1 /* $OpenBSD: ip_mroute.h,v 1.14 2006/04/25 15:49:35 claudio Exp $ */
2 /* $NetBSD: ip_mroute.h,v 1.23 2004/04/21 17:49:46 itojun Exp $ */
3
4 #ifndef _NETINET_IP_MROUTE_H_
5 #define _NETINET_IP_MROUTE_H_
6
7 /*
8 * Definitions for IP multicast forwarding.
9 *
10 * Written by David Waitzman, BBN Labs, August 1988.
11 * Modified by Steve Deering, Stanford, February 1989.
12 * Modified by Ajit Thyagarajan, PARC, August 1993.
13 * Modified by Ajit Thyagarajan, PARC, August 1994.
14 * Modified by Ahmed Helmy, SGI, June 1996.
15 * Modified by Pavlin Radoslavov, ICSI, October 2002.
16 *
17 * MROUTING Revision: 1.2
18 * and PIM-SMv2 and PIM-DM support, advanced API support,
19 * bandwidth metering and signaling.
20 */
21
22 #include <sys/queue.h>
23 #include <sys/timeout.h>
24
25 /*
26 * Multicast Routing set/getsockopt commands.
27 */
28 #define MRT_INIT 100 /* initialize forwarder */
29 #define MRT_DONE 101 /* shut down forwarder */
30 #define MRT_ADD_VIF 102 /* create virtual interface */
31 #define MRT_DEL_VIF 103 /* delete virtual interface */
32 #define MRT_ADD_MFC 104 /* insert forwarding cache entry */
33 #define MRT_DEL_MFC 105 /* delete forwarding cache entry */
34 #define MRT_VERSION 106 /* get kernel version number */
35 #define MRT_ASSERT 107 /* enable assert processing */
36 #define MRT_PIM MRT_ASSERT /* enable PIM processing */
37 #define MRT_API_SUPPORT 109 /* supported MRT API */
38 #define MRT_API_CONFIG 110 /* config MRT API */
39 #define MRT_ADD_BW_UPCALL 111 /* create bandwidth monitor */
40 #define MRT_DEL_BW_UPCALL 112 /* delete bandwidth monitor */
41
42
43 /*
44 * Types and macros for handling bitmaps with one bit per virtual interface.
45 */
46 #define MAXVIFS 32
47 typedef u_int32_t vifbitmap_t;
48 typedef u_int16_t vifi_t; /* type of a vif index */
49
50 #define VIFM_SET(n, m) ((m) |= (1 << (n)))
51 #define VIFM_CLR(n, m) ((m) &= ~(1 << (n)))
52 #define VIFM_ISSET(n, m) ((m) & (1 << (n)))
53 #define VIFM_SETALL(m) ((m) = 0xffffffff)
54 #define VIFM_CLRALL(m) ((m) = 0x00000000)
55 #define VIFM_COPY(mfrom, mto) ((mto) = (mfrom))
56 #define VIFM_SAME(m1, m2) ((m1) == (m2))
57
58 #define VIFF_TUNNEL 0x1 /* vif represents a tunnel end-point */
59 #define VIFF_SRCRT 0x2 /* tunnel uses IP src routing */
60 #define VIFF_REGISTER 0x4 /* used for PIM Register encap/decap */
61
62 /*
63 * Argument structure for MRT_ADD_VIF.
64 * (MRT_DEL_VIF takes a single vifi_t argument.)
65 */
66 struct vifctl {
67 vifi_t vifc_vifi; /* the index of the vif to be added */
68 u_int8_t vifc_flags; /* VIFF_ flags defined above */
69 u_int8_t vifc_threshold; /* min ttl required to forward on vif */
70 u_int32_t vifc_rate_limit; /* max rate */
71 struct in_addr vifc_lcl_addr;/* local interface address */
72 struct in_addr vifc_rmt_addr;/* remote address (tunnels only) */
73 };
74
75 /*
76 * Argument structure for MRT_ADD_MFC and MRT_DEL_MFC.
77 * XXX if you change this, make sure to change struct mfcctl2 as well.
78 */
79 struct mfcctl {
80 struct in_addr mfcc_origin; /* ip origin of mcasts */
81 struct in_addr mfcc_mcastgrp; /* multicast group associated */
82 vifi_t mfcc_parent; /* incoming vif */
83 u_int8_t mfcc_ttls[MAXVIFS]; /* forwarding ttls on vifs */
84 };
85
86 /*
87 * The new argument structure for MRT_ADD_MFC and MRT_DEL_MFC overlays
88 * and extends the old struct mfcctl.
