| 1 | /*	$NetBSD: ip_mroute.h,v 1.36 2025/07/28 21:25:00 kim 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 | /* |
| 25 | * Multicast Routing set/getsockopt commands. |
| 26 | */ |
| 27 | #define	MRT_INIT		100	/* initialize forwarder */ |
| 28 | #define	MRT_DONE		101	/* shut down forwarder */ |
| 29 | #define	MRT_ADD_VIF		102	/* create virtual interface */ |
| 30 | #define	MRT_DEL_VIF		103	/* delete virtual interface */ |
| 31 | #define	MRT_ADD_MFC		104	/* insert forwarding cache entry */ |
| 32 | #define	MRT_DEL_MFC		105	/* delete forwarding cache entry */ |
| 33 | #define	MRT_VERSION		106	/* get kernel version number */ |
| 34 | #define	MRT_ASSERT		107	/* enable assert processing */ |
| 35 | #define MRT_PIM			MRT_ASSERT /* enable PIM processing */ |
| 36 | #define MRT_API_SUPPORT		109	/* supported MRT API */ |
| 37 | #define MRT_API_CONFIG		110	/* config MRT API */ |
| 38 | #define MRT_ADD_BW_UPCALL	111	/* create bandwidth monitor */ |
| 39 | #define MRT_DEL_BW_UPCALL	112	/* delete bandwidth monitor */ |
| 40 | |
| 41 | /* |
| 42 | * Types and macros for handling bitmaps with one bit per virtual interface. |
| 43 | */ |
| 44 | #define	MAXVIFS 32 |
| 45 | typedef u_int32_t vifbitmap_t; |
| 46 | typedef u_int16_t vifi_t;		/* type of a vif index */ |
| 47 | |
| 48 | #define	VIFM_SET(n, m)			((m) |= (1 << (n))) |
| 49 | #define	VIFM_CLR(n, m)			((m) &= ~(1 << (n))) |
| 50 | #define	VIFM_ISSET(n, m)		((m) & (1 << (n))) |
| 51 | #define	VIFM_SETALL(m)			((m) = 0xffffffff) |
| 52 | #define	VIFM_CLRALL(m)			((m) = 0x00000000) |
| 53 | #define	VIFM_COPY(mfrom, mto)		((mto) = (mfrom)) |
| 54 | #define	VIFM_SAME(m1, m2)		((m1) == (m2)) |
| 55 | |
| 56 | #define	VIFF_TUNNEL	0x1		/* vif represents a tunnel end-point */ |
| 57 | #define	VIFF_SRCRT	0x2		/* tunnel uses IP src routing */ |
| 58 | #define VIFF_REGISTER	0x4		/* used for PIM Register encap/decap */ |
| 59 | |
| 60 | /* |
| 61 | * Argument structure for MRT_ADD_VIF. |
| 62 | * (MRT_DEL_VIF takes a single vifi_t argument.) |
| 63 | */ |
| 64 | struct vifctl { |
| 65 | 	vifi_t	 vifc_vifi;	 	/* the index of the vif to be added */ |
| 66 | 	u_int8_t vifc_flags; 	/* VIFF_ flags defined below */ |
| 67 | 	u_int8_t vifc_threshold; 	/* min ttl required to forward on vif */ |
| 68 | 	u_int32_t vifc_rate_limit;	/* max rate */ |
| 69 | 	struct	 in_addr vifc_lcl_addr;/* local interface address */ |
| 70 | 	struct	 in_addr vifc_rmt_addr;/* remote address (tunnels only) */ |
| 71 | }; |
| 72 | |
| 73 | /* |
| 74 | * Argument structure for MRT_ADD_MFC and MRT_DEL_MFC. |
| 75 | * XXX if you change this, make sure to change struct mfcctl2 as well. |
| 76 | */ |
| 77 | struct mfcctl { |
| 78 | 	struct	 in_addr mfcc_origin;	/* ip origin of mcasts */ |
| 79 | 	struct	 in_addr mfcc_mcastgrp;	/* multicast group associated */ |
| 80 | 	vifi_t	 mfcc_parent;		/* incoming vif */ |
| 81 | 	u_int8_t mfcc_ttls[MAXVIFS];	/* forwarding ttls on vifs */ |
| 82 | }; |
| 83 | |
| 84 | /* |
