path: root/manual/llio.texi

@node Low-Level I/O, File System Interface, I/O on Streams, Top
@c %MENU% Low-level, less portable I/O
@chapter Low-Level Input/Output

This chapter describes functions for performing low-level input/output
operations on file descriptors.  These functions include the primitives
for the higher-level I/O functions described in @ref{I/O on Streams}, as
well as functions for performing low-level control operations for which
there are no equivalents on streams.

Stream-level I/O is more flexible and usually more convenient;
therefore, programmers generally use the descriptor-level functions only
when necessary.  These are some of the usual reasons:

@itemize @bullet
@item
For reading binary files in large chunks.

@item
For reading an entire file into core before parsing it.

@item
To perform operations other than data transfer, which can only be done
with a descriptor.  (You can use @code{fileno} to get the descriptor
corresponding to a stream.)

@item
To pass descriptors to a child process.  (The child can create its own
stream to use a descriptor that it inherits, but cannot inherit a stream
directly.)
@end itemize

@menu
* Opening and Closing Files::           How to open and close file
                                         descriptors.
* Truncating Files::                    Change the size of a file.
* I/O Primitives::                      Reading and writing data.
* File Position Primitive::             Setting a descriptor's file
                                         position.
* Descriptors and Streams::             Converting descriptor to stream
                                         or vice-versa.
* Stream/Descriptor Precautions::       Precautions needed if you use both
                                         descriptors and streams.
* Scatter-Gather::                      Fast I/O to discontinous buffers.
* Memory-mapped I/O::                   Using files like memory.
* Waiting for I/O::                     How to check for input or output
					 on multiple file descriptors.
* Synchronizing I/O::                   Making sure all I/O actions completed.
* Asynchronous I/O::                    Perform I/O in parallel.
* Control Operations::                  Various other operations on file
					 descriptors.
* Duplicating Descriptors::             Fcntl commands for duplicating
                                         file descriptors.
* Descriptor Flags::                    Fcntl commands for manipulating
                                         flags associated with file
                                         descriptors.
* File Status Flags::                   Fcntl commands for manipulating
                                         flags associated with open files.
* File Locks::                          Fcntl commands for implementing
                                         file locking.
* Interrupt Input::                     Getting an asynchronous signal when
                                         input arrives.
* IOCTLs::                              Generic I/O Control operations.
@end menu


@node Opening and Closing Files
@section Opening and Closing Files

@cindex opening a file descriptor
@cindex closing a file descriptor
This section describes the primitives for opening and closing files
using file descriptors.  The @code{open} and @code{creat} functions are
declared in the header file @file{fcntl.h}, while @code{close} is
declared in @file{unistd.h}.
@pindex unistd.h
@pindex fcntl.h

@comment fcntl.h
@comment POSIX.1
@deftypefun int open (const char *@var{filename}, int @var{flags}[, mode_t @var{mode}])
The @code{open} function creates and returns a new file descriptor
for the file named by @var{filename}.  Initially, the file position
indicator for the file is at the beginning of the file.  The argument
@var{mode} is used only when a file is created, but it doesn't hurt
to supply the argument in any case.

The @var{flags} argument controls how the file is to be opened.  This is
a bit mask; you create the value by the bitwise OR of the appropriate
parameters (using the @samp{|} operator in C).
@xref{File Status Flags}, for the parameters available.

The normal return value from @code{open} is a non-negative integer file
descriptor.  In the case of an error, a value of @math{-1} is returned
instead.  In addition to the usual file name errors (@pxref{File
Name Errors}), the following @code{errno} error conditions are defined
for this function:

@table @code
@item EACCES
The file exists but is not readable/writable as requested by the @var{flags}
argument, the file does not exist and the directory is unwritable so
it cannot be created.

@item EEXIST
Both @code{O_CREAT} and @code{O_EXCL} are set, and the named file already
exists.

@item EINTR
The @code{open} operation was interrupted by a signal.
@xref{Interrupted Primitives}.

@item EISDIR
The @var{flags} argument specified write access, and the file is a directory.

@item EMFILE
The process has too many files open.
The maximum number of file descriptors is controlled by the
@code{RLIMIT_NOFILE} resource limit; @pxref{Limits on Resources}.

@item ENFILE
The entire system, or perhaps the file system which contains the
directory, cannot support any additional open files at the moment.
(This problem cannot happen on the GNU system.)

@item ENOENT
The named file does not exist, and @code{O_CREAT} is not specified.

@item ENOSPC
The directory or file system that would contain the new file cannot be
extended, because there is no disk space left.

@item ENXIO
@code{O_NONBLOCK} and @code{O_WRONLY} are both set in the @var{flags}
argument, the file named by @var{filename} is a FIFO (@pxref{Pipes and
FIFOs}), and no process has the file open for reading.

@item EROFS
The file resides on a read-only file system and any of @w{@code{O_WRONLY}},
@code{O_RDWR}, and @code{O_TRUNC} are set in the @var{flags} argument,
or @code{O_CREAT} is set and the file does not already exist.
@end table

@c !!! umask

If on a 32 bits machine the sources are translated with
@code{_FILE_OFFSET_BITS == 64} the function @code{open} returns a file
descriptor opened in the large file mode which enables the file handling
functions to use files up to @math{2^63} bytes in size and offset from
@math{-2^63} to @math{2^63}.  This happens transparently for the user
since all of the lowlevel file handling functions are equally replaced.

This function is a cancelation point in multi-threaded programs.  This
is a problem if the thread allocates some resources (like memory, file
descriptors, semaphores or whatever) at the time @code{open} is
called.  If the thread gets canceled these resources stay allocated
until the program ends.  To avoid this calls to @code{open} should be
protected using cancelation handlers.
@c ref pthread_cleanup_push / pthread_cleanup_pop

The @code{open} function is the underlying primitive for the @code{fopen}
and @code{freopen} functions, that create streams.
@end deftypefun

@comment fcntl.h
@comment Unix98
@deftypefun int open64 (const char *@var{filename}, int @var{flags}[, mode_t @var{mode}])
This function is similar to @code{open}.  It returns a file descriptor
which can be used to access the file named by @var{filename}.  The only
the difference is that on 32 bits systems the file is opened in the
large file mode.  I.e., file length and file offsets can exceed 31 bits.

When the sources are translated with @code{_FILE_OFFSET_BITS == 64} this
function is actually available under the name @code{open}.  I.e., the
new, extended API using 64 bit file sizes and offsets transparently
replaces the old API.
@end deftypefun

@comment fcntl.h
@comment POSIX.1
@deftypefn {Obsolete function} int creat (const char *@var{filename}, mode_t @var{mode})
This function is obsolete.  The call:

@smallexample
creat (@var{filename}, @var{mode})
@end smallexample

@noindent
is equivalent to:

@smallexample
open (@var{filename}, O_WRONLY | O_CREAT | O_TRUNC, @var{mode})
@end smallexample

If on a 32 bits machine the sources are translated with
@code{_FILE_OFFSET_BITS == 64} the function @code{creat} returns a file
descriptor opened in the large file mode which enables the file handling
functions to use files up to @math{2^63} in size and offset from
@math{-2^63} to @math{2^63}.  This happens transparently for the user
since all of the lowlevel file handling functions are equally replaced.
@end deftypefn

@comment fcntl.h
@comment Unix98
@deftypefn {Obsolete function} int creat64 (const char *@var{filename}, mode_t @var{mode})
This function is similar to @code{creat}.  It returns a file descriptor
which can be used to access the file named by @var{filename}.  The only
the difference is that on 32 bits systems the file is opened in the
large file mode.  I.e., file length and file offsets can exceed 31 bits.

To use this file descriptor one must not use the normal operations but
instead the counterparts named @code{*64}, e.g., @code{read64}.

When the sources are translated with @code{_FILE_OFFSET_BITS == 64} this
function is actually available under the name @code{open}.  I.e., the
new, extended API using 64 bit file sizes and offsets transparently
replaces the old API.
@end deftypefn

@comment unistd.h
@comment POSIX.1
@deftypefun int close (int @var{filedes})
The function @code{close} closes the file descriptor @var{filedes}.
Closing a file has the following consequences:

@itemize @bullet
@item
The file descriptor is deallocated.

@item
Any record locks owned by the process on the file are unlocked.

@item
When all file descriptors associated with a pipe or FIFO have been closed,
any unread data is discarded.
@end itemize

This function is a cancelation point in multi-threaded programs.  This
is a problem if the thread allocates some resources (like memory, file
descriptors, semaphores or whatever) at the time @code{close} is
called.  If the thread gets canceled these resources stay allocated
until the program ends.  To avoid this calls to @code{close} should be
protected using cancelation handlers.
@c ref pthread_cleanup_push / pthread_cleanup_pop

The normal return value from @code{close} is @math{0}; a value of @math{-1}
is returned in case of failure.  The following @code{errno} error
conditions are defined for this function:

@table @code
@item EBADF
The @var{filedes} argument is not a valid file descriptor.

@item EINTR
The @code{close} call was interrupted by a signal.
@xref{Interrupted Primitives}.
Here is an example of how to handle @code{EINTR} properly:

@smallexample
TEMP_FAILURE_RETRY (close (desc));
@end smallexample

@item ENOSPC
@itemx EIO
@itemx EDQUOT
When the file is accessed by NFS, these errors from @code{write} can sometimes
not be detected until @code{close}.  @xref{I/O Primitives}, for details
on their meaning.
@end table

Please note that there is @emph{no} separate @code{close64} function.
This is not necessary since this function does not determine nor depend
on the mode of the file.  The kernel which performs the @code{close}
operation knows for which mode the descriptor is used and can handle
this situation.
@end deftypefun

To close a stream, call @code{fclose} (@pxref{Closing Streams}) instead
of trying to close its underlying file descriptor with @code{close}.
This flushes any buffered output and updates the stream object to
indicate that it is closed.

@node I/O Primitives
@section Input and Output Primitives

This section describes the functions for performing primitive input and
output operations on file descriptors: @code{read}, @code{write}, and
@code{lseek}.  These functions are declared in the header file
@file{unistd.h}.
@pindex unistd.h

@comment unistd.h
@comment POSIX.1
@deftp {Data Type} ssize_t
This data type is used to represent the sizes of blocks that can be
read or written in a single operation.  It is similar to @code{size_t},
but must be a signed type.
@end deftp

@cindex reading from a file descriptor
@comment unistd.h
@comment POSIX.1
@deftypefun ssize_t read (int @var{filedes}, void *@var{buffer}, size_t @var{size})
The @code{read} function reads up to @var{size} bytes from the file
with descriptor @var{filedes}, storing the results in the @var{buffer}.
(This is not necessarily a character string and there is no terminating
null character added.)

@cindex end-of-file, on a file descriptor
The return value is the number of bytes actually read.  This might be
less than @var{size}; for example, if there aren't that many bytes left
in the file or if there aren't that many bytes immediately available.
The exact behavior depends on what kind of file it is.  Note that
reading less than @var{size} bytes is not an error.

A value of zero indicates end-of-file (except if the value of the
@var{size} argument is also zero).  This is not considered an error.
If you keep calling @code{read} while at end-of-file, it will keep
returning zero and doing nothing else.

If @code{read} returns at least one character, there is no way you can
tell whether end-of-file was reached.  But if you did reach the end, the
next read will return zero.

In case of an error, @code{read} returns @math{-1}.  The following
@code{errno} error conditions are defined for this function:

@table @code
@item EAGAIN
Normally, when no input is immediately available, @code{read} waits for
some input.  But if the @code{O_NONBLOCK} flag is set for the file
(@pxref{File Status Flags}), @code{read} returns immediately without
reading any data, and reports this error.

@strong{Compatibility Note:} Most versions of BSD Unix use a different
error code for this: @code{EWOULDBLOCK}.  In the GNU library,
@code{EWOULDBLOCK} is an alias for @code{EAGAIN}, so it doesn't matter
which name you use.

On some systems, reading a large amount of data from a character special
file can also fail with @code{EAGAIN} if the kernel cannot find enough
physical memory to lock down the user's pages.  This is limited to
devices that transfer with direct memory access into the user's memory,
which means it does not include terminals, since they always use
separate buffers inside the kernel.  This problem never happens in the
GNU system.

Any condition that could result in @code{EAGAIN} can instead result in a
successful @code{read} which returns fewer bytes than requested.
Calling @code{read} again immediately would result in @code{EAGAIN}.

@item EBADF
The @var{filedes} argument is not a valid file descriptor,
or is not open for reading.

@item EINTR
@code{read} was interrupted by a signal while it was waiting for input.
@xref{Interrupted Primitives}.  A signal will not necessary cause
@code{read} to return @code{EINTR}; it may instead result in a
successful @code{read} which returns fewer bytes than requested.

@item EIO
For many devices, and for disk files, this error code indicates
a hardware error.

@code{EIO} also occurs when a background process tries to read from the
controlling terminal, and the normal action of stopping the process by
sending it a @code{SIGTTIN} signal isn't working.  This might happen if
signal is being blocked or ignored, or because the process group is
orphaned.  @xref{Job Control}, for more information about job control,
and @ref{Signal Handling}, for information about signals.
@end table

Please note that there is no function named @code{read64}.  This is not
necessary since this function does not directly modify or handle the
possibly wide file offset.  Since the kernel handles this state
internally the @code{read} function can be used for all cases.

This function is a cancelation point in multi-threaded programs.  This
is a problem if the thread allocates some resources (like memory, file
descriptors, semaphores or whatever) at the time @code{read} is
called.  If the thread gets canceled these resources stay allocated
until the program ends.  To avoid this calls to @code{read} should be
protected using cancelation handlers.
@c ref pthread_cleanup_push / pthread_cleanup_pop

The @code{read} function is the underlying primitive for all of the
functions that read from streams, such as @code{fgetc}.
@end deftypefun

@comment unistd.h
@comment Unix98
@deftypefun ssize_t pread (int @var{filedes}, void *@var{buffer}, size_t @var{size}, off_t @var{offset})
The @code{pread} function is similar to the @code{read} function.  The
first three arguments are identical and also the return values and error
codes correspond.

The difference is the fourth argument and its handling.  The data block
is not read from the current position of the file descriptor
@code{filedes}.  Instead the data is read from the file starting at
position @var{offset}.  The position of the file descriptor itself is
not effected by the operation.  The value is the same as before the call.

When the source file is compiled with @code{_FILE_OFFSET_BITS == 64} the
@code{pread} function is in fact @code{pread64} and the type
@code{off_t} has 64 bits which makes it possible to handle files up to
@math{2^63} bytes in length.

The return value of @code{pread} describes the number of bytes read.
In the error case it returns @math{-1} like @code{read} does and the
error codes are also the same.  Only there are a few more error codes:
@table @code
@item EINVAL
The value given for @var{offset} is negative and therefore illegal.

@item ESPIPE
The file descriptor @var{filedes} is associate with a pipe or a FIFO and
this device does not allow positioning of the file pointer.
@end table

The function is an extension defined in the Unix Single Specification
version 2.
@end deftypefun

@comment unistd.h
@comment Unix98
@deftypefun ssize_t pread64 (int @var{filedes}, void *@var{buffer}, size_t @var{size}, off64_t @var{offset})
This function is similar to the @code{pread} function.  The difference
is that the @var{offset} parameter is of type @code{off64_t} instead of
@code{off_t} which makes it possible on 32 bits machines to address
files larger than @math{2^31} bytes and up to @math{2^63} bytes.  The
file descriptor @code{filedes} must be opened using @code{open64} since
otherwise the large offsets possible with @code{off64_t} will lead to
errors with a descriptor in small file mode.

