1.\" Copyright (c) 1983, 1991, 1993
2.\" The Regents of the University of California. All rights reserved.
3.\"
4.\" Redistribution and use in source and binary forms, with or without
5.\" modification, are permitted provided that the following conditions
6.\" are met:
7.\" 1. Redistributions of source code must retain the above copyright
8.\" notice, this list of conditions and the following disclaimer.
9.\" 2. Redistributions in binary form must reproduce the above copyright
10.\" notice, this list of conditions and the following disclaimer in the
11.\" documentation and/or other materials provided with the distribution.
12.\" 4. Neither the name of the University nor the names of its contributors
13.\" may be used to endorse or promote products derived from this software
14.\" without specific prior written permission.
15.\"
16.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
17.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19.\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
20.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26.\" SUCH DAMAGE.
27.\"
28.\" @(#)getsockopt.2 8.4 (Berkeley) 5/2/95
| 1.\" Copyright (c) 1983, 1991, 1993
2.\" The Regents of the University of California. All rights reserved.
3.\"
4.\" Redistribution and use in source and binary forms, with or without
5.\" modification, are permitted provided that the following conditions
6.\" are met:
7.\" 1. Redistributions of source code must retain the above copyright
8.\" notice, this list of conditions and the following disclaimer.
9.\" 2. Redistributions in binary form must reproduce the above copyright
10.\" notice, this list of conditions and the following disclaimer in the
11.\" documentation and/or other materials provided with the distribution.
12.\" 4. Neither the name of the University nor the names of its contributors
13.\" may be used to endorse or promote products derived from this software
14.\" without specific prior written permission.
15.\"
16.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
17.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19.\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
20.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26.\" SUCH DAMAGE.
27.\"
28.\" @(#)getsockopt.2 8.4 (Berkeley) 5/2/95
|
32.Dt GETSOCKOPT 2
33.Os
34.Sh NAME
35.Nm getsockopt ,
36.Nm setsockopt
37.Nd get and set options on sockets
38.Sh LIBRARY
39.Lb libc
40.Sh SYNOPSIS
41.In sys/types.h
42.In sys/socket.h
43.Ft int
44.Fn getsockopt "int s" "int level" "int optname" "void * restrict optval" "socklen_t * restrict optlen"
45.Ft int
46.Fn setsockopt "int s" "int level" "int optname" "const void *optval" "socklen_t optlen"
47.Sh DESCRIPTION
48The
49.Fn getsockopt
50and
51.Fn setsockopt
52system calls
53manipulate the
54.Em options
55associated with a socket.
56Options may exist at multiple
57protocol levels; they are always present at the uppermost
58.Dq socket
59level.
60.Pp
61When manipulating socket options the level at which the
62option resides and the name of the option must be specified.
63To manipulate options at the socket level,
64.Fa level
65is specified as
66.Dv SOL_SOCKET .
67To manipulate options at any
68other level the protocol number of the appropriate protocol
69controlling the option is supplied.
70For example,
71to indicate that an option is to be interpreted by the
72.Tn TCP
73protocol,
74.Fa level
75should be set to the protocol number of
76.Tn TCP ;
77see
78.Xr getprotoent 3 .
79.Pp
80The
81.Fa optval
82and
83.Fa optlen
84arguments
85are used to access option values for
86.Fn setsockopt .
87For
88.Fn getsockopt
89they identify a buffer in which the value for the
90requested option(s) are to be returned.
91For
92.Fn getsockopt ,
93.Fa optlen
94is a value-result argument, initially containing the
95size of the buffer pointed to by
96.Fa optval ,
97and modified on return to indicate the actual size of
98the value returned.
99If no option value is
100to be supplied or returned,
101.Fa optval
102may be NULL.
103.Pp
104The
105.Fa optname
106argument
107and any specified options are passed uninterpreted to the appropriate
108protocol module for interpretation.
109The include file
110.In sys/socket.h
111contains definitions for
112socket level options, described below.
113Options at other protocol levels vary in format and
114name; consult the appropriate entries in
115section
1164 of the manual.
117.Pp
118Most socket-level options utilize an
119.Vt int
120argument for
121.Fa optval .
122For
123.Fn setsockopt ,
124the argument should be non-zero to enable a boolean option,
125or zero if the option is to be disabled.
