Created attachment 24331 [details] X.org config file When using EXA, I am occasionally getting bad background redraws for some Qt (4.5) widgets - line edit, list views (see attached screenshots)... No (graphics) bugs seem to appear for GTK apps. Also, it doesn't matter, whether I enabled 'desktop effects' in the KDE (4.2). The bug doesn't appear with XAA enabled. Well, to be precise, the acceleration is disabled competely when using XAA, right? : (II) RADEON(0): XAA Render acceleration unsupported on Radeon 9500/9700 and newer. Please use EXA instead. (II) RADEON(0): Render acceleration disabled :-) I've attached the "bad" EXA configuration (a pretty generic one), screenshots what happens and log file. If you need any more asistance, I'll help gladly - I could probably handle compiling of driver by myself and/or even try to find and fix the issue, if someone provided me with hints where to look :-)
Created attachment 24332 [details] screenshot of a amarok directory listview (1)
Created attachment 24333 [details] screenshot of a amarok directory listview (2)
Created attachment 24334 [details] a Qt app edit view notice the double-drawn text on the left side of cursor.. I did this by clicking on the start of text a and pressing right arrow to move cursor. While doing this, the text got this scrambled (as the background wasn't probably reset to white)
Created attachment 24335 [details] X.org log file
Created attachment 24336 [details] kmail dialog This time, the 'uninitialized background' phenomenon occurs - the lineedit on the left contains previous contents along with newly typed text. On the right, the black backrounds contains some white 'dots' (looks like some uninitialized bitmap to me)...
Does Option "EXAOptimizeMigration" "off" help?
yep, that seem to have removed those artifacts... The only thing that bothers me is that redraws of screen (e.g. krusader dir list, or firefox webpage scrolling) is way like 2-3x (i.e. really visible) slower using exa (with composite manager or not), than using no acceleration at all.. It's really uncomfortably slow to be used, thus I guess I'll have to stick with the old-fashined desktop afterall ;-( However I'm not sure, whether it's driver bug or some kind of a "normal acceleration architecture feature"...
This sound a bit similar to my bug 19940.
I'm having a hard time reproducing this with EXA from current upstream xserver Git. Can you try that?
well, the next few days I personally won't have the time to do it - however I'll try as soon as I can (of course, only in case I'll be able to compile Xserver in the way, that won't completely mess up my system :-))
Is this still an issue with xf86-video-ati from git master?
Hello: I'm also seeing artifacts, but I'm not sure I'm seeing those explained in this bug. Maybe you could upgrade your graphics stack and tell us if you see any improvement. Thanks.
Mass closure: This bug has been untouched for more than six years, and is not obviously still valid. Please reopen this bug or file a new report if you continue to experience issues with current releases.
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.