Description of problem: Customer is getting kernel panics due to list_del being called on an already freed list entry. The reason appears to be that the _uid_ses keyring is being cleaned up twice. Version-Release number of selected component (if applicable): kernel 2.6.18-8 How reproducible: Customer sees this regularly when there are a large number of logins and logouts.
Have there been panics like this in the past? What's a good technique for finding the culprit in this case? I've tried systemtap but you can't guarantee the integrity of a disk log after a panic, and getting the stap logs out of a core file has proved ugly.
Is there a vmlinux/vmcore pair that I can look at?