1 .\" This man page is Copyright (C) 1999 Andi Kleen <ak@muc.de>.
2 .\" and Copyright (C) 2008-2014, Michael Kerrisk <mtk.manpages@gmail.com>
4 .\" %%%LICENSE_START(VERBATIM_ONE_PARA)
5 .\" and Copyright (C) 2008, 2012 Michael Kerrisk <mtk.manpages@gmail.com>
6 .\" Permission is granted to distribute possibly modified copies
7 .\" of this page provided the header is included verbatim,
8 .\" and in case of nontrivial modification author and date
9 .\" of the modification is added to the header.
12 .\" Modified, 2003-12-02, Michael Kerrisk, <mtk.manpages@gmail.com>
13 .\" Modified, 2003-09-23, Adam Langley
14 .\" Modified, 2004-05-27, Michael Kerrisk, <mtk.manpages@gmail.com>
15 .\" Added SOCK_SEQPACKET
16 .\" 2008-05-27, mtk, Provide a clear description of the three types of
17 .\" address that can appear in the sockaddr_un structure: pathname,
18 .\" unnamed, and abstract.
20 .TH UNIX 7 2015-02-21 "Linux" "Linux Programmer's Manual"
22 unix \- sockets for local interprocess communication
24 .B #include <sys/socket.h>
26 .B #include <sys/un.h>
28 .IB unix_socket " = socket(AF_UNIX, type, 0);"
30 .IB error " = socketpair(AF_UNIX, type, 0, int *" sv ");"
36 socket family is used to communicate between processes on the same machine
38 Traditionally, UNIX domain sockets can be either unnamed,
39 or bound to a filesystem pathname (marked as being of type socket).
40 Linux also supports an abstract namespace which is independent of the
43 Valid socket types in the UNIX domain are:
45 for a stream-oriented socket;
47 for a datagram-oriented socket that preserves message boundaries
48 (as on most UNIX implementations, UNIX domain datagram
49 sockets are always reliable and don't reorder datagrams);
50 and (since Linux 2.6.4)
52 for a connection-oriented socket that preserves message boundaries
53 and delivers messages in the order that they were sent.
55 UNIX domain sockets support passing file descriptors or process credentials
56 to other processes using ancillary data.
58 A UNIX domain socket address is represented in the following structure:
62 .\" #define UNIX_PATH_MAX 108
65 sa_family_t sun_family; /* AF_UNIX */
66 char sun_path[108]; /* pathname */
77 is 108 bytes in size; see also NOTES, below.
79 Various systems calls (for example,
87 Some other system calls (for example,
93 return an argument of this type.
95 Three types of address are distinguished in the
100 a UNIX domain socket can be bound to a null-terminated
101 filesystem pathname using
103 When the address of a pathname socket is returned
104 (by one of the system calls noted above),
107 offsetof(struct sockaddr_un, sun_path) + strlen(sun_path) + 1
111 contains the null-terminated pathname.
114 expression equates to the same value as
115 .IR sizeof(sa_family_t) ,
116 but some other implementations include other fields before
120 expression more portably describes the size of the address structure.)
122 For further details of pathname sockets, see below.
125 A stream socket that has not been bound to a pathname using
128 Likewise, the two sockets created by
131 When the address of an unnamed socket is returned,
133 .IR "sizeof(sa_family_t)" ,
136 should not be inspected.
137 .\" There is quite some variation across implementations: FreeBSD
138 .\" says the length is 16 bytes, HP-UX 11 says it's zero bytes.
141 an abstract socket address is distinguished (from a pathname socket)
144 is a null byte (\(aq\\0\(aq).
145 The socket's address in this namespace is given by the additional
148 that are covered by the specified length of the address structure.
149 (Null bytes in the name have no special significance.)
150 The name has no connection with filesystem pathnames.
151 When the address of an abstract socket is returned,
155 .IR "sizeof(sa_family_t)"
156 (i.e., greater than 2), and the name of the socket is contained in
158 .IR "(addrlen \- sizeof(sa_family_t))"
161 The abstract socket namespace is a nonportable Linux extension.
163 When binding a socket to a pathname, a few rules should be observed
164 for maximum portability and ease of coding:
168 should be null-terminated.
170 The length of the pathname, including the terminating null byte,
171 should not exceed the size of
176 argument that describes the enclosing
178 structure should have a value of at least:
181 offsetof(struct sockaddr_un, sun_path)+strlen(addr.sun_path)+1
187 .IR "sizeof(struct sockaddr_un)" .
