(split) LDP v3.24 -> v3.29 の定型的な変更内容を反映。

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

.\" Hey Emacs! This file is -*- nroff -*- source.
.\"
.\" Copyright (c) 1992 Drew Eckhardt <drew@cs.colorado.edu>, March 28, 1992
.\" and Copyright (c) Michael Kerrisk, 2001, 2002, 2005
.\" May be distributed under the GNU General Public License.
.\" Modified by Michael Haardt <michael@moria.de>
.\" Modified 24 Jul 1993 by Rik Faith <faith@cs.unc.edu>
.\" Modified 21 Aug 1994 by Michael Chastain <mec@shell.portal.com>:
.\"   New man page (copied from 'fork.2').
.\" Modified 10 June 1995 by Andries Brouwer <aeb@cwi.nl>
.\" Modified 25 April 1998 by Xavier Leroy <Xavier.Leroy@inria.fr>
.\" Modified 26 Jun 2001 by Michael Kerrisk
.\"     Mostly upgraded to 2.4.x
.\"     Added prototype for sys_clone() plus description
.\"     Added CLONE_THREAD with a brief description of thread groups
.\"     Added CLONE_PARENT and revised entire page remove ambiguity
.\"             between "calling process" and "parent process"
.\"     Added CLONE_PTRACE and CLONE_VFORK
.\"     Added EPERM and EINVAL error codes
.\"     Renamed "__clone" to "clone" (which is the prototype in <sched.h>)
.\"     various other minor tidy ups and clarifications.
.\" Modified 26 Jun 2001 by Michael Kerrisk <mtk.manpages@gmail.com>
.\"     Updated notes for 2.4.7+ behavior of CLONE_THREAD
.\" Modified 15 Oct 2002 by Michael Kerrisk <mtk.manpages@gmail.com>
.\"     Added description for CLONE_NEWNS, which was added in 2.4.19
.\" Slightly rephrased, aeb.
.\" Modified 1 Feb 2003 - added CLONE_SIGHAND restriction, aeb.
.\" Modified 1 Jan 2004 - various updates, aeb
.\" Modified 2004-09-10 - added CLONE_PARENT_SETTID etc. - aeb.
.\" 2005-04-12, mtk, noted the PID caching behavior of NPTL's getpid()
.\"     wrapper under BUGS.
.\" 2005-05-10, mtk, added CLONE_SYSVSEM, CLONE_UNTRACED, CLONE_STOPPED.
.\" 2005-05-17, mtk, Substantially enhanced discussion of CLONE_THREAD.
.\" 2008-11-18, mtk, order CLONE_* flags  alphabetically
.\" 2008-11-18, mtk, document CLONE_NEWPID
.\" 2008-11-19, mtk, document CLONE_NEWUTS
.\" 2008-11-19, mtk, document CLONE_NEWIPC
.\" 2008-11-19, Jens Axboe, mtk, document CLONE_IO
.\"
.\" FIXME Document CLONE_NEWUSER, which is new in 2.6.23
.\"       (also supported for unshare()?)
.\" FIXME . 2.6.25 marks the unused CLONE_STOPPED as obsolete, and it will
.\"       probably be removed in the future.
.\"
.\" Japanese Version Copyright (c) 2001 HANATAKA Shinya
.\"     and Copyright(c) 2002, 2005-2008 Akihiro MOTOKI
.\" Translated 2001-08-17, HANATAKA Shinya <hanataka@abyss.rim.or.jp>
.\" Modified 2002-09-24, Akihiro MOTOKI <amotoki@dd.iij4u.or.jp>
.\" Modified 2005-02-02, Akihiro MOTOKI
.\" Updated 2005-04-17, Akihiro MOTOKI
.\" Updated 2005-09-10, Akihiro MOTOKI
.\" Updated 2006-03-05, Akihiro MOTOKI, LDP v2.25
.\" Updated 2007-01-05, Akihiro MOTOKI, LDP v2.43
.\" Updated 2007-05-01, Akihiro MOTOKI, LDP v2.46
.\" Updated 2007-06-13, Akihiro MOTOKI, LDP v2.55
.\" Updated 2008-08-04, Akihiro MOTOKI, LDP v3.05
.\" Updated 2008-11-09, Akihiro MOTOKI, LDP v3.10
.\" Updated 2009-03-02, Akihiro MOTOKI, LDP v3.19
.\" Updated 2010-04-11, Akihiro MOTOKI, LDP v3.24
.\"
.\"WORD:        capability      ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£
.\"WORD:        pending signals ½èÍýÂÔ¤Á¥·¥°¥Ê¥ë
.\"WORD:        namespace       ̾Á°¶õ´Ö
.\"
.TH CLONE 2 2010-09-10 "Linux" "Linux Programmer's Manual"
.\"O .SH NAME
.\"O clone, __clone2 \- create a child process
.SH ̾Á°
clone, __clone2 \- »Ò¥×¥í¥»¥¹¤òºîÀ®¤¹¤ë
.\"O .SH SYNOPSIS
.SH ½ñ¼°
.nf
.\"O .BR "#define _GNU_SOURCE" "             /* See feature_test_macros(7) */"
.BR "#define _GNU_SOURCE" "             /* feature_test_macros(7) »²¾È */"
.\" Actually _BSD_SOURCE || _SVID_SOURCE
.\" See http://sources.redhat.com/bugzilla/show_bug.cgi?id=4749
.B #include <sched.h>

.BI "int clone(int (*" "fn" ")(void *), void *" child_stack ,
.BI "          int " flags ", void *" "arg" ", ... "
.BI "          /* pid_t *" ptid ", struct user_desc *" tls \
", pid_t *" ctid " */ );"
.fi
.\"O .SH DESCRIPTION
.SH ÀâÌÀ
.\"O .BR clone ()
.\"O creates a new process, in a manner similar to
.\"O .BR fork (2).
.\"O It is actually a library function layered on top of the underlying
.\"O .BR clone ()
.\"O system call, hereinafter referred to as
.\"O .BR sys_clone .
.\"O A description of
.\"O .B sys_clone
.\"O is given toward the end of this page.
.BR clone ()
¤Ï
.BR fork (2)
¤ÈƱ¤¸¤è¤¦¤ÊÊýË¡¤Ç¿·¤·¤¤¥×¥í¥»¥¹¤òºîÀ®¤¹¤ë¡£
.BR clone ()
¤Ë¤Ï¡¢¥é¥¤¥Ö¥é¥ê´Ø¿ô¤È¤½¤Î²¼Áؤˤ¢¤¿¤ë
.BR clone ()
¥·¥¹¥Æ¥à¥³¡¼¥ë¤¬Â¸ºß¤¹¤ë¡£°Ê²¼¤ÎÀâÌÀ¤Ç¤Ï¡¢¥·¥¹¥Æ¥à¥³¡¼¥ë¤ÎÊý¤ò
.B sys_clone
¤Èɽ¤¹¤³¤È¤È¤¹¤ë¡£
.B sys_clone
¤Ë´Ø¤¹¤ëÀâÌÀ¤Ï¤³¤Î¥Þ¥Ë¥å¥¢¥ë¤ÎºÇ¸å¤ÎÊý¤Ë¤¢¤ë¡£

.\"O Unlike
.\"O .BR fork (2),
.\"O these calls
.\"O allow the child process to share parts of its execution context with
.\"O the calling process, such as the memory space, the table of file
.\"O descriptors, and the table of signal handlers.
.\"O (Note that on this manual
.\"O page, "calling process" normally corresponds to "parent process".
.\"O But see the description of
.\"O .B CLONE_PARENT
.\"O below.)
.BR fork (2)
¤È¤Ï°Û¤Ê¤ê¡¢¤³¤ì¤é¤Î¥³¡¼¥ë¤Ç¤Ï¡¢»Ò¥×¥í¥»¥¹ (child process)
¤È¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤È¤¬¡¢¥á¥â¥ê¶õ´Ö¡¢
¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Î¥Æ¡¼¥Ö¥ë¡¢¥·¥°¥Ê¥ë¡¦¥Ï¥ó¥É¥é¤Î¥Æ¡¼¥Ö¥ë¤Ê¤É¤Î
¼Â¹Ô¥³¥ó¥Æ¥­¥¹¥È¤Î°ìÉô¤ò¶¦Í­¤Ç¤­¤ë¡£
(¤³¤Î¥Þ¥Ë¥å¥¢¥ë¤Ë¤ª¤±¤ë¡Ö¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¡×¤Ï¡¢Ä̾ï¤Ï
¡Ö¿Æ¥×¥í¥»¥¹¡×¤È°ìÃפ¹¤ë¡£Ã¢¤·¡¢¸å½Ò¤Î
.B CLONE_PARENT
¤Î¹à¤â»²¾È¤Î¤³¤È)

.\"O The main use of
.\"O .BR clone ()
.\"O is to implement threads: multiple threads of control in a program that
.\"O run concurrently in a shared memory space.
.BR clone ()
¤Î¼çÍפʻÈÍÑË¡¤Ï¥¹¥ì¥Ã¥É (threads) ¤ò¼ÂÁõ¤¹¤ë¤³¤È¤Ç¤¢¤ë:
°ì¤Ä¤Î¥×¥í¥°¥é¥à¤ÎÃæ¤ÎÊ£¿ô¤Î¥¹¥ì¥Ã¥É¤Ï¶¦Í­¤µ¤ì¤¿¥á¥â¥ê¶õ´Ö¤Ç
Ʊ»þ¤Ë¼Â¹Ô¤µ¤ì¤ë¡£

.\"O When the child process is created with
.\"O .BR clone (),
.\"O it executes the function
.\"O application
.\"O .IR fn ( arg ).
.\"O (This differs from
.\"O .BR fork (2),
.\"O where execution continues in the child from the point
.\"O of the
.\"O .BR fork (2)
.\"O call.)
.\"O The
.\"O .I fn
.\"O argument is a pointer to a function that is called by the child
.\"O process at the beginning of its execution.
.\"O The
.\"O .I arg
.\"O argument is passed to the
.\"O .I fn
.\"O function.
.BR clone ()
¤Ç»Ò¥×¥í¥»¥¹¤¬ºîÀ®¤µ¤ì¤¿»þ¤Ë¡¢ºîÀ®¤µ¤ì¤¿»Ò¥×¥í¥»¥¹¤Ï´Ø¿ô
.IR fn ( arg )
¤ò¼Â¹Ô¤¹¤ë¡£
(¤³¤ÎÅÀ¤¬
.BR fork (2)
¤È¤Ï°Û¤Ê¤ë¡£
.BR fork (2)
¤Î¾ì¹ç¡¢»Ò¥×¥í¥»¥¹¤Ï
.BR fork (2)
¤¬¸Æ¤Ó½Ð¤µ¤ì¤¿¾ì½ê¤«¤é¼Â¹Ô¤ò³¤±¤ë¡£)
.I fn
°ú¤­¿ô¤Ï¡¢»Ò¥×¥í¥»¥¹¤¬¼Â¹Ô¤ò»Ï¤á¤ë»þ¤Ë»Ò¥×¥í¥»¥¹¤¬¸Æ¤Ó½Ð¤¹
´Ø¿ô¤Ø¤Î¥Ý¥¤¥ó¥¿¤Ç¤¢¤ë¡£
.I arg
°ú¤­¿ô¤Ï¤½¤Î¤Þ¤Þ
.I fn
´Ø¿ô¤Ø¤ÈÅϤµ¤ì¤ë¡£

.\"O When the
.\"O .IR fn ( arg )
.\"O function application returns, the child process terminates.
.\"O The integer returned by
.\"O .I fn
.\"O is the exit code for the child process.
.\"O The child process may also terminate explicitly by calling
.\"O .BR exit (2)
.\"O or after receiving a fatal signal.
.IR fn ( arg )
´Ø¿ô¤¬½ªÎ»¤¹¤ë¤È¡¢»Ò¥×¥í¥»¥¹¤Ï½ªÎ»¤¹¤ë¡£
.I fn
¤Ë¤è¤Ã¤ÆÊÖ¤µ¤ì¤¿À°¿ô¤¬»Ò¥×¥í¥»¥¹¤Î½ªÎ»¥³¡¼¥É¤È¤Ê¤ë¡£
»Ò¥×¥í¥»¥¹¤Ï¡¢
.BR exit (2)
¤ò¸Æ¤ó¤ÇÌÀ¼¨Åª¤Ë½ªÎ»¤¹¤ë¤³¤È¤â¤¢¤ë¤·¡¢Ã×̿Ū¤Ê¥·¥°¥Ê¥ë¤ò¼õ¿®¤·¤¿
¾ì¹ç¤Ë½ªÎ»¤¹¤ë¤³¤È¤â¤¢¤ë¡£

.\"O The
.\"O .I child_stack
.\"O argument specifies the location of the stack used by the child process.
.\"O Since the child and calling process may share memory,
.\"O it is not possible for the child process to execute in the
.\"O same stack as the calling process.
.\"O The calling process must therefore
.\"O set up memory space for the child stack and pass a pointer to this
.\"O space to
.\"O .BR clone ().
.I child_stack
°ú¤­¿ô¤Ï¡¢»Ò¥×¥í¥»¥¹¤Ë¤è¤Ã¤Æ»ÈÍѤµ¤ì¤ë¥¹¥¿¥Ã¥¯¤Î°ÌÃÖ¤ò»ØÄꤹ¤ë¡£
»Ò¥×¥í¥»¥¹¤È¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤Ï¥á¥â¥ê¤ò¶¦Í­¤¹¤ë¤³¤È¤¬¤¢¤ë¤¿¤á¡¢
»Ò¥×¥í¥»¥¹¤Ï¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤ÈƱ¤¸¥¹¥¿¥Ã¥¯¤Ç¼Â¹Ô¤¹¤ë¤³¤È¤¬¤Ç¤­¤Ê¤¤¡£
¤³¤Î¤¿¤á¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤Ï»Ò¥×¥í¥»¥¹¤Î¥¹¥¿¥Ã¥¯¤Î¤¿¤á¤Î¥á¥â¥ê¶õ´Ö¤ò
ÍÑ°Õ¤·¤Æ¡¢¤³¤Î¶õ´Ö¤Ø¤Î¥Ý¥¤¥ó¥¿¤ò
.BR clone ()
¤ØÅϤµ¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£
.\"O Stacks grow downward on all processors that run Linux
.\"O (except the HP PA processors), so
.\"O .I child_stack
.\"O usually points to the topmost address of the memory space set up for
.\"O the child stack.
(HP PA ¥×¥í¥»¥Ã¥µ°Ê³°¤Î) Linux ¤¬Æ°ºî¤¹¤ëÁ´¤Æ¤Î¥×¥í¥»¥Ã¥µ¤Ç¤Ï¡¢
¥¹¥¿¥Ã¥¯¤Ï²¼Êý (¥¢¥É¥ì¥¹¤¬¾®¤µ¤¤Êý¸þ) ¤Ø¤È¿­¤Ó¤ë¡£¤³¤Î¤¿¤á¡¢ÉáÄ̤Ï
.I child_stack
¤Ï»Ò¥×¥í¥»¥¹¤Î¥¹¥¿¥Ã¥¯¤Î¤¿¤á¤ËÍÑ°Õ¤·¤¿¥á¥â¥ê¶õ´Ö¤Î°ìÈÖÂ礭¤¤
¥¢¥É¥ì¥¹¤ò»Ø¤¹¤è¤¦¤Ë¤¹¤ë¡£

