I've recently downloaded ATI's new Radeon 8500 drivers for my 8500 128 Mb (non-LE).

Driver says RADEON R200 QL (Radeon 8500 LE) (LE/non-LE shouldn'tbe a problem I think).

Everything works great, many OpenGL apps no longer have pixel-delocations (and massive texture spikes/invisible areas). Unfortunately, there seems to be a problem in clearing the buffer??. Linux (SUSE 8.2) freezes if I exit many 3D games and apps, some won't even start, and TuxRacer freezes if any key is pressed on the 'press any key' initial screen. Sometimes, 3D apps exit with error stating a FireGL problem in clearing the buffer. There appear to be no other problems while 3D apps are running, only when being destroyed.


EXAMPLE - blender-sample (fails at start)
---------------------------
Created window
Mapped window
GHOST_EventPrinter:rocessEvent, time: 160, type: GHOST_kEventWindowSize
GHOST_EventPrinter:rocessEvent, time: 451, type: GHOST_kEventWindowActivate
GHOST_EventPrinter:rocessEvent, time: 451, type: GHOST_kEventWindowSize
FATAL: fglX11FreeBuffer: firegl_FreeBuffer() failed!
/usr/bin/blender-sample: line 2: 3476 Trace/breakpoint trap /usr/bin/blender /usr/share/doc/packages/blender/geeko.blend
--------------------------------------

When Linux OS freezes, no error messages are found in any logs (fail, warn, messages)

Anybody else with this issue?

flgrxconfig was default pretty much the whole way through, save for monitor settings.

Thanks,

-XE

Make sure you don't have the "newly supported" FSAA enabled.
I'm have no issues of this kind with them but I'm still using my old config [... with no reference to FSAA] so I dunno what gets set by default by fglrxconfig, this time around.

If not the case, you might need to get the 2.9.1x [ ...12 or 13] instead, I know SUSE had them packaged at one point.

I'll hit it with a few new flgrxconfig's tonight when it doesn't matter if the system goes down a few times, but I think I didn't set FSAA to anything.

I was able to avoid the error by telling XF86Config-4 to use an external AGPGART driver. However, everything now runs at a _crawl_ . Therefore, the problem is narrowed down to the internal AGPGART module in ATI's driver. I've read something similar in another post, but it was a side issue and not directly talked about. Some German page mentioned something about the most recent kernel fixing the issue??? However, I'm fairly sure my kernel is up to date. (kernelversion = 2.4)