1 .\" This man page is Copyright (C) 1999 Andi Kleen <ak@muc.de>,
2 .\" Copyright (C) 2008-2014, Michael Kerrisk <mtk.manpages@gmail.com>,
3 .\" and Copyright (C) 2016, Heinrich Schuchardt <xypron.glpk@gmx.de>
5 .\" %%%LICENSE_START(VERBATIM_ONE_PARA)
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 2017-03-13 "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 sequenced-packet socket that is connection-oriented,
53 preserves message boundaries,
54 and delivers messages in the order that they were sent.
56 UNIX domain sockets support passing file descriptors or process credentials
57 to other processes using ancillary data.
59 A UNIX domain socket address is represented in the following structure:
63 .\" #define UNIX_PATH_MAX 108
66 sa_family_t sun_family; /* AF_UNIX */
67 char sun_path[108]; /* pathname */
78 is 108 bytes in size; see also NOTES, below.
80 Various systems calls (for example,
88 Some other system calls (for example,
94 return an argument of this type.
96 Three types of address are distinguished in the
101 a UNIX domain socket can be bound to a null-terminated
102 filesystem pathname using
104 When the address of a pathname socket is returned
105 (by one of the system calls noted above),
108 offsetof(struct sockaddr_un, sun_path) + strlen(sun_path) + 1
112 contains the null-terminated pathname.
115 expression equates to the same value as
116 .IR sizeof(sa_family_t) ,
117 but some other implementations include other fields before
121 expression more portably describes the size of the address structure.)
123 For further details of pathname sockets, see below.
126 A stream socket that has not been bound to a pathname using
129 Likewise, the two sockets created by
132 When the address of an unnamed socket is returned,
134 .IR "sizeof(sa_family_t)" ,
137 should not be inspected.
138 .\" There is quite some variation across implementations: FreeBSD
139 .\" says the length is 16 bytes, HP-UX 11 says it's zero bytes.
142 an abstract socket address is distinguished (from a pathname socket)
145 is a null byte (\(aq\\0\(aq).
146 The socket's address in this namespace is given by the additional
149 that are covered by the specified length of the address structure.
150 (Null bytes in the name have no special significance.)
151 The name has no connection with filesystem pathnames.
152 When the address of an abstract socket is returned,
156 .IR "sizeof(sa_family_t)"
157 (i.e., greater than 2), and the name of the socket is contained in
159 .IR "(addrlen \- sizeof(sa_family_t))"
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 .SS Pathname socket ownership and permissions
225 In the Linux implementation,
226 pathname sockets honor the permissions of the directory they are in.
227 Creation of a new socket will fail if the process does not have write and
228 search (execute) permission on the directory in which the socket is created.
231 connecting to a stream socket object requires write permission on that socket;
232 sending a datagram to a datagram socket likewise
233 requires write permission on that socket.
234 POSIX does not make any statement about the effect of the permissions
235 on a socket file, and on some systems (e.g., older BSDs),
236 the socket permissions are ignored.
237 Portable programs should not rely on
238 this feature for security.
240 When creating a new socket, the owner and group of the socket file
241 are set according to the usual rules.
242 The socket file has all permissions enabled,
243 other than those that are turned off by the process
246 The owner, group, and permissions of a pathname socket can be changed (using
250 .\" However, fchown() and fchmod() do not seem to have an effect
253 Socket permissions have no meaning for abstract sockets:
256 has no effect when binding an abstract socket,
257 and changing the ownership and permissions of the object (via
261 has no effect on the accessibility of the socket.
263 Abstract sockets automatically disappear when all open references
264 to the socket are closed.
266 The abstract socket namespace is a nonportable Linux extension.
269 For historical reasons, these socket options are specified with a
271 type even though they are
280 as the socket family.
283 Enables the receiving of the credentials of the sending process in an
285 When this option is set and the socket is not yet connected
286 a unique name in the abstract namespace will be generated automatically.
287 Expects an integer boolean flag.
294 .IR sizeof(sa_family_t) ,
295 .\" i.e., sizeof(short)
298 socket option was specified for a socket that was
299 not explicitly bound to an address,
300 then the socket is autobound to an abstract address.
