| Message ID | 20170405134622.12227-1-vinschen@redhat.com |
|---|---|
| State | Changes Requested |
| Delegated to: | Jeff Kirsher |
| Headers | show |
On Wed, 2017-04-05 at 15:46 +0200, Corinna Vinschen wrote: > Before libvirt modifies the MAC address and vlan tag for an SRIOV > VF > for use by a virtual machine (either using vfio device assignment > or > macvtap passthru mode), it saves the current MAC address and vlan > tag > so that it can reset them to their original value when the guest is > done. Libvirt can't leave the VF MAC set to the value used by the > now-defunct guest since it may be started again later using a > different VF, but it certainly shouldn't just pick any random > value, > either. So it saves the state of everything prior to using the VF, > and > resets it to that. > > The igb driver initializes the MAC addresses of all VFs to > 00:00:00:00:00:00, and reports that when asked (via an RTM_GETLINK > netlink message, also visible in the list of VFs in the output of > "ip > link show"). But when libvirt attempts to restore the MAC address > back > to 00:00:00:00:00:00 (using an RTM_SETLINK netlink message) the > kernel > responds with "Invalid argument". > > Forbidding a reset back to the original value leaves the VF MAC at > the > value set for the now-defunct virtual machine. Especially on a > system > with NetworkManager enabled, this has very bad consequences, since > NetworkManager forces all interfacess to be IFF_UP all the time - > if > the same virtual machine is restarted using a different VF (or even > on > a different host), there will be multiple interfaces watching for > traffic with the same MAC address. > > To allow libvirt to revert to the original state, we need a way to > remove the administrative set MAC on a VF, to allow normal host > operation again, and to reset/overwrite the VF MAC via VF netdev. > > This patch implements the outlined scenario by allowing to set the > VF MAC to 00:00:00:00:00:00 via RTM_SETLINK on the PF. > igb_ndo_set_vf_mac resets the IGB_VF_FLAG_PF_SET_MAC flag to 0, > so it's possible to reset the VF MAC back to the original value via > the VF netdev. > > Note: Recent patches to libvirt allow for a workaround if the NIC > isn't capable of resetting the administrative MAC back to all 0, > but > in theory the NIC should allow resetting the MAC in the first > place. > > Signed-off-by: Corinna Vinschen <vinschen@redhat.com> > --- > drivers/net/ethernet/intel/igb/igb_main.c | 42 > +++++++++++++++++++++++-------- > 1 file changed, 31 insertions(+), 11 deletions(-) This patch does not apply (not even close). Please make sure to base you patch off my dev-queue branch of my next-queue tree on kernel.org.
On Apr 7 12:06, Jeff Kirsher wrote: > On Wed, 2017-04-05 at 15:46 +0200, Corinna Vinschen wrote: > > Before libvirt modifies the MAC address and vlan tag for an SRIOV > > VF > > for use by a virtual machine (either using vfio device assignment > > or > > macvtap passthru mode), it saves the current MAC address and vlan > > tag > > so that it can reset them to their original value when the guest is > > done. Libvirt can't leave the VF MAC set to the value used by the > > now-defunct guest since it may be started again later using a > > different VF, but it certainly shouldn't just pick any random > > value, > > either. So it saves the state of everything prior to using the VF, > > and > > resets it to that. > > > > The igb driver initializes the MAC addresses of all VFs to > > 00:00:00:00:00:00, and reports that when asked (via an RTM_GETLINK > > netlink message, also visible in the list of VFs in the output of > > "ip > > link show"). But when libvirt attempts to restore the MAC address > > back > > to 00:00:00:00:00:00 (using an RTM_SETLINK netlink message) the > > kernel > > responds with "Invalid argument". > > > > Forbidding a reset back to the original value leaves the VF MAC at > > the > > value set for the now-defunct virtual machine. Especially on a > > system > > with NetworkManager enabled, this has very bad consequences, since > > NetworkManager forces all interfacess to be IFF_UP all the time - > > if > > the same virtual machine is restarted using a different VF (or even > > on > > a different host), there will be multiple interfaces watching for > > traffic with the same MAC address. > > > > To allow libvirt to revert to the original state, we need a way to > > remove the administrative set MAC on a VF, to allow normal host > > operation again, and to reset/overwrite the VF MAC via VF netdev. > > > > This patch implements the outlined scenario by allowing to set the > > VF MAC to 00:00:00:00:00:00 via RTM_SETLINK on the PF. > > igb_ndo_set_vf_mac resets the IGB_VF_FLAG_PF_SET_MAC flag to 0, > > so it's possible to reset the VF MAC back to the original value via > > the VF netdev. > > > > Note: Recent patches to libvirt allow for a workaround if the NIC > > isn't capable of resetting the administrative MAC back to all 0, > > but > > in theory the NIC should allow resetting the MAC in the first > > place. > > > > Signed-off-by: Corinna Vinschen <vinschen@redhat.com> > > --- > > drivers/net/ethernet/intel/igb/igb_main.c | 42 > > +++++++++++++++++++++++-------- > > 1 file changed, 31 insertions(+), 11 deletions(-) > > This patch does not apply (not even close). Please make sure to base > you patch off my dev-queue branch of my next-queue tree on kernel.org. Right, I applied the patch against net-next. Actually, only the first hunk didn't apply to dev-queue because of the change from igb_rar_set_qsel to igb_rar_set_index. I'll send a v3 agains dev-queue in a bit. Corinna
diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c index 26a821f..c7673a2 100644 --- a/drivers/net/ethernet/intel/igb/igb_main.c +++ b/drivers/net/ethernet/intel/igb/igb_main.c @@ -8093,7 +8093,8 @@ static void igb_rar_set_qsel(struct igb_adapter *adapter, u8 *addr, u32 index, rar_high = le16_to_cpup((__le16 *)(addr + 4)); /* Indicate to hardware the Address is Valid. */ - rar_high |= E1000_RAH_AV; + if (is_valid_ether_addr(addr)) + rar_high |= E1000_RAH_AV; if (hw->mac.type == e1000_82575) rar_high |= E1000_RAH_POOL_1 * qsel; @@ -8125,17 +8126,36 @@ static int igb_set_vf_mac(struct igb_adapter *adapter, static int igb_ndo_set_vf_mac(struct net_device *netdev, int vf, u8 *mac) { struct igb_adapter *adapter = netdev_priv(netdev); - if (!is_valid_ether_addr(mac) || (vf >= adapter->vfs_allocated_count)) + + if (vf >= adapter->vfs_allocated_count) + return -EINVAL; + + /* Setting the VF MAC to 0 reverts the IGB_VF_FLAG_PF_SET_MAC + * flag and allows to overwrite the MAC via VF netdev. This + * is necessary to allow libvirt a way to restore the original + * MAC after unbinding vfio-pci and reloading igbvf after shutting + * down a VM. + */ + if (is_zero_ether_addr(mac)) { + adapter->vf_data[vf].flags &= ~IGB_VF_FLAG_PF_SET_MAC; + dev_info(&adapter->pdev->dev, + "remove administratively set MAC on VF %d\n", + vf); + } else if (is_valid_ether_addr (mac)) { + adapter->vf_data[vf].flags |= IGB_VF_FLAG_PF_SET_MAC; + dev_info(&adapter->pdev->dev, "setting MAC %pM on VF %d\n", + mac, vf); + dev_info(&adapter->pdev->dev, + "Reload the VF driver to make this change effective."); + /* Generate additional warning if PF is down */ + if (test_bit(__IGB_DOWN, &adapter->state)) { + dev_warn(&adapter->pdev->dev, + "The VF MAC address has been set, but the PF device is not up.\n"); + dev_warn(&adapter->pdev->dev, + "Bring the PF device up before attempting to use the VF device.\n"); + } + } else { return -EINVAL; - adapter->vf_data[vf].flags |= IGB_VF_FLAG_PF_SET_MAC; - dev_info(&adapter->pdev->dev, "setting MAC %pM on VF %d\n", mac, vf); - dev_info(&adapter->pdev->dev, - "Reload the VF driver to make this change effective."); - if (test_bit(__IGB_DOWN, &adapter->state)) { - dev_warn(&adapter->pdev->dev, - "The VF MAC address has been set, but the PF device is not up.\n"); - dev_warn(&adapter->pdev->dev, - "Bring the PF device up before attempting to use the VF device.\n"); } return igb_set_vf_mac(adapter, vf, mac); }
Before libvirt modifies the MAC address and vlan tag for an SRIOV VF for use by a virtual machine (either using vfio device assignment or macvtap passthru mode), it saves the current MAC address and vlan tag so that it can reset them to their original value when the guest is done. Libvirt can't leave the VF MAC set to the value used by the now-defunct guest since it may be started again later using a different VF, but it certainly shouldn't just pick any random value, either. So it saves the state of everything prior to using the VF, and resets it to that. The igb driver initializes the MAC addresses of all VFs to 00:00:00:00:00:00, and reports that when asked (via an RTM_GETLINK netlink message, also visible in the list of VFs in the output of "ip link show"). But when libvirt attempts to restore the MAC address back to 00:00:00:00:00:00 (using an RTM_SETLINK netlink message) the kernel responds with "Invalid argument". Forbidding a reset back to the original value leaves the VF MAC at the value set for the now-defunct virtual machine. Especially on a system with NetworkManager enabled, this has very bad consequences, since NetworkManager forces all interfacess to be IFF_UP all the time - if the same virtual machine is restarted using a different VF (or even on a different host), there will be multiple interfaces watching for traffic with the same MAC address. To allow libvirt to revert to the original state, we need a way to remove the administrative set MAC on a VF, to allow normal host operation again, and to reset/overwrite the VF MAC via VF netdev. This patch implements the outlined scenario by allowing to set the VF MAC to 00:00:00:00:00:00 via RTM_SETLINK on the PF. igb_ndo_set_vf_mac resets the IGB_VF_FLAG_PF_SET_MAC flag to 0, so it's possible to reset the VF MAC back to the original value via the VF netdev. Note: Recent patches to libvirt allow for a workaround if the NIC isn't capable of resetting the administrative MAC back to all 0, but in theory the NIC should allow resetting the MAC in the first place. Signed-off-by: Corinna Vinschen <vinschen@redhat.com> --- drivers/net/ethernet/intel/igb/igb_main.c | 42 +++++++++++++++++++++++-------- 1 file changed, 31 insertions(+), 11 deletions(-)