vxlan: add ipv6 support
Subsystems (3)
v5 cover vxlan: add ipv6 support
From: Cong Wang <[email protected]>
v5: make David happy on the names of the fields
fix my mistake during rebasing the patches
drop the scope_id patch, because it is broken
export in6addr_loopback
fix a udp checksum bug
rebased on the latest net-next
v4: rename ->sin to ->va_sin
rename ->sin6 to ->va_sin6
rename ->family to ->va_sa
support ll addr
fix more ugly #ifdef
rebased on the latest net-next
v3: fix many coding style issues
fix some ugly #ifdef
rename vxlan_ip to vxlan_addr
rename ->proto to ->family
rename ->ip4/->ip6 to ->sin/->sin6
v2: fix some compile error when !CONFIG_IPV6
improve some code based on Stephen's comments
use sockaddr suggested by David
Cong Wang (5):
vxlan: defer vxlan init as late as possible
ipv6: export ipv6_sock_mc_join and ipv6_sock_mc_drop
ipv6: export in6addr_loopback to modules
vxlan: add ipv6 support
ipv6: Add generic UDP Tunnel segmentation
drivers/net/vxlan.c | 625 ++++++++++++++++++++++++++++++++--------- include/uapi/linux/if_link.h | 2 + net/ipv6/addrconf.c | 9 - net/ipv6/addrconf_core.c | 9 + net/ipv6/ip6_offload.c | 4 +- net/ipv6/mcast.c | 2 + net/ipv6/udp_offload.c | 153 +++++++---- 7 files changed, 609 insertions(+), 195 deletions(-)
Signature · 1 line
1.7.7.6
PATCH v5 1/5 vxlan: defer vxlan init as late as possible
From: Cong Wang <[email protected]>
When vxlan is compiled as builtin, its init code runs before IPv6 init, this could cause problems if we create IPv6 socket in the latter patch.
--- drivers/net/vxlan.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/net/vxlan.c +1 -1
@@ -1635,7 +1635,7 @@ out2: out1:return rc; } -module_init (vxlan_init_module); +late_initcall (vxlan_init_module); static void __exit vxlan_cleanup_module(void) {
Signature · 1 line
1.7.7.6
PATCH v5 2/5 ipv6: export ipv6_sock_mc_join and ipv6_sock_mc_drop
In reply to Cong WangFrom: Cong Wang <[email protected]>
They will be used by vxlan module.
--- net/ipv6/mcast.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)
net/ipv6/mcast.c +2 -0
@@ -200,6 +200,7 @@ int ipv6_sock_mc_join(struct sock *sk, int ifindex, const struct in6_addr *addr)return 0 ; } +EXPORT_SYMBOL (ipv6_sock_mc_join);/* * socket leave on multicast group
@@ -246,6 +247,7 @@ int ipv6_sock_mc_drop(struct sock *sk, int ifindex, const struct in6_addr *addr)return -EADDRNOTAVAIL ; } +EXPORT_SYMBOL (ipv6_sock_mc_drop);/* called with rcu_read_lock() */ static struct inet6_dev *ip6_mc_find_dev_rcu(struct net *net,
Signature · 1 line
1.7.7.6
PATCH v5 3/5 ipv6: export in6addr_loopback to modules
In reply to Cong WangFrom: Cong Wang <[email protected]>
It is needed by vxlan module.
--- net/ipv6/addrconf.c | 9 --------- net/ipv6/addrconf_core.c | 9 +++++++++ 2 files changed, 9 insertions(+), 9 deletions(-)
net/ipv6/addrconf.c +0 -9
@@ -240,15 +240,6 @@ static struct ipv6_devconf ipv6_devconf_dflt __read_mostly = { .accept_dad = 1, }; -/* IPv6 Wildcard Address and Loopback Address defined by RFC2553 */ -const struct in6_addr in6addr_any =IN6ADDR_ANY_INIT ; -const struct in6_addr in6addr_loopback =IN6ADDR_LOOPBACK_INIT ; -const struct in6_addr in6addr_linklocal_allnodes =IN6ADDR_LINKLOCAL_ALLNODES_INIT ; -const struct in6_addr in6addr_linklocal_allrouters =IN6ADDR_LINKLOCAL_ALLROUTERS_INIT ; -const struct in6_addr in6addr_interfacelocal_allnodes =IN6ADDR_INTERFACELOCAL_ALLNODES_INIT ; -const struct in6_addr in6addr_interfacelocal_allrouters =IN6ADDR_INTERFACELOCAL_ALLROUTERS_INIT ; -const struct in6_addr in6addr_sitelocal_allrouters =IN6ADDR_SITELOCAL_ALLROUTERS_INIT ; -/* Check if a valid qdisc is available */ static inline bool addrconf_qdisc_ok (const struct net_device *dev) {
net/ipv6/addrconf_core.c +9 -0
@@ -78,3 +78,12 @@ int __ipv6_addr_type(const struct in6_addr *addr) }EXPORT_SYMBOL (__ipv6_addr_type); +/* IPv6 Wildcard Address and Loopback Address defined by RFC2553 */ +const struct in6_addr in6addr_loopback =IN6ADDR_LOOPBACK_INIT ; +EXPORT_SYMBOL (in6addr_loopback); +const struct in6_addr in6addr_any =IN6ADDR_ANY_INIT ; +const struct in6_addr in6addr_linklocal_allnodes =IN6ADDR_LINKLOCAL_ALLNODES_INIT ; +const struct in6_addr in6addr_linklocal_allrouters =IN6ADDR_LINKLOCAL_ALLROUTERS_INIT ; +const struct in6_addr in6addr_interfacelocal_allnodes =IN6ADDR_INTERFACELOCAL_ALLNODES_INIT ; +const struct in6_addr in6addr_interfacelocal_allrouters =IN6ADDR_INTERFACELOCAL_ALLROUTERS_INIT ; +const struct in6_addr in6addr_sitelocal_allrouters =IN6ADDR_SITELOCAL_ALLROUTERS_INIT ;
Signature · 1 line
1.7.7.6
PATCH v5 4/5 vxlan: add ipv6 support
In reply to Cong WangFrom: Cong Wang <[email protected]>
This patch adds IPv6 support to vxlan device, as the new version RFC already mentions it:
http://tools.ietf.org/html/draft-mahalingam-dutt-dcops-vxlan-03
--- drivers/net/vxlan.c | 623 +++++++++++++++++++++++++++++++++--------- include/uapi/linux/if_link.h | 2 + 2 files changed, 489 insertions(+), 136 deletions(-)
drivers/net/vxlan.c +487 -136
@@ -9,7 +9,6 @@* * TODO * - use IANA UDP port number (when defined) -* - IPv6 (not in RFC) */ #define pr_fmt (fmt) KBUILD_MODNAME ": " fmt
@@ -42,6 +41,11 @@#include <net/inet_ecn.h> #include <net/net_namespace.h> #include <net/netns/generic.h> +#if IS_ENABLED (CONFIG_IPV6 ) +#include <net/addrconf.h> +#include <net/ip6_route.h> +#include <net/ip6_tunnel.h> +#endif #define VXLAN_VERSION "0.1"
