On Mon, Jul 22, 2024 at 10:48 PM Cindy Lu <l...@redhat.com> wrote:
>
> On Mon, 22 Jul 2024 at 20:55, Cindy Lu <l...@redhat.com> wrote:
> >
> > On Mon, 22 Jul 2024 at 17:45, Dragos Tatulea <dtatu...@nvidia.com> wrote:
> > >
> > > On Mon, 2024-07-22 at 15:48 +0800, Jason Wang wrote:
> > > > On Mon, Jul 22, 2024 at 9:06 AM Cindy Lu <l...@redhat.com> wrote:
> > > > >
> > > > > Add the function to support setting the MAC address.
> > > > > For vdpa/mlx5, the function will use mlx5_mpfs_add_mac
> > > > > to set the mac address
> > > > >
> > > > > Tested in ConnectX-6 Dx device
> > > > >
> > > > > Signed-off-by: Cindy Lu <l...@redhat.com>
> > > > > ---
> > > > > drivers/vdpa/mlx5/net/mlx5_vnet.c | 25 +++++++++++++++++++++++++
> > > > > 1 file changed, 25 insertions(+)
> > > > >
> > > > > diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c
> > > > > b/drivers/vdpa/mlx5/net/mlx5_vnet.c
> > > > > index ecfc16151d61..415b527a9c72 100644
> > > > > --- a/drivers/vdpa/mlx5/net/mlx5_vnet.c
> > > > > +++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c
> > > > > @@ -3785,10 +3785,35 @@ static void mlx5_vdpa_dev_del(struct
> > > > > vdpa_mgmt_dev *v_mdev, struct vdpa_device *
> > > > > destroy_workqueue(wq);
> > > > > mgtdev->ndev = NULL;
> > > > > }
> > > > > +static int mlx5_vdpa_set_attr(struct vdpa_mgmt_dev *v_mdev,
> > > > > + struct vdpa_device *dev,
> > > > > + const struct vdpa_dev_set_config
> > > > > *add_config)
> > > > > +{
> > > > > + struct mlx5_vdpa_dev *mvdev;
> > > > > + struct mlx5_vdpa_net *ndev;
> > > > > + struct mlx5_core_dev *mdev;
> > > > > + struct virtio_net_config *config;
> > > > > + struct mlx5_core_dev *pfmdev;
> > > Reverse xmas tree?
> > >
> > will fix this
> > > > > + int err = -EOPNOTSUPP;
> > > > > +
> > > > > + mvdev = to_mvdev(dev);
> > > > > + ndev = to_mlx5_vdpa_ndev(mvdev);
> > > > > + mdev = mvdev->mdev;
> > > > > + config = &ndev->config;
> > > > > +
> > > You still need to take the ndev->reslock.
> > >
> > sure, will do
> > > > > + if (add_config->mask & (1 << VDPA_ATTR_DEV_NET_CFG_MACADDR)) {
> > > > > + pfmdev = pci_get_drvdata(pci_physfn(mdev->pdev));
> > > > > + err = mlx5_mpfs_add_mac(pfmdev, config->mac);
> > > > > + if (!err)
> > > > > + memcpy(config->mac, add_config->net.mac,
> > > > > ETH_ALEN);
> > > What is the expected behaviour when the device is in use?
> > >
> > if the err is 0 then copy the Mac address to config
> > will change this code to make it more clear
> > Thanks
> > Cindy
> sorry for the misunderstanding. The VDPA tool does not support
> changing the MAC address after the guest is loaded. I think I can add
> a check for VIRTIO_CONFIG_S_DRIVER_OK here?
Still racy, and I think we probably don't worry about that case. It's
the fault of the mgmt layer not us.
Thanks
> Thanks
> cindy
> > > > > + }
> > > > > + return err;
> > > >
> > > > Similar to net simulator, how could be serialize the modification to
> > > > mac address:
> > > >
> > > > 1) from vdpa tool
> > > > 2) via control virtqueue
> > > >
> > > > Thanks
> > > >
> > > > > +}
> > > > >
> > > > > static const struct vdpa_mgmtdev_ops mdev_ops = {
> > > > > .dev_add = mlx5_vdpa_dev_add,
> > > > > .dev_del = mlx5_vdpa_dev_del,
> > > > > + .dev_set_attr = mlx5_vdpa_set_attr,
> > > > > };
> > > > >
> > > > > static struct virtio_device_id id_table[] = {
> > > > > --
> > > > > 2.45.0
> > > > >
> > > >
> > > Thanks,
> > > Dragos
>
