socket.rst 72.9 KB
Newer Older
1 2 3 4 5 6
:mod:`socket` --- Low-level networking interface
================================================

.. module:: socket
   :synopsis: Low-level networking interface.

7 8 9
**Source code:** :source:`Lib/socket.py`

--------------
10 11

This module provides access to the BSD *socket* interface. It is available on
12
all modern Unix systems, Windows, MacOS, and probably additional platforms.
13 14 15 16 17 18 19 20 21 22

.. note::

   Some behavior may be platform dependent, since calls are made to the operating
   system socket APIs.

.. index:: object: socket

The Python interface is a straightforward transliteration of the Unix system
call and library interface for sockets to Python's object-oriented style: the
23
:func:`.socket` function returns a :dfn:`socket object` whose methods implement
24 25 26 27 28 29
the various socket system calls.  Parameter types are somewhat higher-level than
in the C interface: as with :meth:`read` and :meth:`write` operations on Python
files, buffer allocation on receive operations is automatic, and buffer length
is implicit on send operations.


Antoine Pitrou's avatar
Antoine Pitrou committed
30 31 32 33 34 35 36 37 38
.. seealso::

   Module :mod:`socketserver`
      Classes that simplify writing network servers.

   Module :mod:`ssl`
      A TLS/SSL wrapper for socket objects.


39 40
Socket families
---------------
41

42 43
Depending on the system and the build options, various socket families
are supported by this module.
44

45 46 47 48 49 50 51
The address format required by a particular socket object is automatically
selected based on the address family specified when the socket object was
created.  Socket addresses are represented as follows:

- The address of an :const:`AF_UNIX` socket bound to a file system node
  is represented as a string, using the file system encoding and the
  ``'surrogateescape'`` error handler (see :pep:`383`).  An address in
52
  Linux's abstract namespace is returned as a :term:`bytes-like object` with
53 54 55
  an initial null byte; note that sockets in this namespace can
  communicate with normal file system sockets, so programs intended to
  run on Linux may need to deal with both types of address.  A string or
56
  bytes-like object can be used for either type of address when
57 58 59 60 61
  passing it as an argument.

   .. versionchanged:: 3.3
      Previously, :const:`AF_UNIX` socket paths were assumed to use UTF-8
      encoding.
62

63
   .. versionchanged:: 3.5
64 65
      Writable :term:`bytes-like object` is now accepted.

66 67
.. _host_port:

68 69 70
- A pair ``(host, port)`` is used for the :const:`AF_INET` address family,
  where *host* is a string representing either a hostname in Internet domain
  notation like ``'daring.cwi.nl'`` or an IPv4 address like ``'100.50.200.5'``,
71
  and *port* is an integer.
72

73 74 75 76 77 78 79
  - For IPv4 addresses, two special forms are accepted instead of a host
    address: ``''`` represents :const:`INADDR_ANY`, which is used to bind to all
    interfaces, and the string ``'<broadcast>'`` represents
    :const:`INADDR_BROADCAST`.  This behavior is not compatible with IPv6,
    therefore, you may want to avoid these if you intend to support IPv6 with your
    Python programs.

80 81 82 83 84 85
- For :const:`AF_INET6` address family, a four-tuple ``(host, port, flowinfo,
  scopeid)`` is used, where *flowinfo* and *scopeid* represent the ``sin6_flowinfo``
  and ``sin6_scope_id`` members in :const:`struct sockaddr_in6` in C.  For
  :mod:`socket` module methods, *flowinfo* and *scopeid* can be omitted just for
  backward compatibility.  Note, however, omission of *scopeid* can cause problems
  in manipulating scoped IPv6 addresses.
86

87 88 89 90 91
  .. versionchanged:: 3.7
     For multicast addresses (with *scopeid* meaningful) *address* may not contain
     ``%scope`` (or ``zone id``) part. This information is superfluous and may
     be safely omitted (recommended).

92
- :const:`AF_NETLINK` sockets are represented as pairs ``(pid, groups)``.
93

94 95 96 97 98
- Linux-only support for TIPC is available using the :const:`AF_TIPC`
  address family.  TIPC is an open, non-IP based networked protocol designed
  for use in clustered computer environments.  Addresses are represented by a
  tuple, and the fields depend on the address type. The general tuple form is
  ``(addr_type, v1, v2, v3 [, scope])``, where:
Benjamin Peterson's avatar
Benjamin Peterson committed
99

100 101 102 103 104
  - *addr_type* is one of :const:`TIPC_ADDR_NAMESEQ`, :const:`TIPC_ADDR_NAME`,
    or :const:`TIPC_ADDR_ID`.
  - *scope* is one of :const:`TIPC_ZONE_SCOPE`, :const:`TIPC_CLUSTER_SCOPE`, and
    :const:`TIPC_NODE_SCOPE`.
  - If *addr_type* is :const:`TIPC_ADDR_NAME`, then *v1* is the server type, *v2* is
105
    the port identifier, and *v3* should be 0.
106

107
    If *addr_type* is :const:`TIPC_ADDR_NAMESEQ`, then *v1* is the server type, *v2*
108 109
    is the lower port number, and *v3* is the upper port number.

110
    If *addr_type* is :const:`TIPC_ADDR_ID`, then *v1* is the node, *v2* is the
111 112
    reference, and *v3* should be set to 0.

113 114 115 116 117
- A tuple ``(interface, )`` is used for the :const:`AF_CAN` address family,
  where *interface* is a string representing a network interface name like
  ``'can0'``. The network interface name ``''`` can be used to receive packets
  from all network interfaces of this family.

118 119 120 121
  - :const:`CAN_ISOTP` protocol require a tuple ``(interface, rx_addr, tx_addr)``
    where both additional parameters are unsigned long integer that represent a
    CAN identifier (standard or extended).

122 123 124 125 126 127 128 129
- A string or a tuple ``(id, unit)`` is used for the :const:`SYSPROTO_CONTROL`
  protocol of the :const:`PF_SYSTEM` family. The string is the name of a
  kernel control using a dynamically-assigned ID. The tuple can be used if ID
  and unit number of the kernel control are known or if a registered ID is
  used.

  .. versionadded:: 3.3

130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147
- :const:`AF_BLUETOOTH` supports the following protocols and address
  formats:

  - :const:`BTPROTO_L2CAP` accepts ``(bdaddr, psm)`` where ``bdaddr`` is
    the Bluetooth address as a string and ``psm`` is an integer.

  - :const:`BTPROTO_RFCOMM` accepts ``(bdaddr, channel)`` where ``bdaddr``
    is the Bluetooth address as a string and ``channel`` is an integer.

  - :const:`BTPROTO_HCI` accepts ``(device_id,)`` where ``device_id`` is
    either an integer or a string with the Bluetooth address of the
    interface. (This depends on your OS; NetBSD and DragonFlyBSD expect
    a Bluetooth address while everything else expects an integer.)

    .. versionchanged:: 3.2
       NetBSD and DragonFlyBSD support added.

  - :const:`BTPROTO_SCO` accepts ``bdaddr`` where ``bdaddr`` is a
148
    :class:`bytes` object containing the Bluetooth address in a
149 150 151
    string format. (ex. ``b'12:23:34:45:56:67'``) This protocol is not
    supported under FreeBSD.

152 153 154 155 156
- :const:`AF_ALG` is a Linux-only socket based interface to Kernel
  cryptography. An algorithm socket is configured with a tuple of two to four
  elements ``(type, name [, feat [, mask]])``, where:

  - *type* is the algorithm type as string, e.g. ``aead``, ``hash``,
157
    ``skcipher`` or ``rng``.
158 159 160 161 162 163

  - *name* is the algorithm name and operation mode as string, e.g.
    ``sha256``, ``hmac(sha256)``, ``cbc(aes)`` or ``drbg_nopr_ctr_aes256``.

  - *feat* and *mask* are unsigned 32bit integers.

164
  .. availability:: Linux 2.6.38, some algorithm types require more recent Kernels.
165 166 167

  .. versionadded:: 3.6

168 169 170 171
- :const:`AF_VSOCK` allows communication between virtual machines and
  their hosts. The sockets are represented as a ``(CID, port)`` tuple
  where the context ID or CID and port are integers.

172
  .. availability:: Linux >= 4.8 QEMU >= 2.8 ESX >= 4.0 ESX Workstation >= 6.5.
173 174 175

  .. versionadded:: 3.7

176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194
- :const:`AF_PACKET` is a low-level interface directly to network devices.
  The packets are represented by the tuple
  ``(ifname, proto[, pkttype[, hatype[, addr]]])`` where:

  - *ifname* - String specifying the device name.
  - *proto* - An in network-byte-order integer specifying the Ethernet
    protocol number.
  - *pkttype* - Optional integer specifying the packet type:

    - ``PACKET_HOST`` (the default) - Packet addressed to the local host.
    - ``PACKET_BROADCAST`` - Physical-layer broadcast packet.
    - ``PACKET_MULTIHOST`` - Packet sent to a physical-layer multicast address.
    - ``PACKET_OTHERHOST`` - Packet to some other host that has been caught by
      a device driver in promiscuous mode.
    - ``PACKET_OUTGOING`` - Packet originating from the local host that is
      looped back to a packet socket.
  - *hatype* - Optional integer specifying the ARP hardware address type.
  - *addr* - Optional bytes-like object specifying the hardware physical
    address, whose interpretation depends on the device.
195

196 197 198 199 200
- :const:`AF_QIPCRTR` is a Linux-only socket based interface for communicating
  with services running on co-processors in Qualcomm platforms. The address
  family is represented as a ``(node, port)`` tuple where the *node* and *port*
  are non-negative integers.

