Bug 122609
Summary: | (FUTEX) Multiple kernel freezes during normal use | ||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Neil Thompson <abraxis> | ||||||||||||||
Component: | kernel | Assignee: | Dave Jones <davej> | ||||||||||||||
Status: | CLOSED NOTABUG | QA Contact: | Brian Brock <bbrock> | ||||||||||||||
Severity: | high | Docs Contact: | |||||||||||||||
Priority: | medium | ||||||||||||||||
Version: | 2 | CC: | alan, pfrields, richard.cunningham, ronny-rhbugzilla, zenczykowski | ||||||||||||||
Target Milestone: | --- | ||||||||||||||||
Target Release: | --- | ||||||||||||||||
Hardware: | All | ||||||||||||||||
OS: | Linux | ||||||||||||||||
Whiteboard: | |||||||||||||||||
Fixed In Version: | Doc Type: | Bug Fix | |||||||||||||||
Doc Text: | Story Points: | --- | |||||||||||||||
Clone Of: | Environment: | ||||||||||||||||
Last Closed: | 2004-11-29 16:26:22 UTC | Type: | --- | ||||||||||||||
Regression: | --- | Mount Type: | --- | ||||||||||||||
Documentation: | --- | CRM: | |||||||||||||||
Verified Versions: | Category: | --- | |||||||||||||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||||||||||
Cloudforms Team: | --- | Target Upstream Version: | |||||||||||||||
Embargoed: | |||||||||||||||||
Attachments: |
|
Description
Neil Thompson
2004-05-06 09:31:22 UTC
Created attachment 100035 [details]
Oopses and log messages from the freezes I've had
Stupid me, in between all my crashes I got the version wrong - I was actually on 1.350, not 1.250 - makes a bit of a difference. I have just upgraded to 1.351, we'll see what happens. Several more freezes on 1.351 - nothing in the logs. Tuxracer kills the box every time. Same problem here with kernel 2.6.5-1.349 and 2.6.5-1.351. The system freezes after different time. All the same whether KDE or XFce is used. In the log files nothing is to be found. SELinux is turned off. Hardware: Board: MSI K7N2 CPU: AMD Athlon(tm) XP 2000+ RAM: 512 MB SD-DDR If further information should be necessary, ask me simply. Is it an ACPI-problem? Created attachment 100107 [details]
Another oops, this time on 1.353
Still freezing on 1.353. Please note that the machine is completely stable
under fairly heavy load (MailScanner with clamav and spamassassin with a good
deal of email coming in) as long as there is no interaction with X.
Created attachment 100141 [details]
Another oops
X restarted with this oops - no freeze
Created attachment 100142 [details]
oops on freeze kernel 1.358
Another oops - immediately after previous - with freeze - on 1.358
Created attachment 100354 [details]
another oops on 2.6.5-1.358
I have now disabled selinux completely - still freezing/crashing regularly.
Sometimes with a log entry and sometimes without. Here's the latest. I'm
starting to get a little despondent - this feels like kernel 0.99.<mumble>,
which was the first one I used.
Is anyone listening?
--- FC2: Sun Java 1.4.2_04 futex freeze --- P3 1GHz, 512MB, freshly upgraded to FC2, kernel 2.6.5-1.358 Sun Java 1.4.2_04 freezes entire kernel, alt+sysrq displays info on screen but sync and unmount and poweroff and term and kill don't do anything. Note: no oops, just freeze! appears to be a problem (according to tasks) with java calling sys_futex and looping in restoresigcontext (according to showpc). Need more data? Does this occur if you add and run the i586 kernel ? I've added the i586 kernel, rebooted to it and banged on the box for about an hour - no freezing yet. Another datum is that I've just built a clone of the box that's failing and the new box is completely stable, so it is probably something to do with the old box - although the old box was stable under FC1. The java case is a specific bug currently being tracked. A possible fix is being looked at now ready for the first FC2 errata kernel. The other one is more curious. The i586 kernel changes a couple of things - it means that memory isnt hit as hard (no fancy memory copying functions for Athlon), and it means that certain kinds of device driver error that were not a problem in 2.4 will also continue to work (including the java bug...). I'd thus be interested in a full list of hardware on the machine (lspci -vxx and lsmod attachment would be great) - and if you too are running the Sun java tools. Also to eliminate one other thing - does it work reliably using the vesa X server ? Created attachment 100453 [details]
lspci and lsmod as requested
I've managed to trigger the fault on the i586 kernel (tuxracer will do it every
time). Now that I have a second machine to play with, I've been able to ssh
into the affected machine - it's not frozen at all. It's just X that has
stopped working. Is there anything I can get from the box while it is in this
state that would help?
I've tested a lot (kernel-2.6.5-1.358 and a vanilla kernel-2.6.6). I think it's a problem with KDE and X.org. If my box is running in runlevel 3 everything is fine. A system load greater than 2 is no problem. In runlevel 5 (with a XFce-session) also everything is fine, but when I'm starting a KDE application, like KMail, the system freezes. By way of trial the x-server was also started again, before KMail or konqueror was started. The same result - the sysytem freezes. Gnome applications don't cause problems. I don't run KDE at all, just Gnome, so that's out. OK, XFce, (Gnome) and the kernel-2.6.5-1.358 is enough - the system will freeze after a few minutes. In the past days the system ran only with the vanilla kernel. With this there were only in connection with KDE applications problems. Which informations should I supply, so that the problem can be solved as soon as possible? any better with the 2.6.9 errata kernel? I eventually gave up and installed XP on the box for my daughter - just before the 2.6.9 kernel came out. That didn't work too well either, and I replaced the motherboard. Before that I had managed to reduce the occurrence of the problem by moving the USB mouse and keyboard around on the various USB connectors on the box. Removing the USB Keyboard completely reduced the freezes to about once a day - provided I could get the box up in the first place. The USB keyboard also caused intermittant problems with the new motherboard. I'd probably put it down to a flaky hardware combination. |