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, rt)
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	rt_addrinfo;		/* ifa_rtrequest */
65struct	socket;
66struct	carp_if;
67struct	carp_softc;
68struct  ifvlantrunk;
69struct	route;			/* if_output */
70struct	vnet;
71struct	ifmedia;
72struct	netmap_adapter;
73struct	netdump_methods;
74
75#ifdef _KERNEL
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_hook);	/* packet filter hooks */
99#define	V_link_pfil_hook	VNET(link_pfil_hook)
100
101#define	HHOOK_IPSEC_INET	0
102#define	HHOOK_IPSEC_INET6	1
103#define	HHOOK_IPSEC_COUNT	2
104VNET_DECLARE(struct hhook_head *, ipsec_hhh_in[HHOOK_IPSEC_COUNT]);
105VNET_DECLARE(struct hhook_head *, ipsec_hhh_out[HHOOK_IPSEC_COUNT]);
106#define	V_ipsec_hhh_in	VNET(ipsec_hhh_in)
107#define	V_ipsec_hhh_out	VNET(ipsec_hhh_out)
108extern epoch_t net_epoch_preempt;
109extern epoch_t net_epoch;
110#endif /* _KERNEL */
111
112typedef enum {
113	IFCOUNTER_IPACKETS = 0,
114	IFCOUNTER_IERRORS,
115	IFCOUNTER_OPACKETS,
116	IFCOUNTER_OERRORS,
117	IFCOUNTER_COLLISIONS,
118	IFCOUNTER_IBYTES,
119	IFCOUNTER_OBYTES,
120	IFCOUNTER_IMCASTS,
121	IFCOUNTER_OMCASTS,
122	IFCOUNTER_IQDROPS,
123	IFCOUNTER_OQDROPS,
124	IFCOUNTER_NOPROTO,
125	IFCOUNTERS /* Array size. */
126} ift_counter;
127
128typedef struct ifnet * if_t;
129
130typedef	void (*if_start_fn_t)(if_t);
131typedef	int (*if_ioctl_fn_t)(if_t, u_long, caddr_t);
132typedef	void (*if_init_fn_t)(void *);
133typedef void (*if_qflush_fn_t)(if_t);
134typedef int (*if_transmit_fn_t)(if_t, struct mbuf *);
135typedef	uint64_t (*if_get_counter_t)(if_t, ift_counter);
136
137struct ifnet_hw_tsomax {
138	u_int	tsomaxbytes;	/* TSO total burst length limit in bytes */
139	u_int	tsomaxsegcount;	/* TSO maximum segment count */
140	u_int	tsomaxsegsize;	/* TSO maximum segment size in bytes */
141};
142
143/* Interface encap request types */
144typedef enum {
145	IFENCAP_LL = 1			/* pre-calculate link-layer header */
146} ife_type;
147
148/*
149 * The structure below allows to request various pre-calculated L2/L3 headers
150 * for different media. Requests varies by type (rtype field).
151 *
152 * IFENCAP_LL type: pre-calculates link header based on address family
153 *   and destination lladdr.
154 *
155 *   Input data fields:
156 *     buf: pointer to destination buffer
157 *     bufsize: buffer size
158 *     flags: IFENCAP_FLAG_BROADCAST if destination is broadcast
159 *     family: address family defined by AF_ constant.
160 *     lladdr: pointer to link-layer address
161 *     lladdr_len: length of link-layer address
162 *     hdata: pointer to L3 header (optional, used for ARP requests).
163 *   Output data fields:
164 *     buf: encap data is stored here
165 *     bufsize: resulting encap length is stored here
166 *     lladdr_off: offset of link-layer address from encap hdr start
167 *     hdata: L3 header may be altered if necessary
168 */
169
170struct if_encap_req {
171	u_char		*buf;		/* Destination buffer (w) */
172	size_t		bufsize;	/* size of provided buffer (r) */
173	ife_type	rtype;		/* request type (r) */
174	uint32_t	flags;		/* Request flags (r) */
175	int		family;		/* Address family AF_* (r) */
176	int		lladdr_off;	/* offset from header start (w) */
177	int		lladdr_len;	/* lladdr length (r) */
178	char		*lladdr;	/* link-level address pointer (r) */
179	char		*hdata;		/* Upper layer header data (rw) */
180};
181
182#define	IFENCAP_FLAG_BROADCAST	0x02	/* Destination is broadcast */
183
184/*
185 * Network interface send tag support. The storage of "struct
186 * m_snd_tag" comes from the network driver and it is free to allocate
187 * as much additional space as it wants for its own use.