When the source file is compiled with @code{_FILE_OFFSET_BITS == 64} on a
32 bits machine this function is actually available under the name
@code{pread} and so transparently replaces the 32 bits interface.
@end deftypefun

@cindex writing to a file descriptor
@comment unistd.h
@comment POSIX.1
@deftypefun ssize_t write (int @var{filedes}, const void *@var{buffer}, size_t @var{size})
The @code{write} function writes up to @var{size} bytes from
@var{buffer} to the file with descriptor @var{filedes}.  The data in
@var{buffer} is not necessarily a character string and a null character is
output like any other character.

The return value is the number of bytes actually written.  This may be
@var{size}, but can always be smaller.  Your program should always call
@code{write} in a loop, iterating until all the data is written.

Once @code{write} returns, the data is enqueued to be written and can be
read back right away, but it is not necessarily written out to permanent
storage immediately.  You can use @code{fsync} when you need to be sure
your data has been permanently stored before continuing.  (It is more
efficient for the system to batch up consecutive writes and do them all
at once when convenient.  Normally they will always be written to disk
within a minute or less.)  Modern systems provide another function
@code{fdatasync} which guarantees integrity only for the file data and
is therefore faster.
@c !!! xref fsync, fdatasync
You can use the @code{O_FSYNC} open mode to make @code{write} always
store the data to disk before returning; @pxref{Operating Modes}.

In the case of an error, @code{write} returns @math{-1}.  The following
@code{errno} error conditions are defined for this function:

@table @code
@item EAGAIN
Normally, @code{write} blocks until the write operation is complete.
But if the @code{O_NONBLOCK} flag is set for the file (@pxref{Control
Operations}), it returns immediately without writing any data, and
reports this error.  An example of a situation that might cause the
process to block on output is writing to a terminal device that supports
flow control, where output has been suspended by receipt of a STOP
character.

@strong{Compatibility Note:} Most versions of BSD Unix use a different
error code for this: @code{EWOULDBLOCK}.  In the GNU library,
@code{EWOULDBLOCK} is an alias for @code{EAGAIN}, so it doesn't matter
which name you use.

On some systems, writing a large amount of data from a character special
file can also fail with @code{EAGAIN} if the kernel cannot find enough
physical memory to lock down the user's pages.  This is limited to
devices that transfer with direct memory access into the user's memory,
which means it does not include terminals, since they always use
separate buffers inside the kernel.  This problem does not arise in the
GNU system.

@item EBADF
The @var{filedes} argument is not a valid file descriptor,
or is not open for writing.

@item EFBIG
The size of the file would become larger than the implementation can support.

@item EINTR
The @code{write} operation was interrupted by a signal while it was
blocked waiting for completion.  A signal will not necessary cause
@code{write} to return @code{EINTR}; it may instead result in a
successful @code{write} which writes fewer bytes than requested.
@xref{Interrupted Primitives}.

@item EIO
For many devices, and for disk files, this error code indicates
a hardware error.

@item ENOSPC
The device containing the file is full.

@item EPIPE
This error is returned when you try to write to a pipe or FIFO that
isn't open for reading by any process.  When this happens, a @code{SIGPIPE}
signal is also sent to the process; see @ref{Signal Handling}.
@end table

Unless you have arranged to prevent @code{EINTR} failures, you should
check @code{errno} after each failing call to @code{write}, and if the
error was @code{EINTR}, you should simply repeat the call.
@xref{Interrupted Primitives}.  The easy way to do this is with the
macro @code{TEMP_FAILURE_RETRY}, as follows:

@smallexample
nbytes = TEMP_FAILURE_RETRY (write (desc, buffer, count));
@end smallexample

Please note that there is no function named @code{write64}.  This is not
necessary since this function does not directly modify or handle the
possibly wide file offset.  Since the kernel handles this state
internally the @code{write} function can be used for all cases.

This function is a cancelation point in multi-threaded programs.  This
is a problem if the thread allocates some resources (like memory, file
descriptors, semaphores or whatever) at the time @code{write} is
called.  If the thread gets canceled these resources stay allocated
until the program ends.  To avoid this calls to @code{write} should be
protected using cancelation handlers.
@c ref pthread_cleanup_push / pthread_cleanup_pop

The @code{write} function is the underlying primitive for all of the
functions that write to streams, such as @code{fputc}.
@end deftypefun

@comment unistd.h
@comment Unix98
@deftypefun ssize_t pwrite (int @var{filedes}, const void *@var{buffer}, size_t @var{size}, off_t @var{offset})
The @code{pwrite} function is similar to the @code{write} function.  The
first three arguments are identical and also the return values and error
codes correspond.

The difference is the fourth argument and its handling.  The data block
is not written to the current position of the file descriptor
@code{filedes}.  Instead the data is written to the file starting at
position @var{offset}.  The position of the file descriptor itself is
not effected by the operation.  The value is the same as before the call.

When the source file is compiled with @code{_FILE_OFFSET_BITS == 64} the
@code{pwrite} function is in fact @code{pwrite64} and the type
@code{off_t} has 64 bits which makes it possible to handle files up to
@math{2^63} bytes in length.

The return value of @code{pwrite} describes the number of written bytes.
In the error case it returns @math{-1} like @code{write} does and the
error codes are also the same.  Only there are a few more error codes:
@table @code
@item EINVAL
The value given for @var{offset} is negative and therefore illegal.

@item ESPIPE
The file descriptor @var{filedes} is associate with a pipe or a FIFO and
this device does not allow positioning of the file pointer.
@end table

The function is an extension defined in the Unix Single Specification
version 2.
@end deftypefun

@comment unistd.h
@comment Unix98
@deftypefun ssize_t pwrite64 (int @var{filedes}, const void *@var{buffer}, size_t @var{size}, off64_t @var{offset})
This function is similar to the @code{pwrite} function.  The difference
is that the @var{offset} parameter is of type @code{off64_t} instead of
@code{off_t} which makes it possible on 32 bits machines to address
files larger than @math{2^31} bytes and up to @math{2^63} bytes.  The
file descriptor @code{filedes} must be opened using @code{open64} since
otherwise the large offsets possible with @code{off64_t} will lead to
errors with a descriptor in small file mode.

When the source file is compiled using @code{_FILE_OFFSET_BITS == 64} on a
32 bits machine this function is actually available under the name
@code{pwrite} and so transparently replaces the 32 bits interface.
@end deftypefun


@node File Position Primitive
@section Setting the File Position of a Descriptor

Just as you can set the file position of a stream with @code{fseek}, you
can set the file position of a descriptor with @code{lseek}.  This
specifies the position in the file for the next @code{read} or
@code{write} operation.  @xref{File Positioning}, for more information
on the file position and what it means.

To read the current file position value from a descriptor, use
@code{lseek (@var{desc}, 0, SEEK_CUR)}.

@cindex file positioning on a file descriptor
@cindex positioning a file descriptor
@cindex seeking on a file descriptor
@comment unistd.h
@comment POSIX.1
@deftypefun off_t lseek (int @var{filedes}, off_t @var{offset}, int @var{whence})
The @code{lseek} function is used to change the file position of the
file with descriptor @var{filedes}.

The @var{whence} argument specifies how the @var{offset} should be
interpreted in the same way as for the @code{fseek} function, and must be
one of the symbolic constants @code{SEEK_SET}, @code{SEEK_CUR}, or
@code{SEEK_END}.

@table @code
@item SEEK_SET
Specifies that @var{whence} is a count of characters from the beginning
of the file.

@item SEEK_CUR
Specifies that @var{whence} is a count of characters from the current
file position.  This count may be positive or negative.

@item SEEK_END
Specifies that @var{whence} is a count of characters from the end of
the file.  A negative count specifies a position within the current
extent of the file; a positive count specifies a position past the
current end.  If you set the position past the current end, and
actually write data, you will extend the file with zeros up to that
position.
@end table

The return value from @code{lseek} is normally the resulting file
position, measured in bytes from the beginning of the file.
You can use this feature together with @code{SEEK_CUR} to read the
current file position.

If you want to append to the file, setting the file position to the
current end of file with @code{SEEK_END} is not sufficient.  Another
process may write more data after you seek but before you write,
extending the file so the position you write onto clobbers their data.
Instead, use the @code{O_APPEND} operating mode; @pxref{Operating Modes}.

You can set the file position past the current end of the file.  This
does not by itself make the file longer; @code{lseek} never changes the
file.  But subsequent output at that position will extend the file.
Characters between the previous end of file and the new position are
filled with zeros.  Extending the file in this way can create a
``hole'': the blocks of zeros are not actually allocated on disk, so the
file takes up less space than it appears so; it is then called a
``sparse file''.
@cindex sparse files
@cindex holes in files

If the file position cannot be changed, or the operation is in some way
invalid, @code{lseek} returns a value of @math{-1}.  The following
@code{errno} error conditions are defined for this function:

@table @code
@item EBADF
The @var{filedes} is not a valid file descriptor.

@item EINVAL
The @var{whence} argument value is not valid, or the resulting
file offset is not valid.  A file offset is invalid.

@item ESPIPE
The @var{filedes} corresponds to an object that cannot be positioned,
such as a pipe, FIFO or terminal device.  (POSIX.1 specifies this error
only for pipes and FIFOs, but in the GNU system, you always get
@code{ESPIPE} if the object is not seekable.)
@end table

When the source file is compiled with @code{_FILE_OFFSET_BITS == 64} the
@code{lseek} function is in fact @code{lseek64} and the type
@code{off_t} has 64 bits which makes it possible to handle files up to
@math{2^63} bytes in length.

This function is a cancelation point in multi-threaded programs.  This
is a problem if the thread allocates some resources (like memory, file
descriptors, semaphores or whatever) at the time @code{lseek} is
called.  If the thread gets canceled these resources stay allocated
until the program ends.  To avoid this calls to @code{lseek} should be
protected using cancelation handlers.
@c ref pthread_cleanup_push / pthread_cleanup_pop

The @code{lseek} function is the underlying primitive for the
@code{fseek}, @code{fseeko}, @code{ftell}, @code{ftello} and
@code{rewind} functions, which operate on streams instead of file
descriptors.
@end deftypefun

@comment unistd.h
@comment Unix98
@deftypefun off64_t lseek64 (int @var{filedes}, off64_t @var{offset}, int @var{whence})
This function is similar to the @code{lseek} function.  The difference
is that the @var{offset} parameter is of type @code{off64_t} instead of
@code{off_t} which makes it possible on 32 bits machines to address
files larger than @math{2^31} bytes and up to @math{2^63} bytes.  The
file descriptor @code{filedes} must be opened using @code{open64} since
otherwise the large offsets possible with @code{off64_t} will lead to
errors with a descriptor in small file mode.

When the source file is compiled with @code{_FILE_OFFSET_BITS == 64} on a
32 bits machine this function is actually available under the name
@code{lseek} and so transparently replaces the 32 bits interface.
@end deftypefun

You can have multiple descriptors for the same file if you open the file
more than once, or if you duplicate a descriptor with @code{dup}.
Descriptors that come from separate calls to @code{open} have independent
file positions; using @code{lseek} on one descriptor has no effect on the
other.  For example,

@smallexample
@group
@{
  int d1, d2;
  char buf[4];
  d1 = open ("foo", O_RDONLY);
  d2 = open ("foo", O_RDONLY);
  lseek (d1, 1024, SEEK_SET);
  read (d2, buf, 4);
@}
@end group
@end smallexample

@noindent
will read the first four characters of the file @file{foo}.  (The
error-checking code necessary for a real program has been omitted here
for brevity.)

By contrast, descriptors made by duplication share a common file
position with the original descriptor that was duplicated.  Anything
which alters the file position of one of the duplicates, including
reading or writing data, affects all of them alike.  Thus, for example,

@smallexample
@{
  int d1, d2, d3;
  char buf1[4], buf2[4];
  d1 = open ("foo", O_RDONLY);
  d2 = dup (d1);
  d3 = dup (d2);
  lseek (d3, 1024, SEEK_SET);
  read (d1, buf1, 4);
  read (d2, buf2, 4);
@}
@end smallexample

@noindent
will read four characters starting with the 1024'th character of
@file{foo}, and then four more characters starting with the 1028'th
character.

@comment sys/types.h
@comment POSIX.1
@deftp {Data Type} off_t
This is an arithmetic data type used to represent file sizes.
In the GNU system, this is equivalent to @code{fpos_t} or @code{long int}.

If the source is compiled with @code{_FILE_OFFSET_BITS == 64} this type
is transparently replaced by @code{off64_t}.
@end deftp

@comment sys/types.h
@comment Unix98
@deftp {Data Type} off64_t
This type is used similar to @code{off_t}.  The difference is that even
on 32 bits machines, where the @code{off_t} type would have 32 bits,
@code{off64_t} has 64 bits and so is able to address files up to
@math{2^63} bytes in length.

When compiling with @code{_FILE_OFFSET_BITS == 64} this type is
available under the name @code{off_t}.
@end deftp

These aliases for the @samp{SEEK_@dots{}} constants exist for the sake
of compatibility with older BSD systems.  They are defined in two
different header files: @file{fcntl.h} and @file{sys/file.h}.

@table @code
@item L_SET
An alias for @code{SEEK_SET}.

@item L_INCR
An alias for @code{SEEK_CUR}.

@item L_XTND
An alias for @code{SEEK_END}.
@end table

@node Descriptors and Streams
@section Descriptors and Streams
@cindex streams, and file descriptors
@cindex converting file descriptor to stream
@cindex extracting file descriptor from stream

Given an open file descriptor, you can create a stream for it with the
@code{fdopen} function.  You can get the underlying file descriptor for
an existing stream with the @code{fileno} function.  These functions are
declared in the header file @file{stdio.h}.
@pindex stdio.h

@comment stdio.h
@comment POSIX.1
@deftypefun {FILE *} fdopen (int @var{filedes}, const char *@var{opentype})
The @code{fdopen} function returns a new stream for the file descriptor
@var{filedes}.

The @var{opentype} argument is interpreted in the same way as for the
@code{fopen} function (@pxref{Opening Streams}), except that
the @samp{b} option is not permitted; this is because GNU makes no
distinction between text and binary files.  Also, @code{"w"} and
@code{"w+"} do not cause truncation of the file; these have affect only
when opening a file, and in this case the file has already been opened.
You must make sure that the @var{opentype} argument matches the actual
mode of the open file descriptor.

The return value is the new stream.  If the stream cannot be created
(for example, if the modes for the file indicated by the file descriptor
do not permit the access specified by the @var{opentype} argument), a
null pointer is returned instead.

In some other systems, @code{fdopen} may fail to detect that the modes
for file descriptor do not permit the access specified by
@code{opentype}.  The GNU C library always checks for this.
@end deftypefun

For an example showing the use of the @code{fdopen} function,
see @ref{Creating a Pipe}.

