ifq.h revision 241037
1/*-
2 * Copyright (c) 1982, 1986, 1989, 1993
3 *	The Regents of the University of California.  All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 *    notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 *    notice, this list of conditions and the following disclaimer in the
12 *    documentation and/or other materials provided with the distribution.
13 * 4. Neither the name of the University nor the names of its contributors
14 *    may be used to endorse or promote products derived from this software
15 *    without specific prior written permission.
16 *
17 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27 * SUCH DAMAGE.
28 *
29 *	From: @(#)if.h	8.1 (Berkeley) 6/10/93
30 * $FreeBSD: head/sys/net/if_var.h 241037 2012-09-28 18:28:27Z glebius $
31 */
32
33#ifndef	_NET_IF_VAR_H_
34#define	_NET_IF_VAR_H_
35
36/*
37 * Structures defining a network interface, providing a packet
38 * transport mechanism (ala level 0 of the PUP protocols).
39 *
40 * Each interface accepts output datagrams of a specified maximum
41 * length, and provides higher level routines with input datagrams
42 * received from its medium.
43 *
44 * Output occurs when the routine if_output is called, with three parameters:
45 *	(*ifp->if_output)(ifp, m, dst, rt)
46 * Here m is the mbuf chain to be sent and dst is the destination address.
47 * The output routine encapsulates the supplied datagram if necessary,
48 * and then transmits it on its medium.
49 *
50 * On input, each interface unwraps the data received by it, and either
51 * places it on the input queue of an internetwork datagram routine
52 * and posts the associated software interrupt, or passes the datagram to a raw
53 * packet input routine.
54 *
55 * Routines exist for locating interfaces by their addresses
56 * or for locating an interface on a certain network, as well as more general
57 * routing and gateway routines maintaining information used to locate
58 * interfaces.  These routines live in the files if.c and route.c
59 */
60
61#ifdef __STDC__
62/*
63 * Forward structure declarations for function prototypes [sic].
64 */
65struct	mbuf;
66struct	thread;
67struct	rtentry;
68struct	rt_addrinfo;
69struct	socket;
70struct	ether_header;
71struct	carp_if;
72struct	carp_softc;
73struct  ifvlantrunk;
74struct	route;
75struct	vnet;
76#endif
77
78#include <sys/queue.h>		/* get TAILQ macros */
79
80#ifdef _KERNEL
81#include <sys/mbuf.h>
82#include <sys/eventhandler.h>
83#include <sys/buf_ring.h>
84#include <net/vnet.h>
85#endif /* _KERNEL */
86#include <sys/lock.h>		/* XXX */
87#include <sys/mutex.h>		/* XXX */
88#include <sys/rwlock.h>		/* XXX */
89#include <sys/sx.h>		/* XXX */
90#include <sys/event.h>		/* XXX */
91#include <sys/_task.h>
92
93#define	IF_DUNIT_NONE	-1
94
95#include <altq/if_altq.h>
96
97TAILQ_HEAD(ifnethead, ifnet);	/* we use TAILQs so that the order of */
98TAILQ_HEAD(ifaddrhead, ifaddr);	/* instantiation is preserved in the list */
99TAILQ_HEAD(ifmultihead, ifmultiaddr);
100TAILQ_HEAD(ifgrouphead, ifg_group);
101
102#ifdef _KERNEL
103VNET_DECLARE(struct pfil_head, link_pfil_hook);	/* packet filter hooks */
104#define	V_link_pfil_hook	VNET(link_pfil_hook)
105#endif /* _KERNEL */
106
107/*
108 * Structure defining a queue for a network interface.
109 */
110struct	ifqueue {
111	struct	mbuf *ifq_head;
112	struct	mbuf *ifq_tail;
113	int	ifq_len;
114	int	ifq_maxlen;
115	int	ifq_drops;
116	struct	mtx ifq_mtx;
117};
118
119/*
120 * Structure defining a network interface.
121 *
122 * (Would like to call this struct ``if'', but C isn't PL/1.)
123 */
124
125struct ifnet {
126	void	*if_softc;		/* pointer to driver state */
127	void	*if_l2com;		/* pointer to protocol bits */
128	struct vnet *if_vnet;		/* pointer to network stack instance */
129	TAILQ_ENTRY(ifnet) if_link; 	/* all struct ifnets are chained */
130	char	if_xname[IFNAMSIZ];	/* external name (name + unit) */
131	const char *if_dname;		/* driver name */
132	int	if_dunit;		/* unit or IF_DUNIT_NONE */
133	u_int	if_refcount;		/* reference count */
134	struct	ifaddrhead if_addrhead;	/* linked list of addresses per if */
135		/*
136		 * if_addrhead is the list of all addresses associated to
137		 * an interface.
138		 * Some code in the kernel assumes that first element
139		 * of the list has type AF_LINK, and contains sockaddr_dl
140		 * addresses which store the link-level address and the name
141		 * of the interface.
142		 * However, access to the AF_LINK address through this
143		 * field is deprecated. Use if_addr or ifaddr_byindex() instead.