@@ -56,6 +60,7 @@#define VXLAN_VID_MASK (VXLAN_N_VID - 1)/* IP header + UDP + VXLAN + Ethernet header */ #define VXLAN_HEADROOM (20 + 8 + 8 + 14) +#define VXLAN6_HEADROOM (40 + 8 + 8 + 14)#define VXLAN_FLAGS 0x08000000/* struct vxlanhdr.vx_flags required value. */
@@ -81,9 +86,15 @@ struct vxlan_net {struct hlist_head vni_list[VNI_HASH_SIZE ]; }; +union vxlan_addr { +struct sockaddr_in sin; +struct sockaddr_in6 sin6; +struct sockaddr sa; +}; +struct vxlan_rdst {struct rcu_head rcu; -__be32 remote_ip; +union vxlan_addr remote_ip;__be16 remote_port;u32 remote_vni;u32 remote_ifindex;
@@ -106,7 +117,7 @@ struct vxlan_dev {struct hlist_node hlist;struct net_device *dev;struct vxlan_rdst default_dst;/* default destination */ -__be32 saddr;/* source address */ +union vxlan_addr saddr;/* source address */ __u16 port_min;/* source port range */ __u16 port_max;__u8 tos;/* TOS override */
@@ -128,6 +139,69 @@ struct vxlan_dev {#define VXLAN_F_L2MISS 0x08#define VXLAN_F_L3MISS 0x10 +static inline +bool vxlan_addr_equal (const union vxlan_addr *a,const union vxlan_addr *b) +{ +#if IS_ENABLED (CONFIG_IPV6 ) +if (a->sa.sa_family != b->sa.sa_family) +return false; +if (a->sa.sa_family ==AF_INET6 ) +return ipv6_addr_equal (&a->sin6.sin6_addr, &b->sin6.sin6_addr); + else +#endif +return a->sin.sin_addr.s_addr == b->sin.sin_addr.s_addr; +} + +static inline bool vxlan_addr_any (const union vxlan_addr *ipa) +{ +#if IS_ENABLED (CONFIG_IPV6 ) +if (ipa->sa.sa_family ==AF_INET6 ) +return ipv6_addr_any (&ipa->sin6.sin6_addr); + else +#endif +return ipa->sin.sin_addr.s_addr ==htonl (INADDR_ANY ); +} + +static inline bool vxlan_addr_multicast (const union vxlan_addr *ipa) +{ +#if IS_ENABLED (CONFIG_IPV6 ) +if (ipa->sa.sa_family ==AF_INET6 ) +return ipv6_addr_is_multicast (&ipa->sin6.sin6_addr); + else +#endif +return IN_MULTICAST (ntohl (ipa->sin.sin_addr.s_addr)); +} + +static int vxlan_nla_get_addr (union vxlan_addr *ip,struct nlattr *nla) +{ +if (nla_len (nla) >=sizeof (struct in6_addr )) { +#if IS_ENABLED (CONFIG_IPV6 ) +nla_memcpy (&ip->sin6.sin6_addr, nla,sizeof (struct in6_addr )); + ip->sa.sa_family =AF_INET6 ; +return 0 ; +#else +return -EAFNOSUPPORT ; +#endif + }else if (nla_len (nla) >=sizeof (__be32)) { + ip->sin.sin_addr.s_addr =nla_get_be32 (nla); + ip->sa.sa_family =AF_INET ; +return 0 ; + }else { +return -EAFNOSUPPORT ; + } +} + +static int vxlan_nla_put_addr (struct sk_buff *skb,int attr, +const union vxlan_addr *ip) +{ +#if IS_ENABLED (CONFIG_IPV6 ) +if (ip->sa.sa_family ==AF_INET6 ) +return nla_put (skb, attr,sizeof (struct in6_addr ), &ip->sin6.sin6_addr); + else +#endif +return nla_put_be32 (skb, attr, ip->sin.sin_addr.s_addr); +} +/* salt for hash table */ static u32 vxlan_salt__read_mostly ;
@@ -174,7 +248,7 @@ static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,if (type ==RTM_GETNEIGH ) { ndm->ndm_family =AF_INET ; - send_ip = rdst->remote_ip !=htonl (INADDR_ANY ); + send_ip = !vxlan_addr_any (&rdst->remote_ip); send_eth = !is_zero_ether_addr (fdb->eth_addr); }else ndm->ndm_family =AF_BRIDGE ;
@@ -186,7 +260,7 @@ static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,if (send_eth &&nla_put (skb,NDA_LLADDR ,ETH_ALEN , &fdb->eth_addr)) goto nla_put_failure; -if (send_ip &&nla_put_be32 (skb,NDA_DST , rdst->remote_ip)) +if (send_ip &&vxlan_nla_put_addr (skb,NDA_DST , &rdst->remote_ip)) goto nla_put_failure; if (rdst->remote_port && rdst->remote_port != vxlan_port &&
@@ -218,7 +292,7 @@ static inline size_t vxlan_nlmsg_size(void) {return NLMSG_ALIGN (sizeof (struct ndmsg )) +nla_total_size (ETH_ALEN )/* NDA_LLADDR */ - +nla_total_size (sizeof (__be32))/* NDA_DST */ + +nla_total_size (sizeof (struct in6_addr ))/* NDA_DST */ +nla_total_size (sizeof (__be32))/* NDA_PORT */ +nla_total_size (sizeof (__be32))/* NDA_VNI */ +nla_total_size (sizeof (__u32))/* NDA_IFINDEX */
@@ -251,14 +325,14 @@ errout:rtnl_set_sk_err (net,RTNLGRP_NEIGH , err); } -static void vxlan_ip_miss (struct net_device *dev,__be32 ipa) +static void vxlan_ip_miss (struct net_device *dev,union vxlan_addr *ipa) {struct vxlan_dev *vxlan =netdev_priv (dev);struct vxlan_fdb f;memset (&f,0 ,sizeof f); f.state =NUD_STALE ; - f.remote.remote_ip = ipa;/* goes to NDA_DST */ + f.remote.remote_ip = *ipa;/* goes to NDA_DST */ f.remote.remote_vni =VXLAN_N_VID ;vxlan_fdb_notify (vxlan, &f,RTM_GETNEIGH );
@@ -313,14 +387,14 @@ static struct vxlan_fdb *vxlan_find_mac(struct vxlan_dev *vxlan, }/* Add/update destinations for multicast */ -static int vxlan_fdb_append (struct vxlan_fdb *f, -__be32 ip,__u32 port,__u32 vni,__u32 ifindex) +static int vxlan_fdb_append (struct vxlan_fdb *f,union vxlan_addr *ip, +__u32 port,__u32 vni,__u32 ifindex) {struct vxlan_rdst *rd_prev, *rd; rd_prev =NULL ;for (rd = &f->remote; rd; rd = rd->remote_next) { -if (rd->remote_ip == ip && +if (vxlan_addr_equal (&rd->remote_ip, ip) && rd->remote_port == port && rd->remote_vni == vni && rd->remote_ifindex == ifindex)
@@ -330,7 +404,7 @@ static int vxlan_fdb_append(struct vxlan_fdb *f, rd =kmalloc (sizeof (*rd),GFP_ATOMIC );if (rd ==NULL )return -ENOBUFS ; - rd->remote_ip = ip; + rd->remote_ip = *ip; rd->remote_port = port; rd->remote_vni = vni; rd->remote_ifindex = ifindex;
@@ -341,7 +415,7 @@ static int vxlan_fdb_append(struct vxlan_fdb *f,/* Add new entry to forwarding table -- assumes lock held */ static int vxlan_fdb_create (struct vxlan_dev *vxlan, -const u8 *mac,__be32 ip, +const u8 *mac,union vxlan_addr *ip,__u16 state,__u16 flags,__u32 port,__u32 vni,__u32 ifindex) {
