On 11/03/2010 03:57 AM, Daniel P. Berrange wrote:
On Tue, Nov 02, 2010 at 04:43:06PM -0600, Eric Blake wrote:
> On 10/25/2010 09:28 AM, Daniel P. Berrange wrote:
>> On Fri, Oct 22, 2010 at 09:14:22AM -0500, Ryan Harper wrote:
>>> Currently libvirt doesn't confirm whether the guest has responded to the
>>> disk removal request. In some cases this can leave the guest with
>>> continued access to the device while the mgmt layer believes that it has
>>> been removed. With a recent qemu monitor command[1] we can
>>> deterministically revoke a guests access to the disk (on the QEMU side)
>>> to ensure no futher access is permitted.
>>>
>> This looks good to me now, ACK, pending my testing with old
QEMU
>> versions under the TCK
>
> Is there any further TCK testing needed for this before we push, or
> should I go ahead and push it now?
I'm not clear that they have accepted the drive_unplug command into QEMU
GIT yet, since there is still a huge ongoing discussion about it on the
qemu mailing list, and the QEMU git server has been unavailable for days
now :-( Until I can confirm and test this, we'll have to wait.
Confirmed that this is in upstream qemu (commit 9063f81415f35), so I'm
going ahead and applying this.
I've got some followups in mind (using virAsprintf instead of snprintf
and a stack-allocation of PATH_MAX bytes), but they can be a separate patch.
--
Eric Blake eblake(a)redhat.com +1-801-349-2682
Libvirt virtualization library
http://libvirt.org