144		 */
145	int	if_pcount;		/* number of promiscuous listeners */
146	struct	carp_if *if_carp;	/* carp interface structure */
147	struct	bpf_if *if_bpf;		/* packet filter structure */
148	u_short	if_index;		/* numeric abbreviation for this if  */
149	short	if_index_reserved;	/* spare space to grow if_index */
150	struct  ifvlantrunk *if_vlantrunk; /* pointer to 802.1q data */
151	int	if_flags;		/* up/down, broadcast, etc. */
152	int	if_capabilities;	/* interface features & capabilities */
153	int	if_capenable;		/* enabled features & capabilities */
154	void	*if_linkmib;		/* link-type-specific MIB data */
155	size_t	if_linkmiblen;		/* length of above data */
156	struct	if_data if_data;
157	struct	ifmultihead if_multiaddrs; /* multicast addresses configured */
158	int	if_amcount;		/* number of all-multicast requests */
159/* procedure handles */
160	int	(*if_output)		/* output routine (enqueue) */
161		(struct ifnet *, struct mbuf *, struct sockaddr *,
162		     struct route *);
163	void	(*if_input)		/* input routine (from h/w driver) */
164		(struct ifnet *, struct mbuf *);
165	void	(*if_start)		/* initiate output routine */
166		(struct ifnet *);
167	int	(*if_ioctl)		/* ioctl routine */
168		(struct ifnet *, u_long, caddr_t);
169	void	(*if_init)		/* Init routine */
170		(void *);
171	int	(*if_resolvemulti)	/* validate/resolve multicast */
172		(struct ifnet *, struct sockaddr **, struct sockaddr *);
173	void	(*if_qflush)		/* flush any queues */
174		(struct ifnet *);
175	int	(*if_transmit)		/* initiate output routine */
176		(struct ifnet *, struct mbuf *);
177	void	(*if_reassign)		/* reassign to vnet routine */
178		(struct ifnet *, struct vnet *, char *);
179	struct	vnet *if_home_vnet;	/* where this ifnet originates from */
180	struct	ifaddr	*if_addr;	/* pointer to link-level address */
181	void	*if_llsoftc;		/* link layer softc */
182	int	if_drv_flags;		/* driver-managed status flags */
183	struct  ifaltq if_snd;		/* output queue (includes altq) */
184	const u_int8_t *if_broadcastaddr; /* linklevel broadcast bytestring */
185
186	void	*if_bridge;		/* bridge glue */
187
188	struct	label *if_label;	/* interface MAC label */
189
190	/* these are only used by IPv6 */
191	void	*if_unused[2];
192	void	*if_afdata[AF_MAX];
193	int	if_afdata_initialized;
194	struct	rwlock if_afdata_lock;
195	struct	task if_linktask;	/* task for link change events */
196	struct	rwlock if_addr_lock;	/* lock to protect address lists */
197
198	LIST_ENTRY(ifnet) if_clones;	/* interfaces of a cloner */
199	TAILQ_HEAD(, ifg_list) if_groups; /* linked list of groups per if */
200					/* protected by if_addr_lock */
201	void	*if_pf_kif;
202	void	*if_lagg;		/* lagg glue */
203	char	*if_description;	/* interface description */
204	u_int	if_fib;			/* interface FIB */
205	u_char	if_alloctype;		/* if_type at time of allocation */
206
207	/*
208	 * Spare fields are added so that we can modify sensitive data
209	 * structures without changing the kernel binary interface, and must
210	 * be used with care where binary compatibility is required.
211	 */
212	char	if_cspare[3];
213	int	if_ispare[4];
214	void	*if_pspare[8];		/* 1 netmap, 7 TDB */
215};
216
217typedef void if_init_f_t(void *);
218
219/*
220 * XXX These aliases are terribly dangerous because they could apply
221 * to anything.
222 */
223#define	if_mtu		if_data.ifi_mtu
224#define	if_type		if_data.ifi_type
225#define if_physical	if_data.ifi_physical
226#define	if_addrlen	if_data.ifi_addrlen
227#define	if_hdrlen	if_data.ifi_hdrlen
228#define	if_metric	if_data.ifi_metric
229#define	if_link_state	if_data.ifi_link_state
230#define	if_baudrate	if_data.ifi_baudrate
231#define	if_hwassist	if_data.ifi_hwassist
232#define	if_ipackets	if_data.ifi_ipackets
233#define	if_ierrors	if_data.ifi_ierrors
234#define	if_opackets	if_data.ifi_opackets
235#define	if_oerrors	if_data.ifi_oerrors
236#define	if_collisions	if_data.ifi_collisions
237#define	if_ibytes	if_data.ifi_ibytes
238#define	if_obytes	if_data.ifi_obytes
239#define	if_imcasts	if_data.ifi_imcasts
240#define	if_omcasts	if_data.ifi_omcasts
241#define	if_iqdrops	if_data.ifi_iqdrops
242#define	if_noproto	if_data.ifi_noproto
243#define	if_lastchange	if_data.ifi_lastchange
244
245/* for compatibility with other BSDs */
246#define	if_addrlist	if_addrhead
247#define	if_list		if_link
248#define	if_name(ifp)	((ifp)->if_xname)
249
250/*
251 * Locks for address lists on the network interface.