@@ -375,13 +449,18 @@ static int vxlan_fdb_create(struct vxlan_dev *vxlan,if (vxlan->addrmax && vxlan->addrcnt >= vxlan->addrmax)return -ENOSPC ; -netdev_dbg (vxlan->dev,"add %pM -> %pI4\n" , mac, &ip); +#if IS_ENABLED (CONFIG_IPV6 ) +if (ip->sa.sa_family ==AF_INET6 ) +netdev_dbg (vxlan->dev,"add %pM -> %pI6\n" , mac, &ip->sin6.sin6_addr); + else +#endif +netdev_dbg (vxlan->dev,"add %pM -> %pI4\n" , mac, &ip->sin.sin_addr.s_addr); f =kmalloc (sizeof (*f),GFP_ATOMIC );if (!f)return -ENOMEM ; notify =1 ; - f->remote.remote_ip = ip; + f->remote.remote_ip = *ip; f->remote.remote_port = port; f->remote.remote_vni = vni; f->remote.remote_ifindex = ifindex;
@@ -433,7 +512,7 @@ static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr *tb[], {struct vxlan_dev *vxlan =netdev_priv (dev);struct net *net =dev_net (vxlan->dev); -__be32 ip; +union vxlan_addr ip;u32 port, vni, ifindex;int err;
@@ -446,10 +525,9 @@ static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr *tb[],if (tb[NDA_DST ] ==NULL )return -EINVAL ; -if (nla_len (tb[NDA_DST ]) !=sizeof (__be32)) -return -EAFNOSUPPORT ; - - ip =nla_get_be32 (tb[NDA_DST ]); + err =vxlan_nla_get_addr (&ip, tb[NDA_DST ]); +if (err) +return err;if (tb[NDA_PORT ]) {if (nla_len (tb[NDA_PORT ]) !=sizeof (u32))
@@ -479,7 +557,7 @@ static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr *tb[], ifindex =0 ;spin_lock_bh (&vxlan->hash_lock); - err =vxlan_fdb_create (vxlan, addr, ip, ndm->ndm_state, flags, port, + err =vxlan_fdb_create (vxlan, addr, &ip, ndm->ndm_state, flags, port, vni, ifindex);spin_unlock_bh (&vxlan->hash_lock);
@@ -543,7 +621,7 @@ skip:* and Tunnel endpoint. */ static void vxlan_snoop (struct net_device *dev, -__be32 src_ip,const u8 *src_mac) +union vxlan_addr *src_ip,const u8 *src_mac) {struct vxlan_dev *vxlan =netdev_priv (dev);struct vxlan_fdb *f;
@@ -552,15 +630,25 @@ static void vxlan_snoop(struct net_device *dev, f =vxlan_find_mac (vxlan, src_mac);if (likely (f)) { f->used = jiffies; -if (likely (f->remote.remote_ip == src_ip)) +if (likely (vxlan_addr_equal (&f->remote.remote_ip, src_ip)))return ; -if (net_ratelimit ()) -netdev_info (dev, -"%pM migrated from %pI4 to %pI4\n" , - src_mac, &f->remote.remote_ip, &src_ip); +if (net_ratelimit ()) { +#if IS_ENABLED (CONFIG_IPV6 ) +if (src_ip->sa.sa_family ==AF_INET6 ) +netdev_info (dev, +"%pM migrated from %pI6 to %pI6\n" , + src_mac, &f->remote.remote_ip.sin6.sin6_addr, + &src_ip->sin6.sin6_addr); + else +#endif +netdev_info (dev, +"%pM migrated from %pI4 to %pI4\n" , + src_mac, &f->remote.remote_ip.sin.sin_addr.s_addr, + &src_ip->sin.sin_addr.s_addr); + } - f->remote.remote_ip = src_ip; + f->remote.remote_ip = *src_ip; f->updated = jiffies; }else {/* learned new entry */
@@ -589,7 +677,8 @@ static bool vxlan_group_used(struct vxlan_net *vn,if (!netif_running (vxlan->dev))continue ; -if (vxlan->default_dst.remote_ip == this->default_dst.remote_ip) +if (vxlan_addr_equal (&vxlan->default_dst.remote_ip, + &this->default_dst.remote_ip))return true; }
@@ -603,7 +692,7 @@ static int vxlan_join_group(struct net_device *dev)struct vxlan_net *vn =net_generic (dev_net (dev), vxlan_net_id);struct sock *sk = vn->sock->sk;struct ip_mreqn mreq = { - .imr_multiaddr.s_addr = vxlan->default_dst.remote_ip, + .imr_multiaddr.s_addr = vxlan->default_dst.remote_ip.sin.sin_addr.s_addr, .imr_ifindex = vxlan->default_dst.remote_ifindex, };int err;
@@ -615,7 +704,13 @@ static int vxlan_join_group(struct net_device *dev)/* Need to drop RTNL to call multicast join */ rtnl_unlock ();lock_sock (sk); - err =ip_mc_join_group (sk, &mreq); +#if IS_ENABLED (CONFIG_IPV6 ) +if (vxlan->default_dst.remote_ip.sa.sa_family ==AF_INET6 ) + err =ipv6_sock_mc_join (sk, vxlan->default_dst.remote_ifindex, + &vxlan->default_dst.remote_ip.sin6.sin6_addr); + else +#endif + err =ip_mc_join_group (sk, &mreq);release_sock (sk);rtnl_lock ();
@@ -631,7 +726,7 @@ static int vxlan_leave_group(struct net_device *dev)int err =0 ;struct sock *sk = vn->sock->sk;struct ip_mreqn mreq = { - .imr_multiaddr.s_addr = vxlan->default_dst.remote_ip, + .imr_multiaddr.s_addr = vxlan->default_dst.remote_ip.sin.sin_addr.s_addr, .imr_ifindex = vxlan->default_dst.remote_ifindex, };
@@ -642,7 +737,13 @@ static int vxlan_leave_group(struct net_device *dev)/* Need to drop RTNL to call multicast leave */ rtnl_unlock ();lock_sock (sk); - err =ip_mc_leave_group (sk, &mreq); +#if IS_ENABLED (CONFIG_IPV6 ) +if (vxlan->default_dst.remote_ip.sa.sa_family ==AF_INET6 ) + err =ipv6_sock_mc_drop (sk, vxlan->default_dst.remote_ifindex, + &vxlan->default_dst.remote_ip.sin6.sin6_addr); + else +#endif + err =ip_mc_leave_group (sk, &mreq);release_sock (sk);rtnl_lock ();
@@ -652,12 +753,16 @@ static int vxlan_leave_group(struct net_device *dev)/* 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 iphdr *oip =NULL ; +#if IS_ENABLED (CONFIG_IPV6 ) +struct ipv6hdr *oip6 =NULL ; +#endif struct vxlanhdr *vxh;struct vxlan_dev *vxlan;struct pcpu_tstats *stats; +union vxlan_addr src_ip;__u32 vni; -int err; +int err =0 ;/* pop off outer UDP header */ __skb_pull (skb,sizeof (struct udphdr ));
@@ -694,7 +799,13 @@ static int vxlan_udp_encap_recv(struct sock *sk, struct sk_buff *skb)skb_reset_mac_header (skb);/* Re-examine inner Ethernet packet */ - oip =ip_hdr (skb); +if (skb->protocol ==htons (ETH_P_IP )) + oip =ip_hdr (skb); +#if IS_ENABLED (CONFIG_IPV6 ) +if (skb->protocol ==htons (ETH_P_IPV6 )) + oip6 =ipv6_hdr (skb); +#endif + skb->protocol =eth_type_trans (skb, vxlan->dev);/* Ignore packet loops (and multicast echo) */
