On 02/27/2013 03:30 AM, Michal Privoznik wrote:
On 27.02.2013 00:22, Eric Blake wrote:
> On 02/26/2013 04:02 AM, Michal Privoznik wrote:
>> Currently, qemuDomainShutdownFlags() chooses the agent method of
>> shutdown whenever the agent is configured. However, this
>> assumption is not enough as the guest agent may be unresponsive
>> at the moment. So unless guest agent method has been explicitly
>> requested, we should fall back to the ACPI method.
>> ---
>>
>> diff to v1:
>> - Rework some conditions as Eric suggested in v1
>>
>> src/qemu/qemu_driver.c | 38 ++++++++++++++++++++++----------------
>> 1 file changed, 22 insertions(+), 16 deletions(-)
>
> ACK.
>
Do you think this one is safe to push now?
It is a bug fix and I think the code is well understood. However, the
issue it's fixing doesn't seem to be so usual to hit. Otherwise there
would be a bug report much sooner than 2012-12-22, right?
https://bugzilla.redhat.com/show_bug.cgi?id=889635
I've hit the bug myself, but never filed a BZ - it's quite annoying that
the shutdown button in virt-manager fails to work if you have the guest
agent wired up in the host XML, but not installed and running in the
guest, at which point you can no longer shut down the guest using
virt-manager. Yes, I think this is safe to push for 1.0.3.
--
Eric Blake eblake redhat com +1-919-301-3266
Libvirt virtualization library
http://libvirt.org