(split) Import translated manuals from JM CVS Repository.

[linuxjm/LDP_man-pages.git] / draft / man2 / fcntl.2

'\" t
.\" Hey Emacs! This file is -*- nroff -*- source.
.\"
.\" This manpage is Copyright (C) 1992 Drew Eckhardt;
.\"                 and Copyright (C) 1993 Michael Haardt, Ian Jackson;
.\"                 and Copyright (C) 1998 Jamie Lokier;
.\"                 and Copyright (C) 2002 Michael Kerrisk.
.\"
.\" Permission is granted to make and distribute verbatim copies of this
.\" manual provided the copyright notice and this permission notice are
.\" preserved on all copies.
.\"
.\" Permission is granted to copy and distribute modified versions of this
.\" manual under the conditions for verbatim copying, provided that the
.\" entire resulting derived work is distributed under the terms of a
.\" permission notice identical to this one.
.\"
.\" Since the Linux kernel and libraries are constantly changing, this
.\" manual page may be incorrect or out-of-date.  The author(s) assume no
.\" responsibility for errors or omissions, or for damages resulting from
.\" the use of the information contained herein.  The author(s) may not
.\" have taken the same level of care in the production of this manual,
.\" which is licensed free of charge, as they might when working
.\" professionally.
.\"
.\" Formatted or processed versions of this manual, if unaccompanied by
.\" the source, must acknowledge the copyright and authors of this work.
.\"
.\" Modified 1993-07-24 by Rik Faith <faith@cs.unc.edu>
.\" Modified 1995-09-26 by Andries Brouwer <aeb@cwi.nl>
.\" and again on 960413 and 980804 and 981223.
.\" Modified 1998-12-11 by Jamie Lokier <jamie@imbolc.ucc.ie>
.\" Applied correction by Christian Ehrhardt - aeb, 990712
.\" Modified 2002-04-23 by Michael Kerrisk <mtk.manpages@gmail.com>
.\"     Added note on F_SETFL and O_DIRECT
.\"     Complete rewrite + expansion of material on file locking
.\"     Incorporated description of F_NOTIFY, drawing on
.\"             Stephen Rothwell's notes in Documentation/dnotify.txt.
.\"     Added description of F_SETLEASE and F_GETLEASE
.\" Corrected and polished, aeb, 020527.
.\" Modified 2004-03-03 by Michael Kerrisk <mtk.manpages@gmail.com>
.\"     Modified description of file leases: fixed some errors of detail
.\"     Replaced the term "lease contestant" by "lease breaker"
.\" Modified, 27 May 2004, Michael Kerrisk <mtk.manpages@gmail.com>
.\"     Added notes on capability requirements
.\" Modified 2004-12-08, added O_NOATIME after note from Martin Pool
.\" 2004-12-10, mtk, noted F_GETOWN bug after suggestion from aeb.
.\" 2005-04-08 Jamie Lokier <jamie@shareable.org>, mtk
.\"     Described behavior of F_SETOWN/F_SETSIG in
.\"     multithreaded processes, and generally cleaned
.\"     up the discussion of F_SETOWN.
.\" 2005-05-20, Johannes Nicolai <johannes.nicolai@hpi.uni-potsdam.de>,
.\"     mtk: Noted F_SETOWN bug for socket file descriptor in Linux 2.4
.\"     and earlier.  Added text on permissions required to send signal.
.\" 2009-09-30, Michael Kerrisk
.\"     Note obsolete F_SETOWN behavior with threads.
.\"     Document F_SETOWN_EX and F_GETOWN_EX
.\"
.\" Japanese Version Copyright (c) 1996 Takeshi Ueno
.\" and Copyright (c) 2005, 2006, 2008 Akihiro MOTOKI
.\"
.\" Translated 1996-07-03, Takeshi Ueno <tueno@vio.co.jp>
.\" Modified 1998-09-10, HANATAKA Shinya <hanataka@abyss.rim.or.jp>
.\" Modified 1999-08-14, HANATAKA Shinya <hanataka@abyss.rim.or.jp>
.\" Updated & Modified 2001-04-03, Yuichi SATO <ysato@h4.dion.ne.jp>
.\" Updated & Modified 2005-03-15, Akihiro MOTOKI <amotoki@dd.iij4u.or.jp>
.\" Updated & Modified 2005-04-22, Akihiro MOTOKI
.\" Updated & Modified 2005-10-14, Akihiro MOTOKI
.\" Updated & Modified 2005-11-19, Akihiro MOTOKI, LDP v2.14
.\" Updated 2006-04-16, Akihiro MOTOKI, LDP v2.29
.\" Updated 2008-02-11, Akihiro MOTOKI, LDP v2.77
.\" Updated 2008-09-19, Akihiro MOTOKI, LDP v3.09
.\" Updated 2010-04-23, Akihiro MOTOKI, LDP v3.24
.\"
.\"WORD:        asynchronous I/O        ÈóƱ´ü I/O
.\"WORD:        descriptor              ¥Ç¥£¥¹¥¯¥ê¥×¥¿
.\"WORD:        open file description   ¥ª¡¼¥×¥ó¥Õ¥¡¥¤¥ëµ­½Ò
.\"WORD:        feature test macro      µ¡Ç½¸¡ºº¥Þ¥¯¥í
.\"WORD:        I/O availability signal I/O ¤¬ÍøÍѲÄǽ¤Ë¤Ê¤Ã¤¿¤³¤È¤ò¼¨¤¹¥·¥°¥Ê¥ë
.\"
.TH FCNTL 2 2009-10-17 "Linux" "Linux Programmer's Manual"
.\"O .SH NAME
.SH ̾Á°
.\"O fcntl \- manipulate file descriptor
fcntl \- ¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤ÎÁàºî¤ò¹Ô¤¦
.\"O .SH SYNOPSIS
.SH ½ñ¼°
.nf
.B #include <unistd.h>
.B #include <fcntl.h>
.sp
.BI "int fcntl(int " fd ", int " cmd ", ... /* " arg " */ );"
.fi
.\"O .SH DESCRIPTION
.SH ÀâÌÀ
.\"O .BR fcntl ()
.\"O performs one of the operations described below on the open file descriptor
.\"O .IR fd .
.\"O The operation is determined by
.\"O .IR cmd .
.BR fcntl ()
¤Ï¡¢¥ª¡¼¥×¥ó¤µ¤ì¤¿¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿
.I fd
¤Ë´Ø¤·¤Æ²¼µ­¤ÎÁàºî¤ò¹Ô¤¦¡£Áàºî¤Ï
.I cmd
¤Ë¤è¤Ã¤Æ·è¤Þ¤ë:

.\"O .BR fcntl ()
.\"O can take an optional third argument.
.\"O Whether or not this argument is required is determined by
.\"O .IR cmd .
.\"O The required argument type is indicated in parentheses after each
.\"O .I cmd
.\"O name (in most cases, the required type is
.\"O .IR long ,
.\"O and we identify the argument using the name
.\"O .IR arg ),
.\"O or
.\"O .I void
.\"O is specified if the argument is not required.
.BR fcntl ()
¤Ï¥ª¥×¥·¥ç¥ó¤È¤·¤ÆÂè»°°ú¤­¿ô¤ò¤È¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
Âè»°°ú¤­¿ô¤¬É¬Íפ«¤É¤¦¤«¤Ï
.I cmd
¤Ë¤è¤ê·è¤Þ¤ë¡£
ɬÍפʰú¤­¿ô¤Î·¿¤Ï
.I cmd
̾¤Î¸å¤í¤Î³ç¸ÌÆâ¤Ç»ØÄꤵ¤ì¤Æ¤¤¤ë
(¤Û¤È¤ó¤É¤Î¾ì¹ç¡¢É¬Íפʷ¿¤Ï
.I long
¤Ç¤¢¤ê¡¢¤³¤Î°ú¤­¿ô¤òɽ¤¹¤Î¤Ë
.I arg
¤È¤¤¤¦Ì¾Á°¤ò»È¤Ã¤Æ¤¤¤ë)¡£
°ú¤­¿ô¤¬É¬Íפʤ¤¾ì¹ç¤Ë¤Ï
.I void
¤¬»ØÄꤵ¤ì¤Æ¤¤¤ë¡£
.\"O .SS "Duplicating a file descriptor"
.SS "¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤ÎÊ£À½"
.TP
.BR F_DUPFD " (\fIlong\fP)"
.\"O Find the lowest numbered available file descriptor
.\"O greater than or equal to
.\"O .I arg
.\"O and make it be a copy of
.\"O .IR fd .
.\"O This is different from
.\"O .BR dup2 (2),
.\"O which uses exactly the descriptor specified.
ÍøÍѲÄǽ¤Ê¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Î¤¦¤Á¡¢
.I arg
°Ê¾å¤ÇºÇ¾®¤Î¤â¤Î¤òõ¤·¡¢
.I fd
¤Î¥³¥Ô¡¼¤È¤¹¤ë¡£¤³¤ì¤ÏÊ̤ηÁ¤Î
.BR dup2 (2)
¤Ç¤¢¤ë¡£
.BR dup2 (2)
¤Ç¤Ï»ØÄꤵ¤ì¤¿¥Ç¥£¥¹¥¯¥ê¥×¥¿¤¬»È¤ï¤ì¤ëÅÀ¤¬°ã¤¦¡£
.IP
.\"O On success, the new descriptor is returned.
À®¸ù¤¹¤ë¤È¡¢¿·¤·¤¤¥Ç¥£¥¹¥¯¥ê¥×¥¿¤¬ÊÖ¤µ¤ì¤ë¡£
.IP
.\"O See
.\"O .BR dup (2)
.\"O for further details.
¾ÜºÙ¤Ï
.BR dup (2)
¤ò»²¾È¤Î¤³¤È¡£
.TP
.\"O .BR F_DUPFD_CLOEXEC " (\fIlong\fP; since Linux 2.6.24)"
.BR F_DUPFD_CLOEXEC " (\fIlong\fP; Linux 2.6.24 °Ê¹ß)"
.\"O As for
.\"O .BR F_DUPFD ,
.\"O but additionally set the
.\"O close-on-exec flag for the duplicate descriptor.
.B F_DUPFD
¤ÈƱÍͤÀ¤¬¡¢¤½¤ì¤Ë²Ã¤¨¤ÆÊ£À½¤µ¤ì¤¿¥Ç¥£¥¹¥¯¥ê¥×¥¿¤ËÂФ·¤Æ
close-on-exec ¥Õ¥é¥°¤ò¥»¥Ã¥È¤¹¤ë¡£
.\"O Specifying this flag permits a program to avoid an additional
.\"O .BR fcntl ()
.\"O .B F_SETFD
.\"O operation to set the
.\"O .B FD_CLOEXEC
.\"O flag.
¤³¤Î¥Õ¥é¥°¤ò»ØÄꤹ¤ë¤³¤È¤Ç¡¢¥×¥í¥°¥é¥à¤Ï
.B FD_CLOEXEC
¥Õ¥é¥°¤ò¥»¥Ã¥È¤¹¤ë¤¿¤á¤Ë
.BR fcntl ()
¤Î
.B F_SETFD
Áàºî¤òÄɲäǹԤ¦É¬Íפ¬¤Ê¤¯¤Ê¤ë¡£
.\"O For an explanation of why this flag is useful,
.\"O see the description of
.\"O .B O_CLOEXEC
.\"O in
.\"O .BR open (2).
¤³¤Î¥Õ¥é¥°¤¬¤Ê¤¼Í­ÍѤ«¤Ë¤Ä¤¤¤Æ¤Ï¡¢
.BR open (2)
¤Î
.B O_CLOEXEC
¤ÎÀâÌÀ¤ò»²¾È¤Î¤³¤È¡£
.\"O .SS "File descriptor flags"
.SS "¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¡¦¥Õ¥é¥°"
.\"O The following commands manipulate the flags associated with
.\"O a file descriptor.
.\"O Currently, only one such flag is defined:
.\"O .BR FD_CLOEXEC ,
.\"O the close-on-exec flag.
.\"O If the
.\"O .B FD_CLOEXEC
.\"O bit is 0, the file descriptor will remain open across an
.\"O .BR execve (2),
.\"O otherwise it will be closed.
°Ê²¼¤Î¥³¥Þ¥ó¥É¤ò»È¤Ã¤Æ¡¢¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Ë´ØÏ¢¤¹¤ë¥Õ¥é¥°
¤òÁàºî¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
¸½ºß¤Î¤È¤³¤í¡¢ÄêµÁ¤µ¤ì¤Æ¤¤¤ë¥Õ¥é¥°¤Ï°ì¤Ä¤À¤±¤Ç¤¢¤ë:
.B FD_CLOEXEC
(close-on-exec ¥Õ¥é¥°)¡£
.B FD_CLOEXEC
¥Ó¥Ã¥È¤¬ 0 ¤Ê¤é¡¢¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Ï
.BR execve (2)
¤ò¹Ô¤Ã¤Æ¤â¥ª¡¼¥×¥ó¤µ¤ì¤¿¤Þ¤Þ¤À¤¬¡¢¤½¤¦¤Ç¤Ê¤¤¾ì¹ç¤Ï¥¯¥í¡¼¥º¤µ¤ì¤ë¡£
.TP
.BR F_GETFD " (\fIvoid\fP)"
.\"O Read the file descriptor flags;
.\"O .I arg
.\"O is ignored.
¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¡¦¥Õ¥é¥°¤òÆɤ߽Ф¹¡£
.I arg
¤Ï̵»ë¤µ¤ì¤ë¡£
.TP
.BR F_SETFD " (\fIlong\fP)"
.\"O Set the file descriptor flags to the value specified by
.\"O .IR arg .
¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¡¦¥Õ¥é¥°¤Ë
.I arg
¤Ç»ØÄꤷ¤¿ÃͤòÀßÄꤹ¤ë¡£
.\"O .SS "File status flags"
.SS "¥Õ¥¡¥¤¥ë¾õÂ֥ե饰"
.\"O Each open file description has certain associated status flags,
.\"O initialized by
.\"O .BR open (2)
.\"O .\" or
.\"O .\" .BR creat (2),
.\"O and possibly modified by
.\"O .BR fcntl ().
.\"O Duplicated file descriptors
.\"O (made with
.\"O .BR dup (2),
.\"O .BR fcntl (F_DUPFD),
.\"O .BR fork (2),
.\"O etc.) refer to the same open file description, and thus
.\"O share the same file status flags.
¥ª¡¼¥×¥ó¥Õ¥¡¥¤¥ëµ­½Ò (open file description) ¤Ë¤Ï¡¢
¥Õ¥¡¥¤¥ëµ­½ÒËè¤ËÀßÄꤵ¤ì¤ë¾õÂ֥ե饰¤¬¤¤¤¯¤Ä¤«¤¢¤ë¡£¤³¤ì¤é¤Î¥Õ¥é¥°¤Ï
.BR open (2)
.\" ¤ä
.\" .BR creat (2)
¤Ë¤è¤Ã¤Æ½é´ü²½¤µ¤ì¡¢
.BR fcntl (2)
¤Ë¤è¤êÊѹ¹¤¹¤ë¤³¤È¤â¤Ç¤­¤ë¡£¤³¤ì¤é¤Ï¡¢
.RB ( dup (2),
.BR fcntl (F_DUPFD),
.BR fork (2)
¤Ê¤É¤Ç) Ê£À½¤µ¤ì¤¿¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿Æ±»Î¤Ï
Ʊ¤¸¥ª¡¼¥×¥ó¥Õ¥¡¥¤¥ëµ­½Ò¤ò»²¾È¤¹¤ë¡£
¤½¤Î¤¿¤á¡¢
Ʊ¤¸¥Õ¥¡¥¤¥ë¾õÂ֥ե饰¤¬¶¦Í­¤µ¤ì¤ë¡£

