(split) LDP man-pages の original/ を v3.32 に更新。

[linuxjm/LDP_man-pages.git] / draft / man7 / capabilities.7

.\" Copyright (c) 2002 by Michael Kerrisk <mtk.manpages@gmail.com>
.\"
.\" 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.\" manual page may be incorrect or out-of-date.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.
.\"
.\" 6 Aug 2002 - Initial Creation
.\" Modified 2003-05-23, Michael Kerrisk, <mtk.manpages@gmail.com>
.\" Modified 2004-05-27, Michael Kerrisk, <mtk.manpages@gmail.com>
.\" 2004-12-08, mtk Added O_NOATIME for CAP_FOWNER
.\" 2005-08-16, mtk, Added CAP_AUDIT_CONTROL and CAP_AUDIT_WRITE
.\" 2008-07-15, Serge Hallyn <serue@us.bbm.com>
.\"     Document file capabilities, per-process capability
.\"     bounding set, changed semantics for CAP_SETPCAP,
.\"     and other changes in 2.6.2[45].
.\"     Add CAP_MAC_ADMIN, CAP_MAC_OVERRIDE, CAP_SETFCAP.
.\" 2008-07-15, mtk
.\"     Add text describing circumstances in which CAP_SETPCAP
.\"     (theoretically) permits a thread to change the
.\"     capability sets of another thread.
.\"     Add section describing rules for programmatically
.\"     adjusting thread capability sets.
.\"     Describe rationale for capability bounding set.
.\"     Document "securebits" flags.
.\"     Add text noting that if we set the effective flag for one file
.\"     capability, then we must also set the effective flag for all
.\"     other capabilities where the permitted or inheritable bit is set.
.\"
.\" Japanese Version Copyright (c) 2005 Akihiro MOTOKI all rights reserved.
.\" Translated 2005-03-09, Akihiro MOTOKI <amotoki@dd.iij4u.or.jp>
.\" Updated 2005-11-04, Akihiro MOTOKI
.\" Updated 2006-04-16, Akihiro MOTOKI, LDP v2.29
.\" Updated 2006-07-20, Akihiro MOTOKI, LDP v2.34
.\" Updated 2007-01-05, Akihiro MOTOKI, LDP v2.43
.\" Updated 2008-12-24, Akihiro MOTOKI, LDP v3.15
.\" Updated 2009-02-27, Akihiro MOTOKI, LDP v3.19
.\" Updated 2010-04-11, Akihiro MOTOKI, LDP v3.24
.\"
.TH CAPABILITIES 7 2010-06-19 "Linux" "Linux Programmer's Manual"
.\"O .SH NAME
.SH ̾Á°
.\"O capabilities \- overview of Linux capabilities
capabilities \- Linux ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£ (capability) ¤Î³µÍ×
.\"O .SH DESCRIPTION
.SH ÀâÌÀ
.\"O For the purpose of performing permission checks,
.\"O traditional UNIX implementations distinguish two categories of processes:
.\"O .I privileged
.\"O processes (whose effective user ID is 0, referred to as superuser or root),
.\"O and
.\"O .I unprivileged
.\"O processes (whose effective UID is nonzero).
¸¢¸Â¤Î¥Á¥§¥Ã¥¯¤ò¹Ô¤¦´ÑÅÀ¤«¤é¸«¤ë¤È¡¢ÅÁÅýŪ¤Ê UNIX ¤Î¼ÂÁõ¤Ç¤Ï
¥×¥í¥»¥¹¤ÏÆó¤Ä¤Î¥«¥Æ¥´¥ê¤ËʬÎà¤Ç¤­¤ë:
.I Æø¢
¥×¥í¥»¥¹ (¼Â¸ú¥æ¡¼¥¶ID ¤¬ 0 ¤Î¥×¥í¥»¥¹¡£¥æ¡¼¥¶ID 0 ¤Ï
¥¹¡¼¥Ñ¡¼¥æ¡¼¥¶¤ä root ¤È¸Æ¤Ð¤ì¤ë) ¤È
.I ÈóÆø¢
¥×¥í¥»¥¹ (¼Â¸ú¥æ¡¼¥¶ID ¤¬ 0 °Ê³°¤Î¥×¥í¥»¥¹) ¤Ç¤¢¤ë¡£
.\"O Privileged processes bypass all kernel permission checks,
.\"O while unprivileged processes are subject to full permission
.\"O checking based on the process's credentials
.\"O (usually: effective UID, effective GID, and supplementary group list).
