#include <linux/in.h>
#include <linux/ip.h>
-#include <linux/list.h>
#include <linux/net.h>
+#include <linux/rculist.h>
#include <linux/udp.h>
#include <net/icmp.h>
#include <net/ip.h>
#include <net/udp.h>
+#include <net/ip_tunnels.h>
+#include <net/udp.h>
+#include <net/rtnetlink.h>
+#include <net/route.h>
+#include <net/dsfield.h>
+#include <net/inet_ecn.h>
+#include <net/net_namespace.h>
+#include <net/netns/generic.h>
+#include <net/vxlan.h>
#include "datapath.h"
-#include "tunnel.h"
#include "vport.h"
-#define VXLAN_FLAGS 0x08000000 /* struct vxlanhdr.vx_flags required value. */
-
-/**
- * struct vxlanhdr - VXLAN header
- * @vx_flags: Must have the exact value %VXLAN_FLAGS.
- * @vx_vni: VXLAN Network Identifier (VNI) in top 24 bits, low 8 bits zeroed.
- */
-struct vxlanhdr {
- __be32 vx_flags;
- __be32 vx_vni;
-};
-
-#define VXLAN_HLEN (sizeof(struct udphdr) + sizeof(struct vxlanhdr))
-
-static inline int vxlan_hdr_len(const struct tnl_mutable_config *mutable,
- const struct ovs_key_ipv4_tunnel *tun_key)
-{
- return VXLAN_HLEN;
-}
+#define OVS_VXLAN_RCV_PRIORITY 8
/**
* struct vxlan_port - Keeps track of open UDP ports
- * @list: list element.
- * @port: The UDP port number in network byte order.
- * @socket: The socket created for this port number.
- * @count: How many ports are using this socket/port.
+ * @vh: vxlan_handler created for the port.
+ * @name: vport name.
*/
struct vxlan_port {
- struct list_head list;
- __be16 port;
- struct socket *vxlan_rcv_socket;
- int count;
+ struct vxlan_handler *vh;
+ char name[IFNAMSIZ];
};
-static LIST_HEAD(vxlan_ports);
-
-static struct vxlan_port *vxlan_port_exists(struct net *net, __be16 port)
-{
- struct vxlan_port *vxlan_port;
-
- list_for_each_entry(vxlan_port, &vxlan_ports, list) {
- if (vxlan_port->port == port &&
- net_eq(sock_net(vxlan_port->vxlan_rcv_socket->sk), net))
- return vxlan_port;
- }
-
- return NULL;
-}
-
-static inline struct vxlanhdr *vxlan_hdr(const struct sk_buff *skb)
+static inline struct vxlan_port *vxlan_vport(const struct vport *vport)
{
- return (struct vxlanhdr *)(udp_hdr(skb) + 1);
-}
-
-/* Compute source port for outgoing packet.
- * Currently we use the flow hash.
- */
-static u16 get_src_port(struct sk_buff *skb)
-{
- int low;
- int high;
- unsigned int range;
- u32 hash = OVS_CB(skb)->flow->hash;
-
- inet_get_local_port_range(&low, &high);
- range = (high - low) + 1;
- return (((u64) hash * range) >> 32) + low;
-}
-
-static struct sk_buff *vxlan_build_header(const struct vport *vport,
- const struct tnl_mutable_config *mutable,
- struct dst_entry *dst,
- struct sk_buff *skb,
- int tunnel_hlen)
-{
- struct udphdr *udph = udp_hdr(skb);
- struct vxlanhdr *vxh = (struct vxlanhdr *)(udph + 1);
- const struct ovs_key_ipv4_tunnel *tun_key = OVS_CB(skb)->tun_key;
- __be64 out_key;
- u32 flags;
-
- tnl_get_param(mutable, tun_key, &flags, &out_key);
-
- udph->dest = mutable->dst_port;
- udph->source = htons(get_src_port(skb));
- udph->check = 0;
- udph->len = htons(skb->len - skb_transport_offset(skb));
-
- vxh->vx_flags = htonl(VXLAN_FLAGS);
- vxh->vx_vni = htonl(be64_to_cpu(out_key) << 8);
-
- /*
- * Allow our local IP stack to fragment the outer packet even if the
- * DF bit is set as a last resort. We also need to force selection of
- * an IP ID here because Linux will otherwise leave it at 0 if the
- * packet originally had DF set.
