1/* $NetBSD: ip_mroute.h,v 1.31 2008/08/07 06:20:14 cegger Exp $ */
2
3#ifndef _NETINET_IP_MROUTE_H_
4#define _NETINET_IP_MROUTE_H_
5
6/*
7 * Definitions for IP multicast forwarding.
8 *
9 * Written by David Waitzman, BBN Labs, August 1988.
10 * Modified by Steve Deering, Stanford, February 1989.
11 * Modified by Ajit Thyagarajan, PARC, August 1993.
12 * Modified by Ajit Thyagarajan, PARC, August 1994.
13 * Modified by Ahmed Helmy, SGI, June 1996.
14 * Modified by Pavlin Radoslavov, ICSI, October 2002.
15 *
16 * MROUTING Revision: 1.2
17 * and PIM-SMv2 and PIM-DM support, advanced API support,
18 * bandwidth metering and signaling.
19 */
20
21#include <sys/queue.h>
22#include <sys/callout.h>
23
24#ifdef _KERNEL
25struct sockopt; /* from <sys/socketvar.h> */
26#endif
27
28/*
29 * Multicast Routing set/getsockopt commands.
30 */
31#define MRT_INIT 100 /* initialize forwarder */
32#define MRT_DONE 101 /* shut down forwarder */
33#define MRT_ADD_VIF 102 /* create virtual interface */
34#define MRT_DEL_VIF 103 /* delete virtual interface */
35#define MRT_ADD_MFC 104 /* insert forwarding cache entry */
36#define MRT_DEL_MFC 105 /* delete forwarding cache entry */
37#define MRT_VERSION 106 /* get kernel version number */
38#define MRT_ASSERT 107 /* enable assert processing */
39#define MRT_PIM MRT_ASSERT /* enable PIM processing */
40#define MRT_API_SUPPORT 109 /* supported MRT API */
41#define MRT_API_CONFIG 110 /* config MRT API */
42#define MRT_ADD_BW_UPCALL 111 /* create bandwidth monitor */
43#define MRT_DEL_BW_UPCALL 112 /* delete bandwidth monitor */
44
45
46/*
47 * Types and macros for handling bitmaps with one bit per virtual interface.
48 */
49#define MAXVIFS 32
50typedef u_int32_t vifbitmap_t;
51typedef u_int16_t vifi_t; /* type of a vif index */
52
53#define VIFM_SET(n, m) ((m) |= (1 << (n)))
54#define VIFM_CLR(n, m) ((m) &= ~(1 << (n)))
55#define VIFM_ISSET(n, m) ((m) & (1 << (n)))
56#define VIFM_SETALL(m) ((m) = 0xffffffff)
57#define VIFM_CLRALL(m) ((m) = 0x00000000)
58#define VIFM_COPY(mfrom, mto) ((mto) = (mfrom))
59#define VIFM_SAME(m1, m2) ((m1) == (m2))
60
61#define VIFF_TUNNEL 0x1 /* vif represents a tunnel end-point */
62#define VIFF_SRCRT 0x2 /* tunnel uses IP src routing */
63#define VIFF_REGISTER 0x4 /* used for PIM Register encap/decap */
64
65/*
66 * Argument structure for MRT_ADD_VIF.
67 * (MRT_DEL_VIF takes a single vifi_t argument.)
68 */
69struct vifctl {
70 vifi_t vifc_vifi; /* the index of the vif to be added */
71 u_int8_t vifc_flags; /* VIFF_ flags defined below */
72 u_int8_t vifc_threshold; /* min ttl required to forward on vif */
73 u_int32_t vifc_rate_limit; /* max rate */
74 struct in_addr vifc_lcl_addr;/* local interface address */
75 struct in_addr vifc_rmt_addr;/* remote address (tunnels only) */
76};
77
78/*
79 * Argument structure for MRT_ADD_MFC and MRT_DEL_MFC.
80 * XXX if you change this, make sure to change struct mfcctl2 as well.
81 */
82struct mfcctl {
83 struct in_addr mfcc_origin; /* ip origin of mcasts */
84 struct in_addr mfcc_mcastgrp; /* multicast group associated */
85 vifi_t mfcc_parent; /* incoming vif */
86 u_int8_t mfcc_ttls[MAXVIFS]; /* forwarding ttls on vifs */
87};
88
89/*
90 * The new argument structure for MRT_ADD_MFC and MRT_DEL_MFC overlays
91 * and extends the old struct mfcctl.