201
  .. versionadded:: 3.8
202

203 204 205 206 207 208
If you use a hostname in the *host* portion of IPv4/v6 socket address, the
program may show a nondeterministic behavior, as Python uses the first address
returned from the DNS resolution.  The socket address will be resolved
differently into an actual IPv4/v6 address, depending on the results from DNS
resolution and/or the host configuration.  For deterministic behavior use a
numeric address in *host* portion.
209

210
All errors raise exceptions.  The normal exceptions for invalid argument types
211 212 213
and out-of-memory conditions can be raised; starting from Python 3.3, errors
related to socket or address semantics raise :exc:`OSError` or one of its
subclasses (they used to raise :exc:`socket.error`).
214

Georg Brandl's avatar
Georg Brandl committed
215 216 217
Non-blocking mode is supported through :meth:`~socket.setblocking`.  A
generalization of this based on timeouts is supported through
:meth:`~socket.settimeout`.
218

219 220 221 222

Module contents
---------------

223
The module :mod:`socket` exports the following elements.
224 225


226 227 228
Exceptions
^^^^^^^^^^

229 230
.. exception:: error

231
   A deprecated alias of :exc:`OSError`.
232

233 234
   .. versionchanged:: 3.3
      Following :pep:`3151`, this class was made an alias of :exc:`OSError`.
235 236


237
.. exception:: herror
238

239
   A subclass of :exc:`OSError`, this exception is raised for
240 241 242 243 244 245
   address-related errors, i.e. for functions that use *h_errno* in the POSIX
   C API, including :func:`gethostbyname_ex` and :func:`gethostbyaddr`.
   The accompanying value is a pair ``(h_errno, string)`` representing an
   error returned by a library call.  *h_errno* is a numeric value, while
   *string* represents the description of *h_errno*, as returned by the
   :c:func:`hstrerror` C function.
246

247 248
   .. versionchanged:: 3.3
      This class was made a subclass of :exc:`OSError`.
249 250 251

.. exception:: gaierror

252
   A subclass of :exc:`OSError`, this exception is raised for
253 254 255 256 257 258
   address-related errors by :func:`getaddrinfo` and :func:`getnameinfo`.
   The accompanying value is a pair ``(error, string)`` representing an error
   returned by a library call.  *string* represents the description of
   *error*, as returned by the :c:func:`gai_strerror` C function.  The
   numeric *error* value will match one of the :const:`EAI_\*` constants
   defined in this module.
259

260 261
   .. versionchanged:: 3.3
      This class was made a subclass of :exc:`OSError`.
262 263 264

.. exception:: timeout

265
   A subclass of :exc:`OSError`, this exception is raised when a timeout
266 267 268 269
   occurs on a socket which has had timeouts enabled via a prior call to
   :meth:`~socket.settimeout` (or implicitly through
   :func:`~socket.setdefaulttimeout`).  The accompanying value is a string
   whose value is currently always "timed out".
270

271 272
   .. versionchanged:: 3.3
      This class was made a subclass of :exc:`OSError`.
273

274 275 276 277

Constants
^^^^^^^^^

278 279 280 281 282
   The AF_* and SOCK_* constants are now :class:`AddressFamily` and
   :class:`SocketKind` :class:`.IntEnum` collections.

   .. versionadded:: 3.4

283 284 285 286 287
.. data:: AF_UNIX
          AF_INET
          AF_INET6

   These constants represent the address (and protocol) families, used for the
288
   first argument to :func:`.socket`.  If the :const:`AF_UNIX` constant is not
289 290
   defined then this protocol is unsupported.  More constants may be available
   depending on the system.
291 292 293 294 295 296 297 298 299


.. data:: SOCK_STREAM
          SOCK_DGRAM
          SOCK_RAW
          SOCK_RDM
          SOCK_SEQPACKET

   These constants represent the socket types, used for the second argument to
300
   :func:`.socket`.  More constants may be available depending on the system.
301 302
   (Only :const:`SOCK_STREAM` and :const:`SOCK_DGRAM` appear to be generally
   useful.)
303

304 305 306 307 308 309 310 311 312 313 314 315
.. data:: SOCK_CLOEXEC
          SOCK_NONBLOCK

   These two constants, if defined, can be combined with the socket types and
   allow you to set some flags atomically (thus avoiding possible race
   conditions and the need for separate calls).

   .. seealso::

      `Secure File Descriptor Handling <http://udrepper.livejournal.com/20407.html>`_
      for a more thorough explanation.

316
   .. availability:: Linux >= 2.6.27.
317 318

   .. versionadded:: 3.2
319 320 321 322 323

.. data:: SO_*
          SOMAXCONN
          MSG_*
          SOL_*
324
          SCM_*
325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341
          IPPROTO_*
          IPPORT_*
          INADDR_*
          IP_*
          IPV6_*
          EAI_*
          AI_*
          NI_*
          TCP_*

   Many constants of these forms, documented in the Unix documentation on sockets
   and/or the IP protocol, are also defined in the socket module. They are
   generally used in arguments to the :meth:`setsockopt` and :meth:`getsockopt`
   methods of socket objects.  In most cases, only those symbols that are defined
   in the Unix header files are defined; for a few symbols, default values are
   provided.

342
   .. versionchanged:: 3.6
343 344
      ``SO_DOMAIN``, ``SO_PROTOCOL``, ``SO_PEERSEC``, ``SO_PASSSEC``,
      ``TCP_USER_TIMEOUT``, ``TCP_CONGESTION`` were added.
345

346 347 348 349
   .. versionchanged:: 3.6.5
      On Windows, ``TCP_FASTOPEN``, ``TCP_KEEPCNT`` appear if run-time Windows
      supports.

350 351 352
   .. versionchanged:: 3.7
      ``TCP_NOTSENT_LOWAT`` was added.

353 354 355
      On Windows, ``TCP_KEEPIDLE``, ``TCP_KEEPINTVL`` appear if run-time Windows
      supports.

356 357 358 359 360 361 362 363
.. data:: AF_CAN
          PF_CAN
          SOL_CAN_*
          CAN_*

   Many constants of these forms, documented in the Linux documentation, are
   also defined in the socket module.

364
   .. availability:: Linux >= 2.6.25.
365 366 367

   .. versionadded:: 3.3

368 369 370 371 372 373 374
.. data:: CAN_BCM
          CAN_BCM_*

   CAN_BCM, in the CAN protocol family, is the broadcast manager (BCM) protocol.
   Broadcast manager constants, documented in the Linux documentation, are also
   defined in the socket module.

375
   .. availability:: Linux >= 2.6.25.
376 377

   .. versionadded:: 3.4
378

379 380 381 382 383 384 385 386
.. data:: CAN_RAW_FD_FRAMES

   Enables CAN FD support in a CAN_RAW socket. This is disabled by default.
   This allows your application to send both CAN and CAN FD frames; however,
   you one must accept both CAN and CAN FD frames when reading from the socket.

   This constant is documented in the Linux documentation.

387
   .. availability:: Linux >= 3.6.
388 389 390

   .. versionadded:: 3.5

391 392 393 394 395
.. data:: CAN_ISOTP

   CAN_ISOTP, in the CAN protocol family, is the ISO-TP (ISO 15765-2) protocol.
   ISO-TP constants, documented in the Linux documentation.

396
   .. availability:: Linux >= 2.6.25.
397 398 399 400

   .. versionadded:: 3.7


401 402 403 404 405 406 407
.. data:: AF_PACKET
          PF_PACKET
          PACKET_*

   Many constants of these forms, documented in the Linux documentation, are
   also defined in the socket module.

408
   .. availability:: Linux >= 2.2.
409 410


411 412 413 414 415 416 417 418
.. data:: AF_RDS
          PF_RDS
          SOL_RDS
          RDS_*

   Many constants of these forms, documented in the Linux documentation, are
   also defined in the socket module.

419
   .. availability:: Linux >= 2.6.30.
420 421 422 423

   .. versionadded:: 3.3


424 425 426
.. data:: SIO_RCVALL
          SIO_KEEPALIVE_VALS
          SIO_LOOPBACK_FAST_PATH
427
          RCVALL_*
428

429
   Constants for Windows' WSAIoctl(). The constants are used as arguments to the
430
   :meth:`~socket.socket.ioctl` method of socket objects.
431

432 433 434
   .. versionchanged:: 3.6
      ``SIO_LOOPBACK_FAST_PATH`` was added.

435

436 437 438 439 440
.. data:: TIPC_*

   TIPC related constants, matching the ones exported by the C socket API. See
   the TIPC documentation for more information.

441 442 443 444 445 446
.. data:: AF_ALG
          SOL_ALG
          ALG_*

   Constants for Linux Kernel cryptography.

447
   .. availability:: Linux >= 2.6.38.
448 449 450

   .. versionadded:: 3.6

451 452 453 454 455 456 457 458

.. data:: AF_VSOCK
          IOCTL_VM_SOCKETS_GET_LOCAL_CID
          VMADDR*
          SO_VM*

   Constants for Linux host/guest communication.

459
   .. availability:: Linux >= 4.8.
460 461 462

   .. versionadded:: 3.7

463 464
.. data:: AF_LINK

465
  .. availability:: BSD, OSX.
466 467

  .. versionadded:: 3.4
468

469 470 471 472 473
.. data:: has_ipv6

   This constant contains a boolean value which indicates if IPv6 is supported on
   this platform.

474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489
.. data:: BDADDR_ANY
          BDADDR_LOCAL

   These are string constants containing Bluetooth addresses with special
   meanings. For example, :const:`BDADDR_ANY` can be used to indicate
   any address when specifying the binding socket with
   :const:`BTPROTO_RFCOMM`.

.. data:: HCI_FILTER
          HCI_TIME_STAMP
          HCI_DATA_DIR

   For use with :const:`BTPROTO_HCI`. :const:`HCI_FILTER` is not
   available for NetBSD or DragonFlyBSD. :const:`HCI_TIME_STAMP` and
   :const:`HCI_DATA_DIR` are not available for FreeBSD, NetBSD, or
   DragonFlyBSD.
490

491 492 493 494 495
.. data:: AF_QIPCRTR

   Constant for Qualcomm's IPC router protocol, used to communicate with
   service providing remote processors.

496
   .. availability:: Linux >= 4.7.
497

498 499 500 501 502 503 504 505 506
Functions
^^^^^^^^^

Creating sockets
''''''''''''''''

The following functions all create :ref:`socket objects <socket-objects>`.