252 */
253#define	IF_ADDR_LOCK_INIT(if)	rw_init(&(if)->if_addr_lock, "if_addr_lock")
254#define	IF_ADDR_LOCK_DESTROY(if)	rw_destroy(&(if)->if_addr_lock)
255#define	IF_ADDR_WLOCK(if)	rw_wlock(&(if)->if_addr_lock)
256#define	IF_ADDR_WUNLOCK(if)	rw_wunlock(&(if)->if_addr_lock)
257#define	IF_ADDR_RLOCK(if)	rw_rlock(&(if)->if_addr_lock)
258#define	IF_ADDR_RUNLOCK(if)	rw_runlock(&(if)->if_addr_lock)
259#define	IF_ADDR_LOCK_ASSERT(if)	rw_assert(&(if)->if_addr_lock, RA_LOCKED)
260#define	IF_ADDR_WLOCK_ASSERT(if) rw_assert(&(if)->if_addr_lock, RA_WLOCKED)
261
262/*
263 * Function variations on locking macros intended to be used by loadable
264 * kernel modules in order to divorce them from the internals of address list
265 * locking.
266 */
267void	if_addr_rlock(struct ifnet *ifp);	/* if_addrhead */
268void	if_addr_runlock(struct ifnet *ifp);	/* if_addrhead */
269void	if_maddr_rlock(struct ifnet *ifp);	/* if_multiaddrs */
270void	if_maddr_runlock(struct ifnet *ifp);	/* if_multiaddrs */
271
272/*
273 * Output queues (ifp->if_snd) and slow device input queues (*ifp->if_slowq)
274 * are queues of messages stored on ifqueue structures
275 * (defined above).  Entries are added to and deleted from these structures
276 * by these macros, which should be called with ipl raised to splimp().
277 */
278#define IF_LOCK(ifq)		mtx_lock(&(ifq)->ifq_mtx)
279#define IF_UNLOCK(ifq)		mtx_unlock(&(ifq)->ifq_mtx)
280#define	IF_LOCK_ASSERT(ifq)	mtx_assert(&(ifq)->ifq_mtx, MA_OWNED)
281#define	_IF_QFULL(ifq)		((ifq)->ifq_len >= (ifq)->ifq_maxlen)
282#define	_IF_DROP(ifq)		((ifq)->ifq_drops++)
283#define	_IF_QLEN(ifq)		((ifq)->ifq_len)
284
285#define	_IF_ENQUEUE(ifq, m) do { 				\
286	(m)->m_nextpkt = NULL;					\
287	if ((ifq)->ifq_tail == NULL) 				\
288		(ifq)->ifq_head = m; 				\
289	else 							\
290		(ifq)->ifq_tail->m_nextpkt = m; 		\
291	(ifq)->ifq_tail = m; 					\
292	(ifq)->ifq_len++; 					\
293} while (0)
294
295#define IF_ENQUEUE(ifq, m) do {					\
296	IF_LOCK(ifq); 						\
297	_IF_ENQUEUE(ifq, m); 					\
298	IF_UNLOCK(ifq); 					\
299} while (0)
300
301#define	_IF_PREPEND(ifq, m) do {				\
302	(m)->m_nextpkt = (ifq)->ifq_head; 			\
303	if ((ifq)->ifq_tail == NULL) 				\
304		(ifq)->ifq_tail = (m); 				\
305	(ifq)->ifq_head = (m); 					\
306	(ifq)->ifq_len++; 					\
307} while (0)
308
309#define IF_PREPEND(ifq, m) do {		 			\
310	IF_LOCK(ifq); 						\
311	_IF_PREPEND(ifq, m); 					\
312	IF_UNLOCK(ifq); 					\
313} while (0)
314
315#define	_IF_DEQUEUE(ifq, m) do { 				\
316	(m) = (ifq)->ifq_head; 					\
317	if (m) { 						\
318		if (((ifq)->ifq_head = (m)->m_nextpkt) == NULL)	\
319			(ifq)->ifq_tail = NULL; 		\
320		(m)->m_nextpkt = NULL; 				\
321		(ifq)->ifq_len--; 				\
322	} 							\
323} while (0)
324
325#define IF_DEQUEUE(ifq, m) do { 				\
326	IF_LOCK(ifq); 						\
327	_IF_DEQUEUE(ifq, m); 					\
328	IF_UNLOCK(ifq); 					\
329} while (0)
330
331#define	_IF_DEQUEUE_ALL(ifq, m) do {				\
332	(m) = (ifq)->ifq_head;					\
333	(ifq)->ifq_head = (ifq)->ifq_tail = NULL;		\
334	(ifq)->ifq_len = 0;					\
335} while (0)
336
337#define	IF_DEQUEUE_ALL(ifq, m) do {				\
338	IF_LOCK(ifq); 						\
339	_IF_DEQUEUE_ALL(ifq, m);				\
340	IF_UNLOCK(ifq); 					\
341} while (0)
342
343#define	_IF_POLL(ifq, m)	((m) = (ifq)->ifq_head)
344#define	IF_POLL(ifq, m)		_IF_POLL(ifq, m)
345
346#define _IF_DRAIN(ifq) do { 					\
347	struct mbuf *m; 					\
348	for (;;) { 						\
349		_IF_DEQUEUE(ifq, m); 				\
350		if (m == NULL) 					\
351			break; 					\
352		m_freem(m); 					\
