Re: [Qemu-devel] [virtio-dev] Re: [PATCH v3 5/7] vhost-user: VHOST_USER_SET_VHOST_PCI msg

[Stefan Hajnoczi <stefanha@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1740 bytes --]

On Wed, Dec 06, 2017 at 06:32:50PM +0800, Wei Wang wrote:
> On 12/06/2017 12:00 AM, Stefan Hajnoczi wrote:
> > On Tue, Dec 05, 2017 at 11:33:14AM +0800, Wei Wang wrote:
> > > Add a new vhost-uer protocol msg, VHOST_USER_SET_VHOST_PCI. This msg is
> > > used to signal the vhost-pci device to start/stop working.
> > > 
> > > Signed-off-by: Wei Wang <wei.w.wang@intel.com>
> > > ---
> > >   hw/net/vhost_net.c                | 37 +++++++++++++++++++++++++++++++++++++
> > >   hw/virtio/vhost-pci-slave.c       |  2 +-
> > >   hw/virtio/vhost-user.c            | 17 +++++++++++++++++
> > >   hw/virtio/vhost.c                 |  7 +++++++
> > >   include/hw/virtio/vhost-backend.h |  2 ++
> > >   include/hw/virtio/vhost-user.h    |  2 ++
> > >   include/hw/virtio/vhost.h         |  2 ++
> > >   include/net/vhost_net.h           |  2 ++
> > >   8 files changed, 70 insertions(+), 1 deletion(-)
> > New protocol messages must be documented in docs/interop/vhost-user.txt.
> 
> OK, I'll add it to the doc after the discussion.
> 
> > 
> > Why is a new message needed?  I'm not sure why it is specific to
> > vhost-pci.
> 
> Yes, it might be useful for other vhost-user slave implementations. Probably
> we can name it "VHOST_USER_SET_SLAVE"?
> The message is used to "link up" or "link down" the slave device. For
> example, when virtio-net leaves, it sends a "VHOST_USER_SET_SLAVE" msg to
> the salve to link down the slave device.
> (a similar msg is VHOST_USER_SET_VRING_ENABLE, but that is for virtqueue
> enable/disable, not for a device level enable/disable)

Why is VHOST_USER_SET_VHOST_PCI necessary when DPDK and other vhost-user
net device slaves already exist today and didn't need it?

Stefan

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 455 bytes --]