@comment stdio.h
@comment POSIX.1
@deftypefun int fileno (FILE *@var{stream})
This function returns the file descriptor associated with the stream
@var{stream}.  If an error is detected (for example, if the @var{stream}
is not valid) or if @var{stream} does not do I/O to a file,
@code{fileno} returns @math{-1}.
@end deftypefun

@cindex standard file descriptors
@cindex file descriptors, standard
There are also symbolic constants defined in @file{unistd.h} for the
file descriptors belonging to the standard streams @code{stdin},
@code{stdout}, and @code{stderr}; see @ref{Standard Streams}.
@pindex unistd.h

@comment unistd.h
@comment POSIX.1
@table @code
@item STDIN_FILENO
@vindex STDIN_FILENO
This macro has value @code{0}, which is the file descriptor for
standard input.
@cindex standard input file descriptor

@comment unistd.h
@comment POSIX.1
@item STDOUT_FILENO
@vindex STDOUT_FILENO
This macro has value @code{1}, which is the file descriptor for
standard output.
@cindex standard output file descriptor

@comment unistd.h
@comment POSIX.1
@item STDERR_FILENO
@vindex STDERR_FILENO
This macro has value @code{2}, which is the file descriptor for
standard error output.
@end table
@cindex standard error file descriptor

@node Stream/Descriptor Precautions
@section Dangers of Mixing Streams and Descriptors
@cindex channels
@cindex streams and descriptors
@cindex descriptors and streams
@cindex mixing descriptors and streams

You can have multiple file descriptors and streams (let's call both
streams and descriptors ``channels'' for short) connected to the same
file, but you must take care to avoid confusion between channels.  There
are two cases to consider: @dfn{linked} channels that share a single
file position value, and @dfn{independent} channels that have their own
file positions.

It's best to use just one channel in your program for actual data
transfer to any given file, except when all the access is for input.
For example, if you open a pipe (something you can only do at the file
descriptor level), either do all I/O with the descriptor, or construct a
stream from the descriptor with @code{fdopen} and then do all I/O with
the stream.

@menu
* Linked Channels::	   Dealing with channels sharing a file position.
* Independent Channels::   Dealing with separately opened, unlinked channels.
* Cleaning Streams::	   Cleaning a stream makes it safe to use
                            another channel.
@end menu

@node Linked Channels
@subsection Linked Channels
@cindex linked channels

Channels that come from a single opening share the same file position;
we call them @dfn{linked} channels.  Linked channels result when you
make a stream from a descriptor using @code{fdopen}, when you get a
descriptor from a stream with @code{fileno}, when you copy a descriptor
with @code{dup} or @code{dup2}, and when descriptors are inherited
during @code{fork}.  For files that don't support random access, such as
terminals and pipes, @emph{all} channels are effectively linked.  On
random-access files, all append-type output streams are effectively
linked to each other.

@cindex cleaning up a stream
If you have been using a stream for I/O, and you want to do I/O using
another channel (either a stream or a descriptor) that is linked to it,
you must first @dfn{clean up} the stream that you have been using.
@xref{Cleaning Streams}.

Terminating a process, or executing a new program in the process,
destroys all the streams in the process.  If descriptors linked to these
streams persist in other processes, their file positions become
undefined as a result.  To prevent this, you must clean up the streams
before destroying them.

@node Independent Channels
@subsection Independent Channels
@cindex independent channels

When you open channels (streams or descriptors) separately on a seekable
file, each channel has its own file position.  These are called
@dfn{independent channels}.

The system handles each channel independently.  Most of the time, this
is quite predictable and natural (especially for input): each channel
can read or write sequentially at its own place in the file.  However,
if some of the channels are streams, you must take these precautions:

@itemize @bullet
@item
You should clean an output stream after use, before doing anything else
that might read or write from the same part of the file.

@item
You should clean an input stream before reading data that may have been
modified using an independent channel.  Otherwise, you might read
obsolete data that had been in the stream's buffer.
@end itemize

If you do output to one channel at the end of the file, this will
certainly leave the other independent channels positioned somewhere
before the new end.  You cannot reliably set their file positions to the
new end of file before writing, because the file can always be extended
by another process between when you set the file position and when you
write the data.  Instead, use an append-type descriptor or stream; they
always output at the current end of the file.  In order to make the
end-of-file position accurate, you must clean the output channel you
were using, if it is a stream.

It's impossible for two channels to have separate file pointers for a
file that doesn't support random access.  Thus, channels for reading or
writing such files are always linked, never independent.  Append-type
channels are also always linked.  For these channels, follow the rules
for linked channels; see @ref{Linked Channels}.

@node Cleaning Streams
@subsection Cleaning Streams

On the GNU system, you can clean up any stream with @code{fclean}:

@comment stdio.h
@comment GNU
@deftypefun int fclean (FILE *@var{stream})
Clean up the stream @var{stream} so that its buffer is empty.  If
@var{stream} is doing output, force it out.  If @var{stream} is doing
input, give the data in the buffer back to the system, arranging to
reread it.
@end deftypefun

On other systems, you can use @code{fflush} to clean a stream in most
cases.

You can skip the @code{fclean} or @code{fflush} if you know the stream
is already clean.  A stream is clean whenever its buffer is empty.  For
example, an unbuffered stream is always clean.  An input stream that is
at end-of-file is clean.  A line-buffered stream is clean when the last
character output was a newline.

There is one case in which cleaning a stream is impossible on most
systems.  This is when the stream is doing input from a file that is not
random-access.  Such streams typically read ahead, and when the file is
not random access, there is no way to give back the excess data already
read.  When an input stream reads from a random-access file,
@code{fflush} does clean the stream, but leaves the file pointer at an
unpredictable place; you must set the file pointer before doing any
further I/O.  On the GNU system, using @code{fclean} avoids both of
these problems.

Closing an output-only stream also does @code{fflush}, so this is a
valid way of cleaning an output stream.  On the GNU system, closing an
input stream does @code{fclean}.

You need not clean a stream before using its descriptor for control
operations such as setting terminal modes; these operations don't affect
the file position and are not affected by it.  You can use any
descriptor for these operations, and all channels are affected
simultaneously.  However, text already ``output'' to a stream but still
buffered by the stream will be subject to the new terminal modes when
subsequently flushed.  To make sure ``past'' output is covered by the
terminal settings that were in effect at the time, flush the output
streams for that terminal before setting the modes.  @xref{Terminal
Modes}.

@node Scatter-Gather
@section Fast Scatter-Gather I/O
@cindex scatter-gather

Some applications may need to read or write data to multiple buffers,
which are seperated in memory.  Although this can be done easily enough
with multiple calls to @code{read} and @code{write}, it is inefficent
because there is overhead associated with each kernel call.

Instead, many platforms provide special high-speed primitives to perform
these @dfn{scatter-gather} operations in a single kernel call.  The GNU C
library will provide an emulation on any system that lacks these
primitives, so they are not a portability threat.  They are defined in
@code{sys/uio.h}.

These functions are controlled with arrays of @code{iovec} structures,
which describe the location and size of each buffer.

@deftp {Data Type} {struct iovec}

The @code{iovec} structure describes a buffer. It contains two fields:

@table @code

@item void *iov_base
Contains the address of a buffer.

@item size_t iov_len
Contains the length of the buffer.

@end table
@end deftp

@deftypefun ssize_t readv (int @var{filedes}, const struct iovec *@var{vector}, int @var{count})

The @code{readv} function reads data from @var{filedes} and scatters it
into the buffers described in @var{vector}, which is taken to be
@var{count} structures long.  As each buffer is filled, data is sent to the
next.

Note that @code{readv} is not guaranteed to fill all the buffers.
It may stop at any point, for the same reasons @code{read} would.

The return value is a count of bytes (@emph{not} buffers) read, @math{0}
indicating end-of-file, or @math{-1} indicating an error.  The possible
errors are the same as in @code{read}.

@end deftypefun

@deftypefun ssize_t writev (int @var{filedes}, const struct iovec *@var{vector}, int @var{count})

The @code{writev} function gathers data from the buffers described in
@var{vector}, which is taken to be @var{count} structures long, and writes
them to @code{filedes}.  As each buffer is written, it moves on to the
next.

Like @code{readv}, @code{writev} may stop midstream under the same
conditions @code{write} would.

The return value is a count of bytes written, or @math{-1} indicating an
error.  The possible errors are the same as in @code{write}.

@end deftypefun

@c Note - I haven't read this anywhere. I surmised it from my knowledge
@c of computer science. Thus, there could be subtleties I'm missing.

Note that if the buffers are small (under about 1kB), high-level streams
may be easier to use than these functions.  However, @code{readv} and
@code{writev} are more efficient when the individual buffers themselves
(as opposed to the total output), are large.  In that case, a high-level
stream would not be able to cache the data effectively.

@node Memory-mapped I/O
@section Memory-mapped I/O

On modern operating systems, it is possible to @dfn{mmap} (pronounced
``em-map'') a file to a region of memory.  When this is done, the file can
be accessed just like an array in the program.

This is more efficent than @code{read} or @code{write}, as only regions
of the file a program actually accesses are loaded.  Accesses to
not-yet-loaded parts of the mmapped region are handled in the same way as
swapped out pages.

Since mmapped pages can be stored back to their file when physical memory
is low, it is possible to mmap files orders of magnitude larger than both
the physical memory @emph{and} swap space.  The only limit is address
space.  The theoretical limit is 4GB on a 32-bit machine - however, the
actual limit will be smaller since some areas will be reserved for other
purposes.

Memory mapping only works on entire pages of memory.  Thus, addresses
for mapping must be page-aligned, and length values will be rounded up.
To determine the size of a page the machine uses one should use

@smallexample
size_t page_size = (size_t) sysconf (_SC_PAGESIZE);
@end smallexample

These functions are declared in @file{sys/mman.h}.

@deftypefun {void *} mmap (void *@var{address}, size_t @var{length},int @var{protect}, int @var{flags}, int @var{filedes}, off_t @var{offset})

The @code{mmap} function creates a new mapping, connected to bytes
(@var{offset}) to (@var{offset} + @var{length}) in the file open on
@var{filedes}.

@var{address} gives a preferred starting address for the mapping.
@code{NULL} expresses no preference. Any previous mapping at that
address is automatically removed. The address you give may still be
changed, unless you use the @code{MAP_FIXED} flag.

@vindex PROT_READ
@vindex PROT_WRITE
@vindex PROT_EXEC
@var{protect} contains flags that control what kind of access is
permitted.  They include @code{PROT_READ}, @code{PROT_WRITE}, and
@code{PROT_EXEC}, which permit reading, writing, and execution,
respectively.  Inappropriate access will cause a segfault (@pxref{Program
Error Signals}).

Note that most hardware designs cannot support write permission without
read permission, and many do not distinguish read and execute permission.
Thus, you may recieve wider permissions than you ask for, and mappings of
write-only files may be denied even if you do not use @code{PROT_READ}.

@var{flags} contains flags that control the nature of the map.
One of @code{MAP_SHARED} or @code{MAP_PRIVATE} must be specified.

They include:

@vtable @code
@item MAP_PRIVATE
This specifies that writes to the region should never be written back
to the attached file.  Instead, a copy is made for the process, and the
region will be swapped normally if memory runs low.  No other process will
see the changes.

Since private mappings effectively revert to ordinary memory
when written to, you must have enough virtual memory for a copy of
the entire mmapped region if you use this mode with @code{PROT_WRITE}.

@item MAP_SHARED
This specifies that writes to the region will be written back to the
file.  Changes made will be shared immediately with other processes
mmaping the same file.

Note that actual writing may take place at any time.  You need to use
@code{msync}, described below, if it is important that other processes
using conventional I/O get a consistent view of the file.

@item MAP_FIXED
This forces the system to use the exact mapping address specified in
@var{address} and fail if it can't.

@c One of these is official - the other is obviously an obsolete synonym
@c Which is which?
@item MAP_ANONYMOUS
@itemx MAP_ANON
This flag tells the system to create an anonymous mapping, not connected
to a file.  @var{filedes} and @var{off} are ignored, and the region is
initialized with zeros.

Anonymous maps are used as the basic primitive to extend the heap on some
systems.  They are also useful to share data between multiple tasks
without creating a file.

On some systems using private anonymous mmaps is more efficent than using
@code{malloc} for large blocks.  This is not an issue with the GNU C library,
as the included @code{malloc} automatically uses @code{mmap} where appropriate.

@c Linux has some other MAP_ options, which I have not discussed here.
@c MAP_DENYWRITE, MAP_EXECUTABLE and MAP_GROWSDOWN don't seem applicable to
@c user programs (and I don't understand the last two). MAP_LOCKED does
@c not appear to be implemented.

@end vtable

@code{mmap} returns the address of the new mapping, or @math{-1} for an
error.

Possible errors include:

@table @code

@item EINVAL

Either @var{address} was unusable, or inconsistent @var{flags} were
given.

@item EACCES

@var{filedes} was not open for the type of access specified in @var{protect}.

@item ENOMEM

Either there is not enough memory for the operation, or the process is
out of address space.

@item ENODEV

This file is of a type that doesn't support mapping.

@item ENOEXEC

The file is on a filesystem that doesn't support mapping.

@c On Linux, EAGAIN will appear if the file has a conflicting mandatory lock.
@c However mandatory locks are not discussed in this manual.
@c
@c Similarly, ETXTBSY will occur if the MAP_DENYWRITE flag (not documented
@c here) is used and the file is already open for writing.

@end table

@end deftypefun

@deftypefun int munmap (void *@var{addr}, size_t @var{length})

@code{munmap} removes any memory maps from (@var{addr}) to (@var{addr} +
@var{length}).  @var{length} should be the length of the mapping.

It is safe to un-map multiple mappings in one command, or include unmapped
space in the range.  It is also possible to unmap only part of an existing
mapping, however only entire pages can be removed.  If @var{length} is not
an even number of pages, it will be rounded up.

It returns @math{0} for success and @math{-1} for an error.

One error is possible:

@table @code

@item EINVAL
The memory range given was outside the user mmap range, or wasn't page
aligned.

@end table

@end deftypefun

@deftypefun int msync (void *@var{address}, size_t @var{length}, int @var{flags})

When using shared mappings, the kernel can write the file at any time
before the mapping is removed.  To be certain data has actually been
written to the file and will be accessable to non-memory-mapped I/O, it
is neccessary to use this function.

It operates on the region @var{address} to (@var{address} + @var{length}).
It may be used on part of a mapping or multiple mappings, however the
region given should not contain any unmapped space.

@var{flags} can contain some options:

@vtable @code

@item MS_SYNC

This flag makes sure the data is actually written @emph{to disk}.
Normally @code{msync} only makes sure that accesses to a file with
conventional I/O reflect the recent changes.

@item MS_ASYNC

This tells @code{msync} to begin the synchronization, but not to wait for
it to complete.

@c Linux also has MS_INVALIDATE, which I don't understand.

@end vtable

@code{msync} returns @math{0} for success and @math{-1} for
error.  Errors include:

@table @code

@item EINVAL
An invalid region was given, or the @var{flags} were invalid.

@item EFAULT
There is no existing mapping in at least part of the given region.

