1 /* 2 * Copyright (c) 2013 Nicira, Inc. 3 * 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. 7 * 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. 12 * 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 16 * 02110-1301, USA 17 */ 18 19 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt 20 21 #include <linux/types.h> 22 #include <linux/kernel.h> 23 #include <linux/skbuff.h> 24 #include <linux/netdevice.h> 25 #include <linux/in.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> 35 36 #include <net/ip.h> 37 #include <net/icmp.h> 38 #include <net/protocol.h> 39 #include <net/ip_tunnels.h> 40 #include <net/arp.h> 41 #include <net/checksum.h> 42 #include <net/dsfield.h> 43 #include <net/inet_ecn.h> 44 #include <net/xfrm.h> 45 #include <net/net_namespace.h> 46 #include <net/netns/generic.h> 47 #include <net/rtnetlink.h> 48 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) 52 { 53 int pkt_len = skb->len; 54 struct iphdr *iph; 55 int err; 56 57 skb_scrub_packet(skb, xnet); 58 59 skb_clear_hash(skb); 60 skb_dst_set(skb, &rt->dst); 61 memset(IPCB(skb), 0, sizeof(*IPCB(skb))); 62 63 /* Push down and install the IP header. */ 64 skb_push(skb, sizeof(struct iphdr)); 65 skb_reset_network_header(skb); 66 67 iph = ip_hdr(skb); 68 69 iph->version = 4; 70 iph->ihl = sizeof(struct iphdr) >> 2; 71 iph->frag_off = df; 72 iph->protocol = proto; 73 iph->tos = tos; 74 iph->daddr = dst; 75 iph->saddr = src; 76 iph->ttl = ttl; 77 __ip_select_ident(dev_net(rt->dst.dev), iph, 78 skb_shinfo(skb)->gso_segs ?: 1); 79 80 err = ip_local_out_sk(sk, skb); 81 if (unlikely(net_xmit_eval(err))) 82 pkt_len = 0; 83 return pkt_len; 84 } 85 EXPORT_SYMBOL_GPL(iptunnel_xmit); 86 87 int iptunnel_pull_header(struct sk_buff *skb, int hdr_len, __be16 inner_proto) 88 { 89 if (unlikely(!pskb_may_pull(skb, hdr_len))) 90 return -ENOMEM; 91 92 skb_pull_rcsum(skb, hdr_len); 93 94 if (inner_proto == htons(ETH_P_TEB)) { 95 struct ethhdr *eh; 96 97 if (unlikely(!pskb_may_pull(skb, ETH_HLEN))) 98 return -ENOMEM; 99 100 eh = (struct ethhdr *)skb->data; 101 if (likely(eth_proto_is_802_3(eh->h_proto))) 102 skb->protocol = eh->h_proto; 103 else 104 skb->protocol = htons(ETH_P_802_2); 105 106 } else { 107 skb->protocol = inner_proto; 108 } 109 110 nf_reset(skb); 111 secpath_reset(skb); 112 skb_clear_hash_if_not_l4(skb); 113 skb_dst_drop(skb); 114 skb->vlan_tci = 0; 115 skb_set_queue_mapping(skb, 0); 116 skb->pkt_type = PACKET_HOST; 117 return 0; 118 } 119 EXPORT_SYMBOL_GPL(iptunnel_pull_header); 120 121 struct sk_buff *iptunnel_handle_offloads(struct sk_buff *skb, 122 bool csum_help, 123 int gso_type_mask) 124 { 125 int err; 126 127 if (likely(!skb->encapsulation)) { 128 skb_reset_inner_headers(skb); 129 skb->encapsulation = 1; 130 } 131 132 if (skb_is_gso(skb)) { 133 err = skb_unclone(skb, GFP_ATOMIC); 134 if (unlikely(err)) 135 goto error; 136 skb_shinfo(skb)->gso_type |= gso_type_mask; 137 return skb; 138 } 139 140 /* 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. 