2 * Copyright (c) 2013 Nicira, Inc.
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of version 2 of the GNU General Public
6 * License as published by the Free Software Foundation.
8 * This program is distributed in the hope that it will be useful, but
9 * WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
11 * General Public License for more details.
13 * You should have received a copy of the GNU General Public License
14 * along with this program; if not, write to the Free Software
15 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
19 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
21 #include <linux/types.h>
22 #include <linux/kernel.h>
23 #include <linux/skbuff.h>
24 #include <linux/netdevice.h>
26 #include <linux/if_arp.h>
27 #include <linux/mroute.h>
28 #include <linux/init.h>
29 #include <linux/in6.h>
30 #include <linux/inetdevice.h>
31 #include <linux/netfilter_ipv4.h>
32 #include <linux/etherdevice.h>
33 #include <linux/if_ether.h>
34 #include <linux/if_vlan.h>
38 #include <net/protocol.h>
39 #include <net/ip_tunnels.h>
41 #include <net/checksum.h>
42 #include <net/dsfield.h>
43 #include <net/inet_ecn.h>
45 #include <net/net_namespace.h>
46 #include <net/netns/generic.h>
47 #include <net/rtnetlink.h>
49 int iptunnel_xmit(struct sock *sk, struct rtable *rt, struct sk_buff *skb,
50 __be32 src, __be32 dst, __u8 proto,
51 __u8 tos, __u8 ttl, __be16 df, bool xnet)
53 int pkt_len = skb->len;
57 skb_scrub_packet(skb, xnet);
60 skb_dst_set(skb, &rt->dst);
61 memset(IPCB(skb), 0, sizeof(*IPCB(skb)));
63 /* Push down and install the IP header. */
64 skb_push(skb, sizeof(struct iphdr));
65 skb_reset_network_header(skb);
70 iph->ihl = sizeof(struct iphdr) >> 2;
72 iph->protocol = proto;
77 __ip_select_ident(dev_net(rt->dst.dev), iph,
78 skb_shinfo(skb)->gso_segs ?: 1);
80 err = ip_local_out_sk(sk, skb);
81 if (unlikely(net_xmit_eval(err)))
85 EXPORT_SYMBOL_GPL(iptunnel_xmit);
87 int iptunnel_pull_header(struct sk_buff *skb, int hdr_len, __be16 inner_proto)
89 if (unlikely(!pskb_may_pull(skb, hdr_len)))
92 skb_pull_rcsum(skb, hdr_len);
94 if (inner_proto == htons(ETH_P_TEB)) {
97 if (unlikely(!pskb_may_pull(skb, ETH_HLEN)))
100 eh = (struct ethhdr *)skb->data;
101 if (likely(eth_proto_is_802_3(eh->h_proto)))
102 skb->protocol = eh->h_proto;
104 skb->protocol = htons(ETH_P_802_2);
107 skb->protocol = inner_proto;
112 skb_clear_hash_if_not_l4(skb);
115 skb_set_queue_mapping(skb, 0);
116 skb->pkt_type = PACKET_HOST;
119 EXPORT_SYMBOL_GPL(iptunnel_pull_header);
121 struct sk_buff *iptunnel_handle_offloads(struct sk_buff *skb,
127 if (likely(!skb->encapsulation)) {
128 skb_reset_inner_headers(skb);
129 skb->encapsulation = 1;
132 if (skb_is_gso(skb)) {
133 err = skb_unclone(skb, GFP_ATOMIC);
136 skb_shinfo(skb)->gso_type |= gso_type_mask;
140 /* If packet is not gso and we are resolving any partial checksum,
141 * clear encapsulation flag. This allows setting CHECKSUM_PARTIAL
142 * on the outer header without confusing devices that implement
143 * NETIF_F_IP_CSUM with encapsulation.