188 */
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_MAX 2
194
195struct if_snd_tag_alloc_header {
196	uint32_t type;		/* send tag type, see IF_SND_TAG_XXX */
197	uint32_t flowid;	/* mbuf hash value */
198	uint32_t flowtype;	/* mbuf hash type */
199};
200
201struct if_snd_tag_alloc_rate_limit {
202	struct if_snd_tag_alloc_header hdr;
203	uint64_t max_rate;	/* in bytes/s */
204};
205
206struct if_snd_tag_rate_limit_params {
207	uint64_t max_rate;	/* in bytes/s */
208	uint32_t queue_level;	/* 0 (empty) .. 65535 (full) */
209#define	IF_SND_QUEUE_LEVEL_MIN 0
210#define	IF_SND_QUEUE_LEVEL_MAX 65535
211	uint32_t reserved;	/* padding */
212};
213
214union if_snd_tag_alloc_params {
215	struct if_snd_tag_alloc_header hdr;
216	struct if_snd_tag_alloc_rate_limit rate_limit;
217	struct if_snd_tag_alloc_rate_limit unlimited;
218};
219
220union if_snd_tag_modify_params {
221	struct if_snd_tag_rate_limit_params rate_limit;
222	struct if_snd_tag_rate_limit_params unlimited;
223};
224
225union if_snd_tag_query_params {
226	struct if_snd_tag_rate_limit_params rate_limit;
227	struct if_snd_tag_rate_limit_params unlimited;
228};
229
230typedef int (if_snd_tag_alloc_t)(struct ifnet *, union if_snd_tag_alloc_params *,
231    struct m_snd_tag **);
232typedef int (if_snd_tag_modify_t)(struct m_snd_tag *, union if_snd_tag_modify_params *);
233typedef int (if_snd_tag_query_t)(struct m_snd_tag *, union if_snd_tag_query_params *);
234typedef void (if_snd_tag_free_t)(struct m_snd_tag *);
235
236/*
237 * Structure defining a network interface.
238 */
239struct ifnet {
240	/* General book keeping of interface lists. */
241	CK_STAILQ_ENTRY(ifnet) if_link; 	/* all struct ifnets are chained (CK_) */
242	LIST_ENTRY(ifnet) if_clones;	/* interfaces of a cloner */
243	CK_STAILQ_HEAD(, ifg_list) if_groups; /* linked list of groups per if (CK_) */
244					/* protected by if_addr_lock */
245	u_char	if_alloctype;		/* if_type at time of allocation */
246
247	/* Driver and protocol specific information that remains stable. */
248	void	*if_softc;		/* pointer to driver state */
249	void	*if_llsoftc;		/* link layer softc */
250	void	*if_l2com;		/* pointer to protocol bits */
251	const char *if_dname;		/* driver name */
252	int	if_dunit;		/* unit or IF_DUNIT_NONE */
253	u_short	if_index;		/* numeric abbreviation for this if  */
254	short	if_index_reserved;	/* spare space to grow if_index */
255	char	if_xname[IFNAMSIZ];	/* external name (name + unit) */
256	char	*if_description;	/* interface description */
257
258	/* Variable fields that are touched by the stack and drivers. */
259	int	if_flags;		/* up/down, broadcast, etc. */
260	int	if_drv_flags;		/* driver-managed status flags */
261	int	if_capabilities;	/* interface features & capabilities */
262	int	if_capenable;		/* enabled features & capabilities */
263	void	*if_linkmib;		/* link-type-specific MIB data */
264	size_t	if_linkmiblen;		/* length of above data */
265	u_int	if_refcount;		/* reference count */
266
267	/* These fields are shared with struct if_data. */
268	uint8_t		if_type;	/* ethernet, tokenring, etc */
269	uint8_t		if_addrlen;	/* media address length */
270	uint8_t		if_hdrlen;	/* media header length */
271	uint8_t		if_link_state;	/* current link state */
272	uint32_t	if_mtu;		/* maximum transmission unit */
273	uint32_t	if_metric;	/* routing metric (external only) */
274	uint64_t	if_baudrate;	/* linespeed */
275	uint64_t	if_hwassist;	/* HW offload capabilities, see IFCAP */
276	time_t		if_epoch;	/* uptime at attach or stat reset */
277	struct timeval	if_lastchange;	/* time of last administrative change */
278
279	struct  ifaltq if_snd;		/* output queue (includes altq) */
280	struct	task if_linktask;	/* task for link change events */
281
282	/* Addresses of different protocol families assigned to this if. */
283	struct mtx if_addr_lock;	/* lock to protect address lists */
284		/*
285		 * if_addrhead is the list of all addresses associated to
286		 * an interface.
287		 * Some code in the kernel assumes that first element
288		 * of the list has type AF_LINK, and contains sockaddr_dl
289		 * addresses which store the link-level address and the name
290		 * of the interface.