92 */
93struct mfcctl2 {
94 /* the mfcctl fields */
95 struct in_addr mfcc_origin; /* ip origin of mcasts */
96 struct in_addr mfcc_mcastgrp; /* multicast group associated*/
97 vifi_t mfcc_parent; /* incoming vif */
98 u_int8_t mfcc_ttls[MAXVIFS]; /* forwarding ttls on vifs */
99
100 /* extension fields */
101 u_int8_t mfcc_flags[MAXVIFS]; /* the MRT_MFC_FLAGS_* flags */
102 struct in_addr mfcc_rp; /* the RP address */
103};
104/*
105 * The advanced-API flags.
106 *
107 * The MRT_MFC_FLAGS_XXX API flags are also used as flags
108 * for the mfcc_flags field.
109 */
110#define MRT_MFC_FLAGS_DISABLE_WRONGVIF (1 << 0) /* disable WRONGVIF signals */
111#define MRT_MFC_FLAGS_BORDER_VIF (1 << 1) /* border vif */
112#define MRT_MFC_RP (1 << 8) /* enable RP address */
113#define MRT_MFC_BW_UPCALL (1 << 9) /* enable bw upcalls */
114#define MRT_MFC_FLAGS_ALL (MRT_MFC_FLAGS_DISABLE_WRONGVIF | \
115 MRT_MFC_FLAGS_BORDER_VIF)
116#define MRT_API_FLAGS_ALL (MRT_MFC_FLAGS_ALL | \
117 MRT_MFC_RP | \
118 MRT_MFC_BW_UPCALL)
119
120/*
121 * Structure for installing or delivering an upcall if the
122 * measured bandwidth is above or below a threshold.
123 *
124 * User programs (e.g. daemons) may have a need to know when the
125 * bandwidth used by some data flow is above or below some threshold.
126 * This interface allows the userland to specify the threshold (in
127 * bytes and/or packets) and the measurement interval. Flows are
128 * all packet with the same source and destination IP address.
129 * At the moment the code is only used for multicast destinations
130 * but there is nothing that prevents its use for unicast.
131 *
132 * The measurement interval cannot be shorter than some Tmin (currently, 3s).
133 * The threshold is set in packets and/or bytes per_interval.
134 *
135 * Measurement works as follows:
136 *
137 * For >= measurements:
138 * The first packet marks the start of a measurement interval.
139 * During an interval we count packets and bytes, and when we
140 * pass the threshold we deliver an upcall and we are done.
141 * The first packet after the end of the interval resets the
142 * count and restarts the measurement.
143 *
144 * For <= measurement:
145 * We start a timer to fire at the end of the interval, and
146 * then for each incoming packet we count packets and bytes.
147 * When the timer fires, we compare the value with the threshold,
148 * schedule an upcall if we are below, and restart the measurement
149 * (reschedule timer and zero counters).
150 */
151
152struct bw_data {
153 struct timeval b_time;
154 u_int64_t b_packets;
155 u_int64_t b_bytes;
156};
157
158struct bw_upcall {
159 struct in_addr bu_src; /* source address */
160 struct in_addr bu_dst; /* destination address */
161 u_int32_t bu_flags; /* misc flags (see below) */
162#define BW_UPCALL_UNIT_PACKETS (1 << 0) /* threshold (in packets) */
163#define BW_UPCALL_UNIT_BYTES (1 << 1) /* threshold (in bytes) */
164#define BW_UPCALL_GEQ (1 << 2) /* upcall if bw >= threshold */
165#define BW_UPCALL_LEQ (1 << 3) /* upcall if bw <= threshold */
166#define BW_UPCALL_DELETE_ALL (1 << 4) /* delete all upcalls for s,d*/
167 struct bw_data bu_threshold; /* the bw threshold */
168 struct bw_data bu_measured; /* the measured bw */
169};
170
171/* max. number of upcalls to deliver together */
172#define BW_UPCALLS_MAX 128
173/* min. threshold time interval for bandwidth measurement */
174#define BW_UPCALL_THRESHOLD_INTERVAL_MIN_SEC 3
175#define BW_UPCALL_THRESHOLD_INTERVAL_MIN_USEC 0
176
177/*
178 * Argument structure used by mrouted to get src-grp pkt counts.