353	} 							\
354} while (0)
355
356#define IF_DRAIN(ifq) do {					\
357	IF_LOCK(ifq);						\
358	_IF_DRAIN(ifq);						\
359	IF_UNLOCK(ifq);						\
360} while(0)
361
362#ifdef _KERNEL
363/* interface link layer address change event */
364typedef void (*iflladdr_event_handler_t)(void *, struct ifnet *);
365EVENTHANDLER_DECLARE(iflladdr_event, iflladdr_event_handler_t);
366/* interface address change event */
367typedef void (*ifaddr_event_handler_t)(void *, struct ifnet *);
368EVENTHANDLER_DECLARE(ifaddr_event, ifaddr_event_handler_t);
369/* new interface arrival event */
370typedef void (*ifnet_arrival_event_handler_t)(void *, struct ifnet *);
371EVENTHANDLER_DECLARE(ifnet_arrival_event, ifnet_arrival_event_handler_t);
372/* interface departure event */
373typedef void (*ifnet_departure_event_handler_t)(void *, struct ifnet *);
374EVENTHANDLER_DECLARE(ifnet_departure_event, ifnet_departure_event_handler_t);
375/* Interface link state change event */
376typedef void (*ifnet_link_event_handler_t)(void *, struct ifnet *, int);
377EVENTHANDLER_DECLARE(ifnet_link_event, ifnet_link_event_handler_t);
378
379/*
380 * interface groups
381 */
382struct ifg_group {
383	char				 ifg_group[IFNAMSIZ];
384	u_int				 ifg_refcnt;
385	void				*ifg_pf_kif;
386	TAILQ_HEAD(, ifg_member)	 ifg_members;
387	TAILQ_ENTRY(ifg_group)		 ifg_next;
388};
389
390struct ifg_member {
391	TAILQ_ENTRY(ifg_member)	 ifgm_next;
392	struct ifnet		*ifgm_ifp;
393};
394
395struct ifg_list {
396	struct ifg_group	*ifgl_group;
397	TAILQ_ENTRY(ifg_list)	 ifgl_next;
398};
399
400/* group attach event */
401typedef void (*group_attach_event_handler_t)(void *, struct ifg_group *);
402EVENTHANDLER_DECLARE(group_attach_event, group_attach_event_handler_t);
403/* group detach event */
404typedef void (*group_detach_event_handler_t)(void *, struct ifg_group *);
405EVENTHANDLER_DECLARE(group_detach_event, group_detach_event_handler_t);
406/* group change event */
407typedef void (*group_change_event_handler_t)(void *, const char *);
408EVENTHANDLER_DECLARE(group_change_event, group_change_event_handler_t);
409
410#define	IF_AFDATA_LOCK_INIT(ifp)	\
411	rw_init(&(ifp)->if_afdata_lock, "if_afdata")
412
413#define	IF_AFDATA_WLOCK(ifp)	rw_wlock(&(ifp)->if_afdata_lock)
414#define	IF_AFDATA_RLOCK(ifp)	rw_rlock(&(ifp)->if_afdata_lock)
415#define	IF_AFDATA_WUNLOCK(ifp)	rw_wunlock(&(ifp)->if_afdata_lock)
416#define	IF_AFDATA_RUNLOCK(ifp)	rw_runlock(&(ifp)->if_afdata_lock)
417#define	IF_AFDATA_LOCK(ifp)	IF_AFDATA_WLOCK(ifp)
418#define	IF_AFDATA_UNLOCK(ifp)	IF_AFDATA_WUNLOCK(ifp)
419#define	IF_AFDATA_TRYLOCK(ifp)	rw_try_wlock(&(ifp)->if_afdata_lock)
420#define	IF_AFDATA_DESTROY(ifp)	rw_destroy(&(ifp)->if_afdata_lock)
421
422#define	IF_AFDATA_LOCK_ASSERT(ifp)	rw_assert(&(ifp)->if_afdata_lock, RA_LOCKED)
423#define	IF_AFDATA_RLOCK_ASSERT(ifp)	rw_assert(&(ifp)->if_afdata_lock, RA_RLOCKED)
424#define	IF_AFDATA_WLOCK_ASSERT(ifp)	rw_assert(&(ifp)->if_afdata_lock, RA_WLOCKED)
425#define	IF_AFDATA_UNLOCK_ASSERT(ifp)	rw_assert(&(ifp)->if_afdata_lock, RA_UNLOCKED)
426
427int	if_handoff(struct ifqueue *ifq, struct mbuf *m, struct ifnet *ifp,
428	    int adjust);
429#define	IF_HANDOFF(ifq, m, ifp)			\
430	if_handoff((struct ifqueue *)ifq, m, ifp, 0)
431#define	IF_HANDOFF_ADJ(ifq, m, ifp, adj)	\
432	if_handoff((struct ifqueue *)ifq, m, ifp, adj)
433
434void	if_start(struct ifnet *);
435
436#define	IFQ_ENQUEUE(ifq, m, err)					\
437do {									\
438	IF_LOCK(ifq);							\
439	if (ALTQ_IS_ENABLED(ifq))					\
