1:mod:`fcntl` --- The ``fcntl`` and ``ioctl`` system calls
2=========================================================
3
4.. module:: fcntl
5   :platform: Unix
6   :synopsis: The fcntl() and ioctl() system calls.
7
8.. sectionauthor:: Jaap Vermeulen
9
10.. index::
11   pair: UNIX; file control
12   pair: UNIX; I/O control
13
14----------------
15
16This module performs file control and I/O control on file descriptors. It is an
17interface to the :c:func:`fcntl` and :c:func:`ioctl` Unix routines.  For a
18complete description of these calls, see :manpage:`fcntl(2)` and
19:manpage:`ioctl(2)` Unix manual pages.
20
21All functions in this module take a file descriptor *fd* as their first
22argument.  This can be an integer file descriptor, such as returned by
23``sys.stdin.fileno()``, or an :class:`io.IOBase` object, such as ``sys.stdin``
24itself, which provides a :meth:`~io.IOBase.fileno` that returns a genuine file
25descriptor.
26
27.. versionchanged:: 3.3
28   Operations in this module used to raise an :exc:`IOError` where they now
29   raise an :exc:`OSError`.
30
31.. versionchanged:: 3.8
32   The fcntl module now contains ``F_ADD_SEALS``, ``F_GET_SEALS``, and
33   ``F_SEAL_*`` constants for sealing of :func:`os.memfd_create` file
34   descriptors.
35
36The module defines the following functions:
37
38
39.. function:: fcntl(fd, cmd, arg=0)
40
41   Perform the operation *cmd* on file descriptor *fd* (file objects providing
42   a :meth:`~io.IOBase.fileno` method are accepted as well).  The values used
43   for *cmd* are operating system dependent, and are available as constants
44   in the :mod:`fcntl` module, using the same names as used in the relevant C
45   header files. The argument *arg* can either be an integer value, or a
46   :class:`bytes` object. With an integer value, the return value of this
47   function is the integer return value of the C :c:func:`fcntl` call.  When
48   the argument is bytes it represents a binary structure, e.g. created by
49   :func:`struct.pack`. The binary data is copied to a buffer whose address is
50   passed to the C :c:func:`fcntl` call.  The return value after a successful
51   call is the contents of the buffer, converted to a :class:`bytes` object.
52   The length of the returned object will be the same as the length of the
53   *arg* argument. This is limited to 1024 bytes. If the information returned
54   in the buffer by the operating system is larger than 1024 bytes, this is
55   most likely to result in a segmentation violation or a more subtle data
56   corruption.
57
58   If the :c:func:`fcntl` fails, an :exc:`OSError` is raised.
59
60   .. audit-event:: fcntl.fcntl fd,cmd,arg fcntl.fcntl
61
62
63.. function:: ioctl(fd, request, arg=0, mutate_flag=True)
64
65   This function is identical to the :func:`~fcntl.fcntl` function, except
66   that the argument handling is even more complicated.
67
68   The *request* parameter is limited to values that can fit in 32-bits.
69   Additional constants of interest for use as the *request* argument can be
70   found in the :mod:`termios` module, under the same names as used in
71   the relevant C header files.
72
73   The parameter *arg* can be one of an integer, an object supporting the
74   read-only buffer interface (like :class:`bytes`) or an object supporting
75   the read-write buffer interface (like :class:`bytearray`).
76
77   In all but the last case, behaviour is as for the :func:`~fcntl.fcntl`
78   function.
79
80   If a mutable buffer is passed, then the behaviour is determined by the value of
81   the *mutate_flag* parameter.
82
83   If it is false, the buffer's mutability is ignored and behaviour is as for a
84   read-only buffer, except that the 1024 byte limit mentioned above is avoided --
85   so long as the buffer you pass is at least as long as what the operating system
86   wants to put there, things should work.
87
88   If *mutate_flag* is true (the default), then the buffer is (in effect) passed
89   to the underlying :func:`ioctl` system call, the latter's return code is
90   passed back to the calling Python, and the buffer's new contents reflect the
91   action of the :func:`ioctl`.  This is a slight simplification, because if the
92   supplied buffer is less than 1024 bytes long it is first copied into a static
93   buffer 1024 bytes long which is then passed to :func:`ioctl` and copied back
94   into the supplied buffer.