126.Dv SO_LINGER
127uses a
128.Vt "struct linger"
129argument, defined in
130.In sys/socket.h ,
131which specifies the desired state of the option and the
132linger interval (see below).
133.Dv SO_SNDTIMEO
134and
135.Dv SO_RCVTIMEO
136use a
137.Vt "struct timeval"
138argument, defined in
139.In sys/time.h .
140.Pp
141The following options are recognized at the socket level.
142For protocol-specific options, see protocol manual pages,
143e.g.
144.Xr ip 4
145or
146.Xr tcp 4 .
147Except as noted, each may be examined with
148.Fn getsockopt
149and set with
150.Fn setsockopt .
151.Bl -column SO_ACCEPTFILTER -offset indent
152.It Dv SO_DEBUG Ta "enables recording of debugging information"
153.It Dv SO_REUSEADDR Ta "enables local address reuse"
154.It Dv SO_REUSEPORT Ta "enables duplicate address and port bindings"
155.It Dv SO_KEEPALIVE Ta "enables keep connections alive"
156.It Dv SO_DONTROUTE Ta "enables routing bypass for outgoing messages"
157.It Dv SO_LINGER Ta "linger on close if data present"
158.It Dv SO_BROADCAST Ta "enables permission to transmit broadcast messages"
159.It Dv SO_OOBINLINE Ta "enables reception of out-of-band data in band"
160.It Dv SO_SNDBUF Ta "set buffer size for output"
161.It Dv SO_RCVBUF Ta "set buffer size for input"
162.It Dv SO_SNDLOWAT Ta "set minimum count for output"
163.It Dv SO_RCVLOWAT Ta "set minimum count for input"
164.It Dv SO_SNDTIMEO Ta "set timeout value for output"
165.It Dv SO_RCVTIMEO Ta "set timeout value for input"
166.It Dv SO_ACCEPTFILTER Ta "set accept filter on listening socket"
167.It Dv SO_NOSIGPIPE Ta
168controls generation of
169.Dv SIGPIPE
170for the socket
171.It Dv SO_TIMESTAMP Ta "enables reception of a timestamp with datagrams"
172.It Dv SO_BINTIME Ta "enables reception of a timestamp with datagrams"
173.It Dv SO_ACCEPTCONN Ta "get listening status of the socket (get only)"
174.It Dv SO_TYPE Ta "get the type of the socket (get only)"
175.It Dv SO_ERROR Ta "get and clear error on the socket (get only)"
176.It Dv SO_SETFIB Ta "set the associated FIB (routing table) for the socket (set only)"
177.El
178.Pp
179The following options are recognized in
180.Fx :
181.Bl -column SO_LISTENINCQLEN -offset indent
182.It Dv SO_LABEL Ta "get MAC label of the socket (get only)"
183.It Dv SO_PEERLABEL Ta "get socket's peer's MAC label (get only)"
184.It Dv SO_LISTENQLIMIT Ta "get backlog limit of the socket (get only)"
185.It Dv SO_LISTENQLEN Ta "get complete queue length of the socket (get only)"
186.It Dv SO_LISTENINCQLEN Ta "get incomplete queue length of the socket (get only)"
187.It Dv SO_USER_COOKIE Ta "set the 'so_user_cookie' value for the socket (uint32_t, set only)"
188.El
189.Pp
190.Dv SO_DEBUG
191enables debugging in the underlying protocol modules.
192.Pp
193.Dv SO_REUSEADDR
194indicates that the rules used in validating addresses supplied
195in a
196.Xr bind 2
197system call should allow reuse of local addresses.
198.Pp
199.Dv SO_REUSEPORT
200allows completely duplicate bindings by multiple processes
201if they all set
202.Dv SO_REUSEPORT
203before binding the port.
204This option permits multiple instances of a program to each
205receive UDP/IP multicast or broadcast datagrams destined for the bound port.
206.Pp
207.Dv SO_KEEPALIVE
208enables the
209periodic transmission of messages on a connected socket.
210Should the
211connected party fail to respond to these messages, the connection is
212considered broken and processes using the socket are notified via a
213.Dv SIGPIPE
214signal when attempting to send data.
215.Pp
216.Dv SO_DONTROUTE
217indicates that outgoing messages should
218bypass the standard routing facilities.
219Instead, messages are directed
220to the appropriate network interface according to the network portion
221of the destination address.