@end table

@end deftypefun

@deftypefun {void *} mremap (void *@var{address}, size_t @var{length}, size_t @var{new_length}, int @var{flag})

This function can be used to change the size of an existing memory
area. @var{address} and @var{length} must cover a region entirely mapped
in the same @code{mmap} statement. A new mapping with the same
characteristics will be returned, but a with the length @var{new_length}
instead.

One option is possible, @code{MREMAP_MAYMOVE}. If it is given in
@var{flags}, the system may remove the existing mapping and create a new
one of the desired length in another location.

The address of the resulting mapping is returned, or @math{-1}. Possible
error codes include:

This function is only available on a few systems.  Except for performing
optional optimizations one should not rely on this function.
@table @code

@item EFAULT
There is no existing mapping in at least part of the original region, or
the region covers two or more distinct mappings.

@item EINVAL
The address given is misaligned or inappropriate.

@item EAGAIN
The region has pages locked, and if extended it would exceed the
process's resource limit for locked pages.  @xref{Limits on Resources}.

@item ENOMEM
The region is private writable, and insufficent virtual memory is
available to extend it.  Also, this error will occur if
@code{MREMAP_MAYMOVE} is not given and the extension would collide with
another mapped region.

@end table
@end deftypefun

Not all file descriptors may be mapped.  Sockets, pipes, and most devices
only allow sequential access and do not fit into the mapping abstraction.
In addition, some regular files may not be mmapable, and older kernels may
not support mapping at all.  Thus, programs using @code{mmap} should
have a fallback method to use should it fail. @xref{Mmap,,,standards,GNU
Coding Standards}.

@c XXX madvice documentation missing

@node Waiting for I/O
@section Waiting for Input or Output
@cindex waiting for input or output
@cindex multiplexing input
@cindex input from multiple files

Sometimes a program needs to accept input on multiple input channels
whenever input arrives.  For example, some workstations may have devices
such as a digitizing tablet, function button box, or dial box that are
connected via normal asynchronous serial interfaces; good user interface
style requires responding immediately to input on any device.  Another
example is a program that acts as a server to several other processes
via pipes or sockets.

You cannot normally use @code{read} for this purpose, because this
blocks the program until input is available on one particular file
descriptor; input on other channels won't wake it up.  You could set
nonblocking mode and poll each file descriptor in turn, but this is very
inefficient.

A better solution is to use the @code{select} function.  This blocks the
program until input or output is ready on a specified set of file
descriptors, or until a timer expires, whichever comes first.  This
facility is declared in the header file @file{sys/types.h}.
@pindex sys/types.h

In the case of a server socket (@pxref{Listening}), we say that
``input'' is available when there are pending connections that could be
accepted (@pxref{Accepting Connections}).  @code{accept} for server
sockets blocks and interacts with @code{select} just as @code{read} does
for normal input.

@cindex file descriptor sets, for @code{select}
The file descriptor sets for the @code{select} function are specified
as @code{fd_set} objects.  Here is the description of the data type
and some macros for manipulating these objects.

@comment sys/types.h
@comment BSD
@deftp {Data Type} fd_set
The @code{fd_set} data type represents file descriptor sets for the
@code{select} function.  It is actually a bit array.
@end deftp

@comment sys/types.h
@comment BSD
@deftypevr Macro int FD_SETSIZE
The value of this macro is the maximum number of file descriptors that a
@code{fd_set} object can hold information about.  On systems with a
fixed maximum number, @code{FD_SETSIZE} is at least that number.  On
some systems, including GNU, there is no absolute limit on the number of
descriptors open, but this macro still has a constant value which
controls the number of bits in an @code{fd_set}; if you get a file
descriptor with a value as high as @code{FD_SETSIZE}, you cannot put
that descriptor into an @code{fd_set}.
@end deftypevr

@comment sys/types.h
@comment BSD
@deftypefn Macro void FD_ZERO (fd_set *@var{set})
This macro initializes the file descriptor set @var{set} to be the
empty set.
@end deftypefn

@comment sys/types.h
@comment BSD
@deftypefn Macro void FD_SET (int @var{filedes}, fd_set *@var{set})
This macro adds @var{filedes} to the file descriptor set @var{set}.
@end deftypefn

@comment sys/types.h
@comment BSD
@deftypefn Macro void FD_CLR (int @var{filedes}, fd_set *@var{set})
This macro removes @var{filedes} from the file descriptor set @var{set}.
@end deftypefn

@comment sys/types.h
@comment BSD
@deftypefn Macro int FD_ISSET (int @var{filedes}, fd_set *@var{set})
This macro returns a nonzero value (true) if @var{filedes} is a member
of the file descriptor set @var{set}, and zero (false) otherwise.
@end deftypefn

Next, here is the description of the @code{select} function itself.

@comment sys/types.h
@comment BSD
@deftypefun int select (int @var{nfds}, fd_set *@var{read-fds}, fd_set *@var{write-fds}, fd_set *@var{except-fds}, struct timeval *@var{timeout})
The @code{select} function blocks the calling process until there is
activity on any of the specified sets of file descriptors, or until the
timeout period has expired.

The file descriptors specified by the @var{read-fds} argument are
checked to see if they are ready for reading; the @var{write-fds} file
descriptors are checked to see if they are ready for writing; and the
@var{except-fds} file descriptors are checked for exceptional
conditions.  You can pass a null pointer for any of these arguments if
you are not interested in checking for that kind of condition.

A file descriptor is considered ready for reading if it is not at end of
file.  A server socket is considered ready for reading if there is a
pending connection which can be accepted with @code{accept};
@pxref{Accepting Connections}.  A client socket is ready for writing when
its connection is fully established; @pxref{Connecting}.

``Exceptional conditions'' does not mean errors---errors are reported
immediately when an erroneous system call is executed, and do not
constitute a state of the descriptor.  Rather, they include conditions
such as the presence of an urgent message on a socket.  (@xref{Sockets},
for information on urgent messages.)

The @code{select} function checks only the first @var{nfds} file
descriptors.  The usual thing is to pass @code{FD_SETSIZE} as the value
of this argument.

The @var{timeout} specifies the maximum time to wait.  If you pass a
null pointer for this argument, it means to block indefinitely until one
of the file descriptors is ready.  Otherwise, you should provide the
time in @code{struct timeval} format; see @ref{High-Resolution
Calendar}.  Specify zero as the time (a @code{struct timeval} containing
all zeros) if you want to find out which descriptors are ready without
waiting if none are ready.

The normal return value from @code{select} is the total number of ready file
descriptors in all of the sets.  Each of the argument sets is overwritten
with information about the descriptors that are ready for the corresponding
operation.  Thus, to see if a particular descriptor @var{desc} has input,
use @code{FD_ISSET (@var{desc}, @var{read-fds})} after @code{select} returns.

If @code{select} returns because the timeout period expires, it returns
a value of zero.

Any signal will cause @code{select} to return immediately.  So if your
program uses signals, you can't rely on @code{select} to keep waiting
for the full time specified.  If you want to be sure of waiting for a
particular amount of time, you must check for @code{EINTR} and repeat
the @code{select} with a newly calculated timeout based on the current
time.  See the example below.  See also @ref{Interrupted Primitives}.

If an error occurs, @code{select} returns @code{-1} and does not modify
the argument file descriptor sets.  The following @code{errno} error
conditions are defined for this function:

@table @code
@item EBADF
One of the file descriptor sets specified an invalid file descriptor.

@item EINTR
The operation was interrupted by a signal.  @xref{Interrupted Primitives}.

@item EINVAL
The @var{timeout} argument is invalid; one of the components is negative
or too large.
@end table
@end deftypefun

@strong{Portability Note:}  The @code{select} function is a BSD Unix
feature.

Here is an example showing how you can use @code{select} to establish a
timeout period for reading from a file descriptor.  The @code{input_timeout}
function blocks the calling process until input is available on the
file descriptor, or until the timeout period expires.

@smallexample
@include select.c.texi
@end smallexample

There is another example showing the use of @code{select} to multiplex
input from multiple sockets in @ref{Server Example}.


@node Synchronizing I/O
@section Synchronizing I/O operations

@cindex synchronizing
In most modern operation systems the normal I/O operations are not
executed synchronously.  I.e., even if a @code{write} system call
returns this does not mean the data is actually written to the media,
e.g., the disk.

In situations where synchronization points are necessary the user can
use special functions which ensure that all operations finished before
they return.

@comment unistd.h
@comment X/Open
@deftypefun int sync (void)
A call to this function will not return as long as there is data which
that is not written to the device.  All dirty buffers in the kernel will
be written and so an overall consistent system can be achieved (if no
other process in parallel writes data).

A prototype for @code{sync} can be found in @file{unistd.h}.

The return value is zero to indicate no error.
@end deftypefun

More often it is wanted that not all data in the system is committed.
Programs want to ensure that data written to a given file are all
committed and in this situation @code{sync} is overkill.

@comment unistd.h
@comment POSIX
@deftypefun int fsync (int @var{fildes})
The @code{fsync} can be used to make sure all data associated with the
open file @var{fildes} is written to the device associated with the
descriptor.  The function call does not return unless all actions have
finished.

A prototype for @code{fsync} can be found in @file{unistd.h}.

This function is a cancelation point in multi-threaded programs.  This
is a problem if the thread allocates some resources (like memory, file
descriptors, semaphores or whatever) at the time @code{fsync} is
called.  If the thread gets canceled these resources stay allocated
until the program ends.  To avoid this calls to @code{fsync} should be
protected using cancelation handlers.
@c ref pthread_cleanup_push / pthread_cleanup_pop

The return value of the function is zero if no error occured.  Otherwise
it is @math{-1} and the global variable @var{errno} is set to the
following values:
@table @code
@item EBADF
The descriptor @var{fildes} is not valid.

@item EINVAL
No synchronization is possible since the system does not implement this.
@end table
@end deftypefun

Sometimes it is not even necessary to write all data associated with a
file descriptor.  E.g., in database files which do not change in size it
is enough to write all the file content data to the device.
Meta-information like the modification time etc. are not that important
and leaving such information uncommitted does not prevent a successful
recovering of the file in case of a problem.

@comment unistd.h
@comment POSIX
@deftypefun int fdatasync (int @var{fildes})
When a call to the @code{fdatasync} function returns it is made sure
that all of the file data is written to the device.  For all pending I/O
operations the parts guaranteeing data integrity finished.

Not all systems implement the @code{fdatasync} operation.  On systems
missing this functionality @code{fdatasync} is emulated by a call to
@code{fsync} since the performed actions are a superset of those
required by @code{fdatasyn}.

The prototype for @code{fdatasync} is in @file{unistd.h}.

The return value of the function is zero if no error occured.  Otherwise
it is @math{-1} and the global variable @var{errno} is set to the
following values:
@table @code
@item EBADF
The descriptor @var{fildes} is not valid.

@item EINVAL
No synchronization is possible since the system does not implement this.
@end table
@end deftypefun


@node Asynchronous I/O
@section Perform I/O Operations in Parallel

The POSIX.1b standard defines a new set of I/O operations which can
reduce the time an application spends waiting at I/O significantly.  The
new functions allow a program to initiate one or more I/O operations and
then immediately resume the normal work while the I/O operations are
executed in parallel.  The functionality is available if the
@file{unistd.h} file defines the symbol @code{_POSIX_ASYNCHRONOUS_IO}.

These functions are part of the library with realtime functions named
@file{librt}.  They are not actually part of the @file{libc} binary.
The implementation of these functions can be done using support in the
kernel (if available) or using an implementation based on threads at
userlevel.  In the latter case it might be necessary to link applications
with the thread library @file{libpthread} in addition to @file{librt}.

All AIO operations operate on files which were opened previously.  There
might be arbitrary many operations for one file running.  The
asynchronous I/O operations are controlled using a data structure named
@code{struct aiocb} (@dfn{AIO control block}).  It is defined in
@file{aio.h} as follows.

@comment aio.h
@comment POSIX.1b
@deftp {Data Type} {struct aiocb}
The POSIX.1b standard mandates that the @code{struct aiocb} structure
contains at least the members described in the following table.  There
might be more elements which are used by the implementation but
depending on these elements is not portable and is highly deprecated.

@table @code
@item int aio_fildes
This element specifies the file descriptor which is used for the
operation.  It must be a legal descriptor since otherwise the operation
fails for obvious reasons.

The device on which the file is opened must allow the seek operation.
I.e., it is not possible to use any of the AIO operations on devices
like terminals where an @code{lseek} call would lead to an error.

@item off_t aio_offset
This element specifies at which offset in the file the operation (input
or output) is performed.  Since the operations are carried out in arbitrary
order and more than one operation for one file descriptor can be
started, one cannot expect a current read/write position of the file
descriptor.

@item volatile void *aio_buf
This is a pointer to the buffer with the data to be written or the place
where the read data is stored.

@item size_t aio_nbytes
This element specifies the length of the buffer pointed to by @code{aio_buf}.

@item int aio_reqprio
If the platform has defined @code{_POSIX_PRIORITIZED_IO} and
@code{_POSIX_PRIORITY_SCHEDULING} the AIO requests are
processed based on the current scheduling priority.  The
@code{aio_reqprio} element can then be used to lower the priority of the
AIO operation.

@item struct sigevent aio_sigevent
This element specifies how the calling process is notified once the
operation terminates.  If the @code{sigev_notify} element is
@code{SIGEV_NONE} no notification is send.  If it is @code{SIGEV_SIGNAL}
the signal determined by @code{sigev_signo} is send.  Otherwise
@code{sigev_notify} must be @code{SIGEV_THREAD}.  In this case a thread
is created which starts executing the function pointed to by
@code{sigev_notify_function}.

@item int aio_lio_opcode
This element is only used by the @code{lio_listio} and
@code{lio_listio64} functions.  Since these functions allow to start an
arbitrary number of operations at once and since each operation can be
input or output (or nothing) the information must be stored in the
control block.  The possible values are:

@vtable @code
@item LIO_READ
Start a read operation.  Read from the file at position
@code{aio_offset} and store the next @code{aio_nbytes} bytes in the
buffer pointed to by @code{aio_buf}.

@item LIO_WRITE
Start a write operation.  Write @code{aio_nbytes} bytes starting at
@code{aio_buf} into the file starting at position @code{aio_offset}.

@item LIO_NOP
Do nothing for this control block.  This value is useful sometimes when
an array of @code{struct aiocb} values contains holes, i.e., some of the
values must not be handled although the whole array is presented to the
@code{lio_listio} function.
@end vtable
@end table

When the sources are compiled using @code{_FILE_OFFSET_BITS == 64} on a
32 bits machine this type is in fact @code{struct aiocb64} since the LFS
interface transparently replaces the @code{struct aiocb} definition.
@end deftp

For use with the AIO functions defined in the LFS there is a similar type
defined which replaces the types of the appropriate members with larger
types but otherwise is equivalent to @code{struct aiocb}.  Especially
all member names are the same.

@comment aio.h
@comment POSIX.1b
@deftp {Data Type} {struct aiocb64}
@table @code
@item int aio_fildes
This element specifies the file descriptor which is used for the
operation.  It must be a legal descriptor since otherwise the operation
fails for obvious reasons.

The device on which the file is opened must allow the seek operation.
I.e., it is not possible to use any of the AIO operations on devices
like terminals where an @code{lseek} call would lead to an error.

