On Wed, Sep 17, 2008 at 09:04:35PM +0200, Ian FREISLICH wrote: > Kostik Belousov wrote: > > On Fri, Sep 12, 2008 at 02:46:39PM +0200, Ian FREISLICH wrote: > > > Interesting, I have the same (card/chip) and same problem: > > >=20 > > > (--) PCI:*(2:0:0) ATI Technologies Inc RV370 5B60 [Radeon X300 (PCIE)] re= > > v 0, Mem _at_ 0xd0000000/27, 0xdfff0000/16, I/O _at_ 0xe800/8, BIOS _at_ 0xdffc0000/= > > 17 > > > (--) PCI: (2:0:1) ATI Technologies Inc RV370 [Radeon X300SE] rev 0, Mem _at_= > > 0xdffe0000/16 > > >=20 > > > in my -STABLE box which is also plagued by these spurious freezes. > > > They always happen when logging out or quitting the window manager. > > > Yesterday the mouse pointer movd but very slowly. The Xorg process > > > using 100% on one CPU and unkillable. In the past, it's locked up > > > the machine entirely, but this time everything else was working. > > > I still had to power cycle it to reboot though. > > >=20 > > > I'll try to get a ktrace next time it happens. > > > > This is quite similar to my problem. I expect your X server to spend time > > in kernel. Please, use the procstat -k to get the kernel stack of the > > process. > > [brane] ~ # procstat -k 1593 > PID TID COMM TDNAME KSTACK > 1593 100062 Xorg initial thread mi_switch ast Xfast_syscall > [brane] ~ # procstat -k 1593 > PID TID COMM TDNAME KSTACK > 1593 100062 Xorg initial thread mi_switch turnstile_wait _mtx_lock_sleep giant_ioctl devfs_ioctl_f kern_ioctl ioctl syscall Xfast_syscall My backtrace is different. Anyway, as I was told, we should wait for 1.5 server and updated Mesa/libGL that could contain the fix.
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:35 UTC