507
.. function:: socket(family=AF_INET, type=SOCK_STREAM, proto=0, fileno=None)
508 509 510

   Create a new socket using the given address family, socket type and protocol
   number.  The address family should be :const:`AF_INET` (the default),
511 512 513 514 515 516
   :const:`AF_INET6`, :const:`AF_UNIX`, :const:`AF_CAN`, :const:`AF_PACKET`,
   or :const:`AF_RDS`. The socket type should be :const:`SOCK_STREAM` (the
   default), :const:`SOCK_DGRAM`, :const:`SOCK_RAW` or perhaps one of the other
   ``SOCK_`` constants. The protocol number is usually zero and may be omitted
   or in the case where the address family is :const:`AF_CAN` the protocol
   should be one of :const:`CAN_RAW`, :const:`CAN_BCM` or :const:`CAN_ISOTP`.
517 518 519 520 521 522 523 524

   If *fileno* is specified, the values for *family*, *type*, and *proto* are
   auto-detected from the specified file descriptor.  Auto-detection can be
   overruled by calling the function with explicit *family*, *type*, or *proto*
   arguments.  This only affects how Python represents e.g. the return value
   of :meth:`socket.getpeername` but not the actual OS resource.  Unlike
   :func:`socket.fromfd`, *fileno* will return the same socket and not a
   duplicate. This may help close a detached socket using
525
   :meth:`socket.close()`.
526 527

   The newly created socket is :ref:`non-inheritable <fd_inheritance>`.
528 529 530 531 532

   .. versionchanged:: 3.3
      The AF_CAN family was added.
      The AF_RDS family was added.

533 534 535 536 537 538
   .. versionchanged:: 3.4
       The CAN_BCM protocol was added.

   .. versionchanged:: 3.4
      The returned socket is now non-inheritable.

539 540
   .. versionchanged:: 3.7
       The CAN_ISOTP protocol was added.
541

542 543 544 545 546 547 548 549 550 551 552 553 554 555
   .. versionchanged:: 3.7
      When :const:`SOCK_NONBLOCK` or :const:`SOCK_CLOEXEC`
      bit flags are applied to *type* they are cleared, and
      :attr:`socket.type` will not reflect them.  They are still passed
      to the underlying system `socket()` call.  Therefore::

          sock = socket.socket(
              socket.AF_INET,
              socket.SOCK_STREAM | socket.SOCK_NONBLOCK)

      will still create a non-blocking socket on OSes that support
      ``SOCK_NONBLOCK``, but ``sock.type`` will be set to
      ``socket.SOCK_STREAM``.

556 557 558 559 560 561 562
.. function:: socketpair([family[, type[, proto]]])

   Build a pair of connected socket objects using the given address family, socket
   type, and protocol number.  Address family, socket type, and protocol number are
   as for the :func:`.socket` function above. The default family is :const:`AF_UNIX`
   if defined on the platform; otherwise, the default is :const:`AF_INET`.

563 564
   The newly created sockets are :ref:`non-inheritable <fd_inheritance>`.

565 566 567 568
   .. versionchanged:: 3.2
      The returned socket objects now support the whole socket API, rather
      than a subset.

569 570 571
   .. versionchanged:: 3.4
      The returned sockets are now non-inheritable.

572 573 574
   .. versionchanged:: 3.5
      Windows support added.

575

576
.. function:: create_connection(address[, timeout[, source_address]])
577

578 579 580 581 582 583 584 585 586 587 588
   Connect to a TCP service listening on the Internet *address* (a 2-tuple
   ``(host, port)``), and return the socket object.  This is a higher-level
   function than :meth:`socket.connect`: if *host* is a non-numeric hostname,
   it will try to resolve it for both :data:`AF_INET` and :data:`AF_INET6`,
   and then try to connect to all possible addresses in turn until a
   connection succeeds.  This makes it easy to write clients that are
   compatible to both IPv4 and IPv6.

   Passing the optional *timeout* parameter will set the timeout on the
   socket instance before attempting to connect.  If no *timeout* is
   supplied, the global default timeout setting returned by
Georg Brandl's avatar
Georg Brandl committed
589
   :func:`getdefaulttimeout` is used.
590

591 592 593 594 595 596 597
   If supplied, *source_address* must be a 2-tuple ``(host, port)`` for the
   socket to bind to as its source address before connecting.  If host or port
   are '' or 0 respectively the OS default behavior will be used.

   .. versionchanged:: 3.2
      *source_address* was added.

598
.. function:: create_server(address, *, family=AF_INET, backlog=None, reuse_port=False, dualstack_ipv6=False)
599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641

   Convenience function which creates a TCP socket bound to *address* (a 2-tuple
   ``(host, port)``) and return the socket object.

   *family* should be either :data:`AF_INET` or :data:`AF_INET6`.
   *backlog* is the queue size passed to :meth:`socket.listen`; when ``0``
   a default reasonable value is chosen.
   *reuse_port* dictates whether to set the :data:`SO_REUSEPORT` socket option.

   If *dualstack_ipv6* is true and the platform supports it the socket will
   be able to accept both IPv4 and IPv6 connections, else it will raise
   :exc:`ValueError`. Most POSIX platforms and Windows are supposed to support
   this functionality.
   When this functionality is enabled the address returned by
   :meth:`socket.getpeername` when an IPv4 connection occurs will be an IPv6
   address represented as an IPv4-mapped IPv6 address.
   If *dualstack_ipv6* is false it will explicitly disable this functionality
   on platforms that enable it by default (e.g. Linux).
   This parameter can be used in conjunction with :func:`has_dualstack_ipv6`:

   ::

     import socket

     addr = ("", 8080)  # all interfaces, port 8080
     if socket.has_dualstack_ipv6():
         s = socket.create_server(addr, family=socket.AF_INET6, dualstack_ipv6=True)
     else:
         s = socket.create_server(addr)

   .. note::
    On POSIX platforms the :data:`SO_REUSEADDR` socket option is set in order to
    immediately reuse previous sockets which were bound on the same *address*
    and remained in TIME_WAIT state.

   .. versionadded:: 3.8

.. function:: has_dualstack_ipv6()

   Return ``True`` if the platform supports creating a TCP socket which can
   handle both IPv4 and IPv6 connections.

   .. versionadded:: 3.8
642

643
.. function:: fromfd(fd, family, type, proto=0)
644 645 646 647 648 649 650 651 652 653

   Duplicate the file descriptor *fd* (an integer as returned by a file object's
   :meth:`fileno` method) and build a socket object from the result.  Address
   family, socket type and protocol number are as for the :func:`.socket` function
   above. The file descriptor should refer to a socket, but this is not checked ---
   subsequent operations on the object may fail if the file descriptor is invalid.
   This function is rarely needed, but can be used to get or set socket options on
   a socket passed to a program as standard input or output (such as a server
   started by the Unix inet daemon).  The socket is assumed to be in blocking mode.

654 655 656 657 658
   The newly created socket is :ref:`non-inheritable <fd_inheritance>`.

   .. versionchanged:: 3.4
      The returned socket is now non-inheritable.

659 660 661 662 663 664

.. function:: fromshare(data)

   Instantiate a socket from data obtained from the :meth:`socket.share`
   method.  The socket is assumed to be in blocking mode.

665
   .. availability:: Windows.
666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681

   .. versionadded:: 3.3


.. data:: SocketType

   This is a Python type object that represents the socket object type. It is the
   same as ``type(socket(...))``.


Other functions
'''''''''''''''

The :mod:`socket` module also offers various network-related services:


682 683 684 685 686 687 688 689
.. function:: close(fd)

   Close a socket file descriptor. This is like :func:`os.close`, but for
   sockets. On some platforms (most noticeable Windows) :func:`os.close`
   does not work for socket file descriptors.

   .. versionadded:: 3.7

690
.. function:: getaddrinfo(host, port, family=0, type=0, proto=0, flags=0)
691

692 693 694 695 696 697
   Translate the *host*/*port* argument into a sequence of 5-tuples that contain
   all the necessary arguments for creating a socket connected to that service.
   *host* is a domain name, a string representation of an IPv4/v6 address
   or ``None``. *port* is a string service name such as ``'http'``, a numeric
   port number or ``None``.  By passing ``None`` as the value of *host*
   and *port*, you can pass ``NULL`` to the underlying C API.
698

699
   The *family*, *type* and *proto* arguments can be optionally specified
700 701 702 703 704 705
   in order to narrow the list of addresses returned.  Passing zero as a
   value for each of these arguments selects the full range of results.
   The *flags* argument can be one or several of the ``AI_*`` constants,
   and will influence how results are computed and returned.
   For example, :const:`AI_NUMERICHOST` will disable domain name resolution
   and will raise an error if *host* is a domain name.
706

707
   The function returns a list of 5-tuples with the following structure:
708

709
   ``(family, type, proto, canonname, sockaddr)``
710

711
   In these tuples, *family*, *type*, *proto* are all integers and are
712
   meant to be passed to the :func:`.socket` function.  *canonname* will be
713 714 715 716 717 718 719 720 721
   a string representing the canonical name of the *host* if
   :const:`AI_CANONNAME` is part of the *flags* argument; else *canonname*
   will be empty.  *sockaddr* is a tuple describing a socket address, whose
   format depends on the returned *family* (a ``(address, port)`` 2-tuple for
   :const:`AF_INET`, a ``(address, port, flow info, scope id)`` 4-tuple for
   :const:`AF_INET6`), and is meant to be passed to the :meth:`socket.connect`
   method.

   The following example fetches address information for a hypothetical TCP
722
   connection to ``example.org`` on port 80 (results may differ on your
723 724
   system if IPv6 isn't enabled)::

725
      >>> socket.getaddrinfo("example.org", 80, proto=socket.IPPROTO_TCP)
726
      [(<AddressFamily.AF_INET6: 10>, <SocketType.SOCK_STREAM: 1>,
727
       6, '', ('2606:2800:220:1:248:1893:25c8:1946', 80, 0, 0)),
728
       (<AddressFamily.AF_INET: 2>, <SocketType.SOCK_STREAM: 1>,
729
       6, '', ('93.184.216.34', 80))]
730

731
   .. versionchanged:: 3.2
732
      parameters can now be passed using keyword arguments.
733

734 735 736 737
   .. versionchanged:: 3.7
      for IPv6 multicast addresses, string representing an address will not
      contain ``%scope`` part.