301 The address consists of a null byte
302 followed by 5 bytes in the character set
304 Thus, there is a limit of 2^20 autobind addresses.
305 (From Linux 2.1.15, when the autobind feature was added,
306 8 bytes were used, and the limit was thus 2^32 autobind addresses.
307 The change to 5 bytes came in Linux 2.3.15.)
309 The following paragraphs describe domain-specific details and
310 unsupported features of the sockets API for UNIX domain sockets on Linux.
312 UNIX domain sockets do not support the transmission of
313 out-of-band data (the
323 flag is not supported by UNIX domain sockets.
326 .\" commit 9f6f9af7694ede6314bed281eec74d588ba9474f
333 was not supported by UNIX domain sockets.
337 socket option does have an effect for UNIX domain sockets, but the
340 For datagram sockets, the
342 value imposes an upper limit on the size of outgoing datagrams.
343 This limit is calculated as the doubled (see
345 option value less 32 bytes used for overhead.
346 .SS Ancillary messages
347 Ancillary data is sent and received using
351 For historical reasons the ancillary message types listed below
354 type even though they are
366 For more information see
370 Send or receive a set of open file descriptors from another process.
371 The data portion contains an integer array of the file descriptors.
372 The passed file descriptors behave as though they have been created with
376 Send or receive UNIX credentials.
377 This can be used for authentication.
378 The credentials are passed as a
381 Thus structure is defined in
388 pid_t pid; /* process ID of the sending process */
389 uid_t uid; /* user ID of the sending process */
390 gid_t gid; /* group ID of the sending process */
397 feature test macro must be defined (before including
399 header files) in order to obtain the definition
402 The credentials which the sender specifies are checked by the kernel.
403 A process with effective user ID 0 is allowed to specify values that do
405 The sender must specify its own process ID (unless it has the capability
407 its real user ID, effective user ID, or saved set-user-ID (unless it has
409 and its real group ID, effective group ID, or saved set-group-ID
416 option must be enabled on the socket.
420 calls return information in
422 The correct syntax is:
427 .IB error " = ioctl(" unix_socket ", " ioctl_type ", &" value ");"
437 socket the function returns the amount of queued unread data in the receive buffer.
438 The socket must not be in LISTEN state, otherwise an error
443 .IR <linux/sockios.h> .
444 .\" FIXME . http://sources.redhat.com/bugzilla/show_bug.cgi?id=12002,
445 .\" filed 2010-09-10, may cause SIOCINQ to be defined in glibc headers
447 you can use the synonymous
451 .\" SIOCOUTQ also has an effect for UNIX domain sockets, but not
452 .\" quite what userland might expect. It seems to return the number
453 .\" of bytes allocated for buffers containing pending output.
454 .\" That number is normally larger than the number of bytes of pending
455 .\" output. Since this info is, from userland's point of view, imprecise,
456 .\" and it may well change, probably best not to document this now.
460 the returned value is the same as
461 for Internet domain datagram socket;
467 The specified local address is already in use or the filesystem socket
468 object already exists.
471 The remote address specified by
473 was not a listening socket.
474 This error can also occur if the target pathname is not a socket.
477 Remote socket was unexpectedly closed.
480 User memory address was not valid.
483 Invalid argument passed.
484 A common cause is that the value
486 was not specified in the
488 field of passed addresses, or the socket was in an
489 invalid state for the applied operation.
493 called on an already connected socket or a target address was
494 specified on a connected socket.
497 The pathname in the remote address specified to
505 Socket operation needs a target address, but the socket is not connected.
508 Stream operation called on non-stream oriented socket or tried to
509 use the out-of-band data option.
512 The sender passed invalid credentials in the
516 Remote socket was closed on a stream socket.
520 This can be avoided by passing the
528 Passed protocol is not
532 Remote socket does not match the local socket type
541 This error can occur for
543 when sending a file descriptor as ancillary data over
544 a UNIX domain socket (see the description of
547 It occurs if the number of "in-flight" file descriptors exceeds the
549 resource limit and the caller does not have the
552 An in-flight file descriptor is one that has been sent using
554 but has not yet been accepted in the recipient process using
557 This error is diagnosed since mainline Linux 4.5
558 (and in some earlier kernel versions where the fix has been backported).
