xref: /freebsd/share/man/man9/pfil.9 (revision 3416500aef140042c64bc149cb1ec6620483bc44)
1.\"	$NetBSD: pfil.9,v 1.22 2003/07/01 13:04:06 wiz Exp $
2.\"
3.\" Copyright (c) 1996 Matthew R. Green
4.\" All rights reserved.
5.\"
6.\" Redistribution and use in source and binary forms, with or without
7.\" modification, are permitted provided that the following conditions
8.\" are met:
9.\" 1. Redistributions of source code must retain the above copyright
10.\"    notice, this list of conditions and the following disclaimer.
11.\" 2. Redistributions in binary form must reproduce the above copyright
12.\"    notice, this list of conditions and the following disclaimer in the
13.\"    documentation and/or other materials provided with the distribution.
14.\" 3. The name of the author may not be used to endorse or promote products
15.\"    derived from this software without specific prior written permission.
16.\"
17.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
18.\" IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
19.\" OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
20.\" IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
21.\" INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
22.\" BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
23.\" LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
24.\" AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
25.\" 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.\" $FreeBSD$
30.\"
31.Dd March 27, 2017
32.Dt PFIL 9
33.Os
34.Sh NAME
35.Nm pfil ,
36.Nm pfil_head_register ,
37.Nm pfil_head_unregister ,
38.Nm pfil_head_get ,
39.Nm pfil_add_hook ,
40.Nm pfil_remove_hook ,
41.Nm pfil_run_hooks ,
42.Nm pfil_rlock ,
43.Nm pfil_runlock ,
44.Nm pfil_wlock ,
45.Nm pfil_wunlock
46.Nd packet filter interface
47.Sh SYNOPSIS
48.In sys/param.h
49.In sys/mbuf.h
50.In net/if.h
51.In net/pfil.h
52.Bd -literal
53typedef int (*pfil_func_t)(void *arg, struct mbuf **mp, struct ifnet *, int dir, struct inpcb);
54.Ft int
55.Fn pfil_head_register "struct pfil_head *head"
56.Ft int
57.Fn pfil_head_unregister "struct pfil_head *head"
58.Ft "struct pfil_head *"
59.Fn pfil_head_get "int af" "u_long dlt"
60.Ft int
61.Fn pfil_add_hook "pfil_func_t" "void *arg" "int flags" "struct pfil_head *"
62.Ft int
63.Fn pfil_remove_hook "pfil_func_t" "void *arg" "int flags" "struct pfil_head *"
64.Ft int
65.Fn pfil_run_hooks "struct pfil_head *head" "struct mbuf **mp" "struct ifnet *" "int dir" "struct inpcb *"
66.Ft void
67.Fn pfil_rlock "struct pfil_head *" "struct rm_priotracker *"
68.Ft void
69.Fn pfil_runlock "struct pfil_head *" "struct rm_priotracker *"
70.Ft void
71.Fn pfil_wlock "struct pfil_head *"
72.Ft void
73.Fn pfil_wunlock "struct pfil_head *"
74.Ed
75.Sh DESCRIPTION
76The
77.Nm
78framework allows for a specified function to be invoked for every
79incoming or outgoing packet for a particular network I/O stream.
80These hooks may be used to implement a firewall or perform packet
81transformations.
82.Pp
83Packet filtering points are registered with
84.Fn pfil_head_register .
85Filtering points are identified by a key
86.Pq Vt "void *"
87and a data link type
88.Pq Vt int
89in the
90.Vt pfil_head
91structure.
92Packet filters use the key and data link type to look up the filtering
93point with which they register themselves.
94The key is unique to the filtering point.
95The data link type is a
96.Xr bpf 4
97DLT constant indicating what kind of header is present on the packet
98at the filtering point.
99Each filtering point uses common per-VNET rmlock by default.
100This can be changed by specifying
101.Vt PFIL_FLAG_PRIVATE_LOCK
102as
103.Vt "flags"
104field in the
105.Vt pfil_head
106structure.
107Note that specifying private lock can break filters sharing the same
108ruleset and/or state between different data link types.