.\"O The file status flags and their semantics are described in
.\"O .BR open (2).
¥Õ¥¡¥¤¥ë¾õÂ֥ե饰¤È¤½¤Î°ÕÌ£¤Ï
.BR open (2)
¤ÇÀâÌÀ¤µ¤ì¤Æ¤¤¤ë¡£
.TP
.BR F_GETFL " (\fIvoid\fP)"
.\"O Read the file status flags;
.\"O .I arg
.\"O is ignored.
¥Õ¥¡¥¤¥ë¾õÂ֥ե饰¤òÆɤ߽Ф¹¡£
.I arg
¤Ï̵»ë¤µ¤ì¤ë¡£
.TP
.BR F_SETFL " (\fIlong\fP)"
.\"O Set the file status flags to the value specified by
.\"O .IR arg .
.\"O File access mode
.\"O .RB ( O_RDONLY ", " O_WRONLY ", " O_RDWR )
.\"O and file creation flags
.\"O (i.e.,
.\"O .BR O_CREAT ", " O_EXCL ", " O_NOCTTY ", " O_TRUNC )
.\"O in
.\"O .I arg
.\"O are ignored.
¥Õ¥¡¥¤¥ë¾õÂ֥ե饰¤Ë
.I arg
¤Ç»ØÄꤵ¤ì¤¿ÃͤòÀßÄꤹ¤ë¡£
.I arg
¤Î¤¦¤Á¡¢¥Õ¥¡¥¤¥ë¤Î¥¢¥¯¥»¥¹¥â¡¼¥É
.RB ( O_RDONLY ", " O_WRONLY ", " O_RDWR )
¤È¥Õ¥¡¥¤¥ëºîÀ®¥Õ¥é¥° (¤¹¤Ê¤ï¤Á
.BR O_CREAT ", " O_EXCL ", " O_NOCTTY ", " O_TRUNC )
¤Ë´Ø¤¹¤ë¥Ó¥Ã¥È¤Ï̵»ë¤µ¤ì¤ë¡£
.\"O On Linux this command can only change the
.\"O .BR O_APPEND ,
.\"O .BR O_ASYNC ,
.\"O .BR O_DIRECT ,
.\"O .BR O_NOATIME ,
.\"O and
.\"O .B O_NONBLOCK
.\"O flags.
Linux ¤Ç¤Ï¡¢¤³¤Î¥³¥Þ¥ó¥É¤ÇÊѹ¹¤Ç¤­¤ë¤Î¤Ï
.BR O_APPEND ,
.BR O_ASYNC ,
.BR O_DIRECT ,
.BR O_NOATIME ,
.B O_NONBLOCK
¥Õ¥é¥°¤À¤±¤Ç¤¢¤ë¡£
.\"O .\" FIXME . According to POSIX.1-2001, O_SYNC should also be modifiable
.\"O .\" via fcntl(2), but currently Linux does not permit this
.\"O .\" See http://bugzilla.kernel.org/show_bug.cgi?id=5994
.\" FIXME . POSIX.1-2001 ¤Ë¤è¤ë¤È¡¢ O_SYNC ¤â fcntl(2) ¤ÇÊѹ¹¤Ç¤­¤ë¤Ù¤­¤À¤¬¡¢
.\" ¸½ºß¤Î¤È¤³¤í Linux ¤Ç¤Ïµö²Ä¤µ¤ì¤Æ¤¤¤Ê¤¤¡£
.\" http://bugzilla.kernel.org/show_bug.cgi?id=5994 »²¾È
.\"O .SS "Advisory locking"
.SS "¥¢¥É¥Ð¥¤¥¶¥ê¡¦¥í¥Ã¥¯"
.\"O .BR F_GETLK ", " F_SETLK " and " F_SETLKW
.\"O are used to acquire, release, and test for the existence of record
.\"O locks (also known as file-segment or file-region locks).
.\"O The third argument,
.\"O .IR lock ,
.\"O is a pointer to a structure that has at least the following fields
.\"O (in unspecified order).
.BR F_GETLK ", " F_SETLK ", " F_SETLKW
¤Ï¡¢¥ì¥³¡¼¥É¡¦¥í¥Ã¥¯¤Î³ÍÆÀ¡¿²òÊü¡¿¥Æ¥¹¥È¤Î¤¿¤á¤Ë»ÈÍѤ¹¤ë
(¥ì¥³¡¼¥É¡¦¥í¥Ã¥¯¤Ï¥Õ¥¡¥¤¥ë¥»¥°¥á¥ó¥È¡¦¥í¥Ã¥¯¤ä
¥Õ¥¡¥¤¥ëÎΰè¥í¥Ã¥¯¤È¤â¸Æ¤Ð¤ì¤ë)¡£
»°ÈÖÌܤΰú¤­¿ô
.I lock
¤Ï¡¢°Ê²¼¤Ë¼¨¤¹¥Õ¥£¡¼¥ë¥É¤ò´Þ¤à¹½Â¤ÂΤؤΥݥ¤¥ó¥¿¤Ç¤¢¤ë
(¥Õ¥£¡¼¥ë¥É¤Î½ç½ø¤Ï´Ø·¸¤Ê¤¯¡¢¹½Â¤ÂΤ˾¤Î¥Õ¥£¡¼¥ë¥É¤¬¤¢¤Ã¤Æ¤â¤è¤¤)¡£
.in +4n
.nf
.sp
struct flock {
    ...
    short l_type;    /* Type of lock: F_RDLCK,
                        F_WRLCK, F_UNLCK */
    short l_whence;  /* How to interpret l_start:
                        SEEK_SET, SEEK_CUR, SEEK_END */
    off_t l_start;   /* Starting offset for lock */
    off_t l_len;     /* Number of bytes to lock */
    pid_t l_pid;     /* PID of process blocking our lock
                        (F_GETLK only) */
    ...
};
.fi
.in
.P
.\"O The
.\"O .IR l_whence ", " l_start ", and " l_len
.\"O fields of this structure specify the range of bytes we wish to lock.
.\"O Bytes past the end of the file may be locked,
.\"O but not bytes before the start of the file.
¤³¤Î¹½Â¤ÂΤÎ
.IR l_whence ", " l_start ", " l_len
¥Õ¥£¡¼¥ë¥É¤Ç¡¢¥í¥Ã¥¯¤ò¹Ô¤¤¤¿¤¤¥Ð¥¤¥ÈÈϰϤò»ØÄꤹ¤ë¡£
¥Õ¥¡¥¤¥ë¤ÎËöÈø¤è¤ê¸å¤í¤Î¥Ð¥¤¥È¤ò¥í¥Ã¥¯¤¹¤ë¤³¤È¤Ï¤Ç¤­¤ë¤¬¡¢
¥Õ¥¡¥¤¥ë¤ÎÀèƬ¤è¤êÁ°¤Î¥Ð¥¤¥È¤ò¥í¥Ã¥¯¤¹¤ë¤³¤È¤Ï¤Ç¤­¤Ê¤¤¡£

.\"O .I l_start
.\"O is the starting offset for the lock, and is interpreted
.\"O relative to either:
.\"O the start of the file (if
.\"O .I l_whence
.\"O is
.\"O .BR SEEK_SET );
.\"O the current file offset (if
.\"O .I l_whence
.\"O is
.\"O .BR SEEK_CUR );
.\"O or the end of the file (if
.\"O .I l_whence
.\"O is
.\"O .BR SEEK_END ).
.\"O In the final two cases,
.\"O .I l_start
.\"O can be a negative number provided the
.\"O offset does not lie before the start of the file.
.I l_start
¤Ï¥í¥Ã¥¯¤ò¹Ô¤¦Îΰè¤Î³«»Ï¥ª¥Õ¥»¥Ã¥È¤Ç¤¢¤ë¡£
¤½¤Î°ÕÌ£¤Ï
.I l_whence
¤Ë¤è¤ê°Û¤Ê¤ë:
.I l_whence
¤¬
.B SEEK_SET
¤Î¾ì¹ç¤Ï¥Õ¥¡¥¤¥ë¤ÎÀèƬ¤«¤é¤Î¥ª¥Õ¥»¥Ã¥È¡¢
.I l_whence
¤¬
.B SEEK_CUR
¤Î¾ì¹ç¤Ï¸½ºß¤Î¥Õ¥¡¥¤¥ë¥ª¥Õ¥»¥Ã¥È¤«¤é¤Î¥ª¥Õ¥»¥Ã¥È¡¢
.I l_whence
¤¬
.B SEEK_END
¤Î¾ì¹ç¤Ï¥Õ¥¡¥¤¥ë¤ÎËöÈø¤«¤é¤Î¥ª¥Õ¥»¥Ã¥È¤È²ò¼á¤µ¤ì¤ë¡£
¸å¤í¤Î£²¤Ä¤Î¾ì¹ç¤Ë¤Ï¡¢
¥Õ¥¡¥¤¥ë¤ÎÀèƬ¤è¤êÁ°¤Ë¤Ê¤é¤Ê¤¤ÈϰϤǡ¢
.I l_start
¤ËÉé¤ÎÃͤò»ØÄꤹ¤ë¤³¤È¤¬¤Ç¤­¤ë¡£

.\"O .I l_len
.\"O specifies the number of bytes to be locked.
.\"O If
.\"O .I l_len
.\"O is positive, then the range to be locked covers bytes
.\"O .I l_start
.\"O up to and including
.\"O .IR l_start + l_len \- 1 .
.\"O Specifying 0 for
.\"O .I l_len
.\"O has the special meaning: lock all bytes starting at the
.\"O location specified by
.\"O .IR l_whence " and " l_start
.\"O through to the end of file, no matter how large the file grows.
.I l_len
¤Ï¥í¥Ã¥¯¤·¤¿¤¤¥Ð¥¤¥È¿ô¤ò¼¨¤¹¡£
.I l_len
¤¬Àµ¤Î¾ì¹ç¡¢¥í¥Ã¥¯¤µ¤ì¤ë¥Ð¥¤¥ÈÈϰϤÏ
.I l_start
°Ê¾å
.IR l_start + l_len \- 1
°Ê²¼¤È¤Ê¤ë¡£
.I l_len
¤Ë 0 ¤ò»ØÄꤷ¤¿¾ì¹ç¤ÏÆÃÊ̤ʰÕÌ£¤ò»ý¤Ä:
.IR l_whence " and " l_start
¤Ç»ØÄꤵ¤ì¤ë°ÌÃÖ¤«¤é¥Õ¥¡¥¤¥ë¤ÎËöÈø¤Þ¤Ç¤ÎÁ´¤Æ¤Î¥Ð¥¤¥È¤ò¥í¥Ã¥¯¤¹¤ë
(¥Õ¥¡¥¤¥ë¤¬¤É¤ó¤Ê¤ËÂ礭¤¯¤Ê¤Ã¤¿¤È¤·¤Æ¤â¥Õ¥¡¥¤¥ë¤ÎËöÈø¤Þ¤Ç¥í¥Ã¥¯¤¹¤ë)¡£

.\"O POSIX.1-2001 allows (but does not require)
.\"O an implementation to support a negative
.\"O .I l_len
.\"O value; if
.\"O .I l_len
.\"O is negative, the interval described by
.\"O .I lock
.\"O covers bytes
.\"O .IR l_start + l_len
.\"O up to and including
.\"O .IR l_start \-1.
.\"O This is supported by Linux since kernel versions 2.4.21 and 2.5.49.
POSIX.1-2001 ¤Ç¤Ï¡¢Éé¤ÎÃͤÎ
.I l_len
¤ò¥µ¥Ý¡¼¥È¤¹¤ë¼ÂÁõ¤òǧ¤á¤Æ¤¤¤ë (ɬ¿Ü¤Ç¤Ï¤Ê¤¤)¡£
.I l_len
¤¬Éé¤Î¾ì¹ç¡¢¥í¥Ã¥¯¤µ¤ì¤ë¥Ð¥¤¥ÈÈϰϤÏ
.IR l_start + l_len
°Ê¾å
.IR l_start \-1
°Ê²¼¤È¤Ê¤ë¡£
¤³¤ÎÆ°ºî¤Ï¥«¡¼¥Í¥ë 2.4.21 °Ê¹ß¤ª¤è¤Ó 2.5.49 °Ê¹ß¤Î Linux ¤Ç
¥µ¥Ý¡¼¥È¤µ¤ì¤Æ¤¤¤ë¡£

.\"O The
.\"O .I l_type
.\"O field can be used to place a read
.\"O .RB ( F_RDLCK )
.\"O or a write
.\"O .RB ( F_WRLCK )
.\"O lock on a file.
.\"O Any number of processes may hold a read lock (shared lock)
.\"O on a file region, but only one process may hold a write lock
.\"O (exclusive lock).
.\"O An exclusive lock excludes all other locks,
.\"O both shared and exclusive.
.I l_type
¥Õ¥£¡¼¥ë¥É¤Ï¡¢¥Õ¥¡¥¤¥ë¤ËÂФ·¤ÆÆɤ߽Ф·¥í¥Ã¥¯
.RB ( F_RDLCK )
¤È½ñ¤­¹þ¤ß¥í¥Ã¥¯
.RB ( F_WRLCK )
¤Î¤É¤Á¤é¤ò
ÀßÄꤹ¤ë¤«¤ò»ØÄꤹ¤ë¡£
¥Õ¥¡¥¤¥ë¤Î¤¢¤ëÎΰè¤ËÂФ·¤Æ¡¢Æɤ߽Ф·¥í¥Ã¥¯ (¶¦Í­¥í¥Ã¥¯) ¤òÊÝ»ý¤Ç¤­¤ë
¥×¥í¥»¥¹¿ô¤ËÀ©¸Â¤Ï¤Ê¤¤¤¬¡¢½ñ¤­¹þ¤ß¥í¥Ã¥¯ (ÇÓ¾¥í¥Ã¥¯) ¤òÊÝ»ý¤Ç¤­¤ë
¤Î¤Ï°ì¤Ä¤Î¥×¥í¥»¥¹¤À¤±¤Ç¤¢¤ë¡£ÇÓ¾¥í¥Ã¥¯¤òÀßÄꤹ¤ë¤È¡¢(¶¦Í­¥í¥Ã¥¯¤«
ÇÓ¾¥í¥Ã¥¯¤Ë¤«¤«¤ï¤é¤º) ¾¤Î¥í¥Ã¥¯¤Ï²¿¤âÀßÄê¤Ç¤­¤Ê¤¤¡£
.\"O A single process can hold only one type of lock on a file region;
.\"O if a new lock is applied to an already-locked region,
.\"O then the existing lock is converted to the new lock type.
.\"O (Such conversions may involve splitting, shrinking, or coalescing with
.\"O an existing lock if the byte range specified by the new lock does not
.\"O precisely coincide with the range of the existing lock.)
°ì¤Ä¤Î¥×¥í¥»¥¹¤Ï¡¢¥Õ¥¡¥¤¥ë¤Î¤¢¤ëÎΰè¤ËÂФ·¤Æ°ì¼ïÎà¤Î¥í¥Ã¥¯¤·¤«ÊÝ»ý¤Ç¤­¤Ê¤¤¡£
¿·µ¬¤Î¥í¥Ã¥¯¤¬¥í¥Ã¥¯¤¬ÀßÄꤵ¤ì¤Æ¤¤¤ëÎΰè¤ËÂФ·¤ÆŬÍѤµ¤ì¤ë¤È¡¢´û¸¤Î¥í¥Ã¥¯
¤Ï¿·µ¬¤Î¥í¥Ã¥¯¤Î¼ïÊ̤ËÊÑ´¹¤µ¤ì¤ë
(¿·µ¬¤Î¥í¥Ã¥¯¤Ç»ØÄꤵ¤ì¤¿¥Ð¥¤¥ÈÈϰϤ¬´û¸¥í¥Ã¥¯¤ÎÈϰϤȰìÃפ¹¤ë¾ì¹ç°Ê³°¤Ç¤Ï¡¢
ÊÑ´¹¤Î²áÄø¤Ç´û¸¤Î¥í¥Ã¥¯¤Îʬ³ä¡¢½Ì¾®¡¢·ë¹ç¤¬¹Ô¤ï¤ì¤ë¤³¤È¤¬¤¢¤ë)¡£
.TP
.BR F_SETLK " (\fIstruct flock *\fP)"
.\"O Acquire a lock (when
.\"O .I l_type
.\"O is
.\"O .B F_RDLCK
.\"O or
.\"O .BR F_WRLCK )
.\"O or release a lock (when
.\"O .I l_type
.\"O is
.\"O .BR F_UNLCK )
.\"O on the bytes specified by the
.\"O .IR l_whence ", " l_start ", and " l_len
.\"O fields of
.\"O .IR lock .
.RI ( l_type
¤¬
.B F_RDLCK
¤«
.B F_WRLCK
¤Î¾ì¹ç¤Ï) ¥í¥Ã¥¯¤Î³ÍÆÀ¤ò¡¢
.RB ( F_UNLCK
¤Î¾ì¹ç¤Ï) ¥í¥Ã¥¯¤Î²òÊü¤ò¡¢
.I flock
¹½Â¤ÂΤΥե£¡¼¥ë¥É
.IR l_whence ", " l_start ", " l_len
¤Ç»ØÄꤵ¤ì¤¿ÈϰϤΥХ¤¥È¤ËÂФ·¤Æ¹Ô¤¦¡£
.\"O If a conflicting lock is held by another process,
.\"O this call returns \-1 and sets
.\"O .I errno
.\"O to
.\"O .B EACCES
.\"O or
.\"O .BR EAGAIN .
»ØÄꤵ¤ì¤¿¥í¥Ã¥¯¤¬Â¾¤Î¥×¥í¥»¥¹¤¬ÀßÄꤷ¤Æ¤¤¤ë¥í¥Ã¥¯¤È¾×Æͤ¹¤ë¾ì¹ç¤Ï¡¢
\-1 ¤òÊÖ¤·¡¢
.I errno
¤Ë
.B EACCES
¤«
.B EAGAIN
¤òÀßÄꤹ¤ë¡£
.TP
.BR F_SETLKW " (\fIstruct flock *\fP)"
.\"O As for
.\"O .BR F_SETLK ,
.\"O but if a conflicting lock is held on the file, then wait for that
.\"O lock to be released.
.B F_SETLK
¤ÈƱÍͤÀ¤¬¡¢¤³¤Á¤é¤Ç¤Ï¤½¤Î¥Õ¥¡¥¤¥ë¤ËÂФ·¤Æ¾×Æͤ¹¤ë¥í¥Ã¥¯¤¬
ŬÍѤµ¤ì¤Æ¤¤¤¿¾ì¹ç¤Ë¡¢¤½¤Î¥í¥Ã¥¯¤¬²òÊü¤µ¤ì¤ë¤Î¤òÂÔ¤ÄÅÀ¤¬°Û¤Ê¤ë¡£
.\"O If a signal is caught while waiting, then the call is interrupted
.\"O and (after the signal handler has returned)
.\"O returns immediately (with return value \-1 and
.\"O .I errno
.\"O set to
.\"O .BR EINTR ;
.\"O see
.\"O .BR signal (7)).
ÂԤäƤ¤¤ë´Ö¤Ë¥·¥°¥Ê¥ë¤ò¼õ¤±¤¿¾ì¹ç¤Ï¡¢¥·¥¹¥Æ¥à¥³¡¼¥ë¤ÏÃæÃǤµ¤ì¡¢
(¥·¥°¥Ê¥ë¥Ï¥ó¥É¥é¤¬Ìá¤Ã¤¿Ä¾¸å¤Ë) ÊÖ¤êÃÍ \-1 ¤òÊÖ¤¹ (¤Þ¤¿
.I errno
¤Ë
.B EINTR
¤¬ÀßÄꤵ¤ì¤ë;
.BR signal (7)
»²¾È)¡£
.TP
.BR F_GETLK " (\fIstruct flock *\fP)"
.\"O On input to this call,
.\"O .I lock
.\"O describes a lock we would like to place on the file.
¤³¤Î¥³¡¼¥ë¤Î¸Æ¤Ó½Ð¤·»þ¤Ë¤Ï¡¢
.I lock
¤Ë¤Ï¤½¤Î¥Õ¥¡¥¤¥ë¤ËŬÍѤ·¤è¤¦¤È¤¹¤ë¥í¥Ã¥¯¤Ë´Ø¤¹¤ë¾ðÊó¤¬Æþ¤Ã¤Æ¤¤¤ë¡£
.\"O If the lock could be placed,
.\"O .BR fcntl ()
.\"O does not actually place it, but returns
.\"O .B F_UNLCK
.\"O in the
.\"O .I l_type
.\"O field of
.\"O .I lock
.\"O and leaves the other fields of the structure unchanged.
¥í¥Ã¥¯¤òŬÍѤǤ­¤ë¾ì¹ç¤Ë¤Ï¡¢
.BR fcntl ()
¤Ï¼ÂºÝ¤Ë¤Ï¥í¥Ã¥¯¤ò¹Ô¤ï¤º¡¢¹½Â¤ÂÎ
.I lock
¤Î
.I l_type
¥Õ¥£¡¼¥ë¥É¤Ë
.B F_UNLCK
¤òÀßÄꤷ¡¢Â¾¤Î¥Õ¥£¡¼¥ë¥É¤ÏÊѹ¹¤»¤º¤Ë¡¢Éüµ¢¤¹¤ë¡£
.\"O If one or more incompatible locks would prevent
.\"O this lock being placed, then
.\"O .BR fcntl ()
.\"O returns details about one of these locks in the
.\"O .IR l_type ", " l_whence ", " l_start ", and " l_len
.\"O fields of
.\"O .I lock
.\"O and sets
.\"O .I l_pid
.\"O to be the PID of the process holding that lock.
°ã¤¦¼ïÊ̤Υí¥Ã¥¯¤¬ (°ì¤Ä¤â¤·¤¯¤ÏÊ£¿ô) ŬÍѤµ¤ì¤Æ¤¤¤Æ
¥í¥Ã¥¯¤òŬÍѤǤ­¤Ê¤¤¤è¤¦¤Ê¾ì¹ç¤Ë¤Ï¡¢
.BR fcntl ()
¤Ï¡¢¸¶°ø¤È¤Ê¤Ã¤¿¥í¥Ã¥¯¤Î°ì¤Ä¤Ë¤Ä¤¤¤Æ¤Î¾ÜºÙ¾ðÊó¤ò¹½Â¤ÂÎ
.I lock
¤Î¥Õ¥£¡¼¥ë¥É
.IR l_type ", " l_whence ", " l_start ", " l_len
¤Ë³ÊǼ¤·¡¢¤Þ¤¿
.I l_pid
¤Ë¥í¥Ã¥¯¤òÊÝ»ý¤·¤Æ¤¤¤ë¥×¥í¥»¥¹¤Î PID ¤òÀßÄꤷ¤Æ¡¢Éüµ¢¤¹¤ë¡£
.P
.\"O In order to place a read lock,
.\"O .I fd
.\"O must be open for reading.
.\"O In order to place a write lock,
.\"O .I fd
.\"O must be open for writing.
.\"O To place both types of lock, open a file read-write.
Æɤ߽Ф·¥í¥Ã¥¯¤òŬÍѤ¹¤ë¤Ë¤Ï¡¢
.I fd
¤ÏÆɤ߽Ф·ÍѤ˥ª¡¼¥×¥ó¤µ¤ì¤Æ¤¤¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£
½ñ¤­¹þ¤ß¥í¥Ã¥¯¤òŬÍѤ¹¤ë¤Ë¤Ï¡¢
.I fd
¤Ï½ñ¤­¹þ¤ßÍѤ˥ª¡¼¥×¥ó¤µ¤ì¤Æ¤¤¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£
Æɤ߽ñ¤­Î¾Êý¤Î¥í¥Ã¥¯¤òŬÍѤ¹¤ë¤Ë¤Ï¡¢Æɤ߽ñ¤­Î¾ÍѤÇ
¥Õ¥¡¥¤¥ë¤ò¥ª¡¼¥×¥ó¤·¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£
.P
.\"O As well as being removed by an explicit
.\"O .BR F_UNLCK ,
.\"O record locks are automatically released when the process
.\"O terminates or if it closes
.\"O .I any
.\"O file descriptor referring to a file on which locks are held.
¥ì¥³¡¼¥É¤Î¥í¥Ã¥¯¤Ï¡¢
.B F_UNLCK
¤Ë¤è¤êÌÀ¼¨Åª¤Ëºï½ü¤µ¤ì¤ë¤À¤±¤Ç¤Ê¤¯¡¢
¥×¥í¥»¥¹¤¬½ªÎ»¤·¤¿¤È¤­¤ä¡¢¥í¥Ã¥¯¤¬Å¬ÍѤµ¤ì¤Æ¤¤¤ë¥Õ¥¡¥¤¥ë¤ò»²¾È¤·¤Æ¤¤¤ë
¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Î¤¤¤º¤ì¤«¤¬¥¯¥í¡¼¥º¤µ¤ì¤¿¾ì¹ç¤Ë¤â²òÊü¤µ¤ì¤ë¡£
¤³¤Î¥í¥Ã¥¯¤Î²òÊü¤Ï¼«Æ°Åª¤Ë¹Ô¤ï¤ì¤ë¡£
.\"O .\" (Additional file descriptors referring to the same file
.\"O .\" may have been obtained by calls to
.\"O .\" .BR open "(2), " dup "(2), " dup2 "(2), or " fcntl ().)
.\" .RB ( open "(2), " dup "(2), " dup2 "(2), " fcntl ()
.\" ¤Î¸Æ¤Ó½Ð¤·¤Ë¤è¤Ã¤ÆƱ¤¸¥Õ¥¡¥¤¥ë¤ò»²¾È¤¹¤ë¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤¬
.\" ¾¤Ë¤â¤Ç¤­¤Æ¤¤¤ë¤«¤â¤·¤ì¤Ê¤¤)
.\"O This is bad: it means that a process can lose the locks on
.\"O a file like
.\"O .I /etc/passwd
.\"O or
.\"O .I /etc/mtab
.\"O when for some reason a library function decides to open, read
.\"O and close it.
¤³¤ÎÆ°ºî¤Ï¤Þ¤º¤¤: ¤¢¤ë¥×¥í¥»¥¹¤¬
.I /etc/passwd
¤ä
.I /etc/mtab
¤È¤¤¤Ã¤¿¥Õ¥¡¥¤¥ë¤Ë¥í¥Ã¥¯¤òŬÍѤ·¤Æ¤¤¤ë¤È¤­¤Ë¡¢
¤¢¤ë¥é¥¤¥Ö¥é¥ê´Ø¿ô¤¬²¿¤«¤ÎÍýͳ¤ÇƱ¤¸¥Õ¥¡¥¤¥ë¤ò open, read, close
¤¹¤ë¤È¡¢¤½¤Î¥Õ¥¡¥¤¥ë¤Ø¤Î¥í¥Ã¥¯¤¬¼º¤ï¤ì¤ë¤³¤È¤Ë¤Ê¤ë¡£
.P
.\"O Record locks are not inherited by a child created via
.\"O .BR fork (2),
.\"O but are preserved across an
.\"O .BR execve (2).
¥ì¥³¡¼¥É¤Î¥í¥Ã¥¯¤Ï
.BR fork (2)
¤ÇºîÀ®¤µ¤ì¤¿»Ò¥×¥í¥»¥¹¤Ë¤Ï·Ñ¾µ¤µ¤ì¤Ê¤¤¤¬¡¢
.BR execve (2)
¤ÎÁ°¸å¤Ç¤ÏÊݸ¤µ¤ì¤ë¡£
.P
.\"O Because of the buffering performed by the
.\"O .BR stdio (3)
.\"O library, the use of record locking with routines in that package
.\"O should be avoided; use
.\"O .BR read (2)
.\"O and
.\"O .BR write (2)
.\"O instead.
.BR stdio (3)
¤Ç¤Ï¥Ð¥Ã¥Õ¥¡¥ê¥ó¥°¤¬¹Ô¤ï¤ì¤ë¤Î¤Ç¡¢
stdio ´ØÏ¢¤Î´Ø¿ô¤Ç¤Ï¥ì¥³¡¼¥É¤Î¥í¥Ã¥¯¤Î»ÈÍѤϲóÈò¤µ¤ì¤ë;
Âå¤ï¤ê¤Ë
.BR read (2)
¤ä
.BR write (2)
¤ò»ÈÍѤ¹¤ë¤³¤È¡£
.\"O .SS "Mandatory locking"
.SS "¶¯À©¥í¥Ã¥¯ (mandatory locking)"
.\"O (Non-POSIX.)
.\"O The above record locks may be either advisory or mandatory,
.\"O and are advisory by default.
¾å½Ò¤Î¥í¥Ã¥¯¤Ë¤Ï¥¢¥É¥Ð¥¤¥¶¥ê¡¦¥í¥Ã¥¯ (advisory lock) ¤È¶¯À©¥í¥Ã¥¯ (mandatory
lock) ¤ÎÆó¼ïÎब¤¢¤ë¤¬¡¢¥Ç¥Õ¥©¥ë¥È¤Ç¤Ï¥¢¥É¥Ð¥¤¥¶¥ê¡¦¥í¥Ã¥¯¤È¤Ê¤ë¡£

