On Fri, Aug 29, 2003 at 08:03:18PM -0400, Carl Nygard wrote: > On Fri, 2003-08-29 at 20:00, Michel D�nzer wrote: > > On Sat, 2003-08-30 at 00:46, Warren Turkal wrote: > > > I am using XFree86 4.3.0 on a Debian Sid system with kernel 2.6.0-test4. > > > Before I file a formal bug, I wanted to verify that it wasn't already fixed > > > in the CVS HEAD. > > > > > > I have a Radeon M7. The XServer locks up the machine every now and then. > > > When it locks up, the machine will still repsond to ACPI events (eg. acpid > > > will still shutdown the machine when I press the power putton). I am not > > > sure if the system is pingable at that point. The locking up is not changed > > > if I add option "NoAccel" "true" to the driver configuration clause in my > > > config file. > > > > In that case it can hardly be an X problem. Sounds rather like the > > kernel (anything interesting in its output? Does the same thing happen > > with 2.4 kernels?) or hardware. > > I just got a Compaq Presario x1000, with roughly the same setup (RH9, > 2.6.0-test4, Radeon Mobility 9200), and I've seen the exact same lockup > problems. Thing is, I've seen it even worse when I boot to XP, so I > wouldn't think it was necessarily related to X or even Linux (at least > in my case, don't know your hardware). > > One check, lots of times if X goes screwey, you can still telnet to your > machine from across the network, and kill X server manually, that should > get gdm/X to restart, if that indeed is the problem. Otherwise, you're > looking at kernel (or hardware) problems.
BTW, i have a case where the glint driver freezes the server for a permedia2 chip. gdb debugging shows that it is busy waiting in the wait for free space routine of the accel engine, but killing the server and restarting it doesn't improve the situation, only a reset of the box will do. In this case, would a hardware problem also be the culprit ? The code in question has hardly changed since earlier versions of X, and it used to work well before. Overheating, dying power supply or graphic chip would be my first diagnostic here, altough i am waiting for the user to provide me output with more debugging info, in order to see which accel hook was called prior to the problem. Friendly, Sven Luther _______________________________________________ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
