Patchwork [net-next] VXLAN: Allow L2 redirection with L3 switching

login
register
mail settings
Submitter David Stevens
Date April 19, 2013, 10:36 a.m.
Message ID <201304191036.r3JAaQ6p005959@lab1.dls>
Download mbox | patch
Permalink /patch/237900/
State Accepted
Delegated to: David Miller
Headers show

Comments

David Stevens - April 19, 2013, 10:36 a.m.
Allow L2 redirection when VXLAN L3 switching is enabled

This patch restricts L3 switching to destination MAC addresses that are
marked as routers in order to allow virtual IP appliances that do L2
redirection to function with VXLAN L3 switching enabled.

We use L3 switching on VXLAN networks to avoid extra hops when the nominal
router for cross-subnet traffic for a VM is remote and the ultimate
destination may be local, or closer to the local node. Currently, the
destination IP address takes precedence over the MAC address in all cases.
Some network appliances receive packets for a virtualized IP address and
redirect by changing the destination MAC address (only) to be the final
destination for packet processing. VXLAN tunnel endpoints with L3 switching
enabled may then overwrite this destination MAC address based on the packet IP
address, resulting in potential loops and, at least, breaking L2 redirections
that travel through tunnel endpoints.

This patch limits L3 switching to the intended case where the original
destination MAC address is a next-hop router and relies on the destination
MAC address for all other cases, thus allowing L2 redirection and L3 switching
to coexist peacefully.

Signed-Off-By: David L Stevens <dlstevens@us.ibm.com>


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
WANG Cong - April 19, 2013, 2:27 p.m.
Hi,

Could you hold on it? The more you add for IPv4, the more I need to
complete for IPv6... You are creating more barriers to my IPv6
patches (not to mention conflicts). 8-)

I highly appreciate if you can understand my situation. Just wait for
a few more days until my patchset is accepted, this will save me *a
lot* of time.

Thanks!

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Miller - April 22, 2013, 8:10 p.m.
From: David L Stevens <dlstevens@us.ibm.com>
Date: Fri, 19 Apr 2013 06:36:26 -0400

> 
> Allow L2 redirection when VXLAN L3 switching is enabled
> 
> This patch restricts L3 switching to destination MAC addresses that are
> marked as routers in order to allow virtual IP appliances that do L2
> redirection to function with VXLAN L3 switching enabled.
> 
> We use L3 switching on VXLAN networks to avoid extra hops when the nominal
> router for cross-subnet traffic for a VM is remote and the ultimate
> destination may be local, or closer to the local node. Currently, the
> destination IP address takes precedence over the MAC address in all cases.
> Some network appliances receive packets for a virtualized IP address and
> redirect by changing the destination MAC address (only) to be the final
> destination for packet processing. VXLAN tunnel endpoints with L3 switching
> enabled may then overwrite this destination MAC address based on the packet IP
> address, resulting in potential loops and, at least, breaking L2 redirections
> that travel through tunnel endpoints.
> 
> This patch limits L3 switching to the intended case where the original
> destination MAC address is a next-hop router and relies on the destination
> MAC address for all other cases, thus allowing L2 redirection and L3 switching
> to coexist peacefully.
> 
> Signed-Off-By: David L Stevens <dlstevens@us.ibm.com>

Applied, thanks David.

Amerigo, I know this conflicts with your ipv6 work, but I'm not making
David wait while you sort out all of those ipv6 symbol export build
failures.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 916a621..a7fd9a0 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -98,6 +98,7 @@  struct vxlan_fdb {
 	unsigned long	  used;
 	struct vxlan_rdst remote;
 	u16		  state;	/* see ndm_state */
+	u8		  flags;	/* see ndm_flags */
 	u8		  eth_addr[ETH_ALEN];
 };
 
@@ -180,7 +181,7 @@  static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,
 		ndm->ndm_family	= AF_BRIDGE;
 	ndm->ndm_state = fdb->state;
 	ndm->ndm_ifindex = vxlan->dev->ifindex;
-	ndm->ndm_flags = NTF_SELF;
+	ndm->ndm_flags = fdb->flags;
 	ndm->ndm_type = NDA_DST;
 
 	if (send_eth && nla_put(skb, NDA_LLADDR, ETH_ALEN, &fdb->eth_addr))
@@ -343,7 +344,8 @@  static int vxlan_fdb_append(struct vxlan_fdb *f,
 static int vxlan_fdb_create(struct vxlan_dev *vxlan,
 			    const u8 *mac, __be32 ip,
 			    __u16 state, __u16 flags,
-			    __u32 port, __u32 vni, __u32 ifindex)
+			    __u32 port, __u32 vni, __u32 ifindex,
+			    __u8 ndm_flags)
 {
 	struct vxlan_fdb *f;
 	int notify = 0;
@@ -360,6 +362,11 @@  static int vxlan_fdb_create(struct vxlan_dev *vxlan,
 			f->updated = jiffies;
 			notify = 1;
 		}
+		if (f->flags != ndm_flags) {
+			f->flags = ndm_flags;
+			f->updated = jiffies;
+			notify = 1;
+		}
 		if ((flags & NLM_F_APPEND) &&
 		    is_multicast_ether_addr(f->eth_addr)) {
 			int rc = vxlan_fdb_append(f, ip, port, vni, ifindex);
@@ -387,6 +394,7 @@  static int vxlan_fdb_create(struct vxlan_dev *vxlan,
 		f->remote.remote_ifindex = ifindex;
 		f->remote.remote_next = NULL;
 		f->state = state;
+		f->flags = ndm_flags;
 		f->updated = f->used = jiffies;
 		memcpy(f->eth_addr, mac, ETH_ALEN);
 
@@ -480,7 +488,7 @@  static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr *tb[],
 
 	spin_lock_bh(&vxlan->hash_lock);
 	err = vxlan_fdb_create(vxlan, addr, ip, ndm->ndm_state, flags, port,
-		vni, ifindex);
+		vni, ifindex, ndm->ndm_flags);
 	spin_unlock_bh(&vxlan->hash_lock);
 
 	return err;
@@ -568,7 +576,9 @@  static void vxlan_snoop(struct net_device *dev,
 		err = vxlan_fdb_create(vxlan, src_mac, src_ip,
 				       NUD_REACHABLE,
 				       NLM_F_EXCL|NLM_F_CREATE,
-				       vxlan_port, vxlan->default_dst.remote_vni, 0);
+				       vxlan_port,
+				       vxlan->default_dst.remote_vni,
+				       0, NTF_SELF);
 		spin_unlock(&vxlan->hash_lock);
 	}
 }
@@ -1098,12 +1108,18 @@  static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
 
 	if ((vxlan->flags & VXLAN_F_PROXY) && ntohs(eth->h_proto) == ETH_P_ARP)
 		return arp_reduce(dev, skb);
-	else if ((vxlan->flags&VXLAN_F_RSC) && ntohs(eth->h_proto) == ETH_P_IP)
-		did_rsc = route_shortcircuit(dev, skb);
 
 	f = vxlan_find_mac(vxlan, eth->h_dest);
+	did_rsc = false;
+
+	if (f && (f->flags & NTF_ROUTER) && (vxlan->flags & VXLAN_F_RSC) &&
+	    ntohs(eth->h_proto) == ETH_P_IP) {
+		did_rsc = route_shortcircuit(dev, skb);
+		if (did_rsc)
+			f = vxlan_find_mac(vxlan, eth->h_dest);
+	}
+
 	if (f == NULL) {
-		did_rsc = false;
 		rdst0 = &vxlan->default_dst;
 
 		if (rdst0->remote_ip == htonl(INADDR_ANY) &&