144 */ 145 if (csum_help) 146 skb->encapsulation = 0; 147 148 if (skb->ip_summed == CHECKSUM_PARTIAL && csum_help) { 149 err = skb_checksum_help(skb); 150 if (unlikely(err)) 151 goto error; 152 } else if (skb->ip_summed != CHECKSUM_PARTIAL) 153 skb->ip_summed = CHECKSUM_NONE; 154 155 return skb; 156 error: 157 kfree_skb(skb); 158 return ERR_PTR(err); 159 } 160 EXPORT_SYMBOL_GPL(iptunnel_handle_offloads); 161 162 /* 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) 165 { 166 int i; 167 168 netdev_stats_to_stats64(tot, &dev->stats); 169 170 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; 174 unsigned int start; 175 176 do { 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)); 183 184 tot->rx_packets += rx_packets; 185 tot->tx_packets += tx_packets; 186 tot->rx_bytes += rx_bytes; 187 tot->tx_bytes += tx_bytes; 188 } 189 190 return tot; 191 } 192 EXPORT_SYMBOL_GPL(ip_tunnel_get_stats64); 193 194 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 }, 203 }; 204 205 static int ip_tun_build_state(struct net_device *dev, struct nlattr *attr, 206 struct lwtunnel_state **ts) 207 { 208 struct ip_tunnel_info *tun_info; 209 struct lwtunnel_state *new_state; 210 struct nlattr *tb[IP_TUN_MAX + 1]; 211 int err; 212 213 err = nla_parse_nested(tb, IP_TUN_MAX, attr, ip_tun_policy); 214 if (err < 0) 215 return err; 216 217 new_state = lwtunnel_state_alloc(sizeof(*tun_info)); 218 if (!new_state) 219 return -ENOMEM; 220 221 new_state->type = LWTUNNEL_ENCAP_IP; 222 223 tun_info = lwt_tun_info(new_state); 224 225 if (tb[IP_TUN_ID]) 226 tun_info->key.tun_id = nla_get_u64(tb[IP_TUN_ID]); 227 228 if (tb[IP_TUN_DST]) 229 tun_info->key.ipv4_dst = nla_get_be32(tb[IP_TUN_DST]); 230 231 if (tb[IP_TUN_SRC]) 232 tun_info->key.ipv4_src = nla_get_be32(tb[IP_TUN_SRC]); 233 234 if (tb[IP_TUN_TTL]) 235 tun_info->key.ipv4_ttl = nla_get_u8(tb[IP_TUN_TTL]); 236 237 if (tb[IP_TUN_TOS]) 238 tun_info->key.ipv4_tos = nla_get_u8(tb[IP_TUN_TOS]); 239 240 if (tb[IP_TUN_SPORT]) 241 tun_info->key.tp_src = nla_get_be16(tb[IP_TUN_SPORT]); 242 243 if (tb[IP_TUN_DPORT]) 244 tun_info->key.tp_dst = nla_get_be16(tb[IP_TUN_DPORT]); 245 246 if (tb[IP_TUN_FLAGS]) 247 tun_info->key.tun_flags = nla_get_u16(tb[IP_TUN_FLAGS]); 248 249 tun_info->mode = IP_TUNNEL_INFO_TX; 250 tun_info->options = NULL; 251 tun_info->options_len = 0; 252 253 *ts = new_state; 254 255 return 0; 256 } 257 258 static int ip_tun_fill_encap_info(struct sk_buff *skb, 259 struct lwtunnel_state *lwtstate) 260 { 261 struct ip_tunnel_info *tun_info = lwt_tun_info(lwtstate); 262 263 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)) 271 return -ENOMEM; 272 273 return 0; 274 } 275 276 static int ip_tun_encap_nlsize(struct lwtunnel_state *lwtstate) 277 { 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 */ 286 } 287 288 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, 292 }; 293 294 static int __init ip_tunnel_core_init(void) 295 { 296 lwtunnel_encap_add_ops(&ip_tun_lwt_ops, LWTUNNEL_ENCAP_IP); 297 298 return 0; 299 } 300 module_init(ip_tunnel_core_init); 301 302 static void __exit ip_tunnel_core_exit(void) 303 { 304 lwtunnel_encap_del_ops(&ip_tun_lwt_ops, LWTUNNEL_ENCAP_IP); 305 } 306 module_exit(ip_tunnel_core_exit); 307