On Thu, 17 Sep 2015 12:39:31 +0200 Greg Kurz <gk...@linux.vnet.ibm.com> wrote: > On Thu, 17 Sep 2015 11:06:01 +0200 > Cornelia Huck <cornelia.h...@de.ibm.com> wrote:
> > +Devices need to be careful in their state processing during load: The > > +load_device() procedure is invoked by the core before subsections have > > +been loaded. Any code that depends on information transmitted in > > subsections > > +therefore has to be invoked in the device's load() function _after_ > > +virtio_load() returned (like e.g. code depending on features). > > + > > From a VMState standpoint, such code would land in a post-load callback most > of > the time. Would that help comprehension if we introduce a virtio_post_load() > function ? It might. I think the main problem is that we need to do some stuff post-load, and it's not really obvious what it is.
