Xenner

Xen, as one of the most successful virtualization solutions out there, has a lot of virtual machines running on top of its PV interface. If someone wants to move away from either Xen in general or Xen's PV interface to an HVM instance, there's significant change inside the guest required to make things work.

So what if we were able to run the Xen PV kernel directly inside an HVM instance?

Xenner is a small shim-kernel that translates Xen PV hypercalls into calls that real hardware understands. It also contains of a small portion of code in QEMU that exposes the Xen PV devices to the guest, so we don't lose performance by translating device access from Xen PV devices to other emulated devices. This way it's possible to run Xen PV guests in an HVM context, without the guest realizing it.

Xenner has been around for a while and had its ups and downs, but never managed to get fully upstreamed. There are also a few implementation detail questions left open. The big question really is: Is anyone seriously interested in maintaining and using this code?

Required attendees
Stefano Stabellini

Topic Lead: Alexander Graf
Alexander has been a steady and long time contributor to the QEMU and KVM projects. He maintains the PowerPC and s390x parts of QEMU as well as the PowerPC port of KVM. He tends to become active, whenever areas seem weird enough for nobody else to touch them, such as nested virtualization, mac os virtualization or ahci. Recently, he has also been involved in kicking off openSUSE for ARM. His motto is なんとかなる.