Progress: * UM-2 [QEMU upstream maintainership] - looking at some late-breaking patches for various more-or-less 9.0-worthy bugs - tagged rc3; looks like we'll need an rc4, though - back to looking at cleanups and refactorings of our reset handling: + add support for a new reset type RESET_TYPE_SNAPSHOT_LOAD so we can use a new-style reset method for the kind of "load new RNG seed into memory" task we can currently only do with qemu_register_reset_nosnapshotload() or with a Machine reset method + pass the reset type into all reset phase methods, not just 'enter' + experimenting with making Machine a subclass of Device, which should let us clean up some inconsistencies (like Machine having its own reset handling that's different from everything else)
-- PMM