179 */
180struct sioc_sg_req {
181 struct in_addr src;
182 struct in_addr grp;
183 u_long pktcnt;
184 u_long bytecnt;
185 u_long wrong_if;
186};
187
188/*
189 * Argument structure used by mrouted to get vif pkt counts.
190 */
191struct sioc_vif_req {
192 vifi_t vifi; /* vif number */
193 u_long icount; /* input packet count on vif */
194 u_long ocount; /* output packet count on vif */
195 u_long ibytes; /* input byte count on vif */
196 u_long obytes; /* output byte count on vif */
197};
198
199
200/*
201 * The kernel's multicast routing statistics.
202 */
203struct mrtstat {
204 u_long mrts_mfc_lookups; /* # forw. cache hash table hits */
205 u_long mrts_mfc_misses; /* # forw. cache hash table misses */
206 u_long mrts_upcalls; /* # calls to mrouted */
207 u_long mrts_no_route; /* no route for packet's origin */
208 u_long mrts_bad_tunnel; /* malformed tunnel options */
209 u_long mrts_cant_tunnel; /* no room for tunnel options */
210 u_long mrts_wrong_if; /* arrived on wrong interface */
211 u_long mrts_upq_ovflw; /* upcall Q overflow */
212 u_long mrts_cache_cleanups; /* # entries with no upcalls */
213 u_long mrts_drop_sel; /* pkts dropped selectively */
214 u_long mrts_q_overflow; /* pkts dropped - Q overflow */
215 u_long mrts_pkt2large; /* pkts dropped - size > BKT SIZE */
216 u_long mrts_upq_sockfull; /* upcalls dropped - socket full */
217};
218
219
220#ifdef _KERNEL
221
222/*
223 * The kernel's virtual-interface structure.
224 */
225struct encaptab;
226struct vif {
227 struct mbuf *tbf_q, **tbf_t; /* packet queue */
228 struct timeval tbf_last_pkt_t; /* arr. time of last pkt */
229 u_int32_t tbf_n_tok; /* no of tokens in bucket */
230 u_int32_t tbf_q_len; /* length of queue at this vif */
231 u_int32_t tbf_max_q_len; /* max. queue length */
232
233 u_int8_t v_flags; /* VIFF_ flags defined above */
234 u_int8_t v_threshold; /* min ttl required to forward on vif */
235 u_int32_t v_rate_limit; /* max rate */
236 struct in_addr v_lcl_addr; /* local interface address */
237 struct in_addr v_rmt_addr; /* remote address (tunnels only) */
238 struct ifnet *v_ifp; /* pointer to interface */
239 u_long v_pkt_in; /* # pkts in on interface */
240 u_long v_pkt_out; /* # pkts out on interface */
241 u_long v_bytes_in; /* # bytes in on interface */
242 u_long v_bytes_out; /* # bytes out on interface */
243 struct route v_route; /* cached route if this is a tunnel */
244 callout_t v_repq_ch; /* for tbf_reprocess_q() */
245#ifdef RSVP_ISI
246 int v_rsvp_on; /* # RSVP listening on this vif */
247 struct socket *v_rsvpd; /* # RSVPD daemon */
248#endif /* RSVP_ISI */
249 const struct encaptab *v_encap_cookie;
250};
251
252/*
253 * The kernel's multicast forwarding cache entry structure.
254 * (A field for the type of service (mfc_tos) is to be added
255 * at a future point.)
256 */
257struct mfc {
258 LIST_ENTRY(mfc) mfc_hash;
259 struct in_addr mfc_origin; /* ip origin of mcasts */
260 struct in_addr mfc_mcastgrp; /* multicast group associated */
261 vifi_t mfc_parent; /* incoming vif */
262 u_int8_t mfc_ttls[MAXVIFS]; /* forwarding ttls on vifs */
263 u_long mfc_pkt_cnt; /* pkt count for src-grp */
264 u_long mfc_byte_cnt; /* byte count for src-grp */
265 u_long mfc_wrong_if; /* wrong if for src-grp */
266 int mfc_expire; /* time to clean entry up */
267 struct timeval mfc_last_assert; /* last time I sent an assert */
268 struct rtdetq *mfc_stall; /* pkts waiting for route */
269 u_int8_t mfc_flags[MAXVIFS]; /* the MRT_MFC_FLAGS_* flags */
270 struct in_addr mfc_rp; /* the RP address */
271 struct bw_meter *mfc_bw_meter; /* list of bandwidth meters */
272};
273
274/*
275 * Structure used to communicate from kernel to multicast router.