@item off64_t aio_offset
This element specified at which offset in the file the operation (input
or output) is performed.  Since the operation are carried in arbitrary
order and more than one operation for one file descriptor can be
started, one cannot expect a current read/write position of the file
descriptor.

@item volatile void *aio_buf
This is a pointer to the buffer with the data to be written or the place
where the ead data is stored.

@item size_t aio_nbytes
This element specifies the length of the buffer pointed to by @code{aio_buf}.

@item int aio_reqprio
If for the platform @code{_POSIX_PRIORITIZED_IO} and
@code{_POSIX_PRIORITY_SCHEDULING} is defined the AIO requests are
processed based on the current scheduling priority.  The
@code{aio_reqprio} element can then be used to lower the priority of the
AIO operation.

@item struct sigevent aio_sigevent
This element specifies how the calling process is notified once the
operation terminates.  If the @code{sigev_notify} element is
@code{SIGEV_NONE} no notification is send.  If it is @code{SIGEV_SIGNAL}
the signal determined by @code{sigev_signo} is send.  Otherwise
@code{sigev_notify} must be @code{SIGEV_THREAD} in which case a thread
which starts executing the function pointeed to by
@code{sigev_notify_function}.

@item int aio_lio_opcode
This element is only used by the @code{lio_listio} and
@code{[lio_listio64} functions.  Since these functions allow to start an
arbitrary number of operations at once and since each operation can be
input or output (or nothing) the information must be stored in the
control block.  See the description of @code{struct aiocb} for a description
of the possible values.
@end table

When the sources are compiled using @code{_FILE_OFFSET_BITS == 64} on a
32 bits machine this type is available under the name @code{struct
aiocb64} since the LFS replaces transparently the old interface.
@end deftp

@menu
* Asynchronous Reads/Writes::    Asynchronous Read and Write Operations.
* Status of AIO Operations::     Getting the Status of AIO Operations.
* Synchronizing AIO Operations:: Getting into a consistent state.
* Cancel AIO Operations::        Cancelation of AIO Operations.
* Configuration of AIO::         How to optimize the AIO implementation.
@end menu

@node Asynchronous Reads/Writes
@subsection Asynchronous Read and Write Operations

@comment aio.h
@comment POSIX.1b
@deftypefun int aio_read (struct aiocb *@var{aiocbp})
This function initiates an asynchronous read operation.  The function
call immediately returns after the operation was enqueued or when an
error was encountered.

The first @code{aiocbp->aio_nbytes} bytes of the file for which
@code{aiocbp->aio_fildes} is a descriptor are written to the buffer
starting at @code{aiocbp->aio_buf}.  Reading starts at the absolute
position @code{aiocbp->aio_offset} in the file.

If prioritized I/O is supported by the platform the
@code{aiocbp->aio_reqprio} value is used to adjust the priority before
the request is actually enqueued.

The calling process is notified about the termination of the read
request according to the @code{aiocbp->aio_sigevent} value.

When @code{aio_read} returns the return value is zero if no error
occurred that can be found before the process is enqueued.  If such an
early error is found the function returns @math{-1} and sets
@code{errno} to one of the following values.

@table @code
@item EAGAIN
The request was not enqueued due to (temporarily) exceeded resource
limitations.
@item ENOSYS
The @code{aio_read} function is not implemented.
@item EBADF
The @code{aiocbp->aio_fildes} descriptor is not valid.  This condition
needs not be recognized before enqueueing the request and so this error
might also be signaled asynchronously.
@item EINVAL
The @code{aiocbp->aio_offset} or @code{aiocbp->aio_reqpiro} value is
invalid.  This condition need not be recognized before enqueueing the
request and so this error might also be signaled asynchrously.
@end table

In the case @code{aio_read} returns zero the current status of the
request can be queried using @code{aio_error} and @code{aio_return}
functions.  As long as the value returned by @code{aio_error} is
@code{EINPROGRESS} the operation has not yet completed.  If
@code{aio_error} returns zero the operation successfully terminated,
otherwise the value is to be interpreted as an error code.  If the
function terminated the result of the operation can be get using a call
to @code{aio_return}.  The returned value is the same as an equivalent
call to @code{read} would have returned.  Possible error codes returned
by @code{aio_error} are:

@table @code
@item EBADF
The @code{aiocbp->aio_fildes} descriptor is not valid.
@item ECANCELED
The operation was canceled before the operation was finished
(@pxref{Cancel AIO Operations})
@item EINVAL
The @code{aiocbp->aio_offset} value is invalid.
@end table

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is in fact @code{aio_read64} since the LFS interface transparently
replaces the normal implementation.
@end deftypefun

@comment aio.h
@comment Unix98
@deftypefun int aio_read64 (struct aiocb *@var{aiocbp})
This function is similar to the @code{aio_read} function.  The only
difference is that on @w{32 bits} machines the file descriptor should
be opened in the large file mode.  Internally @code{aio_read64} uses
functionality equivalent to @code{lseek64} (@pxref{File Position
Primitive}) to position the file descriptor correctly for the reading,
as opposed to @code{lseek} functionality used in @code{aio_read}.

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is available under the name @code{aio_read} and so transparently
replaces the interface for small files on 32 bits machines.
@end deftypefun

To write data asynchronously to a file there exists an equivalent pair
of functions with a very similar interface.

@comment aio.h
@comment POSIX.1b
@deftypefun int aio_write (struct aiocb *@var{aiocbp})
This function initiates an asynchronous write operation.  The function
call immediately returns after the operation was enqueued or if before
this happens an error was encountered.

The first @code{aiocbp->aio_nbytes} bytes from the buffer starting at
@code{aiocbp->aio_buf} are written to the file for which
@code{aiocbp->aio_fildes} is an descriptor, starting at the absolute
position @code{aiocbp->aio_offset} in the file.

If prioritized I/O is supported by the platform the
@code{aiocbp->aio_reqprio} value is used to adjust the priority before
the request is actually enqueued.

The calling process is notified about the termination of the read
request according to the @code{aiocbp->aio_sigevent} value.

When @code{aio_write} returns the return value is zero if no error
occurred that can be found before the process is enqueued.  If such an
early error is found the function returns @math{-1} and sets
@code{errno} to one of the following values.

@table @code
@item EAGAIN
The request was not enqueued due to (temporarily) exceeded resource
limitations.
@item ENOSYS
The @code{aio_write} function is not implemented.
@item EBADF
The @code{aiocbp->aio_fildes} descriptor is not valid.  This condition
needs not be recognized before enqueueing the request and so this error
might also be signaled asynchronously.
@item EINVAL
The @code{aiocbp->aio_offset} or @code{aiocbp->aio_reqpiro} value is
invalid.  This condition needs not be recognized before enqueueing the
request and so this error might also be signaled asynchronously.
@end table

In the case @code{aio_write} returns zero the current status of the
request can be queried using @code{aio_error} and @code{aio_return}
functions.  As long as the value returned by @code{aio_error} is
@code{EINPROGRESS} the operation has not yet completed.  If
@code{aio_error} returns zero the operation successfully terminated,
otherwise the value is to be interpreted as an error code.  If the
function terminated the result of the operation can be get using a call
to @code{aio_return}.  The returned value is the same as an equivalent
call to @code{read} would have returned.  Possible error code returned
by @code{aio_error} are:

@table @code
@item EBADF
The @code{aiocbp->aio_fildes} descriptor is not valid.
@item ECANCELED
The operation was canceled before the operation was finished
(@pxref{Cancel AIO Operations})
@item EINVAL
The @code{aiocbp->aio_offset} value is invalid.
@end table

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is in fact @code{aio_write64} since the LFS interface transparently
replaces the normal implementation.
@end deftypefun

@comment aio.h
@comment Unix98
@deftypefun int aio_write64 (struct aiocb *@var{aiocbp})
This function is similar to the @code{aio_write} function.  The only
difference is that on @w{32 bits} machines the file descriptor should
be opened in the large file mode.  Internally @code{aio_write64} uses
functionality equivalent to @code{lseek64} (@pxref{File Position
Primitive}) to position the file descriptor correctly for the writing,
as opposed to @code{lseek} functionality used in @code{aio_write}.

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is available under the name @code{aio_write} and so transparently
replaces the interface for small files on 32 bits machines.
@end deftypefun

Beside these functions with the more or less traditional interface
POSIX.1b also defines a function with can initiate more than one
operation at once and which can handled freely mixed read and write
operation.  It is therefore similar to a combination of @code{readv} and
@code{writev}.

@comment aio.h
@comment POSIX.1b
@deftypefun int lio_listio (int @var{mode}, struct aiocb *const @var{list}[], int @var{nent}, struct sigevent *@var{sig})
The @code{lio_listio} function can be used to enqueue an arbitrary
number of read and write requests at one time.  The requests can all be
meant for the same file, all for different files or every solution in
between.

@code{lio_listio} gets the @var{nent} requests from the array pointed to
by @var{list}.  What operation has to be performed is determined by the
@code{aio_lio_opcode} member in each element of @var{list}.  If this
field is @code{LIO_READ} an read operation is queued, similar to a call
of @code{aio_read} for this element of the array (except that the way
the termination is signalled is different, as we will see below).  If
the @code{aio_lio_opcode} member is @code{LIO_WRITE} an write operation
is enqueued.  Otherwise the @code{aio_lio_opcode} must be @code{LIO_NOP}
in which case this element of @var{list} is simply ignored.  This
``operation'' is useful in situations where one has a fixed array of
@code{struct aiocb} elements from which only a few need to be handled at
a time.  Another situation is where the @code{lio_listio} call was
cancelled before all requests are processed (@pxref{Cancel AIO
Operations}) and the remaining requests have to be reissued.

The other members of each element of the array pointed to by
@code{list} must have values suitable for the operation as described in
the documentation for @code{aio_read} and @code{aio_write} above.

The @var{mode} argument determines how @code{lio_listio} behaves after
having enqueued all the requests.  If @var{mode} is @code{LIO_WAIT} it
waits until all requests terminated.  Otherwise @var{mode} must be
@code{LIO_NOWAIT} and in this case the function returns immediately after
having enqueued all the requests.  In this case the caller gets a
notification of the termination of all requests according to the
@var{sig} parameter.  If @var{sig} is @code{NULL} no notification is
send.  Otherwise a signal is sent or a thread is started, just as
described in the description for @code{aio_read} or @code{aio_write}.

If @var{mode} is @code{LIO_WAIT} the return value of @code{lio_listio}
is @math{0} when all requests completed successfully.  Otherwise the
function return @math{-1} and @code{errno} is set accordingly.  To find
out which request or requests failed one has to use the @code{aio_error}
function on all the elements of the array @var{list}.

In case @var{mode} is @code{LIO_NOWAIT} the function return @math{0} if
all requests were enqueued correctly.  The current state of the requests
can be found using @code{aio_error} and @code{aio_return} as described
above.  In case @code{lio_listio} returns @math{-1} in this mode the
global variable @code{errno} is set accordingly.  If a request did not
yet terminate a call to @code{aio_error} returns @code{EINPROGRESS}.  If
the value is different the request is finished and the error value (or
@math{0}) is returned and the result of the operation can be retrieved
using @code{aio_return}.

Possible values for @code{errno} are:

@table @code
@item EAGAIN
The resources necessary to queue all the requests are not available in
the moment.  The error status for each element of @var{list} must be
checked which request failed.

Another reason could be that the system wide limit of AIO requests is
exceeded.  This cannot be the case for the implementation on GNU systems
since no arbitrary limits exist.
@item EINVAL
The @var{mode} parameter is invalid or @var{nent} is larger than
@code{AIO_LISTIO_MAX}.
@item EIO
One or more of the request's I/O operations failed.  The error status of
each request should be checked for which one failed.
@item ENOSYS
The @code{lio_listio} function is not supported.
@end table

If the @var{mode} parameter is @code{LIO_NOWAIT} and the caller cancels
an request the error status for this request returned by
@code{aio_error} is @code{ECANCELED}.

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is in fact @code{lio_listio64} since the LFS interface
transparently replaces the normal implementation.
@end deftypefun

@comment aio.h
@comment Unix98
@deftypefun int lio_listio64 (int @var{mode}, struct aiocb *const @var{list}, int @var{nent}, struct sigevent *@var{sig})
This function is similar to the @code{aio_listio} function.  The only
difference is that only @w{32 bits} machines the file descriptor should
be opened in the large file mode.  Internally @code{lio_listio64} uses
functionality equivalent to @code{lseek64} (@pxref{File Position
Primitive}) to position the file descriptor correctly for the reading or
writing, as opposed to @code{lseek} functionality used in
@code{lio_listio}.

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is available under the name @code{lio_listio} and so
transparently replaces the interface for small files on 32 bits
machines.
@end deftypefun

@node Status of AIO Operations
@subsection Getting the Status of AIO Operations

As already described in the documentation of the functions in the last
section it must be possible to get information about the status of a I/O
request.  When the operation is performed really asynchronous (as with
@code{aio_read} and @code{aio_write} and with @code{aio_listio} when the
mode is @code{LIO_NOWAIT}) one sometimes needs to know whether a
specific request already terminated and if yes, what the result was..
The following two function allow to get this kind of information.

@comment aio.h
@comment POSIX.1b
@deftypefun int aio_error (const struct aiocb *@var{aiocbp})
This function determines the error state of the request described by the
@code{struct aiocb} variable pointed to by @var{aiocbp}.  If the
request has not yet terminated the value returned is always
@code{EINPROGRESS}.  Once the request has terminated the value
@code{aio_error} returns is either @math{0} if the request completed
successfully or it returns the value which would be stored in the
@code{errno} variable if the request would have been done using
@code{read}, @code{write}, or @code{fsync}.

The function can return @code{ENOSYS} if it is not implemented.  It
could also return @code{EINVAL} if the @var{aiocbp} parameter does not
refer to an asynchronous operation whose return status is not yet known.

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is in fact @code{aio_error64} since the LFS interface
transparently replaces the normal implementation.
@end deftypefun

@comment aio.h
@comment Unix98
@deftypefun int aio_error64 (const struct aiocb64 *@var{aiocbp})
This function is similar to @code{aio_error} with the only difference
that the argument is a reference to a variable of type @code{struct
aiocb64}.

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is available under the name @code{aio_error} and so
transparently replaces the interface for small files on 32 bits
machines.
@end deftypefun

@comment aio.h
@comment POSIX.1b
@deftypefun ssize_t aio_return (const struct aiocb *@var{aiocbp})
This function can be used to retrieve the return status of the operation
carried out by the request described in the variable pointed to by
@var{aiocbp}.  As long as the error status of this request as returned
by @code{aio_error} is @code{EINPROGRESS} the return of this function is
undefined.

Once the request is finished this function can be used exactly once to
retrieve the return value.  Following calls might lead to undefined
behaviour.  The return value itself is the value which would have been
returned by the @code{read}, @code{write}, or @code{fsync} call.

The function can return @code{ENOSYS} if it is not implemented.  It
could also return @code{EINVAL} if the @var{aiocbp} parameter does not
refer to an asynchronous operation whose return status is not yet known.

