net/mpls/mpls_iptunnel.c

   1 /*
   2  * mpls tunnels An implementation mpls tunnels using the light weight tunnel
   3  *              infrastructure
   4  *
   5  * Authors:     Roopa Prabhu, <roopa@cumulusnetworks.com>
   6  *
   7  *              This program is free software; you can redistribute it and/or
   8  *              modify it under the terms of the GNU General Public License
   9  *              as published by the Free Software Foundation; either version
  10  *              2 of the License, or (at your option) any later version.
  11  *
  12  */
  13 #include <linux/types.h>
  14 #include <linux/skbuff.h>
  15 #include <linux/net.h>
  16 #include <linux/module.h>
  17 #include <linux/mpls.h>
  18 #include <linux/vmalloc.h>
  19 #include <net/ip.h>
  20 #include <net/dst.h>
  21 #include <net/lwtunnel.h>
  22 #include <net/netevent.h>
  23 #include <net/netns/generic.h>
  24 #include <net/ip6_fib.h>
  25 #include <net/route.h>
  26 #include <net/mpls_iptunnel.h>
  27 #include <linux/mpls_iptunnel.h>
  28 #include "internal.h"
  29
  30 static const struct nla_policy mpls_iptunnel_policy[MPLS_IPTUNNEL_MAX + 1] = {
  31         [MPLS_IPTUNNEL_DST]     = { .type = NLA_U32 },
  32 };
  33
  34 static unsigned int mpls_encap_size(struct mpls_iptunnel_encap *en)
  35 {
  36         /* The size of the layer 2.5 labels to be added for this route */
  37         return en->labels * sizeof(struct mpls_shim_hdr);
  38 }
  39
  40 static int mpls_output(struct net *net, struct sock *sk, struct sk_buff *skb)
  41 {
  42         struct mpls_iptunnel_encap *tun_encap_info;
  43         struct mpls_shim_hdr *hdr;
  44         struct net_device *out_dev;
  45         unsigned int hh_len;
  46         unsigned int new_header_size;
  47         unsigned int mtu;
  48         struct dst_entry *dst = skb_dst(skb);
  49         struct rtable *rt = NULL;
  50         struct rt6_info *rt6 = NULL;
  51         int err = 0;
  52         bool bos;
  53         int i;
  54         unsigned int ttl;
  55
  56         /* Obtain the ttl */
  57         if (dst->ops->family == AF_INET) {
  58                 ttl = ip_hdr(skb)->ttl;
  59                 rt = (struct rtable *)dst;
  60         } else if (dst->ops->family == AF_INET6) {
  61                 ttl = ipv6_hdr(skb)->hop_limit;
  62                 rt6 = (struct rt6_info *)dst;
  63         } else {
  64                 goto drop;
  65         }
  66
  67         skb_orphan(skb);
  68
  69         /* Find the output device */
  70         out_dev = dst->dev;
  71         if (!mpls_output_possible(out_dev) ||
  72             !dst->lwtstate || skb_warn_if_lro(skb))
  73                 goto drop;
  74
  75         skb_forward_csum(skb);
  76
  77         tun_encap_info = mpls_lwtunnel_encap(dst->lwtstate);
  78
  79         /* Verify the destination can hold the packet */
  80         new_header_size = mpls_encap_size(tun_encap_info);
  81         mtu = mpls_dev_mtu(out_dev);
  82         if (mpls_pkt_too_big(skb, mtu - new_header_size))
  83                 goto drop;
  84
  85         hh_len = LL_RESERVED_SPACE(out_dev);
  86         if (!out_dev->header_ops)
  87                 hh_len = 0;
  88
  89         /* Ensure there is enough space for the headers in the skb */
  90         if (skb_cow(skb, hh_len + new_header_size))
  91                 goto drop;
  92
  93         skb_push(skb, new_header_size);
  94         skb_reset_network_header(skb);
  95
  96         skb->dev = out_dev;
  97         skb->protocol = htons(ETH_P_MPLS_UC);
  98
  99         /* Push the new labels */
 100         hdr = mpls_hdr(skb);
 101         bos = true;
 102         for (i = tun_encap_info->labels - 1; i >= 0; i--) {
 103                 hdr[i] = mpls_entry_encode(tun_encap_info->label[i],
 104                                            ttl, 0, bos);
 105                 bos = false;
 106         }
 107
 108         if (rt)
 109                 err = neigh_xmit(NEIGH_ARP_TABLE, out_dev, &rt->rt_gateway,
 110                                  skb);