.\"O Advisory locks are not enforced and are useful only between
.\"O cooperating processes.
¥¢¥É¥Ð¥¤¥¶¥ê¡¦¥í¥Ã¥¯¤Ë¶¯À©ÎϤϤʤ¯¡¢¶¨Ä´¤·¤ÆÆ°ºî¤¹¤ë¥×¥í¥»¥¹´Ö¤Ç¤Î¤ß
Í­¸ú¤Ç¤¢¤ë¡£

.\"O Mandatory locks are enforced for all processes.
¶¯À©¥í¥Ã¥¯¤ÏÁ´¤Æ¤Î¥×¥í¥»¥¹¤ËÂФ·¤Æ¸ú²Ì¤¬¤¢¤ë¡£
.\"O If a process tries to perform an incompatible access (e.g.,
.\"O .BR read (2)
.\"O or
.\"O .BR write (2))
.\"O on a file region that has an incompatible mandatory lock,
.\"O then the result depends upon whether the
.\"O .B O_NONBLOCK
.\"O flag is enabled for its open file description.
.\"O If the
.\"O .B O_NONBLOCK
.\"O flag is not enabled, then
.\"O system call is blocked until the lock is removed
.\"O or converted to a mode that is compatible with the access.
.\"O If the
.\"O .B O_NONBLOCK
.\"O flag is enabled, then the system call fails with the error
.\"O .BR EAGAIN .
¤¢¤ë¥×¥í¥»¥¹¤¬¸ß´¹À­¤Î¤Ê¤¤¶¯À©¥í¥Ã¥¯¤¬Å¬ÍѤµ¤ì¤¿¥Õ¥¡¥¤¥ëÎΰè¤ËÂФ·¤Æ
.RB ( read (2)
¤ä
.BR write (2)
¤Ë¤è¤ê) ¸ß´¹À­¤Î¤Ê¤¤¥¢¥¯¥»¥¹¤ò¼Â¹Ô¤·¤è¤¦¤È¤·¤¿¾ì¹ç¡¢
¥¢¥¯¥»¥¹¤Î·ë²Ì¤Ï
¤½¤Î¥Õ¥¡¥¤¥ë¤Î¥ª¡¼¥×¥ó¥Õ¥¡¥¤¥ëµ­½Ò¤Ç
.B O_NONBLOCK
¥Õ¥é¥°¤¬Í­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤ë¤«¤Ë¤è¤ê·è¤Þ¤ë¡£
.B O_NONBLOCK
¥Õ¥é¥°¤¬Í­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤Ê¤¤¤È¤­¤Ï¡¢¥í¥Ã¥¯¤¬ºï½ü¤µ¤ì¤ë¤«¡¢
¥í¥Ã¥¯¤¬¥¢¥¯¥»¥¹¤È¸ß´¹À­¤Î¤¢¤ë¥â¡¼¥É¤ËÊÑ´¹¤µ¤ì¤ë¤Þ¤Ç¡¢
¥·¥¹¥Æ¥à¥³¡¼¥ë¤ÏÄä»ß (block) ¤µ¤ì¤ë¡£
.B O_NONBLOCK
¥Õ¥é¥°¤¬Í­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤ë¤È¤­¤Ï¡¢¥·¥¹¥Æ¥à¥³¡¼¥ë¤Ï¥¨¥é¡¼
.B EAGAIN
¤Ç¼ºÇÔ¤¹¤ë¡£

.\"O To make use of mandatory locks, mandatory locking must be enabled
.\"O both on the file system that contains the file to be locked,
.\"O and on the file itself.
.\"O Mandatory locking is enabled on a file system
.\"O using the "\-o mand" option to
.\"O .BR mount (8),
.\"O or the
.\"O .B MS_MANDLOCK
.\"O flag for
.\"O .BR mount (2).
.\"O Mandatory locking is enabled on a file by disabling
.\"O group execute permission on the file and enabling the set-group-ID
.\"O permission bit (see
.\"O .BR chmod (1)
.\"O and
.\"O .BR chmod (2)).
¶¯À©¥í¥Ã¥¯¤ò»ÈÍѤ¹¤ë¤¿¤á¤Ë¤Ï¡¢¥í¥Ã¥¯ÂоݤΥե¡¥¤¥ë¤¬´Þ¤Þ¤ì¤ë¥Õ¥¡¥¤¥ë¥·¥¹¥Æ¥à
¤È¡¢¥í¥Ã¥¯ÂоݤΥե¡¥¤¥ë¼«¿È¤ÎξÊý¤Ë¤Ä¤¤¤Æ¡¢¶¯À©¥í¥Ã¥¯¤¬Í­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤Ê¤±¤ì
¤Ð¤Ê¤é¤Ê¤¤¡£¥Õ¥¡¥¤¥ë¥·¥¹¥Æ¥à¤Ë¤Ä¤¤¤Æ¶¯À©¥í¥Ã¥¯¤òÍ­¸ú¤Ë¤¹¤ë¤Ë¤Ï¡¢
.BR mount (8)
¤Ë "\-o mand" ¥ª¥×¥·¥ç¥ó¤òÅϤ¹¤«¡¢
.BR mount (2)
¤Ë
.B MS_MANDLOCK
¥Õ¥é¥°¤ò»ØÄꤹ¤ë¡£¥Õ¥¡¥¤¥ë¤Ë¤Ä¤¤¤Æ¶¯À©¥í¥Ã¥¯¤òÍ­¸ú¤Ë¤¹¤ë¤Ë¤Ï¡¢
¤½¤Î¥Õ¥¡¥¤¥ë¤Î¥°¥ë¡¼¥×¼Â¹Ôµö²Ä (group execute permission) ¤ò̵¸ú¤È¤·¡¢
¤«¤Ä set-group-ID µö²Ä¥Ó¥Ã¥È¤òÍ­¸ú¤Ë¤¹¤ë
.RB ( chmod (1)
¤È
.BR chmod (2)
¤ò»²¾È)¡£

.\"O The Linux implementation of mandatory locking is unreliable.
.\"O See BUGS below.
Linux ¤Î¶¯À©¥í¥Ã¥¯¤Î¼ÂÁõ¤Ï¿®ÍêÀ­¤Ë·ç¤±¤ë¤â¤Î¤Ç¤¢¤ë¡£
²¼µ­¤Î¡Ö¥Ð¥°¡×¤ÎÀá¤ò»²¾È¤Î¤³¤È¡£
.\"O .SS "Managing signals"
.SS "¥·¥°¥Ê¥ë¤Î´ÉÍý"
.\"O .BR F_GETOWN ,
.\"O .BR F_SETOWN ,
.\"O .BR F_GETOWN_EX ,
.\"O .BR F_SETOWN_EX ,
.\"O .BR F_GETSIG
.\"O and
.\"O .B F_SETSIG
.\"O are used to manage I/O availability signals:
.BR F_GETOWN ,
.BR F_SETOWN ,
.BR F_GETOWN_EX ,
.BR F_SETOWN_EX ,
.BR F_GETSIG ,
.B F_SETSIG
¤Ï¡¢I/O ¤¬ÍøÍѲÄǽ¤Ë¤Ê¤Ã¤¿¤³¤È¤ò¼¨¤¹¥·¥°¥Ê¥ë¤ò´ÉÍý¤¹¤ë¤¿¤á¤Ë»ÈÍѤµ¤ì¤ë¡£
.TP
.BR F_GETOWN " (\fIvoid\fP)"
.\"O Return (as the function result)
.\"O the process ID or process group currently receiving
.\"O .B SIGIO
.\"O and
.\"O .B SIGURG
.\"O signals for events on file descriptor
.\"O .IR fd .
.\"O Process IDs are returned as positive values;
.\"O process group IDs are returned as negative values (but see BUGS below).
.\"O .I arg
.\"O is ignored.
¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿
.I fd
¤Î¥¤¥Ù¥ó¥È¤ËÂФ¹¤ë¥·¥°¥Ê¥ë
.B SIGIO
¤ª¤è¤Ó
.B SIGURG
¤ò¼õ¤±¤Æ¤¤¤ë¥×¥í¥»¥¹¤Î¥×¥í¥»¥¹ID ¤«¥×¥í¥»¥¹¥°¥ë¡¼¥×¤ò
(´Ø¿ô¤Î·ë²Ì¤È¤·¤Æ) ÊÖ¤¹¡£
¥×¥í¥»¥¹ID ¤ÏÀµ¤ÎÃͤȤ·¤ÆÊÖ¤µ¤ì¤ë¡£
¥×¥í¥»¥¹¥°¥ë¡¼¥×ID ¤ÏÉé¤ÎÃͤȤ·¤ÆÊÖ¤µ¤ì¤ë (²¼µ­¤Î¥Ð¥°¤Î¾Ï¤ò»²¾È)¡£
.I arg
¤Ï̵»ë¤µ¤ì¤ë¡£
.TP
.BR F_SETOWN " (\fIlong\fP)"
.\"O Set the process ID or process group ID that will receive
.\"O .B SIGIO
.\"O and
.\"O .B SIGURG
.\"O signals for events on file descriptor
.\"O .IR fd
.\"O to the ID given in
.\"O .IR arg .
.\"O A process ID is specified as a positive value;
.\"O a process group ID is specified as a negative value.
¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿
.I fd
¤Î¥¤¥Ù¥ó¥ÈȯÀ¸¤òÃΤ餻¤ë¥·¥°¥Ê¥ë
.B SIGIO
¤ä
.B SIGURG
¤ò¼õ¤±¤ë¥×¥í¥»¥¹¤Î
¥×¥í¥»¥¹ ID ¤Þ¤¿¤Ï¥×¥í¥»¥¹¥°¥ë¡¼¥×ID ¤ò
.I arg
¤Ç»ØÄꤵ¤ì¤¿ ID ¤ËÀßÄꤹ¤ë¡£
¥×¥í¥»¥¹ID ¤ÏÀµ¤ÎÃͤȤ·¤Æ»ØÄꤷ¡¢
¥×¥í¥»¥¹¥°¥ë¡¼¥×ID ¤ÏÉé¤ÎÃͤȤ·¤Æ»ØÄꤹ¤ë¡£
.\"O Most commonly, the calling process specifies itself as the owner
.\"O (that is,
.\"O .I arg
.\"O is specified as
.\"O .BR getpid (2)).
¤Û¤È¤ó¤É¤Î¾ì¹ç¡¢¸Æ¤Ó½Ð¤·¸µ¥×¥í¥»¥¹¤Ï½êÍ­¼Ô¤È¤·¤Æ¼«Ê¬¼«¿È¤ò»ØÄꤹ¤ë
(¤Ä¤Þ¤ê
.I arg
¤Ë
.BR getpid (2)
¤ò»ØÄꤹ¤ë)¡£