559 .\" commit 712f4aad406bb1ed67f3f98d04c044191f0ff593
560 In earlier kernel versions,
561 it was possible to place an unlimited number of file descriptors in flight,
562 by sending each file descriptor with
564 and then closing the file descriptor so that it was not accounted against the
568 Other errors can be generated by the generic socket layer or
569 by the filesystem while generating a filesystem socket object.
570 See the appropriate manual pages for more information.
573 and the abstract namespace were introduced with Linux 2.2 and should not
574 be used in portable programs.
575 (Some BSD-derived systems also support credential passing,
576 but the implementation details differ.)
578 Binding to a socket with a filename creates a socket
579 in the filesystem that must be deleted by the caller when it is no
582 The usual UNIX close-behind semantics apply; the socket can be unlinked
583 at any time and will be finally removed from the filesystem when the last
584 reference to it is closed.
586 To pass file descriptors or credentials over a
589 to send or receive at least one byte of nonancillary data in the same
595 UNIX domain stream sockets do not support the notion of out-of-band data.
598 When binding a socket to an address,
599 Linux is one of the implementations that appends a null terminator
600 if none is supplied in
602 In most cases this is unproblematic:
603 when the socket address is retrieved,
604 it will be one byte longer than that supplied when the socket was bound.
605 However, there is one case where confusing behavior can result:
606 if 108 non-null bytes are supplied when a socket is bound,
607 then the addition of the null terminator takes the length of
609 .IR sizeof(sun_path) .
610 Consequently, when retrieving the socket address
613 .\" The behavior on Solaris is quite similar.
616 argument for the retrieving call is specified as
617 .IR "sizeof(struct sockaddr_un)" ,
618 then the returned address structure
620 have a null terminator in
623 In addition, some implementations
624 .\" i.e., traditional BSD
625 don't require a null terminator when binding a socket (the
627 argument is used to determine the length of
629 and when the socket address is retrieved on these implementations,
630 there is no null terminator in
633 Applications that retrieve socket addresses can (portably) code
634 to handle the possibility that there is no null terminator in
636 by respecting the fact that the number of valid bytes in the pathname is:
638 strnlen(addr.sun_path, addrlen \- offsetof(sockaddr_un, sun_path))
639 .\" The following patch to amend kernel behavior was rejected:
640 .\" http://thread.gmane.org/gmane.linux.kernel.api/2437
641 .\" Subject: [patch] Fix handling of overlength pathname in AF_UNIX sun_path
643 .\" And there was a related discussion in the Austin list:
644 .\" http://thread.gmane.org/gmane.comp.standards.posix.austin.general/5735
645 .\" Subject: Having a sun_path with no null terminator
648 .\" FIXME . Track http://austingroupbugs.net/view.php?id=561
650 Alternatively, an application can retrieve
651 the socket address by allocating a buffer of size
652 .I "sizeof(struct sockaddr_un)+1"
653 that is zeroed out before the retrieval.
654 The retrieving call can specify
657 .IR "sizeof(struct sockaddr_un)" ,
658 and the extra zero byte ensures that there will be
659 a null terminator for the string returned in
666 addrlen = sizeof(struct sockaddr_un);
667 addrp = malloc(addrlen + 1);
670 memset(addrp, 0, addrlen + 1);
672 if (getsockname(sfd, (struct sockaddr *) addrp, &addrlen)) == \-1)
675 printf("sun_path = %s\\n", ((struct sockaddr_un *) addrp)\->sun_path);
679 This sort of messiness can be avoided if it is guaranteed
680 that the applications that
682 pathname sockets follow the rules outlined above under
683 .IR "Pathname sockets" .
685 The following code demonstrates the use of sequenced-packet
686 sockets for local interprocess communication.
