If PF changes the MAC address of VF after VF is enabled, VF cannot sense
the change of its own MAC address, and the MAC address in VF's net_device
is still the original MAC address, which will cause the VF's network to
not work properly. Therefore, we should restrict PF to configure VF's MAC
address only when VF is disabled.
Of course, another solution is to notify VF of this event when PF changes
VF's MAC address, but the PSI-to-VSI messaging is not implemented in the
current PF and VF drivers, so this solution is not suitable at present,
If PSI-to-VSI messaging is supported in the future, we can remove the
current restriction to enhance the PF's ability to configure the VF's MAC
address.
Fixes: d4fd0404c1c9 ("enetc: Introduce basic PF and VF ENETC ethernet drivers")
Signed-off-by: Wei Fang <wei.fang@nxp.com>
---
drivers/net/ethernet/freescale/enetc/enetc_pf.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf.c b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
index c95a7c083b0f..a295236cd931 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pf.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
@@ -343,11 +343,18 @@ static int enetc_pf_set_vf_mac(struct net_device *ndev, int vf, u8 *mac)
{
struct enetc_ndev_priv *priv = netdev_priv(ndev);
struct enetc_pf *pf = enetc_si_priv(priv->si);
+ struct device *dev = &pf->si->pdev->dev;
struct enetc_vf_state *vf_state;
if (vf >= pf->total_vfs)
return -EINVAL;
+ if (vf + 1 <= pf->num_vfs) {
+ dev_err(dev, "Cannot set MAC address for an enabled VF\n");
+
+ return -EPERM;
+ }
+
if (!is_valid_ether_addr(mac))
return -EADDRNOTAVAIL;
--
2.34.1
On Thu, Oct 31, 2024 at 02:02:47PM +0800, Wei Fang wrote: > If PF changes the MAC address of VF after VF is enabled, VF cannot sense > the change of its own MAC address, and the MAC address in VF's net_device > is still the original MAC address, which will cause the VF's network to > not work properly. Therefore, we should restrict PF to configure VF's MAC > address only when VF is disabled. > > Of course, another solution is to notify VF of this event when PF changes > VF's MAC address, but the PSI-to-VSI messaging is not implemented in the > current PF and VF drivers, so this solution is not suitable at present, > If PSI-to-VSI messaging is supported in the future, we can remove the > current restriction to enhance the PF's ability to configure the VF's MAC > address. > > Fixes: d4fd0404c1c9 ("enetc: Introduce basic PF and VF ENETC ethernet drivers") > Signed-off-by: Wei Fang <wei.fang@nxp.com> > --- Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
© 2016 - 2024 Red Hat, Inc.