1/*-
2 * SPDX-License-Identifier: BSD-3-Clause
3 *
4 * Copyright (c) 1982, 1986, 1989, 1993
5 *	The Regents of the University of California.  All rights reserved.
6 *
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
9 * are met:
10 * 1. Redistributions of source code must retain the above copyright
11 *    notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 *    notice, this list of conditions and the following disclaimer in the
14 *    documentation and/or other materials provided with the distribution.
15 * 3. Neither the name of the University nor the names of its contributors
16 *    may be used to endorse or promote products derived from this software
17 *    without specific prior written permission.
18 *
19 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * SUCH DAMAGE.
30 *
31 *	From: @(#)if.h	8.1 (Berkeley) 6/10/93
32 * $FreeBSD$
33 */
34
35#ifndef	_NET_IF_VAR_H_
36#define	_NET_IF_VAR_H_
37
38/*
39 * Structures defining a network interface, providing a packet
40 * transport mechanism (ala level 0 of the PUP protocols).
41 *
42 * Each interface accepts output datagrams of a specified maximum
43 * length, and provides higher level routines with input datagrams
44 * received from its medium.
45 *
46 * Output occurs when the routine if_output is called, with three parameters:
47 *	(*ifp->if_output)(ifp, m, dst, ro)
48 * Here m is the mbuf chain to be sent and dst is the destination address.
49 * The output routine encapsulates the supplied datagram if necessary,
50 * and then transmits it on its medium.
51 *
52 * On input, each interface unwraps the data received by it, and either
53 * places it on the input queue of an internetwork datagram routine
54 * and posts the associated software interrupt, or passes the datagram to a raw
55 * packet input routine.
56 *
57 * Routines exist for locating interfaces by their addresses
58 * or for locating an interface on a certain network, as well as more general
59 * routing and gateway routines maintaining information used to locate
60 * interfaces.  These routines live in the files if.c and route.c
61 */
62
63struct	rtentry;		/* ifa_rtrequest */
64struct	socket;
65struct	carp_if;
66struct	carp_softc;
67struct  ifvlantrunk;
68struct	route;			/* if_output */
69struct	vnet;
70struct	ifmedia;
71struct	netmap_adapter;
72struct	debugnet_methods;
73
74#ifdef _KERNEL
75#include <sys/_eventhandler.h>
76#include <sys/mbuf.h>		/* ifqueue only? */
77#include <sys/buf_ring.h>
78#include <net/vnet.h>
79#endif /* _KERNEL */
80#include <sys/ck.h>
81#include <sys/counter.h>
82#include <sys/epoch.h>
83#include <sys/lock.h>		/* XXX */
84#include <sys/mutex.h>		/* struct ifqueue */
85#include <sys/rwlock.h>		/* XXX */
86#include <sys/sx.h>		/* XXX */
87#include <sys/_task.h>		/* if_link_task */
88#define	IF_DUNIT_NONE	-1
89
90#include <net/altq/if_altq.h>
91
92CK_STAILQ_HEAD(ifnethead, ifnet);	/* we use TAILQs so that the order of */
93CK_STAILQ_HEAD(ifaddrhead, ifaddr);	/* instantiation is preserved in the list */
94CK_STAILQ_HEAD(ifmultihead, ifmultiaddr);
95CK_STAILQ_HEAD(ifgrouphead, ifg_group);
96
97#ifdef _KERNEL
98VNET_DECLARE(struct pfil_head *, link_pfil_head);
99#define	V_link_pfil_head	VNET(link_pfil_head)
100#define	PFIL_ETHER_NAME		"ethernet"
101
102#define	HHOOK_IPSEC_INET	0
103#define	HHOOK_IPSEC_INET6	1
104#define	HHOOK_IPSEC_COUNT	2
105VNET_DECLARE(struct hhook_head *, ipsec_hhh_in[HHOOK_IPSEC_COUNT]);
106VNET_DECLARE(struct hhook_head *, ipsec_hhh_out[HHOOK_IPSEC_COUNT]);
107#define	V_ipsec_hhh_in	VNET(ipsec_hhh_in)
108#define	V_ipsec_hhh_out	VNET(ipsec_hhh_out)
109#endif /* _KERNEL */
110
111typedef enum {
112	IFCOUNTER_IPACKETS = 0,
113	IFCOUNTER_IERRORS,
114	IFCOUNTER_OPACKETS,
115	IFCOUNTER_OERRORS,
116	IFCOUNTER_COLLISIONS,
117	IFCOUNTER_IBYTES,
118	IFCOUNTER_OBYTES,
119	IFCOUNTER_IMCASTS,
120	IFCOUNTER_OMCASTS,
121	IFCOUNTER_IQDROPS,
122	IFCOUNTER_OQDROPS,
123	IFCOUNTER_NOPROTO,
124	IFCOUNTERS /* Array size. */
125} ift_counter;
126
127typedef struct ifnet * if_t;
128
129typedef	void (*if_start_fn_t)(if_t);
130typedef	int (*if_ioctl_fn_t)(if_t, u_long, caddr_t);
131typedef	void (*if_init_fn_t)(void *);
132typedef void (*if_qflush_fn_t)(if_t);
133typedef int (*if_transmit_fn_t)(if_t, struct mbuf *);
134typedef	uint64_t (*if_get_counter_t)(if_t, ift_counter);
135
136struct ifnet_hw_tsomax {
137	u_int	tsomaxbytes;	/* TSO total burst length limit in bytes */
138	u_int	tsomaxsegcount;	/* TSO maximum segment count */
139	u_int	tsomaxsegsize;	/* TSO maximum segment size in bytes */
140};
141
142/* Interface encap request types */
143typedef enum {
144	IFENCAP_LL = 1			/* pre-calculate link-layer header */
145} ife_type;
146
147/*
148 * The structure below allows to request various pre-calculated L2/L3 headers
149 * for different media. Requests varies by type (rtype field).
