@@ -57,6 +57,7 @@
#define VXLAN_VID_MASK (VXLAN_N_VID - 1)
/* IP header + UDP + VXLAN + Ethernet header */
#define VXLAN_HEADROOM (20 + 8 + 8 + 14)
+#define VXLAN_HLEN (sizeof(struct udphdr) + sizeof(struct vxlanhdr))
#define VXLAN_FLAGS 0x08000000 /* struct vxlanhdr.vx_flags required value. */
@@ -91,7 +92,12 @@ struct vxlan_sock {
struct list_head handler_list;
};
+struct vxlan_handler;
+typedef int (vxlan_rcv_t)(struct vxlan_handler *vh,
+ struct sk_buff *skb, __be32 key);
+
struct vxlan_handler {
+ vxlan_rcv_t *rcv;
struct list_head node;
struct vxlan_sock *vs;
atomic_t refcnt;
@@ -100,7 +106,8 @@ struct vxlan_handler {
};
static struct vxlan_handler *vs_handler_add(struct vxlan_sock *vs,
- __be16 portno);
+ __be16 portno,
+ vxlan_rcv_t *rcv);
static void vxlan_handler_hold(struct vxlan_handler *vh);
static void vxlan_handler_put(struct vxlan_handler *vh);
@@ -224,6 +231,18 @@ static struct vxlan_dev *vxlan_find_vni(struct net *net, u32 id, __be16 port)
return NULL;
}
+static struct vxlan_dev *vxlan_find_vni_port(struct vxlan_sock *vs, u32 id)
+{
+ struct vxlan_dev *vxlan;
+
+ hlist_for_each_entry_rcu(vxlan, vni_head(vs, id), hlist) {
+ if (vxlan->default_dst.remote_vni == id)
+ return vxlan;
+ }
+
+ return NULL;
+}
+
/* Fill in neighbour message in skbuff. */
static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,
const struct vxlan_fdb *fdb,
@@ -832,23 +851,17 @@ static void vxlan_igmp_work(struct work_struct *work)
/* Callback from net/ipv4/udp.c to receive packets */
static int vxlan_udp_encap_recv(struct sock *sk, struct sk_buff *skb)
{
- struct iphdr *oip;
+ struct vxlan_handler *vh;
+ struct vxlan_sock *vs;
struct vxlanhdr *vxh;
- struct vxlan_dev *vxlan;
- struct pcpu_tstats *stats;
__be16 port;
- __u32 vni;
- int err;
-
- /* pop off outer UDP header */
- __skb_pull(skb, sizeof(struct udphdr));
/* Need Vxlan and inner Ethernet header to be present */
- if (!pskb_may_pull(skb, sizeof(struct vxlanhdr)))
+ if (!pskb_may_pull(skb, VXLAN_HLEN))
goto error;
- /* Drop packets with reserved bits set */
- vxh = (struct vxlanhdr *) skb->data;
+ /* Return packets with reserved bits set */
+ vxh = (struct vxlanhdr *)(udp_hdr(skb) + 1);
if (vxh->vx_flags != htonl(VXLAN_FLAGS) ||
(vxh->vx_vni & htonl(0xff))) {
netdev_dbg(skb->dev, "invalid vxlan flags=%#x vni=%#x\n",
@@ -856,28 +869,46 @@ static int vxlan_udp_encap_recv(struct sock *sk, struct sk_buff *skb)
goto error;
}
- __skb_pull(skb, sizeof(struct vxlanhdr));
+ if (iptunnel_pull_header(skb, VXLAN_HLEN, htons(ETH_P_TEB)))
+ goto drop;
- /* Is this VNI defined? */
- vni = ntohl(vxh->vx_vni) >> 8;
port = inet_sk(sk)->inet_sport;
- vxlan = vxlan_find_vni(sock_net(sk), vni, port);
- if (!vxlan) {
- netdev_dbg(skb->dev, "unknown vni %d port %u\n",
- vni, ntohs(port));
- goto drop;
- }
- if (!pskb_may_pull(skb, ETH_HLEN)) {
- vxlan->dev->stats.rx_length_errors++;
- vxlan->dev->stats.rx_errors++;
+ vs = vxlan_find_sock(sock_net(sk), port);
+ if (!vs)
goto drop;
+
+ list_for_each_entry_rcu(vh, &vs->handler_list, node) {
+ if (vh->rcv(vh, skb, vxh->vx_vni) == PACKET_RCVD)
+ return 0;
}
- skb_reset_mac_header(skb);
+drop:
+ /* Consume bad packet */
+ kfree_skb(skb);
+ return 0;
- /* Re-examine inner Ethernet packet */
- oip = ip_hdr(skb);
+error:
+ /* Return non vxlan pkt */
+ return 1;
+}
+
+static int vxlan_rcv(struct vxlan_handler *vh,
+ struct sk_buff *skb, __be32 vx_vni)
+{
+ struct iphdr *oip;