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is in fact @code{aio_return64} since the LFS interface
transparently replaces the normal implementation.
@end deftypefun

@comment aio.h
@comment Unix98
@deftypefun int aio_return64 (const struct aiocb64 *@var{aiocbp})
This function is similar to @code{aio_return} with the only difference
that the argument is a reference to a variable of type @code{struct
aiocb64}.

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is available under the name @code{aio_return} and so
transparently replaces the interface for small files on 32 bits
machines.
@end deftypefun

@node Synchronizing AIO Operations
@subsection Getting into a Consistent State

When dealing with asynchronous operations it is sometimes necessary to
get into a consistent state.  This would mean for AIO that one wants to
know whether a certain request or a group of request were processed.
This could be done by waiting for the notification sent by the system
after the operation terminated but this sometimes would mean wasting
resources (mainly computation time).  Instead POSIX.1b defines two
functions which will help with most kinds of consistency.

The @code{aio_fsync} and @code{aio_fsync64} functions are only available
if in @file{unistd.h} the symbol @code{_POSIX_SYNCHRONIZED_IO} is
defined.

@cindex synchronizing
@comment aio.h
@comment POSIX.1b
@deftypefun int aio_fsync (int @var{op}, struct aiocb *@var{aiocbp})
Calling this function forces all I/O operations operating queued at the
time of the function call operating on the file descriptor
@code{aiocbp->aio_fildes} into the synchronized I/O completion state
(@pxref{Synchronizing I/O}).  The @code{aio_fsync} function return
immediately but the notification through the method described in
@code{aiocbp->aio_sigevent} will happen only after all requests for this
file descriptor terminated and the file is synchronized.  This also
means that requests for this very same file descriptor which are queued
after the synchronization request are not effected.

If @var{op} is @code{O_DSYNC} the synchronization happens as with a call
to @code{fdatasync}.  Otherwise @var{op} should be @code{O_SYNC} and
the synchronization happens as with @code{fsync}.

As long as the synchronization has not happened a call to
@code{aio_error} with the reference to the object pointed to by
@var{aiocbp} returns @code{EINPROGRESS}.  Once the synchronization is
done @code{aio_error} return @math{0} if the synchronization was not
successful.  Otherwise the value returned is the value to which the
@code{fsync} or @code{fdatasync} function would have set the
@code{errno} variable.  In this case nothing can be assumed about the
consistency for the data written to this file descriptor.

The return value of this function is @math{0} if the request was
successfully filed.  Otherwise the return value is @math{-1} and
@code{errno} is set to one of the following values:

@table @code
@item EAGAIN
The request could not be enqueued due to temporary lack of resources.
@item EBADF
The file descriptor @code{aiocbp->aio_fildes} is not valid or not open
for writing.
@item EINVAL
The implementation does not support I/O synchronization or the @var{op}
parameter is other than @code{O_DSYNC} and @code{O_SYNC}.
@item ENOSYS
This function is not implemented.
@end table

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is in fact @code{aio_return64} since the LFS interface
transparently replaces the normal implementation.
@end deftypefun

@comment aio.h
@comment Unix98
@deftypefun int aio_fsync64 (int @var{op}, struct aiocb64 *@var{aiocbp})
This function is similar to @code{aio_fsync} with the only difference
that the argument is a reference to a variable of type @code{struct
aiocb64}.

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is available under the name @code{aio_fsync} and so
transparently replaces the interface for small files on 32 bits
machines.
@end deftypefun

Another method of synchronization is to wait until one or more requests of a
specific set terminated.  This could be achieved by the @code{aio_*}
functions to notify the initiating process about the termination but in
some situations this is not the ideal solution.  In a program which
constantly updates clients somehow connected to the server it is not
always the best solution to go round robin since some connections might
be slow.  On the other hand letting the @code{aio_*} function notify the
caller might also be not the best solution since whenever the process
works on preparing data for on client it makes no sense to be
interrupted by a notification since the new client will not be handled
before the current client is served.  For situations like this
@code{aio_suspend} should be used.

@comment aio.h
@comment POSIX.1b
@deftypefun int aio_suspend (const struct aiocb *const @var{list}[], int @var{nent}, const struct timespec *@var{timeout})
When calling this function the calling thread is suspended until at
least one of the requests pointed to by the @var{nent} elements of the
array @var{list} has completed.  If any of the requests already has
completed at the time @code{aio_suspend} is called the function returns
immediately.  Whether a request has terminated or not is done by
comparing the error status of the request with @code{EINPROGRESS}.  If
an element of @var{list} is @code{NULL} the entry is simply ignored.

If no request has finished the calling process is suspended.  If
@var{timeout} is @code{NULL} the process is not waked until a request
finished.  If @var{timeout} is not @code{NULL} the process remains
suspended at as long as specified in @var{timeout}.  In this case
@code{aio_suspend} returns with an error.

The return value of the function is @math{0} if one or more requests
from the @var{list} have terminated.  Otherwise the function returns
@math{-1} and @code{errno} is set to one of the following values:

@table @code
@item EAGAIN
None of the requests from the @var{list} completed in the time specified
by @var{timeout}.
@item EINTR
A signal interrupted the @code{aio_suspend} function.  This signal might
also be sent by the AIO implementation while signalling the termination
of one of the requests.
@item ENOSYS
The @code{aio_suspend} function is not implemented.
@end table

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is in fact @code{aio_suspend64} since the LFS interface
transparently replaces the normal implementation.
@end deftypefun

@comment aio.h
@comment Unix98
@deftypefun int aio_suspend64 (const struct aiocb64 *const @var{list}[], int @var{nent}, const struct timespec *@var{timeout})
This function is similar to @code{aio_suspend} with the only difference
that the argument is a reference to a variable of type @code{struct
aiocb64}.

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is available under the name @code{aio_suspend} and so
transparently replaces the interface for small files on 32 bits
machines.
@end deftypefun

@node Cancel AIO Operations
@subsection Cancelation of AIO Operations

When one or more requests are asynchronously processed it might be
useful in some situations to cancel a selected operation, e.g., if it
becomes obvious that the written data is not anymore accurate and would
have to be overwritten soon.  As an example assume an application, which
writes data in files in a situation where new incoming data would have
to be written in a file which will be updated by an enqueued request.
The POSIX AIO implementation provides such a function but this function
is not capable to force the cancelation of the request.  It is up to the
implementation to decide whether it is possible to cancel the operation
or not.  Therefore using this function is merely a hint.

@comment aio.h
@comment POSIX.1b
@deftypefun int aio_cancel (int @var{fildes}, struct aiocb *@var{aiocbp})
The @code{aio_cancel} function can be used to cancel one or more
outstanding requests.  If the @var{aiocbp} parameter is @code{NULL} the
function tries to cancel all outstanding requests which would process
the file descriptor @var{fildes} (i.e.,, whose @code{aio_fildes} member
is @var{fildes}).  If @var{aiocbp} is not @code{NULL} the very specific
request pointed to by @var{aiocbp} is tried to be canceled.

For requests which were successfully canceled the normal notification
about the termination of the request should take place.  I.e., depending
on the @code{struct sigevent} object which controls this, nothing
happens, a signal is sent or a thread is started.  If the request cannot
be canceled it terminates the usual way after performing te operation.

After a request is successfully canceled a call to @code{aio_error} with
a reference to this request as the parameter will return
@code{ECANCELED} and a call to @code{aio_return} will return @math{-1}.
If the request wasn't canceled and is still running the error status is
still @code{EINPROGRESS}.

The return value of the function is @code{AIO_CANCELED} if there were
requests which haven't terminated and which successfully were canceled.
If there is one or more request left which couldn't be canceled the
return value is @code{AIO_NOTCANCELED}.  In this case @code{aio_error}
must be used to find out which of the perhaps multiple requests (in
@var{aiocbp} is @code{NULL}) wasn't successfully canceled.  If all
requests already terminated at the time @code{aio_cancel} is called the
return value is @code{AIO_ALLDONE}.

If an error occurred during the execution of @code{aio_cancel} the
function returns @math{-1} and sets @code{errno} to one of the following
values.

@table @code
@item EBADF
The file descriptor @var{fildes} is not valid.
@item ENOSYS
@code{aio_cancel} is not implemented.
@end table

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is in fact @code{aio_cancel64} since the LFS interface
transparently replaces the normal implementation.
@end deftypefun

@comment aio.h
@comment Unix98
@deftypefun int aio_cancel64 (int @var{fildes}, struct aiocb *@var{aiocbp})
This function is similar to @code{aio_cancel} with the only difference
that the argument is a reference to a variable of type @code{struct
aiocb64}.

When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} this
function is available under the name @code{aio_cancel} and so
transparently replaces the interface for small files on 32 bits
machines.
@end deftypefun

@node Configuration of AIO
@subsection How to optimize the AIO implementation

The POSIX standard does not specify how the AIO functions are
implemented.  They could be system calls but it is also possible to
emulate them at userlevel.

At least the available implementation at the point of this writing is a
userlevel implementation which uses threads for handling the enqueued
requests.  This implementation requires to make some decisions about
limitations but hard limitations are something which better should be
avoided the GNU C library implementation provides a mean to tune the AIO
implementation individually for each use.

@comment aio.h
@comment GNU
@deftp {Data Type} {struct aioinit}
This data type is used to pass the configuration or tunable parameters
to the implementation.  The program has to initialize the members of
this struct and pass it to the implementation using the @code{aio_init}
function.

@table @code
@item int aio_threads
This member specifies the maximal number of threads which must be used
at any one time.
@item int aio_num
This number provides an estimate on the maximal number of simultaneously
enqueued requests.
@item int aio_locks
@c What?
@item int aio_usedba
@c What?
@item int aio_debug
@c What?
@item int aio_numusers
@c What?
@item int aio_reserved[2]
@c What?
@end table
@end deftp

@comment aio.h
@comment GNU
@deftypefun void aio_init (const struct aioinit *@var{init})
This function must be called before any other AIO function.  Calling it
is completely voluntarily since it only is meant to help the AIO
implementation to perform better.

Before calling the @code{aio_init} function the members of a variable of
type @code{struct aioinit} must be initialized.  Then a reference to
this variable is passed as the parameter to @code{aio_init} which itself
may or may not pay attention to the hints.

The function has no return value and no error cases are defined.  It is
a extension which follows a proposal from the SGI implementation in
@w{Irix 6}.  It is not covered by POSIX.1b or Unix98.
@end deftypefun

@node Control Operations
@section Control Operations on Files

@cindex control operations on files
@cindex @code{fcntl} function
This section describes how you can perform various other operations on
file descriptors, such as inquiring about or setting flags describing
the status of the file descriptor, manipulating record locks, and the
like.  All of these operations are performed by the function @code{fcntl}.

The second argument to the @code{fcntl} function is a command that
specifies which operation to perform.  The function and macros that name
various flags that are used with it are declared in the header file
@file{fcntl.h}.  Many of these flags are also used by the @code{open}
function; see @ref{Opening and Closing Files}.
@pindex fcntl.h

@comment fcntl.h
@comment POSIX.1
@deftypefun int fcntl (int @var{filedes}, int @var{command}, @dots{})
The @code{fcntl} function performs the operation specified by
@var{command} on the file descriptor @var{filedes}.  Some commands
require additional arguments to be supplied.  These additional arguments
and the return value and error conditions are given in the detailed
descriptions of the individual commands.

Briefly, here is a list of what the various commands are.

@table @code
@item F_DUPFD
Duplicate the file descriptor (return another file descriptor pointing
to the same open file).  @xref{Duplicating Descriptors}.

@item F_GETFD
Get flags associated with the file descriptor.  @xref{Descriptor Flags}.

@item F_SETFD
Set flags associated with the file descriptor.  @xref{Descriptor Flags}.

@item F_GETFL
Get flags associated with the open file.  @xref{File Status Flags}.

@item F_SETFL
Set flags associated with the open file.  @xref{File Status Flags}.

@item F_GETLK
Get a file lock.  @xref{File Locks}.

@item F_SETLK
Set or clear a file lock.  @xref{File Locks}.

@item F_SETLKW
Like @code{F_SETLK}, but wait for completion.  @xref{File Locks}.

@item F_GETOWN
Get process or process group ID to receive @code{SIGIO} signals.
@xref{Interrupt Input}.

@item F_SETOWN
Set process or process group ID to receive @code{SIGIO} signals.
@xref{Interrupt Input}.
@end table

This function is a cancelation point in multi-threaded programs.  This
is a problem if the thread allocates some resources (like memory, file
descriptors, semaphores or whatever) at the time @code{fcntl} is
called.  If the thread gets canceled these resources stay allocated
until the program ends.  To avoid this calls to @code{fcntl} should be
protected using cancelation handlers.
@c ref pthread_cleanup_push / pthread_cleanup_pop
@end deftypefun


@node Duplicating Descriptors
@section Duplicating Descriptors

@cindex duplicating file descriptors
@cindex redirecting input and output

You can @dfn{duplicate} a file descriptor, or allocate another file
descriptor that refers to the same open file as the original.  Duplicate
descriptors share one file position and one set of file status flags
(@pxref{File Status Flags}), but each has its own set of file descriptor
flags (@pxref{Descriptor Flags}).

The major use of duplicating a file descriptor is to implement
@dfn{redirection} of input or output:  that is, to change the
file or pipe that a particular file descriptor corresponds to.

You can perform this operation using the @code{fcntl} function with the
@code{F_DUPFD} command, but there are also convenient functions
@code{dup} and @code{dup2} for duplicating descriptors.

@pindex unistd.h
@pindex fcntl.h
The @code{fcntl} function and flags are declared in @file{fcntl.h},
while prototypes for @code{dup} and @code{dup2} are in the header file
@file{unistd.h}.

@comment unistd.h
@comment POSIX.1
@deftypefun int dup (int @var{old})
This function copies descriptor @var{old} to the first available
descriptor number (the first number not currently open).  It is
equivalent to @code{fcntl (@var{old}, F_DUPFD, 0)}.
@end deftypefun

@comment unistd.h
@comment POSIX.1
@deftypefun int dup2 (int @var{old}, int @var{new})
This function copies the descriptor @var{old} to descriptor number
@var{new}.

If @var{old} is an invalid descriptor, then @code{dup2} does nothing; it
does not close @var{new}.  Otherwise, the new duplicate of @var{old}
replaces any previous meaning of descriptor @var{new}, as if @var{new}
were closed first.

If @var{old} and @var{new} are different numbers, and @var{old} is a
valid descriptor number, then @code{dup2} is equivalent to:

@smallexample
close (@var{new});
fcntl (@var{old}, F_DUPFD, @var{new})
@end smallexample

However, @code{dup2} does this atomically; there is no instant in the
middle of calling @code{dup2} at which @var{new} is closed and not yet a
duplicate of @var{old}.
@end deftypefun

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int F_DUPFD
This macro is used as the @var{command} argument to @code{fcntl}, to
copy the file descriptor given as the first argument.

The form of the call in this case is:

@smallexample
fcntl (@var{old}, F_DUPFD, @var{next-filedes})
@end smallexample

The @var{next-filedes} argument is of type @code{int} and specifies that
the file descriptor returned should be the next available one greater
than or equal to this value.

