| Summary: | Massive screen corruption w/ compositing after upgrade to xorg 7.6 [Object foo appears more than once] | ||
|---|---|---|---|
| Product: | Mesa | Reporter: | Florian Kriener <florian> |
| Component: | Drivers/DRI/i965 | Assignee: | Chris Wilson <chris> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | major | ||
| Priority: | medium | ||
| Version: | 7.6 | ||
| Hardware: | x86-64 (AMD64) | ||
| OS: | Linux (All) | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
| Attachments: |
Debian package bug report script output
check for duplicate handles in the execbuffer2 check for duplicate handles in the execbuffer2 |
||
[27419.764695] [drm:i915_gem_do_execbuffer] *ERROR* Object ffff88010ce51c00 appears more than once in object list I didn't spot it, what version of mesa where you using? Could you try libgl/experimental? Though, actually that was X doing evil things.... I just checked which version of libgl-mesa-* I have installed and it already is the one from experimental: 7.8.2. Would it help if I provided you with a photo (as screen shots don't show this phenomenon)? I confirm the [drm:i915_gem_do_execbuffer] *ERROR* Object ffff8800b0b08c00 appears more than once in object list but lucky enough I do not experience such massive screen corruption (many smaller issues with multiple screens, suspend/restore, recover of desktop from opengl screensaver, some sporadic freezes with kwin effects, typically at kwin startup after a warm boot or an X server restart). My system is DELL E6500 with Intel Corporation Mobile 4 Series Chipset Integrated Graphics Controller (rev 07) Ubuntu maverick standard kernel (2.6.35) 64bit xserver-xorg-video-intel 2.13.901+git20101124 libdrm 2.4.22+git20101207 mesa 7.9~git20100924 the last three from the glasen PPA that is tracking the latest intel drivers, libdrm and mesa. I now have three of those "appear more than once" things in my system messages. Interestingly, they systematically tend to appear in the following sequence: [ 534.044996] video LNXVIDEO:01: Restoring backlight state [ 534.057188] usb 3-1.3: USB disconnect, address 6 [ 534.074745] [drm:i915_gem_do_execbuffer] *ERROR* Object ffff8800b0b08c00 appears more than once in object list [ 534.142590] Skipping EDID probe due to cached edid (every time with a different object number) guess the usb thing is not related, but the sandwich between the backlight and the edid thing might be connected. Created attachment 40911 [details] [review] check for duplicate handles in the execbuffer2 Can you try this debugging patch for drm and see which application falls over? And even attach gdb? Created attachment 40912 [details] [review] check for duplicate handles in the execbuffer2 I think this should be fixed with: libdrm commit 36d4939343d8789d9066f7245fa2d4fe69119dd8 Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Mon Feb 14 09:39:06 2011 +0000 intel: Remember named bo ... and if asked to open a bo by the same global name, return a fresh reference to the previously allocated buffer. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> |
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.
Created attachment 40241 [details] Debian package bug report script output I have installed the new xserver experimental packages from the debian archive my T400s notebook (intel i915 graphics card) to test it. After restarting the xserver and logging in I experienced heavily flickering massive screen corruption which immediately stopped after turning off compositing and started anew when I turned it on again. I was using the xserver 1.9.2 with intel driver version 2.13.901 at that time and linux 2.6.36.