@@ -702,8 +813,19 @@ static int vxlan_udp_encap_recv(struct sock *sk, struct sk_buff *skb) vxlan->dev->dev_addr) == 0) goto drop; -if (vxlan->flags &VXLAN_F_LEARN ) -vxlan_snoop (skb->dev, oip->saddr,eth_hdr (skb)->h_source); +if (vxlan->flags &VXLAN_F_LEARN ) { +if (oip) { + src_ip.sin.sin_addr.s_addr = oip->saddr; + src_ip.sa.sa_family =AF_INET ; + } +#if IS_ENABLED (CONFIG_IPV6 ) +if (oip6) { + src_ip.sin6.sin6_addr = oip6->saddr; + src_ip.sa.sa_family =AF_INET6 ; + } +#endif +vxlan_snoop (skb->dev, &src_ip,eth_hdr (skb)->h_source); + }__skb_tunnel_rx (skb, vxlan->dev);skb_reset_network_header (skb);
@@ -719,11 +841,24 @@ static int vxlan_udp_encap_recv(struct sock *sk, struct sk_buff *skb) skb->encapsulation =0 ; - err =IP_ECN_decapsulate (oip, skb); +#if IS_ENABLED (CONFIG_IPV6 ) +if (oip6) + err =IP6_ECN_decapsulate (oip6, skb); +#endif +if (oip) + err =IP_ECN_decapsulate (oip, skb); +if (unlikely (err)) { -if (log_ecn_error) -net_info_ratelimited ("non-ECT from %pI4 with TOS=%#x\n" , - &oip->saddr, oip->tos); +if (log_ecn_error) { +#if IS_ENABLED (CONFIG_IPV6 ) +if (oip6) +net_info_ratelimited ("non-ECT from %pI6\n" , + &oip6->saddr); +#endif +if (oip) +net_info_ratelimited ("non-ECT from %pI4 with TOS=%#x\n" , + &oip->saddr, oip->tos); + }if (err >1 ) { ++vxlan->dev->stats.rx_frame_errors; ++vxlan->dev->stats.rx_errors;
@@ -758,6 +893,7 @@ static int arp_reduce(struct net_device *dev, struct sk_buff *skb)u8 *arpptr, *sha;__be32 sip, tip;struct neighbour *n; +union vxlan_addr ipa;if (dev->flags &IFF_NOARP ) goto out;
@@ -799,7 +935,7 @@ static int arp_reduce(struct net_device *dev, struct sk_buff *skb) } f =vxlan_find_mac (vxlan, n->ha); -if (f && f->remote.remote_ip ==htonl (INADDR_ANY )) { +if (f &&vxlan_addr_any (&f->remote.remote_ip)) {/* bridge-local neighbor */ neigh_release (n); goto out;
@@ -817,8 +953,11 @@ static int arp_reduce(struct net_device *dev, struct sk_buff *skb)if (netif_rx_ni (reply) ==NET_RX_DROP ) dev->stats.rx_dropped++; - }else if (vxlan->flags &VXLAN_F_L3MISS ) -vxlan_ip_miss (dev, tip); + }else if (vxlan->flags &VXLAN_F_L3MISS ) { + ipa.sin.sin_addr.s_addr = tip; + ipa.sa.sa_family =AF_INET ; +vxlan_ip_miss (dev, &ipa); + } out:consume_skb (skb);return NETDEV_TX_OK ;
@@ -840,6 +979,14 @@ static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb)return false; pip =ip_hdr (skb); n =neigh_lookup (&arp_tbl, &pip->daddr, dev); +if (!n && vxlan->flags &VXLAN_F_L3MISS ) { +union vxlan_addr ipa; + ipa.sin.sin_addr.s_addr = pip->daddr; + ipa.sa.sa_family =AF_INET ; +vxlan_ip_miss (dev, &ipa); +return false; + } +break ;default :return false;
@@ -856,8 +1003,8 @@ static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb) }neigh_release (n);return diff; - }else if (vxlan->flags &VXLAN_F_L3MISS ) -vxlan_ip_miss (dev, pip->daddr); + } +return false; }
@@ -867,7 +1014,8 @@ static void vxlan_sock_free(struct sk_buff *skb) }/* On transmit, associate with the tunnel socket */ -static void vxlan_set_owner (struct net_device *dev,struct sk_buff *skb) +static inline void vxlan_set_owner (struct net_device *dev, +struct sk_buff *skb) {struct vxlan_net *vn =net_generic (dev_net (dev), vxlan_net_id);struct sock *sk = vn->sock->sk;
@@ -916,15 +1064,26 @@ static void vxlan_encap_bypass(struct sk_buff *skb, struct vxlan_dev *src_vxlan, {struct pcpu_tstats *tx_stats =this_cpu_ptr (src_vxlan->dev->tstats);struct pcpu_tstats *rx_stats =this_cpu_ptr (dst_vxlan->dev->tstats); +union vxlan_addr loopback; skb->pkt_type =PACKET_HOST ; skb->encapsulation =0 ; skb->dev = dst_vxlan->dev;__skb_pull (skb,skb_network_offset (skb)); +if (dst_vxlan->default_dst.remote_ip.sa.sa_family ==AF_INET ) { + loopback.sin.sin_addr.s_addr =htonl (INADDR_LOOPBACK ); + loopback.sa.sa_family =AF_INET ; + } +#if IS_ENABLED (CONFIG_IPV6 ) + else { + loopback.sin6.sin6_addr = in6addr_loopback; + loopback.sa.sa_family =AF_INET6 ; + } +#endif +if (dst_vxlan->flags &VXLAN_F_LEARN ) -vxlan_snoop (skb->dev,htonl (INADDR_LOOPBACK ), -eth_hdr (skb)->h_source); +vxlan_snoop (skb->dev, &loopback,eth_hdr (skb)->h_source);u64_stats_update_begin (&tx_stats->syncp); tx_stats->tx_packets++;
@@ -946,22 +1105,29 @@ static netdev_tx_t vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev, {struct vxlan_dev *vxlan =netdev_priv (dev);struct rtable *rt; -const struct iphdr *old_iph; +const struct iphdr *old_iph =NULL ;struct iphdr *iph;struct vxlanhdr *vxh;struct udphdr *uh;struct flowi4 fl4; -__be32 dst; -__u16 src_port, dst_port; +#if IS_ENABLED (CONFIG_IPV6 ) +struct flowi6 fl6; +struct vxlan_net *vn =net_generic (dev_net (dev), vxlan_net_id); +struct sock *sk = vn->sock->sk; +struct ipv6hdr *ip6h; +#endif +const union vxlan_addr *dst; +struct dst_entry *ndst =NULL ; +__u16 src_port =0 , dst_port;u32 vni;__be16 df =0 ;__u8 tos, ttl; dst_port = rdst->remote_port ? rdst->remote_port : vxlan_port; vni = rdst->remote_vni; - dst = rdst->remote_ip; + dst = &rdst->remote_ip; -if (!dst) { +if (vxlan_addr_any (dst)) {if (did_rsc) {/* short-circuited back to local bridge */ vxlan_encap_bypass (skb, vxlan, vxlan);