| 85 | * The new argument structure for MRT_ADD_MFC and MRT_DEL_MFC overlays |
| 86 | * and extends the old struct mfcctl. |
| 87 | */ |
| 88 | struct mfcctl2 { |
| 89 | 	/* the mfcctl fields */ |
| 90 | 	struct in_addr	mfcc_origin;		/* ip origin of mcasts	 */ |
| 91 | 	struct in_addr	mfcc_mcastgrp;		/* multicast group associated*/ |
| 92 | 	vifi_t		mfcc_parent;		/* incoming vif		 */ |
| 93 | 	u_int8_t	mfcc_ttls[MAXVIFS]; 	/* forwarding ttls on vifs */ |
| 94 | |
| 95 | 	/* extension fields */ |
| 96 | 	u_int8_t	mfcc_flags[MAXVIFS];	/* the MRT_MFC_FLAGS_* flags */ |
| 97 | 	struct in_addr	mfcc_rp;		/* the RP address */ |
| 98 | }; |
| 99 | |
| 100 | /* |
| 101 | * The advanced-API flags. |
| 102 | * |
| 103 | * The MRT_MFC_FLAGS_XXX API flags are also used as flags |
| 104 | * for the mfcc_flags field. |
| 105 | */ |
| 106 | #define	MRT_MFC_FLAGS_DISABLE_WRONGVIF	(1 << 0) /* disable WRONGVIF signals */ |
| 107 | #define	MRT_MFC_FLAGS_BORDER_VIF	(1 << 1) /* border vif		 */ |
| 108 | #define MRT_MFC_RP			(1 << 8) /* enable RP address	 */ |
| 109 | #define MRT_MFC_BW_UPCALL		(1 << 9) /* enable bw upcalls	 */ |
| 110 | #define MRT_MFC_FLAGS_ALL		(MRT_MFC_FLAGS_DISABLE_WRONGVIF | \ |
| 111 | 					 MRT_MFC_FLAGS_BORDER_VIF) |
| 112 | #define MRT_API_FLAGS_ALL		(MRT_MFC_FLAGS_ALL |		 \ |
| 113 | 					 MRT_MFC_RP |			 \ |
| 114 | 					 MRT_MFC_BW_UPCALL) |
| 115 | |
| 116 | /* |
| 117 | * Structure for installing or delivering an upcall if the |
| 118 | * measured bandwidth is above or below a threshold. |
| 119 | * |
| 120 | * User programs (e.g. daemons) may have a need to know when the |
| 121 | * bandwidth used by some data flow is above or below some threshold. |
| 122 | * This interface allows the userland to specify the threshold (in |
| 123 | * bytes and/or packets) and the measurement interval. Flows are |
| 124 | * all packet with the same source and destination IP address. |
| 125 | * At the moment the code is only used for multicast destinations |
| 126 | * but there is nothing that prevents its use for unicast. |
| 127 | * |
| 128 | * The measurement interval cannot be shorter than some Tmin (currently, 3s). |
| 129 | * The threshold is set in packets and/or bytes per_interval. |
| 130 | * |
| 131 | * Measurement works as follows: |
| 132 | * |
| 133 | * For >= measurements: |
| 134 | * The first packet marks the start of a measurement interval. |
| 135 | * During an interval we count packets and bytes, and when we |
| 136 | * pass the threshold we deliver an upcall and we are done. |
| 137 | * The first packet after the end of the interval resets the |
| 138 | * count and restarts the measurement. |
| 139 | * |
| 140 | * For <= measurement: |
| 141 | * We start a timer to fire at the end of the interval, and |
| 142 | * then for each incoming packet we count packets and bytes. |
| 143 | * When the timer fires, we compare the value with the threshold, |
| 144 | * schedule an upcall if we are below, and restart the measurement |