276 * (Note the convenient similarity to an IP packet.)
277 */
278struct igmpmsg {
279 u_int32_t unused1;
280 u_int32_t unused2;
281 u_int8_t im_msgtype; /* what type of message */
282#define IGMPMSG_NOCACHE 1 /* no MFC in the kernel */
283#define IGMPMSG_WRONGVIF 2 /* packet came from wrong interface */
284#define IGMPMSG_WHOLEPKT 3 /* PIM pkt for user level encap. */
285#define IGMPMSG_BW_UPCALL 4 /* BW monitoring upcall */
286 u_int8_t im_mbz; /* must be zero */
287 u_int8_t im_vif; /* vif rec'd on */
288 u_int8_t unused3;
289 struct in_addr im_src, im_dst;
290} __packed;
291
292/*
293 * Argument structure used for pkt info. while upcall is made.
294 */
295struct rtdetq {
296 struct mbuf *m; /* a copy of the packet */
297 struct ifnet *ifp; /* interface pkt came in on */
298#ifdef UPCALL_TIMING
299 struct timeval t; /* timestamp */
300#endif /* UPCALL_TIMING */
301 struct rtdetq *next;
302};
303
304#define MFCTBLSIZ 256
305#define MAX_UPQ 4 /* max. no of pkts in upcall Q */
306
307/*
308 * Token bucket filter code
309 */
310#define MAX_BKT_SIZE 10000 /* 10K bytes size */
311#define MAXQSIZE 10 /* max. no of pkts in token queue */
312
313/*
314 * Structure for measuring the bandwidth and sending an upcall if the
315 * measured bandwidth is above or below a threshold.
316 */
317struct bw_meter {
318 struct bw_meter *bm_mfc_next; /* next bw meter (same mfc) */
319 struct bw_meter *bm_time_next; /* next bw meter (same time) */
320 uint32_t bm_time_hash; /* the time hash value */
321 struct mfc *bm_mfc; /* the corresponding mfc */
322 uint32_t bm_flags; /* misc flags (see below) */
323#define BW_METER_UNIT_PACKETS (1 << 0) /* threshold (in packets) */
324#define BW_METER_UNIT_BYTES (1 << 1) /* threshold (in bytes) */
325#define BW_METER_GEQ (1 << 2) /* upcall if bw >= threshold */
326#define BW_METER_LEQ (1 << 3) /* upcall if bw <= threshold */
327#define BW_METER_USER_FLAGS (BW_METER_UNIT_PACKETS | \
328 BW_METER_UNIT_BYTES | \
329 BW_METER_GEQ | \
330 BW_METER_LEQ)
331
332#define BW_METER_UPCALL_DELIVERED (1 << 24) /* upcall was delivered */
333
334 struct bw_data bm_threshold; /* the upcall threshold */
335 struct bw_data bm_measured; /* the measured bw */
336 struct timeval bm_start_time; /* abs. time */
337};
338
339int ip_mrouter_set(struct socket *, struct sockopt *);
340int ip_mrouter_get(struct socket *, struct sockopt *);
341int mrt_ioctl(struct socket *, u_long, void *);
342int ip_mrouter_done(void);
343void ip_mrouter_detach(struct ifnet *);
344void reset_vif(struct vif *);
345#ifdef RSVP_ISI
346int ip_mforward(struct mbuf *, struct ifnet *, struct ip_moptions *);
347int legal_vif_num(int);
348int ip_rsvp_vif_init(struct socket *, struct mbuf *);
349int ip_rsvp_vif_done(struct socket *, struct mbuf *);
350void ip_rsvp_force_done(struct socket *);
351void rsvp_input(struct mbuf *, int, int);
352#else
353int ip_mforward(struct mbuf *, struct ifnet *);
354#endif
355
356#endif /* _KERNEL */
357
358#endif /* !_NETINET_IP_MROUTE_H_ */
359