89 */
90 struct mfcctl2 {
91 /* the mfcctl fields */
92 struct in_addr mfcc_origin; /* ip origin of mcasts */
93 struct in_addr mfcc_mcastgrp; /* multicast group associated*/
94 vifi_t mfcc_parent; /* incoming vif */
95 u_int8_t mfcc_ttls[MAXVIFS]; /* forwarding ttls on vifs */
96
97 /* extension fields */
98 u_int8_t mfcc_flags[MAXVIFS]; /* the MRT_MFC_FLAGS_* flags */
99 struct in_addr mfcc_rp; /* the RP address */
100 };
101 /*
102 * The advanced-API flags.
103 *
104 * The MRT_MFC_FLAGS_XXX API flags are also used as flags
105 * for the mfcc_flags field.
106 */
107 #define MRT_MFC_FLAGS_DISABLE_WRONGVIF (1 << 0) /* disable WRONGVIF signals */
108 #define MRT_MFC_FLAGS_BORDER_VIF (1 << 1) /* border vif */
109 #define MRT_MFC_RP (1 << 8) /* enable RP address */
110 #define MRT_MFC_BW_UPCALL (1 << 9) /* enable bw upcalls */
111 #define MRT_MFC_FLAGS_ALL (MRT_MFC_FLAGS_DISABLE_WRONGVIF | \
112 MRT_MFC_FLAGS_BORDER_VIF)
113 #define MRT_API_FLAGS_ALL (MRT_MFC_FLAGS_ALL | \
114 MRT_MFC_RP | \
115 MRT_MFC_BW_UPCALL)
116
117 /*
118 * Structure for installing or delivering an upcall if the
119 * measured bandwidth is above or below a threshold.
120 *
121 * User programs (e.g. daemons) may have a need to know when the
122 * bandwidth used by some data flow is above or below some threshold.
123 * This interface allows the userland to specify the threshold (in
124 * bytes and/or packets) and the measurement interval. Flows are
125 * all packet with the same source and destination IP address.
126 * At the moment the code is only used for multicast destinations
127 * but there is nothing that prevents its use for unicast.
128 *
129 * The measurement interval cannot be shorter than some Tmin (currently, 3s).
130 * The threshold is set in packets and/or bytes per_interval.
131 *
132 * Measurement works as follows:
133 *
134 * For >= measurements:
135 * The first packet marks the start of a measurement interval.
136 * During an interval we count packets and bytes, and when we
137 * pass the threshold we deliver an upcall and we are done.
138 * The first packet after the end of the interval resets the
139 * count and restarts the measurement.
140 *
141 * For <= measurement:
142 * We start a timer to fire at the end of the interval, and
143 * then for each incoming packet we count packets and bytes.
144 * When the timer fires, we compare the value with the threshold,
145 * schedule an upcall if we are below, and restart the measurement
146 * (reschedule timer and zero counters).
147 */
148
149 struct bw_data {
150 struct timeval b_time;
151 u_int64_t b_packets;
152 u_int64_t b_bytes;
153 };
154
155 struct bw_upcall {
156 struct in_addr bu_src; /* source address */
157 struct in_addr bu_dst; /* destination address */
158 u_int32_t bu_flags; /* misc flags (see below) */
159 #define BW_UPCALL_UNIT_PACKETS (1 << 0) /* threshold (in packets) */
160 #define BW_UPCALL_UNIT_BYTES (1 << 1) /* threshold (in bytes) */
161 #define BW_UPCALL_GEQ (1 << 2) /* upcall if bw >= threshold */
162 #define BW_UPCALL_LEQ (1 << 3) /* upcall if bw <= threshold */
163 #define BW_UPCALL_DELETE_ALL (1 << 4) /* delete all upcalls for s,d*/
164 struct bw_data bu_threshold; /* the bw threshold */
165 struct bw_data bu_measured; /* the measured bw */
166 };
167
168 /* max. number of upcalls to deliver together */
169 #define BW_UPCALLS_MAX 128
170 /* min. threshold time interval for bandwidth measurement */
171 #define BW_UPCALL_THRESHOLD_INTERVAL_MIN_SEC 3
172 #define BW_UPCALL_THRESHOLD_INTERVAL_MIN_USEC 0
173
174 /*
175 * Argument structure used by mrouted to get src-grp pkt counts.