@@ -975,60 +1141,115 @@ static netdev_tx_t vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev, skb->encapsulation =1 ; } -/* Need space for new headers (invalidates iph ptr) */ -if (skb_cow_head (skb,VXLAN_HEADROOM )) - goto drop; + ttl = vxlan->ttl; + tos = vxlan->tos; +if (dst->sa.sa_family ==AF_INET ) { +/* Need space for new headers (invalidates iph ptr) */ +if (skb_cow_head (skb,VXLAN_HEADROOM )) + goto drop; - old_iph =ip_hdr (skb); + old_iph =ip_hdr (skb); +if (!ttl &&IN_MULTICAST (ntohl (dst->sin.sin_addr.s_addr))) + ttl =1 ; - ttl = vxlan->ttl; -if (!ttl &&IN_MULTICAST (ntohl (dst))) - ttl =1 ; +if (tos ==1 ) + tos =ip_tunnel_get_dsfield (old_iph, skb); - tos = vxlan->tos; -if (tos ==1 ) - tos =ip_tunnel_get_dsfield (old_iph, skb); - - src_port =vxlan_src_port (vxlan, skb); - -memset (&fl4,0 ,sizeof (fl4)); - fl4.flowi4_oif = rdst->remote_ifindex; - fl4.flowi4_tos =RT_TOS (tos); - fl4.daddr = dst; - fl4.saddr = vxlan->saddr; - - rt =ip_route_output_key (dev_net (dev), &fl4); -if (IS_ERR (rt)) { -netdev_dbg (dev,"no route to %pI4\n" , &dst); - dev->stats.tx_carrier_errors++; - goto tx_error; - } + src_port =vxlan_src_port (vxlan, skb); -if (rt->dst.dev == dev) { -netdev_dbg (dev,"circular route to %pI4\n" , &dst); -ip_rt_put (rt); - dev->stats.collisions++; - goto tx_error; - } +memset (&fl4,0 ,sizeof (fl4)); + fl4.flowi4_oif = rdst->remote_ifindex; + fl4.flowi4_tos =RT_TOS (tos); + fl4.daddr = dst->sin.sin_addr.s_addr; + fl4.saddr = vxlan->saddr.sin.sin_addr.s_addr; + + rt =ip_route_output_key (dev_net (dev), &fl4); +if (IS_ERR (rt)) { +netdev_dbg (dev,"no route to %pI4\n" , &dst->sin.sin_addr.s_addr); + dev->stats.tx_carrier_errors++; + goto tx_error; + } -/* Bypass encapsulation if the destination is local */ -if (rt->rt_flags &RTCF_LOCAL && - !(rt->rt_flags & (RTCF_BROADCAST |RTCF_MULTICAST ))) { -struct vxlan_dev *dst_vxlan; +if (rt->dst.dev == dev) { +netdev_dbg (dev,"circular route to %pI4\n" , &dst->sin.sin_addr.s_addr); +ip_rt_put (rt); + dev->stats.collisions++; + goto tx_error; + } + +/* Bypass encapsulation if the destination is local */ +if (rt->rt_flags &RTCF_LOCAL && + !(rt->rt_flags & (RTCF_BROADCAST |RTCF_MULTICAST ))) { +struct vxlan_dev *dst_vxlan; + +ip_rt_put (rt); + dst_vxlan =vxlan_find_vni (dev_net (dev), vni); +if (!dst_vxlan) + goto tx_error; +vxlan_encap_bypass (skb, vxlan, dst_vxlan); +return NETDEV_TX_OK ; + } + + ndst = &rt->dst; + }else { +#if IS_ENABLED (CONFIG_IPV6 ) +const struct ipv6hdr *old_iph6; +u32 flags; + +/* Need space for new headers (invalidates iph ptr) */ +if (skb_cow_head (skb,VXLAN6_HEADROOM )) + goto drop; + + old_iph6 =ipv6_hdr (skb); +if (!ttl &&ipv6_addr_is_multicast (&dst->sin6.sin6_addr)) + ttl =1 ; -ip_rt_put (rt); - dst_vxlan =vxlan_find_vni (dev_net (dev), vni); -if (!dst_vxlan) +if (tos ==1 ) + tos =ipv6_get_dsfield (old_iph6); + + src_port =vxlan_src_port (vxlan, skb); + +memset (&fl6,0 ,sizeof (fl6)); + fl6.flowi6_oif = rdst->remote_ifindex; + fl6.flowi6_tos =RT_TOS (tos); + fl6.daddr = dst->sin6.sin6_addr; + fl6.saddr = vxlan->saddr.sin6.sin6_addr; + fl6.flowi6_proto = skb->protocol; + +if (ip6_dst_lookup (sk, &ndst, &fl6)) { +netdev_dbg (dev,"no route to %pI6\n" , &dst->sin6.sin6_addr); + dev->stats.tx_carrier_errors++; + goto tx_error; + } + +if (ndst->dev == dev) { +netdev_dbg (dev,"circular route to %pI6\n" , &dst->sin6.sin6_addr); +dst_release (ndst); + dev->stats.collisions++; goto tx_error; -vxlan_encap_bypass (skb, vxlan, dst_vxlan); -return NETDEV_TX_OK ; + } + +/* Bypass encapsulation if the destination is local */ + flags = ((struct rt6_info *)ndst)->rt6i_flags; +if (flags &RTF_LOCAL && + !(flags & (RTCF_BROADCAST |RTCF_MULTICAST ))) { +struct vxlan_dev *dst_vxlan; + +dst_release (ndst); + dst_vxlan =vxlan_find_vni (dev_net (dev), vni); +if (!dst_vxlan) + goto tx_error; +vxlan_encap_bypass (skb, vxlan, dst_vxlan); +return NETDEV_TX_OK ; + } +#endif }memset (&(IPCB (skb)->opt),0 ,sizeof (IPCB (skb)->opt));IPCB (skb)->flags &= ~(IPSKB_XFRM_TUNNEL_SIZE |IPSKB_XFRM_TRANSFORMED |IPSKB_REROUTED );skb_dst_drop (skb); -skb_dst_set (skb, &rt->dst); +skb_dst_set (skb, ndst); vxh = (struct vxlanhdr *)__skb_push (skb,sizeof (*vxh)); vxh->vx_flags =htonl (VXLAN_FLAGS );
@@ -1044,27 +1265,63 @@ static netdev_tx_t vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev, uh->len =htons (skb->len); uh->check =0 ; -__skb_push (skb,sizeof (*iph)); -skb_reset_network_header (skb); - iph =ip_hdr (skb); - iph->version =4 ; - iph->ihl =sizeof (struct iphdr ) >>2 ; - iph->frag_off = df; - iph->protocol =IPPROTO_UDP ; - iph->tos =ip_tunnel_ecn_encap (tos, old_iph, skb); - iph->daddr = dst; - iph->saddr = fl4.saddr; - iph->ttl = ttl ? :ip4_dst_hoplimit (&rt->dst); -tunnel_ip_select_ident (skb, old_iph, &rt->dst); - -nf_reset (skb); +if (dst->sa.sa_family ==AF_INET ) { +__skb_push (skb,sizeof (*iph)); +skb_reset_network_header (skb); + iph =ip_hdr (skb); + iph->version =4 ; + iph->ihl =sizeof (struct iphdr ) >>2 ; + iph->frag_off = df; + iph->protocol =IPPROTO_UDP ; + iph->tos =ip_tunnel_ecn_encap (tos, old_iph, skb); + iph->daddr = dst->sin.sin_addr.s_addr; + iph->saddr = fl4.saddr; + iph->ttl = ttl ? :ip4_dst_hoplimit (ndst); +tunnel_ip_select_ident (skb, old_iph, ndst); + }else { +#if IS_ENABLED (CONFIG_IPV6 ) +if (!skb_is_gso (skb) && !