include/net/route.h

   1 /*
   2  * INET         An implementation of the TCP/IP protocol suite for the LINUX
   3  *              operating system.  INET  is implemented using the  BSD Socket
   4  *              interface as the means of communication with the user level.
   5  *
   6  *              Definitions for the IP router.
   7  *
   8  * Version:     @(#)route.h     1.0.4   05/27/93
   9  *
  10  * Authors:     Ross Biro
  11  *              Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
  12  * Fixes:
  13  *              Alan Cox        :       Reformatted. Added ip_rt_local()
  14  *              Alan Cox        :       Support for TCP parameters.
  15  *              Alexey Kuznetsov:       Major changes for new routing code.
  16  *              Mike McLagan    :       Routing by source
  17  *              Robert Olsson   :       Added rt_cache statistics
  18  *
  19  *              This program is free software; you can redistribute it and/or
  20  *              modify it under the terms of the GNU General Public License
  21  *              as published by the Free Software Foundation; either version
  22  *              2 of the License, or (at your option) any later version.
  23  */
  24 #ifndef _ROUTE_H
  25 #define _ROUTE_H
  26
  27 #include <net/dst.h>
  28 #include <net/inetpeer.h>
  29 #include <net/flow.h>
  30 #include <net/inet_sock.h>
  31 #include <linux/in_route.h>
  32 #include <linux/rtnetlink.h>
  33 #include <linux/route.h>
  34 #include <linux/ip.h>
  35 #include <linux/cache.h>
  36 #include <linux/security.h>
  37
  38 #ifndef __KERNEL__
  39 #warning This file is not supposed to be used outside of kernel.
  40 #endif
  41
  42 #define RTO_ONLINK      0x01
  43
  44 #define RTO_CONN        0
  45 /* RTO_CONN is not used (being alias for 0), but preserved not to break
  46  * some modules referring to it. */
  47
  48 #define RT_CONN_FLAGS(sk)   (RT_TOS(inet_sk(sk)->tos) | sock_flag(sk, SOCK_LOCALROUTE))
  49
  50 struct fib_nh;
  51 struct inet_peer;
  52 struct fib_info;
  53 struct rtable {
  54         struct dst_entry        dst;
  55
  56         /* Lookup key. */
  57         __be32                  rt_key_dst;
  58         __be32                  rt_key_src;
  59
  60         int                     rt_genid;
  61         unsigned                rt_flags;
  62         __u16                   rt_type;
  63         __u8                    rt_tos;
  64
  65         __be32                  rt_dst; /* Path destination     */
  66         __be32                  rt_src; /* Path source          */
  67         int                     rt_route_iif;
  68         int                     rt_iif;
  69         int                     rt_oif;
  70         __u32                   rt_mark;
  71
  72         /* Info on neighbour */
  73         __be32                  rt_gateway;
  74
  75         /* Miscellaneous cached information */
  76         __be32                  rt_spec_dst; /* RFC1122 specific destination */
  77         u32                     rt_peer_genid;
  78         struct inet_peer        *peer; /* long-living peer info */
  79         struct fib_info         *fi; /* for client ref to shared metrics */
  80 };
  81
  82 static inline bool rt_is_input_route(struct rtable *rt)
  83 {
  84         return rt->rt_route_iif != 0;
  85 }
  86
  87 static inline bool rt_is_output_route(struct rtable *rt)
  88 {
  89         return rt->rt_route_iif == 0;
  90 }
  91
  92 struct ip_rt_acct {
  93         __u32   o_bytes;
  94         __u32   o_packets;
  95         __u32   i_bytes;
  96         __u32   i_packets;
  97 };
  98
  99 struct rt_cache_stat {
 100         unsigned int in_hit;
 101         unsigned int in_slow_tot;
 102         unsigned int in_slow_mc;
 103         unsigned int in_no_route;
 104         unsigned int in_brd;
 105         unsigned int in_martian_dst;
 106         unsigned int in_martian_src;
 107         unsigned int out_hit;
 108         unsigned int out_slow_tot;
 109         unsigned int out_slow_mc;
 110         unsigned int gc_total;
 111         unsigned int gc_ignored;
