[ovs-dev] [PATCH v2] datapath: Sync vxlan tunneling code with upstream ovs-vxlan.
Pravin B Shelar
pshelar at nicira.com
Mon Aug 26 18:05:35 UTC 2013
Upstream vxlan implementation was changed according to few
comments. Following patch brings back those changes to out
of tree ovs module.
Signed-off-by: Pravin B Shelar <pshelar at nicira.com>
---
v1-v2:
- remove vxlan-dev hash table and vs->refcnt.
- update copyright.
---
datapath/linux/compat/include/net/vxlan.h | 33 ++----
datapath/linux/compat/vxlan.c | 168 ++++++++---------------------
datapath/vport-vxlan.c | 36 +++----
3 files changed, 75 insertions(+), 162 deletions(-)
diff --git a/datapath/linux/compat/include/net/vxlan.h b/datapath/linux/compat/include/net/vxlan.h
index 102bc0c..46cbfb6 100644
--- a/datapath/linux/compat/include/net/vxlan.h
+++ b/datapath/linux/compat/include/net/vxlan.h
@@ -5,35 +5,26 @@
#include <linux/netdevice.h>
#include <linux/udp.h>
+struct vxlan_sock;
+typedef void (vxlan_rcv_t)(struct vxlan_sock *vs, struct sk_buff *skb, __be32 key);
+
/* per UDP socket information */
struct vxlan_sock {
struct hlist_node hlist;
- struct rcu_head rcu;
- struct socket *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;
- void *data;
- struct vxlan_sock *vs;
- atomic_t refcnt;
- struct rcu_head rcu;
+ vxlan_rcv_t *rcv;
+ void *data;
struct work_struct del_work;
- int priority;
+ struct socket *sock;
+ struct rcu_head rcu;
};
-void vxlan_handler_put(struct vxlan_handler *vh);
+struct vxlan_sock *vxlan_sock_add(struct net *net, __be16 port,
+ vxlan_rcv_t *rcv, void *data,
+ bool no_share);
-struct vxlan_handler *vxlan_handler_add(struct net *net,
- __be16 portno, vxlan_rcv_t *rcv,
- void *data, int priority, bool create);
+void vxlan_sock_release(struct vxlan_sock *vs);
-int vxlan_xmit_skb(struct net *net, struct vxlan_handler *vh,
+int vxlan_xmit_skb(struct net *net, struct vxlan_sock *vs,
struct rtable *rt, struct sk_buff *skb,
__be32 src, __be32 dst, __u8 tos, __u8 ttl, __be16 df,
__be16 src_port, __be16 dst_port, __be32 vni);
diff --git a/datapath/linux/compat/vxlan.c b/datapath/linux/compat/vxlan.c
index f3df4e3..0639ab1 100644
--- a/datapath/linux/compat/vxlan.c
+++ b/datapath/linux/compat/vxlan.c
@@ -75,10 +75,13 @@ struct vxlanhdr {
static int vxlan_net_id;
+static int vxlan_init_module(void);
+static void vxlan_cleanup_module(void);
+
/* per-network namespace private data for this module */
struct vxlan_net {
struct hlist_head sock_list[PORT_HASH_SIZE];
- struct mutex sock_lock; /* RTNL lock nests inside this lock. */
+ spinlock_t sock_lock;
};
/* Socket hash table head */
@@ -90,7 +93,8 @@ static inline struct hlist_head *vs_head(struct net *net, __be16 port)
}
/* Find VXLAN socket based on network namespace and UDP port */
-static struct vxlan_sock *vxlan_find_port(struct net *net, __be16 port)
+
+static struct vxlan_sock *vxlan_find_sock(struct net *net, __be16 port)
{
struct vxlan_sock *vs;
@@ -104,7 +108,6 @@ static struct vxlan_sock *vxlan_find_port(struct net *net, __be16 port)
/* Callback from net/ipv4/udp.c to receive packets */
static int vxlan_udp_encap_recv(struct sock *sk, struct sk_buff *skb)
{
- struct vxlan_handler *vh;
struct vxlan_sock *vs;
struct vxlanhdr *vxh;
