Bug 696798 - GPU hang with OpenGL
Summary: GPU hang with OpenGL
Keywords:
Status: CLOSED DUPLICATE of bug 684097
Alias: None
Product: Fedora
Classification: Fedora
Component: xorg-x11-drv-intel
Version: 15
Hardware: x86_64
OS: Linux
unspecified
unspecified
Target Milestone: ---
Assignee: Adam Jackson
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2011-04-14 20:38 UTC by Laurent Aguerreche
Modified: 2018-04-11 07:44 UTC (History)
3 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2011-04-15 22:13:29 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
The teapot hangs from time to time (897.44 KB, video/webm)
2011-04-14 20:40 UTC, Laurent Aguerreche
no flags Details
Source of the example (3.85 KB, text/x-csrc)
2011-04-14 20:41 UTC, Laurent Aguerreche
no flags Details

Description Laurent Aguerreche 2011-04-14 20:38:19 UTC
Description of problem:

For some applications, OpenGL rendering suffers of many small hangs such as this loop:
1/ the OpenGL animation is smooth
2/ suddenly the animation hangs and then advance a bit like five frames of rendering have been dropped
3/ the animation comes back to a smooth rendering but will hand later

Version-Release number of selected component (if applicable):
$ rpm -qa | grep intel
xorg-x11-drv-intel-2.14.0-5.fc15.x86_64

From lspci -v:
00:02.0 VGA compatible controller: Intel Corporation Device 0126 (rev 09) (prog-if 00 [VGA controller])
	Subsystem: Dell Device 050e
	Flags: bus master, fast devsel, latency 0, IRQ 44
	Memory at f1400000 (64-bit, non-prefetchable) [size=4M]
	Memory at e0000000 (64-bit, prefetchable) [size=256M]
	I/O ports at 4000 [size=64]
	Expansion ROM at <unassigned> [disabled]
	Capabilities: [90] MSI: Enable+ Count=1/1 Maskable- 64bit-
	Capabilities: [d0] Power Management version 2
	Capabilities: [a4] PCI Advanced Features
	Kernel driver in use: i915
	Kernel modules: i915

$ uname -a
Linux shaka 2.6.38.2-9.fc15.x86_64 #1 SMP Wed Mar 30 16:55:57 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux


How reproducible:
For some applications, hangs do not appear or in just a few cases. For example, I tried Extreme Tux Racer from Fedora 15 and I saw a small hang before Tux jumps on snow.

For Bullet physics demonstrations, the switch between each demonstration is always very painful.

On a small example I provide to do anaglyph rendering, hangs happen but are visible when I use a lot of CPU power.

Steps to Reproduce:
0. Build the executable:
   gcc -L/usr/lib -lm -lGL -lglut -lGLU stereo-anaglyph.c -o example
1. Runs my small example for anaglyph rendering (./my_example)
2. Runs a screencast (I used GNOME 3 screencast)
3. Press continuously the key 'e' so that the teapot turns 
4. See that after a few seconds -- and sometimes immediately -- some hangs happen (it seems also to depend on the CPU charge).

With Bullet physics, I use Bullet SVN: http://bullet.googlecode.com/svn/trunk
I so build demos and I run 'Demos/AllBulletDemos/AppAllBulletDemos'.


Actual results:
I did a screencast of my teapot example to show some hangs.
For Bullet physics demonstrations, it's even more obvious. I uploaded a screencast at http://dl.free.fr/vOOgY6zur

I am also wondering whether some slowdowns of GNOME 3 may come from the same problem.

Expected results:
No hangs at all...

Comment 1 Laurent Aguerreche 2011-04-14 20:40:21 UTC
Created attachment 492221 [details]
The teapot hangs from time to time

Comment 2 Laurent Aguerreche 2011-04-14 20:41:13 UTC
Created attachment 492222 [details]
Source of the example

Comment 3 Laurent Aguerreche 2011-04-14 20:54:55 UTC
In /var/log/messages, I see sometimes this type of message:

Apr 14 22:11:20 shaka kernel: [20728.991652] [drm:i915_hangcheck_ring_idle] *ERROR* Hangcheck timer elapsed... blt ring idle [waiting on 14616061, at 14616061], missed IRQ?

Comment 4 Laurent Aguerreche 2011-04-14 20:58:22 UTC
Damn, I didn't see those messages in /var/log/messages so my bug report may be related with #684097.

Sorry for the duplicate.

Comment 5 Matěj Cepl 2011-04-15 22:13:29 UTC
(In reply to comment #4)
> Damn, I didn't see those messages in /var/log/messages so my bug report may be
> related with #684097.
> 
> Sorry for the duplicate.

Thanks for finding it. It really looks like a duplicate.

*** This bug has been marked as a duplicate of bug 684097 ***


Note You need to log in before you can comment on or make changes to this bug.