Bug 31590 - Black space between colors on mole hill example
Black space between colors on mole hill example
Status: RESOLVED FIXED
Product: Mesa
Classification: Unclassified
Component: Mesa core
git
Other All
: medium normal
Assigned To: mesa-dev
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2010-11-12 22:48 UTC by Jaime Rave
Modified: 2011-06-08 07:16 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
Screenshot of the issue (39.89 KB, image/png)
2010-11-17 21:23 UTC, Jaime Rave
Details
glxinfo (11.81 KB, text/plain)
2010-11-29 20:20 UTC, Jaime Rave
Details
patch 1 of 2 to fix vbo/display list bugs (2.21 KB, patch)
2011-06-07 15:04 UTC, Brian Paul
Details | Splinter Review
patch 2 of 2 to fix vbo/display list bugs (2.36 KB, patch)
2011-06-07 15:05 UTC, Brian Paul
Details | Splinter Review

Note You need to log in before you can comment on or make changes to this bug.
Description Jaime Rave 2010-11-12 22:48:31 UTC
the molehill example on the opengl page is not been draw correctly, there are some black pieces that separe the colors.

This is a screenshot of how should be draw:
http://www.opengl.org/resources/code/samples/glut_examples/examples/molehill.jpg

The source code of the example can be found on:
http://www.opengl.org/resources/code/samples/glut_examples/examples/molehill.c
Comment 2 Jaime Rave 2010-11-17 21:23:29 UTC
Created attachment 40357 [details]
Screenshot of the issue

This is on a GM45:

00:02.0 0300: 8086:2a42 VGA compatible controller: Intel Corporation Mobile 4 Series Chipset Integrated Graphics Controller (rev 07)


OpenGL renderer string: Mesa DRI Mobile Intel® GM45 Express Chipset GEM 20100330 DEVELOPMENT x86/MMX/SSE2
OpenGL version string: 2.1 Mesa 7.10-devel

This is Ubuntu 10.10 and Linux 2.6.36-1-generic-pae

I'm using mesa and Intel drivers from xorg-edgers ppa.
Comment 3 Jaime Rave 2010-11-29 20:20:27 UTC
Created attachment 40654 [details]
glxinfo

I'm attaching my glxinfo. This is also happening with mesa 7.9 (7.9~git20100924-0ubuntu2).
Comment 4 Eric Anholt 2011-06-07 11:19:29 UTC
Not driver-specific -- appears in swrast as well.
Comment 5 Brian Paul 2011-06-07 11:35:35 UTC
With a debug build I'm seeing:
Mesa: User error: GL_INVALID_OPERATION in glEnd

Looks like there's a problem with NURBS and glBegin/End in a display list.
Jaime, perhaps you could try an older version of Mesa and bisect?  I'm pretty sure this used to work.
Comment 6 Jaime Rave 2011-06-07 12:46:04 UTC
(In reply to comment #5)
> With a debug build I'm seeing:
> Mesa: User error: GL_INVALID_OPERATION in glEnd
> 
> Looks like there's a problem with NURBS and glBegin/End in a display list.
> Jaime, perhaps you could try an older version of Mesa and bisect?  I'm pretty
> sure this used to work.
If you can give me some instructions about how to do it, I will be more than glad to help.
Comment 7 José Fonseca 2011-06-07 13:28:20 UTC
(In reply to comment #5)
> Looks like there's a problem with NURBS and glBegin/End in a display list.
> Jaime, perhaps you could try an older version of Mesa and bisect?  I'm pretty
> sure this used to work.

I tried as far as mesa_7_5_branch and it still fails.
Comment 8 Brian Paul 2011-06-07 13:32:21 UTC
I've found the problem and have a fix.  Just regression testing...
Comment 9 Brian Paul 2011-06-07 15:04:48 UTC
Created attachment 47692 [details] [review]
patch 1 of 2 to fix vbo/display list bugs
Comment 10 Brian Paul 2011-06-07 15:05:12 UTC
Created attachment 47693 [details] [review]
patch 2 of 2 to fix vbo/display list bugs
Comment 11 Brian Paul 2011-06-07 15:06:22 UTC
The patches I just attached fixes the problem with the molehill demo and fixes another bug/crash that I found with a new piglit test.

No regressions found so I'll probably commit them in a day or so.
Feel free to test in the mean time.
Comment 12 Ian Romanick 2011-06-07 17:45:53 UTC
(In reply to comment #11)
> The patches I just attached fixes the problem with the molehill demo and fixes
> another bug/crash that I found with a new piglit test.
> 
> No regressions found so I'll probably commit them in a day or so.
> Feel free to test in the mean time.

The first patch looks trivially okay.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

I'd have to look at the second one a lot more deeply to fully grok the issues.  It generally looks okay, so if it doesn't cause any piglit regressions:

Acked-by: Ian Romanick <ian.d.romanick@intel.com>
Comment 13 José Fonseca 2011-06-08 02:35:23 UTC
Patches look OK to me, Brian.
Comment 14 Brian Paul 2011-06-08 07:16:23 UTC
Fixed with commits
f1cdce95f606584a56eabf3b38eea19ff4c75757
62811057f41bf7c027da6156a44383266c7f3952