150 *
151 * IFENCAP_LL type: pre-calculates link header based on address family
152 *   and destination lladdr.
153 *
154 *   Input data fields:
155 *     buf: pointer to destination buffer
156 *     bufsize: buffer size
157 *     flags: IFENCAP_FLAG_BROADCAST if destination is broadcast
158 *     family: address family defined by AF_ constant.
159 *     lladdr: pointer to link-layer address
160 *     lladdr_len: length of link-layer address
161 *     hdata: pointer to L3 header (optional, used for ARP requests).
162 *   Output data fields:
163 *     buf: encap data is stored here
164 *     bufsize: resulting encap length is stored here
165 *     lladdr_off: offset of link-layer address from encap hdr start
166 *     hdata: L3 header may be altered if necessary
167 */
168
169struct if_encap_req {
170	u_char		*buf;		/* Destination buffer (w) */
171	size_t		bufsize;	/* size of provided buffer (r) */
172	ife_type	rtype;		/* request type (r) */
173	uint32_t	flags;		/* Request flags (r) */
174	int		family;		/* Address family AF_* (r) */
175	int		lladdr_off;	/* offset from header start (w) */
176	int		lladdr_len;	/* lladdr length (r) */
177	char		*lladdr;	/* link-level address pointer (r) */
178	char		*hdata;		/* Upper layer header data (rw) */
179};
180
181#define	IFENCAP_FLAG_BROADCAST	0x02	/* Destination is broadcast */
182
183/*
184 * Network interface send tag support. The storage of "struct
185 * m_snd_tag" comes from the network driver and it is free to allocate
186 * as much additional space as it wants for its own use.
187 */
188struct ktls_session;
189struct m_snd_tag;
190
191#define	IF_SND_TAG_TYPE_RATE_LIMIT 0
192#define	IF_SND_TAG_TYPE_UNLIMITED 1
193#define	IF_SND_TAG_TYPE_TLS 2
194#define	IF_SND_TAG_TYPE_MAX 3
195
196struct if_snd_tag_alloc_header {
197	uint32_t type;		/* send tag type, see IF_SND_TAG_XXX */
198	uint32_t flowid;	/* mbuf hash value */
199	uint32_t flowtype;	/* mbuf hash type */
200	uint8_t numa_domain;	/* numa domain of associated inp */
201};
202
203struct if_snd_tag_alloc_rate_limit {
204	struct if_snd_tag_alloc_header hdr;
205	uint64_t max_rate;	/* in bytes/s */
206	uint32_t flags;		/* M_NOWAIT or M_WAITOK */
207	uint32_t reserved;	/* alignment */
208};
209
210struct if_snd_tag_alloc_tls {
211	struct if_snd_tag_alloc_header hdr;
212	struct inpcb *inp;
213	const struct ktls_session *tls;
214};
215
216struct if_snd_tag_rate_limit_params {
217	uint64_t max_rate;	/* in bytes/s */
218	uint32_t queue_level;	/* 0 (empty) .. 65535 (full) */
219#define	IF_SND_QUEUE_LEVEL_MIN 0
220#define	IF_SND_QUEUE_LEVEL_MAX 65535
221	uint32_t flags;		/* M_NOWAIT or M_WAITOK */
222};
223
224union if_snd_tag_alloc_params {
225	struct if_snd_tag_alloc_header hdr;
226	struct if_snd_tag_alloc_rate_limit rate_limit;
227	struct if_snd_tag_alloc_rate_limit unlimited;
228	struct if_snd_tag_alloc_tls tls;
229};
230
231union if_snd_tag_modify_params {
232	struct if_snd_tag_rate_limit_params rate_limit;
233	struct if_snd_tag_rate_limit_params unlimited;
234};
235
236union if_snd_tag_query_params {
237	struct if_snd_tag_rate_limit_params rate_limit;
238	struct if_snd_tag_rate_limit_params unlimited;
239};
240
241/* Query return flags */
242#define RT_NOSUPPORT	  0x00000000	/* Not supported */
243#define RT_IS_INDIRECT    0x00000001	/*
244					 * Interface like a lagg, select
245					 * the actual interface for
246					 * capabilities.