222.Pp
223.Dv SO_LINGER
224controls the action taken when unsent messages
225are queued on socket and a
226.Xr close 2
227is performed.
228If the socket promises reliable delivery of data and
229.Dv SO_LINGER
230is set,
231the system will block the process on the
232.Xr close 2
233attempt until it is able to transmit the data or until it decides it
234is unable to deliver the information (a timeout period, termed the
235linger interval, is specified in seconds in the
236.Fn setsockopt
237system call when
238.Dv SO_LINGER
239is requested).
240If
241.Dv SO_LINGER
242is disabled and a
243.Xr close 2
244is issued, the system will process the close in a manner that allows
245the process to continue as quickly as possible.
246.Pp
247The option
248.Dv SO_BROADCAST
249requests permission to send broadcast datagrams
250on the socket.
251Broadcast was a privileged operation in earlier versions of the system.
252.Pp
253With protocols that support out-of-band data, the
254.Dv SO_OOBINLINE
255option
256requests that out-of-band data be placed in the normal data input queue
257as received; it will then be accessible with
258.Xr recv 2
259or
260.Xr read 2
261calls without the
262.Dv MSG_OOB
263flag.
264Some protocols always behave as if this option is set.
265.Pp
266.Dv SO_SNDBUF
267and
268.Dv SO_RCVBUF
269are options to adjust the normal
270buffer sizes allocated for output and input buffers, respectively.
271The buffer size may be increased for high-volume connections,
272or may be decreased to limit the possible backlog of incoming data.
273The system places an absolute maximum on these values, which is accessible
274through the
275.Xr sysctl 3
276MIB variable
277.Dq Li kern.ipc.maxsockbuf .
278.Pp
279.Dv SO_SNDLOWAT
280is an option to set the minimum count for output operations.
281Most output operations process all of the data supplied
282by the call, delivering data to the protocol for transmission
283and blocking as necessary for flow control.
284Nonblocking output operations will process as much data as permitted
285subject to flow control without blocking, but will process no data
286if flow control does not allow the smaller of the low water mark value
287or the entire request to be processed.
288A
289.Xr select 2
290operation testing the ability to write to a socket will return true
291only if the low water mark amount could be processed.
292The default value for
293.Dv SO_SNDLOWAT
294is set to a convenient size for network efficiency, often 1024.
295.Pp
296.Dv SO_RCVLOWAT
297is an option to set the minimum count for input operations.
298In general, receive calls will block until any (non-zero) amount of data
299is received, then return with the smaller of the amount available or the amount
300requested.
301The default value for
302.Dv SO_RCVLOWAT
303is 1.
304If
305.Dv SO_RCVLOWAT
306is set to a larger value, blocking receive calls normally
307wait until they have received the smaller of the low water mark value
308or the requested amount.
309Receive calls may still return less than the low water mark if an error
310occurs, a signal is caught, or the type of data next in the receive queue
311is different from that which was returned.
312.Pp
313.Dv SO_SNDTIMEO
314is an option to set a timeout value for output operations.
315It accepts a
316.Vt "struct timeval"
317argument with the number of seconds and microseconds
318used to limit waits for output operations to complete.
319If a send operation has blocked for this much time,
320it returns with a partial count
321or with the error
322.Er EWOULDBLOCK
323if no data were sent.
324In the current implementation, this timer is restarted each time additional
325data are delivered to the protocol,
326implying that the limit applies to output portions ranging in size
327from the low water mark to the high water mark for output.
328.Pp
329.Dv SO_RCVTIMEO
330is an option to set a timeout value for input operations.
331It accepts a
332.Vt "struct timeval"
333argument with the number of seconds and microseconds
334used to limit waits for input operations to complete.
335In the current implementation, this timer is restarted each time additional
336data are received by the protocol,
337and thus the limit is in effect an inactivity timer.
338If a receive operation has been blocked for this much time without
339receiving additional data, it returns with a short count
340or with the error
341.Er EWOULDBLOCK
342if no data were received.
343.Pp
344.Dv SO_SETFIB
345can be used to over-ride the default FIB (routing table) for the given socket.
346The value must be from 0 to one less than the number returned from
347the sysctl
348.Em net.fibs .
349.Pp
350.Dv SO_USER_COOKIE
351can be used to set the uint32_t so_user_cookie field in the socket.
352The value is an uint32_t, and can be used in the kernel code that
353manipulates traffic related to the socket.
