Problems with radeon_dri and Blender

Wed Jul 2 19:41:26 PDT 2008

Roland Scheidegger <sroland at tungstengraphics.com> [08-07-03 02:41]:
> On 02.07.2008 19:25, meino.cramer at gmx.de wrote:
> > Alex Deucher <alexdeucher at gmail.com> [08-07-02 18:56]:
> >> On Wed, Jul 2, 2008 at 12:17 PM,  <meino.cramer at gmx.de> wrote:
> >>> Hi,
> >>>
> >>> When using Blender my radeon_dri driver crashes with a segfault (and
> >>> Blender follows :-/  ). But it only happens when some very specific actions
> >>> are performed and only if the Shading ("Z") is activated for the 3D view.
> >>>
> >>> Currently I am only able to watch this. And dont know how to pinpoint
> >>> the problem and whether the radeon_dri driver or Blender (or both) or
> >>> the hardware habe a problem.
> >>>
> >>> If someone is interested, I will post an *.blend file to trigger the
> >>> segfault.
> >> The best starting point would be if you could get a backtrace from
> >> blender using gdb.
> > 
> > Here ist comes:
> > 
> > solfire:/home/mccramer>gdb blender
> > GNU gdb 6.7.1
> > Copyright (C) 2007 Free Software Foundation, Inc.
> > License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
> > This is free software: you are free to change and redistribute it.
> > There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
> > and "show warranty" for details.
> > This GDB was configured as "i686-pc-linux-gnu"...
> > Using host libthread_db library "/lib/libthread_db.so.1".
> > (gdb) run
> > Starting program: /usr/local/bin/blender
> > [Thread debugging using libthread_db enabled]
> > [New Thread 0xb74af910 (LWP 24683)]
> > Compiled with Python version 2.4.4.
> > Checking for installed Python... got it!
> > 
> > Program received signal SIGSEGV, Segmentation fault.
> > [Switching to Thread 0xb74af910 (LWP 24683)]
> > 0xb733f3e3 in vbo_split_prims () from /usr/lib/dri/radeon_dri.so
> > (gdb) backtrace
> > #0  0xb733f3e3 in vbo_split_prims () from /usr/lib/dri/radeon_dri.so
> > (gdb)
> > 
> > 
> > it is not that much, but....
> 
> Looks similar to http://bugs.freedesktop.org/show_bug.cgi?id=12164
> though this should be fixed in mesa 7.0.3. In any case a backtrace from
> a debug build of the radeon dri driver would be more helpful...
> 
> Roland
> 

Hi,

I recompiled mesa-7.0.3 with "-ggdb" set in the CFLAGS and I also
enabled "debug" for the USEFLAGS of the mesa-package (Gentoo build 
system aka "emerge").

CFLAGS for mesa while compiling it were: CFLAGS="-march=athlon64 -ggdb"
USE flags for mesa were:
USE="debug -doc -motif (-nptl) -pic -xcb" VIDEO_CARDS="radeon -i810 -mach64 -mga -none -r128 -s3virge -savage -sis (-sunffb) -tdfx -trident -via"

Now, gdb shows this backtrace ( I run it three times...) :

solfire:/home/mccramer>gdb blender
GNU gdb 6.7.1
Copyright (C) 2007 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "i686-pc-linux-gnu"...
Using host libthread_db library "/lib/libthread_db.so.1".
(gdb) run
Starting program: /usr/local/bin/blender
[Thread debugging using libthread_db enabled]
[New Thread 0xb745d910 (LWP 19718)]
Compiled with Python version 2.4.4.
Checking for installed Python... got it!

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb745d910 (LWP 19718)]
0xb71fda7e in _tnl_draw_prims () from /usr/lib/dri/radeon_dri.so
(gdb) backtrace
#0  0xb71fda7e in _tnl_draw_prims () from /usr/lib/dri/radeon_dri.so
#1  0xb72df122 in ?? () from /usr/lib/dri/radeon_dri.so
#2  0x093dd578 in ?? ()
#3  0xbf97dc6c in ?? ()
#4  0xbf181170 in ?? ()
#5  0x00000002 in ?? ()
#6  0x00000000 in ?? ()
(gdb) run
The program being debugged has been started already.
Start it from the beginning? (y or n) y
Starting program: /usr/local/bin/blender
[Thread debugging using libthread_db enabled]
[New Thread 0xb74b8910 (LWP 19763)]
Compiled with Python version 2.4.4.
Checking for installed Python... got it!

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb74b8910 (LWP 19763)]
0xb733a688 in vbo_split_inplace () from /usr/lib/dri/radeon_dri.so
(gdb) backtrace
#0  0xb733a688 in vbo_split_inplace () from /usr/lib/dri/radeon_dri.so
#1  0xb7338d2a in vbo_split_prims () from /usr/lib/dri/radeon_dri.so
#2  0xb7258b76 in _tnl_draw_prims () from /usr/lib/dri/radeon_dri.so
#3  0xb733a122 in ?? () from /usr/lib/dri/radeon_dri.so
#4  0x093dd578 in ?? ()
#5  0xbffd8acc in ?? ()
#6  0xbf7dc3d0 in ?? ()
#7  0x00000002 in ?? ()
#8  0x00000000 in ?? ()
(gdb) run
The program being debugged has been started already.
Start it from the beginning? (y or n) y
Starting program: /usr/local/bin/blender
[Thread debugging using libthread_db enabled]
[New Thread 0xb734c910 (LWP 19787)]
Compiled with Python version 2.4.4.
Checking for installed Python... got it!

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb734c910 (LWP 19787)]
0xb71ce688 in vbo_split_inplace () from /usr/lib/dri/radeon_dri.so
(gdb) backtrace
#0  0xb71ce688 in vbo_split_inplace () from /usr/lib/dri/radeon_dri.so
#1  0xb71ccd2a in vbo_split_prims () from /usr/lib/dri/radeon_dri.so
#2  0xb70ecb76 in _tnl_draw_prims () from /usr/lib/dri/radeon_dri.so
#3  0xb71ce122 in ?? () from /usr/lib/dri/radeon_dri.so
#4  0x093dd578 in ?? ()
#5  0xbfb6a65c in ?? ()
#6  0xbf36f360 in ?? ()
#7  0x00000002 in ?? ()
#8  0x00000000 in ?? ()
(gdb)
-- 

The /var/log/Xorg.0.log does not show anything interesting. But if
wanted, I will post it.

Thank you very much for your help in advance!

Kind regards,
 Meino Cramer

Please don't send me any Word- or Powerpoint-Attachments
unless it's absolutely neccessary. - Send simply Text.
See http://www.gnu.org/philosophy/no-word-attachments.html
In a world without fences and walls nobody needs gates and windows.