Bug#379940: Bug#385945: boson: crashes on singleplayer game start

[Steve Langasek]

On Sat, Sep 16, 2006 at 02:14:09AM +0200, Andreas Beckermann wrote:
> > In the report on Sparc, there is additionally an X Error immediately before
> > this, returning a BadRequest for opcode 128 (XFree86-DRI).  I thought
> > perhaps the problem here was that boson's OpenGL support only worked with
> > DRI, so I gave it a try myself; it seems that even with DRI enabled here, I
> > get the same error about needing GL_EXT_framebuffer_object.  I *also* get a
> > warning about DRI being unavailable, even though I know that DRI is enabled
> > on my display and I'm using it successfully!
> [...]

> Well, see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=379940, I am 
> waiting for the additional information here.
> This problem is unfortunately pretty much a "fishing in the dark" kind of 
> thing and can have a lot of reasons. A popular one is that the wrong libGL.so 
> is picked up, which is why boson tells you which one it uses - so if you had 
> nvidia/ati drivers installed before, you may want to check that they are 
> actually gone.

No.  libGL.so.1 is the mesa libGL, and the right one for my system.  This is
not a problem of the wrong driver being available on the system, there is
simply something wrong with boson's DRI detection wrt the DRI-enabled GL in
Debian.

> The output of ldd on program that links against libGL and _is_
> accellerated would help here.

Doubtful, but here's a simple example:

$ ldd /usr/bin/glxgears 
        linux-gate.so.1 =>  (0xffffe000)
        libGL.so.1 => /usr/lib/libGL.so.1 (0xa7f01000)
        libGLU.so.1 => /usr/lib/libGLU.so.1 (0xa7e86000)
        libglut.so.3 => /usr/lib/libglut.so.3 (0xa7e51000)
        libm.so.6 => /lib/tls/i686/cmov/libm.so.6 (0xa7e2c000)
        libc.so.6 => /lib/tls/i686/cmov/libc.so.6 (0xa7cfb000)
        libX11.so.6 => /usr/lib/libX11.so.6 (0xa7c35000)
        libXext.so.6 => /usr/lib/libXext.so.6 (0xa7c27000)
        libXxf86vm.so.1 => /usr/lib/libXxf86vm.so.1 (0xa7c22000)
        libpthread.so.0 => /lib/tls/i686/cmov/libpthread.so.0 (0xa7c0f000)
        libdl.so.2 => /lib/tls/i686/cmov/libdl.so.2 (0xa7c0b000)
        libdrm.so.2 => /usr/lib/libdrm.so.2 (0xa7c04000)
        libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0xa7b25000)
        libgcc_s.so.1 => /lib/libgcc_s.so.1 (0xa7b1a000)
        /lib/ld-linux.so.2 (0xa7f89000)
        libXau.so.6 => /usr/lib/libXau.so.6 (0xa7b16000)
        libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0xa7b11000)
$

> > And indeed, starting a single player game (which works without crashing for
> > me) gets me a very slow display indicative of indirect rendering.  This is
> > with the i810 driver on i386.

> > So I haven't been able to reproduce the original problem, but boson doesn't
> > work very well for me, and I know I know libgl1-mesa-dri does work on my
> > system.

> What exactly makes you think so?

glxinfo, the fact that being an i810 user means I had to go to great pains
to /get/ a mesa that would do DRI right on my Debian system, and that dozens
of other GL-using apps run at appropriate speeds?

$ glxinfo | grep direct
direct rendering: Yes
$ 

Cheers,
-- 
Steve Langasek                   Give me a lever long enough and a Free OS
Debian Developer                   to set it on, and I can move the world.
vorlon at debian.org                                   http://www.debian.org/