354The default value for the field is 0.
355As an example, the value can be used as the skipto target or
356pipe number in
357.Nm ipfw/dummynet .
358.Pp
359.Dv SO_ACCEPTFILTER
360places an
361.Xr accept_filter 9
362on the socket,
363which will filter incoming connections
364on a listening stream socket before being presented for
365.Xr accept 2 .
366Once more,
367.Xr listen 2
368must be called on the socket before
369trying to install the filter on it,
370or else the
371.Fn setsockopt
372system call will fail.
373.Bd -literal
374struct accept_filter_arg {
375 char af_name[16];
376 char af_arg[256-16];
377};
378.Ed
379.Pp
380The
381.Fa optval
382argument
383should point to a
384.Fa struct accept_filter_arg
385that will select and configure the
386.Xr accept_filter 9 .
387The
388.Fa af_name
389argument
390should be filled with the name of the accept filter
391that the application wishes to place on the listening socket.
392The optional argument
393.Fa af_arg
394can be passed to the accept
395filter specified by
396.Fa af_name
397to provide additional configuration options at attach time.
398Passing in an
399.Fa optval
400of NULL will remove the filter.
401.Pp
402The
403.Dv SO_NOSIGPIPE
404option controls generation of the
405.Dv SIGPIPE
406signal normally sent
407when writing to a connected socket where the other end has been
408closed returns with the error
409.Er EPIPE .
410.Pp
411If the
412.Dv SO_TIMESTAMP
413or
414.Dv SO_BINTIME
415option is enabled on a
416.Dv SOCK_DGRAM
417socket, the
418.Xr recvmsg 2
419call will return a timestamp corresponding to when the datagram was received.
420The
421.Va msg_control
422field in the
423.Vt msghdr
424structure points to a buffer that contains a
425.Vt cmsghdr
426structure followed by a
427.Vt "struct timeval"
428for
429.Dv SO_TIMESTAMP
430and
431.Vt "struct bintime"
432for
433.Dv SO_BINTIME .
434The
435.Vt cmsghdr
436fields have the following values for TIMESTAMP:
437.Bd -literal
438 cmsg_len = sizeof(struct timeval);
439 cmsg_level = SOL_SOCKET;
440 cmsg_type = SCM_TIMESTAMP;
441.Ed
442.Pp
443and for
444.Dv SO_BINTIME :
445.Bd -literal
446 cmsg_len = sizeof(struct bintime);
447 cmsg_level = SOL_SOCKET;
448 cmsg_type = SCM_BINTIME;
449.Ed
450.Pp
451.Dv SO_ACCEPTCONN ,
452.Dv SO_TYPE
453and
454.Dv SO_ERROR
455are options used only with
456.Fn getsockopt .
457.Dv SO_ACCEPTCONN
458returns whether the socket is currently accepting connections,
459that is, whether or not the
460.Xr listen 2
461system call was invoked on the socket.
462.Dv SO_TYPE
463returns the type of the socket, such as
464.Dv SOCK_STREAM ;
465it is useful for servers that inherit sockets on startup.
466.Dv SO_ERROR
467returns any pending error on the socket and clears
468the error status.
469It may be used to check for asynchronous errors on connected
470datagram sockets or for other asynchronous errors.
471.Pp
472Finally,
473.Dv SO_LABEL
474returns the MAC label of the socket.
475.Dv SO_PEERLABEL
476returns the MAC label of the socket's peer.
477Note that your kernel must be compiled with MAC support.
478See
479.Xr mac 3
480for more information.
481.Dv SO_LISTENQLIMIT
482returns the maximal number of queued connections, as set by
483.Xr listen 2 .
484.Dv SO_LISTENQLEN
485returns the number of unaccepted complete connections.
486.Dv SO_LISTENINCQLEN
487returns the number of unaccepted incomplete connections.
488.Sh RETURN VALUES
489.Rv -std
490.Sh ERRORS
491The call succeeds unless:
492.Bl -tag -width Er
493.It Bq Er EBADF
494The argument
495.Fa s
496is not a valid descriptor.
497.It Bq Er ENOTSOCK
498The argument
499.Fa s
500is a file, not a socket.
501.It Bq Er ENOPROTOOPT
502The option is unknown at the level indicated.
503.It Bq Er EFAULT
504The address pointed to by
505.Fa optval
506is not in a valid part of the process address space.