.\"O The low byte of
.\"O .I flags
.\"O contains the number of the
.\"O .I "termination signal"
.\"O sent to the parent when the child dies.
.\"O If this signal is specified as anything other than
.\"O .BR SIGCHLD ,
.\"O then the parent process must specify the
.\"O .B __WALL
.\"O or
.\"O .B __WCLONE
.\"O options when waiting for the child with
.\"O .BR wait (2).
.\"O If no signal is specified, then the parent process is not signaled
.\"O when the child terminates.
.I flags
¤Î²¼°Ì 1 ¥Ð¥¤¥È¤Ï»Ò¥×¥í¥»¥¹¤¬»à¤ó¤À¾ì¹ç¤Ë¿Æ¥×¥í¥»¥¹¤Ø¤ÈÁ÷¤é¤ì¤ë
.I "½ªÎ»¥·¥°¥Ê¥ë (termination signal)"
¤ÎÈÖ¹æ¤ò»ØÄꤹ¤ë¡£¤³¤Î¥·¥°¥Ê¥ë¤È¤·¤Æ
.B SIGCHLD
°Ê³°¤¬»ØÄꤵ¤ì¤¿¾ì¹ç¡¢¿Æ¥×¥í¥»¥¹¤Ï¡¢
.BR wait (2)
¤Ç»Ò¥×¥í¥»¥¹¤òÂԤĺݤˡ¢¥ª¥×¥·¥ç¥ó¤È¤·¤Æ
.B __WALL
¤Þ¤¿¤Ï
.B __WCLONE
¤ò»ØÄꤷ¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£
¤É¤Î¥·¥°¥Ê¥ë¤â»ØÄꤵ¤ì¤Ê¤«¤Ã¤¿¾ì¹ç¡¢»Ò¥×¥í¥»¥¹¤¬½ªÎ»¤·¤¿»þ¤Ë¿Æ¥×¥í¥»¥¹
¤Ë¥·¥°¥Ê¥ë¤ÏÁ÷¤é¤ì¤Ê¤¤¡£

.\"O .I flags
.\"O may also be bitwise-or'ed with zero or more of the following constants,
.\"O in order to specify what is shared between the calling process
.\"O and the child process:
.I flags
¤Ë¤Ï¡¢°Ê²¼¤ÎÄê¿ô¤Î¤¦¤Á 0¸Ä°Ê¾å¤ò¥Ó¥Ã¥ÈËè¤ÎÏÀÍýÏ (bitwise-or)
¤ò¤È¤Ã¤¿¤â¤Î¤ò»ØÄê¤Ç¤­¤ë¡£¤³¤ì¤é¤ÎÄê¿ô¤Ï¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤È
»Ò¥×¥í¥»¥¹¤Î´Ö¤Ç²¿¤ò¶¦Í­¤¹¤ë¤«¤ò»ØÄꤹ¤ë:
.TP
.\"O .BR CLONE_CHILD_CLEARTID " (since Linux 2.5.49)"
.BR CLONE_CHILD_CLEARTID " (Linux 2.5.49 °Ê¹ß)"
.\"O Erase child thread ID at location
.\"O .I ctid
.\"O in child memory when the child exits, and do a wakeup on the futex
.\"O at that address.
.\"O The address involved may be changed by the
.\"O .BR set_tid_address (2)
.\"O system call.
.\"O This is used by threading libraries.
»Ò¥×¥í¥»¥¹¤¬½ªÎ»¤·¤¿¤È¤­¤Ë»Ò¥×¥í¥»¥¹¤Î¥á¥â¥êÆâ¤Î
.I ctid
¤¬»Ø¤¹¾ì½ê¤Ë¤¢¤ë»Ò¥×¥í¥»¥¹¤Î¥¹¥ì¥Ã¥É ID ¤ò¾Ãµî¤·¡¢
¤½¤Î¥¢¥É¥ì¥¹¤Ç futex ¤ò wake (µ¯¾²) ¤µ¤»¤ë¡£
¤³¤Î¥¢¥É¥ì¥¹¤Ï
.BR set_tid_address (2)
¥·¥¹¥Æ¥à¥³¡¼¥ë¤ÇÊѹ¹¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
¤³¤Îµ¡Ç½¤Ï¥¹¥ì¥Ã¥É¥é¥¤¥Ö¥é¥ê¤Ç»ÈÍѤµ¤ì¤ë¡£
.TP
.\"O .BR CLONE_CHILD_SETTID " (since Linux 2.5.49)"
.BR CLONE_CHILD_SETTID " (Linux 2.5.49 °Ê¹ß)"
.\"O Store child thread ID at location
.\"O .I ctid
.\"O in child memory.
»Ò¥×¥í¥»¥¹¤Î¥á¥â¥êÆâ¤Î
.I ctid
¤¬»Ø¤¹¾ì½ê¤Ë»Ò¥×¥í¥»¥¹¤Î¥¹¥ì¥Ã¥É ID ¤ò³ÊǼ¤¹¤ë¡£
.TP
.B CLONE_FILES
.\"O If
.\"O .B CLONE_FILES
.\"O is set, the calling process and the child process share the same file
.\"O descriptor table.
.\"O Any file descriptor created by the calling process or by the child
.\"O process is also valid in the other process.
.\"O Similarly, if one of the processes closes a file descriptor,
.\"O or changes its associated flags (using the
.\"O .BR fcntl (2)
.\"O .B F_SETFD
.\"O operation), the other process is also affected.
.B CLONE_FILES
¤¬ÀßÄꤵ¤ì¤¿¾ì¹ç¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤È»Ò¥×¥í¥»¥¹¤Ï¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Î
¥Æ¡¼¥Ö¥ë¤ò¶¦Í­¤¹¤ë¡£
¸Æ¤Ó½Ð¤·¸µ¥×¥í¥»¥¹¤È¤½¤Î»Ò¥×¥í¥»¥¹¤Î°ìÊý¤¬ºîÀ®¤·¤¿
¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Ï¡¢¤â¤¦°ìÊý¤Ë¤ª¤¤¤Æ¤âÍ­¸ú¤Ç¤¢¤ë¡£
Ʊ¤¸¤è¤¦¤Ë¡¢°ìÊý¤Î¥×¥í¥»¥¹¤¬¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤òÊĤ¸¤¿¤ê¡¢
.RB ( fcntl (2)
.B F_SETFD
Áàºî¤ò»È¤Ã¤Æ) ¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Ë´ØÏ¢¤¹¤ë¥Õ¥é¥°¤òÊѹ¹¤·¤¿¤ê¤¹¤ë¤È¡¢
¤â¤¦°ìÊý¤Î¥×¥í¥»¥¹¤Ë¤â±Æ¶Á¤¹¤ë¡£

.\"O If
.\"O .B CLONE_FILES
.\"O is not set, the child process inherits a copy of all file descriptors
.\"O opened in the calling process at the time of
.\"O .BR clone ().
.\"O (The duplicated file descriptors in the child refer to the
.\"O same open file descriptions (see
.\"O .BR open (2))
.\"O as the corresponding file descriptors in the calling process.)
.\"O Subsequent operations that open or close file descriptors,
.\"O or change file descriptor flags,
.\"O performed by either the calling
.\"O process or the child process do not affect the other process.
.B CLONE_FILES
¤¬ÀßÄꤵ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç¡¢»Ò¥×¥í¥»¥¹¤Ï¡¢
.BR clone ()
¤¬¼Â¹Ô¤µ¤ì¤¿»þÅÀ¤Ç¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤¬¥ª¡¼¥×¥ó¤·¤Æ¤¤¤ëÁ´¤Æ¤Î
¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Î¥³¥Ô¡¼¤ò·Ñ¾µ¤¹¤ë
(»Ò¥×¥í¥»¥¹¤ÎÊ£À½¤µ¤ì¤¿¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Ï¡¢
Âбþ¤¹¤ë¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤Î¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤È
Ʊ¤¸¥Õ¥¡¥¤¥ëµ­½Ò
.RB ( open (2)
»²¾È) ¤ò»²¾È¤¹¤ë)¡£
¤³¤ì°Ê¹ß¤Ë¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤È»Ò¥×¥í¥»¥¹¤Î°ìÊý¤¬
¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤ÎÁàºî (¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Î
¥ª¡¼¥×¥ó¡¦¥¯¥í¡¼¥º¤ä¡¢¥Õ¥¡¥¤¥ë¥Ç¥£¥¹¥¯¥ê¥×¥¿¡¦¥Õ¥é¥°¤ÎÊѹ¹)
¤ò¹Ô¤Ã¤Æ¤â¡¢¤â¤¦°ìÊý¤Î¥×¥í¥»¥¹¤Ë¤Ï±Æ¶Á¤òÍ¿¤¨¤Ê¤¤¡£
.TP
.B CLONE_FS
.\"O If
.\"O .B CLONE_FS
.\"O is set, the caller and the child process share the same file system
.\"O information.
.\"O This includes the root of the file system, the current
.\"O working directory, and the umask.
.\"O Any call to
.\"O .BR chroot (2),
.\"O .BR chdir (2),
.\"O or
.\"O .BR umask (2)
.\"O performed by the calling process or the child process also affects the
.\"O other process.
.B CLONE_FS
¤¬ÀßÄꤵ¤ì¤¿¾ì¹ç¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤È»Ò¥×¥í¥»¥¹¤¬Æ±¤¸¥Õ¥¡¥¤¥ë¡¦¥·¥¹¥Æ¥à
¾ðÊó¤ò¶¦Í­¤¹¤ë¡£¥Õ¥¡¥¤¥ë¡¦¥·¥¹¥Æ¥à¾ðÊó¤Ï¡¢¥Õ¥¡¥¤¥ë¡¦¥·¥¹¥Æ¥à¤Î¥ë¡¼¥È (root)¡¢
¥«¥ì¥ó¥È¡¦¥ï¡¼¥­¥ó¥°¡¦¥Ç¥£¥ì¥¯¥È¥ê (current working directory)
¤ä umask ¤Ê¤É¤Ç¤¢¤ë¡£
¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤ä»Ò¥×¥í¥»¥¹¤Î¤É¤Á¤é¤«°ìÊý¤Ë¤è¤Ã¤Æ
.BR chroot (2),
.BR chdir (2),
.BR umask (2)
¤¬¸Æ¤Ó½Ð¤µ¤ì¤ë¤È¡¢¤â¤¦°ìÊý¤Î¥×¥í¥»¥¹¤Ë¤â±Æ¶Á¤¬µÚ¤Ö¡£

.\"O If
.\"O .B CLONE_FS
.\"O is not set, the child process works on a copy of the file system
.\"O information of the calling process at the time of the
.\"O .BR clone ()
.\"O call.
.\"O Calls to
.\"O .BR chroot (2),
.\"O .BR chdir (2),
.\"O .BR umask (2)
.\"O performed later by one of the processes do not affect the other process.
.B CLONE_FS
¤¬ÀßÄꤵ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç¡¢»Ò¥×¥í¥»¥¹¤Ï¡¢
.BR clone ()
¤¬¼Â¹Ô¤µ¤ì¤¿»þÅÀ¤Ç¤Î¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤Î¥Õ¥¡¥¤¥ë¡¦¥·¥¹¥Æ¥à¾ðÊó¤Î¥³¥Ô¡¼¤ò
»ÈÍѤ¹¤ë¡£
¤³¤ì°Ê¹ß¤Ï¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤È»Ò¥×¥í¥»¥¹¤Î°ìÊý¤¬
.BR chroot (2),
.BR chdir (2),
.BR umask (2)
¤ò¸Æ¤Ó½Ð¤·¤Æ¤â¡¢¤â¤¦°ìÊý¤Î¥×¥í¥»¥¹¤Ë¤Ï±Æ¶Á¤òÍ¿¤¨¤Ê¤¤¡£
.TP
.\"O .BR CLONE_IO " (since Linux 2.6.25)"
.BR CLONE_IO " (Linux 2.6.25 °Ê¹ß)"
.\"O If
.\"O .B CLONE_IO
.\"O is set, then the new process shares an I/O context with
.\"O the calling process.
.\"O If this flag is not set, then (as with
.\"O .BR fork (2))
.\"O the new process has its own I/O context.
.B CLONE_IO
¤¬ÀßÄꤵ¤ì¤¿¾ì¹ç¡¢¿·¤·¤¤¥×¥í¥»¥¹¤Ï¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤È
I/O ¥³¥ó¥Æ¥­¥¹¥È¤ò¶¦Í­¤¹¤ë¡£
¤³¤Î¥Õ¥é¥°¤¬ÀßÄꤵ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç¤Ë¤Ï¡¢
.RB ( fork (2)
¤Î¾ì¹ç¤ÈƱÍÍ) ¿·¤·¤¤¥×¥í¥»¥¹¤Ï¼«Ê¬ÀìÍѤΠI/O ¥³¥ó¥Æ¥­¥¹¥È¤ò»ý¤Ä¡£

.\" The following based on text from Jens Axboe
.\"O The I/O context is the I/O scope of the disk scheduler (i.e,
.\"O what the I/O scheduler uses to model scheduling of a process's I/O).
.\"O If processes share the same I/O context,
.\"O they are treated as one by the I/O scheduler.
.\"O As a consequence, they get to share disk time.
I/O ¥³¥ó¥Æ¥­¥¹¥È¤Ï¡¢¥Ç¥£¥¹¥¯¥¹¥±¥¸¥å¡¼¥ë¤Î I/O ¥¹¥³¡¼¥×¤Ç¤¢¤ë
(¸À¤¤´¹¤¨¤ë¤È¡¢I/O ¥³¥ó¥Æ¥­¥¹¥È¤Ï I/O ¥¹¥±¥¸¥å¡¼¥é¤¬¥×¥í¥»¥¹ I/O ¤Î
¥¹¥±¥¸¥å¡¼¥ê¥ó¥°¤ò¥â¥Ç¥ë²½¤¹¤ë¤Î¤Ë»ÈÍѤµ¤ì¤ë)¡£
Ê£¿ô¤Î¥×¥í¥»¥¹¤¬Æ±¤¸ I/O ¥³¥ó¥Æ¥­¥¹¥È¤ò¶¦Í­¤¹¤ë¾ì¹ç¡¢
¤³¤ì¤é¤Î¥×¥í¥»¥¹¤Ï I/O ¥¹¥±¥¸¥å¡¼¥é¤«¤é¤Ï°ì¤Ä¤È¤·¤Æ°·¤ï¤ì¤ë¡£
·ë²Ì¤È¤·¤Æ¡¢¤³¤ì¤é¤Î¥×¥í¥»¥¹¤Ï¥Ç¥£¥¹¥¯¥¢¥¯¥»¥¹¤Î»þ´Ö¤ò¶¦Í­¤¹¤ë¤è¤¦¤Ë¤Ê¤ë¡£
.\"O For some I/O schedulers,
.\"O .\" the anticipatory and CFQ scheduler
.\"O if two processes share an I/O context,
.\"O they will be allowed to interleave their disk access.
.\"O If several threads are doing I/O on behalf of the same process
.\"O .RB ( aio_read (3),
.\"O for instance), they should employ
.\"O .BR CLONE_IO
.\"O to get better I/O performance.
.\"O .\" with CFQ and AS.
¤¤¤¯¤Ä¤«¤Î I/O ¥¹¥±¥¸¥å¡¼¥é¤Ç¤Ï¡¢
.\" ͽ¬¥¹¥±¥¸¥å¡¼¥é¤È CFQ ¥¹¥±¥¸¥å¡¼¥é
Æó¤Ä¤Î¥×¥í¥»¥¹¤¬ I/O ¥³¥ó¥Æ¥­¥¹¥È¤ò¶¦Í­¤·¤Æ¤¤¤ë¾ì¹ç¡¢
¤³¤ì¤é¤Î¥×¥í¥»¥¹¤Ï¥Ç¥£¥¹¥¯¥¢¥¯¥»¥¹¤ò¸ò¸ß¤Ë¹Ô¤¦¤³¤È¤¬¤Ç¤­¤ë¡£
Ʊ¤¸¥×¥í¥»¥¹¤ÎÊ£¿ô¤Î¥¹¥ì¥Ã¥É¤¬ I/O ¤ò¼Â¹Ô¤·¤Æ¤¤¤ë¾ì¹ç
(Î㤨¤Ð
.BR aio_read (3))¡¢
.B CLONE_IO
¤òÍøÍѤ¹¤ë¤³¤È¤Ç I/O À­Ç½¤òÎɤ¯¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
.\" CFQ ¤È AS ¤Î¾ì¹ç

