Skip to contentSkip to conversation

vlan: Propagate MAC address to VLANs unless explicitly set

PATCH Superseded Mike Manning 4 participants 5 messages May 3, 2016 – May 4, 2016 1 message not in the archive

Subsystems (2)

v2 of 5 Superseded Latest is v5 →

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw

PATCH v2 1/1 vlan: Propagate MAC address to VLANs unless explicitly set

The MAC address of the physical interface is only copied to the VLAN when it is first created, resulting in an inconsistency after MAC address changes of only newly created VLANs having an up-to-date MAC.

Continuing to inherit the MAC address unless explicitly changed for the VLAN allows IPv6 EUI64 addresses for the VLAN to reflect the change and thus for DAD to behave as expected for the given MAC.

---
 net/8021q/vlan.c |   22 ++++++++++------------
 1 file changed, 10 insertions(+), 12 deletions(-)
net/8021q/vlan.c +10 -12
@@ -286,22 +286,25 @@ static void vlan_sync_address(struct net
 			      struct net_device *vlandev)
 {
 	struct vlan_dev_priv *vlan = vlan_dev_priv(vlandev);
+	bool real_addr_in_use;
 
 	/* May be called without an actual change */
 	if (ether_addr_equal(vlan->real_dev_addr, dev->dev_addr))
 		return;
 
-	/* vlan address was different from the old address and is equal to
+	real_addr_in_use =
+		ether_addr_equal(vlandev->dev_addr, vlan->real_dev_addr);
+
+	/* vlan address was different from the real address and is equal to
 	 * the new address */
-	if (!ether_addr_equal(vlandev->dev_addr, vlan->real_dev_addr) &&
+	if ((vlandev->flags & IFF_UP) && !real_addr_in_use &&
 	    ether_addr_equal(vlandev->dev_addr, dev->dev_addr))
 		dev_uc_del(dev, vlandev->dev_addr);
 
-	/* vlan address was equal to the old address and is different from
+	/* vlan address was equal to the real address so now also inherit
 	 * the new address */
-	if (ether_addr_equal(vlandev->dev_addr, vlan->real_dev_addr) &&
-	    !ether_addr_equal(vlandev->dev_addr, dev->dev_addr))
-		dev_uc_add(dev, vlandev->dev_addr);
+	if (real_addr_in_use)
+		ether_addr_copy(vlandev->dev_addr, dev->dev_addr);
 
 	ether_addr_copy(vlan->real_dev_addr, dev->dev_addr);
 }
@@ -389,13 +392,8 @@ static int vlan_device_event(struct noti
 
 	case NETDEV_CHANGEADDR:
 		/* Adjust unicast filters on underlying device */
-		vlan_group_for_each_dev(grp, i, vlandev) {
-			flgs = vlandev->flags;
-			if (!(flgs & IFF_UP))
-				continue;
-
+		vlan_group_for_each_dev(grp, i, vlandev)
 			vlan_sync_address(dev, vlandev);
-		}
 		break;
 
 	case NETDEV_CHANGEMTU:
Signature · 1 line
1.7.10.4
maintainer
#2 lore raw

There is, at the very least, a two byte hole in the vlan_dev_priv structure. Put the boolean there as I asked you to.

#3 lore raw
In reply to Mike Manning
Quoted 52 lines · Mike Manning

The MAC address of the physical interface is only copied to the VLAN when it is first created, resulting in an inconsistency after MAC address changes of only newly created VLANs having an up-to-date MAC.

Continuing to inherit the MAC address unless explicitly changed for the VLAN allows IPv6 EUI64 addresses for the VLAN to reflect the change and thus for DAD to behave as expected for the given MAC.

---
 net/8021q/vlan.c |   22 ++++++++++------------
 1 file changed, 10 insertions(+), 12 deletions(-)
net/8021q/vlan.c +9 -7
@@ -286,22 +286,25 @@ static void vlan_sync_address(struct net
 			      struct net_device *vlandev)
 {
 	struct vlan_dev_priv *vlan = vlan_dev_priv(vlandev);
+	bool real_addr_in_use;
 
 	/* May be called without an actual change */
 	if (ether_addr_equal(vlan->real_dev_addr, dev->dev_addr))
 		return;
 
-	/* vlan address was different from the old address and is equal to
+	real_addr_in_use =
+		ether_addr_equal(vlandev->dev_addr, vlan->real_dev_addr);
+
+	/* vlan address was different from the real address and is equal to
 	 * the new address */
-	if (!ether_addr_equal(vlandev->dev_addr, vlan->real_dev_addr) &&
+	if ((vlandev->flags & IFF_UP) && !real_addr_in_use &&
 	    ether_addr_equal(vlandev->dev_addr, dev->dev_addr))
 		dev_uc_del(dev, vlandev->dev_addr);
 
-	/* vlan address was equal to the old address and is different from
+	/* vlan address was equal to the real address so now also inherit
 	 * the new address */
-	if (ether_addr_equal(vlandev->dev_addr, vlan->real_dev_addr) &&
-	    !ether_addr_equal(vlandev->dev_addr, dev->dev_addr))
-		dev_uc_add(dev, vlandev->dev_addr);
+	if (real_addr_in_use)
+		ether_addr_copy(vlandev->dev_addr, dev->dev_addr);
 
 	ether_addr_copy(vlan->real_dev_addr, dev->dev_addr);
 }
@@ -389,13 +392,8 @@ static int vlan_device_event(struct noti
 
 	case NETDEV_CHANGEADDR:
 		/* Adjust unicast filters on underlying device */
-		vlan_group_for_each_dev(grp, i, vlandev) {

net/8021q/vlan.c static int vlan_device_event(struct noti · line 399

Quoted 10 lines
net/8021q/vlan.c +1 -5
-			flgs = vlandev->flags;
-			if (!(flgs & IFF_UP))
-				continue;
-
+		vlan_group_for_each_dev(grp, i, vlandev)
 			vlan_sync_address(dev, vlandev);
-		}
 		break;
 
 	case NETDEV_CHANGEMTU:

The commit message says "unless explicitly changed for the VLAN" but what you really check is "if it is the same as real device MAC address". This, in general, is not the same. (I believe this is what David tries to explain from the start.)

Michal Kubecek

maintainer
#4 lore raw

From: Michal Kubecek <[email protected]>
Date: Wed, 4 May 2016 11:28:02 +0200

Quoted 4 lines

The commit message says "unless explicitly changed for the VLAN" but what you really check is "if it is the same as real device MAC address". This, in general, is not the same. (I believe this is what David tries to explain from the start.)

Even more proof that these MAC checks are rediculous, confusing, and that we need to use a boolean state stores in the vlan private in order to implement the intended semantics properly and more importantly "clearly".

#5 lore raw
Quoted 8 lines · David Miller

From: Michal Kubecek <[email protected]>
Date: Wed, 4 May 2016 11:28:02 +0200

Quoted 4 lines

The commit message says "unless explicitly changed for the VLAN" but what you really check is "if it is the same as real device MAC address". This, in general, is not the same. (I believe this is what David tries to explain from the start.)

Quoted 4 lines

Even more proof that these MAC checks are rediculous, confusing, and that we need to use a boolean state stores in the vlan private in order to implement the intended semantics properly and more importantly "clearly".

Actually last I knew the netdev already has a field called addr_assign_type. You could probably borrow some code from how bonding is currently handling MAC addresses and apply it to VLANs to achieve the effect you are looking for since I believe the default for VLAN is similar to NET_ADDR_STOLEN and when it is set you would switch that over to NET_ADDR_SET.

- Alex