(ndst->dev->features &NETIF_F_IPV6_CSUM )) { +__wsum csum =skb_checksum (skb,0 , skb->len,0 ); + skb->ip_summed =CHECKSUM_UNNECESSARY ; + uh->check =csum_ipv6_magic (&fl6.saddr, &fl6.daddr, skb->len, +IPPROTO_UDP , csum); +if (uh->check ==0 ) + uh->check =CSUM_MANGLED_0 ; + }else { + skb->ip_summed =CHECKSUM_PARTIAL ; + skb->csum_start =skb_transport_header (skb) - skb->head; + skb->csum_offset = offsetof(struct udphdr , check); + uh->check = ~csum_ipv6_magic (&fl6.saddr, &fl6.daddr, + skb->len,IPPROTO_UDP ,0 ); + } + +__skb_push (skb,sizeof (*ip6h)); +skb_reset_network_header (skb); + ip6h =ipv6_hdr (skb); + ip6h->version =6 ; + ip6h->priority =0 ; + ip6h->flow_lbl[0 ] =0 ; + ip6h->flow_lbl[1 ] =0 ; + ip6h->flow_lbl[2 ] =0 ; + ip6h->payload_len =htons (skb->len); + ip6h->nexthdr =IPPROTO_UDP ; + ip6h->hop_limit = ttl ? :ip6_dst_hoplimit (ndst); + ip6h->daddr = fl6.daddr; + ip6h->saddr = fl6.saddr; +#endif + }vxlan_set_owner (dev, skb);if (handle_offloads (skb)) goto drop; -iptunnel_xmit (skb, dev); +#if IS_ENABLED (CONFIG_IPV6 ) +if (dst->sa.sa_family ==AF_INET6 ) +ip6tunnel_xmit (skb, dev); + else +#endif +iptunnel_xmit (skb, dev);return NETDEV_TX_OK ; drop:
@@ -1106,7 +1363,7 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev) did_rsc = false; rdst0 = &vxlan->default_dst; -if (rdst0->remote_ip ==htonl (INADDR_ANY ) && +if (vxlan_addr_any (&rdst0->remote_ip) && (vxlan->flags &VXLAN_F_L2MISS ) && !is_multicast_ether_addr (eth->h_dest))vxlan_fdb_miss (vxlan, eth->h_dest);
@@ -1184,7 +1441,7 @@ static int vxlan_open(struct net_device *dev)struct vxlan_dev *vxlan =netdev_priv (dev);int err; -if (IN_MULTICAST (ntohl (vxlan->default_dst.remote_ip))) { +if (vxlan_addr_multicast (&vxlan->default_dst.remote_ip)) { err =vxlan_join_group (dev);if (err)return err;
@@ -1218,7 +1475,7 @@ static int vxlan_stop(struct net_device *dev) {struct vxlan_dev *vxlan =netdev_priv (dev); -if (IN_MULTICAST (ntohl (vxlan->default_dst.remote_ip))) +if (vxlan_addr_multicast (&vxlan->default_dst.remote_ip))vxlan_leave_group (dev);del_timer_sync (&vxlan->age_timer);
@@ -1268,7 +1525,12 @@ static void vxlan_setup(struct net_device *dev)eth_hw_addr_random (dev);ether_setup (dev); - dev->hard_header_len =ETH_HLEN +VXLAN_HEADROOM ; +#if IS_ENABLED (CONFIG_IPV6 ) +if (vxlan->default_dst.remote_ip.sa.sa_family ==AF_INET6 ) + dev->hard_header_len =ETH_HLEN +VXLAN6_HEADROOM ; + else +#endif + dev->hard_header_len =ETH_HLEN +VXLAN_HEADROOM ; dev->netdev_ops = &vxlan_netdev_ops; dev->destructor = vxlan_free;
@@ -1305,8 +1567,10 @@ static void vxlan_setup(struct net_device *dev)static const struct nla_policy vxlan_policy[IFLA_VXLAN_MAX +1 ] = { [IFLA_VXLAN_ID ] = { .type =NLA_U32 }, [IFLA_VXLAN_REMOTE] = { .len = FIELD_SIZEOF(struct iphdr, daddr) }, + [IFLA_VXLAN_REMOTE6 ] = { .len =sizeof (struct in6_addr ) }, [IFLA_VXLAN_LINK ] = { .type =NLA_U32 }, [IFLA_VXLAN_LOCAL] = { .len = FIELD_SIZEOF(struct iphdr, saddr) }, + [IFLA_VXLAN_LOCAL6 ] = { .len =sizeof (struct in6_addr ) }, [IFLA_VXLAN_TOS ] = { .type =NLA_U8 }, [IFLA_VXLAN_TTL ] = { .type =NLA_U8 }, [IFLA_VXLAN_LEARNING ] = { .type =NLA_U8 },
@@ -1386,11 +1650,31 @@ static int vxlan_newlink(struct net *net, struct net_device *dev, } dst->remote_vni = vni; -if (data[IFLA_VXLAN_REMOTE ]) - dst->remote_ip =nla_get_be32 (data[IFLA_VXLAN_REMOTE ]); +if (data[IFLA_VXLAN_REMOTE ]) { + dst->remote_ip.sin.sin_addr.s_addr =nla_get_be32 (data[IFLA_VXLAN_REMOTE ]); + dst->remote_ip.sa.sa_family =AF_INET ; + }else if (data[IFLA_VXLAN_REMOTE6 ]) { +#if IS_ENABLED (CONFIG_IPV6 ) +nla_memcpy (&dst->remote_ip.sin6.sin6_addr, data[IFLA_VXLAN_REMOTE6 ], +sizeof (struct in6_addr )); + dst->remote_ip.sa.sa_family =AF_INET6 ; +#else +return -EPFNOSUPPORT ; +#endif + } -if (data[IFLA_VXLAN_LOCAL ]) - vxlan->saddr =nla_get_be32 (data[IFLA_VXLAN_LOCAL ]); +if (data[IFLA_VXLAN_LOCAL ]) { + vxlan->saddr.sin.sin_addr.s_addr =nla_get_be32 (data[IFLA_VXLAN_LOCAL ]); + vxlan->saddr.sa.sa_family =AF_INET ; + }else if (data[IFLA_VXLAN_LOCAL6 ]) { +#if IS_ENABLED (CONFIG_IPV6 ) +nla_memcpy (&vxlan->saddr.sin6.sin6_addr, data[IFLA_VXLAN_LOCAL6 ], +sizeof (struct in6_addr )); + vxlan->saddr.sa.sa_family =AF_INET6 ; +#else +return -EPFNOSUPPORT ; +#endif + }if (data[IFLA_VXLAN_LINK ] && (dst->remote_ifindex =nla_get_u32 (data[IFLA_VXLAN_LINK ]))) {
@@ -1468,9 +1752,9 @@ static size_t vxlan_get_size(const struct net_device *dev) {return nla_total_size (sizeof (__u32)) +/* IFLA_VXLAN_ID */ -nla_total_size (sizeof (__be32)) +/* IFLA_VXLAN_REMOTE */ +nla_total_size (sizeof (struct in6_addr )) +/* IFLA_VXLAN_REMOTE{6} */ nla_total_size (sizeof (__u32)) +/* IFLA_VXLAN_LINK */ -nla_total_size (sizeof (__be32))+/* IFLA_VXLAN_LOCAL */ +nla_total_size (sizeof (struct in6_addr )) +/* IFLA_VXLAN_LOCAL{6} */ nla_total_size (sizeof (__u8)) +/* IFLA_VXLAN_TTL */ nla_total_size (sizeof (__u8)) +/* IFLA_VXLAN_TOS */ nla_total_size(sizeof(__u8)) + /* IFLA_VXLAN_LEARNING */
@@ -1496,14 +1780,34 @@ static int vxlan_fill_info(struct sk_buff *skb, const struct net_device *dev)if (nla_put_u32 (skb,IFLA_VXLAN_ID , dst->remote_vni)) goto nla_put_failure; -if (dst->remote_ip &&nla_put_be32 (skb,IFLA_VXLAN_REMOTE , dst->remote_ip)) - goto nla_put_failure; +if (!vxlan_addr_any (&dst->remote_ip)) { +if (dst->remote_ip.sa.sa_family ==AF_INET ) { +if (nla_put_be32 (skb,IFLA_VXLAN_REMOTE , dst->remote_ip.sin.sin_addr.s_addr)) + goto nla_put_failure; + }else { +#if IS_ENABLED (CONFIG_IPV6 ) +if (nla_put (skb,IFLA_VXLAN_REMOTE6 ,sizeof (struct in6_addr ), + &dst->remote_ip.sin6.sin6_addr)) + goto nla_put_failure; +#endif + } + }if (dst->remote_ifindex &&nla_put_u32 (skb,IFLA_VXLAN_LINK , dst->remote_ifindex)) goto nla_put_failure; -if (vxlan->saddr &&nla_put_be32 (skb,IFLA_VXLAN_LOCAL , vxlan->saddr)) - goto nla_put_failure; +if (!vxlan_addr_any (&vxlan->saddr)) { +if (vxlan->saddr.sa.sa_family ==AF_INET ) { +if (nla_put_be32 (skb,IFLA_VXLAN_LOCAL , vxlan->saddr.sin.sin_addr.s_addr)) + goto nla_put_failure; + }else { +#if IS_ENABLED (CONFIG_IPV6 ) +if (nla_put (skb,IFLA_VXLAN_LOCAL6 ,sizeof (struct in6_addr ), + &vxlan->saddr.sin6.sin6_addr)) + goto nla_put_failure; +#endif + } + } if (nla_put_u8(skb, IFLA_VXLAN_TTL, vxlan->ttl) || nla_put_u8(skb, IFLA_VXLAN_TOS, vxlan->tos) ||