.\"O .\" From glibc.info:
.\" glibc.info ¤è¤ê:
.\"O If you set the
.\"O .B O_ASYNC
.\"O status flag on a file descriptor by using the
.\"O .B F_SETFL
.\"O command of
.\"O .BR fcntl (),
.\"O a
.\"O .B SIGIO
.\"O signal is sent whenever input or output becomes possible
.\"O on that file descriptor.
.\"O .B F_SETSIG
.\"O can be used to obtain delivery of a signal other than
.\"O .BR SIGIO .
.BR fcntl ()
¤Î
.B F_SETFL
¥³¥Þ¥ó¥É¤ò»ÈÍѤ·¤Æ¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Ë
.B O_ASYNC
¾õÂ֥ե饰¤òÀßÄꤷ¤¿¾ì¹ç¤Ë¤Ï¡¢¤½¤Î¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Ø¤Î
Æþ½ÐÎϤ¬²Äǽ¤Ë¤Ê¤ëÅÙ¤Ë
.B SIGIO
¥·¥°¥Ê¥ë¤¬Á÷¤é¤ì¤ë¡£
.B F_SETSIG
¤Ï
.B SIGIO
°Ê³°¤ÎÊ̤Υ·¥°¥Ê¥ë¤ÎÇÛÁ÷¤ò¼õ¤±¤é¤ì¤ë¤è¤¦¤Ë
¤¹¤ë¤Î¤Ë¤â»È¤¦¤³¤È¤¬¤Ç¤­¤ë¡£
.\"O If this permission check fails, then the signal is
.\"O silently discarded.
µö²Ä (permission) ¤Î¥Á¥§¥Ã¥¯¤Ç¼ºÇÔ¤·¤¿¾ì¹ç¤Ë¤Ï¡¢
¥·¥°¥Ê¥ë¤ÏÌۤäƼΤƤé¤ì¤ë¡£

.\"O Sending a signal to the owner process (group) specified by
.\"O .B F_SETOWN
.\"O is subject to the same permissions checks as are described for
.\"O .BR kill (2),
.\"O where the sending process is the one that employs
.\"O .B F_SETOWN
.\"O (but see BUGS below).
.B F_SETOWN
¤Ë¤è¤ê»ØÄꤵ¤ì¤¿½êÍ­¼Ô¤Î¥×¥í¥»¥¹ (¤Þ¤¿¤Ï¥×¥í¥»¥¹¥°¥ë¡¼¥×) ¤Ë
¥·¥°¥Ê¥ë¤òÁ÷¤ëºÝ¤Ë¤Ï¡¢
.BR kill (2)
¤Ë½ñ¤«¤ì¤Æ¤¤¤ë¤Î¤ÈƱ¤¸µö²Ä¤Î¥Á¥§¥Ã¥¯¤¬¹Ô¤ï¤ì¤ë¡£
¤³¤Î¤È¤­¡¢¥·¥°¥Ê¥ë¤òÁ÷¿®¤¹¤ë¥×¥í¥»¥¹¤Ï
.B F_SETOWN
¤ò»È¤Ã¤¿¥×¥í¥»¥¹¤Ç¤¢¤ë
(⤷¡¢²¼µ­¤Î¡Ö¥Ð¥°¡×¤Î¾Ï¤ò»²¾È¤Î¤³¤È)¡£

.\"O If the file descriptor
.\"O .I fd
.\"O refers to a socket,
.\"O .B F_SETOWN
.\"O also selects
.\"O the recipient of
.\"O .B SIGURG
.\"O signals that are delivered when out-of-band
.\"O data arrives on that socket.
.\"O .RB ( SIGURG
.\"O is sent in any situation where
.\"O .BR select (2)
.\"O would report the socket as having an "exceptional condition".)
¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤¬¥½¥±¥Ã¥È¤ò»²¾È¤·¤Æ¤¤¤ë¾ì¹ç¤Ï¡¢
.B F_SETOWN
¤ò»ÈÍѤ·¤Æ¡¢¥½¥±¥Ã¥È¤ËÂÓ°è³° (out-of-band) ¥Ç¡¼¥¿¤¬ÆϤ¤¤¿»þ¤Ë
.B SIGURG
¥·¥°¥Ê¥ë¤òÇÛÁ÷¤¹¤ëÁê¼ê¤òÁªÂò¤¹¤ë¤³¤È¤â¤Ç¤­¤ë
.RB ( SIGURG
¤¬Á÷¤é¤ì¤¿¾ì¹ç¤Ë¤Ï
.BR select (2)
¤¬¥½¥±¥Ã¥È¤¬¡ÖÆÃÊ̤ʾõÂ֡פˤ¢¤ë¤ÈÊó¹ð¤¹¤ë¤³¤È¤À¤í¤¦)¡£
.\"O .\" The following appears to be rubbish.  It doesn't seem to
.\"O .\" be true according to the kernel source, and I can write
.\"O .\" a program that gets a terminal-generated SIGIO even though
.\"O .\" it is not the foreground process group of the terminal.
.\"O .\" -- MTK, 8 Apr 05
.\"O .\"
.\"O .\" If the file descriptor
.\"O .\" .I fd
.\"O .\" refers to a terminal device, then SIGIO
.\"O .\" signals are sent to the foreground process group of the terminal.
.\" °Ê²¼¤Îµ­½Ò¤Ï¥´¥ß¤À¤í¤¦¡£
.\" ¥«¡¼¥Í¥ë¥½¡¼¥¹¤ò¸«¤ë¤È¤³¤Îµ­½Ò¤Ï´Ö°ã¤Ã¤Æ¤¤¤ë¤è¤¦¤À¤·¡¢
.\" ¥×¥í¥°¥é¥à¤ò½ñ¤¤¤Æ¤ß¤¿¤È¤³¤í¡¢Ã¼Ëö¤Î¥Õ¥©¥¢¥°¥é¥ó¥É¡¦¥×¥í¥»¥¹¥°¥ë¡¼¥×¤Ç
.\" ¤Ê¤¯¤Æ¤â¡¢Ã¼Ëö¤¬À¸À®¤·¤¿ SIGIO ¥·¥°¥Ê¥ë¤ò¼õ¤±¤È¤ë¡£
.\" -- MTK, 8 Apr 05
.\"
.\" ¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤¬Ã¼Ëö¥Ç¥Ð¥¤¥¹¤ò»²¾È¤·¤Æ¤¤¤ë¾ì¹ç¤Ë¤Ï¡¢
.\" SIGIO ¥·¥°¥Ê¥ë¤Ï¤½¤ÎüËö¤Î¥Õ¥©¥¢¥°¥é¥ó¥É¡¦¥×¥í¥»¥¹¥°¥ë¡¼¥×¤Ø¤ÈÁ÷¤é¤ì¤ë¡£

.\"O The following was true in 2.6.x kernels up to and including
.\"O kernel 2.6.11:
¥Ð¡¼¥¸¥ç¥ó 2.6.11 °ÊÁ°¤Î 2.6.x ¥«¡¼¥Í¥ë¤Ç¤Ï¡¢°Ê²¼¤Ë¼¨¤¹Æ°ºî¤Ç¤¢¤Ã¤¿¡£
.RS
.IP
.\"O If a nonzero value is given to
.\"O .B F_SETSIG
.\"O in a multithreaded process running with a threading library
.\"O that supports thread groups (e.g., NPTL),
.\"O then a positive value given to
.\"O .B F_SETOWN
.\"O has a different meaning:
.\"O .\" The relevant place in the (2.6) kernel source is the
.\"O .\" 'switch' in fs/fcntl.c::send_sigio_to_task() -- MTK, Apr 2005
.\"O instead of being a process ID identifying a whole process,
.\"O it is a thread ID identifying a specific thread within a process.
¥¹¥ì¥Ã¥É¥°¥ë¡¼¥×¤ò¥µ¥Ý¡¼¥È¤·¤Æ¤¤¤ë¥¹¥ì¥Ã¥É¥é¥¤¥Ö¥é¥ê (Î㤨¤Ð NPTL) ¤ò
»È¤Ã¤ÆÆ°ºî¤·¤Æ¤¤¤ë¥Þ¥ë¥Á¥¹¥ì¥Ã¥É¡¦¥×¥í¥»¥¹¤Ç
.B F_SETSIG
¤Ë 0 °Ê³°¤ÎÃͤò»ØÄꤷ¤¿¾ì¹ç¡¢
.B F_SETOWN
¤ËÀµ¤ÎÃͤòÅϤ¹¤È¡¢¤½¤Î°ÕÌ£¤¬°ã¤Ã¤Æ¤¯¤ë:
.\" The relevant place in the (2.6) kernel source is the
.\" 'switch' in fs/fcntl.c::send_sigio_to_task() -- MTK, Apr 2005
¥×¥í¥»¥¹Á´ÂΤò¼¨¤¹¥×¥í¥»¥¹ID ¤Ç¤Ï¤Ê¤¯¡¢¥×¥í¥»¥¹Æâ¤ÎÆÃÄê¤Î
¥¹¥ì¥Ã¥É¤ò¼¨¤¹¥¹¥ì¥Ã¥ÉID ¤È²ò¼á¤µ¤ì¤ë¡£
.\"O Consequently, it may be necessary to pass
.\"O .B F_SETOWN
.\"O the result of
.\"O .BR gettid (2)
.\"O instead of
.\"O .BR getpid (2)
.\"O to get sensible results when
.\"O .B F_SETSIG
.\"O is used.
¤·¤¿¤¬¤Ã¤Æ¡¢
.B F_SETSIG
¤ò»È¤¦¾ì¹ç¤Ë¤Ï¡¢¤­¤Á¤ó¤È·ë²Ì¤ò¼õ¤±¼è¤ë¤Ë¤Ï¡¢
.B F_SETOWN
¤ËÅϤ¹Ãͤò
.BR getpid (2)
¤Ç¤Ï¤Ê¤¯
.BR gettid (2)
¤ÎÊÖ¤êÃͤˤ¹¤ëɬÍפ¬¤¢¤ë¤À¤í¤¦¡£
.\"O (In current Linux threading implementations,
.\"O a main thread's thread ID is the same as its process ID.
.\"O This means that a single-threaded program can equally use
.\"O .BR gettid (2)
.\"O or
.\"O .BR getpid (2)
.\"O in this scenario.)
(¸½¾õ¤Î Linux ¥¹¥ì¥Ã¥É¼ÂÁõ¤Ç¤Ï¡¢¥á¥¤¥ó¡¦¥¹¥ì¥Ã¥É¤Î¥¹¥ì¥Ã¥ÉID ¤Ï
¤½¤Î¥¹¥ì¥Ã¥É¤Î¥×¥í¥»¥¹ID ¤ÈƱ¤¸¤Ç¤¢¤ë¡£¤Ä¤Þ¤ê¡¢
¥·¥°¥Ê¥ë¡¦¥¹¥ì¥Ã¥É¤Î¥×¥í¥°¥é¥à¤Ç¤Ï¤³¤Î¾ì¹ç
.BR gettid (2)
¤È
.BR getpid (2)
¤ÏÁ´¤¯Æ±¤¸¤è¤¦¤Ë»È¤¦¤³¤È¤¬¤Ç¤­¤ë¡£)
.\"O Note, however, that the statements in this paragraph do not apply
.\"O to the
.\"O .B SIGURG
.\"O signal generated for out-of-band data on a socket:
.\"O this signal is always sent to either a process or a process group,
.\"O depending on the value given to
.\"O .BR F_SETOWN .
.\"O .\" send_sigurg()/send_sigurg_to_task() bypasses
.\"O .\" kill_fasync()/send_sigio()/send_sigio_to_task()
.\"O .\" to directly call send_group_sig_info()
.\"O .\"        -- MTK, Apr 2005 (kernel 2.6.11)
¤¿¤À¤·¡¢Ãí°Õ¤¹¤Ù¤­ÅÀ¤È¤·¤Æ¡¢¤³¤ÎÃÊÍî¤Ç½Ò¤Ù¤¿¤³¤È¤Ï¡¢
¥½¥±¥Ã¥È¤ÎÂÓ°è³°¥Ç¡¼¥¿¤¬ÆϤ¤¤¿¤È¤­¤ËÀ¸À®¤µ¤ì¤ë
.B SIGURG
¥·¥°¥Ê¥ë¤Ë¤Ï¤¢¤Æ¤Ï¤Þ¤é¤Ê¤¤¡£
¤³¤Î¥·¥°¥Ê¥ë¤Ï¾ï¤Ë¥×¥í¥»¥¹¤«¥×¥í¥»¥¹¥°¥ë¡¼¥×¤ËÁ÷¤é¤ì¡¢
Á÷¿®Àè¤Ï
.B F_SETOWN
¤ËÅϤµ¤ì¤¿Ãͤˤ·¤¿¤¬¤Ã¤Æ·è¤á¤é¤ì¤ë¡£
.\" send_sigurg()/send_sigurg_to_task() bypasses
.\" kill_fasync()/send_sigio()/send_sigio_to_task()
.\" to directly call send_group_sig_info()
.\"     -- MTK, Apr 2005 (kernel 2.6.11)
.RE
.IP
.\"O The above behavior was accidentally dropped in Linux 2.6.12,
.\"O and won't be restored.
.\"O From Linux 2.6.32 onwards, use
.\"O .BR F_SETOWN_EX
.\"O to target
.\"O .B SIGIO
.\"O and
.\"O .B SIGURG
.\"O signals at a particular thread.
¾åµ­¤ÎÆ°ºî¤Ï¡¢Linux 2.6.12 ¤Ç¿Þ¤é¤º¤âºï½ü¤µ¤ì¡¢
¸µ¤ËÌᤵ¤ì¤Ê¤¤Í½Äê¤Ç¤¢¤ë¡£
Linux 2.6.32 °Ê¹ß¤Ç¡¢ÆÃÄê¤Î¥¹¥ì¥Ã¥É°¸¤Ë¥·¥°¥Ê¥ë
.B SIGIO
¤È
.B SIGURG
¤òÁ÷¤ë¤Ë¤Ï
.B F_SETOWN_EX
¤ò»È¤¦¤³¤È¡£
.TP
.\"O .BR F_GETOWN_EX " (struct f_owner_ex *) (since Linux 2.6.32)"
.BR F_GETOWN_EX " (struct f_owner_ex *) (Linux 2.6.32 °Ê¹ß)"
.\"O Return the current file descriptor owner settings
.\"O as defined by a previous
.\"O .BR F_SETOWN_EX
.\"O operation.
.\"O The information is returned in the structure pointed to by
.\"O .IR arg ,
.\"O which has the following form:
ľÁ°¤Î
.B F_SETOWN_EX
Áàºî¤ÇÄêµÁ¤µ¤ì¤¿¸½ºß¤Î¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Î½êÍ­¼ÔÀßÄê
¤òÊÖ¤¹¡£¾ðÊó¤Ï
.I arg
¤¬»Ø¤¹¹½Â¤ÂΤ˳ÊǼ¤µ¤ì¤ÆÊÖ¤µ¤ì¤ë¡£¹½Â¤ÂΤϰʲ¼¤ÎÄ̤ê¤Ç¤¢¤ë¡£
.nf
.in +4n

struct f_owner_ex {
    int   type;
    pid_t pid;
};

