Re: [Qemu-devel] [RFC-v2 4/6] virtio-scsi: Add start/stop functionality for vhost-scsi

["Michael S. Tsirkin" <mst@redhat.com>]

On Mon, Aug 20, 2012 at 12:31:01PM +0100, Stefan Hajnoczi wrote:
> On Mon, Aug 20, 2012 at 10:04 AM, Paolo Bonzini <pbonzini@redhat.com> wrote:
> > Il 13/08/2012 10:35, Nicholas A. Bellinger ha scritto:
> >> From: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
> >>
> >> This patch starts and stops vhost as the virtio device transitions
> >> through its status phases.  Vhost can only be started once the guest
> >> reports its driver has successfully initialized, which means the
> >> virtqueues have been set up by the guest.
> >>
> >> v2: - Squash virtio-scsi: use the vhost-scsi host device from stefan (nab)
> >>     - Fix up virtio_scsi_properties[] conflict w/ upstream qemu (nab)
> >>     - Drop usage of to_virtio_scsi() in virtio_scsi_set_status()
> >>       (reported by paolo)
> >>     - Use modern VirtIOSCSIConf define in virtio-scsi.h (reported by paolo)
> >>     - Use s->conf->vhost_scsi instead of proxyconf->vhost_scsi in
> >>       virtio_scsi_init() (reported by paolo)
> >>     - Only register QEMU SCSI bus is vhost-scsi is not active (reported
> >>       by paolo)
> >
> > How much of the functionality of virtio-scsi.[ch] is still in use at
> > this point?  Would it make more sense to use a separate vhost-scsi-pci
> > device instead?
> 
> Since the SCSI target lives in the kernel, almost everything is driven
> from tcm_vhost.ko.  tcm_vhost.ko basically implements the full device
> so I see the argument for -device vhost-scsi-pci.

A bit unhappy that there is no virtio in the name since it is
implementing virtio protocol.

> > Especially since advertising VIRTIO_SCSI_F_HOTPLUG and
> > VIRTIO_SCSI_F_CHANGE is probably wrong for vhost-scsi...
> 
> vhost participates in feature bit negotiation, see
> VHOST_GET_FEATURES/VHOST_SET_FEATURES ioctls.
> 
> Stefan