95
96   If the :c:func:`ioctl` fails, an :exc:`OSError` exception is raised.
97
98   An example::
99
100      >>> import array, fcntl, struct, termios, os
101      >>> os.getpgrp()
102      13341
103      >>> struct.unpack('h', fcntl.ioctl(0, termios.TIOCGPGRP, "  "))[0]
104      13341
105      >>> buf = array.array('h', [0])
106      >>> fcntl.ioctl(0, termios.TIOCGPGRP, buf, 1)
107      0
108      >>> buf
109      array('h', [13341])
110
111   .. audit-event:: fcntl.ioctl fd,request,arg fcntl.ioctl
112
113
114.. function:: flock(fd, operation)
115
116   Perform the lock operation *operation* on file descriptor *fd* (file objects providing
117   a :meth:`~io.IOBase.fileno` method are accepted as well). See the Unix manual
118   :manpage:`flock(2)` for details.  (On some systems, this function is emulated
119   using :c:func:`fcntl`.)
120
121   If the :c:func:`flock` fails, an :exc:`OSError` exception is raised.
122
123   .. audit-event:: fcntl.flock fd,operation fcntl.flock
124
125
126.. function:: lockf(fd, cmd, len=0, start=0, whence=0)
127
128   This is essentially a wrapper around the :func:`~fcntl.fcntl` locking calls.
129   *fd* is the file descriptor (file objects providing a :meth:`~io.IOBase.fileno`
130   method are accepted as well) of the file to lock or unlock, and *cmd*
131   is one of the following values:
132
133   * :const:`LOCK_UN` -- unlock
134   * :const:`LOCK_SH` -- acquire a shared lock
135   * :const:`LOCK_EX` -- acquire an exclusive lock
136
137   When *cmd* is :const:`LOCK_SH` or :const:`LOCK_EX`, it can also be
138   bitwise ORed with :const:`LOCK_NB` to avoid blocking on lock acquisition.
139   If :const:`LOCK_NB` is used and the lock cannot be acquired, an
140   :exc:`OSError` will be raised and the exception will have an *errno*
141   attribute set to :const:`EACCES` or :const:`EAGAIN` (depending on the
142   operating system; for portability, check for both values).  On at least some
143   systems, :const:`LOCK_EX` can only be used if the file descriptor refers to a
144   file opened for writing.
145
146   *len* is the number of bytes to lock, *start* is the byte offset at
147   which the lock starts, relative to *whence*, and *whence* is as with
148   :func:`io.IOBase.seek`, specifically:
149
150   * :const:`0` -- relative to the start of the file (:data:`os.SEEK_SET`)
151   * :const:`1` -- relative to the current buffer position (:data:`os.SEEK_CUR`)
152   * :const:`2` -- relative to the end of the file (:data:`os.SEEK_END`)
153
154   The default for *start* is 0, which means to start at the beginning of the file.
155   The default for *len* is 0 which means to lock to the end of the file.  The
156   default for *whence* is also 0.
157
158   .. audit-event:: fcntl.lockf fd,cmd,len,start,whence fcntl.lockf
159
160Examples (all on a SVR4 compliant system)::
161
162   import struct, fcntl, os
163
164   f = open(...)
165   rv = fcntl.fcntl(f, fcntl.F_SETFL, os.O_NDELAY)
166
167   lockdata = struct.pack('hhllhh', fcntl.F_WRLCK, 0, 0, 0, 0, 0)
168   rv = fcntl.fcntl(f, fcntl.F_SETLKW, lockdata)
169
170Note that in the first example the return value variable *rv* will hold an
171integer value; in the second example it will hold a :class:`bytes` object.  The
172structure lay-out for the *lockdata* variable is system dependent --- therefore
173using the :func:`flock` call may be better.
174
175
176.. seealso::
177
178   Module :mod:`os`
179      If the locking flags :data:`~os.O_SHLOCK` and :data:`~os.O_EXLOCK` are
180      present in the :mod:`os` module (on BSD only), the :func:`os.open`
181      function provides an alternative to the :func:`lockf` and :func:`flock`
182      functions.
183