Red Hat Bugzilla – Bug 51893
Random infrequent locks on Dell Lattitude CPiA
Last modified: 2007-04-18 12:35:58 EDT
Description of Problem: System (Dell Latitude CPiA) occassionally locks (could be Xserver locks up. No evidence that the kernel is actually locked. Have not been able to use SysRq keys, but do not believe the sysctl is activated. Hard drive activity indicates that the system is alive) Version-Release number of selected component (if applicable): Kernel and XFree86 from with Roswell. How Reproducible: not very. Happens maybe once a day... Sorry for the low level of usefull information. Video card is the NeoMagic 256AV. The video card shares hardware with the sound card, but the sound card part is not activated. Have not been able to verify its just the X server that locks, so I can't count out kernel bugs yet.
Sysrq is enabled (ie, alt-sysrq should cause a sync), and I have not yet been able to cause a sync/umount/reboot sequence when the laptop locks. I still have no idea how to reproduce this.
I have problems with APM shutting down the PC randomly, please see the APM & Dell C600 bug i have created. Is this happening when you resume from apm shutting down your monitor? I have a problem that when i resume a few minutes later the whole PC freezes and i have to hardboot
I don't tend to suspend the system much, so I have not noticed a correlation... But the last time it locked I had indeed suspended it recently (to remove from the docking station) After reading your bug report (<a href=http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=51494>51494</a>), I can believe that the problems might be different symptoms of the same bug. Will try a newer kernel when I see it in Rawhide. (Note: someone might want to come up with a "rawer than rawhide" tree that's updated daily is is flat guarenteed to not work? Its kinda annoying to wait for two weeks after being told that there is a fix to test it.)
I can pretty offically say that its not an X problem now. Happened this afternoon on a fresh reboot. APM should not have been involved either, as I had never suspended the machine, although the screen saver (DPMS?) had kicked in. Machine locked hard. Network connections timed out, so its not just the X server. Did not remember to try sysrq. Possibly relevant: I was doing some large data copies via the network. Ethernet card was the Xircom Cardbus ethernet+modem (xircom_cb) which appears to be kinda balky with the Roswell kernel. If someone could change the component to "kernel"...
Reassigned bug to kernel maintainer.
I caused it to happen two times in realitively quick succession. From another computer, did a large file (~1G) transfer to the system that has problems (catbert) ( cd /path/to/mp3collection ; tar -cf - . | ssh catbert tar -xf -) Died pretty quickly. Did it again, died a little later, but still died. Sysrq sorta works. B reboots, no activity from either S, U, or K. Both times, the display was not blanked. Went back to catbert and switched from X to a text VT. Didn't die, finished up the transfer. (switched to file-by-file instead of tar to make recovery easier) Changed X to not use DPMS, tried again. It locked. When I went back, the screen was blanked and sysrq did not work. Switched back to a text VT, and coppied the whole collection over again. No lock So: 1) Not DPMS related. 2) Might be X related somehow. (NM256AV, sound drivers currently loaded) 3) Large network useage (Xircom Cardbus card, talking to another machine on the same hub) might be related. 4) Sysrq sometimes partially works. (B). Sometimes doesn't at all. Never fully works. (S,U) Still no good idea what causes it.
No locks today with Roswell2 kernel. I'm gonna close this for now, and reopen it later if it crashes on me again.