2 * Copyright (c) 2015-2016
3 * Alexander V. Chernikov <melifaro@FreeBSD.org>
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
13 * 3. Neither the name of the University nor the names of its contributors
14 * may be used to endorse or promote products derived from this software
15 * without specific prior written permission.
17 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 #ifndef _NET_ROUTE_VAR_H_
33 #define _NET_ROUTE_VAR_H_
36 #include <net/radix.h>
39 #include <sys/epoch.h>
40 #include <netinet/in.h> /* struct sockaddr_in */
41 #include <sys/counter.h>
44 typedef int rnh_preadd_entry_f_t(u_int fibnum, const struct sockaddr *addr,
45 const struct sockaddr *mask, struct nhop_object *nh);
48 struct radix_head head;
49 rn_matchaddr_f_t *rnh_matchaddr; /* longest match for sockaddr */
50 rn_addaddr_f_t *rnh_addaddr; /* add based on sockaddr*/
51 rn_deladdr_f_t *rnh_deladdr; /* remove based on sockaddr */
52 rn_lookup_f_t *rnh_lookup; /* exact match for sockaddr */
53 rn_walktree_t *rnh_walktree; /* traverse tree */
54 rn_walktree_from_t *rnh_walktree_from; /* traverse tree below a */
55 rnh_preadd_entry_f_t *rnh_preadd; /* hook to alter record prior to insertion */
56 rt_gen_t rnh_gen; /* generation counter */
57 int rnh_multipath; /* multipath capable ? */
58 struct radix_node rnh_nodes[3]; /* empty tree for common case */
59 struct rmlock rib_lock; /* config/data path lock */
60 struct radix_mask_head rmhead; /* masks radix head */
61 struct vnet *rib_vnet; /* vnet pointer */
62 int rib_family; /* AF of the rtable */
63 u_int rib_fibnum; /* fib number */
64 struct callout expire_callout; /* Callout for expiring dynamic routes */
65 time_t next_expire; /* Next expire run ts */
66 struct nh_control *nh_control; /* nexthop subsystem data */
67 CK_STAILQ_HEAD(, rib_subscription) rnh_subscribers;/* notification subscribers */
70 #define RIB_RLOCK_TRACKER struct rm_priotracker _rib_tracker
71 #define RIB_LOCK_INIT(rh) rm_init(&(rh)->rib_lock, "rib head lock")
72 #define RIB_LOCK_DESTROY(rh) rm_destroy(&(rh)->rib_lock)
73 #define RIB_RLOCK(rh) rm_rlock(&(rh)->rib_lock, &_rib_tracker)
74 #define RIB_RUNLOCK(rh) rm_runlock(&(rh)->rib_lock, &_rib_tracker)
75 #define RIB_WLOCK(rh) rm_wlock(&(rh)->rib_lock)
76 #define RIB_WUNLOCK(rh) rm_wunlock(&(rh)->rib_lock)
77 #define RIB_LOCK_ASSERT(rh) rm_assert(&(rh)->rib_lock, RA_LOCKED)
78 #define RIB_WLOCK_ASSERT(rh) rm_assert(&(rh)->rib_lock, RA_WLOCKED)
81 #define RIB_MAX_RETRIES 3
82 #define RT_MAXFIBS UINT16_MAX
84 /* Macro for verifying fields in af-specific 'struct route' structures */
85 #define CHK_STRUCT_FIELD_GENERIC(_s1, _f1, _s2, _f2) \
86 _Static_assert(sizeof(((_s1 *)0)->_f1) == sizeof(((_s2 *)0)->_f2), \
87 "Fields " #_f1 " and " #_f2 " size differs"); \
88 _Static_assert(__offsetof(_s1, _f1) == __offsetof(_s2, _f2), \
89 "Fields " #_f1 " and " #_f2 " offset differs");
91 #define _CHK_ROUTE_FIELD(_route_new, _field) \
92 CHK_STRUCT_FIELD_GENERIC(struct route, _field, _route_new, _field)
94 #define CHK_STRUCT_ROUTE_FIELDS(_route_new) \
95 _CHK_ROUTE_FIELD(_route_new, ro_nh) \
96 _CHK_ROUTE_FIELD(_route_new, ro_lle) \
97 _CHK_ROUTE_FIELD(_route_new, ro_prepend)\
98 _CHK_ROUTE_FIELD(_route_new, ro_plen) \
99 _CHK_ROUTE_FIELD(_route_new, ro_flags) \
100 _CHK_ROUTE_FIELD(_route_new, ro_mtu) \
101 _CHK_ROUTE_FIELD(_route_new, spare)
103 #define CHK_STRUCT_ROUTE_COMPAT(_ro_new, _dst_new) \
104 CHK_STRUCT_ROUTE_FIELDS(_ro_new); \
105 _Static_assert(__offsetof(struct route, ro_dst) == __offsetof(_ro_new, _dst_new),\
106 "ro_dst and " #_dst_new " are at different offset")
108 struct rib_head *rt_tables_get_rnh(uint32_t table, sa_family_t family);
109 void rt_mpath_init_rnh(struct rib_head *rnh);
110 int rt_getifa_fib(struct rt_addrinfo *info, u_int fibnum);
111 void rt_setmetrics(const struct rt_addrinfo *info, struct rtentry *rt);
113 struct radix_node *rt_mpath_unlink(struct rib_head *rnh,
114 struct rt_addrinfo *info, struct rtentry *rto, int *perror);
118 VNET_PCPUSTAT_DECLARE(struct rtstat, rtstat);
119 #define RTSTAT_ADD(name, val) \
120 VNET_PCPUSTAT_ADD(struct rtstat, rtstat, name, (val))
121 #define RTSTAT_INC(name) RTSTAT_ADD(name, 1)
125 * Convert a 'struct radix_node *' to a 'struct rtentry *'.
