Merge branch 'ip6_tunnel-add-MPLS-support'

Vadim Fedorenko says:

====================
ip6_tunnel: add MPLS support

The support for MPLS-in-IPv4 was added earlier. This patchset adds
support for MPLS-in-IPv6.

Changes in v2:
- Eliminate ifdefs IS_ENABLE(CONFIG_MPLS)
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

3 files changed, 228 insertions, 109 deletions

diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c
index 4703b09808d0..821d96c720b9 100644
--- a/net/ipv6/ip6_tunnel.c
+++ b/net/ipv6/ip6_tunnel.c
@@ -89,6 +89,11 @@ struct ip6_tnl_net {
 	struct ip6_tnl __rcu *collect_md_tun;
 };
 
+static inline int ip6_tnl_mpls_supported(void)
+{
+	return IS_ENABLED(CONFIG_MPLS);
+}
+
 static struct net_device_stats *ip6_get_stats(struct net_device *dev)
 {
 	struct pcpu_sw_netstats tmp, sum = { 0 };
@@ -718,6 +723,20 @@ ip6ip6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
 	return 0;
 }
 
+static int
+mplsip6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
+	    u8 type, u8 code, int offset, __be32 info)
+{
+	__u32 rel_info = ntohl(info);
+	int err, rel_msg = 0;
+	u8 rel_type = type;
+	u8 rel_code = code;
+
+	err = ip6_tnl_err(skb, IPPROTO_MPLS, opt, &rel_type, &rel_code,
+			  &rel_msg, &rel_info, offset);
+	return err;
+}
+
 static int ip4ip6_dscp_ecn_decapsulate(const struct ip6_tnl *t,
 				       const struct ipv6hdr *ipv6h,
 				       struct sk_buff *skb)
@@ -740,6 +759,14 @@ static int ip6ip6_dscp_ecn_decapsulate(const struct ip6_tnl *t,
 	return IP6_ECN_decapsulate(ipv6h, skb);
 }
 
+static inline int mplsip6_dscp_ecn_decapsulate(const struct ip6_tnl *t,
+					       const struct ipv6hdr *ipv6h,
+					       struct sk_buff *skb)
+{
+	/* ECN is not supported in AF_MPLS */
+	return 0;
+}
+
 __u32 ip6_tnl_get_cap(struct ip6_tnl *t,
 			     const struct in6_addr *laddr,
 			     const struct in6_addr *raddr)
@@ -901,6 +928,11 @@ static const struct tnl_ptk_info tpi_v4 = {
 	.proto = htons(ETH_P_IP),
 };
 
+static const struct tnl_ptk_info tpi_mpls = {
+	/* no tunnel info required for mplsip6. */
+	.proto = htons(ETH_P_MPLS_UC),
+};
+
 static int ipxip6_rcv(struct sk_buff *skb, u8 ipproto,
 		      const struct tnl_ptk_info *tpi,
 		      int (*dscp_ecn_decapsulate)(const struct ip6_tnl *t,
@@ -958,6 +990,12 @@ static int ip6ip6_rcv(struct sk_buff *skb)
 			  ip6ip6_dscp_ecn_decapsulate);
 }
 
+static int mplsip6_rcv(struct sk_buff *skb)
+{
+	return ipxip6_rcv(skb, IPPROTO_MPLS, &tpi_mpls,
+			  mplsip6_dscp_ecn_decapsulate);
+}
+
 struct ipv6_tel_txoption {
 	struct ipv6_txoptions ops;
 	__u8 dst_opt[8];
@@ -1232,6 +1270,8 @@ route_lookup:
 		ipv6_push_frag_opts(skb, &opt.ops, &proto);
 	}
 