738 739 740 741
.. function:: getfqdn([name])

   Return a fully qualified domain name for *name*. If *name* is omitted or empty,
   it is interpreted as the local host.  To find the fully qualified name, the
Benjamin Peterson's avatar
Benjamin Peterson committed
742
   hostname returned by :func:`gethostbyaddr` is checked, followed by aliases for the
743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771
   host, if available.  The first name which includes a period is selected.  In
   case no fully qualified domain name is available, the hostname as returned by
   :func:`gethostname` is returned.


.. function:: gethostbyname(hostname)

   Translate a host name to IPv4 address format.  The IPv4 address is returned as a
   string, such as  ``'100.50.200.5'``.  If the host name is an IPv4 address itself
   it is returned unchanged.  See :func:`gethostbyname_ex` for a more complete
   interface. :func:`gethostbyname` does not support IPv6 name resolution, and
   :func:`getaddrinfo` should be used instead for IPv4/v6 dual stack support.


.. function:: gethostbyname_ex(hostname)

   Translate a host name to IPv4 address format, extended interface. Return a
   triple ``(hostname, aliaslist, ipaddrlist)`` where *hostname* is the primary
   host name responding to the given *ip_address*, *aliaslist* is a (possibly
   empty) list of alternative host names for the same address, and *ipaddrlist* is
   a list of IPv4 addresses for the same interface on the same host (often but not
   always a single address). :func:`gethostbyname_ex` does not support IPv6 name
   resolution, and :func:`getaddrinfo` should be used instead for IPv4/v6 dual
   stack support.


.. function:: gethostname()

   Return a string containing the hostname of the machine where  the Python
772 773 774
   interpreter is currently executing.

   Note: :func:`gethostname` doesn't always return the fully qualified domain
775
   name; use :func:`getfqdn` for that.
776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795


.. function:: gethostbyaddr(ip_address)

   Return a triple ``(hostname, aliaslist, ipaddrlist)`` where *hostname* is the
   primary host name responding to the given *ip_address*, *aliaslist* is a
   (possibly empty) list of alternative host names for the same address, and
   *ipaddrlist* is a list of IPv4/v6 addresses for the same interface on the same
   host (most likely containing only a single address). To find the fully qualified
   domain name, use the function :func:`getfqdn`. :func:`gethostbyaddr` supports
   both IPv4 and IPv6.


.. function:: getnameinfo(sockaddr, flags)

   Translate a socket address *sockaddr* into a 2-tuple ``(host, port)``. Depending
   on the settings of *flags*, the result can contain a fully-qualified domain name
   or numeric address representation in *host*.  Similarly, *port* can contain a
   string port name or a numeric port number.

796 797
   For IPv6 addresses, ``%scope`` is appended to the host part if *sockaddr*
   contains meaningful *scopeid*. Usually this happens for multicast addresses.
798

799 800
   For more information about *flags* you can consult :manpage:`getnameinfo(3)`.

801 802 803
.. function:: getprotobyname(protocolname)

   Translate an Internet protocol name (for example, ``'icmp'``) to a constant
804
   suitable for passing as the (optional) third argument to the :func:`.socket`
805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836
   function.  This is usually only needed for sockets opened in "raw" mode
   (:const:`SOCK_RAW`); for the normal socket modes, the correct protocol is chosen
   automatically if the protocol is omitted or zero.


.. function:: getservbyname(servicename[, protocolname])

   Translate an Internet service name and protocol name to a port number for that
   service.  The optional protocol name, if given, should be ``'tcp'`` or
   ``'udp'``, otherwise any protocol will match.


.. function:: getservbyport(port[, protocolname])

   Translate an Internet port number and protocol name to a service name for that
   service.  The optional protocol name, if given, should be ``'tcp'`` or
   ``'udp'``, otherwise any protocol will match.


.. function:: ntohl(x)

   Convert 32-bit positive integers from network to host byte order.  On machines
   where the host byte order is the same as network byte order, this is a no-op;
   otherwise, it performs a 4-byte swap operation.


.. function:: ntohs(x)

   Convert 16-bit positive integers from network to host byte order.  On machines
   where the host byte order is the same as network byte order, this is a no-op;
   otherwise, it performs a 2-byte swap operation.

837 838 839 840 841 842
   .. deprecated:: 3.7
      In case *x* does not fit in 16-bit unsigned integer, but does fit in a
      positive C int, it is silently truncated to 16-bit unsigned integer.
      This silent truncation feature is deprecated, and will raise an
      exception in future versions of Python.

843 844 845 846 847 848 849 850 851 852 853 854 855 856

.. function:: htonl(x)

   Convert 32-bit positive integers from host to network byte order.  On machines
   where the host byte order is the same as network byte order, this is a no-op;
   otherwise, it performs a 4-byte swap operation.


.. function:: htons(x)

   Convert 16-bit positive integers from host to network byte order.  On machines
   where the host byte order is the same as network byte order, this is a no-op;
   otherwise, it performs a 2-byte swap operation.

857 858 859 860 861 862
   .. deprecated:: 3.7
      In case *x* does not fit in 16-bit unsigned integer, but does fit in a
      positive C int, it is silently truncated to 16-bit unsigned integer.
      This silent truncation feature is deprecated, and will raise an
      exception in future versions of Python.

863 864 865 866

.. function:: inet_aton(ip_string)

   Convert an IPv4 address from dotted-quad string format (for example,
867
   '123.45.67.89') to 32-bit packed binary format, as a bytes object four characters in
868
   length.  This is useful when conversing with a program that uses the standard C
869
   library and needs objects of type :c:type:`struct in_addr`, which is the C type
870 871
   for the 32-bit packed binary this function returns.

872 873 874
   :func:`inet_aton` also accepts strings with less than three dots; see the
   Unix manual page :manpage:`inet(3)` for details.

875
   If the IPv4 address string passed to this function is invalid,
876
   :exc:`OSError` will be raised. Note that exactly what is valid depends on
877
   the underlying C implementation of :c:func:`inet_aton`.
878

879
   :func:`inet_aton` does not support IPv6, and :func:`inet_pton` should be used
880 881 882 883 884
   instead for IPv4/v6 dual stack support.


.. function:: inet_ntoa(packed_ip)

885 886
   Convert a 32-bit packed IPv4 address (a :term:`bytes-like object` four
   bytes in length) to its standard dotted-quad string representation (for example,
887
   '123.45.67.89').  This is useful when conversing with a program that uses the
888
   standard C library and needs objects of type :c:type:`struct in_addr`, which
889 890
   is the C type for the 32-bit packed binary data this function takes as an
   argument.
891

892
   If the byte sequence passed to this function is not exactly 4 bytes in
893
   length, :exc:`OSError` will be raised. :func:`inet_ntoa` does not
894
   support IPv6, and :func:`inet_ntop` should be used instead for IPv4/v6 dual
895
   stack support.
896

897
   .. versionchanged:: 3.5
898 899
      Writable :term:`bytes-like object` is now accepted.

900 901 902

.. function:: inet_pton(address_family, ip_string)

903 904
   Convert an IP address from its family-specific string format to a packed,
   binary format. :func:`inet_pton` is useful when a library or network protocol
905 906
   calls for an object of type :c:type:`struct in_addr` (similar to
   :func:`inet_aton`) or :c:type:`struct in6_addr`.
907 908 909

   Supported values for *address_family* are currently :const:`AF_INET` and
   :const:`AF_INET6`. If the IP address string *ip_string* is invalid,
910
   :exc:`OSError` will be raised. Note that exactly what is valid depends on
911
   both the value of *address_family* and the underlying implementation of
912
   :c:func:`inet_pton`.
913

914
   .. availability:: Unix (maybe not all platforms), Windows.
915

916 917 918
   .. versionchanged:: 3.4
      Windows support added

919 920 921

.. function:: inet_ntop(address_family, packed_ip)

922 923 924 925 926 927
   Convert a packed IP address (a :term:`bytes-like object` of some number of
   bytes) to its standard, family-specific string representation (for
   example, ``'7.10.0.5'`` or ``'5aef:2b::8'``).
   :func:`inet_ntop` is useful when a library or network protocol returns an
   object of type :c:type:`struct in_addr` (similar to :func:`inet_ntoa`) or
   :c:type:`struct in6_addr`.
928 929

   Supported values for *address_family* are currently :const:`AF_INET` and
930 931
   :const:`AF_INET6`. If the bytes object *packed_ip* is not the correct
   length for the specified address family, :exc:`ValueError` will be raised.
932
   :exc:`OSError` is raised for errors from the call to :func:`inet_ntop`.
933

934
   .. availability:: Unix (maybe not all platforms), Windows.
935

936 937 938
   .. versionchanged:: 3.4
      Windows support added

939
   .. versionchanged:: 3.5
940 941
      Writable :term:`bytes-like object` is now accepted.

942

943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959
..
   XXX: Are sendmsg(), recvmsg() and CMSG_*() available on any
   non-Unix platforms?  The old (obsolete?) 4.2BSD form of the
   interface, in which struct msghdr has no msg_control or
   msg_controllen members, is not currently supported.

.. function:: CMSG_LEN(length)

   Return the total length, without trailing padding, of an ancillary
   data item with associated data of the given *length*.  This value
   can often be used as the buffer size for :meth:`~socket.recvmsg` to
   receive a single item of ancillary data, but :rfc:`3542` requires
   portable applications to use :func:`CMSG_SPACE` and thus include
   space for padding, even when the item will be the last in the
   buffer.  Raises :exc:`OverflowError` if *length* is outside the
   permissible range of values.