189 There is some variation in how implementations handle UNIX domain
190 socket addresses that do not follow the above rules.
191 For example, some (but not all) implementations
192 .\" Linux does this, including for the case where the supplied path
194 append a null terminator if none is present in the supplied
197 When coding portable applications,
198 keep in mind that some implementations
202 as short as 92 bytes.
203 .\" Modern BSDs generally have 104, Tru64 and AIX have 104,
204 .\" Solaris and Irix have 108
211 return socket address structures.
212 When applied to UNIX domain sockets, the value-result
214 argument supplied to the call should be initialized as above.
215 Upon return, the argument is set to indicate the
217 size of the address structure.
218 The caller should check the value returned in this argument:
219 if the output value exceeds the input value,
220 then there is no guarantee that a null terminator is present in
224 For historical reasons, these socket options are specified with a
226 type even though they are
235 as the socket family.
238 Enables the receiving of the credentials of the sending process in an
240 When this option is set and the socket is not yet connected
241 a unique name in the abstract namespace will be generated automatically.
242 Expects an integer boolean flag.
249 .IR sizeof(sa_family_t) ,
250 .\" i.e., sizeof(short)
253 socket option was specified for a socket that was
254 not explicitly bound to an address,
255 then the socket is autobound to an abstract address.
256 The address consists of a null byte
257 followed by 5 bytes in the character set
259 Thus, there is a limit of 2^20 autobind addresses.
260 (From Linux 2.1.15, when the autobind feature was added,
261 8 bytes were used, and the limit was thus 2^32 autobind addresses.
262 The change to 5 bytes came in Linux 2.3.15.)
264 The following paragraphs describe domain-specific details and
265 unsupported features of the sockets API for UNIX domain sockets on Linux.
267 UNIX domain sockets do not support the transmission of
268 out-of-band data (the
278 flag is not supported by UNIX domain sockets.
286 is not supported by UNIX domain sockets.
290 socket option does have an effect for UNIX domain sockets, but the
293 For datagram sockets, the
295 value imposes an upper limit on the size of outgoing datagrams.
296 This limit is calculated as the doubled (see
298 option value less 32 bytes used for overhead.
299 .SS Ancillary messages
300 Ancillary data is sent and received using
304 For historical reasons the ancillary message types listed below
307 type even though they are
319 For more information see
323 Send or receive a set of open file descriptors from another process.
324 The data portion contains an integer array of the file descriptors.
325 The passed file descriptors behave as though they have been created with
329 Send or receive UNIX credentials.
330 This can be used for authentication.
331 The credentials are passed as a
334 Thus structure is defined in
341 pid_t pid; /* process ID of the sending process */
342 uid_t uid; /* user ID of the sending process */
343 gid_t gid; /* group ID of the sending process */
350 feature test macro must be defined (before including
352 header files) in order to obtain the definition
355 The credentials which the sender specifies are checked by the kernel.
356 A process with effective user ID 0 is allowed to specify values that do
358 The sender must specify its own process ID (unless it has the capability
360 its user ID, effective user ID, or saved set-user-ID (unless it has
362 and its group ID, effective group ID, or saved set-group-ID
369 option must be enabled on the socket.
373 calls return information in
375 The correct syntax is:
380 .IB error " = ioctl(" unix_socket ", " ioctl_type ", &" value ");"
390 socket the function returns the amount of queued unread data in the receive buffer.
391 The socket must not be in LISTEN state, otherwise an error
396 .IR <linux/sockios.h> .
397 .\" FIXME . http://sources.redhat.com/bugzilla/show_bug.cgi?id=12002,
398 .\" filed 2010-09-10, may cause SIOCINQ to be defined in glibc headers
400 you can use the synonymous
404 .\" SIOCOUTQ also has an effect for UNIX domain sockets, but not
405 .\" quite what userland might expect. It seems to return the number
406 .\" of bytes allocated for buffers containing pending output.
407 .\" That number is normally larger than the number of bytes of pending
408 .\" output. Since this info is, from userland's point of view, imprecise,
409 .\" and it may well change, probably best not to document this now.
413 the returned value is the same as
414 for Internet domain datagram socket;
420 The specified local address is already in use or the filesystem socket
421 object already exists.
424 The remote address specified by
426 was not a listening socket.
427 This error can also occur if the target pathname is not a socket.
430 Remote socket was unexpectedly closed.
433 User memory address was not valid.
436 Invalid argument passed.
437 A common cause is that the value
439 was not specified in the
441 field of passed addresses, or the socket was in an
442 invalid state for the applied operation.