The return value from @code{fcntl} with this command is normally the value
of the new file descriptor.  A return value of @math{-1} indicates an
error.  The following @code{errno} error conditions are defined for
this command:

@table @code
@item EBADF
The @var{old} argument is invalid.

@item EINVAL
The @var{next-filedes} argument is invalid.

@item EMFILE
There are no more file descriptors available---your program is already
using the maximum.  In BSD and GNU, the maximum is controlled by a
resource limit that can be changed; @pxref{Limits on Resources}, for
more information about the @code{RLIMIT_NOFILE} limit.
@end table

@code{ENFILE} is not a possible error code for @code{dup2} because
@code{dup2} does not create a new opening of a file; duplicate
descriptors do not count toward the limit which @code{ENFILE}
indicates.  @code{EMFILE} is possible because it refers to the limit on
distinct descriptor numbers in use in one process.
@end deftypevr

Here is an example showing how to use @code{dup2} to do redirection.
Typically, redirection of the standard streams (like @code{stdin}) is
done by a shell or shell-like program before calling one of the
@code{exec} functions (@pxref{Executing a File}) to execute a new
program in a child process.  When the new program is executed, it
creates and initializes the standard streams to point to the
corresponding file descriptors, before its @code{main} function is
invoked.

So, to redirect standard input to a file, the shell could do something
like:

@smallexample
pid = fork ();
if (pid == 0)
  @{
    char *filename;
    char *program;
    int file;
    @dots{}
    file = TEMP_FAILURE_RETRY (open (filename, O_RDONLY));
    dup2 (file, STDIN_FILENO);
    TEMP_FAILURE_RETRY (close (file));
    execv (program, NULL);
  @}
@end smallexample

There is also a more detailed example showing how to implement redirection
in the context of a pipeline of processes in @ref{Launching Jobs}.


@node Descriptor Flags
@section File Descriptor Flags
@cindex file descriptor flags

@dfn{File descriptor flags} are miscellaneous attributes of a file
descriptor.  These flags are associated with particular file
descriptors, so that if you have created duplicate file descriptors
from a single opening of a file, each descriptor has its own set of flags.

Currently there is just one file descriptor flag: @code{FD_CLOEXEC},
which causes the descriptor to be closed if you use any of the
@code{exec@dots{}} functions (@pxref{Executing a File}).

The symbols in this section are defined in the header file
@file{fcntl.h}.
@pindex fcntl.h

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int F_GETFD
This macro is used as the @var{command} argument to @code{fcntl}, to
specify that it should return the file descriptor flags associated
with the @var{filedes} argument.

The normal return value from @code{fcntl} with this command is a
nonnegative number which can be interpreted as the bitwise OR of the
individual flags (except that currently there is only one flag to use).

In case of an error, @code{fcntl} returns @math{-1}.  The following
@code{errno} error conditions are defined for this command:

@table @code
@item EBADF
The @var{filedes} argument is invalid.
@end table
@end deftypevr


@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int F_SETFD
This macro is used as the @var{command} argument to @code{fcntl}, to
specify that it should set the file descriptor flags associated with the
@var{filedes} argument.  This requires a third @code{int} argument to
specify the new flags, so the form of the call is:

@smallexample
fcntl (@var{filedes}, F_SETFD, @var{new-flags})
@end smallexample

The normal return value from @code{fcntl} with this command is an
unspecified value other than @math{-1}, which indicates an error.
The flags and error conditions are the same as for the @code{F_GETFD}
command.
@end deftypevr

The following macro is defined for use as a file descriptor flag with
the @code{fcntl} function.  The value is an integer constant usable
as a bit mask value.

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int FD_CLOEXEC
@cindex close-on-exec (file descriptor flag)
This flag specifies that the file descriptor should be closed when
an @code{exec} function is invoked; see @ref{Executing a File}.  When
a file descriptor is allocated (as with @code{open} or @code{dup}),
this bit is initially cleared on the new file descriptor, meaning that
descriptor will survive into the new program after @code{exec}.
@end deftypevr

If you want to modify the file descriptor flags, you should get the
current flags with @code{F_GETFD} and modify the value.  Don't assume
that the flags listed here are the only ones that are implemented; your
program may be run years from now and more flags may exist then.  For
example, here is a function to set or clear the flag @code{FD_CLOEXEC}
without altering any other flags:

@smallexample
/* @r{Set the @code{FD_CLOEXEC} flag of @var{desc} if @var{value} is nonzero,}
   @r{or clear the flag if @var{value} is 0.}
   @r{Return 0 on success, or -1 on error with @code{errno} set.} */

int
set_cloexec_flag (int desc, int value)
@{
  int oldflags = fcntl (desc, F_GETFD, 0);
  /* @r{If reading the flags failed, return error indication now.}
  if (oldflags < 0)
    return oldflags;
  /* @r{Set just the flag we want to set.} */
  if (value != 0)
    oldflags |= FD_CLOEXEC;
  else
    oldflags &= ~FD_CLOEXEC;
  /* @r{Store modified flag word in the descriptor.} */
  return fcntl (desc, F_SETFD, oldflags);
@}
@end smallexample

@node File Status Flags
@section File Status Flags
@cindex file status flags

@dfn{File status flags} are used to specify attributes of the opening of a
file.  Unlike the file descriptor flags discussed in @ref{Descriptor
Flags}, the file status flags are shared by duplicated file descriptors
resulting from a single opening of the file.  The file status flags are
specified with the @var{flags} argument to @code{open};
@pxref{Opening and Closing Files}.

File status flags fall into three categories, which are described in the
following sections.

@itemize @bullet
@item
@ref{Access Modes}, specify what type of access is allowed to the
file: reading, writing, or both.  They are set by @code{open} and are
returned by @code{fcntl}, but cannot be changed.

@item
@ref{Open-time Flags}, control details of what @code{open} will do.
These flags are not preserved after the @code{open} call.

@item
@ref{Operating Modes}, affect how operations such as @code{read} and
@code{write} are done.  They are set by @code{open}, and can be fetched or
changed with @code{fcntl}.
@end itemize

The symbols in this section are defined in the header file
@file{fcntl.h}.
@pindex fcntl.h

@menu
* Access Modes::                Whether the descriptor can read or write.
* Open-time Flags::             Details of @code{open}.
* Operating Modes::             Special modes to control I/O operations.
* Getting File Status Flags::   Fetching and changing these flags.
@end menu

@node Access Modes
@subsection File Access Modes

The file access modes allow a file descriptor to be used for reading,
writing, or both.  (In the GNU system, they can also allow none of these,
and allow execution of the file as a program.)  The access modes are chosen
when the file is opened, and never change.

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int O_RDONLY
Open the file for read access.
@end deftypevr

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int O_WRONLY
Open the file for write access.
@end deftypevr

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int O_RDWR
Open the file for both reading and writing.
@end deftypevr

In the GNU system (and not in other systems), @code{O_RDONLY} and
@code{O_WRONLY} are independent bits that can be bitwise-ORed together,
and it is valid for either bit to be set or clear.  This means that
@code{O_RDWR} is the same as @code{O_RDONLY|O_WRONLY}.  A file access
mode of zero is permissible; it allows no operations that do input or
output to the file, but does allow other operations such as
@code{fchmod}.  On the GNU system, since ``read-only'' or ``write-only''
is a misnomer, @file{fcntl.h} defines additional names for the file
access modes.  These names are preferred when writing GNU-specific code.
But most programs will want to be portable to other POSIX.1 systems and
should use the POSIX.1 names above instead.

@comment fcntl.h
@comment GNU
@deftypevr Macro int O_READ
Open the file for reading.  Same as @code{O_RDWR}; only defined on GNU.
@end deftypevr

@comment fcntl.h
@comment GNU
@deftypevr Macro int O_WRITE
Open the file for reading.  Same as @code{O_WRONLY}; only defined on GNU.
@end deftypevr

@comment fcntl.h
@comment GNU
@deftypevr Macro int O_EXEC
Open the file for executing.  Only defined on GNU.
@end deftypevr

To determine the file access mode with @code{fcntl}, you must extract
the access mode bits from the retrieved file status flags.  In the GNU
system, you can just test the @code{O_READ} and @code{O_WRITE} bits in
the flags word.  But in other POSIX.1 systems, reading and writing
access modes are not stored as distinct bit flags.  The portable way to
extract the file access mode bits is with @code{O_ACCMODE}.

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int O_ACCMODE
This macro stands for a mask that can be bitwise-ANDed with the file
status flag value to produce a value representing the file access mode.
The mode will be @code{O_RDONLY}, @code{O_WRONLY}, or @code{O_RDWR}.
(In the GNU system it could also be zero, and it never includes the
@code{O_EXEC} bit.)
@end deftypevr

@node Open-time Flags
@subsection Open-time Flags

The open-time flags specify options affecting how @code{open} will behave.
These options are not preserved once the file is open.  The exception to
this is @code{O_NONBLOCK}, which is also an I/O operating mode and so it
@emph{is} saved.  @xref{Opening and Closing Files}, for how to call
@code{open}.

There are two sorts of options specified by open-time flags.

@itemize @bullet
@item
@dfn{File name translation flags} affect how @code{open} looks up the
file name to locate the file, and whether the file can be created.
@cindex file name translation flags
@cindex flags, file name translation

@item
@dfn{Open-time action flags} specify extra operations that @code{open} will
perform on the file once it is open.
@cindex open-time action flags
@cindex flags, open-time action
@end itemize

Here are the file name translation flags.

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int O_CREAT
If set, the file will be created if it doesn't already exist.
@c !!! mode arg, umask
@cindex create on open (file status flag)
@end deftypevr

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int O_EXCL
If both @code{O_CREAT} and @code{O_EXCL} are set, then @code{open} fails
if the specified file already exists.  This is guaranteed to never
clobber an existing file.
@end deftypevr

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int O_NONBLOCK
@cindex non-blocking open
This prevents @code{open} from blocking for a ``long time'' to open the
file.  This is only meaningful for some kinds of files, usually devices
such as serial ports; when it is not meaningful, it is harmless and
ignored.  Often opening a port to a modem blocks until the modem reports
carrier detection; if @code{O_NONBLOCK} is specified, @code{open} will
return immediately without a carrier.

Note that the @code{O_NONBLOCK} flag is overloaded as both an I/O operating
mode and a file name translation flag.  This means that specifying
@code{O_NONBLOCK} in @code{open} also sets nonblocking I/O mode;
@pxref{Operating Modes}.  To open the file without blocking but do normal
I/O that blocks, you must call @code{open} with @code{O_NONBLOCK} set and
then call @code{fcntl} to turn the bit off.
@end deftypevr

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int O_NOCTTY
If the named file is a terminal device, don't make it the controlling
terminal for the process.  @xref{Job Control}, for information about
what it means to be the controlling terminal.

In the GNU system and 4.4 BSD, opening a file never makes it the
controlling terminal and @code{O_NOCTTY} is zero.  However, other
systems may use a nonzero value for @code{O_NOCTTY} and set the
controlling terminal when you open a file that is a terminal device; so
to be portable, use @code{O_NOCTTY} when it is important to avoid this.
@cindex controlling terminal, setting
@end deftypevr

The following three file name translation flags exist only in the GNU system.

@comment fcntl.h
@comment GNU
@deftypevr Macro int O_IGNORE_CTTY
Do not recognize the named file as the controlling terminal, even if it
refers to the process's existing controlling terminal device.  Operations
on the new file descriptor will never induce job control signals.
@xref{Job Control}.
@end deftypevr

@comment fcntl.h
@comment GNU
@deftypevr Macro int O_NOLINK
If the named file is a symbolic link, open the link itself instead of
the file it refers to.  (@code{fstat} on the new file descriptor will
return the information returned by @code{lstat} on the link's name.)
@cindex symbolic link, opening
@end deftypevr

@comment fcntl.h
@comment GNU
@deftypevr Macro int O_NOTRANS
If the named file is specially translated, do not invoke the translator.
Open the bare file the translator itself sees.
@end deftypevr


The open-time action flags tell @code{open} to do additional operations
which are not really related to opening the file.  The reason to do them
as part of @code{open} instead of in separate calls is that @code{open}
can do them @i{atomically}.

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int O_TRUNC
Truncate the file to zero length.  This option is only useful for
regular files, not special files such as directories or FIFOs.  POSIX.1
requires that you open the file for writing to use @code{O_TRUNC}.  In
BSD and GNU you must have permission to write the file to truncate it,
but you need not open for write access.

This is the only open-time action flag specified by POSIX.1.  There is
no good reason for truncation to be done by @code{open}, instead of by
calling @code{ftruncate} afterwards.  The @code{O_TRUNC} flag existed in
Unix before @code{ftruncate} was invented, and is retained for backward
compatibility.
@end deftypevr

The remaining operating modes are BSD extensions.  They exist only
on some systems.  On other systems, these macros are not defined.

@comment fcntl.h
@comment BSD
@deftypevr Macro int O_SHLOCK
Acquire a shared lock on the file, as with @code{flock}.
@xref{File Locks}.

If @code{O_CREAT} is specified, the locking is done atomically when
creating the file.  You are guaranteed that no other process will get
the lock on the new file first.
@end deftypevr

@comment fcntl.h
@comment BSD
@deftypevr Macro int O_EXLOCK
Acquire an exclusive lock on the file, as with @code{flock}.
@xref{File Locks}.  This is atomic like @code{O_SHLOCK}.
@end deftypevr

@node Operating Modes
@subsection I/O Operating Modes

The operating modes affect how input and output operations using a file
descriptor work.  These flags are set by @code{open} and can be fetched
and changed with @code{fcntl}.

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int O_APPEND
The bit that enables append mode for the file.  If set, then all
@code{write} operations write the data at the end of the file, extending
it, regardless of the current file position.  This is the only reliable
way to append to a file.  In append mode, you are guaranteed that the
data you write will always go to the current end of the file, regardless
of other processes writing to the file.  Conversely, if you simply set
the file position to the end of file and write, then another process can
extend the file after you set the file position but before you write,
resulting in your data appearing someplace before the real end of file.
@end deftypevr

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int O_NONBLOCK
The bit that enables nonblocking mode for the file.  If this bit is set,
@code{read} requests on the file can return immediately with a failure
status if there is no input immediately available, instead of blocking.
Likewise, @code{write} requests can also return immediately with a
failure status if the output can't be written immediately.

Note that the @code{O_NONBLOCK} flag is overloaded as both an I/O
operating mode and a file name translation flag; @pxref{Open-time Flags}.
@end deftypevr

@comment fcntl.h
@comment BSD
@deftypevr Macro int O_NDELAY
This is an obsolete name for @code{O_NONBLOCK}, provided for
compatibility with BSD.  It is not defined by the POSIX.1 standard.
@end deftypevr

The remaining operating modes are BSD and GNU extensions.  They exist only
on some systems.  On other systems, these macros are not defined.

@comment fcntl.h
@comment BSD
@deftypevr Macro int O_ASYNC
The bit that enables asynchronous input mode.  If set, then @code{SIGIO}
signals will be generated when input is available.  @xref{Interrupt Input}.

