OpenGL viewport cut-off with 915gm graphics

I've attached an example of the problem in an Amarok Analyzer. I'll try to get a decent shot from a screensaver too.

And here's the screensaver example. I used the Fireworks 3D OpenGL screensaver.

Interesting; the first one is being cut off about half-way down and the second one about 2/5ths of the way down.

It looks a bit like view-port corruption; are there any OpenGL applications that *do* run normally?

Can you also attach an:

  lspci -nv

and:

  /var/log/Xorg.0.log

from the session with the corruption.

Here's the requested output of lspci -nv

And here's the xorg.0.log file

All OpenGL screensaver from a default flight 7 kubuntu have this, except the following:

KPendulum (GL)
Space (GL)

Sorry, forgot to sudo it

This is the Xorg.0.log file from my system, which also exhibits this bug.

This is the output from lspci -nv, run as root, from my system which also exhibits the bug.

Hey, I have the same problem, with same graphics card. Armagetron is a openGL game and don't have this problem. I come from mandriva 2006 and screensavers look ok there.

I have attached the output from lspci and the Xorg log for my machine -- same graphics card. PlanetPenguin Racer works just fine, however.

Hi! I was trying to run compiz and I've updated some packages from http://xgl.compiz.info with no success :P
But now I'm able to view the screensavers in full screen :)

I'm running a Dell Latitude D620 with a 945GM graphics card and a 1440x900 display on Ubuntu Dapper 6.06 with full updates. I see the same situation on the GL screensavers, but other GL apps like PlanetPenguin Racer work fine. Incidentally, I saw the same problem on a Sony Vaio B100 with a 1024x768 display.

interestingly, installing rss-glx seems to stop this problem, at least with the openGL screensavers.

i got the problem on thinkpad r50e, too.
the problem was noticed while running google earth.
it is now fixed by using the respositries of http://xgl.compiz.info/

thx for the hint @ Edu

I can confirm that the repo's at http://xgl.compiz.info fixes this problem. I noticed an update i810 driver for X.org. I remeber reading somewhere that it was a driver issue. If this is the case, I suggest updating the dapper repo's with the version on compiz.net.

I tried the compiz repository, and while it seems to fix the problems, it also seems to come with a high cost - graphics acceleration is cut in half, according to glxgears. Has anyone else noticed this?

I'm also having OpenGL problems with my 915GM graphics (Thinkpad Z60m). The symptoms are as reported: viewport cut-off in various apps, in google earth: wrong viewport and messed up graphics.

So I decided to give the compiz repository a try. First I only updated the i810 driver, logged out, restarted X, logged in again. Unfortunately acceleration was now broken (as reported by googleearth).

I thought that maybe this driver depended on newer mesa-libs, so I did a full update in adept. Now OpenGL is completely broken. glxgears fails with:
$ glxgears
glxgears: intel_ioctl.c:62: intelEmitIrqLocked: Assertion `((*(int *)intel->driHwLock) & ~0x40000000U) == (0x80000000U|intel->hHWContext)' failed.
Aborted

Google Earth gives the following error message:
do_wait: drmWaitVBlank returned -1, IRQs don't seem to be working correctly.
Try running with LIBGL_THROTTLE_REFRESH and LIBL_SYNC_REFRESH unset.
(and it's *extremely* slow, much slower than with software rendering)

Rebooting doesn't help.

Here's the complete list of updated packages:
xserver-xorg-driver-i810 1:1.5.0.1-0ubuntu2
libdrm2 2.0.2+cvs20060824
libgl1-mesa 6.5.1+cvs20060824
libgl1-mesa-dri 6.5.1+cvs20060824
libglu1-mesa 6.5.1+cvs20060824
libpoppler1 0.5.3-0ubuntu1
libpoppler1-qt 0.5.3-0ubuntu1
libtunepimp2c2a 0.3.0-9.1ubuntu3.1quinn1
libxfixes3 1:4.0-0ubuntu1
libxcomposite1 1:0.3-0ubuntu1
libxss1 1:1.1-0ubuntu1
mesa-utils 6.5.1+cvs20060824
poppler-utils 0.5.3-0ubuntu1

Maybe waiting for the next CVS snapshot will solve my problem. Here's what I found out about this issue:

https://bugs.freedesktop.org/show_bug.cgi?id=8010

According to that page it should be fixed in mesa as of 20060826.

is this a dupe of #23816?

Chris -

Yes, this does seem to be the same bug described in #23816.

Chris Jones wrote:
> is this a dupe of #23816?
>

--
Charles Figura <email address hidden>
Associate Professor of Physics http://mcsp.wartburg.edu/figura
Director, Wartburg Platte Observatory (319) 352-8373
                                          (319) 352-8606 (fax)