.\"O If the kernel is not configured with the
.\"O .B CONFIG_BLOCK
.\"O option, this flag is a no-op.
¥«¡¼¥Í¥ë¤ÎÀßÄ꤬
.B CONFIG_BLOCK
¥ª¥×¥·¥ç¥óÉÕ¤­¤Ç¤Ê¤¤¾ì¹ç¡¢
¤³¤Î¥Õ¥é¥°¤Ï²¿¤Î°ÕÌ£¤â»ý¤¿¤Ê¤¤¡£
.TP
.\"O .BR CLONE_NEWIPC " (since Linux 2.6.19)"
.BR CLONE_NEWIPC " (Linux 2.6.19 °Ê¹ß)"
.\"O If
.\"O .B CLONE_NEWIPC
.\"O is set, then create the process in a new IPC namespace.
.\"O If this flag is not set, then (as with
.\"O .BR fork (2)),
.\"O the process is created in the same IPC namespace as
.\"O the calling process.
.\"O This flag is intended for the implementation of containers.
.B CLONE_NEWIPC
¤¬ÀßÄꤵ¤ì¤¿¾ì¹ç¡¢¿·¤·¤¤ IPC ̾Á°¶õ´Ö (namespace) ¤Ç¥×¥í¥»¥¹¤òºîÀ®¤¹¤ë¡£
¤³¤Î¥Õ¥é¥°¤¬ÀßÄꤵ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç¡¢
.RB ( fork (2)
¤Î¾ì¹ç¤ÈƱÍÍ) ¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤ÈƱ¤¸ IPC ̾Á°¶õ´Ö¤Ç¥×¥í¥»¥¹¤¬
ºîÀ®¤µ¤ì¤ë¡£
¤³¤Î¥Õ¥é¥°¤Ï¡¢¥³¥ó¥Æ¥Ê¤Î¼ÂÁõ¤Ç¤Î»ÈÍѤò°Õ¿Þ¤·¤ÆÍÑ°Õ¤µ¤ì¤¿¤â¤Î¤Ç¤¢¤ë¡£

.\"O An IPC namespace consists of the set of identifiers for
.\"O System V IPC objects.
.\"O (These objects are created using
.\"O .BR msgctl (2),
.\"O .BR semctl (2),
.\"O and
.\"O .BR shmctl (2)).
.\"O Objects created in an IPC namespace are visible to all other processes
.\"O that are members of that namespace,
.\"O but are not visible to processes in other IPC namespaces.
IPC ̾Á°¶õ´Ö¤Ï¡¢System V IPC ¥ª¥Ö¥¸¥§¥¯¥ÈÍѤμ±ÊÌ»Ò (identifiers) ¤Î
½¸¹ç¤Ç¹½À®¤µ¤ì¤ë
(System V IPC ¥ª¥Ö¥¸¥§¥¯¥È¤Ï
.BR msgctl (2),
.BR semctl (2),
.BR shmctl (2)
¤ò»È¤Ã¤ÆºîÀ®¤µ¤ì¤ë)¡£
¤¢¤ë IPC ̾Á°¶õ´Ö¤ËºîÀ®¤µ¤ì¤¿¥ª¥Ö¥¸¥§¥¯¥È¤Ï¡¢
¤½¤Î̾Á°¶õ´Ö¤Î¥á¥ó¥Ð¡¼¤Ç¤¢¤ë¾¤Î¤¹¤Ù¤Æ¤Î¥×¥í¥»¥¹¤«¤é¤â¸«¤¨¤ë¤¬¡¢
°ã¤¦ IPC ̾Á°¶õ´Ö¤Î¥×¥í¥»¥¹¤«¤é¤Ï¸«¤¨¤Ê¤¤¡£

.\"O When an IPC namespace is destroyed
.\"O (i.e, when the last process that is a member of the namespace terminates),
.\"O all IPC objects in the namespace are automatically destroyed.
IPC ̾Á°¶õ´Ö¤¬ÇË´þ¤µ¤ì¤ë»þ
(¤¹¤Ê¤ï¤Á¡¢¤½¤Î̾Á°¶õ´Ö¤Î¥á¥ó¥Ð¡¼¤ÎºÇ¸å¤Î¥×¥í¥»¥¹¤¬½ªÎ»¤¹¤ë»þ)¡¢
¤½¤Î̾Á°¶õ´Ö¤ÎÁ´¤Æ¤Î IPC ¥ª¥Ö¥¸¥§¥¯¥È¤Ï¼«Æ°Åª¤ËÇË´þ¤µ¤ì¤ë¡£

.\"O Use of this flag requires: a kernel configured with the
.\"O .B CONFIG_SYSVIPC
.\"O and
.\"O .B CONFIG_IPC_NS
.\"O options and that the process be privileged
.\"O .RB ( CAP_SYS_ADMIN ).
.\"O This flag can't be specified in conjunction with
.\"O .BR CLONE_SYSVSEM .
¤³¤Î¥Õ¥é¥°¤ò»ÈÍѤ¹¤ë¤¿¤á¤Ë¤Ï¡¢
¥«¡¼¥Í¥ë¤Ç¥ª¥×¥·¥ç¥ó
.B CONFIG_SYSVIPC
¤È
.B CONFIG_IPC_NS
¤òÍ­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤ë¤³¤È¡¢
¥×¥í¥»¥¹¤¬Æø¢
.RB ( CAP_SYS_ADMIN )
¤ò»ý¤Ã¤Æ¤¤¤ë¤³¤È¤¬É¬ÍפǤ¢¤ë¡£
¤³¤Î¥Õ¥é¥°¤Ï
.B CLONE_SYSVSEM
¤ÈÁȤ߹ç¤ï¤»¤Æ»È¤¦¤³¤È¤Ï¤Ç¤­¤Ê¤¤¡£
.TP
.\"O .BR CLONE_NEWNET " (since Linux 2.6.24)"
.BR CLONE_NEWNET " (Linux 2.6.24 °Ê¹ß)"
.\"O (The implementation of this flag is not yet complete,
.\"O but probably will be mostly complete by about Linux 2.6.28.)
(¤³¤Î¥Õ¥é¥°¤Î¼ÂÁõ¤Ï¤Þ¤À´°Î»¤·¤Æ¤¤¤Ê¤¤¤¬¡¢
¤ª¤½¤é¤¯ Linux 2.6.28 ¤¢¤¿¤ê¤Þ¤Ç¤Ë¤Ï¤Û¤Ü´°À®¤¹¤ë¤³¤È¤À¤í¤¦¡£)

.\"O If
.\"O .B CLONE_NEWNET
.\"O is set, then create the process in a new network namespace.
.\"O If this flag is not set, then (as with
.\"O .BR fork (2)),
.\"O the process is created in the same network namespace as
.\"O the calling process.
.\"O This flag is intended for the implementation of containers.
.B CLONE_NEWNET
¤¬ÀßÄꤵ¤ì¤¿¾ì¹ç¡¢¿·¤·¤¤¥Í¥Ã¥È¥ï¡¼¥¯Ì¾Á°¶õ´Ö (network namaspace)
¤Ç¥×¥í¥»¥¹¤òºîÀ®¤¹¤ë¡£
¤³¤Î¥Õ¥é¥°¤¬ÀßÄꤵ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç¡¢
.RB ( fork (2)
¤Î¾ì¹ç¤ÈƱÍÍ) ¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤ÈƱ¤¸¥Í¥Ã¥È¥ï¡¼¥¯Ì¾Á°¶õ´Ö¤Ç¥×¥í¥»¥¹¤¬
ºîÀ®¤µ¤ì¤ë¡£
¤³¤Î¥Õ¥é¥°¤Ï¡¢¥³¥ó¥Æ¥Ê¤Î¼ÂÁõ¤Ç¤Î»ÈÍѤò°Õ¿Þ¤·¤ÆÍÑ°Õ¤µ¤ì¤¿¤â¤Î¤Ç¤¢¤ë¡£

.\"O A network namespace provides an isolated view of the networking stack
.\"O (network device interfaces, IPv4 and IPv6 protocol stacks,
.\"O IP routing tables, firewall rules, the
.\"O .I /proc/net
.\"O and
.\"O .I /sys/class/net
.\"O directory trees, sockets, etc.).
.\"O A physical network device can live in exactly one
.\"O network namespace.
.\"O A virtual network device ("veth") pair provides a pipe-like abstraction
.\"O that can be used to create tunnels between network namespaces,
.\"O and can be used to create a bridge to a physical network device
.\"O in another namespace.
¥Í¥Ã¥È¥ï¡¼¥¯Ì¾Á°¶õ´Ö¤Ï¡¢Ê¬Î¥¤µ¤ì¤¿¥Í¥Ã¥È¥ï¡¼¥¯¥¹¥¿¥Ã¥¯¤òÄ󶡤¹¤ë¤â¤Î¤Ç¤¢¤ë
(¥Í¥Ã¥È¥ï¡¼¥¯¥¹¥¿¥Ã¥¯¤È¤Ï¡¢
¥Í¥Ã¥È¥ï¡¼¥¯¥Ç¥Ð¥¤¥¹¥¤¥ó¥¿¥Õ¥§¡¼¥¹¡¢IPv4 ¤ä IPv6 ¥×¥í¥È¥³¥ë¥¹¥¿¥Ã¥¯¡¢
.IR /proc/net ¡¢
.I /sys/class/net
¥Ç¥£¥ì¥¯¥È¥ê¥Ä¥ê¡¼¡¢¥½¥±¥Ã¥È¤Ê¤É¤Ç¤¢¤ë)¡£
ʪÍý¥Í¥Ã¥È¥ï¡¼¥¯¥Ç¥Ð¥¤¥¹¤¬½ê°¤Ç¤­¤ë¥Í¥Ã¥È¥ï¡¼¥¯Ì¾Á°¶õ´Ö¤Ï°ì¤Ä¤À¤±¤Ç¤¢¤ë¡£
²¾Áۥͥåȥ¥¯¥Ç¥Ð¥¤¥¹ ("veth") ¤Î¥Ú¥¢¤Ë¤è¤ê
¥Ñ¥¤¥×É÷¤ÎÃê¾Ý²½ (abstraction) ¤¬¼Â¸½¤µ¤ì¤Æ¤ª¤ê¡¢
¤³¤ì¤ò»È¤¦¤³¤È¤Ç¡¢¥Í¥Ã¥È¥ï¡¼¥¯Ì¾Á°¶õ´Ö´Ö¤Î¥È¥ó¥Í¥ë¤òºîÀ®¤·¤¿¤ê¡¢
Ê̤Î̾Á°¶õ´Ö¤ÎʪÍý¥Í¥Ã¥È¥ï¡¼¥¯¥Ç¥Ð¥¤¥¹¤Ø¤Î¥Ö¥ê¥Ã¥¸¤òºîÀ®¤·¤¿¤ê
¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£

.\"O When a network namespace is freed
.\"O (i.e., when the last process in the namespace terminates),
.\"O its physical network devices are moved back to the
.\"O initial network namespace (not to the parent of the process).
¥Í¥Ã¥È¥ï¡¼¥¯Ì¾Á°¶õ´Ö¤¬²òÊü¤µ¤ì¤ë»þ
(¤¹¤Ê¤ï¤Á¡¢¤½¤Î̾Á°¶õ´Ö¤ÎºÇ¸å¤Î¥×¥í¥»¥¹¤¬½ªÎ»¤¹¤ë»þ)¡¢
ʪÍý¥Í¥Ã¥È¥ï¡¼¥¯¥Ç¥Ð¥¤¥¹¤Ï½é´ü¥Í¥Ã¥È¥ï¡¼¥¯Ì¾Á°¶õ´Ö
(initial network namespace) ¤ËÌᤵ¤ì¤ë
(¿Æ¥×¥í¥»¥¹¤Î¥Í¥Ã¥È¥ï¡¼¥¯Ì¾Á°¶õ´Ö¤ËÌᤵ¤ì¤ëÌõ¤Ç¤Ï¤Ê¤¤)¡£

.\"O Use of this flag requires: a kernel configured with the
.\"O .B CONFIG_NET_NS
.\"O option and that the process be privileged
.\"O .RB ( CAP_SYS_ADMIN ).
¤³¤Î¥Õ¥é¥°¤ò»ÈÍѤ¹¤ë¤¿¤á¤Ë¤Ï¡¢
¥«¡¼¥Í¥ë¤Ç¥ª¥×¥·¥ç¥ó
.B CONFIG_NET_NS
¤òÍ­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤ë¤³¤È¡¢
¥×¥í¥»¥¹¤¬Æø¢
.RB ( CAP_SYS_ADMIN )
¤ò»ý¤Ã¤Æ¤¤¤ë¤³¤È¤¬É¬ÍפǤ¢¤ë¡£
.TP
.\"O .BR CLONE_NEWNS " (since Linux 2.4.19)"
.BR CLONE_NEWNS " (Linux 2.4.19 °Ê¹ß)"
.\"O Start the child in a new mount namespace.
»Ò¥×¥í¥»¥¹¤ò¿·¤·¤¤¥Þ¥¦¥ó¥È̾Á°¶õ´Ö (mount namespace) ¤Ç³«»Ï¤¹¤ë¡£

.\"O Every process lives in a mount namespace.
.\"O The
.\"O .I namespace
.\"O of a process is the data (the set of mounts) describing the file hierarchy
.\"O as seen by that process.
.\"O After a
.\"O .BR fork (2)
.\"O or
.\"O .BR clone ()
.\"O where the
.\"O .B CLONE_NEWNS
.\"O flag is not set, the child lives in the same mount
.\"O namespace as the parent.
³Æ¥×¥í¥»¥¹¤Ï¤¢¤ë°ì¤Ä¤Î¥Þ¥¦¥ó¥È̾Á°¶õ´ÖÃæ¤Ë¸ºß¤¹¤ë¡£¥×¥í¥»¥¹¤Î
.I "̾Á°¶õ´Ö (namespace)"
¤Ï¡¢¤½¤Î¥×¥í¥»¥¹¤«¤é¸«¤¨¤ë¥Õ¥¡¥¤¥ë³¬Áؤòɽ¤¹¥Ç¡¼¥¿ (mount ¤Î½¸¹ç) ¤Ç¤¢¤ë¡£
.B CLONE_NEWNS
¥Õ¥é¥°¤¬¥»¥Ã¥È¤µ¤ì¤º¤Ë
.BR fork (2)
¤«
.BR clone ()
¤¬¸Æ¤Ð¤ì¤ë¤È¡¢»Ò¥×¥í¥»¥¹¤Ï¿Æ¥×¥í¥»¥¹¤ÈƱ¤¸¥Þ¥¦¥ó¥È̾Á°¶õ´Ö¤ËºîÀ®¤µ¤ì¤ë¡£
.\"O The system calls
.\"O .BR mount (2)
.\"O and
.\"O .BR umount (2)
.\"O change the mount namespace of the calling process, and hence affect
.\"O all processes that live in the same namespace, but do not affect
.\"O processes in a different mount namespace.
¥·¥¹¥Æ¥à¥³¡¼¥ë
.BR mount (2)¡¢
.BR umount (2)
¤¬¸Æ¤Ð¤ì¤ë¤È¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤Î¥Þ¥¦¥ó¥È̾Á°¶õ´Ö¤¬Êѹ¹¤µ¤ì¡¢¤³¤Î·ë²Ì
¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤ÈƱ¤¸Ì¾Á°¶õ´Ö¤Ë¤¤¤ë¥×¥í¥»¥¹¤Ï¤¹¤Ù¤Æ±Æ¶Á¤ò¼õ¤±¤ë¤¬¡¢
°Û¤Ê¤ë¥Þ¥¦¥ó¥È̾Á°¶õ´Ö¤Ë¤¤¤ë¥×¥í¥»¥¹¤Ï±Æ¶Á¤ò¼õ¤±¤Ê¤¤¡£

.\"O After a
.\"O .BR clone ()
.\"O where the
.\"O .B CLONE_NEWNS
.\"O flag is set, the cloned child is started in a new mount namespace,
.\"O initialized with a copy of the namespace of the parent.
.B CLONE_NEWNS
¥Õ¥é¥°¤¬¥»¥Ã¥È¤µ¤ì¤Æ
.BR clone ()
¤¬¸Æ¤Ð¤ì¤ë¤È¡¢clone ¤ÇºîÀ®¤µ¤ì¤¿»Ò¥×¥í¥»¥¹¤Ï¿·¤·¤¤¥Þ¥¦¥ó¥È̾Á°¶õ´Ö¤Ç
³«»Ï¤µ¤ì¤ë¡£¿·¤·¤¤Ì¾Á°¶õ´Ö¤Ï¿Æ¥×¥í¥»¥¹¤Î̾Á°¶õ´Ö¤Î¥³¥Ô¡¼¤Ç½é´ü²½¤µ¤ì¤ë¡£