176 */
177 struct sioc_sg_req {
178 struct in_addr src;
179 struct in_addr grp;
180 u_long pktcnt;
181 u_long bytecnt;
182 u_long wrong_if;
183 };
184
185 /*
186 * Argument structure used by mrouted to get vif pkt counts.
187 */
188 struct sioc_vif_req {
189 vifi_t vifi; /* vif number */
190 u_long icount; /* input packet count on vif */
191 u_long ocount; /* output packet count on vif */
192 u_long ibytes; /* input byte count on vif */
193 u_long obytes; /* output byte count on vif */
194 };
195
196
197 /*
198 * The kernel's multicast routing statistics.
199 */
200 struct mrtstat {
201 u_long mrts_mfc_lookups; /* # forw. cache hash table hits */
202 u_long mrts_mfc_misses; /* # forw. cache hash table misses */
203 u_long mrts_upcalls; /* # calls to mrouted */
204 u_long mrts_no_route; /* no route for packet's origin */
205 u_long mrts_bad_tunnel; /* malformed tunnel options */
206 u_long mrts_cant_tunnel; /* no room for tunnel options */
207 u_long mrts_wrong_if; /* arrived on wrong interface */
208 u_long mrts_upq_ovflw; /* upcall Q overflow */
209 u_long mrts_cache_cleanups; /* # entries with no upcalls */
210 u_long mrts_drop_sel; /* pkts dropped selectively */
211 u_long mrts_q_overflow; /* pkts dropped - Q overflow */
212 u_long mrts_pkt2large; /* pkts dropped - size > BKT SIZE */
213 u_long mrts_upq_sockfull; /* upcalls dropped - socket full */
214 };
215
216
217 #ifdef _KERNEL
218
219 /*
220 * The kernel's virtual-interface structure.
221 */
222 struct vif {
223 struct mbuf *tbf_q, **tbf_t; /* packet queue */
224 struct timeval tbf_last_pkt_t; /* arr. time of last pkt */
225 u_int32_t tbf_n_tok; /* no of tokens in bucket */
226 u_int32_t tbf_q_len; /* length of queue at this vif */
227 u_int32_t tbf_max_q_len; /* max. queue length */
228
229 u_int8_t v_flags; /* VIFF_ flags defined above */
230 u_int8_t v_threshold; /* min ttl required to forward on vif */
231 u_int32_t v_rate_limit; /* max rate */
232 struct in_addr v_lcl_addr; /* local interface address */
233 struct in_addr v_rmt_addr; /* remote address (tunnels only) */
234 struct ifnet *v_ifp; /* pointer to interface */
235 u_long v_pkt_in; /* # pkts in on interface */
236 u_long v_pkt_out; /* # pkts out on interface */
237 u_long v_bytes_in; /* # bytes in on interface */
238 u_long v_bytes_out; /* # bytes out on interface */
239 struct route v_route; /* cached route if this is a tunnel */
240 struct timeout v_repq_ch; /* for tbf_reprocess_q() */
241 #ifdef RSVP_ISI
242 int v_rsvp_on; /* # RSVP listening on this vif */
243 struct socket *v_rsvpd; /* # RSVPD daemon */
244 #endif /* RSVP_ISI */
245 };
246
247 /*
248 * The kernel's multicast forwarding cache entry structure.
249 * (A field for the type of service (mfc_tos) is to be added
250 * at a future point.)
251 */
252 struct mfc {
253 LIST_ENTRY(mfc) mfc_hash;
254 struct in_addr mfc_origin; /* ip origin of mcasts */
255 struct in_addr mfc_mcastgrp; /* multicast group associated */
256 vifi_t mfc_parent; /* incoming vif */
257 u_int8_t mfc_ttls[MAXVIFS]; /* forwarding ttls on vifs */
258 u_long mfc_pkt_cnt; /* pkt count for src-grp */
259 u_long mfc_byte_cnt; /* byte count for src-grp */
260 u_long mfc_wrong_if; /* wrong if for src-grp */
261 int mfc_expire; /* time to clean entry up */
262 struct timeval mfc_last_assert; /* last time I sent an assert */
263 struct rtdetq *mfc_stall; /* pkts waiting for route */
264 u_int8_t mfc_flags[MAXVIFS]; /* the MRT_MFC_FLAGS_* flags */
265 struct in_addr mfc_rp; /* the RP address */
266 struct bw_meter *mfc_bw_meter; /* list of bandwidth meters */
267 };
268
269 /*
270 * Structure used to communicate from kernel to multicast router.
