Bug 9289

Summary:

Kernel oops in GIT version of i915/drm

Product:

DRI

Reporter:

Sean McNamara <smcnam>

Component:

DRM/other

Assignee:

Thomas Hellström <thomas>

Status:

RESOLVED DUPLICATE

QA Contact:

Severity:

critical

Priority:

medium

CC:

geekounet

Version:

DRI git

Keywords:

have-backtrace

Hardware:

x86 (IA32)

OS:

Linux (All)

Whiteboard:

i915 platform:

i915 features:

Attachments:

Description	Flags
Almost same exact crash reproduced by geekounet` on Freenode #dri	none

Description Sean McNamara 2006-12-08 21:59:49 UTC

Hello; this is my first DRI bug report. I am using the following hardware:
http://glue.umd.edu/~jcd/x60.txt on Fedora Core 6 x86_64 SMP (2.6.18-1)
(gcc-4.1). I checked out the latest GIT version of DRM and Mesa3d, and the
latest CVS of xf86-video-i810. I compiled it all according to instructions, and
built libGL using the mesa3d config option linux-dri-x86-64. I use X.Org 7.1.1. 

As an aside, my direct rendering stack was all working "properly" in its OpenGL
1.2 glory, prior to when I commenced to screw things up with development code ;)
The inviting new features of Mesa 6.5.2 enticed me to upgrade before Red Hat has
a chance to incorporate the new release into Fedora.

When attempting to use glxgears or any OpenGL application, I receive a kernel
oops like this one: http://glue.umd.edu/~jcd/oops . I've bounced my config of
the DRM/DRI/mesa stack off of a friend who knows where things should be, and I'm
pretty sure I configured and installed everything properly.

I am more than willing to test changes or run diagnostics; I am freezing my
system's configuration and will host the linked files until this bug is fixed.
You might find me in #dri-devel, #dri, or #kde on FreeNode if you're up for some
on-the-spot testing (even if you lack the chipset I use, the Intel 945GM.) My
screen-name is Allquixotic.

 Thanks!! Sean McNamara.

Comment 1 Sean McNamara 2006-12-09 09:21:05 UTC

Created attachment 8038 [details]
Almost same exact crash reproduced by geekounet` on Freenode #dri

geekounet` put this in #dri from rafb/paste

Comment 2 Sean McNamara 2006-12-09 09:24:42 UTC

NOTE: This bug is no longer marked as x86_64. It is confirmed to also happen on
x86 Gentoo, with an i915GM. So the two configs that it happened on are very
different.

Comment 3 Thomas Hellström 2006-12-10 02:31:54 UTC

Have you compiled and installed the required modified agpgart module at

git://git.freedesktop.org/git/mesa/linux-agp-compat

?

/Thomas

Comment 4 Pierre Guinoiseau 2006-12-10 07:22:49 UTC

(In reply to comment #3)
> Have you compiled and installed the required modified agpgart module at
> 
> git://git.freedesktop.org/git/mesa/linux-agp-compat
> 
> ?
> 
> /Thomas
> 
> 

Thanks, i have installed this module and now it works fine for me.

Comment 5 Thomas Hellström 2006-12-10 09:28:22 UTC

OK, then this is a duplicate of 9120. 
We need to make the system fail more gracefully when the wrong agpgart module is
used.

*** This bug has been marked as a duplicate of 9120 ***

Comment 6 Sean McNamara 2006-12-10 10:56:30 UTC

(In reply to comment #3)
> Have you compiled and installed the required modified agpgart module at
> 
> git://git.freedesktop.org/git/mesa/linux-agp-compat
> 
> ?
> 
> /Thomas
> 
> 

Thomas,

Thanks for your assistance. If I compile the new agp module as indicated and
still can't get it to work, I will post back.

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.