507For
508.Fn getsockopt ,
509this error may also be returned if
510.Fa optlen
511is not in a valid part of the process address space.
512.It Bq Er EINVAL
513Installing an
514.Xr accept_filter 9
515on a non-listening socket was attempted.
516.El
517.Sh SEE ALSO
518.Xr ioctl 2 ,
519.Xr listen 2 ,
520.Xr recvmsg 2 ,
521.Xr socket 2 ,
522.Xr getprotoent 3 ,
523.Xr mac 3 ,
524.Xr sysctl 3 ,
525.Xr ip 4 ,
526.Xr ip6 4 ,
527.Xr sctp 4 ,
528.Xr tcp 4 ,
529.Xr protocols 5 ,
530.Xr sysctl 8 ,
531.Xr accept_filter 9 ,
532.Xr bintime 9
533.Sh HISTORY
534The
535.Fn getsockopt
| 32.Dt GETSOCKOPT 2
33.Os
34.Sh NAME
35.Nm getsockopt ,
36.Nm setsockopt
37.Nd get and set options on sockets
38.Sh LIBRARY
39.Lb libc
40.Sh SYNOPSIS
41.In sys/types.h
42.In sys/socket.h
43.Ft int
44.Fn getsockopt "int s" "int level" "int optname" "void * restrict optval" "socklen_t * restrict optlen"
45.Ft int
46.Fn setsockopt "int s" "int level" "int optname" "const void *optval" "socklen_t optlen"
47.Sh DESCRIPTION
48The
49.Fn getsockopt
50and
51.Fn setsockopt
52system calls
53manipulate the
54.Em options
55associated with a socket.
56Options may exist at multiple
57protocol levels; they are always present at the uppermost
58.Dq socket
59level.
60.Pp
61When manipulating socket options the level at which the
62option resides and the name of the option must be specified.
63To manipulate options at the socket level,
64.Fa level
65is specified as
66.Dv SOL_SOCKET .
67To manipulate options at any
68other level the protocol number of the appropriate protocol
69controlling the option is supplied.
70For example,
71to indicate that an option is to be interpreted by the
72.Tn TCP
73protocol,
74.Fa level
75should be set to the protocol number of
76.Tn TCP ;
77see
78.Xr getprotoent 3 .
79.Pp
80The
81.Fa optval
82and
83.Fa optlen
84arguments
85are used to access option values for
86.Fn setsockopt .
87For
88.Fn getsockopt
89they identify a buffer in which the value for the
90requested option(s) are to be returned.
91For
92.Fn getsockopt ,
93.Fa optlen
94is a value-result argument, initially containing the
95size of the buffer pointed to by
96.Fa optval ,
97and modified on return to indicate the actual size of
98the value returned.
99If no option value is
100to be supplied or returned,
101.Fa optval
102may be NULL.
103.Pp
104The
105.Fa optname
106argument
107and any specified options are passed uninterpreted to the appropriate
108protocol module for interpretation.
109The include file
110.In sys/socket.h
111contains definitions for
112socket level options, described below.
113Options at other protocol levels vary in format and
114name; consult the appropriate entries in
115section
1164 of the manual.
117.Pp
118Most socket-level options utilize an
119.Vt int
120argument for
121.Fa optval .
122For
123.Fn setsockopt ,
124the argument should be non-zero to enable a boolean option,
125or zero if the option is to be disabled.
126.Dv SO_LINGER
127uses a
128.Vt "struct linger"
129argument, defined in
130.In sys/socket.h ,
131which specifies the desired state of the option and the
132linger interval (see below).
133.Dv SO_SNDTIMEO
134and
135.Dv SO_RCVTIMEO
136use a
137.Vt "struct timeval"
138argument, defined in
139.In sys/time.h .
140.Pp
141The following options are recognized at the socket level.
142For protocol-specific options, see protocol manual pages,
143e.g.
144.Xr ip 4
145or
146.Xr tcp 4 .
147Except as noted, each may be examined with
148.Fn getsockopt
149and set with
150.Fn setsockopt .