.\"O Only a privileged process (one having the \fBCAP_SYS_ADMIN\fP capability)
.\"O may specify the
.\"O .B CLONE_NEWNS
.\"O flag.
.\"O It is not permitted to specify both
.\"O .B CLONE_NEWNS
.\"O and
.\"O .B CLONE_FS
.\"O in the same
.\"O .BR clone ()
.\"O call.
Æø¢¥×¥í¥»¥¹ (\fBCAP_SYS_ADMIN\fP ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò»ý¤Ä¥×¥í¥»¥¹) ¤Î¤ß¤¬
.B CLONE_NEWNS
¥Õ¥é¥°¤ò»ØÄꤹ¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
°ì¤Ä¤Î
.BR clone ()
¸Æ¤Ó½Ð¤·¤Ç¡¢
.B CLONE_NEWNS
¤È
.B CLONE_FS
¤ÎξÊý¤ò»ØÄꤹ¤ë¤³¤È¤Ï¤Ç¤­¤Ê¤¤¡£
.TP
.\"O .BR CLONE_NEWPID " (since Linux 2.6.24)"
.BR CLONE_NEWPID " (Linux 2.6.24 °Ê¹ß)"
.\" This explanation draws a lot of details from
.\" http://lwn.net/Articles/259217/
.\" Authors: Pavel Emelyanov <xemul@openvz.org>
.\" and Kir Kolyshkin <kir@openvz.org>
.\"
.\" The primary kernel commit is 30e49c263e36341b60b735cbef5ca37912549264
.\" Author: Pavel Emelyanov <xemul@openvz.org>
.\"O If
.\"O .B CLONE_NEWPID
.\"O is set, then create the process in a new PID namespace.
.\"O If this flag is not set, then (as with
.\"O .BR fork (2)),
.\"O the process is created in the same PID namespace as
.\"O the calling process.
.\"O This flag is intended for the implementation of containers.
.B CLONE_NEWPID
¤¬ÀßÄꤵ¤ì¤¿¾ì¹ç¡¢¿·¤·¤¤ PID ̾Á°¶õ´Ö¤Ç¥×¥í¥»¥¹¤òºîÀ®¤¹¤ë¡£
¤³¤Î¥Õ¥é¥°¤¬ÀßÄꤵ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç¡¢
.RB ( fork (2)
¤Î¾ì¹ç¤ÈƱÍÍ) ¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤ÈƱ¤¸ PID ̾Á°¶õ´Ö¤Ç
¥×¥í¥»¥¹¤¬ºîÀ®¤µ¤ì¤ë¡£
¤³¤Î¥Õ¥é¥°¤Ï¡¢¥³¥ó¥Æ¥Ê¤Î¼ÂÁõ¤Ç¤Î»ÈÍѤò°Õ¿Þ¤·¤ÆÍÑ°Õ¤µ¤ì¤¿¤â¤Î¤Ç¤¢¤ë¡£

.\"O A PID namespace provides an isolated environment for PIDs:
.\"O PIDs in a new namespace start at 1,
.\"O somewhat like a standalone system, and calls to
.\"O .BR fork (2),
.\"O .BR vfork (2),
.\"O or
.\"O .BR clone (2)
.\"O will produce processes with PIDs that are unique within the namespace.
PID ̾Á°¶õ´Ö¤Ï¡¢PID ¤Ë´Ø¤·¤ÆʬΥ¤µ¤ì¤¿´Ä¶­¤òÄ󶡤¹¤ë¤â¤Î¤Ç¤¢¤ë¡£
¿·¤·¤¤Ì¾Á°¶õ´Ö¤Ë¤ª¤±¤ë PID ¤Ï 1 ¤«¤é»Ï¤Þ¤ê
(¤³¤ì¤Ï¥¹¥¿¥ó¥É¥¢¥í¥ó¤Î¥·¥¹¥Æ¥à¤È»÷¤¿¤è¤¦¤Ê´¶¤¸)¡¢
.BR fork (2),
.BR vfork (2),
.BR clone (2)
¤ò¸Æ¤Ó½Ð¤¹¤È¡¢¤½¤Î̾Á°¶õ´Ö¤Ç°ì°Õ¤Ê PID ¤ò»ý¤Ã¤¿¥×¥í¥»¥¹¤¬ºîÀ®¤µ¤ì¤ë¡£

.\"O The first process created in a new namespace
.\"O (i.e., the process created using the
.\"O .BR CLONE_NEWPID
.\"O flag) has the PID 1, and is the "init" process for the namespace.
.\"O Children that are orphaned within the namespace will be reparented
.\"O to this process rather than
.\"O .BR init (8).
.\"O Unlike the traditional
.\"O .B init
.\"O process, the "init" process of a PID namespace can terminate,
.\"O and if it does, all of the processes in the namespace are terminated.
¿·¤·¤¤Ì¾Á°¶õ´Ö¤ÇºîÀ®¤µ¤ì¤ëºÇ½é¤Î¥×¥í¥»¥¹
(¤Ä¤Þ¤ê¡¢
.B CLONE_NEWPID
¥Õ¥é¥°¤ò»È¤Ã¤ÆºîÀ®¤µ¤ì¤¿¥×¥í¥»¥¹) ¤Î PID ¤Ï 1 ¤Ç¤¢¤ê¡¢
¤³¤Î¥×¥í¥»¥¹¤Ï¤½¤Î̾Á°¶õ´Ö¤Ë¤ª¤±¤ë "init" ¥×¥í¥»¥¹¤È¤Ê¤ë¡£
¤³¤Î̾Á°¶õ´Ö¤Ë¤ª¤¤¤Æ¸É»ù (orphaned) ¤È¤Ê¤Ã¤¿»Ò¥×¥í¥»¥¹¤Ë¤Ä¤¤¤Æ¤Ï¡¢
.BR init (8)
¤Ç¤Ï¤Ê¤¯¤³¤Î¥×¥í¥»¥¹¤¬¿Æ¥×¥í¥»¥¹¤È¤Ê¤ë¡£
ÀΤʤ¬¤é¤Î
.B init
¥×¥í¥»¥¹¤È¤Ï°ã¤¤¡¢PID ̾Á°¶õ´Ö¤Î "init" ¥×¥í¥»¥¹¤Ï½ªÎ» (terminated) ¤¹¤ë
¤³¤È¤¬¤Ç¤­¡¢¤½¤Î¾ì¹ç¤Ë¤Ï¡¢¤³¤Î̾Á°¶õ´Ö¤ÎÁ´¤Æ¤Î¥×¥í¥»¥¹¤¬½ªÎ»¤µ¤ì¤ë¡£

.\"O PID namespaces form a hierarchy.
PID ̾Á°¶õ´Ö´Ö¤Ë¤Ï³¬Áع½Â¤¤¬·ÁÀ®¤µ¤ì¤ë¡£
.\"O When a new PID namespace is created,
.\"O the processes in that namespace are visible
.\"O in the PID namespace of the process that created the new namespace;
.\"O analogously, if the parent PID namespace is itself
.\"O the child of another PID namespace,
.\"O then processes in the child and parent PID namespaces will both be
.\"O visible in the grandparent PID namespace.
¿·¤·¤¤ PID ̾Á°¶õ´Ö¤¬ºîÀ®¤µ¤ì¤ë¤È¡¢¤½¤Î̾Á°¶õ´Ö¤Î¥×¥í¥»¥¹¤Ï¡¢
¿·¤·¤¤Ì¾Á°¶õ´Ö¤òºîÀ®¤·¤¿¥×¥í¥»¥¹¤Î PID ̾Á°¶õ´Ö¤Ç¸«¤¨¤ë¡£
ƱÍͤˡ¢¿Æ¤Î PID ̾Á°¶õ´Ö¼«ÂΤ¬Ê̤ΠPID ̾Á°¶õ´Ö¤Î»Ò¶¡¤Î¾ì¹ç¤Ë¤Ï¡¢
»Ò¶¡¤Î PID ̾Á°¶õ´Ö¤È¿Æ¤Î PID ̾Á°¶õ´Ö¤Î¥×¥í¥»¥¹¤Ï¤É¤ì¤â
¿Æ¤Î¿Æ¤Î PID ̾Á°¶õ´Ö¤Ç¤â¸«¤¨¤ë¤³¤È¤Ë¤Ê¤ë¡£
.\"O Conversely, the processes in the "child" PID namespace do not see
.\"O the processes in the parent namespace.
.\"O The existence of a namespace hierarchy means that each process
.\"O may now have multiple PIDs:
.\"O one for each namespace in which it is visible;
.\"O each of these PIDs is unique within the corresponding namespace.
.\"O (A call to
.\"O .BR getpid (2)
.\"O always returns the PID associated with the namespace in which
.\"O the process lives.)
È¿ÂФˡ¢¡Ö»Ò¶¡¡×¤Î PID ̾Á°¶õ´Ö¤Î¥×¥í¥»¥¹¤Ë¤Ï¡¢
¿Æ¤Î̾Á°¶õ´Ö¤Î¥×¥í¥»¥¹¤Ï¸«¤¨¤Ê¤¤¡£
̾Á°¶õ´Ö¤Ë³¬Áع½Â¤¤¬Â¸ºß¤¹¤ë¤È¤¤¤¦¤³¤È¤Ï¡¢¸Ä¡¹¤Î¥×¥í¥»¥¹¤Ï
Ê£¿ô¤Î PID ¤ò»ý¤Ä¤È¤¤¤¦¤³¤È¤ò°ÕÌ£¤·¤Æ¤¤¤ë¡£
¤½¤Î¥×¥í¥»¥¹¤¬¸«¤¨¤ë̾Á°¶õ´Ö°ì¤Ä¤Ë¤Ä¤­ PID ¤¬°ì¤Ä¤¢¤ê¡¢
¤½¤ì¤¾¤ì¤Î PID ¤ÏÂбþ¤¹¤ë̾Á°¶õ´Ö¤Ë¤ª¤¤¤Æ°ì°Õ¤Ç¤¢¤ë¡£
.RB ( getpid (2)
¤ò¸Æ¤Ó½Ð¤¹¤È¡¢¾ï¤Ë¤½¤Î¥×¥í¥»¥¹¤¬Â¸ºß¤·¤Æ¤¤¤ë̾Á°¶õ´Ö¤Ë¤ª¤±¤ë
PID ¤¬ÊÖ¤µ¤ì¤ë¡£)

.\"O After creating the new namespace,
.\"O it is useful for the child to change its root directory
.\"O and mount a new procfs instance at
.\"O .I /proc
.\"O so that tools such as
.\"O .BR ps (1)
.\"O work correctly.
¿·¤·¤¤Ì¾Á°¶õ´Ö¤ÎºîÀ®¸å¤Ë¤Ï¡¢
»Ò¥×¥í¥»¥¹¤Ë¤ª¤¤¤Æ¡¢
.BR ps (1)
¤È¤¤¤Ã¤¿¥Ä¡¼¥ë¤¬Àµ¤·¤¯Æ°ºî¤¹¤ë¤è¤¦¤Ë¡¢
¼«¿È¤Î root ¥Ç¥£¥ì¥¯¥È¥ê¤òÊѹ¹¤·¡¢
.I /proc
¤Ë¿·¤·¤¤ procfs ¥¤¥ó¥¹¥¿¥ó¥¹¤ò¥Þ¥¦¥ó¥È¤¹¤ë¤Î¤¬¤è¤¤¤À¤í¤¦¡£
.\" mount -t proc proc /proc
.\"O (If
.\"O .BR CLONE_NEWNS
.\"O is also included in
.\"O .IR flags ,
.\"O then it isn't necessary to change the root directory:
.\"O a new procfs instance can be mounted directly over
.\"O .IR /proc .)
.RB ( flags
¤Ë
.B CLONE_NEWNS
¤â»ØÄꤵ¤ì¤Æ¤¤¤¿¾ì¹ç¤Ë¤Ï¡¢root ¥Ç¥£¥ì¥¯¥È¥ê¤òÊѹ¹¤¹¤ëɬÍפϤʤ¯¡¢
¤¤¤­¤Ê¤ê¿·¤·¤¤ procfs ¥¤¥ó¥¹¥¿¥ó¥¹¤ò
.I /proc
¤Ë¥Þ¥¦¥ó¥È¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£)

.\"O Use of this flag requires: a kernel configured with the
.\"O .B CONFIG_PID_NS
.\"O option and that the process be privileged
.\"O .RB ( CAP_SYS_ADMIN ).
.\"O This flag can't be specified in conjunction with
.\"O .BR CLONE_THREAD .
¤³¤Î¥Õ¥é¥°¤ò»ÈÍѤ¹¤ë¤¿¤á¤Ë¤Ï¡¢
¥«¡¼¥Í¥ë¤Ç¥ª¥×¥·¥ç¥ó
.B CONFIG_PID_NS
¤òÍ­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤ë¤³¤È¡¢
¥×¥í¥»¥¹¤¬Æø¢
.RB ( CAP_SYS_ADMIN )
¤ò»ý¤Ã¤Æ¤¤¤ë¤³¤È¤¬É¬ÍפǤ¢¤ë¡£
¤³¤Î¥Õ¥é¥°¤Ï
.B CLONE_THREAD
¤ÈÁȤ߹ç¤ï¤»¤Æ»È¤¦¤³¤È¤Ï¤Ç¤­¤Ê¤¤¡£
.TP
.\"O .BR CLONE_NEWUTS " (since Linux 2.6.19)"
.BR CLONE_NEWUTS " (Linux 2.6.19 °Ê¹ß)"
.\"O If
.\"O .B CLONE_NEWUTS
.\"O is set, then create the process in a new UTS namespace,
.\"O whose identifiers are initialized by duplicating the identifiers
.\"O from the UTS namespace of the calling process.
.\"O If this flag is not set, then (as with
.\"O .BR fork (2)),
.\"O the process is created in the same UTS namespace as
.\"O the calling process.
.\"O This flag is intended for the implementation of containers.
.B CLONE_NEWUTS
¤¬ÀßÄꤵ¤ì¤¿¾ì¹ç¡¢¿·¤·¤¤ UTS ̾Á°¶õ´Ö¤Ç¥×¥í¥»¥¹¤òºîÀ®¤¹¤ë¡£
¿·¤·¤¤ UTS ̾Á°¶õ´Ö¤Î¼±Ê̻Ҥνé´üÃͤϡ¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤Î
UTS ̾Á°¶õ´Ö¤Î¼±Ê̻ҤòÊ£À½¤·¤¿¤â¤Î¤È¤Ê¤ë¡£
¤³¤Î¥Õ¥é¥°¤¬ÀßÄꤵ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç¡¢
.RB ( fork (2)
¤Î¾ì¹ç¤ÈƱÍÍ) ¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤ÈƱ¤¸ UTS ̾Á°¶õ´Ö¤Ç
¥×¥í¥»¥¹¤¬ºîÀ®¤µ¤ì¤ë¡£
¤³¤Î¥Õ¥é¥°¤Ï¡¢¥³¥ó¥Æ¥Ê¤Î¼ÂÁõ¤Ç¤Î»ÈÍѤò°Õ¿Þ¤·¤ÆÍÑ°Õ¤µ¤ì¤¿¤â¤Î¤Ç¤¢¤ë¡£

.\"O A UTS namespace is the set of identifiers returned by
.\"O .BR uname (2);
.\"O among these, the domain name and the host name can be modified by
.\"O .BR setdomainname (2)
.\"O and
.\"O .BR
.\"O .BR sethostname (2),
.\"O respectively.
.\"O Changes made to the identifiers in a UTS namespace
.\"O are visible to all other processes in the same namespace,
.\"O but are not visible to processes in other UTS namespaces.
UTS ̾Á°¶õ´Ö¤Ï¡¢
.BR uname (2)
¤¬ÊÖ¤¹¼±Ê̻Ҥν¸¹ç¤Ç¤¢¤ë¡£
¼±Ê̻ҤȤ·¤Æ¤Ï¥É¥á¥¤¥ó̾¤È¥Û¥¹¥È̾¤¬¤¢¤ê¡¢
¤½¤ì¤¾¤ì
.BR setdomainname (2),
.BR sethostname (2)
¤Ç½¤Àµ¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
¤¢¤ë UTS ̾Á°¶õ´Ö¤Ë¤ª¤±¤ë¼±Ê̻ҤÎÊѹ¹¤ÏƱ¤¸Ì¾Á°¶õ´Ö¤Î¾¤Î¤¹¤Ù¤Æ¤Î
¥×¥í¥»¥¹¤Ë¸«¤¨¤ë¤¬¡¢Ê̤ΠUTS ̾Á°¶õ´Ö¤Î¥×¥í¥»¥¹¤Ë¤Ï¸«¤¨¤Ê¤¤¡£