.in
.fi
.\"O The
.\"O .I type
.\"O field will have one of the values
.\"O .BR F_OWNER_TID ,
.\"O .BR F_OWNER_PID ,
.\"O or
.\"O .BR F_OWNER_PGRP .
.I type
¥Õ¥£¡¼¥ë¥É¤Ï¡¢
.B F_OWNER_TID ,
.B F_OWNER_PID ,
.B F_OWNER_PGRP
¤Î¤¤¤º¤ì¤«°ì¤Ä¤ÎÃͤȤʤ롣
.\"O The
.\"O .I pid
.\"O field is a positive integer representing a thread ID, process ID,
.\"O or process group ID.
.\"O See
.\"O .B F_SETOWN_EX
.\"O for more details.
.I pid
¥Õ¥£¡¼¥ë¥É¤Ï¡¢¥¹¥ì¥Ã¥É ID¡¢¥×¥í¥»¥¹ ID¡¢¥×¥í¥»¥¹¥°¥ë¡¼¥× ID ¤ò
ɽ¤¹Àµ¤ÎÀ°¿ô¤Ç¤¢¤ë¡£¾ÜºÙ¤Ï
.B F_SETOWN_EX
¤ò»²¾È¡£
.TP
.\"O .BR F_SETOWN_EX " (struct f_owner_ex *) (since Linux 2.6.32)"
.BR F_SETOWN_EX " (struct f_owner_ex *) (Linux 2.6.32 °Ê¹ß)"
.\"O This operation performs a similar task to
.\"O .BR F_SETOWN .
.\"O It allows the caller to direct I/O availability signals
.\"O to a specific thread, process, or process group.
¤³¤ÎÁàºî¤Ï
.B F_SETOWN
¤ÈƱÍͤνèÍý¤ò¹Ô¤¦¡£
¤³¤ÎÁàºî¤ò»È¤¦¤È¡¢I/O ¤¬ÍøÍѲÄǽ¤Ë¤Ê¤Ã¤¿¤³¤È¤ò¼¨¤¹¥·¥°¥Ê¥ë¤ò¡¢
ÆÃÄê¤Î¥¹¥ì¥Ã¥É¡¢¥×¥í¥»¥¹¡¢¥×¥í¥»¥¹¥°¥ë¡¼¥×¤ËÁ÷¤ë¤³¤È¤¬¤Ç¤­¤ë
¤è¤¦¤Ë¤Ê¤ë¡£
.\"O The caller specifies the target of signals via
.\"O .IR arg ,
.\"O which is a pointer to a
.\"O .IR f_owner_ex
.\"O structure.
.\"O The
.\"O .I type
.\"O field has one of the following values, which define how
.\"O .I pid
.\"O is interpreted:
¸Æ¤Ó½Ð¤·¸µ¤Ï¡¢
.I arg
·Ðͳ¤Ç¥·¥°¥Ê¥ë¤ÎÇÛÁ÷Àè¤ò»ØÄꤹ¤ë¡£
.I arg
¤Ï
.I f_owner_ex
¹½Â¤ÂΤؤΥݥ¤¥ó¥¿¤Ç¤¢¤ë¡£
.I type
¥Õ¥£¡¼¥ë¥É¤Ï°Ê²¼¤Î¤¤¤º¤ì¤«¤ÎÃͤò¼è¤ê¡¢
¤³¤ÎÃͤˤè¤ê
.I pid
¤¬¤É¤Î¤è¤¦¤Ë²ò¼á¤µ¤ì¤ë¤«¤¬µ¬Äꤵ¤ì¤ë¡£
.RS
.TP
.BR F_OWNER_TID
.\"O Send the signal to the thread whose thread ID
.\"O (the value returned by a call to
.\"O .BR clone (2)
.\"O or
.\"O .BR gettid (2))
.\"O is specified in
.\"O .IR pid .
¥¹¥ì¥Ã¥É ID ¤¬
.I pid
¤Ç»ØÄꤵ¤ì¤¿ÃͤΥ¹¥ì¥Ã¥É¤Ë¤½¤Î¥·¥°¥Ê¥ë¤òÁ÷¤ë
(¥¹¥ì¥Ã¥É ID ¤Ï
.BR clone (2)
¤ä
.BR gettid (2)
¤Î¸Æ¤Ó½Ð¤·¤ÇÊÖ¤µ¤ì¤ëÃͤǤ¢¤ë)¡£
.TP
.BR F_OWNER_PID
.\"O Send the signal to the process whose ID
.\"O is specified in
.\"O .IR pid .
ID ¤¬
.I pid
¤Ç»ØÄꤵ¤ì¤¿ÃÍ¤Î¥×¥í¥»¥¹¤Ë¤½¤Î¥·¥°¥Ê¥ë¤òÁ÷¤ë¡£
.TP
.BR F_OWNER_PGRP
.\"O Send the signal to the process group whose ID
.\"O is specified in
.\"O .IR pid .
.\"O (Note that, unlike with
.\"O .BR F_SETOWN ,
.\"O a process group ID is specified as a positive value here.)
ID ¤¬
.I pid
¤Ç»ØÄꤵ¤ì¤¿ÃÍ¤Î¥×¥í¥»¥¹¥°¥ë¡¼¥×¤Ë¤½¤Î¥·¥°¥Ê¥ë¤òÁ÷¤ë¡£
.RB ( F_SETOWN
¤È°Û¤Ê¤ê¡¢¥×¥í¥»¥¹¥°¥ë¡¼¥× ID ¤Ë¤ÏÀµ¤ÎÃͤò»ØÄꤹ¤ëÅÀ¤ËÃí°Õ¤¹¤ë¤³¤È¡£)
.RE
.TP
.BR F_GETSIG " (\fIvoid\fP)"
.\"O Return (as the function result)
.\"O the signal sent when input or output becomes possible.
.\"O A value of zero means
.\"O .B SIGIO
.\"O is sent.
.\"O Any other value (including
.\"O .BR SIGIO )
.\"O is the
.\"O signal sent instead, and in this case additional info is available to
.\"O the signal handler if installed with
.\"O .BR SA_SIGINFO .
.\"O .I arg
.\"O is ignored.
ÆþÎϤä½ÐÎϤ¬²Äǽ¤Ë¤Ê¤Ã¤¿¾ì¹ç¤ËÁ÷¤ë¥·¥°¥Ê¥ë¤ò
(´Ø¿ô¤Î·ë²Ì¤È¤·¤Æ) ÊÖ¤¹¡£
ÃÍ¥¼¥í¤Ï
.B SIGIO
¤òÁ÷¤ë¤³¤È¤ò°ÕÌ£¤¹¤ë¡£
.RB ( SIGIO
¤ò´Þ¤à) ¾¤ÎÃͤϤ¤¤º¤ì¤â¡¢
.B SIGIO
¤ÎÂå¤ï¤ê¤ËÁ÷¤ë¥·¥°¥Ê¥ëÈÖ¹æ¤òɽ¤¹¡£
¸å¼Ô¤Î¾ì¹ç¡¢¥·¥°¥Ê¥ë¥Ï¥ó¥É¥é¤ò
.B SA_SIGINFO
¥Õ¥é¥°ÉÕ¤­¤ÇÀßÄꤹ¤ì¤Ð¡¢¥Ï¥ó¥É¥é¤ÇÄɲäξðÊó¤òÆÀ¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
.I arg
¤Ï̵»ë¤µ¤ì¤ë¡£
.TP
.BR F_SETSIG " (\fIlong\fP)"
.\"O Set the signal sent when input or output becomes possible
.\"O to the value given in
.\"O .IR arg .
.\"O A value of zero means to send the default
.\"O .B SIGIO
.\"O signal.
.\"O Any other value (including
.\"O .BR SIGIO )
.\"O is the signal to send instead, and in this case additional info
.\"O is available to the signal handler if installed with
.\"O .BR SA_SIGINFO .
ÆþÎϤä½ÐÎϤ¬²Äǽ¤Ë¤Ê¤Ã¤¿¾ì¹ç¤ËÁ÷¤ë¥·¥°¥Ê¥ë¤ò
.I arg
¤Ë»ØÄꤵ¤ì¤¿ÃͤËÀßÄꤹ¤ë¡£
ÃÍ¥¼¥í¤Ï
.B SIGIO
¤òÁ÷¤ë¤³¤È¤ò°ÕÌ£¤¹¤ë¡£
.RB ( SIGIO
¤ò´Þ¤à) ¾¤ÎÃͤϤ¤¤º¤ì¤â¡¢
.B SIGIO
¤ÎÂå¤ï¤ê¤ËÁ÷¤ë¥·¥°¥Ê¥ëÈÖ¹æ¤òɽ¤¹¡£
¸å¼Ô¤Î¾ì¹ç¡¢¥·¥°¥Ê¥ë¥Ï¥ó¥É¥é¤ò
.B SA_SIGINFO
¥Õ¥é¥°ÉÕ¤­¤ÇÀßÄꤹ¤ì¤Ð¡¢
¥Ï¥ó¥É¥é¤ÇÄɲäξðÊó¤òÆÀ¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
.\"
.\" The following was true only up until 2.6.11:
.\"
.\"O .\" Additionally, passing a nonzero value to
.\"O .\" .B F_SETSIG
.\"O .\" changes the signal recipient from a whole process to a specific thread
.\"O .\" within a process.
.\"O .\" See the description of
.\"O .\" .B F_SETOWN
.\"O .\" for more details.
.\" ¤Þ¤¿¡¢
.\" .B F_SETSIG
.\" ¤Ë 0 °Ê³°¤ÎÃͤòÅϤ¹¤È¡¢¥·¥°¥Ê¥ë¤Î¼õ¿®¼Ô¤ò¥×¥í¥»¥¹Á´ÂΤ«¤é
.\" ¥×¥í¥»¥¹Æâ¤ÎÆÃÄê¤Î¥¹¥ì¥Ã¥É¤ËÊѹ¹¤µ¤ì¤ë¡£¾ÜºÙ¤Ï
.\" .B F_SETOWN
.\" ¤ÎÀâÌÀ¤ò»²¾È¤Î¤³¤È¡£

.\"O By using
.\"O .B F_SETSIG
.\"O with a nonzero value, and setting
.\"O .B SA_SIGINFO
.\"O for the
.\"O signal handler (see
.\"O .BR sigaction (2)),
.\"O extra information about I/O events is passed to
.\"O the handler in a
.\"O .I siginfo_t
.\"O structure.
.B F_SETSIG
¤Ë¥¼¥í°Ê³°¤ÎÃͤòÀßÄꤷ¡¢¥·¥°¥Ê¥ë¥Ï¥ó¥É¥é¤Ë
.B SA_SIGINFO
¥Õ¥é¥°¤òÀßÄꤹ¤ë¤È¡¢
.RB ( sigaction (2)
¤ò»²¾È) I/O ¥¤¥Ù¥ó¥È¤Ë´Ø¤¹¤ëÄɲäξðÊó¤¬
.I siginfo_t
¹½Â¤ÂΤǥ·¥°¥Ê¥ë¥Ï¥ó¥É¥é¤ØÅϤµ¤ì¤ë¡£
.\"O If the
.\"O .I si_code
.\"O field indicates the source is
.\"O .BR SI_SIGIO ,
.\"O the
.\"O .I si_fd
.\"O field gives the file descriptor associated with the event.
.\"O Otherwise,
.\"O there is no indication which file descriptors are pending, and you
.\"O should use the usual mechanisms
.\"O .RB ( select (2),
.\"O .BR poll (2),
.\"O .BR read (2)
.\"O with
.\"O .B O_NONBLOCK
.\"O set etc.) to determine which file descriptors are available for I/O.
.I si_code
¥Õ¥£¡¼¥ë¥É¤¬¼¨¤¹¥·¥°¥Ê¥ë¤Î¸¶°ø¤¬
.B SI_SIGIO
¤Ç¤¢¤ë¾ì¹ç¡¢
.I si_fd
¥Õ¥£¡¼¥ë¥É¤Ë¤Ï¥¤¥Ù¥ó¥È¤ËÂбþ¤¹¤ë¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤¬Æþ¤Ã¤Æ¤¤¤ë¡£
¤½¤ì°Ê³°¤Î¾ì¹ç¤Ï¡¢¤É¤Î¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤¬ÍøÍѲÄǽ¤«¤ò¼¨¤¹¾ðÊó¤Ï
¤Ê¤¤¤Î¤Ç¡¢¤É¤Î¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Ç I/O ¤¬²Äǽ¤«¤òȽÃǤ¹¤ë¤¿¤á¤Ë¤Ï
Ä̾ï¤Îµ¡¹½
.RB ( select (2),
.BR poll (2),
.B O_NONBLOCK
¤òÀßÄꤷ¤¿
.BR read (2)
¤Ê¤É) ¤ò»ÈÍѤ·¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£

.\"O By selecting a real time signal (value >=
.\"O .BR SIGRTMIN ),
.\"O multiple I/O events may be queued using the same signal numbers.
.\"O (Queuing is dependent on available memory).
.\"O Extra information is available
.\"O if
.\"O .B SA_SIGINFO
.\"O is set for the signal handler, as above.
¥ê¥¢¥ë¥¿¥¤¥à¥·¥°¥Ê¥ë (Ãͤ¬
.B SIGRTMIN
°Ê¾å) ¤òÁªÂò¤·¤Æ¤¤¤ë¾ì¹ç¤Ï¡¢
Ʊ¤¸¥·¥°¥Ê¥ëÈÖ¹æ¤ò»ý¤ÄÊ£¿ô¤Î I/O ¥¤¥Ù¥ó¥È¤¬¥­¥å¡¼¤ËÆþ¤ë¤³¤È¤¬¤¢¤ë
(¥­¥å¡¼¤ËÆþ¤ì¤ë¤«¤É¤¦¤«¤ÏÍøÍѲÄǽ¤Ê¥á¥â¥ê¤Ë°Í¸¤·¤Æ¤¤¤ë)¡£
¾åµ­¤ÈƱÍÍ¡¢ 
.B SA_SIGINFO
¤¬ÀßÄꤵ¤ì¤Æ¤¤¤ë¾ì¹ç¡¢¥·¥°¥Ê¥ë¥Ï¥ó¥É¥é¤Î¤¿¤á¤ÎÄɲäξðÊó¤¬ÆÀ¤é¤ì¤ë¡£

.\"O Note that Linux imposes a limit on the
.\"O number of real-time signals that may be queued to a
.\"O process (see
.\"O .BR getrlimit (2)
.\"O and
.\"O .BR signal (7))
.\"O and if this limit is reached, then the kernel reverts to
.\"O delivering
.\"O .BR SIGIO ,
.\"O and this signal is delivered to the entire
.\"O process rather than to a specific thread.
.\" See fs/fcntl.c::send_sigio_to_task() (2.4/2.6) sources -- MTK, Apr 05
°Ê²¼¤ÎÅÀ¤ËÃí°Õ¤¹¤ë¤³¤È¡£
Linux ¤Ç¤Ï°ì¤Ä¤Î¥×¥í¥»¥¹¤ËÂФ·¤Æ¥­¥å¡¼¤ËÆþ¤ì¤é¤ì¤ë¥ê¥¢¥ë¥¿¥¤¥à
¥·¥°¥Ê¥ë¤Î¿ô¤Ë¾å¸Â¤¬Àߤ±¤é¤ì¤Æ¤ª¤ê
.RB ( getrlimit (2)
¤È
.BR signal (7)
¤ò»²¾È)¡¢¤³¤Î¾å¸Â¤Ë㤹¤ë¤È¥«¡¼¥Í¥ë¤Ï
.B SIGIO
¥·¥°¥Ê¥ë¤òÇÛÁ÷¤¹¤ë¡£¤³¤Î
.B SIGIO
¥·¥°¥Ê¥ë¤Ï¡¢»ØÄꤵ¤ì¤¿¥¹¥ì¥Ã¥É¤Ç¤Ï¤Ê¤¯¥×¥í¥»¥¹Á´ÂΤËÁ÷¤é¤ì¤ë¡£
.PP
.\"O Using these mechanisms, a program can implement fully asynchronous I/O
.\"O without using
.\"O .BR select (2)
.\"O or
.\"O .BR poll (2)
.\"O most of the time.
¤³¤ì¤é¤Îµ¡¹½¤ò»ÈÍѤ¹¤ë¤³¤È¤Ç¡¢¤Û¤È¤ó¤É¤Î¾ì¹ç¤Ç
.BR select (2)
¤ä
.BR poll (2)
¤ò»ÈÍѤ»¤º¤Ë´°Á´¤ÊÈóƱ´ü I/O ¤ò¼ÂÁõ¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
.PP
.\"O The use of
.\"O .BR O_ASYNC ,
.\"O .BR F_GETOWN ,
.\"O .B F_SETOWN
.\"O is specific to BSD and Linux.
.BR O_ASYNC ,
.BR F_GETOWN ,
.B F_SETOWN
¤Î»ÈÍÑ¤Ï BSD ¤È Linux ¤ËÆÃÍ­¤Ç¤¢¤ë¡£
.\"O .BR F_GETOWN_EX ,
.\"O .BR F_SETOWN_EX ,
.\"O .BR F_GETSIG ,
.\"O and
.\"O .B F_SETSIG
.\"O are Linux-specific.
.\"O POSIX has asynchronous I/O and the
.\"O .I aio_sigevent
.\"O structure to achieve similar things; these are also available
.\"O in Linux as part of the GNU C Library (Glibc).
.BR F_GETOWN_EX ,
.BR F_SETOWN_EX ,
.BR F_GETSIG ,
.B F_SETSIG
¤Ï Linux ¸ÇÍ­¤Ç¤¢¤ë¡£POSIX ¤Ë¤Ï¡¢Æ±ÍͤΤ³¤È¤ò¹Ô¤¦¤¿¤á¤Ë¡¢ÈóƱ´ü I/O ¤È
.I aio_sigevent
¹½Â¤ÂΤ¬¤¢¤ë¡£Linux ¤Ç¤Ï¡¢GNU C ¥é¥¤¥Ö¥é¥ê (Glibc) ¤Î°ìÉô¤È¤·¤Æ
¤³¤ì¤é¤âÍøÍѲÄǽ¤Ç¤¢¤ë¡£
.\"O .SS Leases
.SS "¥ê¡¼¥¹ (leases)"
.\"O .B F_SETLEASE
.\"O and
.\"O .B F_GETLEASE
.\"O (Linux 2.4 onwards) are used (respectively) to establish a new lease,
.\"O and retrieve the current lease, on the open file description
.\"O referred to by the file descriptor
.\"O .IR fd .
(Linix 2.4 °Ê¹ß¤ÇÍøÍѲÄǽ)
.B F_SETLEASE
¤Ï¡¢
.I fd
¤¬»²¾È¤¹¤ë¥ª¡¼¥×¥ó¥Õ¥¡¥¤¥ëµ­½Ò¤ËÂФ·¤Æ¿·¤·¤¤¥ê¡¼¥¹¤òÀßÄꤹ¤ë¤Î¤Ë»ÈÍѤµ¤ì¤ë¡£
.B F_GETLEASE
¤Ï¡¢
.I fd
¤¬»²¾È¤¹¤ë¥ª¡¼¥×¥ó¥Õ¥¡¥¤¥ëµ­½Ò¤ËÂФ·¤ÆÀßÄꤵ¤ì¤Æ¤¤¤ë
¸½ºß¤Î¥ê¡¼¥¹¤ò¼èÆÀ¤¹¤ë¤Î¤Ë»ÈÍѤµ¤ì¤ë¡£
.\"O A file lease provides a mechanism whereby the process holding
.\"O the lease (the "lease holder") is notified (via delivery of a signal)
.\"O when a process (the "lease breaker") tries to
.\"O .BR open (2)
.\"O or
.\"O .BR truncate (2)
.\"O the file referred to by that file descriptor.
¥Õ¥¡¥¤¥ë¤Î¥ê¡¼¥¹¤Ë¤è¤ê¡¢
¤¢¤ë¥×¥í¥»¥¹ ("lease breaker") ¤¬¤½¤Î¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤¬»²¾È
¤·¤Æ¤¤¤ë¥Õ¥¡¥¤¥ë¤ËÂФ·¤Æ
.BR open (2)
¤ä
.BR truncate (2)
¤ò¹Ô¤ª¤¦¤È¤·¤¿ºÝ¤Ë¡¢¥ê¡¼¥¹¤òÊÝ»ý¤·¤Æ¤¤¤ë¥×¥í¥»¥¹ ("lease holder") ¤Ø
(¥·¥°¥Ê¥ë¤ÎÇÛÁ÷¤Ë¤è¤ë) ÄÌÃΤ¬¹Ô¤ï¤ì¤ë¤È¤¤¤¦µ¡¹½¤¬Ä󶡤µ¤ì¤ë¡£
.TP
.BR F_SETLEASE " (\fIlong\fP)"
.\"O Set or remove a file lease according to which of the following
.\"O values is specified in the integer
.\"O .IR arg :
.I arg
¤ÎÆâÍƤ˴𤤤ƥե¡¥¤¥ë¤Î¥ê¡¼¥¹¤ÎÀßÄê¡¢ºï½ü¤ò¹Ô¤¦¡£À°¿ô
.I arg
¤Ë¤Ï°Ê²¼¤ÎÃͤ¬»ØÄê¤Ç¤­¤ë:

