On Fri, Jun 24, 2016 at 01:10:45PM +0300, Evgeny Yakovlev wrote: [...] > Thanks for all the comments! > Removing hyperv_* booleans sounds like the proper way forward however it > will break compatibility with how libvirt enables hyperv enlightenments.
It won't break any compatibility, if the property name in the feature_names arrays is the same as the previous boolean property name. The only difference is internal to QEMU: now the property will set a bit inside env->features instead of a boolean field in X86CPU. > I think we will now concentrate on the QOM feature-words and later get back > to reworking hv_* properties. You suggestions will be very helpful for that. OK! -- Eduardo