ÈóÆø¢¥×¥í¥»¥¹¤Ç¤Ï¡¢¥×¥í¥»¥¹¤Î»ñ³Ê¾ðÊó (Ä̾ï¤Ï¡¢¼Â¸úUID ¡¢¼Â¸úGID
¤ÈÄɲäΥ°¥ë¡¼¥×¥ê¥¹¥È) ¤Ë´ð¤Å¤¯¸¢¸Â¥Á¥§¥Ã¥¯¤¬¹Ô¤ï¤ì¤ë¤Î¤ËÂФ·¡¢
Æø¢¥×¥í¥»¥¹¤Ç¤ÏÁ´¤Æ¤Î¥«¡¼¥Í¥ë¤Î¸¢¸Â¥Á¥§¥Ã¥¯¤¬¥Ð¥¤¥Ñ¥¹¤µ¤ì¤ë¡£

.\"O Starting with kernel 2.2, Linux divides the privileges traditionally
.\"O associated with superuser into distinct units, known as
.\"O .IR capabilities ,
.\"O which can be independently enabled and disabled.
.\"O Capabilities are a per-thread attribute.
¥Ð¡¼¥¸¥ç¥ó 2.2 °Ê¹ß¤Î Linux ¤Ç¤Ï¡¢
¤³¤ì¤Þ¤Ç¥¹¡¼¥Ñ¡¼¥æ¡¼¥¶¤Ë·ë¤ÓÉÕ¤±¤é¤ì¤Æ¤­¤¿¸¢¸Â¤ò¡¢
¤¤¤¯¤Ä¤«¤Î¥°¥ë¡¼¥×¤Ëʬ³ä¤·¤Æ¤¤¤ë¡£¤³¤ì¤é¤Î¥°¥ë¡¼¥×¤Ï
.IR ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£ (capability)
¤È¸Æ¤Ð¤ì¡¢¥°¥ë¡¼¥×Ëè¤ËÆÈΩ¤ËÍ­¸ú¡¢Ìµ¸ú¤òÀßÄê¤Ç¤­¤ë¡£
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Ï¥¹¥ì¥Ã¥Éñ°Ì¤Î°À­¤Ç¤¢¤ë¡£
.\"
.\"O .SS Capabilities List
.SS ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Î¥ê¥¹¥È
.\"O The following list shows the capabilities implemented on Linux,
.\"O and the operations or behaviors that each capability permits:
°Ê²¼¤Î¥ê¥¹¥È¤Ï¡¢
Linux ¤Ç¼ÂÁõ¤µ¤ì¤Æ¤¤¤ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤È
³Æ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬µö²Ä¤¹¤ëÁàºî¤ÈÆ°ºî¤ò¤Þ¤È¤á¤¿¤â¤Î¤Ç¤¢¤ë¡£
.TP
.\"O .BR CAP_AUDIT_CONTROL " (since Linux 2.6.11)"
.\"O Enable and disable kernel auditing; change auditing filter rules;
.\"O retrieve auditing status and filtering rules.
.BR CAP_AUDIT_CONTROL " (Linux 2.6.11 °Ê¹ß)"
¥«¡¼¥Í¥ë´Æºº (audit) ¤ÎÍ­¸ú̵¸ú¤ÎÀÚ¤êÂؤ¨¡¢
´Æºº¤Î¥Õ¥£¥ë¥¿¡¦¥ë¡¼¥ë¤ÎÊѹ¹¡¢
´Æºº¤Î¾õ¶·¤ä¥Õ¥£¥ë¥¿¡¦¥ë¡¼¥ë¤Î¼èÆÀ¤¬¤Ç¤­¤ë¡£
.TP
.\"O .BR CAP_AUDIT_WRITE " (since Linux 2.6.11)"
.BR CAP_AUDIT_WRITE " (Linux 2.6.11 °Ê¹ß)"
.\"O Write records to kernel auditing log.