247					 */
248#define RT_IS_SELECTABLE  0x00000002	/*
249					 * No rate table, you select
250					 * rates and the first
251					 * number_of_rates are created.
252					 */
253#define RT_IS_FIXED_TABLE 0x00000004	/* A fixed table is attached */
254#define RT_IS_UNUSABLE	  0x00000008	/* It is not usable for this */
255#define RT_IS_SETUP_REQ	  0x00000010	/* The interface setup must be called before use */
256
257struct if_ratelimit_query_results {
258	const uint64_t *rate_table;	/* Pointer to table if present */
259	uint32_t flags;			/* Flags indicating results */
260	uint32_t max_flows;		/* Max flows using, 0=unlimited */
261	uint32_t number_of_rates;	/* How many unique rates can be created */
262	uint32_t min_segment_burst;	/* The amount the adapter bursts at each send */
263};
264
265typedef int (if_snd_tag_alloc_t)(struct ifnet *, union if_snd_tag_alloc_params *,
266    struct m_snd_tag **);
267typedef int (if_snd_tag_modify_t)(struct m_snd_tag *, union if_snd_tag_modify_params *);
268typedef int (if_snd_tag_query_t)(struct m_snd_tag *, union if_snd_tag_query_params *);
269typedef void (if_snd_tag_free_t)(struct m_snd_tag *);
270typedef void (if_ratelimit_query_t)(struct ifnet *,
271    struct if_ratelimit_query_results *);
272typedef int (if_ratelimit_setup_t)(struct ifnet *, uint64_t, uint32_t);
273
274/*
275 * Structure defining a network interface.
276 */
277struct ifnet {
278	/* General book keeping of interface lists. */
279	CK_STAILQ_ENTRY(ifnet) if_link; 	/* all struct ifnets are chained (CK_) */
280	LIST_ENTRY(ifnet) if_clones;	/* interfaces of a cloner */
281	CK_STAILQ_HEAD(, ifg_list) if_groups; /* linked list of groups per if (CK_) */
282					/* protected by if_addr_lock */
283	u_char	if_alloctype;		/* if_type at time of allocation */
284	uint8_t	if_numa_domain;		/* NUMA domain of device */
285	/* Driver and protocol specific information that remains stable. */
286	void	*if_softc;		/* pointer to driver state */
287	void	*if_llsoftc;		/* link layer softc */
288	void	*if_l2com;		/* pointer to protocol bits */
289	const char *if_dname;		/* driver name */
290	int	if_dunit;		/* unit or IF_DUNIT_NONE */
291	u_short	if_index;		/* numeric abbreviation for this if  */
292	short	if_index_reserved;	/* spare space to grow if_index */
293	char	if_xname[IFNAMSIZ];	/* external name (name + unit) */
294	char	*if_description;	/* interface description */
295
296	/* Variable fields that are touched by the stack and drivers. */
297	int	if_flags;		/* up/down, broadcast, etc. */
298	int	if_drv_flags;		/* driver-managed status flags */
299	int	if_capabilities;	/* interface features & capabilities */
300	int	if_capenable;		/* enabled features & capabilities */
301	void	*if_linkmib;		/* link-type-specific MIB data */
302	size_t	if_linkmiblen;		/* length of above data */
303	u_int	if_refcount;		/* reference count */
304
305	/* These fields are shared with struct if_data. */
306	uint8_t		if_type;	/* ethernet, tokenring, etc */
307	uint8_t		if_addrlen;	/* media address length */
308	uint8_t		if_hdrlen;	/* media header length */
309	uint8_t		if_link_state;	/* current link state */
310	uint32_t	if_mtu;		/* maximum transmission unit */
311	uint32_t	if_metric;	/* routing metric (external only) */
312	uint64_t	if_baudrate;	/* linespeed */
313	uint64_t	if_hwassist;	/* HW offload capabilities, see IFCAP */
314	time_t		if_epoch;	/* uptime at attach or stat reset */
315	struct timeval	if_lastchange;	/* time of last administrative change */
316
317	struct  ifaltq if_snd;		/* output queue (includes altq) */
318	struct	task if_linktask;	/* task for link change events */
319	struct	task if_addmultitask;	/* task for SIOCADDMULTI */
320
321	/* Addresses of different protocol families assigned to this if. */
322	struct mtx if_addr_lock;	/* lock to protect address lists */
323		/*
324		 * if_addrhead is the list of all addresses associated to
325		 * an interface.
326		 * Some code in the kernel assumes that first element
327		 * of the list has type AF_LINK, and contains sockaddr_dl
328		 * addresses which store the link-level address and the name
329		 * of the interface.
330		 * However, access to the AF_LINK address through this
331		 * field is deprecated. Use if_addr or ifaddr_byindex() instead.