| 145 | * (reschedule timer and zero counters). |
| 146 | */ |
| 147 | |
| 148 | struct bw_data { |
| 149 | 	struct timeval	b_time; |
| 150 | 	u_int64_t	b_packets; |
| 151 | 	u_int64_t	b_bytes; |
| 152 | }; |
| 153 | |
| 154 | struct bw_upcall { |
| 155 | 	struct in_addr	bu_src;			/* source address */ |
| 156 | 	struct in_addr	bu_dst;			/* destination address */ |
| 157 | 	u_int32_t	bu_flags;		/* misc flags (see below) */ |
| 158 | #define BW_UPCALL_UNIT_PACKETS (1 << 0)	/* threshold (in packets) */ |
| 159 | #define BW_UPCALL_UNIT_BYTES (1 << 1)	/* threshold (in bytes) */ |
| 160 | #define BW_UPCALL_GEQ (1 << 2)	/* upcall if bw >= threshold */ |
| 161 | #define BW_UPCALL_LEQ (1 << 3)	/* upcall if bw <= threshold */ |
| 162 | #define BW_UPCALL_DELETE_ALL (1 << 4)	/* delete all upcalls for s,d*/ |
| 163 | 	struct bw_data	bu_threshold;		/* the bw threshold	 */ |
| 164 | 	struct bw_data	bu_measured;		/* the measured bw	 */ |
| 165 | }; |
| 166 | |
| 167 | /* max. number of upcalls to deliver together */ |
| 168 | #define BW_UPCALLS_MAX				128 |
| 169 | /* min. threshold time interval for bandwidth measurement */ |
| 170 | #define BW_UPCALL_THRESHOLD_INTERVAL_MIN_SEC	3 |
| 171 | #define BW_UPCALL_THRESHOLD_INTERVAL_MIN_USEC	0 |
| 172 | |
| 173 | /* |
| 174 | * Argument structure used by mrouted to get src-grp pkt counts. |
| 175 | */ |
| 176 | struct sioc_sg_req { |
| 177 | 	struct	in_addr src; |
| 178 | 	struct	in_addr grp; |
| 179 | 	u_long	pktcnt; |
| 180 | 	u_long	bytecnt; |
| 181 | 	u_long	wrong_if; |
| 182 | }; |
| 183 | |
| 184 | /* |
| 185 | * Argument structure used by mrouted to get vif pkt counts. |
| 186 | */ |
| 187 | struct sioc_vif_req { |
| 188 | 	vifi_t	vifi;			/* vif number */ |
| 189 | 	u_long	icount;			/* input packet count on vif */ |
| 190 | 	u_long	ocount;			/* output packet count on vif */ |
| 191 | 	u_long	ibytes;			/* input byte count on vif */ |
| 192 | 	u_long	obytes;			/* output byte count on vif */ |
| 193 | }; |
| 194 | |
| 195 | /* |
| 196 | * The kernel's multicast routing statistics. |
| 197 | */ |
| 198 | struct mrtstat { |
| 199 | 	u_long	mrts_mfc_lookups;	/* # forw. cache hash table hits */ |
| 200 | 	u_long	mrts_mfc_misses;	/* # forw. cache hash table misses */ |
| 201 | 	u_long	mrts_upcalls;		/* # calls to mrouted */ |
| 202 | 	u_long	mrts_no_route;		/* no route for packet's origin */ |
| 203 | 	u_long	mrts_bad_tunnel;	/* malformed tunnel options */ |
| 204 | 	u_long	mrts_cant_tunnel;	/* no room for tunnel options */ |
| 205 | 	u_long	mrts_wrong_if;		/* arrived on wrong interface */ |
| 206 | 	u_long	mrts_upq_ovflw;		/* upcall Q overflow */ |
| 207 | 	u_long	mrts_cache_cleanups;	/* # entries with no upcalls */ |
| 208 | 	u_long	mrts_drop_sel; 	/* pkts dropped selectively */ |
| 209 | 	u_long	mrts_q_overflow; 	/* pkts dropped - Q overflow */ |
| 210 | 	u_long	mrts_pkt2large; 	/* pkts dropped - size > BKT SIZE */ |
| 211 | 	u_long	mrts_upq_sockfull;	/* upcalls dropped - socket full */ |