- */
- skb->local_df = 1;
- __ip_select_ident(ip_hdr(skb), dst, 0);
-
- return skb;
+ return vport_priv(vport);
}
/* Called with rcu_read_lock and BH disabled. */
-static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
+static int vxlan_rcv(struct vxlan_handler *vh, struct sk_buff *skb, __be32 vx_vni)
{
- struct vport *vport;
- struct vxlanhdr *vxh;
- const struct tnl_mutable_config *mutable;
+ struct vport *vport = vh->data;
struct iphdr *iph;
struct ovs_key_ipv4_tunnel tun_key;
__be64 key;
- u32 tunnel_flags = 0;
-
- if (unlikely(!pskb_may_pull(skb, VXLAN_HLEN + ETH_HLEN)))
- goto error;
-
- vxh = vxlan_hdr(skb);
- if (unlikely(vxh->vx_flags != htonl(VXLAN_FLAGS) ||
- vxh->vx_vni & htonl(0xff)))
- goto error;
-
- __skb_pull(skb, VXLAN_HLEN);
- skb_postpull_rcsum(skb, skb_transport_header(skb), VXLAN_HLEN + ETH_HLEN);
-
- key = cpu_to_be64(ntohl(vxh->vx_vni) >> 8);
-
- iph = ip_hdr(skb);
- vport = ovs_tnl_find_port(dev_net(skb->dev), iph->daddr, iph->saddr,
- key, TNL_T_PROTO_VXLAN, &mutable);
- if (unlikely(!vport)) {
- icmp_send(skb, ICMP_DEST_UNREACH, ICMP_PORT_UNREACH, 0);
- goto error;
- }
-
- if (mutable->flags & TNL_F_IN_KEY_MATCH || !mutable->key.daddr)
- tunnel_flags = OVS_TNL_F_KEY;
- else
- key = 0;
/* Save outer tunnel values */
- tnl_tun_key_init(&tun_key, iph, key, tunnel_flags);
- OVS_CB(skb)->tun_key = &tun_key;
-
- ovs_tnl_rcv(vport, skb);
- goto out;
+ iph = ip_hdr(skb);
+ key = cpu_to_be64(ntohl(vx_vni) >> 8);
+ ovs_flow_tun_key_init(&tun_key, iph, key, TUNNEL_KEY);
-error:
- kfree_skb(skb);
-out:
- return 0;
+ ovs_vport_receive(vport, skb, &tun_key);
+ return PACKET_RCVD;
}
-/* Random value. Irrelevant as long as it's not 0 since we set the handler. */
-#define UDP_ENCAP_VXLAN 1
-static int vxlan_socket_init(struct vxlan_port *vxlan_port, struct net *net)
+static int vxlan_get_options(const struct vport *vport, struct sk_buff *skb)
{
- int err;
- struct sockaddr_in sin;
-
- err = sock_create_kern(AF_INET, SOCK_DGRAM, 0,
- &vxlan_port->vxlan_rcv_socket);
- if (err)
- goto error;
-
- /* release net ref. */
- sk_change_net(vxlan_port->vxlan_rcv_socket->sk, net);
-
- sin.sin_family = AF_INET;
- sin.sin_addr.s_addr = htonl(INADDR_ANY);
- sin.sin_port = vxlan_port->port;
-
- err = kernel_bind(vxlan_port->vxlan_rcv_socket, (struct sockaddr *)&sin,
- sizeof(struct sockaddr_in));
- if (err)
- goto error_sock;
-
- udp_sk(vxlan_port->vxlan_rcv_socket->sk)->encap_type = UDP_ENCAP_VXLAN;
- udp_sk(vxlan_port->vxlan_rcv_socket->sk)->encap_rcv = vxlan_rcv;
-
- udp_encap_enable();
+ struct vxlan_port *vxlan_port = vxlan_vport(vport);
+ __be16 dst_port = inet_sport(vxlan_port->vh->vs->sock->sk);
+ if (nla_put_u16(skb, OVS_TUNNEL_ATTR_DST_PORT, ntohs(dst_port)))
+ return -EMSGSIZE;
return 0;
-
-error_sock:
- sk_release_kernel(vxlan_port->vxlan_rcv_socket->sk);
-error:
- pr_warn("cannot register vxlan protocol handler\n");
- return err;
}
