* Paolo Bonzini (pbonz...@redhat.com) wrote: > On 4/27/22 17:16, Dr. David Alan Gilbert wrote: > > * Paolo Bonzini (pbonz...@redhat.com) wrote: > > > On 4/27/22 14:34, Dr. David Alan Gilbert wrote: > > > > If I specify a 'vm' it's not obvious to me whether I'd get NICs and > > > > block devices in the future? > > > > > > VM would not get those (it's global statistics), but the size could > > > balloon > > > if you specify no target at all. > > > > > > > Adding a syntax for 'all' into the vcpus list would fix that? > > > > > > I don't like having special syntax. The current QAPI just doesn't filter > > > what is not in the arguments. > > > > Is there a object that represents the set of all vcpus? > > No.
If it was easy to create one then you could remove all the special casing of vCPUs/VM target? (It feels really like you should call a 'stats' method on the target) > > > Yes, those would have different providers. But a single target can > > > support > > > multiple providers. > > > > Is that just for different implementations - kvm/hcf/tcg etc or do you > > envisage multiple providers on an object in a running VM? > > I think multiple providers are possible for a single object, for example a > device could expose both PCI (how many MSIs, etc.) and SCSI (how many > commands sent/succeeded/failed) statistics. Yeh fair enough. Dave > Paolo > -- Dr. David Alan Gilbert / dgilb...@redhat.com / Manchester, UK
