadFromScreen"
(In reply to comment #7) > Can you narrow down the problem to a particular accel hook? try the following > options to selectively disable hooks: > > Option "EXANoComposite" > Option "EXANoUploadToScreen" > Option "EXANoDownloadFromScreen" > I have just tried these options with latest radeonhd driver from git. I have tested with each option alone, with the three options together, and with both EXANoUploadToScreen and EXANoDownloadFromScreen. In all cases, the result is the same as with none of these options: the background corruption is still there (on a side note, EXANoDownloadFromScreen has a major impact on performance). One more thing I have just noticed, enabling KWin 4.2 desktop effects using XRender composition (since OpenGL is not possible yet on R6xx) reduces significantly the background corruption (I've only tested this with none of the three options). But XRender desktop effects have an impact on performances with the radeonhd driver while they run smoothly with the radeon (xf86-video-ati) driver.
I can confirm that radeon driver does not have any similar issues.
(In reply to comment #8) > (In reply to comment #7) > > Can you narrow down the problem to a particular accel hook? try the following > > options to selectively disable hooks: > > > > Option "EXANoComposite" > > Option "EXANoUploadToScreen" > > Option "EXANoDownloadFromScreen" > > > > I have just tried these options with latest radeonhd driver from git. > > I have tested with each option alone, with the three options together, and with > both EXANoUploadToScreen and EXANoDownloadFromScreen. > > In all cases, the result is the same as with none of these options: the > background corruption is still there (on a side note, EXANoDownloadFromScreen > has a major impact on performance). > > One more thing I have just noticed, enabling KWin 4.2 desktop effects using > XRender composition (since OpenGL is not possible yet on R6xx) reduces > significantly the background corruption (I've only tested this with none of the > three options). But XRender desktop effects have an impact on performances with > the radeonhd driver while they run smoothly with the radeon (xf86-video-ati) > driver. > I can second that behaviour of radeonhd. None of the options affects the issue this bug report is about but has impacts on performance (of course). Using kwin as a composite manager via xrender reduces the problems which is caused in my opinion simply by an increased rate of desktop upgrades. These are caused by all that fancy shadow and fading effects. Greets, Torsten
Does this issue occur with the preferred ati driver (xf86-vide-ati)? If so, please move this to the Driver/Radeon component. Development of radeonhd has pretty much halted and development focus is on the ati driver. Please see http://www.x.org/wiki/radeonhd If the issue does not exist in the ati driver (or if there is no response to this message), this bug will be closed as WONTFIX unless someone contributes a patch.
(In reply to comment #11) > Does this issue occur with the preferred ati driver (xf86-vide-ati)? If so, > please move this to the Driver/Radeon component. > > Development of radeonhd has pretty much halted and development focus is on the > ati driver. Please see http://www.x.org/wiki/radeonhd > > If the issue does not exist in the ati driver (or if there is no response to > this message), this bug will be closed as WONTFIX unless someone contributes a > patch. As said in the original bug report, the radeon driver (which you refer to as xf86-video-ati) was not affected by this issue. I have not used radeonhd for more than 2 years, and it has not seen any commit for more than one year according to its git repository, so WONTFIX is definitively OK for me.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.