332		 */
333	struct	ifaddrhead if_addrhead;	/* linked list of addresses per if */
334	struct	ifmultihead if_multiaddrs; /* multicast addresses configured */
335	int	if_amcount;		/* number of all-multicast requests */
336	struct	ifaddr	*if_addr;	/* pointer to link-level address */
337	void	*if_hw_addr;		/* hardware link-level address */
338	const u_int8_t *if_broadcastaddr; /* linklevel broadcast bytestring */
339	struct	mtx if_afdata_lock;
340	void	*if_afdata[AF_MAX];
341	int	if_afdata_initialized;
342
343	/* Additional features hung off the interface. */
344	u_int	if_fib;			/* interface FIB */
345	struct	vnet *if_vnet;		/* pointer to network stack instance */
346	struct	vnet *if_home_vnet;	/* where this ifnet originates from */
347	struct  ifvlantrunk *if_vlantrunk; /* pointer to 802.1q data */
348	struct	bpf_if *if_bpf;		/* packet filter structure */
349	int	if_pcount;		/* number of promiscuous listeners */
350	void	*if_bridge;		/* bridge glue */
351	void	*if_lagg;		/* lagg glue */
352	void	*if_pf_kif;		/* pf glue */
353	struct	carp_if *if_carp;	/* carp interface structure */
354	struct	label *if_label;	/* interface MAC label */
355	struct	netmap_adapter *if_netmap; /* netmap(4) softc */
356
357	/* Various procedures of the layer2 encapsulation and drivers. */
358	int	(*if_output)		/* output routine (enqueue) */
359		(struct ifnet *, struct mbuf *, const struct sockaddr *,
360		     struct route *);
361	void	(*if_input)		/* input routine (from h/w driver) */
362		(struct ifnet *, struct mbuf *);
363	struct mbuf *(*if_bridge_input)(struct ifnet *, struct mbuf *);
364	int	(*if_bridge_output)(struct ifnet *, struct mbuf *, struct sockaddr *,
365		    struct rtentry *);
366	void (*if_bridge_linkstate)(struct ifnet *ifp);
367	if_start_fn_t	if_start;	/* initiate output routine */
368	if_ioctl_fn_t	if_ioctl;	/* ioctl routine */
369	if_init_fn_t	if_init;	/* Init routine */
370	int	(*if_resolvemulti)	/* validate/resolve multicast */
371		(struct ifnet *, struct sockaddr **, struct sockaddr *);
372	if_qflush_fn_t	if_qflush;	/* flush any queue */
373	if_transmit_fn_t if_transmit;   /* initiate output routine */
374
375	void	(*if_reassign)		/* reassign to vnet routine */
376		(struct ifnet *, struct vnet *, char *);
377	if_get_counter_t if_get_counter; /* get counter values */
378	int	(*if_requestencap)	/* make link header from request */
379		(struct ifnet *, struct if_encap_req *);
380
381	/* Statistics. */
382	counter_u64_t	if_counters[IFCOUNTERS];
383
384	/* Stuff that's only temporary and doesn't belong here. */
385
386	/*
387	 * Network adapter TSO limits:
388	 * ===========================
389	 *
390	 * If the "if_hw_tsomax" field is zero the maximum segment
391	 * length limit does not apply. If the "if_hw_tsomaxsegcount"
392	 * or the "if_hw_tsomaxsegsize" field is zero the TSO segment
393	 * count limit does not apply. If all three fields are zero,
394	 * there is no TSO limit.
395	 *
396	 * NOTE: The TSO limits should reflect the values used in the
397	 * BUSDMA tag a network adapter is using to load a mbuf chain
398	 * for transmission. The TCP/IP network stack will subtract
399	 * space for all linklevel and protocol level headers and
400	 * ensure that the full mbuf chain passed to the network
401	 * adapter fits within the given limits.
402	 */
403	u_int	if_hw_tsomax;		/* TSO maximum size in bytes */
404	u_int	if_hw_tsomaxsegcount;	/* TSO maximum segment count */
405	u_int	if_hw_tsomaxsegsize;	/* TSO maximum segment size in bytes */
406
407	/*
408	 * Network adapter send tag support:
409	 */
410	if_snd_tag_alloc_t *if_snd_tag_alloc;
411	if_snd_tag_modify_t *if_snd_tag_modify;
412	if_snd_tag_query_t *if_snd_tag_query;
413	if_snd_tag_free_t *if_snd_tag_free;
414	if_ratelimit_query_t *if_ratelimit_query;
415	if_ratelimit_setup_t *if_ratelimit_setup;
416
417	/* Ethernet PCP */
418	uint8_t if_pcp;
419
420	/*
421	 * Debugnet (Netdump) hooks to be called while in db/panic.
422	 */
423	struct debugnet_methods *if_debugnet_methods;
424	struct epoch_context	if_epoch_ctx;
425
426	/*
427	 * Spare fields to be added before branching a stable branch, so
428	 * that structure can be enhanced without changing the kernel
429	 * binary interface.
430	 */
431	int	if_ispare[4];		/* general use */
432};
433
434/* for compatibility with other BSDs */
435#define	if_name(ifp)	((ifp)->if_xname)
436
437#define	IF_NODOM	255
438/*
439 * Locks for address lists on the network interface.