| 212 | }; |
| 213 | |
| 214 | /* |
| 215 | * Structure used to communicate from kernel to multicast router. |
| 216 | * (Note the convenient similarity to an IP packet.) |
| 217 | */ |
| 218 | struct igmpmsg { |
| 219 | 	u_int32_t unused1; |
| 220 | 	u_int32_t unused2; |
| 221 | 	u_int8_t im_msgtype;		/* what type of message */ |
| 222 | #define IGMPMSG_NOCACHE		1	/* no MFC in the kernel		 */ |
| 223 | #define IGMPMSG_WRONGVIF	2	/* packet came from wrong interface */ |
| 224 | #define	IGMPMSG_WHOLEPKT	3	/* PIM pkt for user level encap. */ |
| 225 | #define	IGMPMSG_BW_UPCALL	4	/* BW monitoring upcall		 */ |
| 226 | 	u_int8_t im_mbz;		/* must be zero */ |
| 227 | 	u_int8_t im_vif;		/* vif rec'd on */ |
| 228 | 	u_int8_t unused3; |
| 229 | 	struct	 in_addr im_src, im_dst; |
| 230 | }; |
| 231 | #ifdef __CTASSERT |
| 232 | __CTASSERT(sizeof(struct igmpmsg) == 20); |
| 233 | #endif |
| 234 | |
| 235 | #ifdef _KERNEL |
| 236 | |
| 237 | /* |
| 238 | * The kernel's virtual-interface structure. |
| 239 | */ |
| 240 | struct encaptab; |
| 241 | struct vif { |
| 242 | 	struct	 mbuf *tbf_q, **tbf_t;	/* packet queue */ |
| 243 | 	struct	 timeval tbf_last_pkt_t; /* arr. time of last pkt */ |
| 244 | 	u_int32_t tbf_n_tok;		/* no of tokens in bucket */ |
| 245 | 	u_int32_t tbf_q_len;		/* length of queue at this vif */ |
| 246 | 	u_int32_t tbf_max_q_len;	/* max. queue length */ |
| 247 | |
| 248 | 	u_int8_t v_flags;		/* VIFF_ flags defined above */ |
| 249 | 	u_int8_t v_threshold;		/* min ttl required to forward on vif */ |
| 250 | 	u_int32_t v_rate_limit;		/* max rate */ |
| 251 | 	struct	 in_addr v_lcl_addr;	/* local interface address */ |
| 252 | 	struct	 in_addr v_rmt_addr;	/* remote address (tunnels only) */ |
| 253 | 	struct	 ifnet *v_ifp;		/* pointer to interface */ |
| 254 | 	u_long	 v_pkt_in;		/* # pkts in on interface */ |
| 255 | 	u_long	 v_pkt_out;		/* # pkts out on interface */ |
| 256 | 	u_long	 v_bytes_in;		/* # bytes in on interface */ |
| 257 | 	u_long	 v_bytes_out;		/* # bytes out on interface */ |
| 258 | 	struct	 route v_route;	/* cached route if this is a tunnel */ |
| 259 | 	callout_t v_repq_ch;		/* for tbf_reprocess_q() */ |
| 260 | 	const struct encaptab *v_encap_cookie; |
| 261 | }; |
| 262 | |
| 263 | /* |
| 264 | * The kernel's multicast forwarding cache entry structure. |
| 265 | * (A field for the type of service (mfc_tos) is to be added |
| 266 | * at a future point.) |
| 267 | */ |
| 268 | struct mfc { |
| 269 | 	LIST_ENTRY(mfc) mfc_hash; |
| 270 | 	struct	 in_addr mfc_origin;	 	/* ip origin of mcasts */ |
| 271 | 	struct	 in_addr mfc_mcastgrp; 	/* multicast group associated */ |
| 272 | 	vifi_t	 mfc_parent;			/* incoming vif */ |
| 273 | 	u_int8_t mfc_ttls[MAXVIFS]; 		/* forwarding ttls on vifs */ |
| 274 | 	u_long	 mfc_pkt_cnt;			/* pkt count for src-grp */ |
| 275 | 	u_long	 mfc_byte_cnt;			/* byte count for src-grp */ |
| 276 | 	u_long	 mfc_wrong_if;			/* wrong if for src-grp	*/ |