440		ALTQ_ENQUEUE(ifq, m, NULL, err);			\
441	else {								\
442		if (_IF_QFULL(ifq)) {					\
443			m_freem(m);					\
444			(err) = ENOBUFS;				\
445		} else {						\
446			_IF_ENQUEUE(ifq, m);				\
447			(err) = 0;					\
448		}							\
449	}								\
450	if (err)							\
451		(ifq)->ifq_drops++;					\
452	IF_UNLOCK(ifq);							\
453} while (0)
454
455#define	IFQ_DEQUEUE_NOLOCK(ifq, m)					\
456do {									\
457	if (TBR_IS_ENABLED(ifq))					\
458		(m) = tbr_dequeue_ptr(ifq, ALTDQ_REMOVE);		\
459	else if (ALTQ_IS_ENABLED(ifq))					\
460		ALTQ_DEQUEUE(ifq, m);					\
461	else								\
462		_IF_DEQUEUE(ifq, m);					\
463} while (0)
464
465#define	IFQ_DEQUEUE(ifq, m)						\
466do {									\
467	IF_LOCK(ifq);							\
468	IFQ_DEQUEUE_NOLOCK(ifq, m);					\
469	IF_UNLOCK(ifq);							\
470} while (0)
471
472#define	IFQ_POLL_NOLOCK(ifq, m)						\
473do {									\
474	if (TBR_IS_ENABLED(ifq))					\
475		(m) = tbr_dequeue_ptr(ifq, ALTDQ_POLL);			\
476	else if (ALTQ_IS_ENABLED(ifq))					\
477		ALTQ_POLL(ifq, m);					\
478	else								\
479		_IF_POLL(ifq, m);					\
480} while (0)
481
482#define	IFQ_POLL(ifq, m)						\
483do {									\
484	IF_LOCK(ifq);							\
485	IFQ_POLL_NOLOCK(ifq, m);					\
486	IF_UNLOCK(ifq);							\
487} while (0)
488
489#define	IFQ_PURGE_NOLOCK(ifq)						\
490do {									\
491	if (ALTQ_IS_ENABLED(ifq)) {					\
492		ALTQ_PURGE(ifq);					\
493	} else								\
494		_IF_DRAIN(ifq);						\
495} while (0)
496
497#define	IFQ_PURGE(ifq)							\
498do {									\
499	IF_LOCK(ifq);							\
500	IFQ_PURGE_NOLOCK(ifq);						\
501	IF_UNLOCK(ifq);							\
502} while (0)
503
504#define	IFQ_SET_READY(ifq)						\
505	do { ((ifq)->altq_flags |= ALTQF_READY); } while (0)
506
507#define	IFQ_LOCK(ifq)			IF_LOCK(ifq)
508#define	IFQ_UNLOCK(ifq)			IF_UNLOCK(ifq)
509#define	IFQ_LOCK_ASSERT(ifq)		IF_LOCK_ASSERT(ifq)
510#define	IFQ_IS_EMPTY(ifq)		((ifq)->ifq_len == 0)
511#define	IFQ_INC_LEN(ifq)		((ifq)->ifq_len++)
512#define	IFQ_DEC_LEN(ifq)		(--(ifq)->ifq_len)
513#define	IFQ_INC_DROPS(ifq)		((ifq)->ifq_drops++)
514#define	IFQ_SET_MAXLEN(ifq, len)	((ifq)->ifq_maxlen = (len))
515
516/*
517 * The IFF_DRV_OACTIVE test should really occur in the device driver, not in
518 * the handoff logic, as that flag is locked by the device driver.
519 */
520#define	IFQ_HANDOFF_ADJ(ifp, m, adj, err)				\
521do {									\
522	int len;							\
523	short mflags;							\
524									\
525	len = (m)->m_pkthdr.len;					\
526	mflags = (m)->m_flags;						\
527	IFQ_ENQUEUE(&(ifp)->if_snd, m, err);				\
528	if ((err) == 0) {						\
529		(ifp)->if_obytes += len + (adj);			\
530		if (mflags & M_MCAST)					\
531			(ifp)->if_omcasts++;				\
532		if (((ifp)->if_drv_flags & IFF_DRV_OACTIVE) == 0)	\
533			if_start(ifp);					\
534	}								\
535} while (0)
536
537#define	IFQ_HANDOFF(ifp, m, err)					\
538	IFQ_HANDOFF_ADJ(ifp, m, 0, err)
539
540#define	IFQ_DRV_DEQUEUE(ifq, m)						\
541do {									\
542	(m) = (ifq)->ifq_drv_head;					\
543	if (m) {							\
544		if (((ifq)->ifq_drv_head = (m)->m_nextpkt) == NULL)	\
545			(ifq)->ifq_drv_tail = NULL;			\
546		(m)->m_nextpkt = NULL;					\
547		(ifq)->ifq_drv_len--;					\
548	} else {							\
549		IFQ_LOCK(ifq);						\
550		IFQ_DEQUEUE_NOLOCK(ifq, m);				\
551		while ((ifq)->ifq_drv_len < (ifq)->ifq_drv_maxlen) {	\
552			struct mbuf *m0;				\
553			IFQ_DEQUEUE_NOLOCK(ifq, m0);			\
554			if (m0 == NULL)					\
555				break;					\
556			m0->m_nextpkt = NULL;				\
557			if ((ifq)->ifq_drv_tail == NULL)		\
558				(ifq)->ifq_drv_head = m0;		\
