On 07/18/2014 10:11 AM, Peter Krempa wrote:
Add event reason to support transitions to PM suspended state from snapshot revert operations. --- examples/object-events/event-test.c | 3 +++ include/libvirt/libvirt.h.in | 3 +++ tools/virsh-domain.c | 3 ++- 3 files changed, 8 insertions(+), 1 deletion(-)
+++ b/include/libvirt/libvirt.h.in @@ -3693,6 +3693,9 @@ typedef enum { typedef enum { VIR_DOMAIN_EVENT_PMSUSPENDED_MEMORY = 0, /* Guest was PM suspended to memory */ VIR_DOMAIN_EVENT_PMSUSPENDED_DISK = 1, /* Guest was PM suspended to disk */ + VIR_DOMAIN_EVENT_PMSUSPENDED_FROM_SNAPSHOT = 2, /* Guest changed state to + PM suspended due to + snapshot operation */
Umm, is this even possible with qemu? My understanding is that qemu is unable to migrate S3 status, and therefore, the moment you migrate to another machine or to a file, the receiving end of the migration will act as if the guest had a PM wakeup event and is no longer in S3. If you are proposing immediately re-suspending the guest into S3 after resuming it, that's wrong - we should only support states that can be entered immediately, and not states that require an unknown amount of guest CPU cycles. -- Eric Blake eblake redhat com +1-919-301-3266 Libvirt virtualization library http://libvirt.org