Re: [Xen-devel] crash in nvmx_vcpu_destroy

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 21.02.13 at 11:49, Olaf Hering <olaf@xxxxxxxxx> wrote:
> On Wed, Feb 20, Olaf Hering wrote:
> 
>> (XEN) Xen call trace:
>> (XEN)    [<ffff82c4c01dd197>] nvmx_vcpu_destroy+0xb7/0x150
>> (XEN)    [<ffff82c4c01b5ce9>] hvm_vcpu_destroy+0x9/0x40
> 
> For some reason nestedhvm_vcpu_destroy is not in the backtrace.

Tail call optimization?

> And its
> not clear why nestedhvm_vcpu_destroy calls into nvmx_vcpu_destroy
> anyway. nestedhvm is not in the config file.

nHVM can be turned on and off on a domain, so cleanup is needed
unconditionally. This was also the reason why 26507:4f53ddbee940
had to revert 26503:69398345c10e.

Is this perhaps dying on the uninitialized list head? If so, I'm about
to submit a patch that ought to deal with that (albeit its original
purpose is another one).

Jan


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