@@ -1542,38 +1846,82 @@ static struct rtnl_link_ops vxlan_link_ops __read_mostly = { .fill_info = vxlan_fill_info, }; -static __net_init int vxlan_init_net (struct net *net) +/* Create UDP socket for encapsulation receive. AF_INET6 socket +* could be used for both IPv4 and IPv6 communications. +*/ +#if IS_ENABLED (CONFIG_IPV6 ) +static __net_init int create_sock (struct net *net,struct sock **sk) +{ +struct vxlan_net *vn =net_generic (net, vxlan_net_id); +struct sockaddr_in6 vxlan_addr = { + .sin6_family =AF_INET6 , + .sin6_port =htons (vxlan_port), + }; +int rc; + + rc =sock_create_kern (AF_INET6 ,SOCK_DGRAM ,IPPROTO_UDP , &vn->sock); +if (rc <0 ) { +pr_debug ("UDP socket create failed\n" ); +return rc; + } +/* Put in proper namespace */ + *sk = vn->sock->sk; +sk_change_net (*sk, net); + + rc =kernel_bind (vn->sock, (struct sockaddr *)&vxlan_addr, +sizeof (struct sockaddr_in6 )); +if (rc <0 ) { +pr_debug ("bind for UDP socket %pI6:%u (%d)\n" , + &vxlan_addr.sin6_addr,ntohs (vxlan_addr.sin6_port), rc); +sk_release_kernel (*sk); + vn->sock =NULL ; +return rc; + } +return 0 ; +} +#else +static __net_init int create_sock (struct net *net,struct sock **sk) {struct vxlan_net *vn =net_generic (net, vxlan_net_id); -struct sock *sk;struct sockaddr_in vxlan_addr = { .sin_family =AF_INET , + .sin_port =htons (vxlan_port), .sin_addr.s_addr =htonl (INADDR_ANY ), };int rc; -unsigned h; -/* Create UDP socket for encapsulation receive. */ rc =sock_create_kern (AF_INET ,SOCK_DGRAM ,IPPROTO_UDP , &vn->sock);if (rc <0 ) {pr_debug ("UDP socket create failed\n" );return rc; }/* Put in proper namespace */ - sk = vn->sock->sk; -sk_change_net (sk, net); - - vxlan_addr.sin_port =htons (vxlan_port); + *sk = vn->sock->sk; +sk_change_net (*sk, net); - rc =kernel_bind (vn->sock, (struct sockaddr *) &vxlan_addr, -sizeof (vxlan_addr)); + rc =kernel_bind (vn->sock, (struct sockaddr *)&vxlan_addr, +sizeof (struct sockaddr_in ));if (rc <0 ) {pr_debug ("bind for UDP socket %pI4:%u (%d)\n" , &vxlan_addr.sin_addr,ntohs (vxlan_addr.sin_port), rc); -sk_release_kernel (sk); +sk_release_kernel (*sk); vn->sock =NULL ;return rc; } +return 0 ; +} +#endif + +static __net_init int vxlan_init_net (struct net *net) +{ +struct vxlan_net *vn =net_generic (net, vxlan_net_id); +struct sock *sk; +int rc; +unsigned h; + + rc =create_sock (net, &sk); +if (rc <0 ) +return rc;/* Disable multicast loopback */ inet_sk (sk)->mc_loop =0 ;
@@ -1582,6 +1930,9 @@ static __net_init int vxlan_init_net(struct net *net)udp_sk (sk)->encap_type =1 ;udp_sk (sk)->encap_rcv = vxlan_udp_encap_recv;udp_encap_enable (); +#if IS_ENABLED (CONFIG_IPV6 ) +udpv6_encap_enable (); +#endif for (h =0 ; h <VNI_HASH_SIZE ; ++h)INIT_HLIST_HEAD (&vn->vni_list[h]);
include/uapi/linux/if_link.h +2 -0
@@ -310,6 +310,8 @@ enum {IFLA_VXLAN_RSC ,IFLA_VXLAN_L2MISS ,IFLA_VXLAN_L3MISS , +IFLA_VXLAN_REMOTE6 , +IFLA_VXLAN_LOCAL6 , __IFLA_VXLAN_MAX };#define IFLA_VXLAN_MAX (__IFLA_VXLAN_MAX - 1)
Signature · 1 line
1.7.7.6
Cong Wang <[email protected]> wrote on 04/21/2013 10:23:11 AM:
vxlan_encap_bypass(skb, vxlan, vxlan);
Quoted 9 lines
drivers/net/vxlan.c +1
-3@@ -975,60 +1141,115 @@ static netdev_tx_t vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev, skb->encapsulation = 1; } -/* Need space for new headers (invalidates iph ptr) */ -if (skb_cow_head (skb,VXLAN_HEADROOM )) - goto drop; + ttl = vxlan->ttl;
drivers/net/vxlan.c static netdev_tx_t vxlan_xmit_one(struct · line 1154
Quoted 10 lines
drivers/net/vxlan.c +8
-1+ tos = vxlan->tos; +if (dst->sa.sa_family ==AF_INET ) { +/* Need space for new headers (invalidates iph ptr) */ +if (skb_cow_head (skb,VXLAN_HEADROOM )) + goto drop; - old_iph =ip_hdr (skb); + old_iph =ip_hdr (skb); +if (!ttl &&IN_MULTICAST (ntohl (dst->sin.sin_addr.s_addr))) + ttl =1 ;
Why not use vxlan_addr_multicast() and place above as common
code for both v4 and v6?
This has other deficiencies:
1) No L3 miss support for v6
2) No link-local handling for v6
3) No neighbor-discovery reduction (like arp_reduce()) for v6
4) code simplifications/cleanup as described by Bjorn Mork
But I think those can be done in future patches. It might
be appropriate to prohibit link-local addresses with EADDRNOTAVAIL until LL support is there, since I expect you'll get silent drops with the existing code and an LL addr destination.
So, I'll go ahead and say:
and do some follow-up patches for the above if nobody else does first.
+-DLS
PATCH v5 5/5 ipv6: Add generic UDP Tunnel segmentation
In reply to Cong WangFrom: Cong Wang <[email protected]>
Similar to commit 731362674580cb0c696cd1b1a03d8461a10cf90a (tunneling: Add generic Tunnel segmentation)
This patch adds generic tunneling offloading support for IPv6-UDP based tunnels.
This can be used by tunneling protocols like VXLAN.
