Bugzilla – Bug 132271
kernel null dereference with/by X
Last modified: 2005-12-05 10:15:55 UTC
Not sure whether I should bugzilla this, but I had this kernel problem never before I upgraded to 10.0 recently, and therefore doubt it's a hardware issue. If others see it too the info might help you. $USER was browsing gmx webmail with konqueror, and logged out from gmx. The logout completed, but the whole desktop froze. Question whether gmx has anything to do with it. The rest of the box kept running fine, but no keyboard/mouse/console. Ssh login allowed to kill all $USER processes except X and (I think) ksmserver), but didn't unlock the screen (not surprisingly, with X a zombie). Hardware is Duron 600, with 00:00.0 Host bridge: Silicon Integrated Systems [SiS] 730 Host (rev 02) 00:00.1 IDE interface: Silicon Integrated Systems [SiS] 5513 [IDE] (rev d0) 00:01.0 ISA bridge: Silicon Integrated Systems [SiS] SiS85C503/5513 (LPC Bridge) 00:01.1 Ethernet controller: Silicon Integrated Systems [SiS] SiS900 PCI Fast Ethernet (rev 82) 00:02.0 PCI bridge: Silicon Integrated Systems [SiS] Virtual PCI-to-PCI bridge (AGP) 01:00.0 VGA compatible controller: Silicon Integrated Systems [SiS] 630/730 PCI/AGP VGA Display Adapter (rev 31) syslog attached
Created attachment 56440 [details] syslog
It seems the SIS driver tries to grab a huge chunk of memory, which fails, but the driver doesn't check for a NULL return value and happily dereferences this NULL pointer later.
Could you please attach /etc/X11/xorg.conf and /var/log/Xorg.0.log? Thanks.
Created attachment 57588 [details] /etc/X11/xorg.conf
Created attachment 57589 [details] /var/log/Xorg.0.log Not the log file from when the crash happened, that got overwritten ages ago, but I hope it'll give you the same info. Crash hasn't happened again so far.
Ok. The next time it happens, reboot and attach /var/log/Xorg.0.log.old or reboot into runlevel 3 and attach /var/log/Xorg.0.log.
Created attachment 58622 [details] Fix If there isn't anything else after this function broken, too this patch should fix the problem.
The machine is in daily use, but I haven't had another crash since the first, which makes it difficult to test whether the patch fixes the problem.
From my POV, okay to apply to SLES9 branch as well if it has that issue, too.
Takashi, can you please apply this patch to the 10.0 branch?
I'm working on it. Linus' git tree doesn't have this fix, too. Shouldn't we submit to mainstream?
definitely submit to mainstream too.
OK, submitted to LKML. Both 10.0 and HEAD include the patch now.
Thanks. Takashi, could you forward the email to LKML to Dave Airlie <airlied@linux.ie>? He's responsible for the DRM drivers. Just to make sure it won't get lost. I think we can close this now.
OK, let's close it. If you still have a similar problem, please reopen.
Had a lockup with the same symptoms for the second time now, but absolutely nothing in syslog, and a normal Xorg.0.log.old. Perhaps because I pushed the reset button instead of trying to shut the system down cleanly?