559			else						\
560				(ifq)->ifq_drv_tail->m_nextpkt = m0;	\
561			(ifq)->ifq_drv_tail = m0;			\
562			(ifq)->ifq_drv_len++;				\
563		}							\
564		IFQ_UNLOCK(ifq);					\
565	}								\
566} while (0)
567
568#define	IFQ_DRV_PREPEND(ifq, m)						\
569do {									\
570	(m)->m_nextpkt = (ifq)->ifq_drv_head;				\
571	if ((ifq)->ifq_drv_tail == NULL)				\
572		(ifq)->ifq_drv_tail = (m);				\
573	(ifq)->ifq_drv_head = (m);					\
574	(ifq)->ifq_drv_len++;						\
575} while (0)
576
577#define	IFQ_DRV_IS_EMPTY(ifq)						\
578	(((ifq)->ifq_drv_len == 0) && ((ifq)->ifq_len == 0))
579
580#define	IFQ_DRV_PURGE(ifq)						\
581do {									\
582	struct mbuf *m, *n = (ifq)->ifq_drv_head;			\
583	while((m = n) != NULL) {					\
584		n = m->m_nextpkt;					\
585		m_freem(m);						\
586	}								\
587	(ifq)->ifq_drv_head = (ifq)->ifq_drv_tail = NULL;		\
588	(ifq)->ifq_drv_len = 0;						\
589	IFQ_PURGE(ifq);							\
590} while (0)
591
592#ifdef _KERNEL
593static __inline int
594drbr_enqueue(struct ifnet *ifp, struct buf_ring *br, struct mbuf *m)
595{
596	int error = 0;
597
598#ifdef ALTQ
599	if (ALTQ_IS_ENABLED(&ifp->if_snd)) {
600		IFQ_ENQUEUE(&ifp->if_snd, m, error);
601		return (error);
602	}
603#endif
604	error = buf_ring_enqueue(br, m);
605	if (error)
606		m_freem(m);
607
608	return (error);
609}
610
611static __inline void
612drbr_flush(struct ifnet *ifp, struct buf_ring *br)
613{
614	struct mbuf *m;
615
616#ifdef ALTQ
617	if (ifp != NULL && ALTQ_IS_ENABLED(&ifp->if_snd))
618		IFQ_PURGE(&ifp->if_snd);
619#endif
620	while ((m = buf_ring_dequeue_sc(br)) != NULL)
621		m_freem(m);
622}
623
624static __inline void
625drbr_free(struct buf_ring *br, struct malloc_type *type)
626{
627
628	drbr_flush(NULL, br);
629	buf_ring_free(br, type);
630}
631
632static __inline struct mbuf *
633drbr_dequeue(struct ifnet *ifp, struct buf_ring *br)
634{
635#ifdef ALTQ
636	struct mbuf *m;
637
638	if (ALTQ_IS_ENABLED(&ifp->if_snd)) {
639		IFQ_DEQUEUE(&ifp->if_snd, m);
640		return (m);
641	}
642#endif
643	return (buf_ring_dequeue_sc(br));
644}
645
646static __inline struct mbuf *
647drbr_dequeue_cond(struct ifnet *ifp, struct buf_ring *br,
648    int (*func) (struct mbuf *, void *), void *arg)
649{
650	struct mbuf *m;
651#ifdef ALTQ
652	if (ALTQ_IS_ENABLED(&ifp->if_snd)) {
653		IFQ_LOCK(&ifp->if_snd);
654		IFQ_POLL_NOLOCK(&ifp->if_snd, m);
655		if (m != NULL && func(m, arg) == 0) {
656			IFQ_UNLOCK(&ifp->if_snd);
657			return (NULL);
658		}
659		IFQ_DEQUEUE_NOLOCK(&ifp->if_snd, m);
660		IFQ_UNLOCK(&ifp->if_snd);
661		return (m);
662	}
663#endif
664	m = buf_ring_peek(br);
665	if (m == NULL || func(m, arg) == 0)
666		return (NULL);
667
668	return (buf_ring_dequeue_sc(br));
669}
670
671static __inline int
672drbr_empty(struct ifnet *ifp, struct buf_ring *br)
673{
674#ifdef ALTQ
675	if (ALTQ_IS_ENABLED(&ifp->if_snd))
676		return (IFQ_IS_EMPTY(&ifp->if_snd));
677#endif
678	return (buf_ring_empty(br));
679}
680
681static __inline int
682drbr_needs_enqueue(struct ifnet *ifp, struct buf_ring *br)
683{
684#ifdef ALTQ
685	if (ALTQ_IS_ENABLED(&ifp->if_snd))
686		return (1);
687#endif
688	return (!buf_ring_empty(br));
689}
690
691static __inline int
692drbr_inuse(struct ifnet *ifp, struct buf_ring *br)
693{
694#ifdef ALTQ
695	if (ALTQ_IS_ENABLED(&ifp->if_snd))
696		return (ifp->if_snd.ifq_len);
697#endif
698	return (buf_ring_count(br));
699}
700#endif
701/*
702 * 72 was chosen below because it is the size of a TCP/IP
703 * header (40) + the minimum mss (32).
