On Fri, Feb 01, 2008 at 06:01:28PM +0100, Soren Hansen wrote: Sounds reasonable As the focus is shifting toward fully virtualized, the range of emulated hardware will grow. Hopefully underneath this will always be QEmu so these descriptions are likely to be shareable between Xen and KVM. Still we should not rush and try to design the XML description in an hypervisor agnostic way. For example 'virtio' 'ide' 'scsi' are generic terms, you could have different SCSI emulated, or different IDE ... For NICs QEmu can already emulate mutiple kind of cards. I'm not sure there is a choice for the graphic card yet. It's unclear to me how specific those descriptions must be, it looks like a trade-off between being generic (i.e. having similar kind of configs at the XML level for different hypervisor) and more flexibility per engine. On a similar note we also need to add sound card support https://bugzilla.redhat.com/show_bug.cgi?id=236800 and maybe allowing generic USB devices would be nice too. Daniel -- Red Hat Virtualization group http://redhat.com/virtualization/ Daniel Veillard | virtualization library http://libvirt.org/ veillard(a)redhat.com | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/