Bug 28433 - Mesa DRI Intel 845G GEM Drivers returning artifacts in textures that can lockup PC on glxSwapBuffers.
Summary: Mesa DRI Intel 845G GEM Drivers returning artifacts in textures that can lock...
Status: RESOLVED MOVED
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/i915 (show other bugs)
Version: unspecified
Hardware: x86 (IA32) Linux (All)
: medium critical
Assignee: Default DRI bug account
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-06-07 20:39 UTC by Jane
Modified: 2019-09-18 19:32 UTC (History)
0 users

See Also:
i915 platform:
i915 features:

Attachments

Description Jane 2010-06-07 20:39:43 UTC 
I have been using the Mesa DRI Intel 845 Drivers since the very first 2.6 distros with no problem, but this problem started happening with the following distros and drivers:

F11:    Mesa DRI Intel(R) 845G GEM 20090114 x86/MMX/SSE2, 1.3 Mesa 7.5-devel
F12:    Mesa DRI Intel(R) 845G GEM 20090712 2009Q2 RC3, 1.3 Mesa 7.7-devel
U9.04L: Mesa DRI Intel(R) 845G GEM 20090114 x86/MMX/SSE2, 1.3 Mesa 7.3
U9.04:  Mesa DRI Intel(R) 845G GEM 20090326 2009Q1 RC2 x86/MMX/SSE2, 1.3 Mesa 7.4
U9.10L: Mesa DRI Intel(R) 845G GEM 20090712 ex9Q2 RC3 x86/MMX/SSE2, 1.3 Mesa 7.6
M2010:  Mesa DRI Intel(R) 845G GEM 20090712 2009Q2 RC3 x86/MMX/SSE2, 1.3 Mesa 7.5.2

where F stands for Fedora, U for Ubuntu, and M for Mandriva.

Any apps that build textures using glCopyTexImage2D, regardless of size anywhere from 64x64 upto 512x512, and regardless of whether the buffer attributes are RGB or RGBA, although it always lockups the machine if RGB is used.

There is a commonly used open source program called "Radial Blur & Rendering To A Texture" can show how bad the artifacts are with the above drivers:

http://nehe.gamedev.net/data/lessons/lesson.asp?lesson=36
http://nehe.gamedev.net/data/lessons/linuxglx/lesson36.tar.gz

The following are some specifics of one configuration that exhibits this problem:

Intel Corporation 82845G/GL[Brookdale-G/GE Chipset Integrated Graphics Device (rev 01)
Linux ubuntu 2.6.28-11-generic #37-Ubuntu SMP Mon Mar 23 16:40:23 UTC 2009 i686 GNU/Linux
Mesa DRI Intel 845G GEM 20090114 x86/MMX/SSE2, 1.3 Mesa 7.3
Comment 1 Jane 2010-06-09 14:29:37 UTC 
What does it take for a bug report to show up on the FreeDesktop Bugzilla – Bug List? Did I waste a couple hours of my life writing it just to be ignored? Did I show too many drivers at one time thereby turning into a hot-potato between various driver writers? I have never before taken the time to submit a bug-report to an open-source project and may never waste my time again if it does no good...
Comment 2 Tollef Fog Heen 2010-06-11 03:42:00 UTC 
Reassigning away from freedesktop.org, this looks like a mesa issue, not a problem with anything of the freedesktop.org infrastructure.
Comment 3 Timothy Arceri 2018-03-08 02:39:04 UTC 
Reassigning to i915. Probably needs to be retested.
Comment 4 Gert Wollny 2018-03-08 09:31:44 UTC 
I can't see any artefacts on i915 with GM45. In fact there the rendering looks better then with the llvmpipe software renderer and r600. These two create over-bright pixels that are not present with i915.

Versions tested: 
- i915 and llvmpipe: Mesa 13.0.6
- r600 and llvmpipe: Mesa 18.1.0-devel (git-dcd8730445
Comment 5 GitLab Migration User 2019-09-18 19:32:55 UTC 
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/mesa/mesa/issues/680.

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.