-static void vxlan_tunnel_release(struct vxlan_port *vxlan_port)
+static void vxlan_tnl_destroy(struct vport *vport)
{
- vxlan_port->count--;
+ struct vxlan_port *vxlan_port = vxlan_vport(vport);
- if (vxlan_port->count == 0) {
- /* Release old socket */
- sk_release_kernel(vxlan_port->vxlan_rcv_socket->sk);
- list_del(&vxlan_port->list);
- kfree(vxlan_port);
- }
+ vxlan_handler_put(vxlan_port->vh);
+
+ ovs_vport_deferred_free(vport);
}
-static int vxlan_tunnel_setup(struct net *net, struct nlattr *options,
- struct vxlan_port **vxport)
+
+static struct vport *vxlan_tnl_create(const struct vport_parms *parms)
{
+ struct net *net = ovs_dp_get_net(parms->dp);
+ struct nlattr *options = parms->options;
+ struct vxlan_port *vxlan_port;
+ struct vxlan_handler *vh;
+ struct vport *vport;
struct nlattr *a;
- int err;
u16 dst_port;
- struct vxlan_port *vxlan_port = NULL;
-
- *vxport = NULL;
+ int err;
if (!options) {
err = -EINVAL;
- goto out;
+ goto error;
}
-
a = nla_find_nested(options, OVS_TUNNEL_ATTR_DST_PORT);
if (a && nla_len(a) == sizeof(u16)) {
dst_port = nla_get_u16(a);
} else {
/* Require destination port from userspace. */
err = -EINVAL;
- goto out;
- }
-
- /* Verify if we already have a socket created for this port */
- vxlan_port = vxlan_port_exists(net, htons(dst_port));
- if (vxlan_port) {
- vxlan_port->count++;
- err = 0;
- *vxport = vxlan_port;
- goto out;
+ goto error;
}
- /* Add a new socket for this port */
- vxlan_port = kzalloc(sizeof(struct vxlan_port), GFP_KERNEL);
- if (!vxlan_port) {
- err = -ENOMEM;
- goto out;
- }
+ vport = ovs_vport_alloc(sizeof(struct vxlan_port),
+ &ovs_vxlan_vport_ops, parms);
+ if (IS_ERR(vport))
+ return vport;
- vxlan_port->port = htons(dst_port);
- vxlan_port->count = 1;
- list_add_tail(&vxlan_port->list, &vxlan_ports);
+ vxlan_port = vxlan_vport(vport);
+ strncpy(vxlan_port->name, parms->name, IFNAMSIZ);
- err = vxlan_socket_init(vxlan_port, net);
- if (err)
- goto error;
+ vh = vxlan_handler_add(net, htons(dst_port), vxlan_rcv,
+ vport, OVS_VXLAN_RCV_PRIORITY, true);
+ if (IS_ERR(vh)) {
+ ovs_vport_free(vport);
+ return (void *)vh;
+ }
+ vxlan_port->vh = vh;
- *vxport = vxlan_port;
- goto out;
+ return vport;
error:
- list_del(&vxlan_port->list);
- kfree(vxlan_port);
-out:
- return err;
+ return ERR_PTR(err);
}
-static int vxlan_set_options(struct vport *vport, struct nlattr *options)
+static int vxlan_tnl_send(struct vport *vport, struct sk_buff *skb)
{
- int err;
+ struct vxlan_port *vxlan_port = vxlan_vport(vport);
+ __be16 dst_port = inet_sport(vxlan_port->vh->vs->sock->sk);
struct net *net = ovs_dp_get_net(vport->dp);
- struct tnl_vport *tnl_vport = tnl_vport_priv(vport);
- struct tnl_mutable_config *config;
- struct vxlan_port *old_port = NULL;
- struct vxlan_port *vxlan_port = NULL;
-
- config = rtnl_dereference(tnl_vport->mutable);
-
- old_port = vxlan_port_exists(net, config->dst_port);
-