¥«¡¼¥Í¥ë´Æºº¤Î¥í¥°¤Ë¥ì¥³¡¼¥É¤ò½ñ¤­¹þ¤à¡£
.TP
.B CAP_CHOWN
.\"O Make arbitrary changes to file UIDs and GIDs (see
.\"O .BR chown (2)).
¥Õ¥¡¥¤¥ë¤Î UID ¤ÈGID ¤òǤ°Õ¤ËÊѹ¹¤¹¤ë
.RB ( chown (2)
»²¾È)¡£
.TP
.B CAP_DAC_OVERRIDE
.\"O Bypass file read, write, and execute permission checks.
.\"O (DAC is an abbreviation of "discretionary access control".)
¥Õ¥¡¥¤¥ë¤ÎÆɤ߽Ф·¡¢½ñ¤­¹þ¤ß¡¢¼Â¹Ô¤Î¸¢¸Â¥Á¥§¥Ã¥¯¤ò¥Ð¥¤¥Ñ¥¹¤¹¤ë
(DAC ¤Ï "discretionary access control (Ǥ°Õ¤Î¥¢¥¯¥»¥¹À©¸æ)" ¤Îά¤Ç¤¢¤ë)¡£
.TP
.B CAP_DAC_READ_SEARCH
.\"O Bypass file read permission checks and
.\"O directory read and execute permission checks.
¥Õ¥¡¥¤¥ë¤ÎÆɤ߽Ф·¸¢¸Â¤Î¥Á¥§¥Ã¥¯¤È¥Ç¥£¥ì¥¯¥È¥ê¤ÎÆɤ߽Ф·¤È¼Â¹Ô
¤Î¸¢¸Â¥Á¥§¥Ã¥¯¤ò¥Ð¥¤¥Ñ¥¹¤¹¤ë¡£
.TP
.B CAP_FOWNER
.PD 0
.RS
.IP * 2
.\"O Bypass permission checks on operations that normally
.\"O require the file system UID of the process to match the UID of
.\"O the file (e.g.,
.\"O .BR chmod (2),
.\"O .BR utime (2)),
.\"O excluding those operations covered by
.\"O .B CAP_DAC_OVERRIDE
.\"O and
.\"O .BR CAP_DAC_READ_SEARCH ;
Ä̾¥×¥í¥»¥¹¤Î¥Õ¥¡¥¤¥ë¥·¥¹¥Æ¥à UID ¤¬¥Õ¥¡¥¤¥ë¤Î UID ¤Ë°ìÃפ¹¤ë¤³¤È¤¬
Í׵ᤵ¤ì¤ëÁàºî (Î㤨¤Ð
.BR chmod (2),
.BR utime (2))
¤Ë¤ª¤±¤ë¸¢¸Â¥Á¥§¥Ã¥¯¤ò¥Ð¥¤¥Ñ¥¹¤¹¤ë¡£
⤷¡¢
.B CAP_DAC_OVERRIDE
¤«
.B CAP_DAC_READ_SEARCH
¤Ë¤è¤ê¥Á¥§¥Ã¥¯¤¬¹Ô¤ï¤ì¤ëÁàºî¤Ï½ü¤¯¡£
.IP *
.\"O set extended file attributes (see
.\"O .BR chattr (1))
.\"O on arbitrary files;
Ǥ°Õ¤Î¥Õ¥¡¥¤¥ë¤ËÂФ·¤Æ³ÈÄ¥¥Õ¥¡¥¤¥ë°À­¤òÀßÄꤹ¤ë
.RB ( chattr (1)
»²¾È)¡£
.IP *
.\"O set Access Control Lists (ACLs) on arbitrary files;
Ǥ°Õ¤Î¥Õ¥¡¥¤¥ë¤ËÂФ·¤Æ¥¢¥¯¥»¥¹À©¸æ¥ê¥¹¥È (ACL) ¤òÀßÄꤹ¤ë¡£
.IP *
.\"O ignore directory sticky bit on file deletion;
¥Õ¥¡¥¤¥ë¤Îºï½ü¤ÎºÝ¤Ë¥Ç¥£¥ì¥¯¥È¥ê¤Î¥¹¥Æ¥£¥Ã¥­¡¼¥Ó¥Ã¥È¤ò̵»ë¤¹¤ë¡£
.IP *
.\"O specify
.\"O .B O_NOATIME
.\"O for arbitrary files in
.\"O .BR open (2)
.\"O and
.\"O .BR fcntl (2).
