Bug 35557

Summary: [945GM] Text display errors
Product: xorg Reporter: Bryce Harrington <bryce>
Component: Driver/intelAssignee: Chris Wilson <chris>
Status: RESOLVED FIXED QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: medium CC: chisser98, inform
Version: 7.6 (2010.12)   
Hardware: x86 (IA32)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
XorgLog
none
XorgConf.txt
none
CurrentDmesg.txt
none
desktop1.png
none
system info generated by debian bug report script for Andreas
none
screenshot illustrating the issue none

Description Bryce Harrington 2011-03-22 11:56:21 UTC
Forwarding this bug from Ubuntu reporter Jarrett Chisholm:
http://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/737967

[Problem]
Vertical pixel columns are missing from every 20-50th character in text in firefox and other applications.  See screenshot.

[Original Description]
Basically, text in Ubuntu tends not to draw quite right.  Please see the attached screenshot (desktop1.png) for a visual.

I've noticed this happening in Ubuntu, Thunderbird, and Eclipse (the main 3 programs I use).

DistroRelease: Ubuntu 11.04
Package: xorg 1:7.6~3ubuntu11
ProcVersionSignature: Ubuntu 2.6.38-1.28-generic 2.6.38-rc2
Uname: Linux 2.6.38-1-generic i686
Architecture: i386
CompositorRunning: compiz
Date: Fri Mar 18 22:17:57 2011
DistUpgraded: Log time: 2011-02-04 18:07:51.411927
DistroCodename: natty
DistroVariant: ubuntu
GraphicsCard:
 Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller [8086:27a2] (rev 03) (prog-if 00 [VGA controller])
   Subsystem: Dell Device [1028:01d4]
MachineType: Dell Inc. Latitude D520
SourcePackage: xorg
UpgradeStatus: Upgraded to natty on 2011-03-09 (9 days ago)
dmi.bios.date: 05/28/2007
dmi.bios.vendor: Dell Inc.
dmi.bios.version: A06
dmi.board.name: 0NF743
dmi.board.vendor: Dell Inc.
dmi.chassis.type: 8
dmi.chassis.vendor: Dell Inc.
dmi.modalias: dmi:bvnDellInc.:bvrA06:bd05/28/2007:svnDellInc.:pnLatitudeD520:pvr:rvnDellInc.:rn0NF743:rvr:cvnDellInc.:ct8:cvr:
dmi.product.name: Latitude D520
dmi.sys.vendor: Dell Inc.
version.compiz: compiz 1:0.9.4-0ubuntu5
version.libdrm2: libdrm2 2.4.23-1ubuntu3
version.libgl1-mesa-glx: libgl1-mesa-glx 7.10.1-0ubuntu2
version.xserver-xorg: xserver-xorg 1:7.6~3ubuntu11
version.xserver-xorg-video-ati: xserver-xorg-video-ati 1:6.14.0-0ubuntu2
version.xserver-xorg-video-intel: xserver-xorg-video-intel 2:2.14.0-4ubuntu2
version.xserver-xorg-video-nouveau: xserver-xorg-video-nouveau 1:0.0.16+git20110107+b795ca6e-0ubuntu5
Comment 1 Bryce Harrington 2011-03-22 12:34:57 UTC
Created attachment 44721 [details]
XorgLog
Comment 2 Bryce Harrington 2011-03-22 12:35:33 UTC
Created attachment 44722 [details]
XorgConf.txt
Comment 3 Bryce Harrington 2011-03-22 12:35:59 UTC
Created attachment 44723 [details]
CurrentDmesg.txt
Comment 4 Bryce Harrington 2011-03-22 12:36:21 UTC
Created attachment 44724 [details]
desktop1.png
Comment 5 Jarrett 2011-03-22 14:07:47 UTC
Hi Chris,

Let me know if you need any info from me.  I've added myself to the cc list, but here is my e-mail anyway: chisser98@yahoo.com

Cheers
Comment 6 Chris Wilson 2011-04-05 12:45:44 UTC
*** Bug 35995 has been marked as a duplicate of this bug. ***
Comment 7 Chris Wilson 2011-06-26 02:18:07 UTC
I have to admit to being baffled. I can grasp at loose straws for the kernel bugs for misaligning the tiled buffers....

So does Option "Tiling" "False" make any difference?
Comment 8 Jarrett 2011-06-26 05:24:55 UTC
(In reply to comment #7)
> I have to admit to being baffled. I can grasp at loose straws for the kernel
> bugs for misaligning the tiled buffers....
> 
> So does Option "Tiling" "False" make any difference?

Hi Chris,

I'm not sure what option that would be.  However, I've since installed Linux Mint 10 LXDE over my old Ubuntu installation and no longer have any issues with text.

On a side note, I had been upgrading my Ubuntu installation since 8.10 - maybe there was an issue during one of the upgrades that lead to these text issues?  Just a thought.

Cheers

Jarrett
Comment 9 Chris Wilson 2011-06-26 05:34:53 UTC
The crazy idea I had was that font-rendering itself was just wrong, perhaps some stale font data and freetype going mad. But I couldn't convince myself of that as the errors were too inconsistent. On the other hand, the kernel bugs are more likely to cause GPU hangs and even more severe corruption.

The option was a reference to xorg.conf under Section "Driver". But if you can't reproduce it any more, then there is little to chase down and I shall focus on the residual bugs.

Please reopen if you ever see this again! Thank you for the report.
Comment 10 Jarrett 2011-06-26 06:10:19 UTC
(In reply to comment #9)
> The crazy idea I had was that font-rendering itself was just wrong, perhaps
> some stale font data and freetype going mad. But I couldn't convince myself of
> that as the errors were too inconsistent. On the other hand, the kernel bugs
> are more likely to cause GPU hangs and even more severe corruption.
> 
> The option was a reference to xorg.conf under Section "Driver". But if you
> can't reproduce it any more, then there is little to chase down and I shall
> focus on the residual bugs.
> 
> Please reopen if you ever see this again! Thank you for the report.

heh bizarre.  If I ever encounter this again, I'll be sure to reopen the bug.  

Cheers

Jarrett
Comment 11 Andreas Kloeckner 2011-06-26 07:50:19 UTC
Created attachment 48442 [details]
system info generated by debian bug report script for Andreas

I'm still encountering the bug. System info attached.
Comment 12 Chris Wilson 2011-06-26 08:00:49 UTC
Andreas can you please attach a screenshot so I can see if it is the same bug as Jarretts and not one of the other 915GM bugs?
Comment 13 Andreas Kloeckner 2011-06-26 08:51:08 UTC
Created attachment 48443 [details]
screenshot illustrating the issue

At the moment, only the lower part of the 's' is missing. After a while of having that same server running, usually more parts of more letters disappear. (I can post another screenshot next time it gets bad.)
Comment 14 Chris Wilson 2011-06-26 09:10:43 UTC
Ok, that is a different bug, where the glyph cache itself gets corrupted, e.g. bug 36326. What's different about Jarrett's bug is that it involves the individual rendering of certain glyphs, more like in some cases that the kernel is wrong and glyphs overlapped with the wrong operator, and did not demonstrate global corruption of a glyph. Hence its oddity.

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.