291		 * However, access to the AF_LINK address through this
292		 * field is deprecated. Use if_addr or ifaddr_byindex() instead.
293		 */
294	struct	ifaddrhead if_addrhead;	/* linked list of addresses per if */
295	struct	ifmultihead if_multiaddrs; /* multicast addresses configured */
296	int	if_amcount;		/* number of all-multicast requests */
297	struct	ifaddr	*if_addr;	/* pointer to link-level address */
298	void	*if_hw_addr;		/* hardware link-level address */
299	const u_int8_t *if_broadcastaddr; /* linklevel broadcast bytestring */
300	struct	mtx if_afdata_lock;
301	void	*if_afdata[AF_MAX];
302	int	if_afdata_initialized;
303
304	/* Additional features hung off the interface. */
305	u_int	if_fib;			/* interface FIB */
306	struct	vnet *if_vnet;		/* pointer to network stack instance */
307	struct	vnet *if_home_vnet;	/* where this ifnet originates from */
308	struct  ifvlantrunk *if_vlantrunk; /* pointer to 802.1q data */
309	struct	bpf_if *if_bpf;		/* packet filter structure */
310	int	if_pcount;		/* number of promiscuous listeners */
311	void	*if_bridge;		/* bridge glue */
312	void	*if_lagg;		/* lagg glue */
313	void	*if_pf_kif;		/* pf glue */
314	struct	carp_if *if_carp;	/* carp interface structure */
315	struct	label *if_label;	/* interface MAC label */
316	struct	netmap_adapter *if_netmap; /* netmap(4) softc */
317
318	/* Various procedures of the layer2 encapsulation and drivers. */
319	int	(*if_output)		/* output routine (enqueue) */
320		(struct ifnet *, struct mbuf *, const struct sockaddr *,
321		     struct route *);
322	void	(*if_input)		/* input routine (from h/w driver) */
323		(struct ifnet *, struct mbuf *);
324	struct mbuf *(*if_bridge_input)(struct ifnet *, struct mbuf *);
325	int	(*if_bridge_output)(struct ifnet *, struct mbuf *, struct sockaddr *,
326		    struct rtentry *);
327	void (*if_bridge_linkstate)(struct ifnet *ifp);
328	if_start_fn_t	if_start;	/* initiate output routine */
329	if_ioctl_fn_t	if_ioctl;	/* ioctl routine */
330	if_init_fn_t	if_init;	/* Init routine */
331	int	(*if_resolvemulti)	/* validate/resolve multicast */
332		(struct ifnet *, struct sockaddr **, struct sockaddr *);
333	if_qflush_fn_t	if_qflush;	/* flush any queue */
334	if_transmit_fn_t if_transmit;   /* initiate output routine */
335
336	void	(*if_reassign)		/* reassign to vnet routine */
337		(struct ifnet *, struct vnet *, char *);
338	if_get_counter_t if_get_counter; /* get counter values */
339	int	(*if_requestencap)	/* make link header from request */
340		(struct ifnet *, struct if_encap_req *);
341
342	/* Statistics. */
343	counter_u64_t	if_counters[IFCOUNTERS];
344
345	/* Stuff that's only temporary and doesn't belong here. */
346
347	/*
348	 * Network adapter TSO limits:
349	 * ===========================
350	 *
351	 * If the "if_hw_tsomax" field is zero the maximum segment
352	 * length limit does not apply. If the "if_hw_tsomaxsegcount"
353	 * or the "if_hw_tsomaxsegsize" field is zero the TSO segment
354	 * count limit does not apply. If all three fields are zero,
355	 * there is no TSO limit.
356	 *
357	 * NOTE: The TSO limits should reflect the values used in the
358	 * BUSDMA tag a network adapter is using to load a mbuf chain
359	 * for transmission. The TCP/IP network stack will subtract
360	 * space for all linklevel and protocol level headers and
361	 * ensure that the full mbuf chain passed to the network
362	 * adapter fits within the given limits.
363	 */
364	u_int	if_hw_tsomax;		/* TSO maximum size in bytes */
365	u_int	if_hw_tsomaxsegcount;	/* TSO maximum segment count */
366	u_int	if_hw_tsomaxsegsize;	/* TSO maximum segment size in bytes */
367
368	/*
369	 * Network adapter send tag support:
370	 */
371	if_snd_tag_alloc_t *if_snd_tag_alloc;
372	if_snd_tag_modify_t *if_snd_tag_modify;
373	if_snd_tag_query_t *if_snd_tag_query;
374	if_snd_tag_free_t *if_snd_tag_free;
375
376	/* Ethernet PCP */
377	uint8_t if_pcp;
378
379	/*
380	 * Netdump hooks to be called while dumping.