960
   .. availability:: most Unix platforms, possibly others.
961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980

   .. versionadded:: 3.3


.. function:: CMSG_SPACE(length)

   Return the buffer size needed for :meth:`~socket.recvmsg` to
   receive an ancillary data item with associated data of the given
   *length*, along with any trailing padding.  The buffer space needed
   to receive multiple items is the sum of the :func:`CMSG_SPACE`
   values for their associated data lengths.  Raises
   :exc:`OverflowError` if *length* is outside the permissible range
   of values.

   Note that some systems might support ancillary data without
   providing this function.  Also note that setting the buffer size
   using the results of this function may not precisely limit the
   amount of ancillary data that can be received, since additional
   data may be able to fit into the padding area.

981
   .. availability:: most Unix platforms, possibly others.
982 983 984 985

   .. versionadded:: 3.3


986 987
.. function:: getdefaulttimeout()

988
   Return the default timeout in seconds (float) for new socket objects. A value
989 990 991 992 993 994
   of ``None`` indicates that new socket objects have no timeout. When the socket
   module is first imported, the default is ``None``.


.. function:: setdefaulttimeout(timeout)

995
   Set the default timeout in seconds (float) for new socket objects.  When
996 997 998
   the socket module is first imported, the default is ``None``.  See
   :meth:`~socket.settimeout` for possible values and their respective
   meanings.
999 1000


1001 1002
.. function:: sethostname(name)

1003
   Set the machine's hostname to *name*.  This will raise an
1004
   :exc:`OSError` if you don't have enough rights.
1005

1006
   .. availability:: Unix.
1007 1008 1009 1010

   .. versionadded:: 3.3


1011 1012
.. function:: if_nameindex()

1013 1014
   Return a list of network interface information
   (index int, name string) tuples.
1015
   :exc:`OSError` if the system call fails.
1016

1017
   .. availability:: Unix.
1018 1019 1020 1021 1022 1023

   .. versionadded:: 3.3


.. function:: if_nametoindex(if_name)

1024 1025
   Return a network interface index number corresponding to an
   interface name.
1026
   :exc:`OSError` if no interface with the given name exists.
1027

1028
   .. availability:: Unix.
1029 1030 1031 1032 1033 1034

   .. versionadded:: 3.3


.. function:: if_indextoname(if_index)

1035
   Return a network interface name corresponding to an
1036
   interface index number.
1037
   :exc:`OSError` if no interface with the given index exists.
1038

1039
   .. availability:: Unix.
1040 1041 1042 1043

   .. versionadded:: 3.3


1044 1045 1046 1047 1048
.. _socket-objects:

Socket Objects
--------------

1049 1050 1051
Socket objects have the following methods.  Except for
:meth:`~socket.makefile`, these correspond to Unix system calls applicable
to sockets.
1052

1053 1054 1055 1056
.. versionchanged:: 3.2
   Support for the :term:`context manager` protocol was added.  Exiting the
   context manager is equivalent to calling :meth:`~socket.close`.

1057 1058 1059 1060 1061 1062 1063 1064

.. method:: socket.accept()

   Accept a connection. The socket must be bound to an address and listening for
   connections. The return value is a pair ``(conn, address)`` where *conn* is a
   *new* socket object usable to send and receive data on the connection, and
   *address* is the address bound to the socket on the other end of the connection.

1065 1066 1067 1068 1069
   The newly created socket is :ref:`non-inheritable <fd_inheritance>`.

   .. versionchanged:: 3.4
      The socket is now non-inheritable.

1070 1071 1072 1073 1074
   .. versionchanged:: 3.5
      If the system call is interrupted and the signal handler does not raise
      an exception, the method now retries the system call instead of raising
      an :exc:`InterruptedError` exception (see :pep:`475` for the rationale).

1075 1076 1077 1078 1079 1080 1081 1082 1083

.. method:: socket.bind(address)

   Bind the socket to *address*.  The socket must not already be bound. (The format
   of *address* depends on the address family --- see above.)


.. method:: socket.close()

1084 1085 1086 1087 1088 1089 1090 1091 1092
   Mark the socket closed.  The underlying system resource (e.g. a file
   descriptor) is also closed when all file objects from :meth:`makefile()`
   are closed.  Once that happens, all future operations on the socket
   object will fail. The remote end will receive no more data (after
   queued data is flushed).

   Sockets are automatically closed when they are garbage-collected, but
   it is recommended to :meth:`close` them explicitly, or to use a
   :keyword:`with` statement around them.
1093

1094 1095 1096 1097
   .. versionchanged:: 3.6
      :exc:`OSError` is now raised if an error occurs when the underlying
      :c:func:`close` call is made.

1098
   .. note::
Éric Araujo's avatar
Éric Araujo committed
1099

1100 1101 1102 1103 1104
      :meth:`close()` releases the resource associated with a connection but
      does not necessarily close the connection immediately.  If you want
      to close the connection in a timely fashion, call :meth:`shutdown()`
      before :meth:`close()`.

1105 1106 1107 1108 1109 1110

.. method:: socket.connect(address)

   Connect to a remote socket at *address*. (The format of *address* depends on the
   address family --- see above.)

1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123
   If the connection is interrupted by a signal, the method waits until the
   connection completes, or raise a :exc:`socket.timeout` on timeout, if the
   signal handler doesn't raise an exception and the socket is blocking or has
   a timeout. For non-blocking sockets, the method raises an
   :exc:`InterruptedError` exception if the connection is interrupted by a
   signal (or the exception raised by the signal handler).

   .. versionchanged:: 3.5
      The method now waits until the connection completes instead of raising an
      :exc:`InterruptedError` exception if the connection is interrupted by a
      signal, the signal handler doesn't raise an exception and the socket is
      blocking or has a timeout (see the :pep:`475` for the rationale).

1124 1125 1126 1127

.. method:: socket.connect_ex(address)

   Like ``connect(address)``, but return an error indicator instead of raising an
1128
   exception for errors returned by the C-level :c:func:`connect` call (other
1129 1130
   problems, such as "host not found," can still raise exceptions).  The error
   indicator is ``0`` if the operation succeeded, otherwise the value of the
1131
   :c:data:`errno` variable.  This is useful to support, for example, asynchronous
1132 1133 1134
   connects.


1135 1136 1137 1138 1139 1140 1141 1142 1143
.. method:: socket.detach()

   Put the socket object into closed state without actually closing the
   underlying file descriptor.  The file descriptor is returned, and can
   be reused for other purposes.

   .. versionadded:: 3.2


1144 1145 1146 1147 1148 1149 1150 1151 1152 1153
.. method:: socket.dup()

   Duplicate the socket.

   The newly created socket is :ref:`non-inheritable <fd_inheritance>`.

   .. versionchanged:: 3.4
      The socket is now non-inheritable.


1154 1155
.. method:: socket.fileno()

1156 1157
   Return the socket's file descriptor (a small integer), or -1 on failure. This
   is useful with :func:`select.select`.
1158 1159 1160 1161 1162

   Under Windows the small integer returned by this method cannot be used where a
   file descriptor can be used (such as :func:`os.fdopen`).  Unix does not have
   this limitation.

1163 1164 1165 1166 1167 1168 1169 1170 1171
.. method:: socket.get_inheritable()

   Get the :ref:`inheritable flag <fd_inheritance>` of the socket's file
   descriptor or socket's handle: ``True`` if the socket can be inherited in
   child processes, ``False`` if it cannot.

   .. versionadded:: 3.4


1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193
.. method:: socket.getpeername()

   Return the remote address to which the socket is connected.  This is useful to
   find out the port number of a remote IPv4/v6 socket, for instance. (The format
   of the address returned depends on the address family --- see above.)  On some
   systems this function is not supported.


.. method:: socket.getsockname()

   Return the socket's own address.  This is useful to find out the port number of
   an IPv4/v6 socket, for instance. (The format of the address returned depends on
   the address family --- see above.)


.. method:: socket.getsockopt(level, optname[, buflen])

   Return the value of the given socket option (see the Unix man page
   :manpage:`getsockopt(2)`).  The needed symbolic constants (:const:`SO_\*` etc.)
   are defined in this module.  If *buflen* is absent, an integer option is assumed
   and its integer value is returned by the function.  If *buflen* is present, it
   specifies the maximum length of the buffer used to receive the option in, and
1194
   this buffer is returned as a bytes object.  It is up to the caller to decode the
1195
   contents of the buffer (see the optional built-in module :mod:`struct` for a way
1196
   to decode C structures encoded as byte strings).
1197

1198

1199 1200 1201 1202 1203 1204 1205 1206 1207 1208
.. method:: socket.getblocking()

   Return ``True`` if socket is in blocking mode, ``False`` if in
   non-blocking.

   This is equivalent to checking ``socket.gettimeout() == 0``.

   .. versionadded:: 3.7


1209 1210
.. method:: socket.gettimeout()

1211
   Return the timeout in seconds (float) associated with socket operations,
1212 1213 1214 1215
   or ``None`` if no timeout is set.  This reflects the last call to
   :meth:`setblocking` or :meth:`settimeout`.


1216 1217
.. method:: socket.ioctl(control, option)

1218 1219
   :platform: Windows

1220
   The :meth:`ioctl` method is a limited interface to the WSAIoctl system
Georg Brandl's avatar
Georg Brandl committed
1221
   interface.  Please refer to the `Win32 documentation
1222
   <https://msdn.microsoft.com/en-us/library/ms741621%28VS.85%29.aspx>`_ for more
Georg Brandl's avatar
Georg Brandl committed
1223
   information.
1224

1225 1226
   On other platforms, the generic :func:`fcntl.fcntl` and :func:`fcntl.ioctl`
   functions may be used; they accept a socket object as their first argument.
1227

1228 1229 1230 1231 1232 1233
   Currently only the following control codes are supported:
   ``SIO_RCVALL``, ``SIO_KEEPALIVE_VALS``, and ``SIO_LOOPBACK_FAST_PATH``.

   .. versionchanged:: 3.6
      ``SIO_LOOPBACK_FAST_PATH`` was added.