126 * The operation can be done safely (in this code) because a
127 * 'struct rtentry' starts with two 'struct radix_node''s, the first
128 * one representing leaf nodes in the routing tree, which is
129 * what the code in radix.c passes us as a 'struct radix_node'.
131 * But because there are a lot of assumptions in this conversion,
132 * do not cast explicitly, but always use the macro below.
134 #define RNTORT(p) ((struct rtentry *)(p))
137 struct radix_node rt_nodes[2]; /* tree glue, and other values */
139 * XXX struct rtentry must begin with a struct radix_node (or two!)
140 * because the code does some casts of a 'struct radix_node *'
141 * to a 'struct rtentry *'
143 #define rt_key(r) (*((struct sockaddr **)(&(r)->rt_nodes->rn_key)))
144 #define rt_mask(r) (*((struct sockaddr **)(&(r)->rt_nodes->rn_mask)))
145 #define rt_key_const(r) (*((const struct sockaddr * const *)(&(r)->rt_nodes->rn_key)))
146 #define rt_mask_const(r) (*((const struct sockaddr * const *)(&(r)->rt_nodes->rn_mask)))
149 * 2 radix_node structurs above consists of 2x6 pointers, leaving
150 * 4 pointers (32 bytes) of the second cache line on amd64.
153 struct nhop_object *rt_nhop; /* nexthop data */
156 * Destination address storage.
157 * sizeof(struct sockaddr_in6) == 28, however
158 * the dataplane-relevant part (e.g. address) lies
159 * at offset 8..24, making the address not crossing
160 * cacheline boundary.
162 struct sockaddr_in rt_dst4;
163 struct sockaddr_in6 rt_dst6;
164 struct sockaddr rt_dst;
168 int rte_flags; /* up/down?, host/net */
169 u_long rt_weight; /* absolute weight */
170 u_long rt_expire; /* lifetime for route, e.g. redirect */
171 struct rtentry *rt_chain; /* pointer to next rtentry to delete */
172 struct epoch_context rt_epoch_ctx; /* net epoch tracker */
176 * With the split between the routing entry and the nexthop,
177 * rt_flags has to be split between these 2 entries. As rtentry
178 * mostly contains prefix data and is thought to be generic enough
179 * so one can transparently change the nexthop pointer w/o requiring
180 * any other rtentry changes, most of rt_flags shifts to the particular nexthop.
183 * RTF_UP: rtentry, as an indication that it is linked.
184 * RTF_HOST: rtentry, nhop. The latter indication is needed for the datapath
185 * RTF_DYNAMIC: nhop, to make rtentry generic.
186 * RTF_MODIFIED: nhop, to make rtentry generic. (legacy)
187 * -- "native" path (nhop) properties:
188 * RTF_GATEWAY, RTF_STATIC, RTF_PROTO1, RTF_PROTO2, RTF_PROTO3, RTF_FIXEDMTU,
189 * RTF_PINNED, RTF_REJECT, RTF_BLACKHOLE, RTF_BROADCAST
192 /* Nexthop rt flags mask */
193 #define NHOP_RT_FLAG_MASK (RTF_GATEWAY | RTF_HOST | RTF_REJECT | RTF_DYNAMIC | \
194 RTF_MODIFIED | RTF_STATIC | RTF_BLACKHOLE | RTF_PROTO1 | RTF_PROTO2 | \
195 RTF_PROTO3 | RTF_FIXEDMTU | RTF_PINNED | RTF_BROADCAST)
197 /* rtentry rt flag mask */
198 #define RTE_RT_FLAG_MASK (RTF_UP | RTF_HOST)
200 /* Nexthop selection */
201 #define _NH2MP(_nh) ((struct nhgrp_object *)(_nh))
202 #define _SELECT_NHOP(_nh, _flowid) \
203 (_NH2MP(_nh))->nhops[(_flowid) % (_NH2MP(_nh))->mp_size]
204 #define _RT_SELECT_NHOP(_nh, _flowid) \
205 ((!NH_IS_MULTIPATH(_nh)) ? (_nh) : _SELECT_NHOP(_nh, _flowid))
206 #define RT_SELECT_NHOP(_rt, _flowid) _RT_SELECT_NHOP((_rt)->rt_nhop, _flowid)
208 /* rte<>nhop translation */
209 static inline uint16_t
210 fib_rte_to_nh_flags(int rt_flags)
214 res = (rt_flags & RTF_REJECT) ? NHF_REJECT : 0;
215 res |= (rt_flags & RTF_HOST) ? NHF_HOST : 0;
216 res |= (rt_flags & RTF_BLACKHOLE) ? NHF_BLACKHOLE : 0;
217 res |= (rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) ? NHF_REDIRECT : 0;
218 res |= (rt_flags & RTF_BROADCAST) ? NHF_BROADCAST : 0;
219 res |= (rt_flags & RTF_GATEWAY) ? NHF_GATEWAY : 0;
224 void tmproutes_update(struct rib_head *rnh, struct rtentry *rt);
225 void tmproutes_init(struct rib_head *rh);
226 void tmproutes_destroy(struct rib_head *rh);
229 void vnet_rtzone_init(void);
230 void vnet_rtzone_destroy(void);