271 * (Note the convenient similarity to an IP packet.)
272 */
273 struct igmpmsg {
274 u_int32_t unused1;
275 u_int32_t unused2;
276 u_int8_t im_msgtype; /* what type of message */
277 #define IGMPMSG_NOCACHE 1 /* no MFC in the kernel */
278 #define IGMPMSG_WRONGVIF 2 /* packet came from wrong interface */
279 #define IGMPMSG_WHOLEPKT 3 /* PIM pkt for user level encap. */
280 #define IGMPMSG_BW_UPCALL 4 /* BW monitoring upcall */
281 u_int8_t im_mbz; /* must be zero */
282 u_int8_t im_vif; /* vif rec'd on */
283 u_int8_t unused3;
284 struct in_addr im_src, im_dst;
285 };
286
287 /*
288 * Argument structure used for pkt info. while upcall is made.
289 */
290 struct rtdetq {
291 struct mbuf *m; /* a copy of the packet */
292 struct ifnet *ifp; /* interface pkt came in on */
293 #ifdef UPCALL_TIMING
294 struct timeval t; /* timestamp */
295 #endif /* UPCALL_TIMING */
296 struct rtdetq *next;
297 };
298
299 #define MFCTBLSIZ 256
300 #define MAX_UPQ 4 /* max. no of pkts in upcall Q */
301
302 /*
303 * Token bucket filter code
304 */
305 #define MAX_BKT_SIZE 10000 /* 10K bytes size */
306 #define MAXQSIZE 10 /* max. no of pkts in token queue */
307
308 /*
309 * Structure for measuring the bandwidth and sending an upcall if the
310 * measured bandwidth is above or below a threshold.
311 */
312 struct bw_meter {
313 struct bw_meter *bm_mfc_next; /* next bw meter (same mfc) */
314 struct bw_meter *bm_time_next; /* next bw meter (same time) */
315 uint32_t bm_time_hash; /* the time hash value */
316 struct mfc *bm_mfc; /* the corresponding mfc */
317 uint32_t bm_flags; /* misc flags (see below) */
318 #define BW_METER_UNIT_PACKETS (1 << 0) /* threshold (in packets) */
319 #define BW_METER_UNIT_BYTES (1 << 1) /* threshold (in bytes) */
320 #define BW_METER_GEQ (1 << 2) /* upcall if bw >= threshold */
321 #define BW_METER_LEQ (1 << 3) /* upcall if bw <= threshold */
322 #define BW_METER_USER_FLAGS (BW_METER_UNIT_PACKETS | \
323 BW_METER_UNIT_BYTES | \
324 BW_METER_GEQ | \
325 BW_METER_LEQ)
326
327 #define BW_METER_UPCALL_DELIVERED (1 << 24) /* upcall was delivered */
328
329 struct bw_data bm_threshold; /* the upcall threshold */
330 struct bw_data bm_measured; /* the measured bw */
331 struct timeval bm_start_time; /* abs. time */
332 };
333
334 int ip_mrouter_set(struct socket *, int, struct mbuf **);
335 int ip_mrouter_get(struct socket *, int, struct mbuf **);
336 int mrt_ioctl(struct socket *, u_long, caddr_t);
337 int ip_mrouter_done(void);
338 void ip_mrouter_detach(struct ifnet *);
339 void reset_vif(struct vif *);
340 void vif_delete(struct ifnet *);
341 #ifdef RSVP_ISI
342 int ip_mforward(struct mbuf *, struct ifnet *, struct ip_moptions *);
343 int legal_vif_num(int);
344 int ip_rsvp_vif_init(struct socket *, struct mbuf *);
345 int ip_rsvp_vif_done(struct socket *, struct mbuf *);
346 void ip_rsvp_force_done(struct socket *);
347 void rsvp_input(struct mbuf *, int, int);
348 #else
349 int ip_mforward(struct mbuf *, struct ifnet *);
350 #endif /* RSVP_ISI */
351
352 #endif /* _KERNEL */
353 #endif /* _NETINET_IP_MROUTE_H_ */