.\"O Use of this flag requires: a kernel configured with the
.\"O .B CONFIG_UTS_NS
.\"O option and that the process be privileged
.\"O .RB ( CAP_SYS_ADMIN ).
¤³¤Î¥Õ¥é¥°¤ò»ÈÍѤ¹¤ë¤¿¤á¤Ë¤Ï¡¢
¥«¡¼¥Í¥ë¤Ç¥ª¥×¥·¥ç¥ó
.B CONFIG_UTS_NS
¤òÍ­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤ë¤³¤È¡¢
¥×¥í¥»¥¹¤¬Æø¢
.RB ( CAP_SYS_ADMIN )
¤ò»ý¤Ã¤Æ¤¤¤ë¤³¤È¤¬É¬ÍפǤ¢¤ë¡£
.TP
.\"O .BR CLONE_PARENT " (since Linux 2.3.12)"
.BR CLONE_PARENT " (Linux 2.3.12 °Ê¹ß)"
.\"O If
.\"O .B CLONE_PARENT
.\"O is set, then the parent of the new child (as returned by
.\"O .BR getppid (2))
.\"O will be the same as that of the calling process.
.B CLONE_PARENT
¤¬ÀßÄꤵ¤ì¤¿¾ì¹ç¡¢¿·¤·¤¤»Ò¶¡¤Î
.RB ( getppid (2)
¤ÇÊÖ¤µ¤ì¤ë) ¿Æ¥×¥í¥»¥¹¤Ï¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤Î¿Æ¥×¥í¥»¥¹¤ÈƱ¤¸¤Ë¤Ê¤ë¡£

.\"O If
.\"O .B CLONE_PARENT
.\"O is not set, then (as with
.\"O .BR fork (2))
.\"O the child's parent is the calling process.
.B CLONE_PARENT
¤¬ÀßÄꤵ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç¡¢
.RB ( fork (2)
¤ÈƱÍͤË) ¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤¬¤½¤Î»Ò¶¡¤Î¿Æ¤Ë¤Ê¤ë¡£

.\"O Note that it is the parent process, as returned by
.\"O .BR getppid (2),
.\"O which is signaled when the child terminates, so that
.\"O if
.\"O .B CLONE_PARENT
.\"O is set, then the parent of the calling process, rather than the
.\"O calling process itself, will be signaled.
»Ò¶¡¤¬½ªÎ»¤·¤¿»þ¤Ë¥·¥°¥Ê¥ë¤¬Á÷¤é¤ì¤ë¤Î¤Ï
.BR getppid (2)
¤¬ÊÖ¤¹¿Æ¥×¥í¥»¥¹¤Ç¤¢¤ëÅÀ¤ËÃí°Õ¤¹¤ë¤³¤È¡£¤³¤Î¤¿¤á
.B CLONE_PARENT
¤¬ÀßÄꤵ¤ì¤¿¾ì¹ç¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤Ç¤Ï¤Ê¤¯¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤Î
¿Æ¥×¥í¥»¥¹¤Ë¥·¥°¥Ê¥ë¤¬Á÷¤é¤ì¤ë¡£
.TP
.\"O .BR CLONE_PARENT_SETTID " (since Linux 2.5.49)"
.BR CLONE_PARENT_SETTID " (Linux 2.5.49 °Ê¹ß)"
.\"O Store child thread ID at location
.\"O .I ptid
.\"O in parent and child memory.
.\"O (In Linux 2.5.32-2.5.48 there was a flag
.\"O .B CLONE_SETTID
.\"O that did this.)
¿Æ¥×¥í¥»¥¹¤È»Ò¥×¥í¥»¥¹¤Î¥á¥â¥êÆâ¤Î
.I ptid
¤¬»Ø¤¹Îΰè¤Ë»Ò¥×¥í¥»¥¹¤Î¥¹¥ì¥Ã¥É ID ¤ò³ÊǼ¤¹¤ë¡£
(Linux 2.5.32-2.5.48 ¤Ç¤Ï¡¢
Ʊ¤¸¤³¤È¤ò¤¹¤ë
.B CLONE_SETTID
¤È¤¤¤¦¥Õ¥é¥°¤¬Â¸ºß¤·¤¿¡£)
.TP
.\"O .BR CLONE_PID " (obsolete)"
.BR CLONE_PID " (ÇÑ»ßͽÄê)"
.\"O If
.\"O .B CLONE_PID
.\"O is set, the child process is created with the same process ID as
.\"O the calling process.
.\"O This is good for hacking the system, but otherwise
.\"O of not much use.
.\"O Since 2.3.21 this flag can be
.\"O specified only by the system boot process (PID 0).
.\"O It disappeared in Linux 2.5.16.
.B CLONE_PID
¤¬ÀßÄꤵ¤ì¤¿¾ì¹ç¡¢»Ò¥×¥í¥»¥¹¤Ï¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤ÈƱ¤¸¥×¥í¥»¥¹ ID
¤ÇºîÀ®¤µ¤ì¤ë¡£¤³¤ì¤Ï¥·¥¹¥Æ¥à¤ò¥Ï¥Ã¥­¥ó¥°¤¹¤ë¤Î¤Ë¤ÏÊØÍø¤À¤¬¡¢
¤½¤ì°Ê³°¤Ë¤Ï¤¢¤Þ¤ê»È¤ï¤ì¤Ê¤¤¡£ Linux 2.3.21 °Ê¹ß¤Ç¤Ï¡¢
¥·¥¹¥Æ¥à¤Î¥Ö¡¼¥È¥×¥í¥»¥¹ (PID 0) ¤À¤±¤¬¤³¤Î¥Õ¥é¥°¤ò»ØÄê¤Ç¤­¤ë¡£
Linux 2.5.16 ¤Çºï½ü¤µ¤ì¤¿¡£
.TP
.B CLONE_PTRACE
.\"O If
.\"O .B CLONE_PTRACE
.\"O is specified, and the calling process is being traced,
.\"O then trace the child also (see
.\"O .BR ptrace (2)).
.B CLONE_PTRACE
¤¬»ØÄꤵ¤ì¡¢¤«¤Ä¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤¬ÄÉÀ× (trace) ¤µ¤ì¤Æ¤¤¤¿¾ì¹ç¡¢»Ò¥×¥í¥»¥¹¤â
ƱÍͤËÄÉÀפµ¤ì¤ë¡£
.RB ( ptrace (2)
¤ò»²¾È¤Î¤³¤È)
.TP
.\"O .BR CLONE_SETTLS " (since Linux 2.5.32)"
.BR CLONE_SETTLS " (Linux 2.5.32 °Ê¹ß)"
.\"O The
.\"O .I newtls
.\"O argument is the new TLS (Thread Local Storage) descriptor.
.\"O (See
.\"O .BR set_thread_area (2).)
.I newtls
°ú¤­¿ô¤Ï¡¢¿·¤·¤¤ TLS (Thread Local Storage) ¥Ç¥£¥¹¥¯¥ê¥×¥¿¤Ç¤¢¤ë¡£
.RB ( set_thread_area (2)
¤ò»²¾È¤Î¤³¤È)
.TP
.B CLONE_SIGHAND
.\"O If
.\"O .B CLONE_SIGHAND
.\"O is set, the calling process and the child process share the same table of
.\"O signal handlers.
.\"O If the calling process or child process calls
.\"O .BR sigaction (2)
.\"O to change the behavior associated with a signal, the behavior is
.\"O changed in the other process as well.
.\"O However, the calling process and child
.\"O processes still have distinct signal masks and sets of pending
.\"O signals.
.\"O So, one of them may block or unblock some signals using
.\"O .BR sigprocmask (2)
.\"O without affecting the other process.
.B CLONE_SIGHAND
¤¬ÀßÄꤵ¤ì¤¿¾ì¹ç¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤È»Ò¥×¥í¥»¥¹¤ÏƱ¤¸¥·¥°¥Ê¥ë¡¦¥Ï¥ó
¥É¥é¤Î¥Æ¡¼¥Ö¥ë¤ò¶¦Í­¤¹¤ë¡£¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤Þ¤¿¤Ï»Ò¥×¥í¥»¥¹¤Î¤É¤Á¤é¤«¤¬
.BR sigaction (2)
¤ò¸Æ¤Ó½Ð¤·¤Æ¥·¥°¥Ê¥ë¤ËÂбþ¤¹¤ëÆ°ºî¤òÊѹ¹¤·¤¿¾ì¹ç¡¢
¤â¤¦°ìÊý¤Î¥×¥í¥»¥¹¤Î¥·¥°¥Ê¥ëÆ°ºî¤âÊѹ¹¤µ¤ì¤ë¡£
⤷¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤È»Ò¥×¥í¥»¥¹¤Ï¡¢
¥×¥í¥»¥¹Ëè¤Ë¡¢¥·¥°¥Ê¥ë¡¦¥Þ¥¹¥¯ (signal mask) ¤È½èÍýÂÔ¤Á¥·¥°¥Ê¥ë¤Î½¸¹ç
¤ò»ý¤Ã¤Æ¤¤¤ë¡£¤³¤Î¤¿¤á¡¢¤¢¤ë¥×¥í¥»¥¹¤Ï¡¢
.BR sigprocmask (2)
¤ò»ÈÍѤ·¤Æ¡¢¤â¤¦°ìÊý¤Î¥×¥í¥»¥¹¤Ë±Æ¶Á¤òÍ¿¤¨¤º¤Ë
¥·¥°¥Ê¥ë¤ò¶Ø»ß (block) ¤·¤¿¤êµö²Ä (unblock) ¤·¤¿¤ê¤Ç¤­¤ë¡£

.\"O If
.\"O .B CLONE_SIGHAND
.\"O is not set, the child process inherits a copy of the signal handlers
.\"O of the calling process at the time
.\"O .BR clone ()
.\"O is called.
.\"O Calls to
.\"O .BR sigaction (2)
.\"O performed later by one of the processes have no effect on the other
.\"O process.
.B CLONE_SIGHAND
¤¬ÀßÄꤵ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç¡¢»Ò¥×¥í¥»¥¹¤Ï
.BR clone ()
¤¬¼Â¹Ô¤µ¤ì¤¿»þÅÀ¤Ç¤Î¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤Î¥·¥°¥Ê¥ë¡¦¥Ï¥ó¥É¥é¤Î
¥³¥Ô¡¼¤ò·Ñ¾µ¤¹¤ë¡£¤³¤ì°Ê¹ß¤Ï¡¢°ìÊý¤Î¥×¥í¥»¥¹¤¬
.BR sigaction (2)
¤ò¸Æ¤Ó½Ð¤·¤Æ¤â¡¢¤â¤¦°ìÊý¤Î¥×¥í¥»¥¹¤Ë¤Ï±Æ¶Á¤òÍ¿¤¨¤Ê¤¤¡£

.\"O Since Linux 2.6.0-test6,
.\"O .I flags
.\"O must also include
.\"O .B CLONE_VM
.\"O if
.\"O .B CLONE_SIGHAND
.\"O is specified
Linux 2.6.0-test6 °Ê¹ß¤Ç¤Ï¡¢
.B CLONE_SIGHAND
¤ò»ØÄꤹ¤ë¾ì¹ç¡¢
.B CLONE_VM
¤â
.I flags
¤Ë´Þ¤á¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£
.TP
.\"O .BR CLONE_STOPPED " (since Linux 2.6.0-test2)"
.BR CLONE_STOPPED " (Linux 2.6.0-test2 °Ê¹ß)"
.\"O If
.\"O .B CLONE_STOPPED
.\"O is set, then the child is initially stopped (as though it was sent a
.\"O .B SIGSTOP
.\"O signal), and must be resumed by sending it a
.\"O .B SIGCONT
.\"O signal.
.B CLONE_STOPPED
¤¬ÀßÄꤵ¤ì¤ë¤È¡¢»Ò¥×¥í¥»¥¹¤ÏºÇ½é
.RB ( SIGSTOP
¥·¥°¥Ê¥ë¤òÁ÷¤é¤ì¤¿¤«¤Î¤è¤¦¤Ë) Ää»ß¤·¤¿¾õÂ֤Ȥʤ롣
»Ò¥×¥í¥»¥¹¤òºÆ³«¤µ¤»¤ë¤Ë¤Ï
.B SIGCONT
¥·¥°¥Ê¥ë¤òÁ÷¿®¤·¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£

.\"O .I "From Linux 2.6.25 this flag is deprecated."
.\"O You probably never wanted to use it,
.\"O you certainly shouldn't be using it, and soon it will go away.
.I "Linux 2.6.25 °Ê¹ß¡¢¤³¤Î¥Õ¥é¥°¤ÏÈó¿ä¾©¤Ç¤¢¤ë¡£"
¤³¤Î¥Õ¥é¥°¤ò»È¤¤¤¿¤¤¤È»×¤¦¤³¤È¤Ï·è¤·¤Æ¤Ê¤¤¤À¤í¤¦¤·¡¢
³Î¼Â¤Ë»È¤ï¤Ê¤¤¤è¤¦¤Ë¤¹¤Ù¤­¤Ç¤¢¤ë¡£
¤³¤Î¥Õ¥é¥°¤Ï¶á¤¤¤¦¤Á¤Ë¤Ê¤¯¤Ê¤ë¤³¤È¤À¤í¤¦¡£
.\" glibc 2.8 removed this defn from bits/sched.h
.TP
.\"O .BR CLONE_SYSVSEM " (since Linux 2.5.10)"
.BR CLONE_SYSVSEM " (Linux 2.5.10 °Ê¹ß)"
.\"O If
.\"O .B CLONE_SYSVSEM
.\"O is set, then the child and the calling process share
.\"O a single list of System V semaphore undo values (see
.\"O .BR semop (2)).
.\"O If this flag is not set, then the child has a separate undo list,
.\"O which is initially empty.
.B CLONE_SYSVSEM
¤¬¥»¥Ã¥È¤µ¤ì¤ë¤È¡¢»Ò¥×¥í¥»¥¹¤È¸Æ¤Ó½Ð¤·¸µ¥×¥í¥»¥¹¤Ï°ì¤Ä¤Î
System V ¥»¥Þ¥Õ¥©¤Î¥¢¥ó¥É¥¥Ãͥꥹ¥È
.RB ( semop (2)
»²¾È) ¤ò¶¦Í­¤¹¤ë¡£¤³¤Î¥Õ¥é¥°¤¬¥»¥Ã¥È¤µ¤ì¤Æ¤¤¤Ê¤±¤ì¤Ð¡¢
»Ò¥×¥í¥»¥¹¤ÏÆȼ«¤Î¥¢¥ó¥É¥¥¥ê¥¹¥È¤ò»ý¤Ä (¥ê¥¹¥È¤Î½é´üÃͤ϶õ¤Ç¤¢¤ë)¡£
.TP
.\"O .BR CLONE_THREAD " (since Linux 2.4.0-test8)"
.BR CLONE_THREAD " (Linux 2.4.0-test8°Ê¹ß)"
.\"O If
.\"O .B CLONE_THREAD
.\"O is set, the child is placed in the same thread group as the calling process.
.\"O To make the remainder of the discussion of
.\"O .B CLONE_THREAD
.\"O more readable, the term "thread" is used to refer to the
.\"O processes within a thread group.
.B CLONE_THREAD
¤¬ÀßÄꤵ¤ì¤¿¾ì¹ç¡¢»Ò¥×¥í¥»¥¹¤Ï¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤ÈƱ¤¸¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤Ë
ÃÖ¤«¤ì¤ë¡£
.B CLONE_THREAD
¤Ë¤Ä¤¤¤Æ¤Î°Ê¹ß¤ÎµÄÏÀ¤òÆɤߤ䤹¤¯¤¹¤ë¤¿¤á¡¢
¡Ö¥¹¥ì¥Ã¥É¡×¤È¤¤¤¦ÍѸì¤Ï¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤ÎÃæ¤Î¥×¥í¥»¥¹¤ò
»²¾È¤¹¤ë¤Î¤Ë»È¤¦¤³¤È¤È¤¹¤ë¡£