151.Bl -column SO_ACCEPTFILTER -offset indent
152.It Dv SO_DEBUG Ta "enables recording of debugging information"
153.It Dv SO_REUSEADDR Ta "enables local address reuse"
154.It Dv SO_REUSEPORT Ta "enables duplicate address and port bindings"
155.It Dv SO_KEEPALIVE Ta "enables keep connections alive"
156.It Dv SO_DONTROUTE Ta "enables routing bypass for outgoing messages"
157.It Dv SO_LINGER Ta "linger on close if data present"
158.It Dv SO_BROADCAST Ta "enables permission to transmit broadcast messages"
159.It Dv SO_OOBINLINE Ta "enables reception of out-of-band data in band"
160.It Dv SO_SNDBUF Ta "set buffer size for output"
161.It Dv SO_RCVBUF Ta "set buffer size for input"
162.It Dv SO_SNDLOWAT Ta "set minimum count for output"
163.It Dv SO_RCVLOWAT Ta "set minimum count for input"
164.It Dv SO_SNDTIMEO Ta "set timeout value for output"
165.It Dv SO_RCVTIMEO Ta "set timeout value for input"
166.It Dv SO_ACCEPTFILTER Ta "set accept filter on listening socket"
167.It Dv SO_NOSIGPIPE Ta
168controls generation of
169.Dv SIGPIPE
170for the socket
171.It Dv SO_TIMESTAMP Ta "enables reception of a timestamp with datagrams"
172.It Dv SO_BINTIME Ta "enables reception of a timestamp with datagrams"
173.It Dv SO_ACCEPTCONN Ta "get listening status of the socket (get only)"
174.It Dv SO_TYPE Ta "get the type of the socket (get only)"
175.It Dv SO_ERROR Ta "get and clear error on the socket (get only)"
176.It Dv SO_SETFIB Ta "set the associated FIB (routing table) for the socket (set only)"
177.El
178.Pp
179The following options are recognized in
180.Fx :
181.Bl -column SO_LISTENINCQLEN -offset indent
182.It Dv SO_LABEL Ta "get MAC label of the socket (get only)"
183.It Dv SO_PEERLABEL Ta "get socket's peer's MAC label (get only)"
184.It Dv SO_LISTENQLIMIT Ta "get backlog limit of the socket (get only)"
185.It Dv SO_LISTENQLEN Ta "get complete queue length of the socket (get only)"
186.It Dv SO_LISTENINCQLEN Ta "get incomplete queue length of the socket (get only)"
187.It Dv SO_USER_COOKIE Ta "set the 'so_user_cookie' value for the socket (uint32_t, set only)"
188.El
189.Pp
190.Dv SO_DEBUG
191enables debugging in the underlying protocol modules.
192.Pp
193.Dv SO_REUSEADDR
194indicates that the rules used in validating addresses supplied
195in a
196.Xr bind 2
197system call should allow reuse of local addresses.
198.Pp
199.Dv SO_REUSEPORT
200allows completely duplicate bindings by multiple processes
201if they all set
202.Dv SO_REUSEPORT
203before binding the port.
204This option permits multiple instances of a program to each
205receive UDP/IP multicast or broadcast datagrams destined for the bound port.
206.Pp
207.Dv SO_KEEPALIVE
208enables the
209periodic transmission of messages on a connected socket.
210Should the
211connected party fail to respond to these messages, the connection is
212considered broken and processes using the socket are notified via a
213.Dv SIGPIPE
214signal when attempting to send data.
215.Pp
216.Dv SO_DONTROUTE
217indicates that outgoing messages should
218bypass the standard routing facilities.
219Instead, messages are directed
220to the appropriate network interface according to the network portion
221of the destination address.
222.Pp
223.Dv SO_LINGER
224controls the action taken when unsent messages
225are queued on socket and a
226.Xr close 2
227is performed.
228If the socket promises reliable delivery of data and
229.Dv SO_LINGER
230is set,
231the system will block the process on the
232.Xr close 2
233attempt until it is able to transmit the data or until it decides it
234is unable to deliver the information (a timeout period, termed the
235linger interval, is specified in seconds in the
236.Fn setsockopt
237system call when
238.Dv SO_LINGER
239is requested).
240If
241.Dv SO_LINGER
242is disabled and a
243.Xr close 2
244is issued, the system will process the close in a manner that allows
245the process to continue as quickly as possible.
246.Pp
247The option
248.Dv SO_BROADCAST
249requests permission to send broadcast datagrams
250on the socket.
251Broadcast was a privileged operation in earlier versions of the system.
