blob: 562e78fb06abd5da83a46ee208e69a73e7e7da88 [file] [log] [blame]
Serhiy Storchaka32dc1412013-10-09 14:20:06 +03001:mod:`fcntl` --- The ``fcntl`` and ``ioctl`` system calls
2=========================================================
Georg Brandl8ec7f652007-08-15 14:28:01 +00003
4.. module:: fcntl
5 :platform: Unix
6 :synopsis: The fcntl() and ioctl() system calls.
7.. sectionauthor:: Jaap Vermeulen
8
9
10.. index::
Georg Brandl1f620872010-04-25 10:29:17 +000011 pair: UNIX; file control
12 pair: UNIX; I/O control
Georg Brandl8ec7f652007-08-15 14:28:01 +000013
14This module performs file control and I/O control on file descriptors. It is an
Sandro Tosi98ed08f2012-01-14 16:42:02 +010015interface to the :c:func:`fcntl` and :c:func:`ioctl` Unix routines.
Georg Brandl8ec7f652007-08-15 14:28:01 +000016
17All functions in this module take a file descriptor *fd* as their first
18argument. This can be an integer file descriptor, such as returned by
19``sys.stdin.fileno()``, or a file object, such as ``sys.stdin`` itself, which
Serhiy Storchaka32dc1412013-10-09 14:20:06 +030020provides a :meth:`~io.IOBase.fileno` which returns a genuine file descriptor.
Georg Brandl8ec7f652007-08-15 14:28:01 +000021
22The module defines the following functions:
23
24
25.. function:: fcntl(fd, op[, arg])
26
R David Murraya6912192013-11-07 10:52:53 -050027 Perform the operation *op* on file descriptor *fd* (file objects providing
28 a :meth:`~io.IOBase.fileno` method are accepted as well). The values used
29 for for *op* are operating system dependent, and are available as constants
30 in the :mod:`fcntl` module, using the same names as used in the relevant C
31 header files. The argument *arg* is optional, and defaults to the integer
Georg Brandl8ec7f652007-08-15 14:28:01 +000032 value ``0``. When present, it can either be an integer value, or a string.
33 With the argument missing or an integer value, the return value of this function
Sandro Tosi98ed08f2012-01-14 16:42:02 +010034 is the integer return value of the C :c:func:`fcntl` call. When the argument is
Georg Brandl8ec7f652007-08-15 14:28:01 +000035 a string it represents a binary structure, e.g. created by :func:`struct.pack`.
36 The binary data is copied to a buffer whose address is passed to the C
Sandro Tosi98ed08f2012-01-14 16:42:02 +010037 :c:func:`fcntl` call. The return value after a successful call is the contents
Georg Brandl8ec7f652007-08-15 14:28:01 +000038 of the buffer, converted to a string object. The length of the returned string
39 will be the same as the length of the *arg* argument. This is limited to 1024
40 bytes. If the information returned in the buffer by the operating system is
41 larger than 1024 bytes, this is most likely to result in a segmentation
42 violation or a more subtle data corruption.
43
Sandro Tosi98ed08f2012-01-14 16:42:02 +010044 If the :c:func:`fcntl` fails, an :exc:`IOError` is raised.
Georg Brandl8ec7f652007-08-15 14:28:01 +000045
46
47.. function:: ioctl(fd, op[, arg[, mutate_flag]])
48
Serhiy Storchaka32dc1412013-10-09 14:20:06 +030049 This function is identical to the :func:`~fcntl.fcntl` function, except that the
Georg Brandl8ec7f652007-08-15 14:28:01 +000050 operations are typically defined in the library module :mod:`termios` and the
51 argument handling is even more complicated.
52
Gregory P. Smitha5cfcad2008-03-19 23:03:25 +000053 The op parameter is limited to values that can fit in 32-bits.
R David Murraya6912192013-11-07 10:52:53 -050054 Additional constants of interest for use as the *op* argument can be
55 found in the :mod:`termios` module, under the same names as used in
56 the relevant C header files.
Gregory P. Smitha5cfcad2008-03-19 23:03:25 +000057
Georg Brandl8ec7f652007-08-15 14:28:01 +000058 The parameter *arg* can be one of an integer, absent (treated identically to the
59 integer ``0``), an object supporting the read-only buffer interface (most likely
60 a plain Python string) or an object supporting the read-write buffer interface.
61
Serhiy Storchaka32dc1412013-10-09 14:20:06 +030062 In all but the last case, behaviour is as for the :func:`~fcntl.fcntl`
63 function.
Georg Brandl8ec7f652007-08-15 14:28:01 +000064
65 If a mutable buffer is passed, then the behaviour is determined by the value of
66 the *mutate_flag* parameter.
67
68 If it is false, the buffer's mutability is ignored and behaviour is as for a
69 read-only buffer, except that the 1024 byte limit mentioned above is avoided --
70 so long as the buffer you pass is as least as long as what the operating system
71 wants to put there, things should work.
