Occasionally when moving a window, screen corruption similar to that shown in the attachment will occur. It is frequently around the button or toolbar area but has appeared in an email text pane. It appears to be at the widget/control level as shading+unshading or moving off and on screen do not clear it. HOWEVER 'effects' with compositing are enabled so the usual window redrawing rules probably don't apply. Minimising and restoring the window DOES clear it. lspci reports: 01:05.0 VGA compatible controller: ATI Technologies Inc Radeon XPRESS 200M 5955 (PCIE)
Created attachment 225375 [details] sample corruption screenshot
Created attachment 225377 [details] emerge --info
Please attach your xorg.conf and Xorg.0.log Which versions of kde, qt, xorg-server and of the driver are you using?
Created attachment 228347 [details] Xorg log
No xorg.conf as this is a 'pure' HAL setup. Xorg.0.log attached. KDE 4.4.2, Qt 4.6.2 (Qt-core: 4.6.2-r1), xorg-server 1.7.6, xf86-video-ati 6.13.0.
Created attachment 228687 [details] screen shot showing corruption The attachment 'full-screen.png' shows corruption and might shed some light on the cause. I clicked on the konqueror browser window horizontal scroll bar to pan to the right of the web page. As can be seen, this has caused the green and purple background colour *from another window* *running a different, non-KDE application* to be copied into the browser window. So it seems that the root cause of the corruption is colour data somehow being taken from the wrong part of the screen. The other application is 'fullrecall' which does is not a KDE application and does not use the Qt widget set.
I can confirm this, I see the same corruption. It's not a KDE issue though; here it started at the moment of the x11-base/xorg-server-1.7.6 stabilization / upgrade. I have desktop effects disabled, and one of the apps showing the worst corruption is freeciv. Both x11-drivers/xf86-video-ati-6.13.0 and xf86-video-ati-6.12.6 are affected. Files follow...
Created attachment 228693 [details] sample screenshot for gtk+ application
Created attachment 228697 [details] Xorg log
Created attachment 228699 [details] xorg.conf
Created attachment 228701 [details] emerge --info output
Also give dmesg output. You seem to have kms disabled, but some kind of boot splash? Ensure that you don't use radeonfb, or else things will not work properly.
As a sidenote, since the x11-base/xorg-server-1.7.6 upgrade my display also irregulary locks up when the screensaver becomes active. No visible screensaver activity, unchanged image, no pointer visible, and no response to keyboard and mouse at all. The power button acpi event is still processed and shuts the machine down though. No useful info in the logs... I suspect the OpenGL screensavers; since I took them out of the random screensaver selection, I have not seen this again. May be related to the corruption, or not...
Created attachment 228703 [details] dmesg boot messages
Created attachment 228707 [details] kernel configuration
not sure why freedesktop-bugs@ was CC'd, it's not a general dumping ground...
Ok that may be the problem here - radeonfb.
I guess there are two or more different issues in this bug. What the original reporter sees is probably due to a radeonfb conflict. Regarding OpenGL applications, these will only work properly on R600+ since kernel 2.6.32.
(In reply to comment #17) > Ok that may be the problem here - radeonfb. > Negative. I made a 2.6.31-gentoo-r10 kernel without radeonfb, and still have the same problems. Files following...
Created attachment 228727 [details] kernel configuration (2.6.31-gentoo-r10, no radeonfb)
Created attachment 228729 [details] dmesg boot messages (2.6.31-gentoo-r10, no radeonfb)
Created attachment 228731 [details] Xorg log (2.6.31-gentoo-r10, no radeonfb)
(In reply to comment #18) > Regarding OpenGL applications, these will only work properly on R600+ since > kernel 2.6.32. The corruption is unchanged with 2.6.33-gentoo-r1 ... (please tell me if I should post the logs again)
This is not a Gentoo specific problem. I happens in Ubuntu also. I recommend closing this bug, the freedesktop guys should take care of this. Talking to them will be more effective but there's already a bug off this, and it's well known. All we can do is wait. Tip -- adding fbdev and vesa to to your VIDEO_CARDS variable alongside radeon might increase your your 2d acceleration.
(In reply to comment #24) > This is not a Gentoo specific problem. I happens in Ubuntu also. > > I recommend closing this bug, the freedesktop guys should take care of this. > Talking to them will be more effective but there's already a bug off this, and > it's well known. > > All we can do is wait. Could you please post the URL of the upstream bug here?
https://bugs.freedesktop.org/show_bug.cgi?id=20469 https://bugs.freedesktop.org/show_bug.cgi?id=20469 https://bugs.freedesktop.org/show_bug.cgi?id=21244 https://bugs.freedesktop.org/show_bug.cgi?id=21840 https://bugs.freedesktop.org/show_bug.cgi?id=22519 https://bugs.freedesktop.org/show_bug.cgi?id=23015 https://bugs.freedesktop.org/show_bug.cgi?id=23928 https://bugs.freedesktop.org/show_bug.cgi?id=26187 https://bugs.freedesktop.org/show_bug.cgi?id=26789 https://bugs.freedesktop.org/show_bug.cgi?id=27223 https://bugs.freedesktop.org/show_bug.cgi?id=27223 https://bugs.freedesktop.org/show_bug.cgi?id=27627 https://bugs.freedesktop.org/show_bug.cgi?id=27805 https://bugs.freedesktop.org/show_bug.cgi?id=21963 https://bugs.freedesktop.org/show_bug.cgi?id=18397 https://bugs.freedesktop.org/show_bug.cgi?id=27707 https://bugs.freedesktop.org/show_bug.cgi?id=23383 https://bugs.freedesktop.org/show_bug.cgi?id=27529
It's DRM. Disable DRM and the problem does not manifest.
Disabling DRM is *not* recommended. Unfortunately, there seems to be nothing for us to fix, please track the bugs upstream directly. Thanks