@@ -124,14 +127,12 @@ static int vxlan_udp_encap_recv(struct sock *sk, struct sk_buff *skb)
if (iptunnel_pull_header(skb, VXLAN_HLEN, htons(ETH_P_TEB)))
goto drop;
- vs = vxlan_find_port(sock_net(sk), inet_sport(sk));
+ vs = vxlan_find_sock(sock_net(sk), inet_sport(sk));
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;
- }
+ vs->rcv(vs, skb, vxh->vx_vni);
+ return 0;
drop:
/* Consume bad packet */
@@ -211,7 +212,7 @@ static int handle_offloads(struct sk_buff *skb)
return 0;
}
-int vxlan_xmit_skb(struct net *net, struct vxlan_handler *vh,
+int vxlan_xmit_skb(struct net *net, struct vxlan_sock *vs,
struct rtable *rt, struct sk_buff *skb,
__be32 src, __be32 dst, __u8 tos, __u8 ttl, __be16 df,
__be16 src_port, __be16 dst_port, __be32 vni)
@@ -249,7 +250,7 @@ int vxlan_xmit_skb(struct net *net, struct vxlan_handler *vh,
uh->len = htons(skb->len);
uh->check = 0;
- vxlan_set_owner(vh->vs->sock->sk, skb);
+ vxlan_set_owner(vs->sock->sk, skb);
err = handle_offloads(skb);
if (err)
@@ -259,8 +260,19 @@ int vxlan_xmit_skb(struct net *net, struct vxlan_handler *vh,
IPPROTO_UDP, tos, ttl, df);
}
-static struct vxlan_sock *vxlan_socket_create(struct net *net, __be16 port)
+static void vxlan_del_work(struct work_struct *work)
+{
+ struct vxlan_sock *vs = container_of(work, struct vxlan_sock, del_work);
+
+ sk_release_kernel(vs->sock->sk);
+ kfree_rcu(vs, rcu);
+ vxlan_cleanup_module();
+}
+
+static struct vxlan_sock *vxlan_socket_create(struct net *net, __be16 port,
+ vxlan_rcv_t *rcv, void *data)
{
+ struct vxlan_net *vn = net_generic(net, vxlan_net_id);
struct vxlan_sock *vs;
struct sock *sk;
struct sockaddr_in vxlan_addr = {
@@ -271,8 +283,12 @@ static struct vxlan_sock *vxlan_socket_create(struct net *net, __be16 port)
int rc;
vs = kmalloc(sizeof(*vs), GFP_KERNEL);
- if (!vs)
+ if (!vs) {
+ pr_debug("memory alocation failure\n");
return ERR_PTR(-ENOMEM);
+ }
+
+ INIT_WORK(&vs->del_work, vxlan_del_work);
/* Create UDP socket for encapsulation receive. */
rc = sock_create_kern(AF_INET, SOCK_DGRAM, IPPROTO_UDP, &vs->sock);
@@ -287,81 +303,36 @@ static struct vxlan_sock *vxlan_socket_create(struct net *net, __be16 port)
sk_change_net(sk, net);
rc = kernel_bind(vs->sock, (struct sockaddr *) &vxlan_addr,
- sizeof(vxlan_addr));
+ sizeof(vxlan_addr));
if (rc < 0) {
pr_debug("bind for UDP socket %pI4:%u (%d)\n",
- &vxlan_addr.sin_addr, ntohs(vxlan_addr.sin_port), rc);
+ &vxlan_addr.sin_addr, ntohs(vxlan_addr.sin_port), rc);
sk_release_kernel(sk);
kfree(vs);
return ERR_PTR(rc);
}
+ vs->rcv = rcv;
+ vs->data = data;
/* Disable multicast loopback */
inet_sk(sk)->mc_loop = 0;
- INIT_LIST_HEAD(&vs->handler_list);
+ spin_lock(&vn->sock_lock);
hlist_add_head_rcu(&vs->hlist, vs_head(net, port));
+ spin_unlock(&vn->sock_lock);
/* Mark socket as an encapsulation socket. */
udp_sk(sk)->encap_type = 1;
udp_sk(sk)->encap_rcv = vxlan_udp_encap_recv;
udp_encap_enable();
-
return vs;
}
-static void rcu_free_vs_callback(struct rcu_head *rcu)
-{
- struct vxlan_sock *vs = container_of(rcu, struct vxlan_sock, rcu);
-
- kfree(vs);
-}
-
-static void vxlan_socket_del(struct vxlan_sock *vs)
-{
- if (list_empty(&vs->handler_list)) {
- hlist_del_rcu(&vs->hlist);
-
- sk_release_kernel(vs->sock->sk);
- call_rcu(&vs->rcu, rcu_free_vs_callback);
