Am 16.11.2012 07:22, schrieb Stefan Hajnoczi: > On Thu, Nov 15, 2012 at 10:08 PM, Anthony Liguori <aligu...@us.ibm.com> wrote: >> Stefan Hajnoczi <stefa...@redhat.com> writes: >> >>> The virtio-blk-data-plane feature is easy to integrate into >>> hw/virtio-blk.c. The data plane can be started and stopped similar to >>> vhost-net. >>> >>> Users can take advantage of the virtio-blk-data-plane feature using the >>> new -device virtio-blk-pci,x-data-plane=on property. >> >> I don't think this should be a property of virtio-blk-pci but rather a >> separate device. > > The hw/virtio-blk.c code still needs to be used since > hw/dataplane/virtio-blk.c is only a subset of virtio-blk. > > So we're talking about adding a new virtio-blk-data-plane-pci device > type to hw/virtio-pci.c?
A new device sounds wrong to me, it's the very same thing from a guest perspective. Which makes me wonder if in the final version it shouldn't be a -blockdev option rather than a -device one... Kevin