On 06/08/2012 02:04 AM, Michal Privoznik wrote: > This feature has been requested for a very long time. However, > we had to wait for guest agent to obtain reliable results as > user might create totally different structure of interfaces than > seen from outside (e.g. bonding, virtual interfaces, etc.). > That's the main reason why sniffing for domain traffic can > return bogus results. Fortunately, qemu guest agent implement > requested part for a while so nothing holds us back anymore. How hard would it be to wire this API up to _also_ have the option of using our nwfilter IP learning code (first packet detection mode has been here for a while, and DHCP snooping mode was just added)? Use of the flags parameter should make it possible to force which method we attempt, use of flags==0 chooses the best method possible (GA if present, otherwise fall back to nwfilter IP learning). > > This API is called virDomainInterfacesAddresses (okay, maybe > too many plurals) and returns a dynamically allocated array > of virDomainInterface struct. The great disadvantage once > this gets released, it's written in stone and we cannot change > or add an item into it. Therefore we might add a padding into > it - something like reserved for future use. On the other hand, > everything important is already there - what else we will want > to add? :) But that's why we invented virTypedParameter. If you would return an allocated array of virTypedParameter instead of a hard-coded virDomainInterface struct, then you have the flexibility to add new named parameters down the road without ABI concerns. I haven't looked closely at your proposal yet, but just reading this one paragraph makes me think we need to rework it into using typed parameters. -- Eric Blake eblake@xxxxxxxxxx +1-919-301-3266 Libvirt virtualization library http://libvirt.org
Attachment:
signature.asc
Description: OpenPGP digital signature
-- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list