| 277 | 	int	 mfc_expire;			/* time to clean entry up */ |
| 278 | 	struct	 timeval mfc_last_assert;	/* last time I sent an assert */ |
| 279 | 	struct	 rtdetq *mfc_stall;		/* pkts waiting for route */ |
| 280 | 	u_int8_t mfc_flags[MAXVIFS];		/* the MRT_MFC_FLAGS_* flags */ |
| 281 | 	struct	 in_addr mfc_rp;		/* the RP address	 */ |
| 282 | 	struct	 bw_meter *mfc_bw_meter;	/* list of bandwidth meters */ |
| 283 | }; |
| 284 | |
| 285 | /* |
| 286 | * Argument structure used for pkt info. while upcall is made. |
| 287 | */ |
| 288 | struct rtdetq { |
| 289 | 	struct	mbuf *m;		/* a copy of the packet */ |
| 290 | 	struct	ifnet *ifp;		/* interface pkt came in on */ |
| 291 | #ifdef UPCALL_TIMING |
| 292 | 	struct	timeval t;		/* timestamp */ |
| 293 | #endif /* UPCALL_TIMING */ |
| 294 | 	struct	rtdetq *next; |
| 295 | }; |
| 296 | |
| 297 | #define	MFCTBLSIZ	256 |
| 298 | #define	MAX_UPQ		4		/* max. no of pkts in upcall Q */ |
| 299 | |
| 300 | /* |
| 301 | * Token bucket filter code |
| 302 | */ |
| 303 | #define	MAX_BKT_SIZE 10000		/* 10K bytes size */ |
| 304 | #define	MAXQSIZE 10		/* max. no of pkts in token queue */ |
| 305 | |
| 306 | /* |
| 307 | * Structure for measuring the bandwidth and sending an upcall if the |
| 308 | * measured bandwidth is above or below a threshold. |
| 309 | */ |
| 310 | struct bw_meter { |
| 311 | 	struct bw_meter	*bm_mfc_next;		/* next bw meter (same mfc) */ |
| 312 | 	struct bw_meter	*bm_time_next;		/* next bw meter (same time) */ |
| 313 | 	uint32_t	bm_time_hash;		/* the time hash value */ |
| 314 | 	struct mfc	*bm_mfc;		/* the corresponding mfc */ |
| 315 | 	uint32_t	bm_flags;		/* misc flags (see below) */ |
| 316 | #define BW_METER_UNIT_PACKETS	(1 << 0)	/* threshold (in packets) */ |
| 317 | #define BW_METER_UNIT_BYTES	(1 << 1)	/* threshold (in bytes) */ |
| 318 | #define BW_METER_GEQ		(1 << 2)	/* upcall if bw >= threshold */ |
| 319 | #define BW_METER_LEQ		(1 << 3)	/* upcall if bw <= threshold */ |
| 320 | #define BW_METER_USER_FLAGS 	(BW_METER_UNIT_PACKETS |		\ |
| 321 | 				 BW_METER_UNIT_BYTES |			\ |
| 322 | 				 BW_METER_GEQ |				\ |
| 323 | 				 BW_METER_LEQ) |
| 324 | |
| 325 | #define BW_METER_UPCALL_DELIVERED (1 << 24)	/* upcall was delivered */ |
| 326 | |
| 327 | 	struct bw_data	bm_threshold;		/* the upcall threshold	 */ |
| 328 | 	struct bw_data	bm_measured;		/* the measured bw	 */ |
| 329 | 	struct timeval	bm_start_time;		/* abs. time		 */ |
| 330 | }; |
| 331 | |
| 332 | struct sockopt; /* from <sys/socketvar.h> */ |
| 333 | |
| 334 | int	ip_mrouter_set(struct socket *, struct sockopt *); |
| 335 | int	ip_mrouter_get(struct socket *, struct sockopt *); |
| 336 | int	mrt_ioctl(struct socket *, u_long, void *); |
| 337 | int	ip_mrouter_done(void); |
| 338 | void	ip_mrouter_detach(struct ifnet *); |
| 339 | void	reset_vif(struct vif *); |
| 340 | int	ip_mforward(struct mbuf *, struct ifnet *); |
| 341 | |
| 342 | #endif /* _KERNEL */ |
| 343 | |
| 344 | #endif /* !_NETINET_IP_MROUTE_H_ */ |