Re: [PATCH] kvm: notify host when guest paniced

[Avi Kivity <avi@redhat.com>]

On 02/29/2012 12:31 PM, Wen Congyang wrote:
> At 02/29/2012 06:05 PM, Avi Kivity Wrote:
> > On 02/29/2012 11:58 AM, Daniel P. Berrange wrote:
> >>>
> >>> How about using a virtio-serial channel for this?  You can transfer any
> >>> amount of information (including the dump itself).
> >>
> >> When the guest OS has crashed, any dumps will be done from the host
> >> OS using libvirt's core dump mechanism. The guest OS isn't involved
> >> and is likely too dead to be of any use anyway. Likewise it is
> >> quite probably too dead to work a virtio-serial channel or any
> >> similarly complex device. We're really just after the simplest
> >> possible notification that the guest kernel has paniced.
> > 
> > If it's alive enough to panic, it's alive enough to kexec its kdump
> > kernel.  After that it can do anything.
> > 
> > Guest-internal dumps are more useful IMO that host-initiated dumps.  In
>
> Yes, guest-internal dump is better than host dump. But the user may not
> start guest-internal dump or guest-internal dump failed. So we need the
> following feature:
> 1. If the guest-internal dump does not work, the guest's status is 'crashed'.
>    And then the user does the host dump.
> 2. If the guest-internal dump is working, the guest's status should be
>    'dumping'. The user see this status and know the guest has paniced, and
>     the guest-internal dump is working.

I agree.  There is room for host dump, and we do want notification about
what the guest is doing.  The question is whether we should reuse
virtio-serial for guest-host communication in this case.  It's more
complicated, but allows us to avoid touching the hypervisor.

-- 
error compiling committee.c: too many arguments to function