Description of problem: After running an arbitrary workload involving network traffic for some time (1-2 days), a xen guest running the 2.6.9-67 x86_64 xenU kernel locks up with both vcpu's spinning at 100%. Version-Release number of selected component (if applicable): kernel-2.6.9-67 How reproducible: Reproduces after running a test workload (involving network traffic) for 1-2 days, but not consistently. Steps to Reproduce: 1. Run a test workload involving network traffic 2. monitor cpu usage of guest 3. wait... until guest cpu usage goes to 100% Actual results: Guest kernel spins on both vcpu's Expected results: Guest kernel doesn't spin Additional info: The problem is due to a race between the scheduler and network interrupts. On one vcpu, the scheduler takes the runqueue spinlock of the other vcpu to schedule a process, and attempts to lock mm_unpinned_lock. On the other vcpu, another process is holding mm_unpinned_lock (because it is starting or exiting), and is interrupted by a network interrupt. The network interrupt handler attempts to wake up the same process that the first vcpu is trying to schedule, and will try to get the runqueue spinlock that the first vcpu is already holding. I was not able to obtain a full kernel stack from the interrupt, but do have kernel stacks of the tasks on the vcpu's, if needed.
Created attachment 309000 [details] fix for mm_unpinned_lock / runq deadlock
The patch looks good to me.
Updating PM score.
This request was evaluated by Red Hat Product Management for inclusion in a Red Hat Enterprise Linux maintenance release. Product Management has requested further review of this request by Red Hat Engineering, for potential inclusion in a Red Hat Enterprise Linux Update release for currently deployed products. This request is not yet committed for inclusion in an Update release.
OK. I posted this patch to upstream Xen, and they pointed out that not only does 2.6.18 probably not have this issue, but also that the patch isn't entirely sufficient, since you might still deadlock with the mm->page_table_lock. Upstream 2.6.18 does this a different way (which the RHEL-5 kernel already has); I've backported the changes, and built a test kernel available at http://people.redhat.com/clalance/bz450953/. Can the reporter please download the test kernel, try it out, and report back test results? Thank you, Chris Lalancette
Created attachment 324399 [details] Backport of upstream xen-unstable c/s 10343 and 10995 This is a backport of xen-unstable c/s 10343 and 10995, that I believe will fix the issue. Instead of changing the locking, this actually defers the mm_pin time until activate_mm, which means that we should avoid the deadlock in the first place.
Chris, the upstream patches are indeed better. I'll go NACK Larry's patches :)
Committed in 78.20.EL . RPMS are available at http://people.redhat.com/vgoyal/rhel4/
Any updates here? Has this issue been resolved in the RHEL 4.8 Beta? later kernel?
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHSA-2009-1024.html