146 skb->encapsulation = 0;
148 if (skb->ip_summed == CHECKSUM_PARTIAL && csum_help) {
149 err = skb_checksum_help(skb);
152 } else if (skb->ip_summed != CHECKSUM_PARTIAL)
153 skb->ip_summed = CHECKSUM_NONE;
160 EXPORT_SYMBOL_GPL(iptunnel_handle_offloads);
162 /* Often modified stats are per cpu, other are shared (netdev->stats) */
163 struct rtnl_link_stats64 *ip_tunnel_get_stats64(struct net_device *dev,
164 struct rtnl_link_stats64 *tot)
168 netdev_stats_to_stats64(tot, &dev->stats);
170 for_each_possible_cpu(i) {
171 const struct pcpu_sw_netstats *tstats =
172 per_cpu_ptr(dev->tstats, i);
173 u64 rx_packets, rx_bytes, tx_packets, tx_bytes;
177 start = u64_stats_fetch_begin_irq(&tstats->syncp);
178 rx_packets = tstats->rx_packets;
179 tx_packets = tstats->tx_packets;
180 rx_bytes = tstats->rx_bytes;
181 tx_bytes = tstats->tx_bytes;
182 } while (u64_stats_fetch_retry_irq(&tstats->syncp, start));
184 tot->rx_packets += rx_packets;
185 tot->tx_packets += tx_packets;
186 tot->rx_bytes += rx_bytes;
187 tot->tx_bytes += tx_bytes;
192 EXPORT_SYMBOL_GPL(ip_tunnel_get_stats64);
194 static const struct nla_policy ip_tun_policy[IP_TUN_MAX + 1] = {
195 [IP_TUN_ID] = { .type = NLA_U64 },
196 [IP_TUN_DST] = { .type = NLA_U32 },
197 [IP_TUN_SRC] = { .type = NLA_U32 },
198 [IP_TUN_TTL] = { .type = NLA_U8 },
199 [IP_TUN_TOS] = { .type = NLA_U8 },
200 [IP_TUN_SPORT] = { .type = NLA_U16 },
201 [IP_TUN_DPORT] = { .type = NLA_U16 },
202 [IP_TUN_FLAGS] = { .type = NLA_U16 },
205 static int ip_tun_build_state(struct net_device *dev, struct nlattr *attr,
206 struct lwtunnel_state **ts)
208 struct ip_tunnel_info *tun_info;
209 struct lwtunnel_state *new_state;
210 struct nlattr *tb[IP_TUN_MAX + 1];
213 err = nla_parse_nested(tb, IP_TUN_MAX, attr, ip_tun_policy);
217 new_state = lwtunnel_state_alloc(sizeof(*tun_info));
221 new_state->type = LWTUNNEL_ENCAP_IP;
223 tun_info = lwt_tun_info(new_state);
226 tun_info->key.tun_id = nla_get_u64(tb[IP_TUN_ID]);
229 tun_info->key.ipv4_dst = nla_get_be32(tb[IP_TUN_DST]);
232 tun_info->key.ipv4_src = nla_get_be32(tb[IP_TUN_SRC]);
235 tun_info->key.ipv4_ttl = nla_get_u8(tb[IP_TUN_TTL]);
238 tun_info->key.ipv4_tos = nla_get_u8(tb[IP_TUN_TOS]);
240 if (tb[IP_TUN_SPORT])
241 tun_info->key.tp_src = nla_get_be16(tb[IP_TUN_SPORT]);
243 if (tb[IP_TUN_DPORT])
244 tun_info->key.tp_dst = nla_get_be16(tb[IP_TUN_DPORT]);
246 if (tb[IP_TUN_FLAGS])
247 tun_info->key.tun_flags = nla_get_u16(tb[IP_TUN_FLAGS]);
249 tun_info->mode = IP_TUNNEL_INFO_TX;
250 tun_info->options = NULL;
251 tun_info->options_len = 0;
258 static int ip_tun_fill_encap_info(struct sk_buff *skb,
259 struct lwtunnel_state *lwtstate)
261 struct ip_tunnel_info *tun_info = lwt_tun_info(lwtstate);
263 if (nla_put_u64(skb, IP_TUN_ID, tun_info->key.tun_id) ||
264 nla_put_be32(skb, IP_TUN_DST, tun_info->key.ipv4_dst) ||
265 nla_put_be32(skb, IP_TUN_SRC, tun_info->key.ipv4_src) ||
266 nla_put_u8(skb, IP_TUN_TOS, tun_info->key.ipv4_tos) ||
267 nla_put_u8(skb, IP_TUN_TTL, tun_info->key.ipv4_ttl) ||
268 nla_put_u16(skb, IP_TUN_SPORT, tun_info->key.tp_src) ||
269 nla_put_u16(skb, IP_TUN_DPORT, tun_info->key.tp_dst) ||
270 nla_put_u16(skb, IP_TUN_FLAGS, tun_info->key.tun_flags))
276 static int ip_tun_encap_nlsize(struct lwtunnel_state *lwtstate)
278 return nla_total_size(8) /* IP_TUN_ID */
279 + nla_total_size(4) /* IP_TUN_DST */
280 + nla_total_size(4) /* IP_TUN_SRC */
281 + nla_total_size(1) /* IP_TUN_TOS */
282 + nla_total_size(1) /* IP_TUN_TTL */
283 + nla_total_size(2) /* IP_TUN_SPORT */
284 + nla_total_size(2) /* IP_TUN_DPORT */
285 + nla_total_size(2); /* IP_TUN_FLAGS */
288 static const struct lwtunnel_encap_ops ip_tun_lwt_ops = {
289 .build_state = ip_tun_build_state,
290 .fill_encap = ip_tun_fill_encap_info,
291 .get_encap_size = ip_tun_encap_nlsize,
294 static int __init ip_tunnel_core_init(void)
296 lwtunnel_encap_add_ops(&ip_tun_lwt_ops, LWTUNNEL_ENCAP_IP);
300 module_init(ip_tunnel_core_init);
302 static void __exit ip_tunnel_core_exit(void)
304 lwtunnel_encap_del_ops(&ip_tun_lwt_ops, LWTUNNEL_ENCAP_IP);
306 module_exit(ip_tunnel_core_exit);