net/ipv4/ip_tunnel_core.c

   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(ntohs(eh->h_proto) >= ETH_P_802_3_MIN))
 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         for_each_possible_cpu(i) {
 169                 const struct pcpu_sw_netstats *tstats =
 170                                                    per_cpu_ptr(dev->tstats, i);
 171                 u64 rx_packets, rx_bytes, tx_packets, tx_bytes;
 172                 unsigned int start;
 173
 174                 do {
 175                         start = u64_stats_fetch_begin_irq(&tstats->syncp);
 176                         rx_packets = tstats->rx_packets;
 177                         tx_packets = tstats->tx_packets;
 178                         rx_bytes = tstats->rx_bytes;
 179                         tx_bytes = tstats->tx_bytes;
 180                 } while (u64_stats_fetch_retry_irq(&tstats->syncp, start));
 181
 182                 tot->rx_packets += rx_packets;
 183                 tot->tx_packets += tx_packets;
 184                 tot->rx_bytes   += rx_bytes;
 185                 tot->tx_bytes   += tx_bytes;
 186         }
 187
 188         tot->multicast = dev->stats.multicast;
 189
 190         tot->rx_crc_errors = dev->stats.rx_crc_errors;
 191         tot->rx_fifo_errors = dev->stats.rx_fifo_errors;
 192         tot->rx_length_errors = dev->stats.rx_length_errors;
 193         tot->rx_frame_errors = dev->stats.rx_frame_errors;
 194         tot->rx_errors = dev->stats.rx_errors;
 195
 196         tot->tx_fifo_errors = dev->stats.tx_fifo_errors;
 197         tot->tx_carrier_errors = dev->stats.tx_carrier_errors;
 198         tot->tx_dropped = dev->stats.tx_dropped;
 199         tot->tx_aborted_errors = dev->stats.tx_aborted_errors;
 200         tot->tx_errors = dev->stats.tx_errors;
 201
 202         tot->collisions  = dev->stats.collisions;
 203
 204         return tot;
 205 }
 206 EXPORT_SYMBOL_GPL(ip_tunnel_get_stats64);