440 */
441#define	IF_ADDR_LOCK_INIT(if)	mtx_init(&(if)->if_addr_lock, "if_addr_lock", NULL, MTX_DEF)
442#define	IF_ADDR_LOCK_DESTROY(if)	mtx_destroy(&(if)->if_addr_lock)
443
444#define	IF_ADDR_WLOCK(if)	mtx_lock(&(if)->if_addr_lock)
445#define	IF_ADDR_WUNLOCK(if)	mtx_unlock(&(if)->if_addr_lock)
446#define	IF_ADDR_LOCK_ASSERT(if)	MPASS(in_epoch(net_epoch_preempt) || mtx_owned(&(if)->if_addr_lock))
447#define	IF_ADDR_WLOCK_ASSERT(if) mtx_assert(&(if)->if_addr_lock, MA_OWNED)
448
449#ifdef _KERNEL
450/* interface link layer address change event */
451typedef void (*iflladdr_event_handler_t)(void *, struct ifnet *);
452EVENTHANDLER_DECLARE(iflladdr_event, iflladdr_event_handler_t);
453/* interface address change event */
454typedef void (*ifaddr_event_handler_t)(void *, struct ifnet *);
455EVENTHANDLER_DECLARE(ifaddr_event, ifaddr_event_handler_t);
456typedef void (*ifaddr_event_ext_handler_t)(void *, struct ifnet *,
457    struct ifaddr *, int);
458EVENTHANDLER_DECLARE(ifaddr_event_ext, ifaddr_event_ext_handler_t);
459#define	IFADDR_EVENT_ADD	0
460#define	IFADDR_EVENT_DEL	1
461/* new interface arrival event */
462typedef void (*ifnet_arrival_event_handler_t)(void *, struct ifnet *);
463EVENTHANDLER_DECLARE(ifnet_arrival_event, ifnet_arrival_event_handler_t);
464/* interface departure event */
465typedef void (*ifnet_departure_event_handler_t)(void *, struct ifnet *);
466EVENTHANDLER_DECLARE(ifnet_departure_event, ifnet_departure_event_handler_t);
467/* Interface link state change event */
468typedef void (*ifnet_link_event_handler_t)(void *, struct ifnet *, int);
469EVENTHANDLER_DECLARE(ifnet_link_event, ifnet_link_event_handler_t);
470/* Interface up/down event */
471#define IFNET_EVENT_UP		0
472#define IFNET_EVENT_DOWN	1
473#define IFNET_EVENT_PCP		2	/* priority code point, PCP */
474
475typedef void (*ifnet_event_fn)(void *, struct ifnet *ifp, int event);
476EVENTHANDLER_DECLARE(ifnet_event, ifnet_event_fn);
477
478/*
479 * interface groups
480 */
481struct ifg_group {
482	char				 ifg_group[IFNAMSIZ];
483	u_int				 ifg_refcnt;
484	void				*ifg_pf_kif;
485	CK_STAILQ_HEAD(, ifg_member)	 ifg_members; /* (CK_) */
486	CK_STAILQ_ENTRY(ifg_group)		 ifg_next; /* (CK_) */
487};
488
489struct ifg_member {
490	CK_STAILQ_ENTRY(ifg_member)	 ifgm_next; /* (CK_) */
491	struct ifnet		*ifgm_ifp;
492};
493
494struct ifg_list {
495	struct ifg_group	*ifgl_group;
496	CK_STAILQ_ENTRY(ifg_list)	 ifgl_next; /* (CK_) */
497};
498
499#ifdef _SYS_EVENTHANDLER_H_
500/* group attach event */
501typedef void (*group_attach_event_handler_t)(void *, struct ifg_group *);
502EVENTHANDLER_DECLARE(group_attach_event, group_attach_event_handler_t);
503/* group detach event */
504typedef void (*group_detach_event_handler_t)(void *, struct ifg_group *);
505EVENTHANDLER_DECLARE(group_detach_event, group_detach_event_handler_t);
506/* group change event */
507typedef void (*group_change_event_handler_t)(void *, const char *);
508EVENTHANDLER_DECLARE(group_change_event, group_change_event_handler_t);
509#endif /* _SYS_EVENTHANDLER_H_ */
510
511#define	IF_AFDATA_LOCK_INIT(ifp)	\
512	mtx_init(&(ifp)->if_afdata_lock, "if_afdata", NULL, MTX_DEF)
513
514#define	IF_AFDATA_WLOCK(ifp)	mtx_lock(&(ifp)->if_afdata_lock)
515#define	IF_AFDATA_WUNLOCK(ifp)	mtx_unlock(&(ifp)->if_afdata_lock)
516#define	IF_AFDATA_LOCK(ifp)	IF_AFDATA_WLOCK(ifp)
517#define	IF_AFDATA_UNLOCK(ifp)	IF_AFDATA_WUNLOCK(ifp)
518#define	IF_AFDATA_TRYLOCK(ifp)	mtx_trylock(&(ifp)->if_afdata_lock)
519#define	IF_AFDATA_DESTROY(ifp)	mtx_destroy(&(ifp)->if_afdata_lock)
520
521#define	IF_AFDATA_LOCK_ASSERT(ifp)	MPASS(in_epoch(net_epoch_preempt) || mtx_owned(&(ifp)->if_afdata_lock))
522#define	IF_AFDATA_WLOCK_ASSERT(ifp)	mtx_assert(&(ifp)->if_afdata_lock, MA_OWNED)
523#define	IF_AFDATA_UNLOCK_ASSERT(ifp)	mtx_assert(&(ifp)->if_afdata_lock, MA_NOTOWNED)
524
525/*
526 * 72 was chosen below because it is the size of a TCP/IP
527 * header (40) + the minimum mss (32).
