> Am Dienstag, 09. Dezember 2003 02:22 schrieb Michel Dänzer: >> On Fri, 2003-12-05 at 00:24, [EMAIL PROTECTED] wrote: >> > I am trying to get a Radeon 9000 pro to work using the latest DRI >> CVS tree, but encountering a few problems. >> > >> > The system in question is a pair of opterons running on a Tyan S2885 >> board, with Suse 9.0 Professional (AMD64) installed and the latest >> 2.4.21-149 kernel. >> > >> > I get the same results whether building/installing the whole tree, >> or building/installing just the radeon.o module. >> > >> > After manually modprobing agpgart, and starting X, glxinfo reports >> that direct rendering is enabled, and I can run either multiple GL >> apps or instances of apps that do not use textures (like glxgears, >> or bubble3d from the xscreensaver modules), or a single instance of >> a GL app that uses textures. >> > >> > Attempting to run more than one GL app with at least one of the apps >> using textures results in the apps freezing up, and the keyboard >> locking. >> > The mouse cursor still responds, but not the buttons. >> >> If you have network access to the box, and you can still log in after >> this happens, which process hogs the CPU, and what happens when you >> kill it?
I've tried. X is sucking up all of one CPU. It seems that the app(s) using GL textures has/have died. Everything is veeeeeeeeery slow. If I run "top", top itself claims to take up > 50% of the other CPU. As this is a dual opteron system, it shouldn't be lack of sheer horsepower. "kill -9 <pid of X>" just zombifies X. Nothing I can kill will give me back console access. If I reboot the system from my ssh login (i.e. - init 6), there is a long pause, and then the box emits a continual loud "BEEEEEEEEEEEEEEEEEEEEEEEEEP" until I power cycle it. > Appear on all (?) SMP systems so far for some months... > Even Quake3 didn't run in SMP mode (quake3-smp) anymore (after Ian's > context rearrangements). > > I havent't had enough time and a broken system disks to examine even > further in the past weeks. > > dual Athlon MP 1900+ > R200, 64 MB > > Please try two "gears", then "gears" and "ipers" and then two instances > of "ipers". I can't completely verify this diagnosis. If I compile/run a uniprocessor kernel (same config otherwise as the SMP one), I get what so far appears to be a longer (and still varying) but finite amount of time before the same behavior appears. So far I have been able to start a few more texturing GL apps before the hang occurs, but it still occurs. Still, nothing obvious turns up in /var/log/messages or XFree86.0.log or ..X.err. With the uniprocessor setup, I have been able to kill X after logging in remotely, as well as whatever GL app may/may not be sucking up process time. So far, (3 tries; I'll keep collecting statistical data.) I've been able to restart the machine once from the remote login. The other times I tried to restart it remotely, it "froze" harder and refused any further connection attempts. ------------- [EMAIL PROTECTED] PENGUIN COMPUTING www.penguincomputing.com ------------------------------------------------------- This SF.net email is sponsored by: IBM Linux Tutorials. Become an expert in LINUX or just sharpen your skills. Sign up for IBM's Free Linux Tutorials. Learn everything from the bash shell to sys admin. Click now! http://ads.osdn.com/?ad_id78&alloc_id371&op=click _______________________________________________ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
