Re: [RFC PATCH 11/12] vdpa: Add virtio-net mac address via CVQ at start

[Eugenio Perez Martin]

On Mon, Jul 18, 2022 at 10:55 AM Jason Wang <jasowang@redhat.com> wrote:
>
>
> 在 2022/7/16 19:34, Eugenio Pérez 写道:
> > This is needed so the destination vdpa device see the same state a the
> > guest set in the source.
> >
> > Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
> > ---
> >   net/vhost-vdpa.c | 53 +++++++++++++++++++++++++++++++++++++++++++++++-
> >   1 file changed, 52 insertions(+), 1 deletion(-)
> >
> > diff --git a/net/vhost-vdpa.c b/net/vhost-vdpa.c
> > index 0183fce353..2873be2ba4 100644
> > --- a/net/vhost-vdpa.c
> > +++ b/net/vhost-vdpa.c
> > @@ -383,7 +383,7 @@ static virtio_net_ctrl_ack 
> > vhost_vdpa_net_svq_add(VhostShadowVirtqueue *svq,
> >       return VIRTIO_NET_OK;
> >   }
> >
> > -static int vhost_vdpa_start_control_svq(struct vhost_vdpa *v)
> > +static int vhost_vdpa_enable_control_svq(struct vhost_vdpa *v)
> >   {
> >       struct vhost_vring_state state = {
> >           .index = v->dev->vq_index,
> > @@ -395,6 +395,57 @@ static int vhost_vdpa_start_control_svq(struct 
> > vhost_vdpa *v)
> >       return r < 0 ? -errno : r;
> >   }
> >
> > +static int vhost_vdpa_start_control_svq(struct vhost_vdpa *v)
> > +{
> > +
> > +    VirtIONet *n = VIRTIO_NET(v->dev->vdev);
> > +    uint64_t features = v->dev->vdev->host_features;
> > +    VhostShadowVirtqueue *svq = g_ptr_array_index(v->shadow_vqs, 0);
> > +    VhostVDPAState *s = container_of(v, VhostVDPAState, vhost_vdpa);
> > +    int r;
> > +
> > +    r = vhost_vdpa_enable_control_svq(v);
> > +    if (unlikely(r < 0)) {
> > +        return r;
> > +    }
> > +
> > +    if (features & BIT_ULL(VIRTIO_NET_F_CTRL_MAC_ADDR)) {
> > +        const struct virtio_net_ctrl_hdr ctrl = {
> > +            .class = VIRTIO_NET_CTRL_MAC,
> > +            .cmd = VIRTIO_NET_CTRL_MAC_ADDR_SET,
> > +        };
> > +        uint8_t mac[6];
> > +        const struct iovec out[] = {
> > +            {
> > +                .iov_base = (void *)&ctrl,
> > +                .iov_len = sizeof(ctrl),
> > +            },{
> > +                .iov_base = mac,
> > +                .iov_len = sizeof(mac),
> > +            },
> > +        };
> > +        struct iovec dev_buffers[2] = {
> > +            { .iov_base = s->cvq_cmd_out_buffer },
> > +            { .iov_base = s->cvq_cmd_in_buffer },
> > +        };
> > +        bool ok;
> > +        virtio_net_ctrl_ack state;
> > +
> > +        ok = vhost_vdpa_net_cvq_map_sg(s, out, ARRAY_SIZE(out), 
> > dev_buffers);
> > +        if (unlikely(!ok)) {
> > +            return -1;
> > +        }
> > +
> > +        memcpy(mac, n->mac, sizeof(mac));
> > +        state = vhost_vdpa_net_svq_add(svq, dev_buffers);
> > +        vhost_vdpa_cvq_unmap_buf(v, dev_buffers[0].iov_base);
> > +        vhost_vdpa_cvq_unmap_buf(v, dev_buffers[1].iov_base);
>
>
> Any reason we do per buffer unmap instead of the sg unmap here?
>

I think I don't get this comment.

vhost_vdpa_net_handle_ctrl_avail also unmap each buffer individually,
and I need a function to unmap one of them at a time. I could create a
function to unmap a whole sg, but I'm not sure how much value it adds.

Thanks!

> Thanks
>
>
> > +        return state == VIRTIO_NET_OK ? 0 : 1;
> > +    }
> > +
> > +    return 0;
> > +}
> > +
> >   /**
> >    * Do not forward commands not supported by SVQ. Otherwise, the device 
> > could
> >    * accept it and qemu would not know how to update the device model.
>