72
73 If *mutate_flag* is true, then the buffer is (in effect) passed to the
74 underlying :func:`ioctl` system call, the latter's return code is passed back to
75 the calling Python, and the buffer's new contents reflect the action of the
76 :func:`ioctl`. This is a slight simplification, because if the supplied buffer
77 is less than 1024 bytes long it is first copied into a static buffer 1024 bytes
78 long which is then passed to :func:`ioctl` and copied back into the supplied
79 buffer.
80
81 If *mutate_flag* is not supplied, then from Python 2.5 it defaults to true,
82 which is a change from versions 2.3 and 2.4. Supply the argument explicitly if
83 version portability is a priority.
84
85 An example::
86
87 >>> import array, fcntl, struct, termios, os
88 >>> os.getpgrp()
89 13341
90 >>> struct.unpack('h', fcntl.ioctl(0, termios.TIOCGPGRP, " "))[0]
91 13341
92 >>> buf = array.array('h', [0])
93 >>> fcntl.ioctl(0, termios.TIOCGPGRP, buf, 1)
94 0
95 >>> buf
96 array('h', [13341])
97
98
99.. function:: flock(fd, op)
100
101 Perform the lock operation *op* on file descriptor *fd* (file objects providing
Serhiy Storchaka32dc1412013-10-09 14:20:06 +0300102 a :meth:`~io.IOBase.fileno` method are accepted as well). See the Unix manual
Georg Brandl4a20b1a2009-06-04 10:22:31 +0000103 :manpage:`flock(2)` for details. (On some systems, this function is emulated
Sandro Tosi98ed08f2012-01-14 16:42:02 +0100104 using :c:func:`fcntl`.)
Georg Brandl8ec7f652007-08-15 14:28:01 +0000105
106
107.. function:: lockf(fd, operation, [length, [start, [whence]]])
108
Serhiy Storchaka32dc1412013-10-09 14:20:06 +0300109 This is essentially a wrapper around the :func:`~fcntl.fcntl` locking calls.
110 *fd* is the file descriptor of the file to lock or unlock, and *operation*
111 is one of the following values:
Georg Brandl8ec7f652007-08-15 14:28:01 +0000112
113 * :const:`LOCK_UN` -- unlock
114 * :const:`LOCK_SH` -- acquire a shared lock
115 * :const:`LOCK_EX` -- acquire an exclusive lock
116
117 When *operation* is :const:`LOCK_SH` or :const:`LOCK_EX`, it can also be
Georg Brandlf725b952008-01-05 19:44:22 +0000118 bitwise ORed with :const:`LOCK_NB` to avoid blocking on lock acquisition.
Georg Brandl8ec7f652007-08-15 14:28:01 +0000119 If :const:`LOCK_NB` is used and the lock cannot be acquired, an
120 :exc:`IOError` will be raised and the exception will have an *errno*
121 attribute set to :const:`EACCES` or :const:`EAGAIN` (depending on the
122 operating system; for portability, check for both values). On at least some
123 systems, :const:`LOCK_EX` can only be used if the file descriptor refers to a
124 file opened for writing.
125
Serhiy Storchaka32dc1412013-10-09 14:20:06 +0300126 *length* is the number of bytes to lock, *start* is the byte offset at
127 which the lock starts, relative to *whence*, and *whence* is as with
128 :func:`io.IOBase.seek`, specifically:
Georg Brandl8ec7f652007-08-15 14:28:01 +0000129
Serhiy Storchaka32dc1412013-10-09 14:20:06 +0300130 * :const:`0` -- relative to the start of the file (:data:`os.SEEK_SET`)
131 * :const:`1` -- relative to the current buffer position (:data:`os.SEEK_CUR`)
132 * :const:`2` -- relative to the end of the file (:data:`os.SEEK_END`)
Georg Brandl8ec7f652007-08-15 14:28:01 +0000133
134 The default for *start* is 0, which means to start at the beginning of the file.
135 The default for *length* is 0 which means to lock to the end of the file. The
136 default for *whence* is also 0.
137
138Examples (all on a SVR4 compliant system)::
139
Benjamin Petersona7b55a32009-02-20 03:31:23 +0000140 import struct, fcntl, os
Georg Brandl8ec7f652007-08-15 14:28:01 +0000141
142 f = open(...)
143 rv = fcntl.fcntl(f, fcntl.F_SETFL, os.O_NDELAY)
144
145 lockdata = struct.pack('hhllhh', fcntl.F_WRLCK, 0, 0, 0, 0, 0)
146 rv = fcntl.fcntl(f, fcntl.F_SETLKW, lockdata)
147
148Note that in the first example the return value variable *rv* will hold an
149integer value; in the second example it will hold a string value. The structure
150lay-out for the *lockdata* variable is system dependent --- therefore using the
151:func:`flock` call may be better.
152
153
154.. seealso::
155
156 Module :mod:`os`
Serhiy Storchaka32dc1412013-10-09 14:20:06 +0300157 If the locking flags :data:`~os.O_SHLOCK` and :data:`~os.O_EXLOCK` are
158 present in the :mod:`os` module (on BSD only), the :func:`os.open`
159 function provides an alternative to the :func:`lockf` and :func:`flock`
160 functions.
Georg Brandl8ec7f652007-08-15 14:28:01 +0000161