+ struct vxlan_dev *vxlan;
+ struct pcpu_tstats *stats;
+ __u32 vni;
+ int err;
+
+ vni = ntohl(vx_vni) >> 8;
+ /* Is this VNI defined? */
+ vxlan = vxlan_find_vni_port(vh->vs, vni);
+ if (!vxlan)
+ return PACKET_REJECT;
+
+ skb_reset_mac_header(skb);
skb->protocol = eth_type_trans(skb, vxlan->dev);
/* Ignore packet loops (and multicast echo) */
@@ -885,11 +916,12 @@ static int vxlan_udp_encap_recv(struct sock *sk, struct sk_buff *skb)
vxlan->dev->dev_addr) == 0)
goto drop;
+ /* Re-examine inner Ethernet packet */
+ oip = ip_hdr(skb);
if ((vxlan->flags & VXLAN_F_LEARN) &&
vxlan_snoop(skb->dev, oip->saddr, eth_hdr(skb)->h_source))
goto drop;
- __skb_tunnel_rx(skb, vxlan->dev);
skb_reset_network_header(skb);
/* If the NIC driver gave us an encapsulated packet with
@@ -923,16 +955,11 @@ static int vxlan_udp_encap_recv(struct sock *sk, struct sk_buff *skb)
netif_rx(skb);
- return 0;
-error:
- /* Put UDP header back */
- __skb_push(skb, sizeof(struct udphdr));
-
- return 1;
+ return PACKET_RCVD;
drop:
/* Consume bad packet */
kfree_skb(skb);
- return 0;
+ return PACKET_RCVD;
}
static int arp_reduce(struct net_device *dev, struct sk_buff *skb)
@@ -1354,7 +1381,7 @@ static int vxlan_init(struct net_device *dev)
struct vxlan_handler *vh;
/* If we have a socket with same port already, reuse it */
- vh = vs_handler_add(vs, vxlan->dst_port);
+ vh = vs_handler_add(vs, vxlan->dst_port, vxlan_rcv);
vxlan_vh_add_dev(vh, vxlan);
} else {
/* otherwise make new socket outside of RTNL */
@@ -1671,15 +1698,25 @@ static void vh_del_work(struct work_struct *work)
}
static struct vxlan_handler *vs_handler_add(struct vxlan_sock *vs,
- __be16 portno)
+ __be16 portno,
+ vxlan_rcv_t *rcv)
{
struct vxlan_handler *vh;
+ /* Try existing vxlan hanlders for this socket. */
+ list_for_each_entry(vh, &vs->handler_list, node) {
+ if (vh->rcv == rcv) {
+ atomic_inc(&vh->refcnt);
+ return vh;
+ }
+ }
+
vh = kzalloc(sizeof(*vh), GFP_KERNEL);
if (!vh) {
vxlan_socket_del(vs);
return ERR_PTR(-ENOMEM);
}
+ vh->rcv = rcv;
vh->vs = vs;
atomic_set(&vh->refcnt, 1);
INIT_WORK(&vh->del_work, vh_del_work);
@@ -1688,7 +1725,8 @@ static struct vxlan_handler *vs_handler_add(struct vxlan_sock *vs,
return vh;
}
-static struct vxlan_handler *vxlan_handler_add(struct net *net, __be16 portno)
+static struct vxlan_handler *vxlan_handler_add(struct net *net,
+ __be16 portno, vxlan_rcv_t *rcv)
{
struct vxlan_net *vn = net_generic(net, vxlan_net_id);
struct vxlan_handler *vh;
@@ -1705,7 +1743,7 @@ static struct vxlan_handler *vxlan_handler_add(struct net *net, __be16 portno)
vh = ERR_PTR(-ENOENT);
goto out;
}
- vh = vs_handler_add(vs, portno);
+ vh = vs_handler_add(vs, portno, rcv);
out:
mutex_unlock(&vn->sock_lock);
@@ -1747,7 +1785,7 @@ static void vxlan_handler_work(struct work_struct *work)
__be16 port = vxlan->dst_port;
struct vxlan_handler *vh = NULL;
- vh = vxlan_handler_add(net, port);
+ vh = vxlan_handler_add(net, port, vxlan_rcv);
vxlan_vh_add_dev(vh, vxlan);
dev_put(dev);
}
Once we have ovs-vxlan functionality, one UDP port can be shared between vxlan and openvswitch module. Following patch adds this functionality so that vxlan_handler_add() user can pass its own rcv handler function. Signed-off-by: Pravin B Shelar <pshelar@nicira.com> --- split from first patch of version 3. --- drivers/net/vxlan.c | 118 +++++++++++++++++++++++++++++++++----------------- 1 files changed, 78 insertions(+), 40 deletions(-)