On Wed, Nov 01, 2017 at 08:58:32AM +0100, Gerd Hoffmann wrote: > Hi, > > > No one confirmed whether or not this new field needs to be added to > > the > > VMSTATE sections, and if so, how todo this in a back compatible > > manner.... > > Good question. Sky isn't falling if we don't send this over, worst > case is a minor keyboard glitch, affecting only the SysRq key, right? > > Annoying things like sticky modifier keys should not happen. So I'd > tend to simply leave things as-is.
Yeah, we would only gitch that key combo. It should be (practically) impossible to hit, because from the user input POV its just a single key, that internally we turn into a sequence of keys when injecting. > > When sending it over we need a subsection for it. Simplest way would > be to only send over the section in the unlikely case that the field is > non-zero. But this might fail migration to older qemu versions in case > someone hits a modifier key the wrong moment. Alternatively send the > sibsection only on new machine types using a compat property, but that > needs a bit more boilerplate. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