381	 */
382	struct netdump_methods *if_netdump_methods;
383	struct epoch_context	if_epoch_ctx;
384	void 		       *if_unused[4];
385
386	/*
387	 * Spare fields to be added before branching a stable branch, so
388	 * that structure can be enhanced without changing the kernel
389	 * binary interface.
390	 */
391	int	if_ispare[4];		/* general use */
392};
393
394/* for compatibility with other BSDs */
395#define	if_name(ifp)	((ifp)->if_xname)
396
397/*
398 * Locks for address lists on the network interface.
399 */
400#define	IF_ADDR_LOCK_INIT(if)	mtx_init(&(if)->if_addr_lock, "if_addr_lock", NULL, MTX_DEF)
401#define	IF_ADDR_LOCK_DESTROY(if)	mtx_destroy(&(if)->if_addr_lock)
402#define	IF_ADDR_RLOCK(if)       struct epoch_tracker if_addr_et; epoch_enter_preempt(net_epoch_preempt, &if_addr_et);
403#define	IF_ADDR_RUNLOCK(if)     epoch_exit_preempt(net_epoch_preempt, &if_addr_et);
404
405#define	IF_ADDR_WLOCK(if)	mtx_lock(&(if)->if_addr_lock)
406#define	IF_ADDR_WUNLOCK(if)	mtx_unlock(&(if)->if_addr_lock)
407#define	IF_ADDR_LOCK_ASSERT(if)	MPASS(in_epoch(net_epoch_preempt) || mtx_owned(&(if)->if_addr_lock))
408#define	IF_ADDR_WLOCK_ASSERT(if) mtx_assert(&(if)->if_addr_lock, MA_OWNED)
409#define	NET_EPOCH_ENTER() struct epoch_tracker nep_et; epoch_enter_preempt(net_epoch_preempt, &nep_et)
410#define	NET_EPOCH_ENTER_ET(et) epoch_enter_preempt(net_epoch_preempt, &(et))
411#define	NET_EPOCH_EXIT() epoch_exit_preempt(net_epoch_preempt, &nep_et)
412#define	NET_EPOCH_EXIT_ET(et) epoch_exit_preempt(net_epoch_preempt, &(et))
413#define	NET_EPOCH_WAIT() epoch_wait_preempt(net_epoch_preempt)
414
415
416/*
417 * Function variations on locking macros intended to be used by loadable
418 * kernel modules in order to divorce them from the internals of address list
419 * locking.
420 */
421void	if_addr_rlock(struct ifnet *ifp);	/* if_addrhead */
422void	if_addr_runlock(struct ifnet *ifp);	/* if_addrhead */
423void	if_maddr_rlock(if_t ifp);	/* if_multiaddrs */
424void	if_maddr_runlock(if_t ifp);	/* if_multiaddrs */
425
426#ifdef _KERNEL
427#ifdef _SYS_EVENTHANDLER_H_
428/* interface link layer address change event */
429typedef void (*iflladdr_event_handler_t)(void *, struct ifnet *);
430EVENTHANDLER_DECLARE(iflladdr_event, iflladdr_event_handler_t);
431/* interface address change event */
432typedef void (*ifaddr_event_handler_t)(void *, struct ifnet *);
433EVENTHANDLER_DECLARE(ifaddr_event, ifaddr_event_handler_t);
434typedef void (*ifaddr_event_ext_handler_t)(void *, struct ifnet *,
435    struct ifaddr *, int);
436EVENTHANDLER_DECLARE(ifaddr_event_ext, ifaddr_event_ext_handler_t);
437#define	IFADDR_EVENT_ADD	0
438#define	IFADDR_EVENT_DEL	1
439/* new interface arrival event */
440typedef void (*ifnet_arrival_event_handler_t)(void *, struct ifnet *);
441EVENTHANDLER_DECLARE(ifnet_arrival_event, ifnet_arrival_event_handler_t);
442/* interface departure event */
443typedef void (*ifnet_departure_event_handler_t)(void *, struct ifnet *);
444EVENTHANDLER_DECLARE(ifnet_departure_event, ifnet_departure_event_handler_t);
445/* Interface link state change event */
446typedef void (*ifnet_link_event_handler_t)(void *, struct ifnet *, int);
447EVENTHANDLER_DECLARE(ifnet_link_event, ifnet_link_event_handler_t);