.\"O Thread groups were a feature added in Linux 2.4 to support the
.\"O POSIX threads notion of a set of threads that share a single PID.
.\"O Internally, this shared PID is the so-called
.\"O thread group identifier (TGID) for the thread group.
.\"O Since Linux 2.4, calls to
.\"O .BR getpid (2)
.\"O return the TGID of the caller.
¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤Ï¡¢ ¥¹¥ì¥Ã¥É½¸¹ç¤Ç°ì¤Ä¤Î PID ¤ò¶¦Í­¤¹¤ë¤È¤¤¤¦ POSIX
¥¹¥ì¥Ã¥É¤Î³µÇ°¤ò¥µ¥Ý¡¼¥È¤¹¤ë¤¿¤á¤Ë Linux 2.4 ¤Ë²Ã¤¨¤é¤ì¤¿µ¡Ç½¤Ç¤¢¤Ã¤¿¡£
ÆâÉôŪ¤Ë¤Ï¡¢¤³¤Î¶¦Í­ PID ¤Ï¤¤¤ï¤æ¤ë¤½¤Î¥¹¥ì¥Ã¥É¥°¥ë¡¼¥×¤Î
¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¼±ÊÌ»Ò (TGID) ¤Ç¤¢¤ë¡£
Linux 2.4 °Ê¹ß¤Ç¤Ï¡¢
.BR getpid (2)
¤Î¸Æ¤Ó½Ð¤·¤Ç¤Ï¤½¤Î¥×¥í¥»¥¹¤Î¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥× ID ¤òÊÖ¤¹¡£

.\"O The threads within a group can be distinguished by their (system-wide)
.\"O unique thread IDs (TID).
.\"O A new thread's TID is available as the function result
.\"O returned to the caller of
.\"O .BR clone (),
.\"O and a thread can obtain
.\"O its own TID using
.\"O .BR gettid (2).
¤¢¤ë¥°¥ë¡¼¥×¤Ë°¤¹¤ë¥¹¥ì¥Ã¥É¤Ï (¥·¥¹¥Æ¥àÁ´ÂΤÇ) °ì°Õ¤Ê¥¹¥ì¥Ã¥É ID (TID)
¤Ç¶èÊ̤Ǥ­¤ë¡£¿·¤·¤¤¥¹¥ì¥Ã¥É¤Î TID ¤Ï
.BR clone ()
¤Î¸Æ¤Ó½Ð¤·¸µ¤Ø´Ø¿ô¤Î·ë²Ì¤È¤·¤ÆÊÖ¤µ¤ì¡¢
¥¹¥ì¥Ã¥É¤Ï¼«Ê¬¼«¿È¤Î TID ¤ò
.BR gettid (2)
¤Ç¼èÆÀ¤Ç¤­¤ë¡£

.\"O When a call is made to
.\"O .BR clone ()
.\"O without specifying
.\"O .BR CLONE_THREAD ,
.\"O then the resulting thread is placed in a new thread group
.\"O whose TGID is the same as the thread's TID.
.\"O This thread is the
.\"O .I leader
.\"O of the new thread group.
.B CLONE_THREAD
¤ò»ØÄꤻ¤º¤Ë
.BR clone ()
¤Î¸Æ¤Ó½Ð¤·¤¬¹Ô¤ï¤ì¤ë¤È¡¢
À¸À®¤µ¤ì¤¿¥¹¥ì¥Ã¥É¤Ï¤½¤Î¥¹¥ì¥Ã¥É¤Î TID ¤ÈƱ¤¸ÃͤΠTGID ¤ò»ý¤Ä
¿·¤·¤¤¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤ËÃÖ¤«¤ì¤ë¡£¤³¤Î¥¹¥ì¥Ã¥É¤Ï
¿·¤·¤¤¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤Î¡Ö¥ê¡¼¥À¡¼¡×¤Ç¤¢¤ë¡£

.\"O A new thread created with
.\"O .B CLONE_THREAD
.\"O has the same parent process as the caller of
.\"O .BR clone ()
.\"O (i.e., like
.\"O .BR CLONE_PARENT ),
.\"O so that calls to
.\"O .BR getppid (2)
.\"O return the same value for all of the threads in a thread group.
.B CLONE_THREAD
¤ò»ØÄꤷ¤ÆºîÀ®¤µ¤ì¤¿¿·¤·¤¤¥¹¥ì¥Ã¥É¤Ï¡¢
.RB ( CLONE_PARENT
¤Î¾ì¹ç¤ÈƱÍͤË)
.BR clone ()
¤ò¸Æ¤Ó½Ð¤·¸µ¤ÈƱ¤¸¿Æ¥×¥í¥»¥¹¤ò»ý¤Ä¡£
¤½¤Î¤¿¤á¡¢
.BR getppid (2)
¤ò¸Æ¤Ö¤È¡¢°ì¤Ä¤Î¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤Ë°¤¹¥¹¥ì¥Ã¥É¤ÏÁ´¤ÆƱ¤¸ÃͤòÊÖ¤¹¡£
.\"O When a
.\"O .B CLONE_THREAD
.\"O thread terminates, the thread that created it using
.\"O .BR clone ()
.\"O is not sent a
.\"O .B SIGCHLD
.\"O (or other termination) signal;
.\"O nor can the status of such a thread be obtained
.\"O using
.\"O .BR wait (2).
.\"O (The thread is said to be
.\"O .IR detached .)
.B CLONE_THREAD
¤Çºî¤é¤ì¤¿¥¹¥ì¥Ã¥É¤¬½ªÎ»¤·¤¿ºÝ¤Ë¡¢
¤½¤Î¥¹¥ì¥Ã¥É¤ò
.BR clone (2)
¤ò»È¤Ã¤ÆÀ¸À®¤·¤¿¥¹¥ì¥Ã¥É¤Ë¤Ï
.B SIGCHLD
(¤â¤·¤¯¤Ï¾¤Î½ªÎ»¥·¥°¥Ê¥ë) ¤ÏÁ÷¿®¤µ¤ì¤Ê¤¤¡£
¤Þ¤¿¡¢
.BR wait (2)
¤ò»È¤Ã¤Æ½ªÎ»¤·¤¿¥¹¥ì¥Ã¥É¤Î¾õÂÖ¤ò¼èÆÀ¤¹¤ë¤³¤È¤â¤Ç¤­¤Ê¤¤
(¤½¤Î¤è¤¦¤Ê¥¹¥ì¥Ã¥É¤Ï
.I detached
(ʬΥ¤µ¤ì¤¿) ¤È¤¤¤ï¤ì¤ë)¡£

.\"O After all of the threads in a thread group terminate
.\"O the parent process of the thread group is sent a
.\"O .B SIGCHLD
.\"O (or other termination) signal.
¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤Ë°¤¹Á´¤Æ¤Î¥¹¥ì¥Ã¥É¤¬½ªÎ»¤·¤¿¸å¡¢
¤½¤Î¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤Î¿Æ¥×¥í¥»¥¹¤Ë
.B SIGCHLD
(¤â¤·¤¯¤Ï¾¤Î½ªÎ»¥·¥°¥Ê¥ë) ¤¬Á÷¤é¤ì¤ë¡£

.\"O If any of the threads in a thread group performs an
.\"O .BR execve (2),
.\"O then all threads other than the thread group leader are terminated,
.\"O and the new program is executed in the thread group leader.
¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤Ë°¤¹¤¤¤º¤ì¤«¤Î¥¹¥ì¥Ã¥É¤¬
.BR execve (2)
¤ò¼Â¹Ô¤¹¤ë¤È¡¢¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¡¦¥ê¡¼¥À¡¼°Ê³°¤ÎÁ´¤Æ¤Î¥¹¥ì¥Ã¥É¤Ï
½ªÎ»¤µ¤ì¡¢¿·¤·¤¤¥×¥í¥»¥¹¤¬¤½¤Î¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¡¦¥ê¡¼¥À¡¼¤Î²¼¤Ç
¼Â¹Ô¤µ¤ì¤ë¡£

.\"O If one of the threads in a thread group creates a child using
.\"O .BR fork (2),
.\"O then any thread in the group can
.\"O .BR wait (2)
.\"O for that child.
¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤Ë°¤¹¥¹¥ì¥Ã¥É¤Î°ì¤Ä¤¬
.BR fork (2)
¤ò»È¤Ã¤Æ»Ò¥×¥í¥»¥¹¤òºîÀ®¤·¤¿¾ì¹ç¡¢
¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤Î¤É¤Î¥¹¥ì¥Ã¥É¤Ç¤¢¤Ã¤Æ¤â
¤½¤Î»Ò¶¡¤ò
.BR wait (2)
¤Ç¤­¤ë¡£

.\"O Since Linux 2.5.35,
.\"O .I flags
.\"O must also include
.\"O .B CLONE_SIGHAND
.\"O if
.\"O .B CLONE_THREAD
.\"O is specified.
Linux 2.5.35 °Ê¹ß¤Ç¤Ï¡¢
.B CLONE_THREAD
¤ò»ØÄꤹ¤ë¾ì¹ç¡¢
.I flags
¤Ë
.B CLONE_SIGHAND
¤â´Þ¤Þ¤ì¤Æ¤¤¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£

.\"O Signals may be sent to a thread group as a whole (i.e., a TGID) using
.\"O .BR kill (2),
.\"O or to a specific thread (i.e., TID) using
.\"O .BR tgkill (2).
.BR kill (2)
¤ò»È¤Ã¤Æ¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×Á´ÂÎ (¤Ä¤Þ¤ê TGID) ¤Ë¥·¥°¥Ê¥ë¤òÁ÷¤ë¤³¤È¤â¤Ç¤­¤ì¤Ð¡¢
.BR tgkill (2)
¤ò»È¤Ã¤ÆÆÃÄê¤Î¥¹¥ì¥Ã¥É (¤Ä¤Þ¤ê TID) ¤Ë¥·¥°¥Ê¥ë¤òÁ÷¤ë¤³¤È¤â¤Ç¤­¤ë¡£

.\"O Signal dispositions and actions are process-wide:
.\"O if an unhandled signal is delivered to a thread, then
.\"O it will affect (terminate, stop, continue, be ignored in)
.\"O all members of the thread group.
¥·¥°¥Ê¥ë¤ÎÇÛÁ÷¤È½èÍý¤Ï¥×¥í¥»¥¹Á´ÂΤ˱ƶÁ¤¹¤ë:
¥Ï¥ó¥É¥é¤òÀßÄꤷ¤Æ¤¤¤Ê¤¤¥·¥°¥Ê¥ë¤¬¤¢¤ë¥¹¥ì¥Ã¥É¤ËÇÛÁ÷¤µ¤ì¤ë¤È¡¢
¤½¤Î¥·¥°¥Ê¥ë¤Ï¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤ÎÁ´¥á¥ó¥Ð¡¼¤Ë±Æ¶Á¤òµÚ¤Ü¤¹
(½ªÎ»¤·¤¿¤ê¡¢Ää»ß¤·¤¿¤ê¡¢Æ°ºî¤ò·Ñ³¤·¤¿¤ê¡¢Ìµ»ë¤µ¤ì¤¿¤ê¤¹¤ë)¡£

.\"O Each thread has its own signal mask, as set by
.\"O .BR sigprocmask (2),
.\"O but signals can be pending either: for the whole process
.\"O (i.e., deliverable to any member of the thread group),
.\"O when sent with
.\"O .BR kill (2);
.\"O or for an individual thread, when sent with
.\"O .BR tgkill (2).
.\"O A call to
.\"O .BR sigpending (2)
.\"O returns a signal set that is the union of the signals pending for the
.\"O whole process and the signals that are pending for the calling thread.
³Æ¡¹¤Î¥¹¥ì¥Ã¥É¤ÏÆȼ«¤Î¥·¥°¥Ê¥ë¥Þ¥¹¥¯¤ò»ý¤Ã¤Æ¤ª¤ê¡¢
.BR sigprocmask (2)
¤ÇÀßÄê¤Ç¤­¤ë¡£
¤À¤¬¡¢½èÍýÂÔ¤Á¤Î¥·¥°¥Ê¥ë¤Ë¤Ï¡¢
.BR kill (2)
¤ÇÁ÷¿®¤µ¤ì¤ë¥×¥í¥»¥¹Á´ÂΤËÂФ¹¤ë¤â¤Î (¤Ä¤Þ¤ê¡¢¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤Î
¤É¤Î¥á¥ó¥Ð¡¼¤Ë¤âÇÛÁ÷¤Ç¤­¤ë¤â¤Î) ¤È¡¢
.BR tgkill (2)
¤ÇÁ÷¿®¤µ¤ì¤ë¸Ä¡¹¤Î¥¹¥ì¥Ã¥É¤ËÂФ¹¤ë¤â¤Î¤¬¤¢¤ê¤¨¤ë¡£
.BR sigpending (2)
¤ò¸Æ¤Ó½Ð¤¹¤È¡¢¥×¥í¥»¥¹Á´ÂΤËÂФ¹¤ë½èÍýÂÔ¤Á¥·¥°¥Ê¥ë¤È¸Æ¤Ó½Ð¤·¸µ¤Î
¥¹¥ì¥Ã¥É¤ËÂФ¹¤ë½èÍýÂÔ¤Á¥·¥°¥Ê¥ë¤ò·ë¹ç¤·¤¿¥·¥°¥Ê¥ë½¸¹ç¤¬ÊÖ¤µ¤ì¤ë¡£

.\"O If
.\"O .BR kill (2)
.\"O is used to send a signal to a thread group,
.\"O and the thread group has installed a handler for the signal, then
.\"O the handler will be invoked in exactly one, arbitrarily selected
.\"O member of the thread group that has not blocked the signal.
.\"O If multiple threads in a group are waiting to accept the same signal using
.\"O .BR sigwaitinfo (2),
.\"O the kernel will arbitrarily select one of these threads
.\"O to receive a signal sent using
.\"O .BR kill (2).
.BR kill (2)
¤ò»È¤Ã¤Æ¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤Ë¥·¥°¥Ê¥ë¤¬Á÷¤é¤ì¤¿¾ì¹ç¤Ç¡¢
¤½¤Î¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤¬¤½¤Î¥·¥°¥Ê¥ë¤ËÂФ¹¤ë¥·¥°¥Ê¥ë¡¦¥Ï¥ó¥É¥é¤¬
ÅÐÏ¿¤µ¤ì¤Æ¤¤¤¿¤È¤­¤Ë¤Ï¡¢¥·¥°¥Ê¥ë¡¦¥Ï¥ó¥É¥é¤Ï¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤Î
¥á¥ó¥Ð¡¼¤Î¤¦¤Á¡¢¤¿¤À°ì¤Ä¤Î¥¹¥ì¥Ã¥É¤Ç¤À¤±µ¯Æ°¤µ¤ì¤ë¡£¥Ï¥ó¥É¥é¤¬
µ¯Æ°¤µ¤ì¤ë¥¹¥ì¥Ã¥É¤Ï¡¢¤½¤Î¥·¥°¥Ê¥ë¤ò¶Ø»ß (block) ¤·¤Æ¤¤¤Ê¤¤
¥á¥ó¥Ð¡¼¤ÎÃ椫¤é°ì¤Ä¤À¤±¤¬¾¡¼ê¤Ë (arbitrarily) Áª¤Ð¤ì¤ë¡£
¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤Ë°¤¹Ê£¿ô¤Î¥¹¥ì¥Ã¥É¤¬
.BR sigwaitinfo (2)
¤ò»È¤Ã¤ÆƱ¤¸¥·¥°¥Ê¥ë¤òÂԤäƤ¤¤ë¾ì¹ç¡¢
¤³¤ì¤é¤Î¥¹¥ì¥Ã¥É¤ÎÃ椫¤é°ì¤Ä¤ò¥«¡¼¥Í¥ë¤¬¾¡¼ê¤ËÁªÂò¤·¡¢
¤½¤Î¥¹¥ì¥Ã¥É¤¬
.B kill (2)
¤ò»È¤Ã¤ÆÁ÷¿®¤µ¤ì¤¿¥·¥°¥Ê¥ë¤ò¼õ¿®¤¹¤ë¡£
.TP
.\"O .BR CLONE_UNTRACED " (since Linux 2.5.46)"
.BR CLONE_UNTRACED " (Linux 2.5.46 °Ê¹ß)"
.\"O If
.\"O .B CLONE_UNTRACED
.\"O is specified, then a tracing process cannot force
.\"O .B CLONE_PTRACE
.\"O on this child process.
.B CLONE_UNTRACED
¤¬»ØÄꤵ¤ì¤ë¤È¡¢ trace ¤ò¹Ô¤Ã¤Æ¤¤¤ë¥×¥í¥»¥¹¤Ï
¤³¤Î»Ò¥×¥í¥»¥¹¤Ë
.B CLONE_PTRACE
¤òŬÍѤ¹¤ë¤³¤È¤¬¤Ç¤­¤Ê¤¤¡£
.TP
.B CLONE_VFORK
.\"O If
.\"O .B CLONE_VFORK
.\"O is set, the execution of the calling process is suspended
.\"O until the child releases its virtual memory
.\"O resources via a call to
.\"O .BR execve (2)
.\"O or
.\"O .BR _exit (2)
.\"O (as with
.\"O .BR vfork (2)).
.B CLONE_VFORK
¤¬ÀßÄꤵ¤ì¤¿¾ì¹ç¡¢
.RB ( vfork (2)
¤ÈƱÍͤË) »Ò¥×¥í¥»¥¹¤¬
.BR execve (2)
¤Þ¤¿¤Ï
.BR _exit (2)
¤Ë¤è¤Ã¤Æ²¾ÁÛ¥á¥â¥ê¤ò²òÊü¤¹¤ë¤Þ¤Ç¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤Î¼Â¹Ô¤ÏÄä»ß¤µ¤ì¤ë¡£

