Some time late in Ubuntu Maverick 'alpha' period, visual effects that involve "fading" became unusably slow. The following javascript web slideshow, when viewed with Firefox, makes this effect very obvious: http://www.electricprism.com/aeron/slideshow/example1.html but it is also apparent with e.g. F-Spot slideshow. Disabling KMS by adding "radeon.modeset=0" to the kernel command line makes the problem go away. The current version of X server on which I can reproduce the problem is 1:7.5+6ubuntu3. I observed the effect on three different machines with different Radeon models. It looks somewhat similar to the bug #23513, but in my case, 3D performance is unaffected. I can provide more information on request, though in fact I think that anyone can reproduce it at will; I see it when I boot from "live USB" containing Maverick. Eugene
That page seems to work fine here. What hardware are you using? Can you attach your xorg log and dmesg output? Any chance you can narrow down when this started or what update caused it?
Created attachment 39445 [details] Xorg logfile
I can confirm that behaviour. It also happens at me, but only with firefox. Opera works fine. xorg-server: 1.9.0-r2 radeon: git @evergreen-accel branch kernel: drm26 git kernel hardware: Radeon HD 5750
(In reply to comment #3) > I can confirm that behaviour. > > It also happens at me, but only with firefox. Opera works fine. > > xorg-server: 1.9.0-r2 > radeon: git @evergreen-accel branch > kernel: drm26 git kernel > hardware: Radeon HD 5750 Are you on ubuntu as well? Also, evergreen_accel has been merged to master. Does master work any better?
Oh, I didn't knew that. Is there a place where such things are annouced? Thats why there were no updates at the evergreen_accel branch :D I'm not on ubuntu, I'm using gentoo. I'll test the master branch when I'm back at my linux later this evening.
Created attachment 39452 [details] Xorg.log from Compaq 8510p running maverick from "live USB" As requested, I am attaching Xorg.log. This is just one case, I observe the same behavior on RS690 and on M56GL.
Created attachment 39453 [details] dmesg from Compaq 8510p Also including `dmesg` output as requested. Sorry I cannot give the exact point when the problem appeared: it is easy to not notice until you try to watch a slideshow.
Ok, it doesn't happen at me with the master branch. Sorry for that.
Eugene, can you try git master of xf86-video-ati? Does that help? Next could you try previous xf86-video-ati releases (6.13.0, 6.13.1, 6.13.2) and see where the problem started?
Alex, can I use some binary build(s) for this test? I remeber building from the source is not quick and I am somewhat time-constrained at the moment. (Please forgive me, it was not a good thing to say to the developers, and I really appreciate *your* time and effort, but...)
I figured that I could first try prebuild packages from ubuntu edgers-ppa, and lo and behold! running 1:6.13.99+git20101012.0d1f9fd0-0ubuntu0sarvatt with KMS enabled the problem is gone! So, sorry for bothering you, I will now go report on launchpad as it is apparently ubuntu-specific. Eugene
Problem due old driver Version!
Reported to Ubuntu as https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/660999
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.