448/* Interface up/down event */
449#define IFNET_EVENT_UP		0
450#define IFNET_EVENT_DOWN	1
451#define IFNET_EVENT_PCP		2	/* priority code point, PCP */
452
453typedef void (*ifnet_event_fn)(void *, struct ifnet *ifp, int event);
454EVENTHANDLER_DECLARE(ifnet_event, ifnet_event_fn);
455#endif /* _SYS_EVENTHANDLER_H_ */
456
457/*
458 * interface groups
459 */
460struct ifg_group {
461	char				 ifg_group[IFNAMSIZ];
462	u_int				 ifg_refcnt;
463	void				*ifg_pf_kif;
464	CK_STAILQ_HEAD(, ifg_member)	 ifg_members; /* (CK_) */
465	CK_STAILQ_ENTRY(ifg_group)		 ifg_next; /* (CK_) */
466};
467
468struct ifg_member {
469	CK_STAILQ_ENTRY(ifg_member)	 ifgm_next; /* (CK_) */
470	struct ifnet		*ifgm_ifp;
471};
472
473struct ifg_list {
474	struct ifg_group	*ifgl_group;
475	CK_STAILQ_ENTRY(ifg_list)	 ifgl_next; /* (CK_) */
476};
477
478#ifdef _SYS_EVENTHANDLER_H_
479/* group attach event */
480typedef void (*group_attach_event_handler_t)(void *, struct ifg_group *);
481EVENTHANDLER_DECLARE(group_attach_event, group_attach_event_handler_t);
482/* group detach event */
483typedef void (*group_detach_event_handler_t)(void *, struct ifg_group *);
484EVENTHANDLER_DECLARE(group_detach_event, group_detach_event_handler_t);
485/* group change event */
486typedef void (*group_change_event_handler_t)(void *, const char *);
487EVENTHANDLER_DECLARE(group_change_event, group_change_event_handler_t);
488#endif /* _SYS_EVENTHANDLER_H_ */
489
490#define	IF_AFDATA_LOCK_INIT(ifp)	\
491	mtx_init(&(ifp)->if_afdata_lock, "if_afdata", NULL, MTX_DEF)
492
493#define	IF_AFDATA_WLOCK(ifp)	mtx_lock(&(ifp)->if_afdata_lock)
494#define	IF_AFDATA_RLOCK(ifp)	struct epoch_tracker if_afdata_et; epoch_enter_preempt(net_epoch_preempt, &if_afdata_et)
495#define	IF_AFDATA_WUNLOCK(ifp)	mtx_unlock(&(ifp)->if_afdata_lock)
496#define	IF_AFDATA_RUNLOCK(ifp)	epoch_exit_preempt(net_epoch_preempt, &if_afdata_et)
497#define	IF_AFDATA_LOCK(ifp)	IF_AFDATA_WLOCK(ifp)
498#define	IF_AFDATA_UNLOCK(ifp)	IF_AFDATA_WUNLOCK(ifp)
499#define	IF_AFDATA_TRYLOCK(ifp)	mtx_trylock(&(ifp)->if_afdata_lock)
500#define	IF_AFDATA_DESTROY(ifp)	mtx_destroy(&(ifp)->if_afdata_lock)
501
502#define	IF_AFDATA_LOCK_ASSERT(ifp)	MPASS(in_epoch(net_epoch_preempt) || mtx_owned(&(ifp)->if_afdata_lock))
503#define	IF_AFDATA_RLOCK_ASSERT(ifp)	MPASS(in_epoch(net_epoch_preempt));
504#define	IF_AFDATA_WLOCK_ASSERT(ifp)	mtx_assert(&(ifp)->if_afdata_lock, MA_OWNED)
505#define	IF_AFDATA_UNLOCK_ASSERT(ifp)	mtx_assert(&(ifp)->if_afdata_lock, MA_NOTOWNED)
506
507/*
508 * 72 was chosen below because it is the size of a TCP/IP
509 * header (40) + the minimum mss (32).
510 */
511#define	IF_MINMTU	72
512#define	IF_MAXMTU	65535
513
514#define	TOEDEV(ifp)	((ifp)->if_llsoftc)
515
516/*
517 * The ifaddr structure contains information about one address
518 * of an interface.  They are maintained by the different address families,
519 * are allocated and attached when an address is set, and are linked
520 * together so all addresses for an interface can be located.
521 *
522 * NOTE: a 'struct ifaddr' is always at the beginning of a larger
523 * chunk of malloc'ed memory, where we store the three addresses
524 * (ifa_addr, ifa_dstaddr and ifa_netmask) referenced here.