.\"O If
.\"O .B CLONE_VFORK
.\"O is not set then both the calling process and the child are schedulable
.\"O after the call, and an application should not rely on execution occurring
.\"O in any particular order.
.B CLONE_VFORK
¤¬ÀßÄꤵ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç¡¢
.BR clone ()
¸Æ¤Ó½Ð¤·¸å¤Ï¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤È»Ò¥×¥í¥»¥¹¤Î
ξÊý¤¬¥¹¥±¥¸¥å¡¼¥ëÂоݤȤʤꡢ¥¢¥×¥ê¥±¡¼¥·¥ç¥ó¤Ï¤³¤ì¤é¤Î¥×¥í¥»¥¹¤Î
¼Â¹Ô½ç½ø¤Ë°Í¸¤·¤Ê¤¤¤è¤¦¤Ë¤¹¤Ù¤­¤Ç¤¢¤ë¡£
.TP
.B CLONE_VM
.\"O If
.\"O .B CLONE_VM
.\"O is set, the calling process and the child process run in the same memory
.\"O space.
.\"O In particular, memory writes performed by the calling process
.\"O or by the child process are also visible in the other process.
.\"O Moreover, any memory mapping or unmapping performed with
.\"O .BR mmap (2)
.\"O or
.\"O .BR munmap (2)
.\"O by the child or calling process also affects the other process.
.B CLONE_VM
¤¬ÀßÄꤵ¤ì¤¿¾ì¹ç¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤È»Ò¥×¥í¥»¥¹¤ÏƱ¤¸¥á¥â¥ê¶õ´Ö¤Ç
¼Â¹Ô¤µ¤ì¤ë¡£Æäˡ¢¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤ä»Ò¥×¥í¥»¥¹¤Î°ìÊý¤¬¥á¥â¥ê¤Ë
½ñ¤­¹þ¤ó¤ÀÆâÍƤϤ⤦°ìÊý¤Î¥×¥í¥»¥¹¤«¤é¤â¸«¤ë¤³¤È¤¬¤Ç¤­¤ë¡£¤µ¤é¤Ë¡¢
»Ò¥×¥í¥»¥¹¤ä¸Æ¤Ó½Ð¤·¸µ¤Î¥×¥í¥»¥¹¤Î°ìÊý¤¬
.BR mmap (2)
¤ä
.BR munmap (2)
¤ò»È¤Ã¤Æ¥á¥â¥ê¤ò¥Þ¥Ã¥×¤·¤¿¤ê¥¢¥ó¥Þ¥Ã¥×¤·¤¿¾ì¹ç¡¢
¤â¤¦°ìÊý¤Î¥×¥í¥»¥¹¤Ë¤â±Æ¶Á¤¬µÚ¤Ö¡£

.\"O If
.\"O .B CLONE_VM
.\"O is not set, the child process runs in a separate copy of the memory
.\"O space of the calling process at the time of
.\"O .BR clone ().
.\"O Memory writes or file mappings/unmappings performed by one of the
.\"O processes do not affect the other, as with
.\"O .BR fork (2).
.B CLONE_VM
¤¬ÀßÄꤵ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç¡¢»Ò¥×¥í¥»¥¹¤Ï
.BR clone ()
¤¬¼Â¹Ô¤µ¤ì¤¿»þÅÀ¤Ç¤Î¡¢¿Æ¥×¥í¥»¥¹¤Î¥á¥â¥ê¶õ´Ö¤ò¥³¥Ô¡¼¤·¤¿
Ê̤Υá¥â¥ê¶õ´Ö¤Ç¼Â¹Ô¤µ¤ì¤ë¡£
°ìÊý¤Î¥×¥í¥»¥¹¤¬¹Ô¤Ã¤¿¥á¥â¥ê¤Ø¤Î½ñ¤­¹þ¤ß¤ä
¥Õ¥¡¥¤¥ë¤Î¥Þ¥Ã¥×/¥¢¥ó¥Þ¥Ã¥×¤Ï¡¢
.BR fork (2)
¤Î¾ì¹ç¤ÈƱÍÍ¡¢¤â¤¦°ìÊý¤Î¥×¥í¥»¥¹¤Ë¤Ï±Æ¶Á¤·¤Ê¤¤¡£
.SS "sys_clone"
.\"O The
.\"O .B sys_clone
.\"O system call corresponds more closely to
.\"O .BR fork (2)
.\"O in that execution in the child continues from the point of the
.\"O call.
.\"O Thus,
.\"O .B sys_clone
.\"O only requires the
.\"O .I flags
.\"O and
.\"O .I child_stack
.\"O arguments, which have the same meaning as for
.\"O .BR clone ().
.\"O (Note that the order of these arguments differs from
.\"O .BR clone ().)
.B sys_clone
¥·¥¹¥Æ¥à¥³¡¼¥ë¤Ï¡¢¤è¤ê
.BR fork (2)
¤Ë¶á¤¤¤«¤¿¤Á¤Ë¤Ê¤Ã¤Æ¤ª¤ê¡¢»Ò¥×¥í¥»¥¹¤Î¼Â¹Ô¤¬¸Æ¤Ó½Ð¤·¤¬¹Ô¤ï¤ì¤¿¾ì½ê¤«¤é
³¤±¤é¤ì¤ë¡£
¤½¤Î¤¿¤á¡¢
.B sys_clone
¤¬É¬ÍפȤ¹¤ë°ú¤­¿ô¤Ï
.I flags
¤È
.I child_stack
¤À¤±¤Ç¤¢¤ê¡¢¤½¤ì¤é¤Ï
.BR clone ()
¤ÈƱ¤¸°ÕÌ£¤ò»ý¤Ä
(¤³¤ì¤é¤Î°ú¤­¿ô¤Î½çÈÖ¤Ï
.BR clone ()
¤È¤Ï°Û¤Ê¤ë¤³¤È¤ËÃí°Õ¤»¤è)¡£

.\"O Another difference for
.\"O .B sys_clone
.\"O is that the
.\"O .I child_stack
.\"O argument may be zero, in which case copy-on-write semantics ensure that the
.\"O child gets separate copies of stack pages when either process modifies
.\"O the stack.
.\"O In this case, for correct operation, the
.\"O .B CLONE_VM
.\"O option should not be specified.
.B sys_clone
¤Î¤â¤¦°ì¤Ä¤Î°ã¤¤¤Ï¡¢
.I child_stack
°ú¤­¿ô¤¬¥¼¥í¤Ç¤âÎɤ¤¤³¤È¤Ç¤¢¤ë¡£¤³¤Î¾ì¹ç¤Ë¤Ï¡¢¤É¤Á¤é¤«¤Î¥×¥í¥»¥¹¤¬
¥¹¥¿¥Ã¥¯¤òÊѹ¹¤·¤¿»þ¤Ë¡¢½ñ¤­¹þ¤ß»þ¥³¥Ô¡¼ (copy-on-write) Êý¼°¤Ë¤è¤ê
»Ò¥×¥í¥»¥¹¤¬¥¹¥¿¥Ã¥¯¡¦¥Ú¡¼¥¸¤ÎÆÈΩ¤·¤¿¥³¥Ô¡¼¤òÆÀ¤é¤ì¤ë¤³¤È¤¬Êݾڤµ¤ì¤ë¡£
¤³¤Î¾ì¹ç¡¢Àµ¾ï¤ËÆ°ºî¤µ¤»¤ë¤¿¤á¤Ë¤Ï¡¢
.B CLONE_VM
¥ª¥×¥·¥ç¥ó¤ò»ØÄꤷ¤Æ¤Ï¤Ê¤é¤Ê¤¤¡£

.\"O In Linux 2.4 and earlier,
.\"O .BR clone ()
.\"O does not take arguments
.\"O .IR ptid ,
.\"O .IR tls ,
.\"O and
.\"O .IR ctid .
Linux 2.4 °ÊÁ°¤Ç¤Ï¡¢
.BR clone ()
¤Ï°ú¤­¿ô
.IR ptid ,
.IR tls ,
.IR ctid
¤ò¼è¤é¤Ê¤¤¡£
.\"O .SH "RETURN VALUE"
.SH ÊÖ¤êÃÍ
.\"O .\" gettid(2) returns current->pid;
.\"O .\" getpid(2) returns current->tgid;
.\"O On success, the thread ID of the child process is returned
.\"O in the caller's thread of execution.
.\"O On failure, \-1 is returned
.\"O in the caller's context, no child process will be created, and
.\"O .I errno
.\"O will be set appropriately.
.\" gettid(2) ¤Ï current->pid ¤òÊÖ¤¹
.\" getpid(2) ¤Ï current->tgid ¤òÊÖ¤¹
À®¸ù¤·¤¿¾ì¹ç¡¢¸Æ¤Ó½Ð¤·¸µ¤Î¼Â¹Ô¥¹¥ì¥Ã¥É¤Ë¤Ï»Ò¥×¥í¥»¥¹¤Î¥¹¥ì¥Ã¥ÉID ¤¬ÊÖ¤µ¤ì¤ë¡£
¼ºÇÔ¤·¤¿¾ì¹ç¡¢ ¸Æ¤Ó½Ð¤·¸µ¤Î¥³¥ó¥Æ¥­¥¹¥È¤Ë¤Ï \-1 ¤¬ÊÖ¤µ¤ì¡¢»Ò¥×¥í¥»¥¹¤Ï
ºîÀ®¤µ¤ì¤º¡¢
.I errno
¤¬Å¬ÀÚ¤ËÀßÄꤵ¤ì¤ë¡£
.\"O .SH ERRORS
.SH ¥¨¥é¡¼
.TP
.B EAGAIN
.\"O Too many processes are already running.
¤¹¤Ç¤Ë¼Â¹ÔÃæ¤Î¥×¥í¥»¥¹¤¬Â¿¤¹¤®¤ë¡£
.TP
.B EINVAL
.\"O .B CLONE_SIGHAND
.\"O was specified, but
.\"O .B CLONE_VM
.\"O was not.
.\"O (Since Linux 2.6.0-test6.)
.B CLONE_SIGHAND
¤¬»ØÄꤵ¤ì¤Æ¤¤¤¿¤¬¡¢
.B CLONE_VM
¤¬»ØÄꤵ¤ì¤Æ¤¤¤Ê¤«¤Ã¤¿¡£
(Linux 2.6.0-test6 °Ê¹ß)
.TP
.B EINVAL
.\"O .B CLONE_THREAD
.\"O was specified, but
.\"O .B CLONE_SIGHAND
.\"O was not.
.\"O (Since Linux 2.5.35.)
.B CLONE_THREAD
¤¬»ØÄꤵ¤ì¤Æ¤¤¤¿¤¬¡¢
.B CLONE_SIGHAND
¤¬»ØÄꤵ¤ì¤Æ¤¤¤Ê¤«¤Ã¤¿¡£
(Linux 2.5.35 °Ê¹ß)
.\" .TP
.\" .B EINVAL
.\"O .\" Precisely one of
.\"O .\" .B CLONE_DETACHED
.\"O .\" and
.\"O .\" .B CLONE_THREAD
.\"O .\" was specified.
.\"O .\" (Since Linux 2.6.0-test6.)
.\" .B CLONE_DETACHED
.\" ¤È
.\" .B CLONE_THREAD
.\" ¤Î¤¤¤º¤ì¤«°ìÊý¤À¤±¤¬»ØÄꤵ¤ì¤¿¡£
.\" (Linux 2.6.0-test6 °Ê¹ß)
.TP
.B EINVAL
.\"O Both
.\"O .B CLONE_FS
.\"O and
.\"O .B CLONE_NEWNS
.\"O were specified in
.\"O .IR flags .
.B CLONE_FS
¤È
.B CLONE_NEWNS
¤ÎξÊý¤¬
.I flags
¤Ë»ØÄꤵ¤ì¤¿¡£
.TP
.B EINVAL
.\"O Both
.\"O .B CLONE_NEWIPC
.\"O and
.\"O .B CLONE_SYSVSEM
.\"O were specified in
.\"O .IR flags .
.B CLONE_NEWIPC
¤È
.B CLONE_SYSVSEM
¤ÎξÊý¤¬
.I flags
¤Ë»ØÄꤵ¤ì¤¿¡£
.TP
.B EINVAL
.\"O Both
.\"O .BR CLONE_NEWPID
.\"O and
.\"O .BR CLONE_THREAD
.\"O were specified in
.\"O .IR flags .
.B CLONE_NEWPID
¤È
.B CLONE_THREAD
¤ÎξÊý¤¬
.I flags
¤Ë»ØÄꤵ¤ì¤¿¡£
.TP
.B EINVAL
.\"O Returned by
.\"O .BR clone ()
.\"O when a zero value is specified for
.\"O .IR child_stack .
.I child_stack
¤Ë¥¼¥í¤ò»ØÄꤷ¤¿¾ì¹ç¤Ë
.BR clone ()
¤¬ÊÖ¤¹¡£
.TP
.B EINVAL
.\"O .BR CLONE_NEWIPC
.\"O was specified in
.\"O .IR flags ,
.\"O but the kernel was not configured with the
.\"O .B CONFIG_SYSVIPC
.\"O and
.\"O .BR CONFIG_IPC_NS
.\"O options.
.I flags
¤Ë
.B CLONE_NEWIPC
¤¬»ØÄꤵ¤ì¤¿¤¬¡¢¥«¡¼¥Í¥ë¤Ç¥ª¥×¥·¥ç¥ó
.B CONFIG_SYSVIPC
¤È
.B CONFIG_IPC_NS
¤¬Í­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤Ê¤«¤Ã¤¿¡£
.TP
.B EINVAL
.\"O .BR CLONE_NEWNET
.\"O was specified in
.\"O .IR flags ,
.\"O but the kernel was not configured with the
.\"O .B CONFIG_NET_NS
.\"O option.
.I flags
¤Ë
.B CLONE_NEWNET
¤¬»ØÄꤵ¤ì¤¿¤¬¡¢¥«¡¼¥Í¥ë¤Ç¥ª¥×¥·¥ç¥ó
.B CONFIG_NET_NS
¤¬Í­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤Ê¤«¤Ã¤¿¡£
.TP
.B EINVAL
.\"O .BR CLONE_NEWPID
.\"O was specified in
.\"O .IR flags ,
.\"O but the kernel was not configured with the
.\"O .B CONFIG_PID_NS
.\"O option.
.I flags
¤Ë
.B CLONE_NEWPID
¤¬»ØÄꤵ¤ì¤¿¤¬¡¢¥«¡¼¥Í¥ë¤Ç¥ª¥×¥·¥ç¥ó
.B CONFIG_PID_NS
¤¬Í­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤Ê¤«¤Ã¤¿¡£
.TP
.B EINVAL
.\"O .BR CLONE_NEWUTS
.\"O was specified in
.\"O .IR flags ,
.\"O but the kernel was not configured with the
.\"O .B CONFIG_UTS
.\"O option.
.I flags
¤Ë
.B CLONE_NEWUTS
¤¬»ØÄꤵ¤ì¤¿¤¬¡¢¥«¡¼¥Í¥ë¤Ç¥ª¥×¥·¥ç¥ó
.B CONFIG_UTS
¤¬Í­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤Ê¤«¤Ã¤¿¡£
.TP
.B ENOMEM
.\"O Cannot allocate sufficient memory to allocate a task structure for the
.\"O child, or to copy those parts of the caller's context that need to be
.\"O copied.
»Ò¥×¥í¥»¥¹¤Î¤¿¤á¤Ë³ÎÊݤ¹¤Ù¤­¥¿¥¹¥¯¹½Â¤ÂΤ䡢¸Æ¤Ó½Ð¤·¸µ¤Î¥³¥ó¥Æ¥­¥¹¥È¤Î
°ìÉô¤ò¥³¥Ô¡¼¤¹¤ë¤Î¤ËɬÍפʥá¥â¥ê¤ò½½Ê¬¤Ë³ä¤êÅö¤Æ¤ë¤³¤È¤¬¤Ç¤­¤Ê¤¤¡£
.TP
.B EPERM
.\"O .BR CLONE_NEWIPC ,
.\"O .BR CLONE_NEWNET ,
.\"O .BR CLONE_NEWNS ,
.\"O .BR CLONE_NEWPID ,
.\"O or
.\"O .BR CLONE_NEWUTS
.\"O was specified by an unprivileged process (process without \fBCAP_SYS_ADMIN\fP).
ÈóÆø¢¥×¥í¥»¥¹ (\fBCAP_SYS_ADMIN\fP ¤ò»ý¤¿¤Ê¤¤¥×¥í¥»¥¹) ¤¬
.BR CLONE_NEWIPC ,
.BR CLONE_NEWNET ,
.BR CLONE_NEWNS ,
.BR CLONE_NEWPID ,
.BR CLONE_NEWUTS
¤ò»ØÄꤷ¤¿¡£
.TP
.B EPERM
.\"O .B CLONE_PID
.\"O was specified by a process other than process 0.
PID ¤¬ 0 °Ê³°¤Î¥×¥í¥»¥¹¤Ë¤è¤Ã¤Æ
.B CLONE_PID
¤¬»ØÄꤵ¤ì¤¿¡£
.\"O .SH VERSIONS
.SH ¥Ð¡¼¥¸¥ç¥ó
.\"O There is no entry for
.\"O .BR clone ()
.\"O in libc5.
.\"O glibc2 provides
.\"O .BR clone ()
.\"O as described in this manual page.
libc5 ¤Ë¤Ï
.BR clone ()
¤Ï¤Ê¤¤¡£glibc2 ¤Ç¤Ï
.BR clone ()
¤¬Ä󶡤µ¤ì¤Æ¤ª¤ê¡¢¤³¤Î¥Þ¥Ë¥å¥¢¥ë¥Ú¡¼¥¸¤Ëµ­ºÜ¤ÎÄ̤ê¤Ç¤¢¤ë¡£
.\"O .SH "CONFORMING TO"
.SH ½àµò
.\"O The
.\"O .BR clone ()
.\"O and
.\"O .B sys_clone
.\"O calls are Linux-specific and should not be used in programs
.\"O intended to be portable.
.BR clone ()
¤È
.B sys_clone
¥³¡¼¥ë¤Ï Linux ÆÃÍ­¤Ç¤¢¤ê¡¢°Ü¿¢¤ò¹Íθ¤·¤¿¥×¥í¥°¥é¥à¤Ç¤Ï»ÈÍѤ¹¤Ù¤­
¤Ç¤Ï¤Ê¤¤¡£
.\"O .SH NOTES
.SH Ãí°Õ
.\"O In the kernel 2.4.x series,
.\"O .B CLONE_THREAD
.\"O generally does not make the parent of the new thread the same
.\"O as the parent of the calling process.
.\"O However, for kernel versions 2.4.7 to 2.4.18 the
.\"O .B CLONE_THREAD
.\"O flag implied the
.\"O .B CLONE_PARENT
.\"O flag (as in kernel 2.6).
¥«¡¼¥Í¥ë 2.4.x ·ÏÎó¤Ç¤Ï¡¢°ìÈÌŪ¤Ë¤Ï
.B CLONE_THREAD
¥Õ¥é¥°¤ò»ØÄꤷ¤Æ¤â¿·¤·¤¤¥¹¥ì¥Ã¥É¤Î¿Æ¤ò
¸Æ¤Ó½Ð¤·¸µ¥×¥í¥»¥¹¤Î¿Æ¤ÈƱ¤¸¤Ë¤Ï¤·¤Ê¤¤¡£
¤·¤«¤·¡¢¥Ð¡¼¥¸¥ç¥ó 2.4.7¡Á2.4.18 ¤Î¥«¡¼¥Í¥ë¤Ç¤Ï¡¢
(¥«¡¼¥Í¥ë 2.6 ¤ÈƱ¤¸¤è¤¦¤Ë) CLONE_THREAD ¥Õ¥é¥°¤ò»ØÄꤹ¤ë¤È¡¢
°ÅÌۤΤ¦¤Á¤Ë CLONE_PARENT ¥Õ¥é¥°¤ò»ØÄꤷ¤¿¤³¤È¤Ë¤Ê¤ë¡£