--- net/ipv6/ip6_offload.c | 4 +- net/ipv6/udp_offload.c | 153 +++++++++++++++++++++++++++++++++--------------- 2 files changed, 108 insertions(+), 49 deletions(-)
net/ipv6/ip6_offload.c +3 -1
@@ -91,6 +91,7 @@ static struct sk_buff *ipv6_gso_segment(struct sk_buff *skb,unsigned int unfrag_ip6hlen;u8 *prevhdr;int offset =0 ; +bool tunnel; if (unlikely(skb_shinfo(skb)->gso_type & ~(SKB_GSO_UDP |
@@ -105,6 +106,7 @@ static struct sk_buff *ipv6_gso_segment(struct sk_buff *skb,if (unlikely (!pskb_may_pull (skb,sizeof (*ipv6h)))) goto out; + tunnel = skb->encapsulation; ipv6h =ipv6_hdr (skb);__skb_pull (skb,sizeof (*ipv6h)); segs =ERR_PTR (-EPROTONOSUPPORT );
@@ -125,7 +127,7 @@ static struct sk_buff *ipv6_gso_segment(struct sk_buff *skb, ipv6h =ipv6_hdr (skb); ipv6h->payload_len =htons (skb->len - skb->mac_len -sizeof (*ipv6h)); -if (proto ==IPPROTO_UDP ) { +if (!tunnel && proto ==IPPROTO_UDP ) { unfrag_ip6hlen =ip6_find_1stfragopt (skb, &prevhdr); fptr = (struct frag_hdr *)(skb_network_header (skb) + unfrag_ip6hlen);
net/ipv6/udp_offload.c +105 -48
174 lines not on this page · raw message
Signature · 1 line
1.7.7.6
Quoted 33 lines · On Sun, 21 Apr 2013 22:23:07 +0800 Cong Wang <[email protected]> wrote:
From: Cong Wang <[email protected]>
v5: make David happy on the names of the fields
fix my mistake during rebasing the patches
drop thescope_idpatch, because it is broken
exportin6addr_loopback
fix a udp checksum bug
rebased on the latest net-nextv4: rename ->sin to ->va_sin
rename ->sin6 to ->va_sin6
rename ->family to ->va_sa
support ll addr
fix more ugly #ifdef
rebased on the latest net-nextv3: fix many coding style issues
fix some ugly #ifdef
renamevxlan_iptovxlan_addr
rename ->proto to ->family
rename ->ip4/->ip6 to ->sin/->sin6v2: fix some compile error when !CONFIG_IPV6
improve some code based on Stephen's comments
use sockaddr suggested by DavidCong Wang (5):
vxlan: defer vxlan init as late as possible
ipv6: exportipv6_sock_mc_joinandipv6_sock_mc_drop
ipv6: exportin6addr_loopbackto modules
vxlan: add ipv6 support
ipv6: Add generic UDP Tunnel segmentation
Quoted 9 lines
drivers/net/vxlan.c | 625 ++++++++++++++++++++++++++++++++--------- include/uapi/linux/if_link.h | 2 + net/ipv6/addrconf.c | 9 - net/ipv6/addrconf_core.c | 9 + net/ipv6/ip6_offload.c | 4 +- net/ipv6/mcast.c | 2 + net/ipv6/udp_offload.c | 153 +++++++---- 7 files changed, 609 insertions(+), 195 deletions(-)
Looks good.
David Miller
maintainerThis is broken. Every time I see someone export new things from IPV6 and then try to use those symbols in some other unrelated module, it is a huge red flag.
You can't call into IPV6 protected symbols unless VXLAN and IPV6 are configured identically.
So with your changes, with VXLAN=y and IPV6=m, you'll get link errors. I could see this just by looking at your patch, I didn't have to even try to build it.
Please do not fix this by adding Kconfig dependencies, you have to find another way. In bonding and bridging, we've made it such that you can configure them in any combination whatsoever with ipv6 and everything works properly. Most of them time this can be accomplished by moving things into the explicit "obj-y" objects in net/ipv6/Makefile
If you are adding stateful dependencies upon ipv6 (you want to inspect the ipv6 routes or something like that), I'm sorry but I really don't want any hard dependies on ipv6's internal state, we can't export that properly.
Quoted 10 lines · David Miller
This is broken. Every time I see someone export new things from IPV6 and then try to use those symbols in some other unrelated module, it is a huge red flag.
You can't call into IPV6 protected symbols unless VXLAN and IPV6 are configured identically.
So with your changes, with VXLAN=y and IPV6=m, you'll get link errors. I could see this just by looking at your patch, I didn't have to even try to build it.
Yeah, the IPv6 multicast API's we export indeed have such problem.
Quoted 7 lines
Please do not fix this by adding
Kconfigdependencies, you have to find another way. In bonding and bridging, we've made it such that you can configure them in any combination whatsoever with ipv6 and everything works properly. Most of them time this can be accomplished by moving things into the explicit "obj-y" objects innet/ipv6/Makefile
One quick solution is just linking mcast.o statically, because it is not easy to separate core functions from mcast.c like addrconf_core.c.
Quoted 5 lines
If you are adding stateful dependencies upon ipv6 (you want to inspect the ipv6 routes or something like that), I'm sorry but I really don't want any hard dependies on ipv6's internal state, we can't export that properly.
I don't think we need that.
Thanks!
Quoted 13 lines · Cong Wang
Quoted 7 lines · David Miller
Please do not fix this by adding
Kconfigdependencies, you have to find another way. In bonding and bridging, we've made it such that you can configure them in any combination whatsoever with ipv6 and everything works properly. Most of them time this can be accomplished by moving things into the explicit "obj-y" objects innet/ipv6/MakefileOne quick solution is just linking mcast.o statically, because it is not easy to separate core functions from
mcast.clikeaddrconf_core.c.Quoted 1 line
Quoted 7 lines
Quoted 4 lines
If you are adding stateful dependencies upon ipv6 (you want to inspect the ipv6 routes or something like that), I'm sorry but I really don't want any hard dependies on ipv6's internal state, we can't export that properly.
I don't think we need that.
After several tries, I think it is not easy to do at all, it relies on some icmp functions at least, which are still compiled as a module. So, I can't think out any easier solution than simply adding Kconfig dependency.
Or maybe I should raise the question again: should we forbid compiling IPv6 as a module from now on? At least some popular distributions already use CONFIG_IPV6=y. The only IPv6 things we really need to compile as a module is probably just procfs/sysfs stuffs.
Thanks.
Quoted 4 lines · On Tue, 23 Apr 2013 14:27:35 +0800 Cong Wang <[email protected]> wrote:
After several tries, I think it is not easy to do at all, it relies on some icmp functions at least, which are still compiled as a module. So, I can't think out any easier solution than simply adding
Kconfigdependency.
Why is IPv6 support depending on ICMP functions? Everything doesn't have to be hyper optimized, at least not initially.
Quoted 10 lines · Stephen Hemminger
Quoted 4 lines · On Tue, 23 Apr 2013 14:27:35 +0800 Cong Wang <[email protected]> wrote:
After several tries, I think it is not easy to do at all, it relies on some icmp functions at least, which are still compiled as a module. So, I can't think out any easier solution than simply adding
Kconfigdependency.Why is IPv6 support depending on ICMP functions? Everything doesn't have to be hyper optimized, at least not initially.
VXLAN relies on IPv6 multicast which is on top of ICMPv6. At very least igmp6_send() calls icmpv6_flow_init() and icmp6_dst_alloc() in the code.
Bridge multicast snooping only parses some multicast packets, which seems much easier than maintaining multicast group ownership.
Yeah, I definitely agree on that we can fix this later, if David has no objections.
Thanks.
David Miller
maintainerFrom: Cong Wang <[email protected]>
Date: Tue, 23 Apr 2013 14:27:35 +0800
Quoted 5 lines
Or maybe I should raise the question again: should we forbid compiling IPv6 as a module from now on? At least some popular distributions already use
CONFIG_IPV6=y. The only IPv6 things we really need to compile as a module is probably just procfs/sysfs stuffs.
We're not removing IPV6 modularity.
Please surprise me and fix this properly.
Thanks.