704 */
705#define	IF_MINMTU	72
706#define	IF_MAXMTU	65535
707
708#define	TOEDEV(ifp)	((ifp)->if_llsoftc)
709
710#endif /* _KERNEL */
711
712/*
713 * The ifaddr structure contains information about one address
714 * of an interface.  They are maintained by the different address families,
715 * are allocated and attached when an address is set, and are linked
716 * together so all addresses for an interface can be located.
717 *
718 * NOTE: a 'struct ifaddr' is always at the beginning of a larger
719 * chunk of malloc'ed memory, where we store the three addresses
720 * (ifa_addr, ifa_dstaddr and ifa_netmask) referenced here.
721 */
722struct ifaddr {
723	struct	sockaddr *ifa_addr;	/* address of interface */
724	struct	sockaddr *ifa_dstaddr;	/* other end of p-to-p link */
725#define	ifa_broadaddr	ifa_dstaddr	/* broadcast address interface */
726	struct	sockaddr *ifa_netmask;	/* used to determine subnet */
727	struct	if_data if_data;	/* not all members are meaningful */
728	struct	ifnet *ifa_ifp;		/* back-pointer to interface */
729	struct	carp_softc *ifa_carp;	/* pointer to CARP data */
730	TAILQ_ENTRY(ifaddr) ifa_link;	/* queue macro glue */
731	void	(*ifa_rtrequest)	/* check or clean routes (+ or -)'d */
732		(int, struct rtentry *, struct rt_addrinfo *);
733	u_short	ifa_flags;		/* mostly rt_flags for cloning */
734	u_int	ifa_refcnt;		/* references to this structure */
735	int	ifa_metric;		/* cost of going out this interface */
736	int (*ifa_claim_addr)		/* check if an addr goes to this if */
737		(struct ifaddr *, struct sockaddr *);
738	struct mtx ifa_mtx;
739};
740#define	IFA_ROUTE	RTF_UP		/* route installed */
741#define IFA_RTSELF	RTF_HOST	/* loopback route to self installed */
742
743/* for compatibility with other BSDs */
744#define	ifa_list	ifa_link
745
746#ifdef _KERNEL
747#define	IFA_LOCK(ifa)		mtx_lock(&(ifa)->ifa_mtx)
748#define	IFA_UNLOCK(ifa)		mtx_unlock(&(ifa)->ifa_mtx)
749
750void	ifa_free(struct ifaddr *ifa);
751void	ifa_init(struct ifaddr *ifa);
752void	ifa_ref(struct ifaddr *ifa);
753#endif
754
755/*
756 * Multicast address structure.  This is analogous to the ifaddr
757 * structure except that it keeps track of multicast addresses.
758 */
759struct ifmultiaddr {
760	TAILQ_ENTRY(ifmultiaddr) ifma_link; /* queue macro glue */
761	struct	sockaddr *ifma_addr; 	/* address this membership is for */
762	struct	sockaddr *ifma_lladdr;	/* link-layer translation, if any */
763	struct	ifnet *ifma_ifp;	/* back-pointer to interface */
764	u_int	ifma_refcount;		/* reference count */
765	void	*ifma_protospec;	/* protocol-specific state, if any */
766	struct	ifmultiaddr *ifma_llifma; /* pointer to ifma for ifma_lladdr */
767};
768
769#ifdef _KERNEL
770
771extern	struct rwlock ifnet_rwlock;
772extern	struct sx ifnet_sxlock;
773
774#define	IFNET_LOCK_INIT() do {						\
775	rw_init_flags(&ifnet_rwlock, "ifnet_rw",  RW_RECURSE);		\
776	sx_init_flags(&ifnet_sxlock, "ifnet_sx",  SX_RECURSE);		\
777} while(0)
778
779#define	IFNET_WLOCK() do {						\
780	sx_xlock(&ifnet_sxlock);					\
781	rw_wlock(&ifnet_rwlock);					\
782} while (0)
783
784#define	IFNET_WUNLOCK() do {						\
785	rw_wunlock(&ifnet_rwlock);					\
786	sx_xunlock(&ifnet_sxlock);					\
787} while (0)
788
789/*
790 * To assert the ifnet lock, you must know not only whether it's for read or
791 * write, but also whether it was acquired with sleep support or not.
792 */
793#define	IFNET_RLOCK_ASSERT()		sx_assert(&ifnet_sxlock, SA_SLOCKED)
794#define	IFNET_RLOCK_NOSLEEP_ASSERT()	rw_assert(&ifnet_rwlock, RA_RLOCKED)
795#define	IFNET_WLOCK_ASSERT() do {					\
796	sx_assert(&ifnet_sxlock, SA_XLOCKED);				\
797	rw_assert(&ifnet_rwlock, RA_WLOCKED);				\
798} while (0)
799
800#define	IFNET_RLOCK()		sx_slock(&ifnet_sxlock)
801#define	IFNET_RLOCK_NOSLEEP()	rw_rlock(&ifnet_rwlock)
802#define	IFNET_RUNLOCK()		sx_sunlock(&ifnet_sxlock)
803#define	IFNET_RUNLOCK_NOSLEEP()	rw_runlock(&ifnet_rwlock)
804
805/*
806 * Look up an ifnet given its index; the _ref variant also acquires a
807 * reference that must be freed using if_rele().  It is almost always a bug
808 * to call ifnet_byindex() instead if ifnet_byindex_ref().