525 */
526struct ifaddr {
527	struct	sockaddr *ifa_addr;	/* address of interface */
528	struct	sockaddr *ifa_dstaddr;	/* other end of p-to-p link */
529#define	ifa_broadaddr	ifa_dstaddr	/* broadcast address interface */
530	struct	sockaddr *ifa_netmask;	/* used to determine subnet */
531	struct	ifnet *ifa_ifp;		/* back-pointer to interface */
532	struct	carp_softc *ifa_carp;	/* pointer to CARP data */
533	CK_STAILQ_ENTRY(ifaddr) ifa_link;	/* queue macro glue */
534	void	(*ifa_rtrequest)	/* check or clean routes (+ or -)'d */
535		(int, struct rtentry *, struct rt_addrinfo *);
536	u_short	ifa_flags;		/* mostly rt_flags for cloning */
537#define	IFA_ROUTE	RTF_UP		/* route installed */
538#define	IFA_RTSELF	RTF_HOST	/* loopback route to self installed */
539	u_int	ifa_refcnt;		/* references to this structure */
540
541	counter_u64_t	ifa_ipackets;
542	counter_u64_t	ifa_opackets;
543	counter_u64_t	ifa_ibytes;
544	counter_u64_t	ifa_obytes;
545	struct	epoch_context	ifa_epoch_ctx;
546};
547
548struct ifaddr *	ifa_alloc(size_t size, int flags);
549void	ifa_free(struct ifaddr *ifa);
550void	ifa_ref(struct ifaddr *ifa);
551
552/*
553 * Multicast address structure.  This is analogous to the ifaddr
554 * structure except that it keeps track of multicast addresses.
555 */
556#define IFMA_F_ENQUEUED		0x1
557struct ifmultiaddr {
558	CK_STAILQ_ENTRY(ifmultiaddr) ifma_link; /* queue macro glue */
559	struct	sockaddr *ifma_addr; 	/* address this membership is for */
560	struct	sockaddr *ifma_lladdr;	/* link-layer translation, if any */
561	struct	ifnet *ifma_ifp;	/* back-pointer to interface */
562	u_int	ifma_refcount;		/* reference count */
563	int	ifma_flags;
564	void	*ifma_protospec;	/* protocol-specific state, if any */
565	struct	ifmultiaddr *ifma_llifma; /* pointer to ifma for ifma_lladdr */
566	struct	epoch_context	ifma_epoch_ctx;
567};
568
569extern	struct rwlock ifnet_rwlock;
570extern	struct sx ifnet_sxlock;
571
572#define	IFNET_WLOCK()		sx_xlock(&ifnet_sxlock)
573#define	IFNET_WUNLOCK()		sx_xunlock(&ifnet_sxlock)
574#define	IFNET_RLOCK_ASSERT()	sx_assert(&ifnet_sxlock, SA_SLOCKED)
575#define	IFNET_RLOCK_NOSLEEP_ASSERT()	MPASS(in_epoch(net_epoch_preempt))
576#define	IFNET_WLOCK_ASSERT()	sx_assert(&ifnet_sxlock, SA_XLOCKED)
577#define	IFNET_RLOCK()		sx_slock(&ifnet_sxlock)
578#define	IFNET_RLOCK_NOSLEEP()	struct epoch_tracker ifnet_rlock_et; epoch_enter_preempt(net_epoch_preempt, &ifnet_rlock_et)
579#define	IFNET_RUNLOCK()		sx_sunlock(&ifnet_sxlock)
580#define	IFNET_RUNLOCK_NOSLEEP()	epoch_exit_preempt(net_epoch_preempt, &ifnet_rlock_et)
581
582/*
583 * Look up an ifnet given its index; the _ref variant also acquires a
584 * reference that must be freed using if_rele().  It is almost always a bug
585 * to call ifnet_byindex() instead of ifnet_byindex_ref().
586 */
587struct ifnet	*ifnet_byindex(u_short idx);
588struct ifnet	*ifnet_byindex_locked(u_short idx);
589struct ifnet	*ifnet_byindex_ref(u_short idx);
590
591/*
592 * Given the index, ifaddr_byindex() returns the one and only
593 * link-level ifaddr for the interface. You are not supposed to use
594 * it to traverse the list of addresses associated to the interface.
595 */
596struct ifaddr	*ifaddr_byindex(u_short idx);
597
598VNET_DECLARE(struct ifnethead, ifnet);
599VNET_DECLARE(struct ifgrouphead, ifg_head);
600VNET_DECLARE(int, if_index);
601VNET_DECLARE(struct ifnet *, loif);	/* first loopback interface */
602
603#define	V_ifnet		VNET(ifnet)
604#define	V_ifg_head	VNET(ifg_head)
605#define	V_if_index	VNET(if_index)
606#define	V_loif		VNET(loif)
607
608#ifdef MCAST_VERBOSE
609#define MCDPRINTF printf
610#else
611#define MCDPRINTF(...)