528 */
529#define	IF_MINMTU	72
530#define	IF_MAXMTU	65535
531
532#define	TOEDEV(ifp)	((ifp)->if_llsoftc)
533
534/*
535 * The ifaddr structure contains information about one address
536 * of an interface.  They are maintained by the different address families,
537 * are allocated and attached when an address is set, and are linked
538 * together so all addresses for an interface can be located.
539 *
540 * NOTE: a 'struct ifaddr' is always at the beginning of a larger
541 * chunk of malloc'ed memory, where we store the three addresses
542 * (ifa_addr, ifa_dstaddr and ifa_netmask) referenced here.
543 */
544struct ifaddr {
545	struct	sockaddr *ifa_addr;	/* address of interface */
546	struct	sockaddr *ifa_dstaddr;	/* other end of p-to-p link */
547#define	ifa_broadaddr	ifa_dstaddr	/* broadcast address interface */
548	struct	sockaddr *ifa_netmask;	/* used to determine subnet */
549	struct	ifnet *ifa_ifp;		/* back-pointer to interface */
550	struct	carp_softc *ifa_carp;	/* pointer to CARP data */
551	CK_STAILQ_ENTRY(ifaddr) ifa_link;	/* queue macro glue */
552	u_short	ifa_flags;		/* mostly rt_flags for cloning */
553#define	IFA_ROUTE	RTF_UP		/* route installed */
554#define	IFA_RTSELF	RTF_HOST	/* loopback route to self installed */
555	u_int	ifa_refcnt;		/* references to this structure */
556
557	counter_u64_t	ifa_ipackets;
558	counter_u64_t	ifa_opackets;
559	counter_u64_t	ifa_ibytes;
560	counter_u64_t	ifa_obytes;
561	struct	epoch_context	ifa_epoch_ctx;
562};
563
564struct ifaddr *	ifa_alloc(size_t size, int flags);
565void	ifa_free(struct ifaddr *ifa);
566void	ifa_ref(struct ifaddr *ifa);
567
568/*
569 * Multicast address structure.  This is analogous to the ifaddr
570 * structure except that it keeps track of multicast addresses.
571 */
572#define IFMA_F_ENQUEUED		0x1
573struct ifmultiaddr {
574	CK_STAILQ_ENTRY(ifmultiaddr) ifma_link; /* queue macro glue */
575	struct	sockaddr *ifma_addr; 	/* address this membership is for */
576	struct	sockaddr *ifma_lladdr;	/* link-layer translation, if any */
577	struct	ifnet *ifma_ifp;	/* back-pointer to interface */
578	u_int	ifma_refcount;		/* reference count */
579	int	ifma_flags;
580	void	*ifma_protospec;	/* protocol-specific state, if any */
581	struct	ifmultiaddr *ifma_llifma; /* pointer to ifma for ifma_lladdr */
582	struct	epoch_context	ifma_epoch_ctx;
583};
584
585extern	struct rwlock ifnet_rwlock;
586extern	struct sx ifnet_sxlock;
587
588#define	IFNET_WLOCK() do {						\
589	sx_xlock(&ifnet_sxlock);					\
590	rw_wlock(&ifnet_rwlock);					\
591} while (0)
592
593#define	IFNET_WUNLOCK() do {						\
594	rw_wunlock(&ifnet_rwlock);					\
595	sx_xunlock(&ifnet_sxlock);					\
596} while (0)
597
598/*
599 * To assert the ifnet lock, you must know not only whether it's for read or
600 * write, but also whether it was acquired with sleep support or not.
601 */
602#define	IFNET_RLOCK_ASSERT()		sx_assert(&ifnet_sxlock, SA_SLOCKED)
603#define	IFNET_WLOCK_ASSERT() do {					\
604	sx_assert(&ifnet_sxlock, SA_XLOCKED);				\
605	rw_assert(&ifnet_rwlock, RA_WLOCKED);				\
606} while (0)
607
608#define	IFNET_RLOCK()		sx_slock(&ifnet_sxlock)
609#define	IFNET_RUNLOCK()		sx_sunlock(&ifnet_sxlock)
610
611/*
612 * Look up an ifnet given its index; the _ref variant also acquires a
613 * reference that must be freed using if_rele().  It is almost always a bug
614 * to call ifnet_byindex() instead of ifnet_byindex_ref().
615 */
616struct ifnet	*ifnet_byindex(u_short idx);
617struct ifnet	*ifnet_byindex_ref(u_short idx);
618
619/*
620 * Given the index, ifaddr_byindex() returns the one and only
621 * link-level ifaddr for the interface. You are not supposed to use
622 * it to traverse the list of addresses associated to the interface.