+	skb_set_inner_ipproto(skb, proto);
+
 	skb_push(skb, sizeof(struct ipv6hdr));
 	skb_reset_network_header(skb);
 	ipv6h = ipv6_hdr(skb);
@@ -1253,22 +1293,22 @@ tx_err_dst_release:
 EXPORT_SYMBOL(ip6_tnl_xmit);
 
 static inline int
-ip4ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
+ipxip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev,
+		u8 protocol)
 {
 	struct ip6_tnl *t = netdev_priv(dev);
+	struct ipv6hdr *ipv6h;
 	const struct iphdr  *iph;
 	int encap_limit = -1;
+	__u16 offset;
 	struct flowi6 fl6;
-	__u8 dsfield;
+	__u8 dsfield, orig_dsfield;
 	__u32 mtu;
 	u8 tproto;
 	int err;
 
-	iph = ip_hdr(skb);
-	memset(&(IPCB(skb)->opt), 0, sizeof(IPCB(skb)->opt));
-
 	tproto = READ_ONCE(t->parms.proto);
-	if (tproto != IPPROTO_IPIP && tproto != 0)
+	if (tproto != protocol && tproto != 0)
 		return -1;
 
 	if (t->parms.collect_md) {
@@ -1281,129 +1321,100 @@ ip4ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
 			return -1;
 		key = &tun_info->key;
 		memset(&fl6, 0, sizeof(fl6));
-		fl6.flowi6_proto = IPPROTO_IPIP;
+		fl6.flowi6_proto = protocol;
 		fl6.saddr = key->u.ipv6.src;
 		fl6.daddr = key->u.ipv6.dst;
 		fl6.flowlabel = key->label;
 		dsfield =  key->tos;
+		switch (protocol) {
+		case IPPROTO_IPIP:
+			iph = ip_hdr(skb);
+			orig_dsfield = ipv4_get_dsfield(iph);
+			break;
+		case IPPROTO_IPV6:
+			ipv6h = ipv6_hdr(skb);
+			orig_dsfield = ipv6_get_dsfield(ipv6h);
+			break;
+		default:
+			orig_dsfield = dsfield;
+			break;
+		}
 	} else {
 		if (!(t->parms.flags & IP6_TNL_F_IGN_ENCAP_LIMIT))
 			encap_limit = t->parms.encap_limit;
+		if (protocol == IPPROTO_IPV6) {
+			offset = ip6_tnl_parse_tlv_enc_lim(skb,
+						skb_network_header(skb));
+			/* ip6_tnl_parse_tlv_enc_lim() might have
+			 * reallocated skb->head
+			 */
+			if (offset > 0) {
+				struct ipv6_tlv_tnl_enc_lim *tel;
 
-		memcpy(&fl6, &t->fl.u.ip6, sizeof(fl6));
-		fl6.flowi6_proto = IPPROTO_IPIP;
-
-		if (t->parms.flags & IP6_TNL_F_USE_ORIG_TCLASS)
-			dsfield = ipv4_get_dsfield(iph);
-		else
-			dsfield = ip6_tclass(t->parms.flowinfo);
-		if (t->parms.flags & IP6_TNL_F_USE_ORIG_FWMARK)
-			fl6.flowi6_mark = skb->mark;
-		else
-			fl6.flowi6_mark = t->parms.fwmark;
-	}
-
-	fl6.flowi6_uid = sock_net_uid(dev_net(dev), NULL);
-	dsfield = INET_ECN_encapsulate(dsfield, ipv4_get_dsfield(iph));
-
-	if (iptunnel_handle_offloads(skb, SKB_GSO_IPXIP6))
-		return -1;
-
-	skb_set_inner_ipproto(skb, IPPROTO_IPIP);
-
-	err = ip6_tnl_xmit(skb, dev, dsfield, &fl6, encap_limit, &mtu,
-			   IPPROTO_IPIP);
-	if (err != 0) {
-		/* XXX: send ICMP error even if DF is not set. */
-		if (err == -EMSGSIZE)
-			icmp_send(skb, ICMP_DEST_UNREACH, ICMP_FRAG_NEEDED,
-				  htonl(mtu));
-		return -1;
-	}
-
-	return 0;
-}
-
-static inline int
-ip6ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
-{
-	struct ip6_tnl *t = netdev_priv(dev);
-	struct ipv6hdr *ipv6h;
-	int encap_limit = -1;
-	__u16 offset;
-	struct flowi6 fl6;
-	__u8 dsfield;
-	__u32 mtu;
-	u8 tproto;
-	int err;
-
-	ipv6h = ipv6_hdr(skb);
-	tproto = READ_ONCE(t->parms.proto);
-	if ((tproto != IPPROTO_IPV6 && tproto != 0) ||
-	    ip6_tnl_addr_conflict(t, ipv6h))
-		return -1;
-
-	if (t->parms.collect_md) {
-		struct ip_tunnel_info *tun_info;
-		const struct ip_tunnel_key *key;
-
-		tun_info = skb_tunnel_info(skb);
-		if (unlikely(!tun_info || !(tun_info->mode & IP_TUNNEL_INFO_TX) ||
-			     ip_tunnel_info_af(tun_info) != AF_INET6))
-			return -1;
-		key = &tun_info->key;
-		memset(&fl6, 0, sizeof(fl6));
-		fl6.flowi6_proto = IPPROTO_IPV6;
-		fl6.saddr = key->u.ipv6.src;
-		fl6.daddr = key->u.ipv6.dst;
-		fl6.flowlabel = key->label;
-		dsfield = key->tos;
-	} else {
-		offset = ip6_tnl_parse_tlv_enc_lim(skb, skb_network_header(skb));
-		/* ip6_tnl_parse_tlv_enc_lim() might have reallocated skb->head */
-		ipv6h = ipv6_hdr(skb);
-		if (offset > 0) {
-			struct ipv6_tlv_tnl_enc_lim *tel;
-
-			tel = (void *)&skb_network_header(skb)[offset];
-			if (tel->encap_limit == 0) {
-				icmpv6_send(skb, ICMPV6_PARAMPROB,
-					    ICMPV6_HDR_FIELD, offset + 2);
-				return -1;
+				tel = (void *)&skb_network_header(skb)[offset];
+				if (tel->encap_limit == 0) {
+					icmpv6_send(skb, ICMPV6_PARAMPROB,
+						ICMPV6_HDR_FIELD, offset + 2);
+					return -1;
+				}
+				encap_limit = tel->encap_limit - 1;
 			}
-			encap_limit = tel->encap_limit - 1;
-		} else if (!(t->parms.flags & IP6_TNL_F_IGN_ENCAP_LIMIT)) {
-			encap_limit = t->parms.encap_limit;
 		}
 
 		memcpy(&fl6, &t->fl.u.ip6, sizeof(fl6));
-		fl6.flowi6_proto = IPPROTO_IPV6;
+		fl6.flowi6_proto = protocol;
 
-		if (t->parms.flags & IP6_TNL_F_USE_ORIG_TCLASS)
-			dsfield = ipv6_get_dsfield(ipv6h);
-		else
-			dsfield = ip6_tclass(t->parms.flowinfo);
-		if (t->parms.flags & IP6_TNL_F_USE_ORIG_FLOWLABEL)
-			fl6.flowlabel |= ip6_flowlabel(ipv6h);
 		if (t->parms.flags & IP6_TNL_F_USE_ORIG_FWMARK)
 			fl6.flowi6_mark = skb->mark;
 		else
 			fl6.flowi6_mark = t->parms.fwmark;
+		switch (protocol) {
+		case IPPROTO_IPIP:
+			iph = ip_hdr(skb);
+			orig_dsfield = ipv4_get_dsfield(iph);
+			if (t->parms.flags & IP6_TNL_F_USE_ORIG_TCLASS)
+				dsfield = orig_dsfield;
+			else
+				dsfield = ip6_tclass(t->parms.flowinfo);
+			break;
+		case IPPROTO_IPV6:
+			ipv6h = ipv6_hdr(skb);
+			orig_dsfield = ipv6_get_dsfield(ipv6h);
+			if (t->parms.flags & IP6_TNL_F_USE_ORIG_TCLASS)
+				dsfield = orig_dsfield;
+			else
+				dsfield = ip6_tclass(t->parms.flowinfo);
+			if (t->parms.flags & IP6_TNL_F_USE_ORIG_FLOWLABEL)
+				fl6.flowlabel |= ip6_flowlabel(ipv6h);
+			break;
+		default:
+			orig_dsfield = dsfield = ip6_tclass(t->parms.flowinfo);
+			break;
+		}
 	}
 
 	fl6.flowi6_uid = sock_net_uid(dev_net(dev), NULL);
-	dsfield = INET_ECN_encapsulate(dsfield, ipv6_get_dsfield(ipv6h));
+	dsfield = INET_ECN_encapsulate(dsfield, orig_dsfield);
 
 	if (iptunnel_handle_offloads(skb, SKB_GSO_IPXIP6))
 		return -1;
 
-	skb_set_inner_ipproto(skb, IPPROTO_IPV6);
-
 	err = ip6_tnl_xmit(skb, dev, dsfield, &fl6, encap_limit, &mtu,
-			   IPPROTO_IPV6);
+			   protocol);
 	if (err != 0) {
+		/* XXX: send ICMP error even if DF is not set. */
 		if (err == -EMSGSIZE)
-			icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu);
+			switch (protocol) {
+			case IPPROTO_IPIP:
+				icmp_send(skb, ICMP_DEST_UNREACH,
+					  ICMP_FRAG_NEEDED, htonl(mtu));
+				break;
+			case IPPROTO_IPV6:
+				icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu);
+				break;
+			default:
+				break;
+			}
 		return -1;
 	}
 
@@ -1415,6 +1426,7 @@ ip6_tnl_start_xmit(struct sk_buff *skb, struct net_device *dev)
 {
 	struct ip6_tnl *t = netdev_priv(dev);
 	struct net_device_stats *stats = &t->dev->stats;
+	u8 ipproto;
 	int ret;
 
 	if (!pskb_inet_may_pull(skb))
@@ -1422,15 +1434,21 @@ ip6_tnl_start_xmit(struct sk_buff *skb, struct net_device *dev)
 
 	switch (skb->protocol) {
 	case htons(ETH_P_IP):
-		ret = ip4ip6_tnl_xmit(skb, dev);
+		ipproto = IPPROTO_IPIP;
 		break;
 	case htons(ETH_P_IPV6):
-		ret = ip6ip6_tnl_xmit(skb, dev);
+		if (ip6_tnl_addr_conflict(t, ipv6_hdr(skb)))
+			goto tx_err;
+		ipproto = IPPROTO_IPV6;
+		break;
+	case htons(ETH_P_MPLS_UC):
+		ipproto = IPPROTO_MPLS;
 		break;
 	default:
 		goto tx_err;
 	}
 
+	ret = ipxip6_tnl_xmit(skb, dev, ipproto);
 	if (ret < 0)
 		goto tx_err;
 
@@ -2218,6 +2236,12 @@ static struct xfrm6_tunnel ip6ip6_handler __read_mostly = {
 	.priority	=	1,
 };
 
+static struct xfrm6_tunnel mplsip6_handler __read_mostly = {
+	.handler	= mplsip6_rcv,
+	.err_handler	= mplsip6_err,
+	.priority	=	1,
+};
+
 static void __net_exit ip6_tnl_destroy_tunnels(struct net *net, struct list_head *list)
 {
 	struct ip6_tnl_net *ip6n = net_generic(net, ip6_tnl_net_id);
@@ -2332,6 +2356,15 @@ static int __init ip6_tunnel_init(void)
 		pr_err("%s: can't register ip6ip6\n", __func__);
 		goto out_ip6ip6;
 	}
+
+	if (ip6_tnl_mpls_supported()) {
+		err = xfrm6_tunnel_register(&mplsip6_handler, AF_MPLS);
+		if (err < 0) {
+			pr_err("%s: can't register mplsip6\n", __func__);
+			goto out_mplsip6;
+		}
+	}
+
 	err = rtnl_link_register(&ip6_link_ops);
 	if (err < 0)
 		goto rtnl_link_failed;
@@ -2339,6 +2372,9 @@ static int __init ip6_tunnel_init(void)
 	return 0;
 
 rtnl_link_failed:
+	if (ip6_tnl_mpls_supported())
+		xfrm6_tunnel_deregister(&mplsip6_handler, AF_MPLS);
+out_mplsip6:
 	xfrm6_tunnel_deregister(&ip6ip6_handler, AF_INET6);
 out_ip6ip6:
 	xfrm6_tunnel_deregister(&ip4ip6_handler, AF_INET);
@@ -2361,6 +2397,9 @@ static void __exit ip6_tunnel_cleanup(void)
 	if (xfrm6_tunnel_deregister(&ip6ip6_handler, AF_INET6))
 		pr_info("%s: can't deregister ip6ip6\n", __func__);
 
+	if (ip6_tnl_mpls_supported() &&
+	    xfrm6_tunnel_deregister(&mplsip6_handler, AF_MPLS))
+		pr_info("%s: can't deregister mplsip6\n", __func__);
 	unregister_pernet_device(&ip6_tnl_net_ops);
 }
 
diff --git a/net/ipv6/tunnel6.c b/net/ipv6/tunnel6.c
index 21e7b95ddbfa..06c02ebe6b9b 100644
--- a/net/ipv6/tunnel6.c
+++ b/net/ipv6/tunnel6.c
@@ -21,8 +21,14 @@
 
 static struct xfrm6_tunnel __rcu *tunnel6_handlers __read_mostly;
 static struct xfrm6_tunnel __rcu *tunnel46_handlers __read_mostly;
+static struct xfrm6_tunnel __rcu *tunnelmpls6_handlers __read_mostly;
 static DEFINE_MUTEX(tunnel6_mutex);
 
+static inline int xfrm6_tunnel_mpls_supported(void)
+{
+	return IS_ENABLED(CONFIG_MPLS);
+}
+
 int xfrm6_tunnel_register(struct xfrm6_tunnel *handler, unsigned short family)
 {
 	struct xfrm6_tunnel __rcu **pprev;
@@ -32,8 +38,21 @@ int xfrm6_tunnel_register(struct xfrm6_tunnel *handler, unsigned short family)
 
 	mutex_lock(&tunnel6_mutex);
 
-	for (pprev = (family == AF_INET6) ? &tunnel6_handlers : &tunnel46_handlers;
-	     (t = rcu_dereference_protected(*pprev,
+	switch (family) {
+	case AF_INET6:
+		pprev = &tunnel6_handlers;
+		break;
+	case AF_INET:
+		pprev = &tunnel46_handlers;
+		break;
+	case AF_MPLS:
+		pprev = &tunnelmpls6_handlers;
+		break;
+	default:
+		goto err;
+	}
+
+	for (; (t = rcu_dereference_protected(*pprev,
 			lockdep_is_held(&tunnel6_mutex))) != NULL;
 	     pprev = &t->next) {
 		if (t->priority > priority)
@@ -62,8 +81,21 @@ int xfrm6_tunnel_deregister(struct xfrm6_tunnel *handler, unsigned short family)
 
 	mutex_lock(&tunnel6_mutex);
 
-	for (pprev = (family == AF_INET6) ? &tunnel6_handlers : &tunnel46_handlers;
-	     (t = rcu_dereference_protected(*pprev,
+	switch (family) {
+	case AF_INET6:
+		pprev = &tunnel6_handlers;
+		break;
+	case AF_INET:
+		pprev = &tunnel46_handlers;
+		break;
+	case AF_MPLS:
+		pprev = &tunnelmpls6_handlers;
+		break;
+	default:
+		goto err;
+	}
+
+	for (; (t = rcu_dereference_protected(*pprev,
 			lockdep_is_held(&tunnel6_mutex))) != NULL;
 	     pprev = &t->next) {
 		if (t == handler) {
@@ -73,6 +105,7 @@ int xfrm6_tunnel_deregister(struct xfrm6_tunnel *handler, unsigned short family)
 		}
 	}
 
+err:
 	mutex_unlock(&tunnel6_mutex);
 
 	synchronize_net();
@@ -86,6 +119,24 @@ EXPORT_SYMBOL(xfrm6_tunnel_deregister);
 	     handler != NULL;				\
 	     handler = rcu_dereference(handler->next))	\
 
+static int tunnelmpls6_rcv(struct sk_buff *skb)
+{
+	struct xfrm6_tunnel *handler;
+
+	if (!pskb_may_pull(skb, sizeof(struct ipv6hdr)))
+		goto drop;
+
+	for_each_tunnel_rcu(tunnelmpls6_handlers, handler)
+		if (!handler->handler(skb))
+			return 0;
+
+	icmpv6_send(skb, ICMPV6_DEST_UNREACH, ICMPV6_PORT_UNREACH, 0);
+
+drop:
+	kfree_skb(skb);
+	return 0;
+}
+
 static int tunnel6_rcv(struct sk_buff *skb)
 {
 	struct xfrm6_tunnel *handler;
@@ -146,6 +197,18 @@ static int tunnel46_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
 	return -ENOENT;
 }
 
+static int tunnelmpls6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
+			   u8 type, u8 code, int offset, __be32 info)
+{
+	struct xfrm6_tunnel *handler;
+
+	for_each_tunnel_rcu(tunnelmpls6_handlers, handler)
+		if (!handler->err_handler(skb, opt, type, code, offset, info))
+			return 0;
+
+	return -ENOENT;
+}
+
 static const struct inet6_protocol tunnel6_protocol = {
 	.handler	= tunnel6_rcv,
 	.err_handler	= tunnel6_err,
@@ -158,6 +221,12 @@ static const struct inet6_protocol tunnel46_protocol = {
 	.flags          = INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
 };
 
+static const struct inet6_protocol tunnelmpls6_protocol = {
+	.handler	= tunnelmpls6_rcv,
+	.err_handler	= tunnelmpls6_err,
+	.flags          = INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
+};
+
 static int __init tunnel6_init(void)
 {
 	if (inet6_add_protocol(&tunnel6_protocol, IPPROTO_IPV6)) {
@@ -169,6 +238,13 @@ static int __init tunnel6_init(void)
 		inet6_del_protocol(&tunnel6_protocol, IPPROTO_IPV6);
 		return -EAGAIN;
 	}
+	if (xfrm6_tunnel_mpls_supported() &&
+	    inet6_add_protocol(&tunnelmpls6_protocol, IPPROTO_MPLS)) {
+		pr_err("%s: can't add protocol\n", __func__);
+		inet6_del_protocol(&tunnel6_protocol, IPPROTO_IPV6);
+		inet6_del_protocol(&tunnel46_protocol, IPPROTO_IPIP);
+		return -EAGAIN;
+	}
 	return 0;
 }
 
@@ -178,6 +254,9 @@ static void __exit tunnel6_fini(void)
 		pr_err("%s: can't remove protocol\n", __func__);
 	if (inet6_del_protocol(&tunnel6_protocol, IPPROTO_IPV6))
 		pr_err("%s: can't remove protocol\n", __func__);
+	if (xfrm6_tunnel_mpls_supported() &&
+	    inet6_del_protocol(&tunnelmpls6_protocol, IPPROTO_MPLS))
+		pr_err("%s: can't remove protocol\n", __func__);
 }
 
 module_init(tunnel6_init);
diff --git a/net/mpls/af_mpls.c b/net/mpls/af_mpls.c
index a42e4ed5ab0e..fd30ea61336e 100644
--- a/net/mpls/af_mpls.c
+++ b/net/mpls/af_mpls.c
@@ -1593,7 +1593,8 @@ static int mpls_dev_notify(struct notifier_block *this, unsigned long event,
 		    dev->type == ARPHRD_IPGRE ||
 		    dev->type == ARPHRD_IP6GRE ||
 		    dev->type == ARPHRD_SIT ||
-		    dev->type == ARPHRD_TUNNEL) {
+		    dev->type == ARPHRD_TUNNEL ||
+		    dev->type == ARPHRD_TUNNEL6) {
 			mdev = mpls_add_dev(dev);
 			if (IS_ERR(mdev))
 				return notifier_from_errno(PTR_ERR(mdev));