.\"O For a while there was
.\"O .B CLONE_DETACHED
.\"O (introduced in 2.5.32):
.\"O parent wants no child-exit signal.
.\"O In 2.6.2 the need to give this
.\"O together with
.\"O .B CLONE_THREAD
.\"O disappeared.
.\"O This flag is still defined, but has no effect.
.B CLONE_DETACHED
¤È¤¤¤¦¥Õ¥é¥°¤¬¡¢2.5.32 ¤ÇƳÆþ¤µ¤ì¤Æ°ÊÍ褷¤Ð¤é¤¯¤Î´Ö¸ºß¤·¤¿¡£
¤³¤Î¥Õ¥é¥°¤Ï¿Æ¥×¥í¥»¥¹¤¬»Ò¥×¥í¥»¥¹½ªÎ»¤Î¥·¥°¥Ê¥ë¤òɬÍפȤ·¤Ê¤¤¤³¤È¤ò
ɽ¤¹¤â¤Î¤Ç¤¢¤ë¡£
2.6.2 ¤Ç¡¢ CLONE_DETATCHED ¤ò CLONE_THREAD ¤È°ì½ï¤Ë»ØÄꤹ¤ëɬÍפϤʤ¯¤Ê¤Ã¤¿¡£
¤³¤Î¥Õ¥é¥°¤Ï¤Þ¤ÀÄêµÁ¤µ¤ì¤Æ¤¤¤ë¤¬¡¢²¿¤Î¸ú²Ì¤â¤Ê¤¤¡£

.\"O On i386,
.\"O .BR clone ()
.\"O should not be called through vsyscall, but directly through
.\"O .IR "int $0x80" .
i386 ¾å¤Ç¤Ï¡¢
.BR clone ()
¤Ï vsyscall ·Ðͳ¤Ç¤Ï¤Ê¤¯¡¢Ä¾ÀÜ
.I "int $0x80"
·Ðͳ¤Ç¸Æ¤Ó½Ð¤¹¤Ù¤­¤Ç¤¢¤ë¡£

.\"O On ia64, a different system call is used:
ia64 ¤Ç¤Ï¡¢Ê̤Υ·¥¹¥Æ¥à¥³¡¼¥ë¤¬»ÈÍѤµ¤ì¤ë:
.nf

.BI "int __clone2(int (*" "fn" ")(void *), "
.BI "             void *" child_stack_base ", size_t " stack_size ,
.BI "             int " flags ", void *" "arg" ", ... "
.BI "          /* pid_t *" ptid ", struct user_desc *" tls \
", pid_t *" ctid " */ );"
.fi
.PP
.\"O The
.\"O .BR __clone2 ()
.\"O system call operates in the same way as
.\"O .BR clone (),
.\"O except that
.\"O .I child_stack_base
.\"O points to the lowest address of the child's stack area,
.\"O and
.\"O .I stack_size
.\"O specifies the size of the stack pointed to by
.\"O .IR child_stack_base .
.BR __clone2 ()
¥·¥¹¥Æ¥à¥³¡¼¥ë¤Ï
.BR clone ()
¤ÈƱ¤¸¤è¤¦¤ËÆ°ºî¤¹¤ë¤¬¡¢°Ê²¼¤ÎÅÀ¤¬°Û¤Ê¤ë:
.I child_stack_base
¤Ï»Ò¥×¥í¥»¥¹¤Î¥¹¥¿¥Ã¥¯¥¨¥ê¥¢¤ÎºÇ¾®¤Î¥¢¥É¥ì¥¹¤ò»Ø¤·¡¢
.I stack_size
¤Ï
.I child_stack_base
¤¬»Ø¤·¼¨¤¹¥¹¥¿¥Ã¥¯¥¨¥ê¥¢¤ÎÂ礭¤µ¤ò¼¨¤¹¡£
.\"O .SH BUGS
.SH ¥Ð¥°
.\"O Versions of the GNU C library that include the NPTL threading library
.\"O contain a wrapper function for
.\"O .BR getpid (2)
.\"O that performs caching of PIDs.
NPTL ¥¹¥ì¥Ã¥É¡¦¥é¥¤¥Ö¥é¥ê¤ò´Þ¤ó¤Ç¤¤¤ë GNU C ¥é¥¤¥Ö¥é¥ê¤Î¤¤¤¯¤Ä¤«¤Î¥Ð¡¼¥¸¥ç¥ó
¤Ë¤Ï¡¢
.BR getpid (2)
¤Î¥é¥Ã¥Ñ¡¼´Ø¿ô¤¬´Þ¤Þ¤ì¤Æ¤ª¤ê¡¢¤³¤Î¥é¥Ã¥Ñ¡¼´Ø¿ô¤Ï PID ¤ò¥­¥ã¥Ã¥·¥å¤¹¤ë¡£
.\"O This caching relies on support in the glibc wrapper for
.\"O .BR clone (),
.\"O but as currently implemented,
.\"O the cache may not be up to date in some circumstances.
¤³¤Î¥­¥ã¥Ã¥·¥å½èÍý¤¬Àµ¤·¤¯Æ°ºî¤¹¤ë¤¿¤á¤Ë¤Ï glibc ¤Î
.BR clone (2)
¤Î¥é¥Ã¥Ñ¡¼´Ø¿ô¤Ç¤Î½õ¤±¤¬É¬ÍפÀ¤¬¡¢¸½¾õ¤Î¼ÂÁõ¤Ç¤Ï¡¢
¤¢¤ë¾õ¶·²¼¤Ë¤ª¤¤¤Æ¥­¥ã¥Ã¥·¥å¤¬ºÇ¿·¤È¤Ê¤é¤Ê¤¤²ÄǽÀ­¤¬¤¢¤ë¡£
.\"O In particular,
.\"O if a signal is delivered to the child immediately after the
.\"O .BR clone ()
.\"O call, then a call to
.\"O .BR getpid ()
.\"O in a handler for the signal may return the PID
.\"O of the calling process ("the parent"),
.\"O if the clone wrapper has not yet had a chance to update the PID
.\"O cache in the child.
Æäˡ¢
.BR clone ()
¤Î¸Æ¤Ó½Ð¤·Ä¾¸å¤Ë¥·¥°¥Ê¥ë¤¬»Ò¥×¥í¥»¥¹¤ËÇÛÁ÷¤µ¤ì¤¿¾ì¹ç¤Ë¡¢
¤½¤Î¥·¥°¥Ê¥ë¤ËÂФ¹¤ë¥Ï¥ó¥É¥éÆâ¤Ç
.BR getpid ()
¤ò¸Æ¤Ó½Ð¤¹¤È¡¢¤½¤ì¤Þ¤Ç¤Ë clone ¤Î¥é¥Ã¥Ñ¡¼´Ø¿ô¤¬»Ò¥×¥í¥»¥¹¤Î PID ¥­¥ã¥Ã¥·¥å¤ò
¹¹¿·¤¹¤ëµ¡²ñ¤¬ÆÀ¤é¤ì¤Æ¤¤¤Ê¤±¤ì¤Ð¡¢¸Æ¤Ó½Ð¤·¸µ¥×¥í¥»¥¹ ("¿Æ¥×¥í¥»¥¹") ¤Î PID ¤¬
ÊÖ¤µ¤ì¤ë²ÄǽÀ­¤¬¤¢¤ë¡£
.\"O (This discussion ignores the case where the child was created using
.\"O .BR CLONE_THREAD ,
.\"O when
.\"O .BR getpid ()
.\"O .I should
.\"O return the same value in the child and in the process that called
.\"O .BR clone (),
.\"O since the caller and the child are in the same thread group.
.\"O The stale-cache problem also does not occur if the
.\"O .I flags
.\"O argument includes
.\"O .BR CLONE_VM .)
(¤³¤ÎµÄÏÀ¤Ç¤Ï¡¢»Ò¥×¥í¥»¥¹¤¬
.B CLONE_THREAD
¤ò»È¤Ã¤ÆºîÀ®¤µ¤ì¤¿¾ì¹ç¤Î¤³¤È¤Ï̵»ë¤·¤Æ¤¤¤ë¡£
»Ò¥×¥í¥»¥¹¤¬
.B CLONE_THREAD
¤òºî¤Ã¤ÆºîÀ®¤µ¤ì¤¿¾ì¹ç¤Ë¤Ï¡¢
¸Æ¤Ó½Ð¤·¸µ¤È»Ò¥×¥í¥»¥¹¤ÏƱ¤¸¥¹¥ì¥Ã¥É¡¦¥°¥ë¡¼¥×¤Ë°¤¹¤Î¤Ç¡¢
.BR getpid ()
¤Ï»Ò¥×¥í¥»¥¹¤È
.BR clone ()
¤ò¸Æ¤Ó½Ð¤·¤¿¥×¥í¥»¥¹¤ÇƱ¤¸ÃͤòÊÖ¤¹¤Î¤¬¡ÖÀµ¤·¤¤¡×¡£
¥­¥ã¥Ã¥·¥å¤¬ºÇ¿·¤È¤Ê¤é¤Ê¤¤ÌäÂê (stale-cache problem) ¤Ï¡¢
.I flags
¤Ë
.B CLONE_VM
¤¬´Þ¤Þ¤ì¤Æ¤¤¤ë¾ì¹ç¤Ë¤âȯÀ¸¤·¤Ê¤¤¡£)
.\"O To get the truth, it may be necessary to use code such as the following:
ËÜÅö¤ÎÃͤòÆÀ¤ë¤¿¤á¤Ë¤Ï¡¢¼¡¤Î¤è¤¦¤Ê¥³¡¼¥É¤ò»È¤¦É¬Íפ¬¤¢¤ë¤«¤â¤·¤ì¤Ê¤¤¡£
.nf

    #include <syscall.h>

    pid_t mypid;

    mypid = syscall(SYS_getpid);
.fi
.\" See also the following bug reports
.\" https://bugzilla.redhat.com/show_bug.cgi?id=417521
.\" http://sourceware.org/bugzilla/show_bug.cgi?id=6910
.\"O .SH "SEE ALSO"
.SH ´ØÏ¢¹àÌÜ
.BR fork (2),
.BR futex (2),
.BR getpid (2),
.BR gettid (2),
.BR set_thread_area (2),
.BR set_tid_address (2),
.BR tkill (2),
.BR unshare (2),
.BR wait (2),
.BR capabilities (7),
.BR pthreads (7)