809 */
810struct ifnet	*ifnet_byindex(u_short idx);
811struct ifnet	*ifnet_byindex_locked(u_short idx);
812struct ifnet	*ifnet_byindex_ref(u_short idx);
813
814/*
815 * Given the index, ifaddr_byindex() returns the one and only
816 * link-level ifaddr for the interface. You are not supposed to use
817 * it to traverse the list of addresses associated to the interface.
818 */
819struct ifaddr	*ifaddr_byindex(u_short idx);
820
821VNET_DECLARE(struct ifnethead, ifnet);
822VNET_DECLARE(struct ifgrouphead, ifg_head);
823VNET_DECLARE(int, if_index);
824VNET_DECLARE(struct ifnet *, loif);	/* first loopback interface */
825VNET_DECLARE(int, useloopback);
826
827#define	V_ifnet		VNET(ifnet)
828#define	V_ifg_head	VNET(ifg_head)
829#define	V_if_index	VNET(if_index)
830#define	V_loif		VNET(loif)
831#define	V_useloopback	VNET(useloopback)
832
833extern	int ifqmaxlen;
834
835int	if_addgroup(struct ifnet *, const char *);
836int	if_delgroup(struct ifnet *, const char *);
837int	if_addmulti(struct ifnet *, struct sockaddr *, struct ifmultiaddr **);
838int	if_allmulti(struct ifnet *, int);
839struct	ifnet* if_alloc(u_char);
840void	if_attach(struct ifnet *);
841void	if_dead(struct ifnet *);
842int	if_delmulti(struct ifnet *, struct sockaddr *);
843void	if_delmulti_ifma(struct ifmultiaddr *);
844void	if_detach(struct ifnet *);
845void	if_vmove(struct ifnet *, struct vnet *);
846void	if_purgeaddrs(struct ifnet *);
847void	if_delallmulti(struct ifnet *);
848void	if_down(struct ifnet *);
849struct ifmultiaddr *
850	if_findmulti(struct ifnet *, struct sockaddr *);
851void	if_free(struct ifnet *);
852void	if_initname(struct ifnet *, const char *, int);
853void	if_link_state_change(struct ifnet *, int);
854int	if_printf(struct ifnet *, const char *, ...) __printflike(2, 3);
855void	if_qflush(struct ifnet *);
856void	if_ref(struct ifnet *);
857void	if_rele(struct ifnet *);
858int	if_setlladdr(struct ifnet *, const u_char *, int);
859void	if_up(struct ifnet *);
860int	ifioctl(struct socket *, u_long, caddr_t, struct thread *);
861int	ifpromisc(struct ifnet *, int);
862struct	ifnet *ifunit(const char *);
863struct	ifnet *ifunit_ref(const char *);
864
865void	ifq_init(struct ifaltq *, struct ifnet *ifp);
866void	ifq_delete(struct ifaltq *);
867
868int	ifa_add_loopback_route(struct ifaddr *, struct sockaddr *);
869int	ifa_del_loopback_route(struct ifaddr *, struct sockaddr *);
870
871struct	ifaddr *ifa_ifwithaddr(struct sockaddr *);
872int		ifa_ifwithaddr_check(struct sockaddr *);
873struct	ifaddr *ifa_ifwithbroadaddr(struct sockaddr *);
874struct	ifaddr *ifa_ifwithdstaddr(struct sockaddr *);
875struct	ifaddr *ifa_ifwithnet(struct sockaddr *, int);
876struct	ifaddr *ifa_ifwithroute(int, struct sockaddr *, struct sockaddr *);
877struct	ifaddr *ifa_ifwithroute_fib(int, struct sockaddr *, struct sockaddr *, u_int);
878
879struct	ifaddr *ifaof_ifpforaddr(struct sockaddr *, struct ifnet *);
880
881int	if_simloop(struct ifnet *ifp, struct mbuf *m, int af, int hlen);
882
883typedef	void *if_com_alloc_t(u_char type, struct ifnet *ifp);
884typedef	void if_com_free_t(void *com, u_char type);
885void	if_register_com_alloc(u_char type, if_com_alloc_t *a, if_com_free_t *f);
886void	if_deregister_com_alloc(u_char type);
887
888#define IF_LLADDR(ifp)							\
889    LLADDR((struct sockaddr_dl *)((ifp)->if_addr->ifa_addr))
890
891#ifdef DEVICE_POLLING
892enum poll_cmd {	POLL_ONLY, POLL_AND_CHECK_STATUS };
893
894typedef	int poll_handler_t(struct ifnet *ifp, enum poll_cmd cmd, int count);
895int    ether_poll_register(poll_handler_t *h, struct ifnet *ifp);
896int    ether_poll_deregister(struct ifnet *ifp);
897#endif /* DEVICE_POLLING */
898
899#endif /* _KERNEL */
900
901#endif /* !_NET_IF_VAR_H_ */
902