.RS
.TP
.B F_RDLCK
.\"O Take out a read lease.
.\"O This will cause the calling process to be notified when
.\"O the file is opened for writing or is truncated.
.\"O .\" The following became true in kernel 2.6.10:
.\"O .\" See the man-pages-2.09 Changelog for further info.
.\"O A read lease can only be placed on a file descriptor that
.\"O is opened read-only.
Æɤ߽Ф·¥ê¡¼¥¹¤ò¼èÆÀ¤¹¤ë¡£¤³¤ì¤Ë¤è¤ê¡¢
¤½¤Î¥Õ¥¡¥¤¥ë¤¬½ñ¤­¹þ¤ßÍѤ˥ª¡¼¥×¥ó¤µ¤ì¤¿¤ê¡¢¥Õ¥¡¥¤¥ë¤¬ÀÚ¤êµÍ¤á¤é¤ì¤¿¾ì¹ç¤Ë¡¢
¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤ËÄÌÃΤ¬¹Ô¤ï¤ì¤ë¤è¤¦¤Ë¤Ê¤ë¡£
.\" °Ê²¼¤ÎÆâÍƤϥ«¡¼¥Í¥ë 2.6.10 ¤Ç¼ÂÁõ¤µ¤ì¤¿¡£
.\" ¤è¤ê¾Ü¤·¤¤¾ðÊó¤Ï man-pages-2.09 ¤Î Changelog ¤ò»²¾È¡£
Æɤ߽Ф·¥ê¡¼¥¹¤òÀßÄê¤Ç¤­¤ë¤Î¤Ï¡¢Æɤ߽Ф·ÀìÍѤǥª¡¼¥×¥ó¤µ¤ì¤Æ¤¤¤ë
¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤ËÂФ·¤Æ¤Î¤ß¤Ç¤¢¤ë¡£
.TP
.B F_WRLCK
.\"O Take out a write lease.
.\"O This will cause the caller to be notified when
.\"O the file is opened for reading or writing or is truncated.
.\"O A write lease may be placed on a file only if there are no
.\"O other open file descriptors for the file.
½ñ¤­¹þ¤ß¥ê¡¼¥¹¤ò¼èÆÀ¤¹¤ë¡£¤³¤ì¤Ë¤è¤ê¡¢
(Æɤ߽Ф·ÍѤ«½ñ¤­¹þ¤ßÍѤˤ«¤«¤ï¤é¤º) ¤½¤Î¥Õ¥¡¥¤¥ë¤¬¥ª¡¼¥×¥ó¤µ¤ì¤¿¤ê¡¢
¥Õ¥¡¥¤¥ë¤¬ÀÚ¤êµÍ¤á¤é¤ì¤¿¾ì¹ç¤Ë¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤ËÄÌÃΤ¬¹Ô¤ï¤ì¤ë¤è¤¦¤Ë¤Ê¤ë¡£
½ñ¤­¹þ¤ß¥ê¡¼¥¹¤Ï¡¢¤½¤Î¥Õ¥¡¥¤¥ë¤ËÂФ¹¤ë¥ª¡¼¥×¥ó¤µ¤ì¤¿¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤¬
¾¤Ë¤Ê¤¤¾ì¹ç¤Ë¤Î¤ßÀßÄê¤Ç¤­¤ë¡£
.TP
.B F_UNLCK
.\"O Remove our lease from the file.
¤½¤Î¥Õ¥¡¥¤¥ë¤«¤é¥ê¡¼¥¹¤òºï½ü¤¹¤ë¡£
.RE
.P
.\"O Leases are associated with an open file description (see
.\"O .BR open (2)).
.\"O This means that duplicate file descriptors (created by, for example,
.\"O .BR fork (2)
.\"O or
.\"O .BR dup (2))
.\"O refer to the same lease, and this lease may be modified
.\"O or released using any of these descriptors.
.\"O Furthermore, the lease is released by either an explicit
.\"O .B F_UNLCK
.\"O operation on any of these duplicate descriptors, or when all
.\"O such descriptors have been closed.
¥ê¡¼¥¹¤Ï¥ª¡¼¥×¥ó¥Õ¥¡¥¤¥ëµ­½Ò¤ËÂФ·¤Æ´ØÏ¢ÉÕ¤±¤é¤ì¤ë
.RB ( open (2)
»²¾È)¡£
¤Ä¤Þ¤ê¡¢
.RB ( fork (2)
¤ä
.BR dup (2)
¤Ê¤É¤Ë¤è¤êºîÀ®¤µ¤ì¤¿) ¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤ÎÊ£À½¤ÏƱ¤¸¥ê¡¼¥¹¤ò»²¾È¤·¡¢
Ê£À½¤â´Þ¤á¤¿¤É¤Î¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤ò»È¤Ã¤Æ¤â¤³¤Î¥ê¡¼¥¹¤òÊѹ¹¤·¤¿¤ê
²òÊü¤·¤¿¤ê¤Ç¤­¤ë¡£
¤Þ¤¿¡¢¤³¤ì¤é¤Î¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Î¤¤¤º¤ì¤«¤ËÂФ·¤Æ
.B F_UNLCK
Áàºî¤¬ÌÀ¼¨Åª¤Ë¼Â¹Ô¤µ¤ì¤¿¾ì¹ç¤ä¡¢¤¹¤Ù¤Æ¤Î¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤¬
ÊĤ¸¤é¤ì¤¿¾ì¹ç¤Ë¤â¡¢¥ê¡¼¥¹¤Ï²òÊü¤µ¤ì¤ë¡£
.P
.\"O Leases may only be taken out on regular files.
.\"O An unprivileged process may only take out a lease on a file whose
.\"O UID (owner) matches the file system UID of the process.
.\"O A process with the
.\"O .B CAP_LEASE
.\"O capability may take out leases on arbitrary files.
¥ê¡¼¥¹¤Î¼èÆÀ¤ÏÄ̾ï¤Î¥Õ¥¡¥¤¥ë (regular file) ¤ËÂФ·¤Æ¤Î¤ß²Äǽ¤Ç¤¢¤ë¡£
ÈóÆø¢¥×¥í¥»¥¹¤¬¥ê¡¼¥¹¤ò¼èÆÀ¤Ç¤­¤ë¤Î¤Ï¡¢UID (½êÍ­¼Ô) ¤¬¥×¥í¥»¥¹¤Î
¥Õ¥¡¥¤¥ë¥·¥¹¥Æ¥à UID ¤È°ìÃפ¹¤ë¥Õ¥¡¥¤¥ë¤ËÂФ·¤Æ¤À¤±¤Ç¤¢¤ë¡£
.B CAP_LEASE
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò»ý¤Ä¥×¥í¥»¥¹¤ÏǤ°Õ¤Î¥Õ¥¡¥¤¥ë¤ËÂФ·¤Æ¥ê¡¼¥¹¤ò¼èÆÀ¤Ç¤­¤ë¡£
.TP
.BR F_GETLEASE " (\fIvoid\fP)"
.\"O Indicates what type of lease is associated with the file descriptor
.\"O .I fd
.\"O by returning either
.\"O .BR F_RDLCK ", " F_WRLCK ", or " F_UNLCK ,
.\"O indicating, respectively, a read lease , a write lease, or no lease.
.\"O .I arg
.\"O is ignored.
¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿
.I fd
¤ËÂФ·¤ÆÀßÄꤵ¤ì¤Æ¤¤¤ë¥ê¡¼¥¹¤Î¼ïÊ̤ò¼èÆÀ¤¹¤ë¡£
.BR F_RDLCK ", " F_WRLCK ", " F_UNLCK
¤Î¤¤¤º¤ì¤«¤¬ÊÖ¤µ¤ì¤ë¡£
.BR F_RDLCK ", " F_WRLCK
¤Ï¤½¤ì¤¾¤ì¡¢Æɤ߽Ф·¥ê¡¼¥¹¡¢½ñ¤­¹þ¤ß¥ê¡¼¥¹¤¬ÀßÄꤵ¤ì¤Æ¤¤¤ë¤³¤È¤ò¼¨¤·¡¢
.B F_UNLCK
¤Ï¥ê¡¼¥¹¤¬²¿¤âÀßÄꤵ¤ì¤Æ¤¤¤Ê¤¤¤³¤È¤ò¼¨¤¹¡£
.I arg
¤Ï̵»ë¤µ¤ì¤ë¡£
.PP
.\"O When a process (the "lease breaker") performs an
.\"O .BR open (2)
.\"O or
.\"O .BR truncate (2)
.\"O that conflicts with a lease established via
.\"O .BR F_SETLEASE ,
.\"O the system call is blocked by the kernel and
.\"O the kernel notifies the lease holder by sending it a signal
.\"O .RB ( SIGIO
.\"O by default).
¤¢¤ë¥×¥í¥»¥¹ ("lease folder") ¤¬
.B F_SETLEASE
¤ÇÀßÄꤵ¤ì¤¿¥ê¡¼¥¹¤ÈÌ·½â¤¹¤ë¤è¤¦¤Ê
.BR open (2)
¤ä
.BR truncate (2)
¤ò¼Â¹Ô¤·¤¿¾ì¹ç¡¢
¤½¤Î¥·¥¹¥Æ¥à¥³¡¼¥ë¤Ï¥«¡¼¥Í¥ë¤Ë¤è¤Ã¤ÆÄä»ß¤µ¤ì¡¢
¥«¡¼¥Í¥ë¤Ï lease holder ¤Ë¥·¥°¥Ê¥ë (¥Ç¥Õ¥©¥ë¥È¤Ç¤Ï
.BR SIGIO )
¤òÁ÷¤Ã¤ÆÄÌÃΤò¹Ô¤¦¡£
.\"O The lease holder should respond to receipt of this signal by doing
.\"O whatever cleanup is required in preparation for the file to be
.\"O accessed by another process (e.g., flushing cached buffers) and
.\"O then either remove or downgrade its lease.
lease holder ¤Ï¤³¤Î¥·¥°¥Ê¥ë¤ò¼õ¿®¤·¤¿¤È¤­¤Ë¤Ï¤­¤Á¤ó¤ÈÂбþ¤¹¤Ù¤­¤Ç¤¢¤ë¡£
¶ñÂÎŪ¤Ë¤Ï¡¢ÊÌ¤Î¥×¥í¥»¥¹¤¬¤½¤Î¥Õ¥¡¥¤¥ë¤Ë¥¢¥¯¥»¥¹¤¹¤ë¤¿¤á¤Î½àÈ÷¤È¤·¤Æ
ɬÍפʸåÊÒÉÕ¤± (Î㤨¤Ð¡¢¥­¥ã¥Ã¥·¥å¤µ¤ì¤¿¥Ð¥Ã¥Õ¥¡¤Î¥Õ¥é¥Ã¥·¥å) ¤ò
¤¹¤Ù¤Æ¹Ô¤Ã¤Æ¤«¤é¡¢¤½¤Î¥Õ¥¡¥¤¥ë¤Î¥ê¡¼¥¹¤Îºï½ü¤Þ¤¿¤Ï³Ê²¼¤²¤ò¹Ô¤¦¡£
.\"O A lease is removed by performing an
.\"O .B F_SETLEASE
.\"O command specifying
.\"O .I arg
.\"O as
.\"O .BR F_UNLCK .
¥ê¡¼¥¹¤òºï½ü¤ò¤¹¤ë¤Ë¤Ï¡¢
.I arg
¤Ë
.B F_UNLCK
¤ò»ØÄꤷ¤Æ
.B F_SETLEASE
¤ò¼Â¹Ô¤¹¤ë¡£
.\"O If the lease holder currently holds a write lease on the file,
.\"O and the lease breaker is opening the file for reading,
.\"O then it is sufficient for the lease holder to downgrade
.\"O the lease to a read lease.
.\"O This is done by performing an
.\"O .B F_SETLEASE
.\"O command specifying
.\"O .I arg
.\"O as
.\"O .BR F_RDLCK .
lease holder ¤¬¥Õ¥¡¥¤¥ë¤Ë½ñ¤­¹þ¤ß¥ê¡¼¥¹¤òÊÝ»ý¤·¤Æ¤¤¤Æ¡¢
lease breaker ¤¬Æɤ߽Ф·ÍѤˤ½¤Î¥Õ¥¡¥¤¥ë¤ò¥ª¡¼¥×¥ó¤·¤Æ¤¤¤ë¾ì¹ç¡¢
lease holder ¤¬ÊÝ»ý¤·¤Æ¤¤¤ë¥ê¡¼¥¹¤òÆɤ߽Ф·¥ê¡¼¥¹¤Ë³Ê²¼¤²¤¹¤ì¤Ð
½½Ê¬¤Ç¤¢¤ë¡£¤³¤ì¤ò¤¹¤ë¤Ë¤Ï¡¢
.I arg
¤Ë
.B F_RDLCK
¤ò»ØÄꤷ¤Æ
.B F_SETLEASE
¤ò¼Â¹Ô¤¹¤ë¡£

.\"O If the lease holder fails to downgrade or remove the lease within
.\"O the number of seconds specified in
.\"O .I /proc/sys/fs/lease-break-time
.\"O then the kernel forcibly removes or downgrades the lease holder's lease.
lease holder ¤¬
.I /proc/sys/fs/lease-break-time
¤Ç»ØÄꤵ¤ì¤¿Éÿô°ÊÆâ¤Ë¥ê¡¼¥¹¤Î³Ê²¼¤²¤«ºï½ü¤ò¹Ô¤¨¤Ê¤«¤Ã¤¿¾ì¹ç¡¢
¥«¡¼¥Í¥ë¤Ï¶¯À©Åª¤Ë¤½¤Î lease holder ¤Î¥ê¡¼¥¹¤òºï½ü¤â¤·¤¯¤Ï³Ê²¼¤²¤ò¹Ô¤¦¡£

.\"O Once the lease has been voluntarily or forcibly removed or downgraded,
.\"O and assuming the lease breaker has not unblocked its system call,
.\"O the kernel permits the lease breaker's system call to proceed.
°ìÅ٥꡼¥¹¤Îºï½ü¤«³Ê²¼¤²¤¬¼«È¯Åª¤â¤·¤¯¤Ï¶¯À©Åª¤Ë¹Ô¤ï¤ì¤ë¤È¡¢
lease breaker ¤¬¤Þ¤À¥·¥¹¥Æ¥à¥³¡¼¥ë¤òºÆ³«¤·¤Æ¤¤¤Ê¤¤¾ì¹ç¤Ë¤Ï¡¢
¥«¡¼¥Í¥ë¤¬ lease breaker ¤Î¥·¥¹¥Æ¥à¥³¡¼¥ë¤Î³¹Ô¤òµö²Ä¤¹¤ë¡£

.\"O If the lease breaker's blocked
.\"O .BR open (2)
.\"O or
.\"O .BR truncate (2)
.\"O is interrupted by a signal handler,
.\"O then the system call fails with the error
.\"O .BR EINTR ,
.\"O but the other steps still occur as described above.
lease breaker ¤¬¼Â¹Ô¤·¤¿
.BR open (2)
¤ä
.BR truncate (2)
¤¬Ää»ßÃæ¤Ë¥·¥°¥Ê¥ë¥Ï¥ó¥É¥é¤Ë¤è¤êÃæÃǤµ¤ì¤¿¾ì¹ç¡¢
¤½¤Î¥·¥¹¥Æ¥à¥³¡¼¥ë¤Ï
.B EINTR
¥¨¥é¡¼¤Ç¼ºÇÔ¤¹¤ë¤¬¡¢¾å¤Ç½Ò¤Ù¤¿Â¾¤Î½èÍý¤Ï
¤½¤Î¤Þ¤Þ¹Ô¤ï¤ì¤ë¡£
.\"O If the lease breaker is killed by a signal while blocked in
.\"O .BR open (2)
.\"O or
.\"O .BR truncate (2),
.\"O then the other steps still occur as described above.
.BR open (2)
¤ä
.BR truncate (2)
¤¬Ää»ßÃæ¤Ë lease breaker ¤¬¥·¥°¥Ê¥ë¤Ë¤è¤ê kill ¤µ¤ì¤¿¾ì¹ç¡¢
¾å¤Ç½Ò¤Ù¤¿Â¾¤Î½èÍý¤Ï¤½¤Î¤Þ¤Þ¹Ô¤ï¤ì¤ë¡£
.\"O If the lease breaker specifies the
.\"O .B O_NONBLOCK
.\"O flag when calling
.\"O .BR open (2),
.\"O then the call immediately fails with the error
.\"O .BR EWOULDBLOCK ,
.\"O but the other steps still occur as described above.
lease breaker ¤¬
.BR open (2)
¤ò¸Æ¤ÖºÝ¤Ë
.B O_NONBLOCK
¥Õ¥é¥°¤ò»ØÄꤷ¤¿¾ì¹ç¡¢¤½¤Î¥·¥¹¥Æ¥à¥³¡¼¥ë¤Ï
.B EWOULDBLOCK
¥¨¥é¡¼¤Çľ¤Á¤Ë¼ºÇÔ¤¹¤ë¤¬¡¢¾å¤Ç½Ò¤Ù¤¿Â¾¤Î½èÍý¤Ï¤½¤Î¤Þ¤Þ¹Ô¤ï¤ì¤ë¡£