612#endif
613
614int	if_addgroup(struct ifnet *, const char *);
615int	if_delgroup(struct ifnet *, const char *);
616int	if_addmulti(struct ifnet *, struct sockaddr *, struct ifmultiaddr **);
617int	if_allmulti(struct ifnet *, int);
618struct	ifnet* if_alloc(u_char);
619void	if_attach(struct ifnet *);
620void	if_dead(struct ifnet *);
621int	if_delmulti(struct ifnet *, struct sockaddr *);
622void	if_delmulti_ifma(struct ifmultiaddr *);
623void	if_delmulti_ifma_flags(struct ifmultiaddr *, int flags);
624void	if_detach(struct ifnet *);
625void	if_purgeaddrs(struct ifnet *);
626void	if_delallmulti(struct ifnet *);
627void	if_down(struct ifnet *);
628struct ifmultiaddr *
629	if_findmulti(struct ifnet *, const struct sockaddr *);
630void	if_freemulti(struct ifmultiaddr *ifma);
631void	if_free(struct ifnet *);
632void	if_initname(struct ifnet *, const char *, int);
633void	if_link_state_change(struct ifnet *, int);
634int	if_printf(struct ifnet *, const char *, ...) __printflike(2, 3);
635void	if_ref(struct ifnet *);
636void	if_rele(struct ifnet *);
637int	if_setlladdr(struct ifnet *, const u_char *, int);
638int	if_tunnel_check_nesting(struct ifnet *, struct mbuf *, uint32_t, int);
639void	if_up(struct ifnet *);
640int	ifioctl(struct socket *, u_long, caddr_t, struct thread *);
641int	ifpromisc(struct ifnet *, int);
642struct	ifnet *ifunit(const char *);
643struct	ifnet *ifunit_ref(const char *);
644
645int	ifa_add_loopback_route(struct ifaddr *, struct sockaddr *);
646int	ifa_del_loopback_route(struct ifaddr *, struct sockaddr *);
647int	ifa_switch_loopback_route(struct ifaddr *, struct sockaddr *);
648
649struct	ifaddr *ifa_ifwithaddr(const struct sockaddr *);
650int		ifa_ifwithaddr_check(const struct sockaddr *);
651struct	ifaddr *ifa_ifwithbroadaddr(const struct sockaddr *, int);
652struct	ifaddr *ifa_ifwithdstaddr(const struct sockaddr *, int);
653struct	ifaddr *ifa_ifwithnet(const struct sockaddr *, int, int);
654struct	ifaddr *ifa_ifwithroute(int, const struct sockaddr *, struct sockaddr *,
655    u_int);
656struct	ifaddr *ifaof_ifpforaddr(const struct sockaddr *, struct ifnet *);
657int	ifa_preferred(struct ifaddr *, struct ifaddr *);
658
659int	if_simloop(struct ifnet *ifp, struct mbuf *m, int af, int hlen);
660
661typedef	void *if_com_alloc_t(u_char type, struct ifnet *ifp);
662typedef	void if_com_free_t(void *com, u_char type);
663void	if_register_com_alloc(u_char type, if_com_alloc_t *a, if_com_free_t *f);
664void	if_deregister_com_alloc(u_char type);
665void	if_data_copy(struct ifnet *, struct if_data *);
666uint64_t if_get_counter_default(struct ifnet *, ift_counter);
667void	if_inc_counter(struct ifnet *, ift_counter, int64_t);
668
669#define IF_LLADDR(ifp)							\
670    LLADDR((struct sockaddr_dl *)((ifp)->if_addr->ifa_addr))
671
672uint64_t if_setbaudrate(if_t ifp, uint64_t baudrate);
673uint64_t if_getbaudrate(if_t ifp);
674int if_setcapabilities(if_t ifp, int capabilities);
675int if_setcapabilitiesbit(if_t ifp, int setbit, int clearbit);
676int if_getcapabilities(if_t ifp);
677int if_togglecapenable(if_t ifp, int togglecap);
678int if_setcapenable(if_t ifp, int capenable);
679int if_setcapenablebit(if_t ifp, int setcap, int clearcap);
680int if_getcapenable(if_t ifp);
681const char *if_getdname(if_t ifp);
682int if_setdev(if_t ifp, void *dev);
683int if_setdrvflagbits(if_t ifp, int if_setflags, int clear_flags);
684int if_getdrvflags(if_t ifp);
685int if_setdrvflags(if_t ifp, int flags);
686int if_clearhwassist(if_t ifp);
687int if_sethwassistbits(if_t ifp, int toset, int toclear);