109Filtering points may be unregistered with the
110.Fn pfil_head_unregister
111function.
112.Pp
113Packet filters register/unregister themselves with a filtering point
114with the
115.Fn pfil_add_hook
116and
117.Fn pfil_remove_hook
118functions, respectively.
119The head is looked up using the
120.Fn pfil_head_get
121function, which takes the key and data link type that the packet filter
122expects.
123Filters may provide an argument to be passed to the filter when
124invoked on a packet.
125.Pp
126When a filter is invoked, the packet appears just as if it
127.Dq came off the wire .
128That is, all protocol fields are in network byte order.
129The filter is called with its specified argument, the pointer to the
130pointer to the
131.Vt mbuf
132containing the packet, the pointer to the network
133interface that the packet is traversing, and the direction
134.Dv ( PFIL_IN
135or
136.Dv PFIL_OUT )
137that the packet is traveling.
138The filter may change which mbuf the
139.Vt "mbuf\ **"
140argument references.
141The filter returns an error (errno) if the packet processing is to stop, or 0
142if the processing is to continue.
143If the packet processing is to stop, it is the responsibility of the
144filter to free the packet.
145.Pp
146Every filter hook is called with
147.Nm
148read lock held.
149All heads uses the same lock within the same VNET instance.
150Packet filter can use this lock instead of own locking model to
151improve performance.
152Since
153.Nm
154uses
155.Xr rmlock 9
156.Fn pfil_rlock
157and
158.Fn pfil_runlock
159require
160.Va struct rm_priotracker
161to be passed as argument.
162Filter can acquire and release writer lock via
163.Fn pfil_wlock
164and
165.Fn pfil_wunlock
166functions.
167See
168.Xr rmlock 9
169for more details.
170.Sh FILTERING POINTS
171Currently, filtering points are implemented for the following link types:
172.Pp
173.Bl -tag -width "AF_INET6" -offset XXX -compact
174.It AF_INET
175IPv4 packets.
176.It AF_INET6
177IPv6 packets.
178.It AF_LINK
179Link-layer packets.
180.El
181.Sh RETURN VALUES
182If successful,
183.Fn pfil_head_get
184returns the
185.Vt pfil_head
186structure for the given key/dlt.
187The
188.Fn pfil_add_hook
189and
190.Fn pfil_remove_hook
191functions
192return 0 if successful.
193If called with flag
194.Dv PFIL_WAITOK ,
195.Fn pfil_remove_hook
196is expected to always succeed.
197.Pp
198The
199.Fn pfil_head_unregister
200function
201might sleep!
202.Sh SEE ALSO
203.Xr bpf 4 ,
204.Xr if_bridge 4 ,
205.Xr rmlock 9
206.Sh HISTORY
207The
208.Nm
209interface first appeared in
210.Nx 1.3 .
211The
212.Nm
213input and output lists were originally implemented as
214.In sys/queue.h
215.Dv LIST
216structures;
217however this was changed in
218.Nx 1.4
219to
220.Dv TAILQ
221structures.
222This change was to allow the input and output filters to be processed in
223reverse order, to allow the same path to be taken, in or out of the kernel.
224.Pp
225The
226.Nm
227interface was changed in 1.4T to accept a 3rd parameter to both
228.Fn pfil_add_hook
229and
230.Fn pfil_remove_hook ,
231introducing the capability of per-protocol filtering.
232This was done primarily in order to support filtering of IPv6.
233.Pp
234In 1.5K, the
235.Nm
236framework was changed to work with an arbitrary number of filtering points,
237as well as be less IP-centric.
238.Pp
239Fine-grained locking was added in
240.Fx 5.2 .
241.Nm
242lock export was added in
243.Fx 10.0 .
244.Sh BUGS
245When a
246.Vt pfil_head
247is being modified, no traffic is diverted
248(to avoid deadlock).
249This means that traffic may be dropped unconditionally for a short period
250of time.
251.Fn pfil_run_hooks
252will return
253.Er ENOBUFS
254to indicate this.
255