623 */
624struct ifaddr	*ifaddr_byindex(u_short idx);
625
626VNET_DECLARE(struct ifnethead, ifnet);
627VNET_DECLARE(struct ifgrouphead, ifg_head);
628VNET_DECLARE(int, if_index);
629VNET_DECLARE(struct ifnet *, loif);	/* first loopback interface */
630
631#define	V_ifnet		VNET(ifnet)
632#define	V_ifg_head	VNET(ifg_head)
633#define	V_if_index	VNET(if_index)
634#define	V_loif		VNET(loif)
635
636#ifdef MCAST_VERBOSE
637#define MCDPRINTF printf
638#else
639#define MCDPRINTF(...)
640#endif
641
642int	if_addgroup(struct ifnet *, const char *);
643int	if_delgroup(struct ifnet *, const char *);
644int	if_addmulti(struct ifnet *, struct sockaddr *, struct ifmultiaddr **);
645int	if_allmulti(struct ifnet *, int);
646struct	ifnet* if_alloc(u_char);
647struct	ifnet* if_alloc_dev(u_char, device_t dev);
648struct	ifnet* if_alloc_domain(u_char, int numa_domain);
649void	if_attach(struct ifnet *);
650void	if_dead(struct ifnet *);
651int	if_delmulti(struct ifnet *, struct sockaddr *);
652void	if_delmulti_ifma(struct ifmultiaddr *);
653void	if_delmulti_ifma_flags(struct ifmultiaddr *, int flags);
654void	if_detach(struct ifnet *);
655void	if_purgeaddrs(struct ifnet *);
656void	if_delallmulti(struct ifnet *);
657void	if_down(struct ifnet *);
658struct ifmultiaddr *
659	if_findmulti(struct ifnet *, const struct sockaddr *);
660void	if_freemulti(struct ifmultiaddr *ifma);
661void	if_free(struct ifnet *);
662void	if_initname(struct ifnet *, const char *, int);
663void	if_link_state_change(struct ifnet *, int);
664int	if_printf(struct ifnet *, const char *, ...) __printflike(2, 3);
665void	if_ref(struct ifnet *);
666void	if_rele(struct ifnet *);
667int	if_setlladdr(struct ifnet *, const u_char *, int);
668int	if_tunnel_check_nesting(struct ifnet *, struct mbuf *, uint32_t, int);
669void	if_up(struct ifnet *);
670int	ifioctl(struct socket *, u_long, caddr_t, struct thread *);
671int	ifpromisc(struct ifnet *, int);
672struct	ifnet *ifunit(const char *);
673struct	ifnet *ifunit_ref(const char *);
674
675int	ifa_add_loopback_route(struct ifaddr *, struct sockaddr *);
676int	ifa_del_loopback_route(struct ifaddr *, struct sockaddr *);
677int	ifa_switch_loopback_route(struct ifaddr *, struct sockaddr *);
678
679struct	ifaddr *ifa_ifwithaddr(const struct sockaddr *);
680int		ifa_ifwithaddr_check(const struct sockaddr *);
681struct	ifaddr *ifa_ifwithbroadaddr(const struct sockaddr *, int);
682struct	ifaddr *ifa_ifwithdstaddr(const struct sockaddr *, int);
683struct	ifaddr *ifa_ifwithnet(const struct sockaddr *, int, int);
684struct	ifaddr *ifa_ifwithroute(int, const struct sockaddr *,
685    const struct sockaddr *, u_int);
686struct	ifaddr *ifaof_ifpforaddr(const struct sockaddr *, struct ifnet *);
687int	ifa_preferred(struct ifaddr *, struct ifaddr *);
688
689int	if_simloop(struct ifnet *ifp, struct mbuf *m, int af, int hlen);
690
691typedef	void *if_com_alloc_t(u_char type, struct ifnet *ifp);
692typedef	void if_com_free_t(void *com, u_char type);
693void	if_register_com_alloc(u_char type, if_com_alloc_t *a, if_com_free_t *f);
694void	if_deregister_com_alloc(u_char type);
695void	if_data_copy(struct ifnet *, struct if_data *);
696uint64_t if_get_counter_default(struct ifnet *, ift_counter);
697void	if_inc_counter(struct ifnet *, ift_counter, int64_t);
698
699#define IF_LLADDR(ifp)							\
700    LLADDR((struct sockaddr_dl *)((ifp)->if_addr->ifa_addr))
701
702uint64_t if_setbaudrate(if_t ifp, uint64_t baudrate);
703uint64_t if_getbaudrate(if_t ifp);
704int if_setcapabilities(if_t ifp, int capabilities);
705int if_setcapabilitiesbit(if_t ifp, int setbit, int clearbit);
706int if_getcapabilities(if_t ifp);
707int if_togglecapenable(if_t ifp, int togglecap);
708int if_setcapenable(if_t ifp, int capenable);