.\"O The default signal used to notify the lease holder is
.\"O .BR SIGIO ,
.\"O but this can be changed using the
.\"O .B F_SETSIG
.\"O command to
.\"O .BR fcntl ().
.\"O If a
.\"O .B F_SETSIG
.\"O command is performed (even one specifying
.\"O .BR SIGIO ),
.\"O and the signal
.\"O handler is established using
.\"O .BR SA_SIGINFO ,
.\"O then the handler will receive a
.\"O .I siginfo_t
.\"O structure as its second argument, and the
.\"O .I si_fd
.\"O field of this argument will hold the descriptor of the leased file
.\"O that has been accessed by another process.
.\"O (This is useful if the caller holds leases against multiple files).
lease holder ¤Ø¤ÎÄÌÃΤ˻Ȥï¤ì¤ë¥Ç¥Õ¥©¥ë¥È¤Î¥·¥°¥Ê¥ë¤Ï
.B SIGIO
¤À¤¬¡¢
.BR fcntl ()
¤Î
.B F_SETSIG
¥³¥Þ¥ó¥É¤ÇÊѹ¹¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
.B F_SETSIG
¥³¥Þ¥ó¥É¤¬¼Â¹Ô¤µ¤ì
.RB ( SIGIO
¤ò»ØÄꤵ¤ì¤¿¾ì¹ç¤â´Þ¤à)¡¢
.B SA_SIGINFO
¥Õ¥é¥°ÉÕ¤­¤Ç¥·¥°¥Ê¥ë¥Ï¥ó¥É¥é¤¬ÀßÄꤵ¤ì¤Æ¤¤¤ë¾ì¹ç¤Ë¤Ï¡¢
¥Ï¥ó¥É¥é¤ÎÂèÆó°ú¤­¿ô¤È¤·¤Æ
.I siginfo_t
¹½Â¤ÂΤ¬ÅϤµ¤ì¡¢¤³¤Î°ú¤­¿ô¤Î
.I si_fd
¥Õ¥£¡¼¥ë¥É¤Ë¤ÏÊÌ¤Î¥×¥í¥»¥¹¤¬¥¢¥¯¥»¥¹¤·¤¿¥ê¡¼¥¹ÀßÄêºÑ¤ß¥Õ¥¡¥¤¥ë¤Î
¥Ç¥£¥¹¥¯¥ê¥×¥¿¤¬Æþ¤Ã¤Æ¤¤¤ë
(¤³¤Îµ¡Ç½¤ÏÊ£¿ô¤Î¥Õ¥¡¥¤¥ë¤ËÂФ·¤Æ¥ê¡¼¥¹¤òÀßÄꤹ¤ë¾ì¹ç¤ËÍ­ÍѤǤ¢¤ë)¡£
.\"O .SS "File and directory change notification (dnotify)"
.SS "¥Õ¥¡¥¤¥ë¤ä¥Ç¥£¥ì¥¯¥È¥ê¤ÎÊѹ¹¤ÎÄÌÃÎ (dnotify)"
.TP
.BR F_NOTIFY " (\fIlong\fP)"
.\"O (Linux 2.4 onwards)
(Linux 2.4 °Ê¹ß)
.\"O Provide notification when the directory referred to by
.\"O .I fd
.\"O or any of the files that it contains is changed.
.\"O The events to be notified are specified in
.\"O .IR arg ,
.\"O which is a bit mask specified by ORing together zero or more of
.\"O the following bits:
.I fd
¤Ç»²¾È¤µ¤ì¤ë¥Ç¥£¥ì¥¯¥È¥ê¤«¡¢¤½¤ÎÃæ¤Ë¤¢¤ë¥Õ¥¡¥¤¥ë¤ËÊѹ¹¤¬¤¢¤Ã¤¿¾ì¹ç¤Ë
ÄÌÃΤò¹Ô¤¦¡£¤É¤Î¥¤¥Ù¥ó¥È¤òÄÌÃΤ¹¤ë¤«¤Ï
.I arg
¤Ç»ØÄꤹ¤ë¡£
.I arg
¤Ï¥Ó¥Ã¥È¥Þ¥¹¥¯¤Ç¡¢°Ê²¼¤Î¥Ó¥Ã¥È¤Î 0¸Ä°Ê¾å¤ÎÏÀÍýϤò¤È¤Ã¤¿¤â¤Î¤ò»ØÄꤹ¤ë¡£
.RS
.sp
.PD 0
.TP 12
.B DN_ACCESS
.\"O A file was accessed (read, pread, readv)
¥Õ¥¡¥¤¥ë¤Ø¤Î¥¢¥¯¥»¥¹¤¬¤¢¤Ã¤¿ (read, pread, readv)
.TP
.B DN_MODIFY
.\"O A file was modified (write, pwrite, writev, truncate, ftruncate).
¥Õ¥¡¥¤¥ë¤ÎÆâÍƤ¬Êѹ¹¤µ¤ì¤¿ (write, pwrite, writev, truncate, ftruncate).
.TP
.B DN_CREATE
.\"O A file was created (open, creat, mknod, mkdir, link, symlink, rename).
¥Õ¥¡¥¤¥ë¤¬ºîÀ®¤µ¤ì¤¿ (open, creat, mknod, mkdir, link, symlink, rename).
.TP
.B DN_DELETE
.\"O A file was unlinked (unlink, rename to another directory, rmdir).
¥Õ¥¡¥¤¥ë¤¬ºï½ü (unlink) ¤µ¤ì¤¿ (unlink, Ê̤Υǥ£¥ì¥¯¥È¥ê¤Ø¤Î rename, rmdir)
.TP
.B DN_RENAME
.\"O A file was renamed within this directory (rename).
¥Ç¥£¥ì¥¯¥È¥êÆâ¤Ç¤Î¥Õ¥¡¥¤¥ë̾¤ÎÊѹ¹¤¬¤¢¤Ã¤¿ (rename)
.TP
.B DN_ATTRIB
.\"O The attributes of a file were changed (chown, chmod, utime[s]).
¥Õ¥¡¥¤¥ë°À­¤¬Êѹ¹¤µ¤ì¤¿ (chown, chmod, utime[s])
.PD
.RE
.IP
.\"O (In order to obtain these definitions, the
.\"O .B _GNU_SOURCE
.\"O feature test macro must be defined.)
(¾åµ­¤ÎÄêµÁ¤òÍøÍѤ¹¤ë¤Ë¤Ï
.B _GNU_SOURCE
µ¡Ç½¸¡ºº¥Þ¥¯¥í¤òÄêµÁ¤·¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£)

.\"O Directory notifications are normally "one-shot", and the application
.\"O must reregister to receive further notifications.
¥Ç¥£¥ì¥¯¥È¥ê¤ÎÊѹ¹ÄÌÃΤÏÄ̾ï¡Ö°ì²ó¸Â¤ê (one-shot)¡×¤Ç¤¢¤ê¡¢
¥¢¥×¥ê¥±¡¼¥·¥ç¥ó¦¤Ç¤½¤Î¸å¤µ¤é¤ËÄÌÃΤò¼õ¿®¤·¤¿¤¤¾ì¹ç¤Ï
ºÆÅÐÏ¿¤·¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£
.\"O Alternatively, if
.\"O .B DN_MULTISHOT
.\"O is included in
.\"O .IR arg ,
.\"O then notification will remain in effect until explicitly removed.
.I arg
¤Ë
.B DN_MULTISHOT
¤¬´Þ¤Þ¤ì¤Æ¤¤¤¿¾ì¹ç¤Ë¤Ï¡¢
Êѹ¹ÄÌÃΤÏÌÀ¼¨Åª¤Ë²ò½ü¤µ¤ì¤ë¤Þ¤ÇÍ­¸ú¾õÂÖ¤¬·Ñ³¤¹¤ë¡£

.\"O .\" The following does seem a poor API-design choice...
.\" °Ê²¼¤Ï API ¤ÎÀ߷פ¬¤Þ¤º¤¤¤È»×¤¦¤Î¤À¤¬...
.\"O A series of
.\"O .B F_NOTIFY
.\"O requests is cumulative, with the events in
.\"O .I arg
.\"O being added to the set already monitored.
.B F_NOTIFY
Í×µá¤ÏÀÑ»»¤µ¤ì¤Æ¤¤¤¯¡£¤Ä¤Þ¤ê¡¢
.I arg
¤Ç»ØÄꤵ¤ì¤¿¥¤¥Ù¥ó¥È¤¬¤¹¤Ç¤Ë¥â¥Ë¥¿¤µ¤ì¤Æ¤¤¤ë
¥¤¥Ù¥ó¥È½¸¹ç¤Ë²Ã»»¤µ¤ì¤ë·Á¤Ë¤Ê¤ë¡£
.\"O To disable notification of all events, make an
.\"O .B F_NOTIFY
.\"O call specifying
.\"O .I arg
.\"O as 0.
¤¹¤Ù¤Æ¤Î¥¤¥Ù¥ó¥È¤ÎÄÌÃΤò̵¸ú¤Ë¤¹¤ë¤Ë¤Ï¡¢
.I arg
¤Ë 0 ¤ò»ØÄꤷ¤Æ
.B F_NOTIFY
¤ò¸Æ¤Ó½Ð¤¹É¬Íפ¬¤¢¤ë¡£

.\"O Notification occurs via delivery of a signal.
ÄÌÃΤϥ·¥°¥Ê¥ë¤ÎÇÛÁ÷¤Ç¹Ô¤ï¤ì¤ë¡£
.\"O The default signal is
.\"O .BR SIGIO ,
.\"O but this can be changed using the
.\"O .B F_SETSIG
.\"O command to
.\"O .BR fcntl ().
¥Ç¥Õ¥©¥ë¥È¤Î¥·¥°¥Ê¥ë¤Ï
.B SIGIO
¤À¤¬¡¢
.BR fcntl ()
¤Î
.B F_SETSIG
¥³¥Þ¥ó¥É¤ÇÊѹ¹¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
.\"O In the latter case, the signal handler receives a
.\"O .I siginfo_t
.\"O structure as its second argument (if the handler was
.\"O established using
.\"O .BR SA_SIGINFO )
.\"O and the
.\"O .I si_fd
.\"O field of this structure contains the file descriptor which
.\"O generated the notification (useful when establishing notification
.\"O on multiple directories).
¸å¼Ô¤Î¾ì¹ç¤Ë¤Ï¡¢
.RB ( SA_SIGINFO
¥Õ¥é¥°ÉÕ¤­¤Ç¥·¥°¥Ê¥ë¥Ï¥ó¥É¥é¤¬ÀßÄꤵ¤ì¤Æ¤¤¤ë¾ì¹ç¤Ë¤Ï)
¥Ï¥ó¥É¥é¤ÎÂèÆó°ú¤­¿ô¤È¤·¤Æ
.I siginfo_t
¹½Â¤ÂΤ¬ÅϤµ¤ì¡¢¤³¤Î¹½Â¤ÂΤÎ
.I si_fd
¥Õ¥£¡¼¥ë¥É¤Ë¤ÏÄÌÃΤιԤï¤ì¤¿¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤¬Æþ¤Ã¤Æ¤¤¤ë
(¤³¤Îµ¡Ç½¤ÏÊ£¿ô¤Î¥Ç¥£¥ì¥¯¥È¥ê¤ËÂФ·¤ÆÄÌÃΤòÀßÄꤹ¤ë¾ì¹ç¤ËÍ­ÍѤǤ¢¤ë)¡£

.\"O Especially when using
.\"O .BR DN_MULTISHOT ,
.\"O a real time signal should be used for notification,
.\"O so that multiple notifications can be queued.
ÆäË
.B DN_MULTISHOT
¤ò»È¤¦¾ì¹ç¤Ï¡¢ÄÌÃΤˤϥꥢ¥ë¥¿¥¤¥à¥·¥°¥Ê¥ë¤ò»È¤¦¤Ù¤­¤Ç¤¢¤ë¡£
¤½¤ì¤Ï¡¢¥ê¥¢¥ë¥¿¥¤¥à¥·¥°¥Ê¥ë¤ò»È¤¦¤³¤È¤Ç¡¢Ê£¿ô¤ÎÄÌÃΤò¥­¥å¡¼¤ËÆþ¤ì¤ë
¤³¤È¤¬¤Ç¤­¤ë¤«¤é¤Ç¤¢¤ë¡£

.\"O .B NOTE:
.\"O New applications should use the
.\"O .I inotify
.\"O interface (available since kernel 2.6.13),
.\"O which provides a much superior interface for obtaining notifications of
.\"O file system events.
.\"O See
.\"O .BR inotify (7).
.B Ãí°Õ:
¿·¤·¤¯¥¢¥×¥ê¥±¡¼¥·¥ç¥ó¤ò½ñ¤¯ºÝ¤Ë¤Ï¡¢(¥«¡¼¥Í¥ë 2.6.13 °Ê¹ß¤ÇÍøÍѲÄǽ¤È¤Ê¤Ã¤¿)
.I inotify
¥¤¥ó¥¿¥Õ¥§¡¼¥¹¤ò»ÈÍѤ¹¤Ù¤­¤Ç¤¢¤ë¡£
.I inotify
¤Ï¥Õ¥¡¥¤¥ë¥·¥¹¥Æ¥à¥¤¥Ù¥ó¥È¤ÎÄÌÃΤò¼èÆÀ¤¹¤ë¤¿¤á¤Î
¤º¤Ã¤ÈÍ¥¤ì¤¿¥¤¥ó¥¿¥Õ¥§¡¼¥¹¤Ç¤¢¤ë¡£
.BR inotify (7)
¤ò»²¾È¡£
.\"O .SH "RETURN VALUE"
.SH ÊÖ¤êÃÍ
.\"O For a successful call, the return value depends on the operation:
À®¸ù¤·¤¿¾ì¹ç¤ÎÊÖ¤êÃͤÏÁàºî¤Î¼ïÎà¤Ë¤è¤ê°ã¤¦:
.TP 0.9i
.B F_DUPFD
.\"O The new descriptor.
¿·¤·¤¤¥Ç¥£¥¹¥¯¥ê¥×¥¿¤òÊÖ¤¹¡£
.TP
.B F_GETFD
.\"O Value of flags.
¥Õ¥é¥°¤ÎÃͤòÊÖ¤¹¡£
.TP
.B F_GETFL
.\"O Value of flags.
¥Õ¥é¥°¤ÎÃͤòÊÖ¤¹¡£
.TP
.B F_GETLEASE
.\"O Type of lease held on file descriptor.
¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤ËÂФ·¤ÆÊÝ»ý¤µ¤ì¤Æ¤¤¤ë¥ê¡¼¥¹¤Î¼ïÊ̤òÊÖ¤¹¡£
.TP
.B F_GETOWN
.\"O Value of descriptor owner.
¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Î½êÍ­¼Ô¤òÊÖ¤¹¡£
.TP
.B F_GETSIG
.\"O Value of signal sent when read or write becomes possible, or zero
.\"O for traditional
.\"O .B SIGIO
.\"O behavior.
Æɤ߹þ¤ß¤ä½ñ¤­½Ð¤·¤¬²Äǽ¤Ë¤Ê¤Ã¤¿»þ¤ËÁ÷¤é¤ì¤ë¥·¥°¥Ê¥ë¤ÎÃÍ¡¢¤â¤·¤¯¤Ï
ÅÁÅýŪ¤Ê
.B SIGIO
Æ°ºî¤Î¾ì¹ç¤Ë¤Ï¥¼¥í¤òÊÖ¤¹¡£
.TP
.\"O All other commands
.\"O Zero.
¾¤ÎÁ´¤Æ¤Î¥³¥Þ¥ó¥É¤Ï 0 ¤òÊÖ¤¹¡£
.PP
.\"O On error, \-1 is returned, and
.\"O .I errno
.\"O is set appropriately.
¥¨¥é¡¼¤Î»þ¤Ï \-1 ¤¬ÊÖ¤µ¤ì¡¢
.I errno
¤ËŬÀÚ¤ÊÃͤ¬ÀßÄꤵ¤ì¤ë¡£
.\"O .SH ERRORS
.SH ¥¨¥é¡¼
.TP
.\"O .BR EACCES " or " EAGAIN
.BR EACCES " ¤« " EAGAIN
.\"O Operation is prohibited by locks held by other processes.
¾¤Î¥×¥í¥»¥¹¤¬ÊÝ»ý¤·¤Æ¤¤¤ë¥í¥Ã¥¯¤Ë¤è¤Ã¤ÆÁàºî¤¬¶Ø»ß¤µ¤ì¤Æ¤¤¤ë¡£
.TP
.B EAGAIN
.\"O The operation is prohibited because the file has been memory-mapped by
.\"O another process.
¤½¤Î¥Õ¥¡¥¤¥ë¤Ï¾¤Î¥×¥í¥»¥¹¤Ë¤è¤Ã¤Æ¥á¥â¥ê¡¦¥Þ¥Ã¥×¤µ¤ì¤Æ¤¤¤ë¤¿¤á¡¢
Áàºî¤¬¶Ø»ß¤µ¤ì¤Æ¤¤¤ë¡£
.TP
.B EBADF
.\"O .I fd
.\"O is not an open file descriptor, or the command was
.\"O .B F_SETLK
.\"O or
.\"O .B F_SETLKW
.\"O and the file descriptor open mode doesn't match with the
.\"O type of lock requested.
.I fd
¤¬¥ª¡¼¥×¥ó¤µ¤ì¤¿¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Ç¤Ê¤¤¡£
¤¢¤ë¤¤¤Ï¥³¥Þ¥ó¥É¤¬
.B F_SETLK
¤Þ¤¿¤Ï
.B F_SETLKW
¤À¤Ã¤¿¤¬¡¢ÂоݤΥե¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Î¥ª¡¼¥×¥ó¥â¡¼¥É¤¬
ɬÍפȤʤë¥í¥Ã¥¯¤Î·¿¤Ë¥Þ¥Ã¥Á¤·¤Æ¤¤¤Ê¤¤¡£
.TP
.\"O .B EDEADLK
.\"O It was detected that the specified
.\"O .B F_SETLKW
.\"O command would cause a deadlock.
.B EDEADLK
»ØÄꤵ¤ì¤¿
.B F_SETLKW
¥³¥Þ¥ó¥É¤ò¼Â¹Ô¤·¤¿¾ì¹ç¤Ë¤Ï¥Ç¥Ã¥É¥í¥Ã¥¯¤Ë¤Ê¤ë¤³¤È¤¬¸¡½Ð¤µ¤ì¤¿¡£
.TP
.B EFAULT
.\"O .I lock
.\"O is outside your accessible address space.
.I lock
¤¬ÍøÍѲÄǽ¤Ê¥¢¥É¥ì¥¹¶õ´Ö¤Î³°Éô¤Ë¤¢¤ë¡£
.TP
.B EINTR
.\"O For
.\"O .BR F_SETLKW ,
.\"O the command was interrupted by a signal; see
.\"O .BR signal (7).
.\"O For
.\"O .BR F_GETLK " and " F_SETLK ,
.\"O the command was interrupted by a signal before the lock was checked or
.\"O acquired.
.\"O Most likely when locking a remote file (e.g., locking over
.\"O NFS), but can sometimes happen locally.
.B F_SETLKW
¥³¥Þ¥ó¥É¤¬¥·¥°¥Ê¥ë¤Ë¤è¤ê³ä¤ê¹þ¤Þ¤ì¤¿
.RB ( signal (7)
»²¾È)¡£
.BR F_GETLK " ¤È " F_SETLK
¤Î¾ì¹ç¡¢¥í¥Ã¥¯¤ò³Îǧ¤·¤¿¤ê¼èÆÀ¤·¤¿¤ê¤¹¤ëÁ°¤Ë¥·¥°¥Ê¥ë¤Ë¤è¤Ã¤Æ³ä¤ê¹þ¤Þ¤ì¤¿¡£
¤³¤ì¤Ï¤¿¤¤¤Æ¤¤¥ê¥â¡¼¥È¤Î¥Õ¥¡¥¤¥ë¤ò¥í¥Ã¥¯¤¹¤ë¾ì¹ç
(Î㤨¤Ð NFS ¾å¤Ç¥í¥Ã¥¯¤¹¤ë¾ì¹ç) ¤Ëµ¯¤³¤ë¡£
¤·¤«¤·¥í¡¼¥«¥ë¤Ç¤âµ¯¤³¤ë¾ì¹ç¤¬¤¢¤ë¡£
.TP
.B EINVAL
.\"O For
.\"O .BR F_DUPFD ,
.\"O .I arg
.\"O is negative or is greater than the maximum allowable value.
.\"O For
.\"O .BR F_SETSIG ,
.\"O .I arg
.\"O is not an allowable signal number.
.BR F_DUPFD ¤Ç¡¢
.I arg
¤¬É餫¡¢¤â¤·¤¯¤Ïµö¤µ¤ì¤ëºÇÂçÃͤè¤ê¤âÂ礭¤¤¡£
.B F_SETSIG
¤Î¾ì¹ç¡¢
.I arg
¤¬ÍøÍѲÄǽ¤Ê¥·¥°¥Ê¥ëÈÖ¹æ¤Ç¤Ï¤Ê¤¤¡£
.TP
.B EMFILE
.\"O For
.\"O .BR F_DUPFD ,
.\"O the process already has the maximum number of file descriptors open.
.BR F_DUPFD ¤Ç¡¢
¥×¥í¥»¥¹¤¬¤¹¤Ç¤ËºÇÂç¿ô¤Þ¤Ç¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤ò¥ª¡¼¥×¥ó¤·¤Æ¤¤¤ë¡£
.TP
.B ENOLCK
.\"O Too many segment locks open, lock table is full, or a remote locking
.\"O protocol failed (e.g., locking over NFS).
¥ª¡¼¥×¥ó¤µ¤ì¤Æ¤¤¤ë¥í¥Ã¥¯¤Î¿ô¤¬Â¿²á¤®¤Æ¡¢¥í¥Ã¥¯¡¦¥Æ¡¼¥Ö¥ë¤¬¤¤¤Ã¤Ñ¤¤¤Ç¤¢¤ë¡£
¤Þ¤¿¤Ï remote locking protocol (Î㤨¤Ð NFS ¾å¤Î¥í¥Ã¥¯) ¤¬¼ºÇÔ¤·¤¿¡£
.TP
.B EPERM
.\"O Attempted to clear the
.\"O .B O_APPEND
.\"O flag on a file that has the append-only attribute set.
ÄɲÃÀìÍÑ°À­¤¬ÀßÄꤵ¤ì¤¿¥Õ¥¡¥¤¥ë¤Î
.B O_APPEND
¥Õ¥é¥°¤ò¥¯¥ê¥¢¤·¤è¤¦¤È»î¤ß¤¿¡£
.\"O .SH "CONFORMING TO"
.SH ½àµò
.\"O SVr4, 4.3BSD, POSIX.1-2001.
SVr4, 4.3BSD, POSIX.1-2001.
.\"O Only the operations
.\"O .BR F_DUPFD ,
.\"O .BR F_GETFD ,
.\"O .BR F_SETFD ,
.\"O .BR F_GETFL ,
.\"O .BR F_SETFL ,
.\"O .BR F_GETLK ,
.\"O .BR F_SETLK ,
.\"O .BR F_SETLKW ,
.\"O .BR F_GETOWN ,
.\"O and
.\"O .B F_SETOWN
.\"O are specified in POSIX.1-2001.
POSIX.1-2001 ¤Çµ¬Äꤵ¤ì¤Æ¤¤¤ëÁàºî¤Ï¡¢
.BR F_DUPFD ,
.BR F_GETFD ,
.BR F_SETFD ,
.BR F_GETFL ,
.BR F_SETFL ,
.BR F_GETLK ,
.BR F_SETLK ,
.BR F_SETLKW ,
.BR F_GETOWN ,
.B F_SETOWN
¤À¤±¤Ç¤¢¤ë¡£