- }
-}
-
-static int vxlan_init_module(void);
-static void vxlan_cleanup_module(void);
-
-static void rcu_free_vh_callback(struct rcu_head *rcu)
-{
- struct vxlan_handler *vh = container_of(rcu, struct vxlan_handler, rcu);
-
- kfree(vh);
-}
-
-static void vh_del_work(struct work_struct *work)
-{
- struct vxlan_handler *vh = container_of(work, struct vxlan_handler, del_work);
- struct vxlan_sock *vs = vh->vs;
- struct net *net = sock_net(vs->sock->sk);
- struct vxlan_net *vn = net_generic(net, vxlan_net_id);
-
- mutex_lock(&vn->sock_lock);
-
- list_del_rcu(&vh->node);
- call_rcu(&vh->rcu, rcu_free_vh_callback);
- vxlan_socket_del(vs);
-
- mutex_unlock(&vn->sock_lock);
-
- vxlan_cleanup_module();
-}
-
-struct vxlan_handler *vxlan_handler_add(struct net *net,
- __be16 portno, vxlan_rcv_t *rcv,
- void *data, int priority, bool create)
+struct vxlan_sock *vxlan_sock_add(struct net *net, __be16 port,
+ vxlan_rcv_t *rcv, void *data,
+ bool no_share)
{
struct vxlan_net *vn;
struct vxlan_sock *vs;
- struct vxlan_handler *vh;
- struct vxlan_handler *new;
int err;
err = vxlan_init_module();
@@ -369,64 +340,19 @@ struct vxlan_handler *vxlan_handler_add(struct net *net,
return ERR_PTR(err);
vn = net_generic(net, vxlan_net_id);
- mutex_lock(&vn->sock_lock);
- /* Look to see if can reuse socket */
- vs = vxlan_find_port(net, portno);
- if (!vs) {
- vs = vxlan_socket_create(net, portno);
- if (IS_ERR(vs)) {
- new = (void *) vs;
- goto out;
- }
- }
-
- /* Try existing vxlan hanlders for this socket. */
- list_for_each_entry(vh, &vs->handler_list, node) {
- if (vh->rcv == rcv) {
- if (create) {
- vxlan_socket_del(vs);
- new = ERR_PTR(-EEXIST);
- goto out;
- }
- atomic_inc(&vh->refcnt);
- new = vh;
- goto out;
- }
- }
-
- new = kzalloc(sizeof(*new), GFP_KERNEL);
- if (!new) {
- vxlan_socket_del(vs);
- new = ERR_PTR(-ENOMEM);
- goto out;
- }
-
- new->rcv = rcv;
- new->vs = vs;
- atomic_set(&new->refcnt, 1);
- INIT_WORK(&new->del_work, vh_del_work);
- new->data = data;
- new->priority = priority;
-
- list_for_each_entry(vh, &vs->handler_list, node) {
- if (vh->priority > priority) {
- list_add_tail_rcu(&new->node, &vh->node);
- goto out;
- }
- }
-
- list_add_tail_rcu(&new->node, &vs->handler_list);
-out:
- mutex_unlock(&vn->sock_lock);
- return new;
+ vs = vxlan_socket_create(net, port, rcv, data);
+ return vs;
}
-void vxlan_handler_put(struct vxlan_handler *vh)
+void vxlan_sock_release(struct vxlan_sock *vs)
{
- BUG_ON(!vh->vs);
+ struct vxlan_net *vn = net_generic(sock_net(vs->sock->sk), vxlan_net_id);
+
+ spin_lock(&vn->sock_lock);
+ hlist_del_rcu(&vs->hlist);
+ spin_unlock(&vn->sock_lock);
- if (atomic_dec_and_test(&vh->refcnt))
- queue_work(&vh->del_work);
+ queue_work(&vs->del_work);
}
static int vxlan_init_net(struct net *net)
@@ -434,7 +360,7 @@ static int vxlan_init_net(struct net *net)
struct vxlan_net *vn = net_generic(net, vxlan_net_id);
unsigned int h;
- mutex_init(&vn->sock_lock);
+ spin_lock_init(&vn->sock_lock);
for (h = 0; h < PORT_HASH_SIZE; ++h)
INIT_HLIST_HEAD(&vn->sock_list[h]);
diff --git a/datapath/vport-vxlan.c b/datapath/vport-vxlan.c
index f3ef947..c614eff 100644
--- a/datapath/vport-vxlan.c
+++ b/datapath/vport-vxlan.c
@@ -1,6 +1,6 @@
/*
- * Copyright (c) 2011 Nicira, Inc.