252.Pp
253With protocols that support out-of-band data, the
254.Dv SO_OOBINLINE
255option
256requests that out-of-band data be placed in the normal data input queue
257as received; it will then be accessible with
258.Xr recv 2
259or
260.Xr read 2
261calls without the
262.Dv MSG_OOB
263flag.
264Some protocols always behave as if this option is set.
265.Pp
266.Dv SO_SNDBUF
267and
268.Dv SO_RCVBUF
269are options to adjust the normal
270buffer sizes allocated for output and input buffers, respectively.
271The buffer size may be increased for high-volume connections,
272or may be decreased to limit the possible backlog of incoming data.
273The system places an absolute maximum on these values, which is accessible
274through the
275.Xr sysctl 3
276MIB variable
277.Dq Li kern.ipc.maxsockbuf .
278.Pp
279.Dv SO_SNDLOWAT
280is an option to set the minimum count for output operations.
281Most output operations process all of the data supplied
282by the call, delivering data to the protocol for transmission
283and blocking as necessary for flow control.
284Nonblocking output operations will process as much data as permitted
285subject to flow control without blocking, but will process no data
286if flow control does not allow the smaller of the low water mark value
287or the entire request to be processed.
288A
289.Xr select 2
290operation testing the ability to write to a socket will return true
291only if the low water mark amount could be processed.
292The default value for
293.Dv SO_SNDLOWAT
294is set to a convenient size for network efficiency, often 1024.
295.Pp
296.Dv SO_RCVLOWAT
297is an option to set the minimum count for input operations.
298In general, receive calls will block until any (non-zero) amount of data
299is received, then return with the smaller of the amount available or the amount
300requested.
301The default value for
302.Dv SO_RCVLOWAT
303is 1.
304If
305.Dv SO_RCVLOWAT
306is set to a larger value, blocking receive calls normally
307wait until they have received the smaller of the low water mark value
308or the requested amount.
309Receive calls may still return less than the low water mark if an error
310occurs, a signal is caught, or the type of data next in the receive queue
311is different from that which was returned.
312.Pp
313.Dv SO_SNDTIMEO
314is an option to set a timeout value for output operations.
315It accepts a
316.Vt "struct timeval"
317argument with the number of seconds and microseconds
318used to limit waits for output operations to complete.
319If a send operation has blocked for this much time,
320it returns with a partial count
321or with the error
322.Er EWOULDBLOCK
323if no data were sent.
324In the current implementation, this timer is restarted each time additional
325data are delivered to the protocol,
326implying that the limit applies to output portions ranging in size
327from the low water mark to the high water mark for output.
328.Pp
329.Dv SO_RCVTIMEO
330is an option to set a timeout value for input operations.
331It accepts a
332.Vt "struct timeval"
333argument with the number of seconds and microseconds
334used to limit waits for input operations to complete.
335In the current implementation, this timer is restarted each time additional
336data are received by the protocol,
337and thus the limit is in effect an inactivity timer.
338If a receive operation has been blocked for this much time without
339receiving additional data, it returns with a short count
340or with the error
341.Er EWOULDBLOCK
342if no data were received.
343.Pp
344.Dv SO_SETFIB
345can be used to over-ride the default FIB (routing table) for the given socket.
346The value must be from 0 to one less than the number returned from
347the sysctl
348.Em net.fibs .
349.Pp
350.Dv SO_USER_COOKIE
351can be used to set the uint32_t so_user_cookie field in the socket.
352The value is an uint32_t, and can be used in the kernel code that
353manipulates traffic related to the socket.
354The default value for the field is 0.
355As an example, the value can be used as the skipto target or
356pipe number in
357.Nm ipfw/dummynet .
358.Pp
359.Dv SO_ACCEPTFILTER
360places an
361.Xr accept_filter 9
362on the socket,
363which will filter incoming connections
364on a listening stream socket before being presented for
365.Xr accept 2 .
366Once more,
367.Xr listen 2
368must be called on the socket before
369trying to install the filter on it,
370or else the
371.Fn setsockopt
372system call will fail.
373.Bd -literal
374struct accept_filter_arg {
375 char af_name[16];
376 char af_arg[256-16];
377};
378.Ed
379.Pp
380The
381.Fa optval
382argument
383should point to a
384.Fa struct accept_filter_arg
385that will select and configure the
386.Xr accept_filter 9 .
387The
388.Fa af_name
389argument
390should be filled with the name of the accept filter
391that the application wishes to place on the listening socket.
