On Tue, Jan 25, 2011 at 11:18:57AM -0800, Dan Nicholson wrote: > On Tue, Jan 25, 2011 at 9:31 AM, Ken Moffat <[email protected]> wrote: > > Thanks, Dan. > > > > Follow up question(s), sorry for venturing O/T for this list - > > > > Is it reasonable for me to raise either of the following two > > problems as Mesa bugs, or should I raise them against mesa-demos ? > > BTW, I haven't tested against Mesa-7.9 plus dj's patch to add these > > progs. I guess that might be my next step, but I'm severely lacking > > in available time at the moment. > > > > (a) radeon 7500 - > > > > with Mesa-7.9, glxgears works, but crashes if I try to resize its > > window (both clicking on the maximize window control, and dragging > > its borders). > > > > (b) radeon 9200se - > > > > with both Mesa-7.9 and 7.10, and mesa-demos-8.0.1, glxgears does > > not produce the window with the gears and dies after reporting > > Error: Unable to find RGB, Z, double-buffered visual > > > > In both cases, the same version of the ati driver and xorg-server, > > with Mesa-7.8.2 (which includes the demos), work without problems. > > Those both have to be driver stack bugs. glxgears is an incredibly > simple program that hasn't changed in a long time. There are basically > two important parts of the 3d stack when doing direct rendering - the > mesa driver and the kernel driver. The xorg driver (xf86-video-ati) > and server could certainly interact poorly, but most likely this is > somewhere flowing down from libGL -> r*00_dri/radeon_dri -> > libdrm/libdrm_radeon -> radeon.ko. > > So, you can try booting different kernels to see if the bug is in the > drm driver. What kernel are you using? Does dmesg say anything > interesting? Or, you can build an older version of mesa, set > LIBGL_DRIVERS_PATH=$builddir/lib (might have the envvar wrong), and > keep the kernel the same.
Again, thanks for the repsonse. For the 9200se, I've reverted to 7.8.2 at the moment. I'm sure I'm running 2.6.37 on that box, certainly it's been the same kernel across all three Mesa builds. I haven't looked at dmesg recently, will do when time permits. > > Although, the error about missing visual could be a server interaction > through glx. I don't know if you can get a backtrace on the 7500 > crash, but that would be valuable. > For the 7500 I'm unsure which kernel it was running, probably some late 2.6.37-rc. Again, when I've got the time I'll see if I can get a bt. ĸen -- das eine Mal als Tragödie, das andere Mal als Farce -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