1234
.. method:: socket.listen([backlog])
1235

1236 1237 1238 1239
   Enable a server to accept connections.  If *backlog* is specified, it must
   be at least 0 (if it is lower, it is set to 0); it specifies the number of
   unaccepted connections that the system will allow before refusing new
   connections. If not specified, a default reasonable value is chosen.
1240

1241 1242
   .. versionchanged:: 3.5
      The *backlog* parameter is now optional.
1243

1244 1245
.. method:: socket.makefile(mode='r', buffering=None, *, encoding=None, \
                            errors=None, newline=None)
1246 1247 1248

   .. index:: single: I/O control; buffering

1249 1250
   Return a :term:`file object` associated with the socket.  The exact returned
   type depends on the arguments given to :meth:`makefile`.  These arguments are
1251 1252
   interpreted the same way as by the built-in :func:`open` function, except
   the only supported *mode* values are ``'r'`` (default), ``'w'`` and ``'b'``.
1253

1254
   The socket must be in blocking mode; it can have a timeout, but the file
1255
   object's internal buffer may end up in an inconsistent state if a timeout
1256 1257 1258 1259 1260
   occurs.

   Closing the file object returned by :meth:`makefile` won't close the
   original socket unless all other file objects have been closed and
   :meth:`socket.close` has been called on the socket object.
1261 1262 1263 1264 1265 1266

   .. note::

      On Windows, the file-like object created by :meth:`makefile` cannot be
      used where a file object with a file descriptor is expected, such as the
      stream arguments of :meth:`subprocess.Popen`.
1267

1268 1269 1270

.. method:: socket.recv(bufsize[, flags])

1271
   Receive data from the socket.  The return value is a bytes object representing the
1272 1273 1274 1275 1276 1277 1278 1279 1280
   data received.  The maximum amount of data to be received at once is specified
   by *bufsize*.  See the Unix manual page :manpage:`recv(2)` for the meaning of
   the optional argument *flags*; it defaults to zero.

   .. note::

      For best match with hardware and network realities, the value of  *bufsize*
      should be a relatively small power of 2, for example, 4096.

1281 1282 1283 1284 1285
   .. versionchanged:: 3.5
      If the system call is interrupted and the signal handler does not raise
      an exception, the method now retries the system call instead of raising
      an :exc:`InterruptedError` exception (see :pep:`475` for the rationale).

1286 1287 1288

.. method:: socket.recvfrom(bufsize[, flags])

1289 1290
   Receive data from the socket.  The return value is a pair ``(bytes, address)``
   where *bytes* is a bytes object representing the data received and *address* is the
1291 1292 1293 1294
   address of the socket sending the data.  See the Unix manual page
   :manpage:`recv(2)` for the meaning of the optional argument *flags*; it defaults
   to zero. (The format of *address* depends on the address family --- see above.)

1295 1296 1297 1298 1299
   .. versionchanged:: 3.5
      If the system call is interrupted and the signal handler does not raise
      an exception, the method now retries the system call instead of raising
      an :exc:`InterruptedError` exception (see :pep:`475` for the rationale).

1300 1301 1302 1303
   .. versionchanged:: 3.7
      For multicast IPv6 address, first item of *address* does not contain
      ``%scope`` part anymore. In order to get full IPv6 address use
      :func:`getnameinfo`.
1304

1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365
.. method:: socket.recvmsg(bufsize[, ancbufsize[, flags]])

   Receive normal data (up to *bufsize* bytes) and ancillary data from
   the socket.  The *ancbufsize* argument sets the size in bytes of
   the internal buffer used to receive the ancillary data; it defaults
   to 0, meaning that no ancillary data will be received.  Appropriate
   buffer sizes for ancillary data can be calculated using
   :func:`CMSG_SPACE` or :func:`CMSG_LEN`, and items which do not fit
   into the buffer might be truncated or discarded.  The *flags*
   argument defaults to 0 and has the same meaning as for
   :meth:`recv`.

   The return value is a 4-tuple: ``(data, ancdata, msg_flags,
   address)``.  The *data* item is a :class:`bytes` object holding the
   non-ancillary data received.  The *ancdata* item is a list of zero
   or more tuples ``(cmsg_level, cmsg_type, cmsg_data)`` representing
   the ancillary data (control messages) received: *cmsg_level* and
   *cmsg_type* are integers specifying the protocol level and
   protocol-specific type respectively, and *cmsg_data* is a
   :class:`bytes` object holding the associated data.  The *msg_flags*
   item is the bitwise OR of various flags indicating conditions on
   the received message; see your system documentation for details.
   If the receiving socket is unconnected, *address* is the address of
   the sending socket, if available; otherwise, its value is
   unspecified.

   On some systems, :meth:`sendmsg` and :meth:`recvmsg` can be used to
   pass file descriptors between processes over an :const:`AF_UNIX`
   socket.  When this facility is used (it is often restricted to
   :const:`SOCK_STREAM` sockets), :meth:`recvmsg` will return, in its
   ancillary data, items of the form ``(socket.SOL_SOCKET,
   socket.SCM_RIGHTS, fds)``, where *fds* is a :class:`bytes` object
   representing the new file descriptors as a binary array of the
   native C :c:type:`int` type.  If :meth:`recvmsg` raises an
   exception after the system call returns, it will first attempt to
   close any file descriptors received via this mechanism.

   Some systems do not indicate the truncated length of ancillary data
   items which have been only partially received.  If an item appears
   to extend beyond the end of the buffer, :meth:`recvmsg` will issue
   a :exc:`RuntimeWarning`, and will return the part of it which is
   inside the buffer provided it has not been truncated before the
   start of its associated data.

   On systems which support the :const:`SCM_RIGHTS` mechanism, the
   following function will receive up to *maxfds* file descriptors,
   returning the message data and a list containing the descriptors
   (while ignoring unexpected conditions such as unrelated control
   messages being received).  See also :meth:`sendmsg`. ::

      import socket, array

      def recv_fds(sock, msglen, maxfds):
          fds = array.array("i")   # Array of ints
          msg, ancdata, flags, addr = sock.recvmsg(msglen, socket.CMSG_LEN(maxfds * fds.itemsize))
          for cmsg_level, cmsg_type, cmsg_data in ancdata:
              if (cmsg_level == socket.SOL_SOCKET and cmsg_type == socket.SCM_RIGHTS):
                  # Append data, ignoring any truncated integers at the end.
                  fds.fromstring(cmsg_data[:len(cmsg_data) - (len(cmsg_data) % fds.itemsize)])
          return msg, list(fds)

1366
   .. availability:: most Unix platforms, possibly others.
1367 1368 1369

   .. versionadded:: 3.3

1370 1371 1372 1373 1374
   .. versionchanged:: 3.5
      If the system call is interrupted and the signal handler does not raise
      an exception, the method now retries the system call instead of raising
      an :exc:`InterruptedError` exception (see :pep:`475` for the rationale).

1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407

.. method:: socket.recvmsg_into(buffers[, ancbufsize[, flags]])

   Receive normal data and ancillary data from the socket, behaving as
   :meth:`recvmsg` would, but scatter the non-ancillary data into a
   series of buffers instead of returning a new bytes object.  The
   *buffers* argument must be an iterable of objects that export
   writable buffers (e.g. :class:`bytearray` objects); these will be
   filled with successive chunks of the non-ancillary data until it
   has all been written or there are no more buffers.  The operating
   system may set a limit (:func:`~os.sysconf` value ``SC_IOV_MAX``)
   on the number of buffers that can be used.  The *ancbufsize* and
   *flags* arguments have the same meaning as for :meth:`recvmsg`.

   The return value is a 4-tuple: ``(nbytes, ancdata, msg_flags,
   address)``, where *nbytes* is the total number of bytes of
   non-ancillary data written into the buffers, and *ancdata*,
   *msg_flags* and *address* are the same as for :meth:`recvmsg`.

   Example::

      >>> import socket
      >>> s1, s2 = socket.socketpair()
      >>> b1 = bytearray(b'----')
      >>> b2 = bytearray(b'0123456789')
      >>> b3 = bytearray(b'--------------')
      >>> s1.send(b'Mary had a little lamb')
      22
      >>> s2.recvmsg_into([b1, memoryview(b2)[2:9], b3])
      (22, [], 0, None)
      >>> [b1, b2, b3]
      [bytearray(b'Mary'), bytearray(b'01 had a 9'), bytearray(b'little lamb---')]

1408
   .. availability:: most Unix platforms, possibly others.
1409 1410 1411 1412

   .. versionadded:: 3.3


1413 1414
.. method:: socket.recvfrom_into(buffer[, nbytes[, flags]])

1415 1416
   Receive data from the socket, writing it into *buffer* instead of creating a
   new bytestring.  The return value is a pair ``(nbytes, address)`` where *nbytes* is
1417 1418 1419 1420 1421 1422 1423 1424 1425
   the number of bytes received and *address* is the address of the socket sending
   the data.  See the Unix manual page :manpage:`recv(2)` for the meaning of the
   optional argument *flags*; it defaults to zero.  (The format of *address*
   depends on the address family --- see above.)


.. method:: socket.recv_into(buffer[, nbytes[, flags]])

   Receive up to *nbytes* bytes from the socket, storing the data into a buffer
1426
   rather than creating a new bytestring.  If *nbytes* is not specified (or 0),
Benjamin Peterson's avatar
Benjamin Peterson committed
1427 1428 1429
   receive up to the size available in the given buffer.  Returns the number of
   bytes received.  See the Unix manual page :manpage:`recv(2)` for the meaning
   of the optional argument *flags*; it defaults to zero.
1430 1431


1432
.. method:: socket.send(bytes[, flags])
1433 1434 1435 1436 1437

   Send data to the socket.  The socket must be connected to a remote socket.  The
   optional *flags* argument has the same meaning as for :meth:`recv` above.
   Returns the number of bytes sent. Applications are responsible for checking that
   all data has been sent; if only some of the data was transmitted, the
1438 1439
   application needs to attempt delivery of the remaining data. For further
   information on this topic, consult the :ref:`socket-howto`.
1440

1441 1442 1443 1444 1445
   .. versionchanged:: 3.5
      If the system call is interrupted and the signal handler does not raise
      an exception, the method now retries the system call instead of raising
      an :exc:`InterruptedError` exception (see :pep:`475` for the rationale).