Asynchronous input mode is a BSD feature.
@end deftypevr

@comment fcntl.h
@comment BSD
@deftypevr Macro int O_FSYNC
The bit that enables synchronous writing for the file.  If set, each
@code{write} call will make sure the data is reliably stored on disk before
returning. @c !!! xref fsync

Synchronous writing is a BSD feature.
@end deftypevr

@comment fcntl.h
@comment BSD
@deftypevr Macro int O_SYNC
This is another name for @code{O_FSYNC}.  They have the same value.
@end deftypevr

@comment fcntl.h
@comment GNU
@deftypevr Macro int O_NOATIME
If this bit is set, @code{read} will not update the access time of the
file.  @xref{File Times}.  This is used by programs that do backups, so
that backing a file up does not count as reading it.
Only the owner of the file or the superuser may use this bit.

This is a GNU extension.
@end deftypevr

@node Getting File Status Flags
@subsection Getting and Setting File Status Flags

The @code{fcntl} function can fetch or change file status flags.

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int F_GETFL
This macro is used as the @var{command} argument to @code{fcntl}, to
read the file status flags for the open file with descriptor
@var{filedes}.

The normal return value from @code{fcntl} with this command is a
nonnegative number which can be interpreted as the bitwise OR of the
individual flags.  Since the file access modes are not single-bit values,
you can mask off other bits in the returned flags with @code{O_ACCMODE}
to compare them.

In case of an error, @code{fcntl} returns @math{-1}.  The following
@code{errno} error conditions are defined for this command:

@table @code
@item EBADF
The @var{filedes} argument is invalid.
@end table
@end deftypevr

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int F_SETFL
This macro is used as the @var{command} argument to @code{fcntl}, to set
the file status flags for the open file corresponding to the
@var{filedes} argument.  This command requires a third @code{int}
argument to specify the new flags, so the call looks like this:

@smallexample
fcntl (@var{filedes}, F_SETFL, @var{new-flags})
@end smallexample

You can't change the access mode for the file in this way; that is,
whether the file descriptor was opened for reading or writing.

The normal return value from @code{fcntl} with this command is an
unspecified value other than @math{-1}, which indicates an error.  The
error conditions are the same as for the @code{F_GETFL} command.
@end deftypevr

If you want to modify the file status flags, you should get the current
flags with @code{F_GETFL} and modify the value.  Don't assume that the
flags listed here are the only ones that are implemented; your program
may be run years from now and more flags may exist then.  For example,
here is a function to set or clear the flag @code{O_NONBLOCK} without
altering any other flags:

@smallexample
@group
/* @r{Set the @code{O_NONBLOCK} flag of @var{desc} if @var{value} is nonzero,}
   @r{or clear the flag if @var{value} is 0.}
   @r{Return 0 on success, or -1 on error with @code{errno} set.} */

int
set_nonblock_flag (int desc, int value)
@{
  int oldflags = fcntl (desc, F_GETFL, 0);
  /* @r{If reading the flags failed, return error indication now.} */
  if (oldflags == -1)
    return -1;
  /* @r{Set just the flag we want to set.} */
  if (value != 0)
    oldflags |= O_NONBLOCK;
  else
    oldflags &= ~O_NONBLOCK;
  /* @r{Store modified flag word in the descriptor.} */
  return fcntl (desc, F_SETFL, oldflags);
@}
@end group
@end smallexample

@node File Locks
@section File Locks

@cindex file locks
@cindex record locking
The remaining @code{fcntl} commands are used to support @dfn{record
locking}, which permits multiple cooperating programs to prevent each
other from simultaneously accessing parts of a file in error-prone
ways.

@cindex exclusive lock
@cindex write lock
An @dfn{exclusive} or @dfn{write} lock gives a process exclusive access
for writing to the specified part of the file.  While a write lock is in
place, no other process can lock that part of the file.

@cindex shared lock
@cindex read lock
A @dfn{shared} or @dfn{read} lock prohibits any other process from
requesting a write lock on the specified part of the file.  However,
other processes can request read locks.

The @code{read} and @code{write} functions do not actually check to see
whether there are any locks in place.  If you want to implement a
locking protocol for a file shared by multiple processes, your application
must do explicit @code{fcntl} calls to request and clear locks at the
appropriate points.

Locks are associated with processes.  A process can only have one kind
of lock set for each byte of a given file.  When any file descriptor for
that file is closed by the process, all of the locks that process holds
on that file are released, even if the locks were made using other
descriptors that remain open.  Likewise, locks are released when a
process exits, and are not inherited by child processes created using
@code{fork} (@pxref{Creating a Process}).

When making a lock, use a @code{struct flock} to specify what kind of
lock and where.  This data type and the associated macros for the
@code{fcntl} function are declared in the header file @file{fcntl.h}.
@pindex fcntl.h

@comment fcntl.h
@comment POSIX.1
@deftp {Data Type} {struct flock}
This structure is used with the @code{fcntl} function to describe a file
lock.  It has these members:

@table @code
@item short int l_type
Specifies the type of the lock; one of @code{F_RDLCK}, @code{F_WRLCK}, or
@code{F_UNLCK}.

@item short int l_whence
This corresponds to the @var{whence} argument to @code{fseek} or
@code{lseek}, and specifies what the offset is relative to.  Its value
can be one of @code{SEEK_SET}, @code{SEEK_CUR}, or @code{SEEK_END}.

@item off_t l_start
This specifies the offset of the start of the region to which the lock
applies, and is given in bytes relative to the point specified by
@code{l_whence} member.

@item off_t l_len
This specifies the length of the region to be locked.  A value of
@code{0} is treated specially; it means the region extends to the end of
the file.

@item pid_t l_pid
This field is the process ID (@pxref{Process Creation Concepts}) of the
process holding the lock.  It is filled in by calling @code{fcntl} with
the @code{F_GETLK} command, but is ignored when making a lock.
@end table
@end deftp

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int F_GETLK
This macro is used as the @var{command} argument to @code{fcntl}, to
specify that it should get information about a lock.  This command
requires a third argument of type @w{@code{struct flock *}} to be passed
to @code{fcntl}, so that the form of the call is:

@smallexample
fcntl (@var{filedes}, F_GETLK, @var{lockp})
@end smallexample

If there is a lock already in place that would block the lock described
by the @var{lockp} argument, information about that lock overwrites
@code{*@var{lockp}}.  Existing locks are not reported if they are
compatible with making a new lock as specified.  Thus, you should
specify a lock type of @code{F_WRLCK} if you want to find out about both
read and write locks, or @code{F_RDLCK} if you want to find out about
write locks only.

There might be more than one lock affecting the region specified by the
@var{lockp} argument, but @code{fcntl} only returns information about
one of them.  The @code{l_whence} member of the @var{lockp} structure is
set to @code{SEEK_SET} and the @code{l_start} and @code{l_len} fields
set to identify the locked region.

If no lock applies, the only change to the @var{lockp} structure is to
update the @code{l_type} to a value of @code{F_UNLCK}.

The normal return value from @code{fcntl} with this command is an
unspecified value other than @math{-1}, which is reserved to indicate an
error.  The following @code{errno} error conditions are defined for
this command:

@table @code
@item EBADF
The @var{filedes} argument is invalid.

@item EINVAL
Either the @var{lockp} argument doesn't specify valid lock information,
or the file associated with @var{filedes} doesn't support locks.
@end table
@end deftypevr

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int F_SETLK
This macro is used as the @var{command} argument to @code{fcntl}, to
specify that it should set or clear a lock.  This command requires a
third argument of type @w{@code{struct flock *}} to be passed to
@code{fcntl}, so that the form of the call is:

@smallexample
fcntl (@var{filedes}, F_SETLK, @var{lockp})
@end smallexample

If the process already has a lock on any part of the region, the old lock
on that part is replaced with the new lock.  You can remove a lock
by specifying a lock type of @code{F_UNLCK}.

If the lock cannot be set, @code{fcntl} returns immediately with a value
of @math{-1}.  This function does not block waiting for other processes
to release locks.  If @code{fcntl} succeeds, it return a value other
than @math{-1}.

The following @code{errno} error conditions are defined for this
function:

@table @code
@item EAGAIN
@itemx EACCES
The lock cannot be set because it is blocked by an existing lock on the
file.  Some systems use @code{EAGAIN} in this case, and other systems
use @code{EACCES}; your program should treat them alike, after
@code{F_SETLK}.  (The GNU system always uses @code{EAGAIN}.)

@item EBADF
Either: the @var{filedes} argument is invalid; you requested a read lock
but the @var{filedes} is not open for read access; or, you requested a
write lock but the @var{filedes} is not open for write access.

@item EINVAL
Either the @var{lockp} argument doesn't specify valid lock information,
or the file associated with @var{filedes} doesn't support locks.

@item ENOLCK
The system has run out of file lock resources; there are already too
many file locks in place.

Well-designed file systems never report this error, because they have no
limitation on the number of locks.  However, you must still take account
of the possibility of this error, as it could result from network access
to a file system on another machine.
@end table
@end deftypevr

@comment fcntl.h
@comment POSIX.1
@deftypevr Macro int F_SETLKW
This macro is used as the @var{command} argument to @code{fcntl}, to
specify that it should set or clear a lock.  It is just like the
@code{F_SETLK} command, but causes the process to block (or wait)
until the request can be specified.

This command requires a third argument of type @code{struct flock *}, as
for the @code{F_SETLK} command.

The @code{fcntl} return values and errors are the same as for the
@code{F_SETLK} command, but these additional @code{errno} error conditions
are defined for this command:

@table @code
@item EINTR
The function was interrupted by a signal while it was waiting.
@xref{Interrupted Primitives}.

@item EDEADLK
The specified region is being locked by another process.  But that
process is waiting to lock a region which the current process has
locked, so waiting for the lock would result in deadlock.  The system
does not guarantee that it will detect all such conditions, but it lets
you know if it notices one.
@end table
@end deftypevr


The following macros are defined for use as values for the @code{l_type}
member of the @code{flock} structure.  The values are integer constants.

@table @code
@comment fcntl.h
@comment POSIX.1
@vindex F_RDLCK
@item F_RDLCK
This macro is used to specify a read (or shared) lock.

@comment fcntl.h
@comment POSIX.1
@vindex F_WRLCK
@item F_WRLCK
This macro is used to specify a write (or exclusive) lock.

@comment fcntl.h
@comment POSIX.1
@vindex F_UNLCK
@item F_UNLCK
This macro is used to specify that the region is unlocked.
@end table

As an example of a situation where file locking is useful, consider a
program that can be run simultaneously by several different users, that
logs status information to a common file.  One example of such a program
might be a game that uses a file to keep track of high scores.  Another
example might be a program that records usage or accounting information
for billing purposes.

Having multiple copies of the program simultaneously writing to the
file could cause the contents of the file to become mixed up.  But
you can prevent this kind of problem by setting a write lock on the
file before actually writing to the file.

If the program also needs to read the file and wants to make sure that
the contents of the file are in a consistent state, then it can also use
a read lock.  While the read lock is set, no other process can lock
that part of the file for writing.

@c ??? This section could use an example program.

Remember that file locks are only a @emph{voluntary} protocol for
controlling access to a file.  There is still potential for access to
the file by programs that don't use the lock protocol.

@node Interrupt Input
@section Interrupt-Driven Input

@cindex interrupt-driven input
If you set the @code{O_ASYNC} status flag on a file descriptor
(@pxref{File Status Flags}), a @code{SIGIO} signal is sent whenever
input or output becomes possible on that file descriptor.  The process
or process group to receive the signal can be selected by using the
@code{F_SETOWN} command to the @code{fcntl} function.  If the file
descriptor is a socket, this also selects the recipient of @code{SIGURG}
signals that are delivered when out-of-band data arrives on that socket;
see @ref{Out-of-Band Data}.  (@code{SIGURG} is sent in any situation
where @code{select} would report the socket as having an ``exceptional
condition''.  @xref{Waiting for I/O}.)

If the file descriptor corresponds to a terminal device, then @code{SIGIO}
signals are sent to the foreground process group of the terminal.
@xref{Job Control}.

@pindex fcntl.h
The symbols in this section are defined in the header file
@file{fcntl.h}.

@comment fcntl.h
@comment BSD
@deftypevr Macro int F_GETOWN
This macro is used as the @var{command} argument to @code{fcntl}, to
specify that it should get information about the process or process
group to which @code{SIGIO} signals are sent.  (For a terminal, this is
actually the foreground process group ID, which you can get using
@code{tcgetpgrp}; see @ref{Terminal Access Functions}.)

The return value is interpreted as a process ID; if negative, its
absolute value is the process group ID.

The following @code{errno} error condition is defined for this command:

@table @code
@item EBADF
The @var{filedes} argument is invalid.
@end table
@end deftypevr

@comment fcntl.h
@comment BSD
@deftypevr Macro int F_SETOWN
This macro is used as the @var{command} argument to @code{fcntl}, to
specify that it should set the process or process group to which
@code{SIGIO} signals are sent.  This command requires a third argument
of type @code{pid_t} to be passed to @code{fcntl}, so that the form of
the call is:

@smallexample
fcntl (@var{filedes}, F_SETOWN, @var{pid})
@end smallexample

The @var{pid} argument should be a process ID.  You can also pass a
negative number whose absolute value is a process group ID.

The return value from @code{fcntl} with this command is @math{-1}
in case of error and some other value if successful.  The following
@code{errno} error conditions are defined for this command:

@table @code
@item EBADF
The @var{filedes} argument is invalid.

@item ESRCH
There is no process or process group corresponding to @var{pid}.
@end table
@end deftypevr

@c ??? This section could use an example program.

@node IOCTLs
@section Generic I/O Control operations
@cindex generic i/o control operations
@cindex IOCTLs

The GNU system can handle most input/output operations on many different
devices and objects in terms of a few file primitives - @code{read},
@code{write} and @code{lseek}.  However, most devices also have a few
peculiar operations which do not fit into this model. Such as:

@itemize @bullet

@item
Changing the character font used on a terminal.

@item
Telling a magnetic tape system to rewind or fast forward.  (Since they
cannot move in byte increments, @code{lseek} is inapplicable).

@item
Ejecting a disk from a drive.

@item
Playing an audio track from a CD-ROM drive.

@item
Maintaining routing tables for a network.

@end itemize

Although some such objects such as sockets and terminals
@footnote{Actually, the terminal-specific functions are implemented with
IOCTLs on many platforms.} have special functions of their own, it would
not be practical to create functions for all these cases.

Instead these minor operations, known as @dfn{IOCTL}s, are assigned code
numbers and multiplexed through the @code{ioctl} function, defined in
@code{sys/ioctl.h}.  The code numbers themselves are defined in many
different headers.

@deftypefun int ioctl (int @var{filedes}, int @var{command}, @dots{})

The @code{ioctl} function performs the generic I/O operation
@var{command} on @var{filedes}.

A third argument is usually present, either a single number or a pointer
to a structure.  The meaning of this argument, the returned value, and
any error codes depends upon the command used.  Often @math{-1} is
returned for a failure.

@end deftypefun

On some systems, IOCTLs used by different devices share the same numbers.
Thus, although use of an inappropriate IOCTL @emph{usually} only produces
an error, you should not attempt to use device-specific IOCTLs on an
unknown device.

Most IOCTLs are OS-specific and/or only used in special system utilities,
and are thus beyond the scope of this document.  For an example of the use
of an IOCTL, see @ref{Out-of-Band Data}.