687 It consists of two programs.
688 The server program waits for a connection from the client program.
689 The client sends each of its command-line arguments in separate messages.
690 The server treats the incoming messages as integers and adds them up.
691 The client sends the command string "END".
692 The server sends back a message containing the sum of the client's integers.
693 The client prints the sum and exits.
694 The server waits for the next client to connect.
695 To stop the server, the client is called with the command-line argument "DOWN".
697 The following output was recorded while running the server in the background
698 and repeatedly executing the client.
699 Execution of the server program ends when it receives the "DOWN" command.
707 $ \fB./client 11 \-5\fP
709 $ \fB./client DOWN\fP
721 #define SOCKET_NAME "/tmp/9Lq7BNBnBycd6nxy.socket"
722 #define BUFFER_SIZE 12
731 #include <sys/socket.h>
734 #include "connection.h"
737 main(int argc, char *argv[])
739 struct sockaddr_un name;
742 int connection_socket;
745 char buffer[BUFFER_SIZE];
748 * In case the program exited inadvertently on the last run,
754 /* Create local socket. */
756 connection_socket = socket(AF_UNIX, SOCK_SEQPACKET, 0);
757 if (connection_socket == \-1) {
763 * For portability clear the whole structure, since some
764 * implementations have additional (nonstandard) fields in
768 memset(&name, 0, sizeof(struct sockaddr_un));
770 /* Bind socket to socket name. */
772 name.sun_family = AF_UNIX;
773 strncpy(name.sun_path, SOCKET_NAME, sizeof(name.sun_path) \- 1);
775 ret = bind(connection_socket, (const struct sockaddr *) &name,
776 sizeof(struct sockaddr_un));
783 * Prepare for accepting connections. The backlog size is set
784 * to 20. So while one request is being processed other requests
788 ret = listen(connection_socket, 20);
794 /* This is the main loop for handling connections. */
798 /* Wait for incoming connection. */
800 data_socket = accept(connection_socket, NULL, NULL);
801 if (data_socket == \-1) {
809 /* Wait for next data packet. */
811 ret = read(data_socket, buffer, BUFFER_SIZE);
817 /* Ensure buffer is 0\-terminated. */
819 buffer[BUFFER_SIZE \- 1] = 0;
821 /* Handle commands. */
823 if (!strncmp(buffer, "DOWN", BUFFER_SIZE)) {
828 if (!strncmp(buffer, "END", BUFFER_SIZE)) {
832 /* Add received summand. */
834 result += atoi(buffer);
839 sprintf(buffer, "%d", result);
840 ret = write(data_socket, buffer, BUFFER_SIZE);
851 /* Quit on DOWN command. */
858 close(connection_socket);
860 /* Unlink the socket. */
875 #include <sys/socket.h>
878 #include "connection.h"
881 main(int argc, char *argv[])
883 struct sockaddr_un addr;
887 char buffer[BUFFER_SIZE];
889 /* Create local socket. */
891 data_socket = socket(AF_UNIX, SOCK_SEQPACKET, 0);
892 if (data_socket == \-1) {
898 * For portability clear the whole structure, since some
899 * implementations have additional (nonstandard) fields in
903 memset(&addr, 0, sizeof(struct sockaddr_un));
905 /* Connect socket to socket address */
907 addr.sun_family = AF_UNIX;
908 strncpy(addr.sun_path, SOCKET_NAME, sizeof(addr.sun_path) \- 1);
910 ret = connect (data_socket, (const struct sockaddr *) &addr,
911 sizeof(struct sockaddr_un));
913 fprintf(stderr, "The server is down.\\n");
917 /* Send arguments. */
919 for (i = 1; i < argc; ++i) {
920 ret = write(data_socket, argv[i], strlen(argv[i]) + 1);
927 /* Request result. */
929 strcpy (buffer, "END");
930 ret = write(data_socket, buffer, strlen(buffer) + 1);
936 /* Receive result. */
938 ret = read(data_socket, buffer, BUFFER_SIZE);
944 /* Ensure buffer is 0\-terminated. */
946 buffer[BUFFER_SIZE \- 1] = 0;
948 printf("Result = %s\\n", buffer);
958 For an example of the use of
968 .BR capabilities (7),