 112         unsigned int gc_goal_miss;
 113         unsigned int gc_dst_overflow;
 114         unsigned int in_hlist_search;
 115         unsigned int out_hlist_search;
 116 };
 117
 118 extern struct ip_rt_acct __percpu *ip_rt_acct;
 119
 120 struct in_device;
 121 extern int              ip_rt_init(void);
 122 extern void             ip_rt_redirect(__be32 old_gw, __be32 dst, __be32 new_gw,
 123                                        __be32 src, struct net_device *dev);
 124 extern void             rt_cache_flush(struct net *net, int how);
 125 extern void             rt_cache_flush_batch(struct net *net);
 126 extern struct rtable *__ip_route_output_key(struct net *, const struct flowi4 *flp);
 127 extern struct rtable *ip_route_output_flow(struct net *, struct flowi4 *flp,
 128                                            struct sock *sk);
 129 extern struct dst_entry *ipv4_blackhole_route(struct net *net, struct dst_entry *dst_orig);
 130
 131 static inline struct rtable *ip_route_output_key(struct net *net, struct flowi4 *flp)
 132 {
 133         return ip_route_output_flow(net, flp, NULL);
 134 }
 135
 136 static inline struct rtable *ip_route_output(struct net *net, __be32 daddr,
 137                                              __be32 saddr, u8 tos, int oif)
 138 {
 139         struct flowi4 fl4 = {
 140                 .flowi4_oif = oif,
 141                 .daddr = daddr,
 142                 .saddr = saddr,
 143                 .flowi4_tos = tos,
 144         };
 145         return ip_route_output_key(net, &fl4);
 146 }
 147
 148 static inline struct rtable *ip_route_output_ports(struct net *net, struct sock *sk,
 149                                                    __be32 daddr, __be32 saddr,
 150                                                    __be16 dport, __be16 sport,
 151                                                    __u8 proto, __u8 tos, int oif)
 152 {
 153         struct flowi4 fl4;
 154
 155         flowi4_init_output(&fl4, oif, sk ? sk->sk_mark : 0, tos,
 156                            RT_SCOPE_UNIVERSE, proto,
 157                            sk ? inet_sk_flowi_flags(sk) : 0,
 158                            daddr, saddr, dport, sport);
 159         if (sk)
 160                 security_sk_classify_flow(sk, flowi4_to_flowi(&fl4));
 161         return ip_route_output_flow(net, &fl4, sk);
 162 }
 163
 164 static inline struct rtable *ip_route_output_gre(struct net *net,
 165                                                  __be32 daddr, __be32 saddr,
 166                                                  __be32 gre_key, __u8 tos, int oif)
 167 {
 168         struct flowi4 fl4 = {
 169                 .flowi4_oif = oif,
 170                 .daddr = daddr,
 171                 .saddr = saddr,
 172                 .flowi4_tos = tos,
 173                 .flowi4_proto = IPPROTO_GRE,
 174                 .fl4_gre_key = gre_key,
 175         };
 176         return ip_route_output_key(net, &fl4);
 177 }
 178
 179 extern int ip_route_input_common(struct sk_buff *skb, __be32 dst, __be32 src,
 180                                  u8 tos, struct net_device *devin, bool noref);
 181
 182 static inline int ip_route_input(struct sk_buff *skb, __be32 dst, __be32 src,
 183                                  u8 tos, struct net_device *devin)
 184 {
 185         return ip_route_input_common(skb, dst, src, tos, devin, false);
 186 }
 187
 188 static inline int ip_route_input_noref(struct sk_buff *skb, __be32 dst, __be32 src,
 189                                        u8 tos, struct net_device *devin)
 190 {
 191         return ip_route_input_common(skb, dst, src, tos, devin, true);
 192 }
 193
 194 extern unsigned short   ip_rt_frag_needed(struct net *net, struct iphdr *iph, unsigned short new_mtu, struct net_device *dev);
 195 extern void             ip_rt_send_redirect(struct sk_buff *skb);
 196
 197 extern unsigned         inet_addr_type(struct net *net, __be32 addr);
 198 extern unsigned         inet_dev_addr_type(struct net *net, const struct net_device *dev, __be32 addr);
 199 extern void             ip_rt_multicast_event(struct in_device *);
