libGL crashes Enemy Territory Quake Wars(F8 x86_64)

[Dan Williams]

On Fri, 2008-02-01 at 11:24 -0800, Sean Bruno wrote:
> On Fri, 2008-02-01 at 14:11 -0500, Dan Williams wrote:
> > On Fri, 2008-02-01 at 10:36 -0800, Sean Bruno wrote:
> > > I reported a crash to the idSoftware folks that appears to be a Fedora
> > > x86_64 specific libGL crash.
> > > 
> > > I'm not really able to provide much in the way of debugging traces other
> > > than show that a crash is occuring in libGL.  According to id, x86_64
> > > Fedora is the only distro reporting this crash:
> > 
> > You appear to be using the nvidia binary driver, hence you need to talk
> > to nvidia to get them to fix their driver.  It's not a problem Fedora
> > can fix because nobody but nvidia has the source to their proprietary
> > driver.
> > 
> > Dan
> > 
> > > 
> > > Core was generated by `./etqw.x86 +set developer 1 +set logfile 1'.
> > > Program terminated with signal 11, Segmentation fault.
> > > #0  0x082e51fc in __gxx_personality_v0 ()
> > > at ../../../../gcc-4.1.2/libstdc++-v3/libsupc++/eh_personality.cc:349
> > > 349     ../../../../gcc-4.1.2/libstdc++-v3/libsupc++/eh_personality.cc:
> > > No such file or directory.
> > >         in ../../../../gcc-4.1.2/libstdc++-v3/libsupc
> > > ++/eh_personality.cc
> > > (gdb) set logging on
> > > Copying output to gdb.txt.
> > > (gdb) set logging off
> > > Done logging to gdb.txt.
> > > (gdb) whe
> > > #0  0x082e51fc in __gxx_personality_v0 ()
> > > at ../../../../gcc-4.1.2/libstdc++-v3/libsupc++/eh_personality.cc:349
> > > #1  <signal handler called>
> > > #2  0x02eed5e2 in ?? () from /usr/lib/libGLcore.so.1
> > > #3  0x00000001 in ?? ()
> > > #4  0x00001403 in ?? ()
> > > #5  0xf6dfb008 in ?? ()
> > > #6  0x00000072 in ?? ()
> > > #7  0x0c7c0680 in ?? ()
> > > #8  0x02f7cdb8 in ?? () from /usr/lib/libGLcore.so.1
> > > #9  0xf6e64000 in ?? ()
> > > #10 0x00000000 in ?? ()
> > > 
> > 
> 
> Forgive my ignorance, but how can you tell that from my back trace?  

Google for /usr/lib/libGLcore.so.1 :)  I couldn't find it in any of the
usual suspects (mesa, xorg, etc) so a quick google search found that
it's part of the nvidia binary driver packages.

So that's not to say that this problem _isn't_ ultimately caused by
something in Fedora (glibc changes, etc).  But in the end when something
in Fedora changes there isn't a good way to do anything about binary
packages until the vendor does a rebuild or starts to debug the issue
themselves, since they have the source.

Dan