446 called on an already connected socket or a target address was
447 specified on a connected socket.
450 The pathname in the remote address specified to
458 Socket operation needs a target address, but the socket is not connected.
461 Stream operation called on non-stream oriented socket or tried to
462 use the out-of-band data option.
465 The sender passed invalid credentials in the
469 Remote socket was closed on a stream socket.
473 This can be avoided by passing the
481 Passed protocol is not
485 Remote socket does not match the local socket type
493 Other errors can be generated by the generic socket layer or
494 by the filesystem while generating a filesystem socket object.
495 See the appropriate manual pages for more information.
498 and the abstract namespace were introduced with Linux 2.2 and should not
499 be used in portable programs.
500 (Some BSD-derived systems also support credential passing,
501 but the implementation details differ.)
503 In the Linux implementation, sockets which are visible in the
504 filesystem honor the permissions of the directory they are in.
505 Their owner, group, and permissions can be changed.
506 Creation of a new socket will fail if the process does not have write and
507 search (execute) permission on the directory the socket is created in.
508 Connecting to the socket object requires read/write permission.
509 This behavior differs from many BSD-derived systems which
510 ignore permissions for UNIX domain sockets.
511 Portable programs should not rely on
512 this feature for security.
514 Binding to a socket with a filename creates a socket
515 in the filesystem that must be deleted by the caller when it is no
518 The usual UNIX close-behind semantics apply; the socket can be unlinked
519 at any time and will be finally removed from the filesystem when the last
520 reference to it is closed.
522 To pass file descriptors or credentials over a
525 to send or receive at least one byte of nonancillary data in the same
531 UNIX domain stream sockets do not support the notion of out-of-band data.
534 When binding a socket to an address,
535 Linux is one of the implementations that appends a null terminator
536 if none is supplied in
538 In most cases this is unproblematic:
539 when the socket address is retrieved,
540 it will be one byte longer than that supplied when the socket was bound.
541 However, there is one case where confusing behavior can result:
542 if 108 non-null bytes are supplied when a socket is bound,
543 then the addition of the null terminator takes the length of
545 .IR sizeof(sun_path) .
546 Consequently, when retrieving the socket address
549 .\" The behavior on Solaris is quite similar.
552 argument for the retrieving call is specified as
553 .IR "sizeof(struct sockaddr_un)" ,
554 then the returned address structure
556 have a null terminator in
559 In addition, some implementations
560 .\" i.e., traditional BSD
561 don't require a null terminator when binding a socket (the
563 argument is used to determine the length of
565 and when the socket address is retrieved on these implementations,
566 there is no null terminator in
569 Applications that retrieve socket addresses can (portably) code
570 to handle the possibility that there is no null terminator in
572 by respecting the fact that the number of valid bytes in the pathname is:
574 strnlen(addr.sun_path, addrlen \- offsetof(sockaddr_un, sun_path))
575 .\" The following patch to amend kernel behavior was rejected:
576 .\" http://thread.gmane.org/gmane.linux.kernel.api/2437
577 .\" Subject: [patch] Fix handling of overlength pathname in AF_UNIX sun_path
579 .\" And there was a related discussion in the Austin list:
580 .\" http://thread.gmane.org/gmane.comp.standards.posix.austin.general/5735
581 .\" Subject: Having a sun_path with no null terminator
584 .\" FIXME . Track http://austingroupbugs.net/view.php?id=561
586 Alternatively, an application can retrieve
587 the socket address by allocating a buffer of size
588 .I "sizeof(struct sockaddr_un)+1"
589 that is zeroed out before the retrieval.
590 The retrieving call can specify
593 .IR "sizeof(struct sockaddr_un)" ,
594 and the extra zero byte ensures that there will be
595 a null terminator for the string returned in
602 addrlen = sizeof(struct sockaddr_un);
603 addrp = malloc(addrlen + 1);
606 memset(addrp, 0, addrlen + 1);
608 if (getsockname(sfd, (struct sockaddr *) addrp, &addrlen)) == \-1)
611 printf("sun_path = %s\\n", ((struct sockaddr_un *) addrp)\->sun_path);
615 This sort of messiness can be avoided if it is guaranteed
616 that the applications that
618 pathname sockets follow the rules outlined above under
619 .IR "Pathname sockets" .
624 For an example of the use of
634 .BR capabilities (7),