709int if_setcapenablebit(if_t ifp, int setcap, int clearcap);
710int if_getcapenable(if_t ifp);
711const char *if_getdname(if_t ifp);
712int if_setdev(if_t ifp, void *dev);
713int if_setdrvflagbits(if_t ifp, int if_setflags, int clear_flags);
714int if_getdrvflags(if_t ifp);
715int if_setdrvflags(if_t ifp, int flags);
716int if_clearhwassist(if_t ifp);
717int if_sethwassistbits(if_t ifp, int toset, int toclear);
718int if_sethwassist(if_t ifp, int hwassist_bit);
719int if_gethwassist(if_t ifp);
720int if_setsoftc(if_t ifp, void *softc);
721void *if_getsoftc(if_t ifp);
722int if_setflags(if_t ifp, int flags);
723int if_gethwaddr(if_t ifp, struct ifreq *);
724int if_setmtu(if_t ifp, int mtu);
725int if_getmtu(if_t ifp);
726int if_getmtu_family(if_t ifp, int family);
727int if_setflagbits(if_t ifp, int set, int clear);
728int if_getflags(if_t ifp);
729int if_sendq_empty(if_t ifp);
730int if_setsendqready(if_t ifp);
731int if_setsendqlen(if_t ifp, int tx_desc_count);
732int if_sethwtsomax(if_t ifp, u_int if_hw_tsomax);
733int if_sethwtsomaxsegcount(if_t ifp, u_int if_hw_tsomaxsegcount);
734int if_sethwtsomaxsegsize(if_t ifp, u_int if_hw_tsomaxsegsize);
735u_int if_gethwtsomax(if_t ifp);
736u_int if_gethwtsomaxsegcount(if_t ifp);
737u_int if_gethwtsomaxsegsize(if_t ifp);
738int if_input(if_t ifp, struct mbuf* sendmp);
739int if_sendq_prepend(if_t ifp, struct mbuf *m);
740struct mbuf *if_dequeue(if_t ifp);
741int if_setifheaderlen(if_t ifp, int len);
742void if_setrcvif(struct mbuf *m, if_t ifp);
743void if_setvtag(struct mbuf *m, u_int16_t tag);
744u_int16_t if_getvtag(struct mbuf *m);
745int if_vlantrunkinuse(if_t ifp);
746caddr_t if_getlladdr(if_t ifp);
747void *if_gethandle(u_char);
748void if_bpfmtap(if_t ifp, struct mbuf *m);
749void if_etherbpfmtap(if_t ifp, struct mbuf *m);
750void if_vlancap(if_t ifp);
751
752/*
753 * Traversing through interface address lists.
754 */
755struct sockaddr_dl;
756typedef u_int iflladdr_cb_t(void *, struct sockaddr_dl *, u_int);
757u_int if_foreach_lladdr(if_t, iflladdr_cb_t, void *);
758u_int if_foreach_llmaddr(if_t, iflladdr_cb_t, void *);
759u_int if_lladdr_count(if_t);
760u_int if_llmaddr_count(if_t);
761
762int if_getamcount(if_t ifp);
763struct ifaddr * if_getifaddr(if_t ifp);
764
765/* Functions */
766void if_setinitfn(if_t ifp, void (*)(void *));
767void if_setioctlfn(if_t ifp, int (*)(if_t, u_long, caddr_t));
768void if_setstartfn(if_t ifp, void (*)(if_t));
769void if_settransmitfn(if_t ifp, if_transmit_fn_t);
770void if_setqflushfn(if_t ifp, if_qflush_fn_t);
771void if_setgetcounterfn(if_t ifp, if_get_counter_t);
772
773/* Revisit the below. These are inline functions originally */
774int drbr_inuse_drv(if_t ifp, struct buf_ring *br);
775struct mbuf* drbr_dequeue_drv(if_t ifp, struct buf_ring *br);
776int drbr_needs_enqueue_drv(if_t ifp, struct buf_ring *br);
777int drbr_enqueue_drv(if_t ifp, struct buf_ring *br, struct mbuf *m);
778
779/* TSO */
780void if_hw_tsomax_common(if_t ifp, struct ifnet_hw_tsomax *);
781int if_hw_tsomax_update(if_t ifp, struct ifnet_hw_tsomax *);
782
783/* accessors for struct ifreq */
784void *ifr_data_get_ptr(void *ifrp);
785void *ifr_buffer_get_buffer(void *data);
786size_t ifr_buffer_get_length(void *data);
787
788int ifhwioctl(u_long, struct ifnet *, caddr_t, struct thread *);
789
790#ifdef DEVICE_POLLING
791enum poll_cmd { POLL_ONLY, POLL_AND_CHECK_STATUS };
792
793typedef	int poll_handler_t(if_t ifp, enum poll_cmd cmd, int count);
794int    ether_poll_register(poll_handler_t *h, if_t ifp);
795int    ether_poll_deregister(if_t ifp);
796#endif /* DEVICE_POLLING */
797
798#endif /* _KERNEL */
799
800#include <net/ifq.h>	/* XXXAO: temporary unconditional include */
801
802#endif /* !_NET_IF_VAR_H_ */
803