Bug 11075

Summary: Strange effects with 3D applications and KDETV
Product: Mesa Reporter: Taiyo Rawle <taiyo_rawle>
Component: Drivers/DRI/i965Assignee: Eric Anholt <eric>
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: medium Keywords: NEEDINFO
Version: 6.5   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
URL: http://www.flickr.com/photo_zoom.gne?id=518784044&size=o
Whiteboard:
i915 platform: i915 features:

Description Taiyo Rawle 2007-05-28 16:12:38 UTC
I hope this is the right place to report this bug - I've only noticed it on G965 based systems.

When some 3D applications and KDETV are on the screen at the same time then frames from KDETV will "leak" from thier window and fill most of the screen. See the URL for a screenshot - KDETV is in a small window in the bottom right of the screen and the Xscreensaver preferences window is next to it showing the GLschool preview - the screensaver menu and the preview is visible though the rest of the desktop except for the taskbar is hidden behind the "leaked" frame from KDETV. I also noticed this with GtkRadiant (a Quake map editor) which is a Gtk application that has a 3D view. I haven't been able to recreate this effect wit 2D only applications. So far GtkRadiant and the Glschool screensaver are the only two 3D applications I've noticed this with. This doesn't crash X, KDETV or the 3D application and the areas of the screen effected by the "leaked" video frame return to normal when refreshed (e.g. when I drag a window over them). I was using KDETV with a BT848 based analogue PCI tuner card.

System information: 
Fedora 7 test 4 on a core 2 duo system.

uname -a: 
Linux localhost.localdomain 2.6.21-1.3194.fc7 #1 SMP Wed May 23 22:47:07 EDT 2007 x86_64 x86_64 x86_64 GNU/Linux

glxinfo: 
OpenGL vendor string: Tungsten Graphics, Inc
OpenGL renderer string: Mesa DRI Intel(R) 965G 4.1.3002
OpenGL version string: 1.4 Mesa 6.5.2
Comment 1 Eric Anholt 2009-07-02 15:05:45 UTC
I feel ashamed of replying after such a long time.  My suspicion would be that this is fixed with DRI2, with private backbuffers and the server doing the swaps on its own with a consistent view of the cilprects.  Could you retest at this point?
Comment 2 Eric Anholt 2009-10-19 14:00:17 UTC
I'm going to go ahead and mark it closed at this point on the assumption that it's fixed.  Please reopen if the problem continues.

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.