The crash itself seems to point at a broken task structures tree while scheduling during a pipe read. Though this is rather cause not reason. Right now, nothing obvious strikes. There were a few Xen related patches waiting to come in through stable but I am not sure those would help here.
Meanwhile it would be good if we could gather one or two more crash stack traces along with "Xen version" grabbed from dmesg (which is the version of the hypervisor running on the host).
The crash itself seems to point at a broken task structures tree while scheduling during a pipe read. Though this is rather cause not reason. Right now, nothing obvious strikes. There were a few Xen related patches waiting to come in through stable but I am not sure those would help here.
Meanwhile it would be good if we could gather one or two more crash stack traces along with "Xen version" grabbed from dmesg (which is the version of the hypervisor running on the host).