Am 30.01.2012 16:55, schrieb Paolo Bonzini: > On 01/30/2012 04:43 PM, Kevin Wolf wrote: >>>>>> >>>>>> If we had force=true in the initial eject command, bdrv_close is called, >>>>>> which in turn goes through bdrv_dev_change_media_cb where an event is >>>>>> emitted. >>>> >>>> Can't this race with the guest eject? >> Can't see how, there's nothing asynchronous in the path. > > The button press is sent down to the guest even with -f.
Right, but there's no race. The guest will always be later. Kevin