- err = vxlan_tunnel_setup(net, options, &vxlan_port);
- if (err)
- goto out;
-
- err = ovs_tnl_set_options(vport, options);
+ struct rtable *rt;
+ __be16 src_port;
+ __be32 saddr;
+ __be16 df;
+ int port_min;
+ int port_max;
+ int err;
- if (err)
- vxlan_tunnel_release(vxlan_port);
- else {
- /* Release old socket */
- vxlan_tunnel_release(old_port);
+ if (unlikely(!OVS_CB(skb)->tun_key)) {
+ err = -EINVAL;
+ goto error;
}
-out:
- return err;
-}
-static const struct tnl_ops ovs_vxlan_tnl_ops = {
- .tunnel_type = TNL_T_PROTO_VXLAN,
- .ipproto = IPPROTO_UDP,
- .hdr_len = vxlan_hdr_len,
- .build_header = vxlan_build_header,
-};
-
-static void vxlan_tnl_destroy(struct vport *vport)
-{
- struct vxlan_port *vxlan_port;
- struct tnl_vport *tnl_vport = tnl_vport_priv(vport);
- struct tnl_mutable_config *config;
-
- config = rtnl_dereference(tnl_vport->mutable);
+ forward_ip_summed(skb, true);
+
+ /* Route lookup */
+ saddr = OVS_CB(skb)->tun_key->ipv4_src;
+ rt = find_route(ovs_dp_get_net(vport->dp),
+ &saddr,
+ OVS_CB(skb)->tun_key->ipv4_dst,
+ IPPROTO_UDP,
+ OVS_CB(skb)->tun_key->ipv4_tos,
+ skb_get_mark(skb));
+ if (IS_ERR(rt)) {
+ err = PTR_ERR(rt);
+ goto error;
+ }
- vxlan_port = vxlan_port_exists(ovs_dp_get_net(vport->dp),
- config->dst_port);
+ df = OVS_CB(skb)->tun_key->tun_flags & TUNNEL_DONT_FRAGMENT ?
+ htons(IP_DF) : 0;
- vxlan_tunnel_release(vxlan_port);
+ skb->local_df = 1;
- ovs_tnl_destroy(vport);
+ inet_get_local_port_range(&port_min, &port_max);
+ src_port = vxlan_src_port(port_min, port_max, skb);
+
+ err = vxlan_xmit_skb(net, vxlan_port->vh, rt, skb,
+ saddr, OVS_CB(skb)->tun_key->ipv4_dst,
+ OVS_CB(skb)->tun_key->ipv4_tos,
+ OVS_CB(skb)->tun_key->ipv4_ttl, df,
+ src_port, dst_port,
+ htonl(be64_to_cpu(OVS_CB(skb)->tun_key->tun_id) << 8));
+ if (err < 0)
+ ip_rt_put(rt);
+error:
+ return err;
}
-static struct vport *vxlan_tnl_create(const struct vport_parms *parms)
+static const char *vxlan_get_name(const struct vport *vport)
{
- int err;
- struct vport *vport;
- struct vxlan_port *vxlan_port = NULL;
-
- err = vxlan_tunnel_setup(ovs_dp_get_net(parms->dp), parms->options,
- &vxlan_port);
- if (err)
- return ERR_PTR(err);
-
- vport = ovs_tnl_create(parms, &ovs_vxlan_vport_ops, &ovs_vxlan_tnl_ops);
-
- if (IS_ERR(vport))
- vxlan_tunnel_release(vxlan_port);
-
- return vport;
+ struct vxlan_port *vxlan_port = vxlan_vport(vport);
+ return vxlan_port->name;
}
const struct vport_ops ovs_vxlan_vport_ops = {
.type = OVS_VPORT_TYPE_VXLAN,
- .flags = VPORT_F_TUN_ID,
.create = vxlan_tnl_create,
.destroy = vxlan_tnl_destroy,
- .set_addr = ovs_tnl_set_addr,
- .get_name = ovs_tnl_get_name,
- .get_addr = ovs_tnl_get_addr,
- .get_options = ovs_tnl_get_options,
- .set_options = vxlan_set_options,
- .send = ovs_tnl_send,
+ .get_name = vxlan_get_name,
+ .get_options = vxlan_get_options,
+ .send = vxlan_tnl_send,
};
#else
#warning VXLAN tunneling will not be available on kernels before 2.6.26