Re: [Xen-devel] kernel panic with no call trace

[Andrew Cooper <andrew.cooper3@xxxxxxxxxx>]

On 06/09/2017 03:39, Minjun Hong wrote:
> Hello~~
> I'm struggling to resolve a kernel panic problem during developing
> scheduler code.
> But I have not made any progress since I can not get any meaningful
> information from the serial log.
> When the panic occurred, always there is no call trace and only panic
> notification like following:
>
> (XEN)
> (XEN) ****************************************
> (XEN) Panic on CPU 0:
> (XEN) cpu:20, vcpu:20 in csched_schedule(1891)
> (XEN) cpu:21, vcpu:21 in csched_schedule(1891)
> (XEN) cpumask_test_cpu(cpu, prv->in_cosched) in csched_schedule(1907)
> (XEN) cpumask_test_cpu(cpu, prv->in_cosched) in csched_schedule(1907)
> (XEN) cpumask_test_cpu(cpu, prv->in_cosched) in csched_schedule(1907)
> (XEN) cpumask_test_cpu(cpu, prv->in_cosched) in csched_schedule(1907)
> (XEN) FATAL PAGE FAULT
> (XEN) [error_code=0000]
> (XEN) Faulting linear address: ffff830078efcc98
> (XEN) ****************************************
> (XEN)
> (XEN) Reboot in five seconds...
>
> I'm using Xen-4.5.0 on my server having 2 Intel Xeon E5-2620 v4 cpus,
> 128 GB RAM(16 GB DDR4 * 4) and 1 TB HDD and, using Ubuntu 14.04 LTS.
>
> Is there any method to make the call trace show up or 
> when there is no call trace, please tell me from where I should start
> to debug.
>
> Thanks in advance and I wait for your comments.

There is a call trace, but as you've clearly added printk()'s to the
scheduler, the calltrace will be getting lost in the spew of logging
beforehand.

From what you've printed, you've fallen over a bad pointer which isn't
present, although the offset into the directmap does look semi
plausible.  Either way, you've got memory corruption of some kind.

~Andrew

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel