ip4ip6: Support for GSO/GRO
[cascardo/linux.git] / net / ipv6 / ip6_tunnel.c
index 1f20345..823dad1 100644 (file)
@@ -238,6 +238,7 @@ static void ip6_dev_free(struct net_device *dev)
 {
        struct ip6_tnl *t = netdev_priv(dev);
 
+       gro_cells_destroy(&t->gro_cells);
        dst_cache_destroy(&t->dst_cache);
        free_percpu(dev->tstats);
        free_netdev(dev);
@@ -753,97 +754,157 @@ int ip6_tnl_rcv_ctl(struct ip6_tnl *t,
 }
 EXPORT_SYMBOL_GPL(ip6_tnl_rcv_ctl);
 
-/**
- * ip6_tnl_rcv - decapsulate IPv6 packet and retransmit it locally
- *   @skb: received socket buffer
- *   @protocol: ethernet protocol ID
- *   @dscp_ecn_decapsulate: the function to decapsulate DSCP code and ECN
- *
- * Return: 0
- **/
-
-static int ip6_tnl_rcv(struct sk_buff *skb, __u16 protocol,
-                      __u8 ipproto,
-                      int (*dscp_ecn_decapsulate)(const struct ip6_tnl *t,
-                                                  const struct ipv6hdr *ipv6h,
-                                                  struct sk_buff *skb))
+static int __ip6_tnl_rcv(struct ip6_tnl *tunnel, struct sk_buff *skb,
+                        const struct tnl_ptk_info *tpi,
+                        struct metadata_dst *tun_dst,
+                        int (*dscp_ecn_decapsulate)(const struct ip6_tnl *t,
+                                               const struct ipv6hdr *ipv6h,
+                                               struct sk_buff *skb),
+                        bool log_ecn_err)
 {
-       struct ip6_tnl *t;
+       struct pcpu_sw_netstats *tstats;
        const struct ipv6hdr *ipv6h = ipv6_hdr(skb);
-       u8 tproto;
        int err;
 
-       rcu_read_lock();
-       t = ip6_tnl_lookup(dev_net(skb->dev), &ipv6h->saddr, &ipv6h->daddr);
-       if (t) {
-               struct pcpu_sw_netstats *tstats;
+       if ((!(tpi->flags & TUNNEL_CSUM) &&
+            (tunnel->parms.i_flags & TUNNEL_CSUM)) ||
+           ((tpi->flags & TUNNEL_CSUM) &&
+            !(tunnel->parms.i_flags & TUNNEL_CSUM))) {
+               tunnel->dev->stats.rx_crc_errors++;
+               tunnel->dev->stats.rx_errors++;
+               goto drop;
+       }
 
-               tproto = ACCESS_ONCE(t->parms.proto);
-               if (tproto != ipproto && tproto != 0) {
-                       rcu_read_unlock();
-                       goto discard;
+       if (tunnel->parms.i_flags & TUNNEL_SEQ) {
+               if (!(tpi->flags & TUNNEL_SEQ) ||
+                   (tunnel->i_seqno &&
+                    (s32)(ntohl(tpi->seq) - tunnel->i_seqno) < 0)) {
+                       tunnel->dev->stats.rx_fifo_errors++;
+                       tunnel->dev->stats.rx_errors++;
+                       goto drop;
                }
+               tunnel->i_seqno = ntohl(tpi->seq) + 1;
+       }
 
-               if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb)) {
-                       rcu_read_unlock();
-                       goto discard;
-               }
+       skb->protocol = tpi->proto;
 
-               if (!ip6_tnl_rcv_ctl(t, &ipv6h->daddr, &ipv6h->saddr)) {
-                       t->dev->stats.rx_dropped++;
-                       rcu_read_unlock();
-                       goto discard;
+       /* Warning: All skb pointers will be invalidated! */
+       if (tunnel->dev->type == ARPHRD_ETHER) {
+               if (!pskb_may_pull(skb, ETH_HLEN)) {
+                       tunnel->dev->stats.rx_length_errors++;
+                       tunnel->dev->stats.rx_errors++;
+                       goto drop;
                }
-               skb->mac_header = skb->network_header;
-               skb_reset_network_header(skb);
-               skb->protocol = htons(protocol);
-               memset(skb->cb, 0, sizeof(struct inet6_skb_parm));
-
-               __skb_tunnel_rx(skb, t->dev, t->net);
-
-               err = dscp_ecn_decapsulate(t, ipv6h, skb);
-               if (unlikely(err)) {
-                       if (log_ecn_error)
-                               net_info_ratelimited("non-ECT from %pI6 with dsfield=%#x\n",
-                                                    &ipv6h->saddr,
-                                                    ipv6_get_dsfield(ipv6h));
-                       if (err > 1) {
-                               ++t->dev->stats.rx_frame_errors;
-                               ++t->dev->stats.rx_errors;
-                               rcu_read_unlock();
-                               goto discard;
-                       }
+
+               ipv6h = ipv6_hdr(skb);
+               skb->protocol = eth_type_trans(skb, tunnel->dev);
+               skb_postpull_rcsum(skb, eth_hdr(skb), ETH_HLEN);
+       } else {
+               skb->dev = tunnel->dev;
+       }
+
+       skb_reset_network_header(skb);
+       memset(skb->cb, 0, sizeof(struct inet6_skb_parm));
+
+       __skb_tunnel_rx(skb, tunnel->dev, tunnel->net);
+
+       err = dscp_ecn_decapsulate(tunnel, ipv6h, skb);
+       if (unlikely(err)) {
+               if (log_ecn_err)
+                       net_info_ratelimited("non-ECT from %pI6 with DS=%#x\n",
+                                            &ipv6h->saddr,
+                                            ipv6_get_dsfield(ipv6h));
+               if (err > 1) {
+                       ++tunnel->dev->stats.rx_frame_errors;
+                       ++tunnel->dev->stats.rx_errors;
+                       goto drop;
                }
+       }
 
-               tstats = this_cpu_ptr(t->dev->tstats);
-               u64_stats_update_begin(&tstats->syncp);
-               tstats->rx_packets++;
-               tstats->rx_bytes += skb->len;
-               u64_stats_update_end(&tstats->syncp);
+       tstats = this_cpu_ptr(tunnel->dev->tstats);
+       u64_stats_update_begin(&tstats->syncp);
+       tstats->rx_packets++;
+       tstats->rx_bytes += skb->len;
+       u64_stats_update_end(&tstats->syncp);
 
-               netif_rx(skb);
+       skb_scrub_packet(skb, !net_eq(tunnel->net, dev_net(tunnel->dev)));
 
-               rcu_read_unlock();
-               return 0;
+       gro_cells_receive(&tunnel->gro_cells, skb);
+       return 0;
+
+drop:
+       kfree_skb(skb);
+       return 0;
+}
+
+int ip6_tnl_rcv(struct ip6_tnl *t, struct sk_buff *skb,
+               const struct tnl_ptk_info *tpi,
+               struct metadata_dst *tun_dst,
+               bool log_ecn_err)
+{
+       return __ip6_tnl_rcv(t, skb, tpi, NULL, ip6ip6_dscp_ecn_decapsulate,
+                            log_ecn_err);
+}
+EXPORT_SYMBOL(ip6_tnl_rcv);
+
+static const struct tnl_ptk_info tpi_v6 = {
+       /* no tunnel info required for ipxip6. */
+       .proto = htons(ETH_P_IPV6),
+};
+
+static const struct tnl_ptk_info tpi_v4 = {
+       /* no tunnel info required for ipxip6. */
+       .proto = htons(ETH_P_IP),
+};
+
+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,
+                                                 const struct ipv6hdr *ipv6h,
+                                                 struct sk_buff *skb))
+{
+       struct ip6_tnl *t;
+       const struct ipv6hdr *ipv6h = ipv6_hdr(skb);
+       int ret = -1;
+
+       rcu_read_lock();
+       t = ip6_tnl_lookup(dev_net(skb->dev), &ipv6h->saddr, &ipv6h->daddr);
+
+       if (t) {
+               u8 tproto = ACCESS_ONCE(t->parms.proto);
+
+               if (tproto != ipproto && tproto != 0)
+                       goto drop;
+               if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb))
+                       goto drop;
+               if (!ip6_tnl_rcv_ctl(t, &ipv6h->daddr, &ipv6h->saddr))
+                       goto drop;
+               if (iptunnel_pull_header(skb, 0, tpi->proto, false))
+                       goto drop;
+               ret = __ip6_tnl_rcv(t, skb, tpi, NULL, dscp_ecn_decapsulate,
+                                   log_ecn_error);
        }
+
        rcu_read_unlock();
-       return 1;
 
-discard:
+       return ret;
+
+drop:
+       rcu_read_unlock();
        kfree_skb(skb);
        return 0;
 }
 
 static int ip4ip6_rcv(struct sk_buff *skb)
 {
-       return ip6_tnl_rcv(skb, ETH_P_IP, IPPROTO_IPIP,
-                          ip4ip6_dscp_ecn_decapsulate);
+       return ipxip6_rcv(skb, IPPROTO_IPIP, &tpi_v4,
+                         ip4ip6_dscp_ecn_decapsulate);
 }
 
 static int ip6ip6_rcv(struct sk_buff *skb)
 {
-       return ip6_tnl_rcv(skb, ETH_P_IPV6, IPPROTO_IPV6,
-                          ip6ip6_dscp_ecn_decapsulate);
+       return ipxip6_rcv(skb, IPPROTO_IPV6, &tpi_v6,
+                         ip6ip6_dscp_ecn_decapsulate);
 }
 
 struct ipv6_tel_txoption {
@@ -918,13 +979,14 @@ int ip6_tnl_xmit_ctl(struct ip6_tnl *t,
 EXPORT_SYMBOL_GPL(ip6_tnl_xmit_ctl);
 
 /**
- * ip6_tnl_xmit2 - encapsulate packet and send
+ * ip6_tnl_xmit - encapsulate packet and send
  *   @skb: the outgoing socket buffer
  *   @dev: the outgoing tunnel device
  *   @dsfield: dscp code for outer header
- *   @fl: flow of tunneled packet
+ *   @fl6: flow of tunneled packet
  *   @encap_limit: encapsulation limit
  *   @pmtu: Path MTU is stored if packet is too big
+ *   @proto: next header value
  *
  * Description:
  *   Build new header and do some sanity checks on the packet before sending
@@ -936,12 +998,9 @@ EXPORT_SYMBOL_GPL(ip6_tnl_xmit_ctl);
  *   %-EMSGSIZE message too big. return mtu in this case.
  **/
 
-static int ip6_tnl_xmit2(struct sk_buff *skb,
-                        struct net_device *dev,
-                        __u8 dsfield,
-                        struct flowi6 *fl6,
-                        int encap_limit,
-                        __u32 *pmtu)
+int ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev, __u8 dsfield,
+                struct flowi6 *fl6, int encap_limit, __u32 *pmtu,
+                __u8 proto)
 {
        struct ip6_tnl *t = netdev_priv(dev);
        struct net *net = t->net;
@@ -951,8 +1010,8 @@ static int ip6_tnl_xmit2(struct sk_buff *skb,
        struct dst_entry *dst = NULL, *ndst = NULL;
        struct net_device *tdev;
        int mtu;
-       unsigned int max_headroom = sizeof(struct ipv6hdr);
-       u8 proto;
+       unsigned int psh_hlen = sizeof(struct ipv6hdr) + t->encap_hlen;
+       unsigned int max_headroom = psh_hlen;
        int err = -1;
 
        /* NBMA tunnel */
@@ -1005,7 +1064,7 @@ static int ip6_tnl_xmit2(struct sk_buff *skb,
                                     t->parms.name);
                goto tx_err_dst_release;
        }
-       mtu = dst_mtu(dst) - sizeof(*ipv6h);
+       mtu = dst_mtu(dst) - psh_hlen;
        if (encap_limit >= 0) {
                max_headroom += 8;
                mtu -= 8;
@@ -1014,12 +1073,23 @@ static int ip6_tnl_xmit2(struct sk_buff *skb,
                mtu = IPV6_MIN_MTU;
        if (skb_dst(skb))
                skb_dst(skb)->ops->update_pmtu(skb_dst(skb), NULL, skb, mtu);
-       if (skb->len > mtu) {
+       if (skb->len > mtu && !skb_is_gso(skb)) {
                *pmtu = mtu;
                err = -EMSGSIZE;
                goto tx_err_dst_release;
        }
 
+       if (t->err_count > 0) {
+               if (time_before(jiffies,
+                               t->err_time + IP6TUNNEL_ERR_TIMEO)) {
+                       t->err_count--;
+
+                       dst_link_failure(skb);
+               } else {
+                       t->err_count = 0;
+               }
+       }
+
        skb_scrub_packet(skb, !net_eq(t->net, dev_net(dev)));
 
        /*
@@ -1045,9 +1115,6 @@ static int ip6_tnl_xmit2(struct sk_buff *skb,
                dst_cache_set_ip6(&t->dst_cache, ndst, &fl6->saddr);
        skb_dst_set(skb, dst);
 
-       skb->transport_header = skb->network_header;
-
-       proto = fl6->flowi6_proto;
        if (encap_limit >= 0) {
                init_tel_txopt(&opt, encap_limit);
                ipv6_push_nfrag_opts(skb, &opt.ops, &proto, NULL);
@@ -1058,6 +1125,18 @@ static int ip6_tnl_xmit2(struct sk_buff *skb,
                skb->encapsulation = 1;
        }
 
+       /* Calculate max headroom for all the headers and adjust
+        * needed_headroom if necessary.
+        */
+       max_headroom = LL_RESERVED_SPACE(dst->dev) + sizeof(struct ipv6hdr)
+                       + dst->header_len + t->hlen;
+       if (max_headroom > dev->needed_headroom)
+               dev->needed_headroom = max_headroom;
+
+       err = ip6_tnl_encap(skb, t, &proto, fl6);
+       if (err)
+               return err;
+
        skb_push(skb, sizeof(struct ipv6hdr));
        skb_reset_network_header(skb);
        ipv6h = ipv6_hdr(skb);
@@ -1076,6 +1155,7 @@ tx_err_dst_release:
        dst_release(dst);
        return err;
 }
+EXPORT_SYMBOL(ip6_tnl_xmit);
 
 static inline int
 ip4ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
@@ -1099,7 +1179,6 @@ ip4ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
                encap_limit = t->parms.encap_limit;
 
        memcpy(&fl6, &t->fl.u.ip6, sizeof(fl6));
-       fl6.flowi6_proto = IPPROTO_IPIP;
 
        dsfield = ipv4_get_dsfield(iph);
 
@@ -1109,7 +1188,13 @@ ip4ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
        if (t->parms.flags & IP6_TNL_F_USE_ORIG_FWMARK)
                fl6.flowi6_mark = skb->mark;
 
-       err = ip6_tnl_xmit2(skb, dev, dsfield, &fl6, encap_limit, &mtu);
+       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)
@@ -1153,7 +1238,6 @@ ip6ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
                encap_limit = t->parms.encap_limit;
 
        memcpy(&fl6, &t->fl.u.ip6, sizeof(fl6));
-       fl6.flowi6_proto = IPPROTO_IPV6;
 
        dsfield = ipv6_get_dsfield(ipv6h);
        if (t->parms.flags & IP6_TNL_F_USE_ORIG_TCLASS)
@@ -1163,7 +1247,13 @@ ip6ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
        if (t->parms.flags & IP6_TNL_F_USE_ORIG_FWMARK)
                fl6.flowi6_mark = skb->mark;
 
-       err = ip6_tnl_xmit2(skb, dev, dsfield, &fl6, encap_limit, &mtu);
+       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);
        if (err != 0) {
                if (err == -EMSGSIZE)
                        icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu);
@@ -1174,7 +1264,7 @@ ip6ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
 }
 
 static netdev_tx_t
-ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
+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;
@@ -1208,6 +1298,7 @@ static void ip6_tnl_link_config(struct ip6_tnl *t)
        struct net_device *dev = t->dev;
        struct __ip6_tnl_parm *p = &t->parms;
        struct flowi6 *fl6 = &t->fl.u.ip6;
+       int t_hlen;
 
        memcpy(dev->dev_addr, &p->laddr, sizeof(struct in6_addr));
        memcpy(dev->broadcast, &p->raddr, sizeof(struct in6_addr));
@@ -1231,6 +1322,10 @@ static void ip6_tnl_link_config(struct ip6_tnl *t)
        else
                dev->flags &= ~IFF_POINTOPOINT;
 
+       t->tun_hlen = 0;
+       t->hlen = t->encap_hlen + t->tun_hlen;
+       t_hlen = t->hlen + sizeof(struct ipv6hdr);
+
        if (p->flags & IP6_TNL_F_CAP_XMIT) {
                int strict = (ipv6_addr_type(&p->raddr) &
                              (IPV6_ADDR_MULTICAST|IPV6_ADDR_LINKLOCAL));
@@ -1244,9 +1339,9 @@ static void ip6_tnl_link_config(struct ip6_tnl *t)
 
                if (rt->dst.dev) {
                        dev->hard_header_len = rt->dst.dev->hard_header_len +
-                               sizeof(struct ipv6hdr);
+                               t_hlen;
 
-                       dev->mtu = rt->dst.dev->mtu - sizeof(struct ipv6hdr);
+                       dev->mtu = rt->dst.dev->mtu - t_hlen;
                        if (!(t->parms.flags & IP6_TNL_F_IGN_ENCAP_LIMIT))
                                dev->mtu -= 8;
 
@@ -1370,6 +1465,8 @@ ip6_tnl_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
        struct net *net = t->net;
        struct ip6_tnl_net *ip6n = net_generic(net, ip6_tnl_net_id);
 
+       memset(&p1, 0, sizeof(p1));
+
        switch (cmd) {
        case SIOCGETTUNNEL:
                if (dev == ip6n->fb_tnl_dev) {
@@ -1464,8 +1561,7 @@ ip6_tnl_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
  *   %-EINVAL if mtu too small
  **/
 
-static int
-ip6_tnl_change_mtu(struct net_device *dev, int new_mtu)
+int ip6_tnl_change_mtu(struct net_device *dev, int new_mtu)
 {
        struct ip6_tnl *tnl = netdev_priv(dev);
 
@@ -1481,6 +1577,7 @@ ip6_tnl_change_mtu(struct net_device *dev, int new_mtu)
        dev->mtu = new_mtu;
        return 0;
 }
+EXPORT_SYMBOL(ip6_tnl_change_mtu);
 
 int ip6_tnl_get_iflink(const struct net_device *dev)
 {
@@ -1490,16 +1587,74 @@ int ip6_tnl_get_iflink(const struct net_device *dev)
 }
 EXPORT_SYMBOL(ip6_tnl_get_iflink);
 
+int ip6_tnl_encap_add_ops(const struct ip6_tnl_encap_ops *ops,
+                         unsigned int num)
+{
+       if (num >= MAX_IPTUN_ENCAP_OPS)
+               return -ERANGE;
+
+       return !cmpxchg((const struct ip6_tnl_encap_ops **)
+                       &ip6tun_encaps[num],
+                       NULL, ops) ? 0 : -1;
+}
+EXPORT_SYMBOL(ip6_tnl_encap_add_ops);
+
+int ip6_tnl_encap_del_ops(const struct ip6_tnl_encap_ops *ops,
+                         unsigned int num)
+{
+       int ret;
+
+       if (num >= MAX_IPTUN_ENCAP_OPS)
+               return -ERANGE;
+
+       ret = (cmpxchg((const struct ip6_tnl_encap_ops **)
+                      &ip6tun_encaps[num],
+                      ops, NULL) == ops) ? 0 : -1;
+
+       synchronize_net();
+
+       return ret;
+}
+EXPORT_SYMBOL(ip6_tnl_encap_del_ops);
+
+int ip6_tnl_encap_setup(struct ip6_tnl *t,
+                       struct ip_tunnel_encap *ipencap)
+{
+       int hlen;
+
+       memset(&t->encap, 0, sizeof(t->encap));
+
+       hlen = ip6_encap_hlen(ipencap);
+       if (hlen < 0)
+               return hlen;
+
+       t->encap.type = ipencap->type;
+       t->encap.sport = ipencap->sport;
+       t->encap.dport = ipencap->dport;
+       t->encap.flags = ipencap->flags;
+
+       t->encap_hlen = hlen;
+       t->hlen = t->encap_hlen + t->tun_hlen;
+
+       return 0;
+}
+EXPORT_SYMBOL_GPL(ip6_tnl_encap_setup);
+
 static const struct net_device_ops ip6_tnl_netdev_ops = {
        .ndo_init       = ip6_tnl_dev_init,
        .ndo_uninit     = ip6_tnl_dev_uninit,
-       .ndo_start_xmit = ip6_tnl_xmit,
+       .ndo_start_xmit = ip6_tnl_start_xmit,
        .ndo_do_ioctl   = ip6_tnl_ioctl,
        .ndo_change_mtu = ip6_tnl_change_mtu,
        .ndo_get_stats  = ip6_get_stats,
        .ndo_get_iflink = ip6_tnl_get_iflink,
 };
 
+#define IPXIPX_FEATURES (NETIF_F_SG |          \
+                        NETIF_F_FRAGLIST |     \
+                        NETIF_F_HIGHDMA |      \
+                        NETIF_F_GSO_SOFTWARE | \
+                        NETIF_F_HW_CSUM)
 
 /**
  * ip6_tnl_dev_setup - setup virtual tunnel device
@@ -1511,20 +1666,18 @@ static const struct net_device_ops ip6_tnl_netdev_ops = {
 
 static void ip6_tnl_dev_setup(struct net_device *dev)
 {
-       struct ip6_tnl *t;
-
        dev->netdev_ops = &ip6_tnl_netdev_ops;
        dev->destructor = ip6_dev_free;
 
        dev->type = ARPHRD_TUNNEL6;
-       dev->hard_header_len = LL_MAX_HEADER + sizeof(struct ipv6hdr);
-       dev->mtu = ETH_DATA_LEN - sizeof(struct ipv6hdr);
-       t = netdev_priv(dev);
-       if (!(t->parms.flags & IP6_TNL_F_IGN_ENCAP_LIMIT))
-               dev->mtu -= 8;
        dev->flags |= IFF_NOARP;
        dev->addr_len = sizeof(struct in6_addr);
+       dev->features |= NETIF_F_LLTX;
        netif_keep_dst(dev);
+
+       dev->features           |= IPXIPX_FEATURES;
+       dev->hw_features        |= IPXIPX_FEATURES;
+
        /* This perm addr will be used as interface identifier by IPv6 */
        dev->addr_assign_type = NET_ADDR_RANDOM;
        eth_random_addr(dev->perm_addr);
@@ -1541,6 +1694,7 @@ ip6_tnl_dev_init_gen(struct net_device *dev)
 {
        struct ip6_tnl *t = netdev_priv(dev);
        int ret;
+       int t_hlen;
 
        t->dev = dev;
        t->net = dev_net(dev);
@@ -1549,13 +1703,32 @@ ip6_tnl_dev_init_gen(struct net_device *dev)
                return -ENOMEM;
 
        ret = dst_cache_init(&t->dst_cache, GFP_KERNEL);
-       if (ret) {
-               free_percpu(dev->tstats);
-               dev->tstats = NULL;
-               return ret;
-       }
+       if (ret)
+               goto free_stats;
+
+       ret = gro_cells_init(&t->gro_cells, dev);
+       if (ret)
+               goto destroy_dst;
+
+       t->tun_hlen = 0;
+       t->hlen = t->encap_hlen + t->tun_hlen;
+       t_hlen = t->hlen + sizeof(struct ipv6hdr);
+
+       dev->type = ARPHRD_TUNNEL6;
+       dev->hard_header_len = LL_MAX_HEADER + t_hlen;
+       dev->mtu = ETH_DATA_LEN - t_hlen;
+       if (!(t->parms.flags & IP6_TNL_F_IGN_ENCAP_LIMIT))
+               dev->mtu -= 8;
 
        return 0;
+
+destroy_dst:
+       dst_cache_destroy(&t->dst_cache);
+free_stats:
+       free_percpu(dev->tstats);
+       dev->tstats = NULL;
+
+       return ret;
 }
 
 /**
@@ -1643,13 +1816,55 @@ static void ip6_tnl_netlink_parms(struct nlattr *data[],
                parms->proto = nla_get_u8(data[IFLA_IPTUN_PROTO]);
 }
 
+static bool ip6_tnl_netlink_encap_parms(struct nlattr *data[],
+                                       struct ip_tunnel_encap *ipencap)
+{
+       bool ret = false;
+
+       memset(ipencap, 0, sizeof(*ipencap));
+
+       if (!data)
+               return ret;
+
+       if (data[IFLA_IPTUN_ENCAP_TYPE]) {
+               ret = true;
+               ipencap->type = nla_get_u16(data[IFLA_IPTUN_ENCAP_TYPE]);
+       }
+
+       if (data[IFLA_IPTUN_ENCAP_FLAGS]) {
+               ret = true;
+               ipencap->flags = nla_get_u16(data[IFLA_IPTUN_ENCAP_FLAGS]);
+       }
+
+       if (data[IFLA_IPTUN_ENCAP_SPORT]) {
+               ret = true;
+               ipencap->sport = nla_get_be16(data[IFLA_IPTUN_ENCAP_SPORT]);
+       }
+
+       if (data[IFLA_IPTUN_ENCAP_DPORT]) {
+               ret = true;
+               ipencap->dport = nla_get_be16(data[IFLA_IPTUN_ENCAP_DPORT]);
+       }
+
+       return ret;
+}
+
 static int ip6_tnl_newlink(struct net *src_net, struct net_device *dev,
                           struct nlattr *tb[], struct nlattr *data[])
 {
        struct net *net = dev_net(dev);
        struct ip6_tnl *nt, *t;
+       struct ip_tunnel_encap ipencap;
 
        nt = netdev_priv(dev);
+
+       if (ip6_tnl_netlink_encap_parms(data, &ipencap)) {
+               int err = ip6_tnl_encap_setup(nt, &ipencap);
+
+               if (err < 0)
+                       return err;
+       }
+
        ip6_tnl_netlink_parms(data, &nt->parms);
 
        t = ip6_tnl_locate(net, &nt->parms, 0);
@@ -1666,10 +1881,17 @@ static int ip6_tnl_changelink(struct net_device *dev, struct nlattr *tb[],
        struct __ip6_tnl_parm p;
        struct net *net = t->net;
        struct ip6_tnl_net *ip6n = net_generic(net, ip6_tnl_net_id);
+       struct ip_tunnel_encap ipencap;
 
        if (dev == ip6n->fb_tnl_dev)
                return -EINVAL;
 
+       if (ip6_tnl_netlink_encap_parms(data, &ipencap)) {
+               int err = ip6_tnl_encap_setup(t, &ipencap);
+
+               if (err < 0)
+                       return err;
+       }
        ip6_tnl_netlink_parms(data, &p);
 
        t = ip6_tnl_locate(net, &p, 0);
@@ -1710,6 +1932,14 @@ static size_t ip6_tnl_get_size(const struct net_device *dev)
                nla_total_size(4) +
                /* IFLA_IPTUN_PROTO */
                nla_total_size(1) +
+               /* IFLA_IPTUN_ENCAP_TYPE */
+               nla_total_size(2) +
+               /* IFLA_IPTUN_ENCAP_FLAGS */
+               nla_total_size(2) +
+               /* IFLA_IPTUN_ENCAP_SPORT */
+               nla_total_size(2) +
+               /* IFLA_IPTUN_ENCAP_DPORT */
+               nla_total_size(2) +
                0;
 }
 
@@ -1727,6 +1957,17 @@ static int ip6_tnl_fill_info(struct sk_buff *skb, const struct net_device *dev)
            nla_put_u32(skb, IFLA_IPTUN_FLAGS, parm->flags) ||
            nla_put_u8(skb, IFLA_IPTUN_PROTO, parm->proto))
                goto nla_put_failure;
+
+       if (nla_put_u16(skb, IFLA_IPTUN_ENCAP_TYPE,
+                       tunnel->encap.type) ||
+       nla_put_be16(skb, IFLA_IPTUN_ENCAP_SPORT,
+                    tunnel->encap.sport) ||
+       nla_put_be16(skb, IFLA_IPTUN_ENCAP_DPORT,
+                    tunnel->encap.dport) ||
+       nla_put_u16(skb, IFLA_IPTUN_ENCAP_FLAGS,
+                   tunnel->encap.flags))
+               goto nla_put_failure;
+
        return 0;
 
 nla_put_failure:
@@ -1750,6 +1991,10 @@ static const struct nla_policy ip6_tnl_policy[IFLA_IPTUN_MAX + 1] = {
        [IFLA_IPTUN_FLOWINFO]           = { .type = NLA_U32 },
        [IFLA_IPTUN_FLAGS]              = { .type = NLA_U32 },
        [IFLA_IPTUN_PROTO]              = { .type = NLA_U8 },
+       [IFLA_IPTUN_ENCAP_TYPE]         = { .type = NLA_U16 },
+       [IFLA_IPTUN_ENCAP_FLAGS]        = { .type = NLA_U16 },
+       [IFLA_IPTUN_ENCAP_SPORT]        = { .type = NLA_U16 },
+       [IFLA_IPTUN_ENCAP_DPORT]        = { .type = NLA_U16 },
 };
 
 static struct rtnl_link_ops ip6_link_ops __read_mostly = {