.\"O .B F_DUPFD_CLOEXEC
.\"O is specified in POSIX.1-2008.
.B F_DUPFD_CLOEXEC
¤Ï POSIX.1-2008 ¤Çµ¬Äꤵ¤ì¤Æ¤¤¤ë¡£

.\"O .BR F_GETOWN_EX ,
.\"O .BR F_SETOWN_EX ,
.\"O .BR F_GETSIG ,
.\"O .BR F_SETSIG ,
.\"O .BR F_NOTIFY ,
.\"O .BR F_GETLEASE ,
.\"O and
.\"O .B F_SETLEASE
.\"O are Linux-specific.
.\"O (Define the
.\"O .B _GNU_SOURCE
.\"O macro to obtain these definitions.)
.BR F_GETOWN_EX ,
.BR F_SETOWN_EX ,
.BR F_GETSIG ,
.BR F_SETSIG ,
.BR F_NOTIFY ,
.BR F_GETLEASE ,
.B F_SETLEASE
¤Ï Linux ¸ÇÍ­¤Ç¤¢¤ë (¤³¤ì¤é¤ÎÄêµÁ¤òÍ­¸ú¤Ë¤¹¤ë¤Ë¤Ï
.B _GNU_SOURCE
¥Þ¥¯¥í¤òÄêµÁ¤¹¤ë¤³¤È)¡£
.\" .PP
.\"O .\" SVr4 documents additional EIO, ENOLINK and EOVERFLOW error conditions.
.\" SVr4 ¤Ë¤Ï¾¤Ë EFAULT, EINTR, EIO, ENOLINK, EOVERFLOW ¥¨¥é¡¼¾õÂ֤ˤĤ¤¤Æ¤Î
.\" µ­½Ò¤¬¤¢¤ë¡£
.\"O .SH NOTES
.SH Ãí°Õ
.\"O The errors returned by
.\"O .BR dup2 (2)
.\"O are different from those returned by
.\"O .BR F_DUPFD .
¥¨¥é¡¼¤ÎºÝ¤ÎÊÖ¤êÃͤ¬
.BR dup2 (2)
¤È
.B F_DUPFD
¤Ç¤Ï°Û¤Ê¤Ã¤Æ¤¤¤ë¡£

.\"O Since kernel 2.0, there is no interaction between the types of lock
.\"O placed by
.\"O .BR flock (2)
.\"O and
.\"O .BR fcntl ().
¥«¡¼¥Í¥ë 2.0 °Ê¹ß¤Ç¤Ï¡¢
.BR flock (2)
¤È
.BR fcntl ()
¤¬ÀßÄꤹ¤ë¥í¥Ã¥¯¼ïÊ̤δ֤ËÁê¸ßºîÍѤϤʤ¤¡£

.\"O Several systems have more fields in
.\"O .I "struct flock"
.\"O such as, for example,
.\"O .IR l_sysid .
.\"O .\" e.g., Solaris 8 documents this field in fcntl(2), and Irix 6.5
.\"O .\" documents it in fcntl(5).  mtk, May 2007
.\"O Clearly,
.\"O .I l_pid
.\"O alone is not going to be very useful if the process holding the lock
.\"O may live on a different machine.
¥·¥¹¥Æ¥à¤Ë¤è¤Ã¤Æ¤Ï¡¢
.I "struct flock"
¤Ë¾åµ­°Ê³°¤Î¥Õ¥£¡¼¥ë¥É¤¬¤¢¤ë¤â¤Î¤â¤¢¤ë (Î㤨¤Ð
.IR l_sysid )¡£
.\" e.g., Solaris 8 documents this field in fcntl(2), and Irix 6.5
.\" documents it in fcntl(5).  mtk, May 2007
¤Ï¤Ã¤­¤ê¤È¸À¤¨¤ë¤³¤È¤Ï¡¢¥í¥Ã¥¯¤òÊÝ»ý¤·¤Æ¤¤¤ë¥×¥í¥»¥¹¤¬Ê̤Υޥ·¥ó¤Ë¸ºß
¤¹¤ë¾ì¹ç¤Ë¤Ï¡¢
.I l_pid
¤À¤±¤Ï¤¢¤Þ¤êÌò¤Ë¤¿¤¿¤Ê¤¤¤À¤í¤¦¤È¤¤¤¦¤³¤È¤Ç¤¢¤ë¡£
.\"O .SH BUGS
.SH ¥Ð¥°
.\"O A limitation of the Linux system call conventions on some
.\"O architectures (notably i386) means that if a (negative)
.\"O process group ID to be returned by
.\"O .B F_GETOWN
.\"O falls in the range \-1 to \-4095, then the return value is wrongly
.\"O interpreted by glibc as an error in the system call;
.\"O .\" glibc source: sysdeps/unix/sysv/linux/i386/sysdep.h
.\"O that is, the return value of
.\"O .BR fcntl ()
.\"O will be \-1, and
.\"O .I errno
.\"O will contain the (positive) process group ID.
¤¤¤¯¤Ä¤«¤Î¥¢¡¼¥­¥Æ¥¯¥Á¥ã (ÆÃ¤Ë i386) ¤Ë¤ª¤±¤ë Linux ¥·¥¹¥Æ¥à¥³¡¼¥ë¤Î´·½¬
¤Î¤¿¤á°Ê²¼¤ÎÀ©¸Â¤¬Â¸ºß¤¹¤ë¡£
.B F_GETOWN
¤¬ÊÖ¤¹ (Éé¤Î) ¥×¥í¥»¥¹¥°¥ë¡¼¥×ID ¤¬ \-1 ¤«¤é \-4095 ¤ÎÈϰϤËÆþ¤Ã¤¿¾ì¹ç¡¢
glibc ¤Ï¤³¤ÎÊÖ¤êÃͤò¥·¥¹¥Æ¥à¥³¡¼¥ë¤Ç¥¨¥é¡¼¤¬µ¯¤³¤Ã¤¿¤È
´Ö°ã¤Ã¤Æ²ò¼á¤·¤Æ¤·¤Þ¤¦¡£¤Ä¤Þ¤ê¡¢
.BR fcntl ()
¤ÎÊÖ¤êÃÍ¤Ï \-1 ¤È¤Ê¤ê¡¢
.I errno
¤Ë¤Ï (Àµ¤Î) ¥×¥í¥»¥¹¥°¥ë¡¼¥×ID ¤¬ÀßÄꤵ¤ì¤ë¤³¤È¤Ë¤Ê¤ë¡£
.\"O The Linux-specific
.\"O .BR F_SETOWN_EX
.\"O and
.\"O .BR F_GETOWN_EX
.\"O operations avoid this problem.
Linux ¸ÇÍ­¤Î
.B F_SETOWN_EX
¤È
.B F_GETOWN_EX
¤Ç¤Ï¤³¤ÎÌäÂê¤ò²óÈò¤Ç¤­¤ë¡£
.\"O .\" mtk, Dec 04: some limited testing on alpha and ia64 seems to
.\"O .\" indicate that ANY negative PGID value will cause F_GETOWN
.\"O .\" to misinterpret the return as an error. Some other architectures
.\"O .\" seem to have the same range check as i386.
.\" mtk, Dec 04: alpha ¤È ia64 ¤Ç¤Ï¡¢¾¯¤·¥Æ¥¹¥È¤·¤Æ¤ß¤¿¸Â¤ê¡¢
.\" ¡Ö¤É¤ó¤Ê¡×Éé¤Î PGID ¤Ç¤â F_GETOWN ¤Ç¥¨¥é¡¼¤¬µ¯¤³¤Ã¤¿¤È´Ö°ã¤Ã¤Æ
.\" ²ò¼á¤µ¤ì¤Æ¤·¤Þ¤¦¤è¤¦¤À¡£Â¾¤Î¤¤¤¯¤Ä¤«¤Î¥¢¡¼¥­¥Æ¥¯¥Á¥ã¤Ç¤Ï¡¢
.\" i386 ¤ÈƱÍͤÎÈϰϤΥÁ¥§¥Ã¥¯¤¬¹Ô¤ï¤ì¤Æ¤¤¤ë¤è¤¦¤À¡£

.\"O In Linux 2.4 and earlier, there is bug that can occur
.\"O when an unprivileged process uses
.\"O .B F_SETOWN
.\"O to specify the owner
.\"O of a socket file descriptor
.\"O as a process (group) other than the caller.
.\"O In this case,
.\"O .BR fcntl ()
.\"O can return \-1 with
.\"O .I errno
.\"O set to
.\"O .BR EPERM ,
.\"O even when the owner process (group) is one that the caller
.\"O has permission to send signals to.
.\"O Despite this error return, the file descriptor owner is set,
.\"O and signals will be sent to the owner.
Linux 2.4 °ÊÁ°¤Ç¤Ï¡¢ÈóÆø¢¥×¥í¥»¥¹¤¬
.B F_SETOWN
¤ò»È¤Ã¤Æ¡¢¥½¥±¥Ã¥È¤Î¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Î½êÍ­¼Ô¤Ë
¸Æ¤Ó½Ð¤·¸µ°Ê³°¤Î¥×¥í¥»¥¹ (¤ä¥×¥í¥»¥¹¥°¥ë¡¼¥×) ¤ò»ØÄꤹ¤ë¤È
ȯÀ¸¤¹¤ë¥Ð¥°¤¬¤¢¤ë¡£¤³¤Î¾ì¹ç¡¢
¸Æ¤Ó½Ð¤·¸µ¤¬½êÍ­¼Ô¤È¤·¤Æ»ØÄꤷ¤¿¥×¥í¥»¥¹ (¤ä¥×¥í¥»¥¹¥°¥ë¡¼¥×) ¤Ë
¥·¥°¥Ê¥ë¤òÁ÷¤ëµö²Ä¤ò»ý¤Ã¤Æ¤¤¤¿¤È¤·¤Æ¤â¡¢
.BR fcntl ()
¤¬ \-1 ¤òÊÖ¤·
.I errno
¤Ë
.B EPERM
¤òÀßÄꤹ¤ë¤³¤È¤¬¤¢¤ë¡£
¤³¤Î¥¨¥é¡¼¤¬Ê֤俤ˤ⤫¤«¤ï¤é¤º¡¢¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Î½êÍ­¼Ô
¤ÏÀßÄꤵ¤ì¡¢¥·¥°¥Ê¥ë¤Ï¤½¤Î½êÍ­¼Ô¤ËÁ÷¤é¤ì¤ë¡£

.\"O The implementation of mandatory locking in all known versions of Linux
.\"O is subject to race conditions which render it unreliable:
.\"O .\" http://marc.info/?l=linux-kernel&m=119013491707153&w=2
.\"O a
.\"O .BR write (2)
.\"O call that overlaps with a lock may modify data after the mandatory lock is
.\"O acquired;
.\"O a
.\"O .BR read (2)
.\"O call that overlaps with a lock may detect changes to data that were made
.\"O only after a write lock was acquired.
.\"O Similar races exist between mandatory locks and
.\"O .BR mmap (2).
.\"O It is therefore inadvisable to rely on mandatory locking.
¤³¤ì¤Þ¤Ç¤Î Linux ¤ÎÁ´¤Æ¤Î¥Ð¡¼¥¸¥ç¥ó¤Ë¤ª¤±¤ë¶¯À©¥í¥Ã¥¯¤Î¼ÂÁõ¤Ï¡¢
¶¥¹ç¾ò·ï²¼¤Ç¶¯À©¥í¥Ã¥¯¤¬ÉÔ´°Á´¤Ë¤Ê¤ë¤è¤¦¤Ê¾ì¹ç¤¬¤¢¤ë¡£
.\" http://marc.info/?l=linux-kernel&m=119013491707153&w=2
¥í¥Ã¥¯¤È½Å¤Ê¤Ã¤Æ¼Â¹Ô¤µ¤ì¤¿
.BR write (2)
¤Î¸Æ¤Ó½Ð¤·¤Ï¶¯À©¥í¥Ã¥¯¤¬³ÍÆÀ¤µ¤ì¤¿¸å¤Ë¤â¥Ç¡¼¥¿¤òÊѹ¹¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
¥í¥Ã¥¯¤È½Å¤Ê¤Ã¤Æ¼Â¹Ô¤µ¤ì¤¿
.BR read (2)
¤Î¸Æ¤Ó½Ð¤·¤Ï¶¯À©¥í¥Ã¥¯¤¬³ÍÆÀ¤µ¤ì¤¿¸å¤Ë¤Ê¤Ã¤Æ¹Ô¤ï¤ì¤¿¥Ç¡¼¥¿¤ÎÊѹ¹¤ò
¸¡½Ð¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
ƱÍͤζ¥¹ç¾ò·ï¤¬¶¯À©¥í¥Ã¥¯¤È
.BR mmap (2)
¤Î´Ö¤Ë¤â¸ºß¤¹¤ë¡£¤½¤ì¤æ¤¨¡¢¶¯À©¥í¥Ã¥¯¤ËÍê¤ë¤Î¤Ï¤ªÁ¦¤á¤Ç¤­¤Ê¤¤¡£
.\"O .SH "SEE ALSO"
.SH ´ØÏ¢¹àÌÜ
.BR dup2 (2),
.BR flock (2),
.BR open (2),
.BR socket (2),
.BR lockf (3),
.BR capabilities (7),
.BR feature_test_macros (7)
.P
.\"O See also
.\"O .IR locks.txt ,
.\"O .IR mandatory-locking.txt ,
.\"O and
.\"O .I dnotify.txt
.\"O in the kernel source directory
.\"O .IR Documentation/filesystems/ .
.\"O (On older kernels, these files are directly under the
.\"O .I Documentation/
.\"O directory, and
.\"O .I mandatory-locking.txt
.\"O is called
.\"O .IR mandatory.txt .)
¥«¡¼¥Í¥ë¥½¡¼¥¹¤Î
.IR Documentation/filesystems/
¥Ç¥£¥ì¥¯¥È¥êÆâ¤Î
.IR locks.txt ,
.IR mandatory-locking.txt ,
.I dnotify.txt
¤â»²¾È¤Î¤³¤È¡£
(°ÊÁ°¤Î¥«¡¼¥Í¥ë¤Ç¤Ï¡¢¤³¤ì¤é¤Î¥Õ¥¡¥¤¥ë¤Ï
.I Documentation/
¥Ç¥£¥ì¥¯¥È¥êľ²¼¤Ë¤¢¤ê¡¢
.I mandatory-locking.txt
¤Ï
.I mandatory.txt
¤È¤¤¤¦Ì¾Á°¤Ç¤¢¤Ã¤¿¡£)