688int if_sethwassist(if_t ifp, int hwassist_bit);
689int if_gethwassist(if_t ifp);
690int if_setsoftc(if_t ifp, void *softc);
691void *if_getsoftc(if_t ifp);
692int if_setflags(if_t ifp, int flags);
693int if_gethwaddr(if_t ifp, struct ifreq *);
694int if_setmtu(if_t ifp, int mtu);
695int if_getmtu(if_t ifp);
696int if_getmtu_family(if_t ifp, int family);
697int if_setflagbits(if_t ifp, int set, int clear);
698int if_getflags(if_t ifp);
699int if_sendq_empty(if_t ifp);
700int if_setsendqready(if_t ifp);
701int if_setsendqlen(if_t ifp, int tx_desc_count);
702int if_sethwtsomax(if_t ifp, u_int if_hw_tsomax);
703int if_sethwtsomaxsegcount(if_t ifp, u_int if_hw_tsomaxsegcount);
704int if_sethwtsomaxsegsize(if_t ifp, u_int if_hw_tsomaxsegsize);
705u_int if_gethwtsomax(if_t ifp);
706u_int if_gethwtsomaxsegcount(if_t ifp);
707u_int if_gethwtsomaxsegsize(if_t ifp);
708int if_input(if_t ifp, struct mbuf* sendmp);
709int if_sendq_prepend(if_t ifp, struct mbuf *m);
710struct mbuf *if_dequeue(if_t ifp);
711int if_setifheaderlen(if_t ifp, int len);
712void if_setrcvif(struct mbuf *m, if_t ifp);
713void if_setvtag(struct mbuf *m, u_int16_t tag);
714u_int16_t if_getvtag(struct mbuf *m);
715int if_vlantrunkinuse(if_t ifp);
716caddr_t if_getlladdr(if_t ifp);
717void *if_gethandle(u_char);
718void if_bpfmtap(if_t ifp, struct mbuf *m);
719void if_etherbpfmtap(if_t ifp, struct mbuf *m);
720void if_vlancap(if_t ifp);
721
722/*
723 * Traversing through interface address lists.
724 */
725struct sockaddr_dl;
726typedef u_int iflladdr_cb_t(void *, struct sockaddr_dl *, u_int);
727u_int if_foreach_lladdr(if_t, iflladdr_cb_t, void *);
728u_int if_foreach_llmaddr(if_t, iflladdr_cb_t, void *);
729int if_multiaddr_count(if_t ifp, int max);
730
731/* Obsoleted multicast management functions. */
732int if_setupmultiaddr(if_t ifp, void *mta, int *cnt, int max);
733int if_multiaddr_array(if_t ifp, void *mta, int *cnt, int max);
734int if_multi_apply(struct ifnet *ifp, int (*filter)(void *, struct ifmultiaddr *, int), void *arg);
735
736int if_getamcount(if_t ifp);
737struct ifaddr * if_getifaddr(if_t ifp);
738
739/* Functions */
740void if_setinitfn(if_t ifp, void (*)(void *));
741void if_setioctlfn(if_t ifp, int (*)(if_t, u_long, caddr_t));
742void if_setstartfn(if_t ifp, void (*)(if_t));
743void if_settransmitfn(if_t ifp, if_transmit_fn_t);
744void if_setqflushfn(if_t ifp, if_qflush_fn_t);
745void if_setgetcounterfn(if_t ifp, if_get_counter_t);
746
747/* Revisit the below. These are inline functions originally */
748int drbr_inuse_drv(if_t ifp, struct buf_ring *br);
749struct mbuf* drbr_dequeue_drv(if_t ifp, struct buf_ring *br);
750int drbr_needs_enqueue_drv(if_t ifp, struct buf_ring *br);
751int drbr_enqueue_drv(if_t ifp, struct buf_ring *br, struct mbuf *m);
752
753/* TSO */
754void if_hw_tsomax_common(if_t ifp, struct ifnet_hw_tsomax *);
755int if_hw_tsomax_update(if_t ifp, struct ifnet_hw_tsomax *);
756
757/* accessors for struct ifreq */
758void *ifr_data_get_ptr(void *ifrp);
759void *ifr_buffer_get_buffer(void *data);
760size_t ifr_buffer_get_length(void *data);
761
762int ifhwioctl(u_long, struct ifnet *, caddr_t, struct thread *);
763
764#ifdef DEVICE_POLLING
765enum poll_cmd { POLL_ONLY, POLL_AND_CHECK_STATUS };
766
767typedef	int poll_handler_t(if_t ifp, enum poll_cmd cmd, int count);
768int    ether_poll_register(poll_handler_t *h, if_t ifp);
769int    ether_poll_deregister(if_t ifp);
770#endif /* DEVICE_POLLING */
771
772#endif /* _KERNEL */
773
774#include <net/ifq.h>	/* XXXAO: temporary unconditional include */
775
776#endif /* !_NET_IF_VAR_H_ */
777