1446

1447
.. method:: socket.sendall(bytes[, flags])
1448 1449 1450

   Send data to the socket.  The socket must be connected to a remote socket.  The
   optional *flags* argument has the same meaning as for :meth:`recv` above.
1451
   Unlike :meth:`send`, this method continues to send data from *bytes* until
1452 1453 1454 1455
   either all data has been sent or an error occurs.  ``None`` is returned on
   success.  On error, an exception is raised, and there is no way to determine how
   much data, if any, was successfully sent.

1456
   .. versionchanged:: 3.5
1457
      The socket timeout is no more reset each time data is sent successfully.
1458 1459
      The socket timeout is now the maximum total duration to send all data.

1460 1461 1462 1463 1464
   .. versionchanged:: 3.5
      If the system call is interrupted and the signal handler does not raise
      an exception, the method now retries the system call instead of raising
      an :exc:`InterruptedError` exception (see :pep:`475` for the rationale).

1465

1466 1467
.. method:: socket.sendto(bytes, address)
            socket.sendto(bytes, flags, address)
1468 1469 1470 1471 1472 1473 1474

   Send data to the socket.  The socket should not be connected to a remote socket,
   since the destination socket is specified by *address*.  The optional *flags*
   argument has the same meaning as for :meth:`recv` above.  Return the number of
   bytes sent. (The format of *address* depends on the address family --- see
   above.)

1475 1476 1477 1478 1479
   .. versionchanged:: 3.5
      If the system call is interrupted and the signal handler does not raise
      an exception, the method now retries the system call instead of raising
      an :exc:`InterruptedError` exception (see :pep:`475` for the rationale).

1480

1481 1482 1483 1484 1485
.. method:: socket.sendmsg(buffers[, ancdata[, flags[, address]]])

   Send normal and ancillary data to the socket, gathering the
   non-ancillary data from a series of buffers and concatenating it
   into a single message.  The *buffers* argument specifies the
1486 1487
   non-ancillary data as an iterable of
   :term:`bytes-like objects <bytes-like object>`
1488 1489 1490 1491 1492 1493 1494
   (e.g. :class:`bytes` objects); the operating system may set a limit
   (:func:`~os.sysconf` value ``SC_IOV_MAX``) on the number of buffers
   that can be used.  The *ancdata* argument specifies the ancillary
   data (control messages) as an iterable of zero or more tuples
   ``(cmsg_level, cmsg_type, cmsg_data)``, where *cmsg_level* and
   *cmsg_type* are integers specifying the protocol level and
   protocol-specific type respectively, and *cmsg_data* is a
1495
   bytes-like object holding the associated data.  Note that
1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511
   some systems (in particular, systems without :func:`CMSG_SPACE`)
   might support sending only one control message per call.  The
   *flags* argument defaults to 0 and has the same meaning as for
   :meth:`send`.  If *address* is supplied and not ``None``, it sets a
   destination address for the message.  The return value is the
   number of bytes of non-ancillary data sent.

   The following function sends the list of file descriptors *fds*
   over an :const:`AF_UNIX` socket, on systems which support the
   :const:`SCM_RIGHTS` mechanism.  See also :meth:`recvmsg`. ::

      import socket, array

      def send_fds(sock, msg, fds):
          return sock.sendmsg([msg], [(socket.SOL_SOCKET, socket.SCM_RIGHTS, array.array("i", fds))])

1512
   .. availability:: most Unix platforms, possibly others.
1513 1514 1515

   .. versionadded:: 3.3

1516 1517 1518 1519 1520
   .. versionchanged:: 3.5
      If the system call is interrupted and the signal handler does not raise
      an exception, the method now retries the system call instead of raising
      an :exc:`InterruptedError` exception (see :pep:`475` for the rationale).

1521 1522 1523 1524 1525
.. method:: socket.sendmsg_afalg([msg], *, op[, iv[, assoclen[, flags]]])

   Specialized version of :meth:`~socket.sendmsg` for :const:`AF_ALG` socket.
   Set mode, IV, AEAD associated data length and flags for :const:`AF_ALG` socket.

1526
   .. availability:: Linux >= 2.6.38.
1527 1528 1529

   .. versionadded:: 3.6

1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540
.. method:: socket.sendfile(file, offset=0, count=None)

   Send a file until EOF is reached by using high-performance
   :mod:`os.sendfile` and return the total number of bytes which were sent.
   *file* must be a regular file object opened in binary mode. If
   :mod:`os.sendfile` is not available (e.g. Windows) or *file* is not a
   regular file :meth:`send` will be used instead. *offset* tells from where to
   start reading the file. If specified, *count* is the total number of bytes
   to transmit as opposed to sending the file until EOF is reached. File
   position is updated on return or also in case of error in which case
   :meth:`file.tell() <io.IOBase.tell>` can be used to figure out the number of
1541 1542
   bytes which were sent. The socket must be of :const:`SOCK_STREAM` type.
   Non-blocking sockets are not supported.
1543 1544

   .. versionadded:: 3.5
1545

1546 1547 1548 1549 1550 1551 1552 1553
.. method:: socket.set_inheritable(inheritable)

   Set the :ref:`inheritable flag <fd_inheritance>` of the socket's file
   descriptor or socket's handle.

   .. versionadded:: 3.4


1554 1555
.. method:: socket.setblocking(flag)

1556 1557
   Set blocking or non-blocking mode of the socket: if *flag* is false, the
   socket is set to non-blocking, else to blocking mode.
1558

1559
   This method is a shorthand for certain :meth:`~socket.settimeout` calls:
1560

1561
   * ``sock.setblocking(True)`` is equivalent to ``sock.settimeout(None)``
1562

1563
   * ``sock.setblocking(False)`` is equivalent to ``sock.settimeout(0.0)``
1564

1565 1566 1567 1568
   .. versionchanged:: 3.7
      The method no longer applies :const:`SOCK_NONBLOCK` flag on
      :attr:`socket.type`.

1569

1570
.. method:: socket.settimeout(value)
1571

1572 1573 1574 1575 1576 1577
   Set a timeout on blocking socket operations.  The *value* argument can be a
   nonnegative floating point number expressing seconds, or ``None``.
   If a non-zero value is given, subsequent socket operations will raise a
   :exc:`timeout` exception if the timeout period *value* has elapsed before
   the operation has completed.  If zero is given, the socket is put in
   non-blocking mode. If ``None`` is given, the socket is put in blocking mode.
1578

1579
   For further information, please consult the :ref:`notes on socket timeouts <socket-timeouts>`.
1580

1581 1582 1583 1584
   .. versionchanged:: 3.7
      The method no longer toggles :const:`SOCK_NONBLOCK` flag on
      :attr:`socket.type`.

1585

1586 1587 1588
.. method:: socket.setsockopt(level, optname, value: int)
.. method:: socket.setsockopt(level, optname, value: buffer)
.. method:: socket.setsockopt(level, optname, None, optlen: int)
1589 1590 1591 1592 1593

   .. index:: module: struct

   Set the value of the given socket option (see the Unix manual page
   :manpage:`setsockopt(2)`).  The needed symbolic constants are defined in the
1594
   :mod:`socket` module (:const:`SO_\*` etc.).  The value can be an integer,
1595
   ``None`` or a :term:`bytes-like object` representing a buffer. In the later
1596 1597
   case it is up to the caller to ensure that the bytestring contains the
   proper bits (see the optional built-in module :mod:`struct` for a way to
1598
   encode C structures as bytestrings). When value is set to ``None``,
1599 1600 1601
   optlen argument is required. It's equivalent to call setsockopt C
   function with optval=NULL and optlen=optlen.

1602

1603
   .. versionchanged:: 3.5
1604 1605
      Writable :term:`bytes-like object` is now accepted.

1606 1607 1608
   .. versionchanged:: 3.6
      setsockopt(level, optname, None, optlen: int) form added.

1609 1610 1611 1612 1613 1614

.. method:: socket.shutdown(how)

   Shut down one or both halves of the connection.  If *how* is :const:`SHUT_RD`,
   further receives are disallowed.  If *how* is :const:`SHUT_WR`, further sends
   are disallowed.  If *how* is :const:`SHUT_RDWR`, further sends and receives are
1615
   disallowed.
1616

1617 1618 1619

.. method:: socket.share(process_id)

1620 1621 1622 1623 1624 1625
   Duplicate a socket and prepare it for sharing with a target process.  The
   target process must be provided with *process_id*.  The resulting bytes object
   can then be passed to the target process using some form of interprocess
   communication and the socket can be recreated there using :func:`fromshare`.
   Once this method has been called, it is safe to close the socket since
   the operating system has already duplicated it for the target process.
1626

1627
   .. availability:: Windows.
1628 1629 1630 1631

   .. versionadded:: 3.3


Georg Brandl's avatar
Georg Brandl committed
1632 1633
Note that there are no methods :meth:`read` or :meth:`write`; use
:meth:`~socket.recv` and :meth:`~socket.send` without *flags* argument instead.
1634 1635

Socket objects also have these (read-only) attributes that correspond to the
1636
values given to the :class:`~socket.socket` constructor.
1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653


.. attribute:: socket.family

   The socket family.


.. attribute:: socket.type

   The socket type.


.. attribute:: socket.proto

   The socket protocol.


1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708

.. _socket-timeouts:

Notes on socket timeouts
------------------------

A socket object can be in one of three modes: blocking, non-blocking, or
timeout.  Sockets are by default always created in blocking mode, but this
can be changed by calling :func:`setdefaulttimeout`.

* In *blocking mode*, operations block until complete or the system returns
  an error (such as connection timed out).

* In *non-blocking mode*, operations fail (with an error that is unfortunately
  system-dependent) if they cannot be completed immediately: functions from the
  :mod:`select` can be used to know when and whether a socket is available for
  reading or writing.

