Paolo Bonzini <pbonz...@redhat.com> writes: > Il 14/12/2012 01:57, Rusty Russell ha scritto: >> With the new code we only need the head from that structure. > > We also need to do again all validation of the elements if we fetch it > back from the data. Sometimes the parsed data is saved elsewhere (e.g. > in a SCSIRequest struct that is serialized by the SCSI subsystem) and > that data may be inconsistent with whatever you read from guest memory. > It's a can of worms. > >>> I'm certainly in favor of cleaning up the savevm format and probably >>> leaving the existing load/save functions as-is for legacy purposes. >>> I'll leave that as an exercise for someone else though :-) >> >> What is the rule about new versions? Can we introduce a new save >> version at any time, or only at major qemu version changes? > > Any time, but we provide a backwards-compatible loader for older > versions.
Well.. if we can avoid bumping the version, we should. Bumping the version breaks migration from new -> old so it's preferrable that we don't do it unless it's absolutely required. Regards, Anthony Liguori > > Paolo