392The optional argument
393.Fa af_arg
394can be passed to the accept
395filter specified by
396.Fa af_name
397to provide additional configuration options at attach time.
398Passing in an
399.Fa optval
400of NULL will remove the filter.
401.Pp
402The
403.Dv SO_NOSIGPIPE
404option controls generation of the
405.Dv SIGPIPE
406signal normally sent
407when writing to a connected socket where the other end has been
408closed returns with the error
409.Er EPIPE .
410.Pp
411If the
412.Dv SO_TIMESTAMP
413or
414.Dv SO_BINTIME
415option is enabled on a
416.Dv SOCK_DGRAM
417socket, the
418.Xr recvmsg 2
419call will return a timestamp corresponding to when the datagram was received.
420The
421.Va msg_control
422field in the
423.Vt msghdr
424structure points to a buffer that contains a
425.Vt cmsghdr
426structure followed by a
427.Vt "struct timeval"
428for
429.Dv SO_TIMESTAMP
430and
431.Vt "struct bintime"
432for
433.Dv SO_BINTIME .
434The
435.Vt cmsghdr
436fields have the following values for TIMESTAMP:
437.Bd -literal
438 cmsg_len = sizeof(struct timeval);
439 cmsg_level = SOL_SOCKET;
440 cmsg_type = SCM_TIMESTAMP;
441.Ed
442.Pp
443and for
444.Dv SO_BINTIME :
445.Bd -literal
446 cmsg_len = sizeof(struct bintime);
447 cmsg_level = SOL_SOCKET;
448 cmsg_type = SCM_BINTIME;
449.Ed
450.Pp
451.Dv SO_ACCEPTCONN ,
452.Dv SO_TYPE
453and
454.Dv SO_ERROR
455are options used only with
456.Fn getsockopt .
457.Dv SO_ACCEPTCONN
458returns whether the socket is currently accepting connections,
459that is, whether or not the
460.Xr listen 2
461system call was invoked on the socket.
462.Dv SO_TYPE
463returns the type of the socket, such as
464.Dv SOCK_STREAM ;
465it is useful for servers that inherit sockets on startup.
466.Dv SO_ERROR
467returns any pending error on the socket and clears
468the error status.
469It may be used to check for asynchronous errors on connected
470datagram sockets or for other asynchronous errors.
471.Pp
472Finally,
473.Dv SO_LABEL
474returns the MAC label of the socket.
475.Dv SO_PEERLABEL
476returns the MAC label of the socket's peer.
477Note that your kernel must be compiled with MAC support.
478See
479.Xr mac 3
480for more information.
481.Dv SO_LISTENQLIMIT
482returns the maximal number of queued connections, as set by
483.Xr listen 2 .
484.Dv SO_LISTENQLEN
485returns the number of unaccepted complete connections.
486.Dv SO_LISTENINCQLEN
487returns the number of unaccepted incomplete connections.
488.Sh RETURN VALUES
489.Rv -std
490.Sh ERRORS
491The call succeeds unless:
492.Bl -tag -width Er
493.It Bq Er EBADF
494The argument
495.Fa s
496is not a valid descriptor.
497.It Bq Er ENOTSOCK
498The argument
499.Fa s
500is a file, not a socket.
501.It Bq Er ENOPROTOOPT
502The option is unknown at the level indicated.
503.It Bq Er EFAULT
504The address pointed to by
505.Fa optval
506is not in a valid part of the process address space.
507For
508.Fn getsockopt ,
509this error may also be returned if
510.Fa optlen
511is not in a valid part of the process address space.
512.It Bq Er EINVAL
513Installing an
514.Xr accept_filter 9
515on a non-listening socket was attempted.
516.El
517.Sh SEE ALSO
518.Xr ioctl 2 ,
519.Xr listen 2 ,
520.Xr recvmsg 2 ,
521.Xr socket 2 ,
522.Xr getprotoent 3 ,
523.Xr mac 3 ,
524.Xr sysctl 3 ,
525.Xr ip 4 ,
526.Xr ip6 4 ,
527.Xr sctp 4 ,
528.Xr tcp 4 ,
529.Xr protocols 5 ,
530.Xr sysctl 8 ,
531.Xr accept_filter 9 ,
532.Xr bintime 9
533.Sh HISTORY
534The
535.Fn getsockopt
|