* In *timeout mode*, operations fail if they cannot be completed within the
  timeout specified for the socket (they raise a :exc:`timeout` exception)
  or if the system returns an error.

.. note::
   At the operating system level, sockets in *timeout mode* are internally set
   in non-blocking mode.  Also, the blocking and timeout modes are shared between
   file descriptors and socket objects that refer to the same network endpoint.
   This implementation detail can have visible consequences if e.g. you decide
   to use the :meth:`~socket.fileno()` of a socket.

Timeouts and the ``connect`` method
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

The :meth:`~socket.connect` operation is also subject to the timeout
setting, and in general it is recommended to call :meth:`~socket.settimeout`
before calling :meth:`~socket.connect` or pass a timeout parameter to
:meth:`create_connection`.  However, the system network stack may also
return a connection timeout error of its own regardless of any Python socket
timeout setting.

Timeouts and the ``accept`` method
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

If :func:`getdefaulttimeout` is not :const:`None`, sockets returned by
the :meth:`~socket.accept` method inherit that timeout.  Otherwise, the
behaviour depends on settings of the listening socket:

* if the listening socket is in *blocking mode* or in *timeout mode*,
  the socket returned by :meth:`~socket.accept` is in *blocking mode*;

* if the listening socket is in *non-blocking mode*, whether the socket
  returned by :meth:`~socket.accept` is in blocking or non-blocking mode
  is operating system-dependent.  If you want to ensure cross-platform
  behaviour, it is recommended you manually override this setting.


1709 1710 1711 1712 1713 1714 1715
.. _socket-example:

Example
-------

Here are four minimal example programs using the TCP/IP protocol: a server that
echoes all data that it receives back (servicing only one client), and a client
1716
using it.  Note that a server must perform the sequence :func:`.socket`,
Georg Brandl's avatar
Georg Brandl committed
1717 1718
:meth:`~socket.bind`, :meth:`~socket.listen`, :meth:`~socket.accept` (possibly
repeating the :meth:`~socket.accept` to service more than one client), while a
1719
client only needs the sequence :func:`.socket`, :meth:`~socket.connect`.  Also
1720 1721
note that the server does not :meth:`~socket.sendall`/:meth:`~socket.recv` on
the socket it is listening on but on the new socket returned by
Georg Brandl's avatar
Georg Brandl committed
1722
:meth:`~socket.accept`.
1723 1724 1725 1726 1727 1728

The first two examples support IPv4 only. ::

   # Echo server program
   import socket

Christian Heimes's avatar
Christian Heimes committed
1729
   HOST = ''                 # Symbolic name meaning all available interfaces
1730
   PORT = 50007              # Arbitrary non-privileged port
1731 1732 1733 1734 1735 1736 1737 1738 1739 1740
   with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
       s.bind((HOST, PORT))
       s.listen(1)
       conn, addr = s.accept()
       with conn:
           print('Connected by', addr)
           while True:
               data = conn.recv(1024)
               if not data: break
               conn.sendall(data)
1741 1742 1743 1744 1745 1746 1747 1748

::

   # Echo client program
   import socket

   HOST = 'daring.cwi.nl'    # The remote host
   PORT = 50007              # The same port as used by the server
1749 1750 1751 1752
   with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
       s.connect((HOST, PORT))
       s.sendall(b'Hello, world')
       data = s.recv(1024)
1753
   print('Received', repr(data))
1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765

The next two examples are identical to the above two, but support both IPv4 and
IPv6. The server side will listen to the first address family available (it
should listen to both instead). On most of IPv6-ready systems, IPv6 will take
precedence and the server may not accept IPv4 traffic. The client side will try
to connect to the all addresses returned as a result of the name resolution, and
sends traffic to the first one connected successfully. ::

   # Echo server program
   import socket
   import sys

1766
   HOST = None               # Symbolic name meaning all available interfaces
1767 1768
   PORT = 50007              # Arbitrary non-privileged port
   s = None
1769 1770
   for res in socket.getaddrinfo(HOST, PORT, socket.AF_UNSPEC,
                                 socket.SOCK_STREAM, 0, socket.AI_PASSIVE):
1771 1772
       af, socktype, proto, canonname, sa = res
       try:
1773
           s = socket.socket(af, socktype, proto)
1774
       except OSError as msg:
1775 1776
           s = None
           continue
1777
       try:
1778 1779
           s.bind(sa)
           s.listen(1)
1780
       except OSError as msg:
1781 1782 1783
           s.close()
           s = None
           continue
1784 1785
       break
   if s is None:
1786
       print('could not open socket')
1787 1788
       sys.exit(1)
   conn, addr = s.accept()
1789 1790 1791 1792 1793 1794
   with conn:
       print('Connected by', addr)
       while True:
           data = conn.recv(1024)
           if not data: break
           conn.send(data)
1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807

::

   # Echo client program
   import socket
   import sys

   HOST = 'daring.cwi.nl'    # The remote host
   PORT = 50007              # The same port as used by the server
   s = None
   for res in socket.getaddrinfo(HOST, PORT, socket.AF_UNSPEC, socket.SOCK_STREAM):
       af, socktype, proto, canonname, sa = res
       try:
1808
           s = socket.socket(af, socktype, proto)
1809
       except OSError as msg:
1810 1811
           s = None
           continue
1812
       try:
1813
           s.connect(sa)
1814
       except OSError as msg:
1815 1816 1817
           s.close()
           s = None
           continue
1818 1819
       break
   if s is None:
1820
       print('could not open socket')
1821
       sys.exit(1)
1822 1823 1824
   with s:
       s.sendall(b'Hello, world')
       data = s.recv(1024)
1825
   print('Received', repr(data))
1826

1827
The next example shows how to write a very simple network sniffer with raw
1828
sockets on Windows. The example requires administrator privileges to modify
1829 1830 1831 1832 1833 1834
the interface::

   import socket

   # the public network interface
   HOST = socket.gethostbyname(socket.gethostname())
1835

1836 1837 1838
   # create a raw socket and bind it to the public interface
   s = socket.socket(socket.AF_INET, socket.SOCK_RAW, socket.IPPROTO_IP)
   s.bind((HOST, 0))
1839

1840 1841
   # Include IP headers
   s.setsockopt(socket.IPPROTO_IP, socket.IP_HDRINCL, 1)
1842

1843 1844
   # receive all packages
   s.ioctl(socket.SIO_RCVALL, socket.RCVALL_ON)
1845

1846
   # receive a package
1847
   print(s.recvfrom(65565))
1848

Christian Heimes's avatar
Christian Heimes committed
1849
   # disabled promiscuous mode
1850
   s.ioctl(socket.SIO_RCVALL, socket.RCVALL_OFF)
1851

1852
The next example shows how to use the socket interface to communicate to a CAN
1853 1854 1855 1856 1857 1858
network using the raw socket protocol. To use CAN with the broadcast
manager protocol instead, open a socket with::

    socket.socket(socket.AF_CAN, socket.SOCK_DGRAM, socket.CAN_BCM)

After binding (:const:`CAN_RAW`) or connecting (:const:`CAN_BCM`) the socket, you
1859
can use the :meth:`socket.send`, and the :meth:`socket.recv` operations (and
1860 1861
their counterparts) on the socket object as usual.

1862
This last example might require special privileges::
1863 1864 1865 1866 1867

   import socket
   import struct


1868
   # CAN frame packing/unpacking (see 'struct can_frame' in <linux/can.h>)
1869 1870

   can_frame_fmt = "=IB3x8s"
1871
   can_frame_size = struct.calcsize(can_frame_fmt)
1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882

   def build_can_frame(can_id, data):
       can_dlc = len(data)
       data = data.ljust(8, b'\x00')
       return struct.pack(can_frame_fmt, can_id, can_dlc, data)

   def dissect_can_frame(frame):
       can_id, can_dlc, data = struct.unpack(can_frame_fmt, frame)
       return (can_id, can_dlc, data[:can_dlc])


1883
   # create a raw socket and bind it to the 'vcan0' interface
1884 1885 1886 1887
   s = socket.socket(socket.AF_CAN, socket.SOCK_RAW, socket.CAN_RAW)
   s.bind(('vcan0',))

   while True:
1888
       cf, addr = s.recvfrom(can_frame_size)
1889 1890 1891 1892 1893

       print('Received: can_id=%x, can_dlc=%x, data=%s' % dissect_can_frame(cf))

       try:
           s.send(cf)
1894
       except OSError:
1895 1896 1897 1898
           print('Error sending CAN frame')

       try:
           s.send(build_can_frame(0x01, b'\x01\x02\x03'))
1899
       except OSError:
1900
           print('Error sending CAN frame')
1901

1902 1903 1904
Running an example several times with too small delay between executions, could
lead to this error::

1905
   OSError: [Errno 98] Address already in use
1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920

This is because the previous execution has left the socket in a ``TIME_WAIT``
state, and can't be immediately reused.

There is a :mod:`socket` flag to set, in order to prevent this,
:data:`socket.SO_REUSEADDR`::

   s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
   s.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
   s.bind((HOST, PORT))

the :data:`SO_REUSEADDR` flag tells the kernel to reuse a local socket in
``TIME_WAIT`` state, without waiting for its natural timeout to expire.


1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935
.. seealso::

   For an introduction to socket programming (in C), see the following papers:

   - *An Introductory 4.3BSD Interprocess Communication Tutorial*, by Stuart Sechrest

   - *An Advanced 4.3BSD Interprocess Communication Tutorial*, by Samuel J.  Leffler et
     al,

   both in the UNIX Programmer's Manual, Supplementary Documents 1 (sections
   PS1:7 and PS1:8).  The platform-specific reference material for the various
   socket-related system calls are also a valuable source of information on the
   details of socket semantics.  For Unix, refer to the manual pages; for Windows,
   see the WinSock (or Winsock 2) specification.  For IPv6-ready APIs, readers may
   want to refer to :rfc:`3493` titled Basic Socket Interface Extensions for IPv6.