- * Copyright (c) 2012 Cisco Systems, Inc.
+ * Copyright (c) 2013 Nicira, Inc.
+ * Copyright (c) 2013 Cisco Systems, Inc.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of version 2 of the GNU General Public
@@ -44,15 +44,13 @@
#include "datapath.h"
#include "vport.h"
-#define OVS_VXLAN_RCV_PRIORITY 8
-
/**
* struct vxlan_port - Keeps track of open UDP ports
- * @vh: vxlan_handler created for the port.
+ * @vs: vxlan_sock created for the port.
* @name: vport name.
*/
struct vxlan_port {
- struct vxlan_handler *vh;
+ struct vxlan_sock *vs;
char name[IFNAMSIZ];
};
@@ -62,11 +60,11 @@ static inline struct vxlan_port *vxlan_vport(const struct vport *vport)
}
/* Called with rcu_read_lock and BH disabled. */
-static int vxlan_rcv(struct vxlan_handler *vh, struct sk_buff *skb, __be32 vx_vni)
+static void vxlan_rcv(struct vxlan_sock *vs, struct sk_buff *skb, __be32 vx_vni)
{
- struct vport *vport = vh->data;
- struct iphdr *iph;
struct ovs_key_ipv4_tunnel tun_key;
+ struct vport *vport = vs->data;
+ struct iphdr *iph;
__be64 key;
/* Save outer tunnel values */
@@ -75,13 +73,12 @@ static int vxlan_rcv(struct vxlan_handler *vh, struct sk_buff *skb, __be32 vx_vn
ovs_flow_tun_key_init(&tun_key, iph, key, TUNNEL_KEY);
ovs_vport_receive(vport, skb, &tun_key);
- return PACKET_RCVD;
}
static int vxlan_get_options(const struct vport *vport, struct sk_buff *skb)
{
struct vxlan_port *vxlan_port = vxlan_vport(vport);
- __be16 dst_port = inet_sport(vxlan_port->vh->vs->sock->sk);
+ __be16 dst_port = inet_sport(vxlan_port->vs->sock->sk);
if (nla_put_u16(skb, OVS_TUNNEL_ATTR_DST_PORT, ntohs(dst_port)))
return -EMSGSIZE;
@@ -92,7 +89,7 @@ static void vxlan_tnl_destroy(struct vport *vport)
{
struct vxlan_port *vxlan_port = vxlan_vport(vport);
- vxlan_handler_put(vxlan_port->vh);
+ vxlan_sock_release(vxlan_port->vs);
ovs_vport_deferred_free(vport);
}
@@ -102,7 +99,7 @@ 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 vxlan_sock *vs;
struct vport *vport;
struct nlattr *a;
u16 dst_port;
@@ -129,13 +126,12 @@ static struct vport *vxlan_tnl_create(const struct vport_parms *parms)
vxlan_port = vxlan_vport(vport);
strncpy(vxlan_port->name, parms->name, IFNAMSIZ);
- vh = vxlan_handler_add(net, htons(dst_port), vxlan_rcv,
- vport, OVS_VXLAN_RCV_PRIORITY, true);
- if (IS_ERR(vh)) {
+ vs = vxlan_sock_add(net, htons(dst_port), vxlan_rcv, vport, true);
+ if (IS_ERR(vs)) {
ovs_vport_free(vport);
- return (void *)vh;
+ return (void *)vs;
}
- vxlan_port->vh = vh;
+ vxlan_port->vs = vs;
return vport;
@@ -146,7 +142,7 @@ error:
static int vxlan_tnl_send(struct vport *vport, struct sk_buff *skb)
{
struct vxlan_port *vxlan_port = vxlan_vport(vport);
- __be16 dst_port = inet_sport(vxlan_port->vh->vs->sock->sk);
+ __be16 dst_port = inet_sport(vxlan_port->vs->sock->sk);
struct net *net = ovs_dp_get_net(vport->dp);
struct rtable *rt;
__be16 src_port;
@@ -184,7 +180,7 @@ static int vxlan_tnl_send(struct vport *vport, struct sk_buff *skb)
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,
+ err = vxlan_xmit_skb(net, vxlan_port->vs, 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,
--
1.7.1
More information about the dev
mailing list