Bug 22483 - [865G] Random freezes, with intel 2.7.99.901
Summary: [865G] Random freezes, with intel 2.7.99.901
Status: RESOLVED FIXED
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/intel (show other bugs)
Version: 7.4 (2008.09)
Hardware: x86-64 (AMD64) Linux (All)
: medium critical
Assignee: Eric Anholt
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2009-06-25 15:40 UTC by Götz
Modified: 2009-07-16 12:01 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
intel_gpu_dump (172.14 KB, application/octet-stream)
2009-06-25 15:40 UTC, Götz
no flags Details
intel_gpu_dump after logout freeze (83.12 KB, application/octet-stream)
2009-07-02 11:46 UTC, Götz
no flags Details
screenshot of the display corruption (16.25 KB, image/png)
2009-07-16 09:35 UTC, Götz
no flags Details

Description Götz 2009-06-25 15:40:07 UTC
Created attachment 27142 [details]
intel_gpu_dump

00:02.0 VGA compatible controller: Intel Corporation 82865G Integrated Graphics
Controller (rev 02)

linux: 2.6.30-10-generic 
xserver: 2:1.6.1.901+git20090622+server1.6-branch.dbac41b6-0ubuntu0sarvatt~jaunty
intel: 2:2.7.99.901+git20090619.534e73ad-0ubuntu0sarvatt2~jaunty
libdrm: 2.4.11+git20090519.f355ad89-0ubuntu0sarvatt~jaunty
mesa: 7.5.0~git20090622+mesa-7-5-branch.abfd56c2-0ubuntu0sarvatt

Using 865G I get about three freezes in one day. Mostly when using some OpenGL apps, but also after leaving the pc idle some minutes, don't know exactly how much time.

After the freeze, there is no useful log, but I attached the output from intel_gpu_dump. Also, my xorg.conf has only Option "DontZap" "False".
Comment 1 Götz 2009-07-02 11:46:16 UTC
Created attachment 27342 [details]
intel_gpu_dump after logout freeze

linux: 2.6.30-10-generic 
xserver: 2:1.6.1.902~git20090701+server-1.6-branch.a7a93c12-0ubuntu0sarvatt~jaunty
intel: 2:2.7.99.901+git20090701.1e4784bf-0ubuntu0sarvatt
libdrm: 2.4.11+git20090519.f355ad89-0ubuntu0sarvatt~jaunty
mesa: 7.5.0~git20090629+mesa-7-5-branch.4e43126a-0ubuntu0sarvatt

Now the freeze also happens when I logout from the session (KDE 4.2).
Comment 2 Eric Anholt 2009-07-15 13:44:11 UTC
Please be sure to include Xorg.0.log and dmesg.  Your logout freeze I suspect is X just crashing due to the server regen bug.  However, the first dump shows off an issue that I knew about at one point and had forgotten, so I'll consider this bug to be about that problem.
Comment 3 Eric Anholt 2009-07-15 15:34:45 UTC
Apparently I had the change 90% done sitting around in my tree.

commit a1e6abb5ca89d699144d10fdc4309b3b78f2f7a9
Author: Eric Anholt <eric@anholt.net>
Date:   Wed Jul 15 14:15:10 2009 -0700

    Use batch_start_atomic to fix batchbuffer wrapping problems with 8xx render.
    
    Bug #22483.
Comment 4 Tobias Doerffel 2009-07-16 03:57:16 UTC
I'm sorry to inform you that your commit (a1e6abb5ca89d699144d10fdc4309b3b78f2f7a9) completely breaks things for me (00:02.0 VGA compatible controller: Intel Corporation Mobile 945GME Express Integrated Graphics Controller (rev 03)).

The problem is that intel_batch_start_atomic() gets called twice: once inside i915_composite() then again in i830_composite() - this causes an assertion in intel_batch_start_atomic() (i830_batchbuffer.h:59). So every simple compositing operation makes X crash. Reverting your commit makes everything work fine. Commenting out intel_batch_start_atomic() and intel_batch_end_atomic() in i915_render.c doesn't seem to help.
Comment 5 Götz 2009-07-16 09:35:10 UTC
Created attachment 27769 [details]
screenshot of the display corruption

I didn't include the Xorg.0.log and dmesg because the where the same before and after the freezes.

With the a1e6abb5ca89d699144d10fdc4309b3b78f2f7a9 commit I don't get more freezes, but many display corruptions with and w/o composition, but not in OpenGL apps. Moving the mouse above places in the screen makes those places disappear like in the screenshot, when opening new apps they look the same.
Comment 6 Eric Anholt 2009-07-16 11:57:58 UTC
commit 8dd7ccf37e2de6d80b556e6d18af244cefc1e417
Author: Eric Anholt <eric@anholt.net>
Date:   Thu Jul 16 11:40:51 2009 -0700

    Fix 915-class Render after the 8xx-class Render fix.
    
    The two shared i830_composite.c, so giving i830 atomic batch support
    triggered anger about starting i830's atomic area while in i915's atomic
    area.  Instead, split the emit-a-primitive stuff from the state emission.
Comment 7 Eric Anholt 2009-07-16 12:01:38 UTC
Oh,and Götz, unless the bad rendering is new with this change, can you move it to a new bug report?


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.