 111         else if (rt6)
 112                 err = neigh_xmit(NEIGH_ND_TABLE, out_dev, &rt6->rt6i_gateway,
 113                                  skb);
 114         if (err)
 115                 net_dbg_ratelimited("%s: packet transmission failed: %d\n",
 116                                     __func__, err);
 117
 118         return 0;
 119
 120 drop:
 121         kfree_skb(skb);
 122         return -EINVAL;
 123 }
 124
 125 static int mpls_build_state(struct net_device *dev, struct nlattr *nla,
 126                             unsigned int family, const void *cfg,
 127                             struct lwtunnel_state **ts)
 128 {
 129         struct mpls_iptunnel_encap *tun_encap_info;
 130         struct nlattr *tb[MPLS_IPTUNNEL_MAX + 1];
 131         struct lwtunnel_state *newts;
 132         int tun_encap_info_len;
 133         int ret;
 134
 135         ret = nla_parse_nested(tb, MPLS_IPTUNNEL_MAX, nla,
 136                                mpls_iptunnel_policy);
 137         if (ret < 0)
 138                 return ret;
 139
 140         if (!tb[MPLS_IPTUNNEL_DST])
 141                 return -EINVAL;
 142
 143         tun_encap_info_len = sizeof(*tun_encap_info);
 144
 145         newts = lwtunnel_state_alloc(tun_encap_info_len);
 146         if (!newts)
 147                 return -ENOMEM;
 148
 149         newts->len = tun_encap_info_len;
 150         tun_encap_info = mpls_lwtunnel_encap(newts);
 151         ret = nla_get_labels(tb[MPLS_IPTUNNEL_DST], MAX_NEW_LABELS,
 152                              &tun_encap_info->labels, tun_encap_info->label);
 153         if (ret)
 154                 goto errout;
 155         newts->type = LWTUNNEL_ENCAP_MPLS;
 156         newts->flags |= LWTUNNEL_STATE_OUTPUT_REDIRECT;
 157
 158         *ts = newts;
 159
 160         return 0;
 161
 162 errout:
 163         kfree(newts);
 164         *ts = NULL;
 165
 166         return ret;
 167 }
 168
 169 static int mpls_fill_encap_info(struct sk_buff *skb,
 170                                 struct lwtunnel_state *lwtstate)
 171 {
 172         struct mpls_iptunnel_encap *tun_encap_info;
 173
 174         tun_encap_info = mpls_lwtunnel_encap(lwtstate);
 175
 176         if (nla_put_labels(skb, MPLS_IPTUNNEL_DST, tun_encap_info->labels,
 177                            tun_encap_info->label))
 178                 goto nla_put_failure;
 179
 180         return 0;
 181
 182 nla_put_failure:
 183         return -EMSGSIZE;
 184 }
 185
 186 static int mpls_encap_nlsize(struct lwtunnel_state *lwtstate)
 187 {
 188         struct mpls_iptunnel_encap *tun_encap_info;
 189
 190         tun_encap_info = mpls_lwtunnel_encap(lwtstate);
 191
 192         return nla_total_size(tun_encap_info->labels * 4);
 193 }
 194
 195 static int mpls_encap_cmp(struct lwtunnel_state *a, struct lwtunnel_state *b)
 196 {
 197         struct mpls_iptunnel_encap *a_hdr = mpls_lwtunnel_encap(a);
 198         struct mpls_iptunnel_encap *b_hdr = mpls_lwtunnel_encap(b);
 199         int l;
 200
 201         if (a_hdr->labels != b_hdr->labels)
 202                 return 1;
 203
 204         for (l = 0; l < MAX_NEW_LABELS; l++)
 205                 if (a_hdr->label[l] != b_hdr->label[l])
 206                         return 1;
 207         return 0;
 208 }
 209
 210 static const struct lwtunnel_encap_ops mpls_iptun_ops = {
 211         .build_state = mpls_build_state,
 212         .output = mpls_output,
 213         .fill_encap = mpls_fill_encap_info,
 214         .get_encap_size = mpls_encap_nlsize,
 215         .cmp_encap = mpls_encap_cmp,
 216 };
 217
 218 static int __init mpls_iptunnel_init(void)
 219 {
 220         return lwtunnel_encap_add_ops(&mpls_iptun_ops, LWTUNNEL_ENCAP_MPLS);
 221 }
 222 module_init(mpls_iptunnel_init);
 223
 224 static void __exit mpls_iptunnel_exit(void)
 225 {
 226         lwtunnel_encap_del_ops(&mpls_iptun_ops, LWTUNNEL_ENCAP_MPLS);
 227 }
 228 module_exit(mpls_iptunnel_exit);
 229
 230 MODULE_ALIAS_RTNL_LWT(MPLS);
 231 MODULE_DESCRIPTION("MultiProtocol Label Switching IP Tunnels");
 232 MODULE_LICENSE("GPL v2");