 200 extern int              ip_rt_ioctl(struct net *, unsigned int cmd, void __user *arg);
 201 extern void             ip_rt_get_source(u8 *src, struct rtable *rt);
 202 extern int              ip_rt_dump(struct sk_buff *skb,  struct netlink_callback *cb);
 203
 204 struct in_ifaddr;
 205 extern void fib_add_ifaddr(struct in_ifaddr *);
 206 extern void fib_del_ifaddr(struct in_ifaddr *, struct in_ifaddr *);
 207
 208 static inline void ip_rt_put(struct rtable * rt)
 209 {
 210         if (rt)
 211                 dst_release(&rt->dst);
 212 }
 213
 214 #define IPTOS_RT_MASK   (IPTOS_TOS_MASK & ~3)
 215
 216 extern const __u8 ip_tos2prio[16];
 217
 218 static inline char rt_tos2priority(u8 tos)
 219 {
 220         return ip_tos2prio[IPTOS_TOS(tos)>>1];
 221 }
 222
 223 static inline struct rtable *ip_route_connect(__be32 dst, __be32 src, u32 tos,
 224                                               int oif, u8 protocol,
 225                                               __be16 sport, __be16 dport,
 226                                               struct sock *sk, bool can_sleep)
 227 {
 228         struct net *net = sock_net(sk);
 229         struct rtable *rt;
 230         struct flowi4 fl4;
 231         __u8 flow_flags;
 232
 233         flow_flags = 0;
 234         if (inet_sk(sk)->transparent)
 235                 flow_flags |= FLOWI_FLAG_ANYSRC;
 236         if (protocol == IPPROTO_TCP)
 237                 flow_flags |= FLOWI_FLAG_PRECOW_METRICS;
 238         if (can_sleep)
 239                 flow_flags |= FLOWI_FLAG_CAN_SLEEP;
 240
 241         flowi4_init_output(&fl4, oif, sk->sk_mark, tos, RT_SCOPE_UNIVERSE,
 242                            protocol, flow_flags, dst, src, dport, sport);
 243
 244         if (!dst || !src) {
 245                 rt = __ip_route_output_key(net, &fl4);
 246                 if (IS_ERR(rt))
 247                         return rt;
 248                 fl4.daddr = rt->rt_dst;
 249                 fl4.saddr = rt->rt_src;
 250                 ip_rt_put(rt);
 251         }
 252         security_sk_classify_flow(sk, flowi4_to_flowi(&fl4));
 253         return ip_route_output_flow(net, &fl4, sk);
 254 }
 255
 256 static inline struct rtable *ip_route_newports(struct rtable *rt,
 257                                                u8 protocol, __be16 orig_sport,
 258                                                __be16 orig_dport, __be16 sport,
 259                                                __be16 dport, struct sock *sk)
 260 {
 261         if (sport != orig_sport || dport != orig_dport) {
 262                 struct flowi4 fl4;
 263                 __u8 flow_flags;
 264
 265                 flow_flags = 0;
 266                 if (inet_sk(sk)->transparent)
 267                         flow_flags |= FLOWI_FLAG_ANYSRC;
 268                 if (protocol == IPPROTO_TCP)
 269                         flow_flags |= FLOWI_FLAG_PRECOW_METRICS;
 270                 flowi4_init_output(&fl4, rt->rt_oif, rt->rt_mark, rt->rt_tos,
 271                                    RT_SCOPE_UNIVERSE, protocol, flow_flags,
 272                                    rt->rt_dst, rt->rt_src, dport, sport);
 273                 ip_rt_put(rt);
 274                 security_sk_classify_flow(sk, flowi4_to_flowi(&fl4));
 275                 return ip_route_output_flow(sock_net(sk), &fl4, sk);
 276         }
 277         return rt;
 278 }
 279
 280 extern void rt_bind_peer(struct rtable *rt, int create);
 281
 282 static inline struct inet_peer *rt_get_peer(struct rtable *rt)
 283 {
 284         if (rt->peer)
 285                 return rt->peer;
 286
 287         rt_bind_peer(rt, 0);
 288         return rt->peer;
 289 }
 290
 291 static inline int inet_iif(const struct sk_buff *skb)
 292 {
 293         return skb_rtable(skb)->rt_iif;
 294 }
 295
 296 extern int sysctl_ip_default_ttl;
 297
 298 static inline int ip4_dst_hoplimit(const struct dst_entry *dst)
 299 {
 300         int hoplimit = dst_metric_raw(dst, RTAX_HOPLIMIT);
 301
 302         if (hoplimit == 0)
 303                 hoplimit = sysctl_ip_default_ttl;
 304         return hoplimit;
 305 }
 306
 307 #endif  /* _ROUTE_H */