.BR open (2)
¤ä
.BR fcntl (2)
¤ÇǤ°Õ¤Î¥Õ¥¡¥¤¥ë¤ËÂФ·¤Æ
.B O_NOATIME
¤ò»ØÄꤹ¤ë¡£
.RE
.PD
.TP
.B CAP_FSETID
.\"O Don't clear set-user-ID and set-group-ID permission
.\"O bits when a file is modified;
.\"O set the set-group-ID bit for a file whose GID does not match
.\"O the file system or any of the supplementary GIDs of the calling process.
¥Õ¥¡¥¤¥ë¤¬Êѹ¹¤µ¤ì¤¿¤È¤­¤Ë set-user-ID ¤Èset-group-ID ¤Îµö²Ä¥Ó¥Ã¥È¤ò¥¯¥ê¥¢
¤·¤Ê¤¤¡£¸Æ¤Ó½Ð¤·¸µ¥×¥í¥»¥¹¤Î¥Õ¥¡¥¤¥ë¥·¥¹¥Æ¥à GID ¤ÈÄɲäΠGID ¤Î¤¤¤º¤ì¤È¤â
GID ¤¬°ìÃפ·¤Ê¤¤¥Õ¥¡¥¤¥ë¤ËÂФ·¤Æ set-group-ID ¥Ó¥Ã¥È¤òÀßÄꤹ¤ë¡£
.TP
.B CAP_IPC_LOCK
.\"O Lock memory
.\"O .RB ( mlock (2),
.\"O .BR mlockall (2),
.\"O .BR mmap (2),
.\"O .BR shmctl (2)).
¥á¥â¥ê¡¼¤Î¥í¥Ã¥¯
.RB ( mlock (2),
.BR mlockall (2),
.BR mmap (2),
.BR shmctl (2))
¤ò¹Ô¤¦¡£
.TP
.B CAP_IPC_OWNER
.\"O Bypass permission checks for operations on System V IPC objects.
System V IPC ¥ª¥Ö¥¸¥§¥¯¥È¤ËÂФ¹¤ëÁàºî¤Ë´Ø¤·¤Æ¸¢¸Â¥Á¥§¥Ã¥¯¤ò¥Ð¥¤¥Ñ¥¹¤¹¤ë¡£
.TP
.B CAP_KILL
.\"O Bypass permission checks for sending signals (see
.\"O .BR kill (2)).
.\"O This includes use of the
.\"O .BR ioctl (2)
.\"O .B KDSIGACCEPT
.\"O operation.
¥·¥°¥Ê¥ë¤òÁ÷¿®¤¹¤ëºÝ¤Ë¸¢¸Â¥Á¥§¥Ã¥¯¤ò¥Ð¥¤¥Ñ¥¹¤¹¤ë
.RB ( kill (2)
»²¾È)¡£¤³¤ì¤Ë¤Ï
.BR ioctl (2)
¤Î
.B KDSIGACCEPT
Áàºî¤Î»ÈÍѤâ´Þ¤Þ¤ì¤ë¡£
.\" FIXME CAP_KILL also has an effect for threads + setting child
.\"       termination signal to other than SIGCHLD: without this
.\"       capability, the termination signal reverts to SIGCHLD
.\"       if the child does an exec().  What is the rationale
.\"       for this?
.TP
.\"O .BR CAP_LEASE " (since Linux 2.4)"
.BR CAP_LEASE " (Linux 2.4 °Ê¹ß)"
.\"O Establish leases on arbitrary files (see
.\"O .BR fcntl (2)).
Ǥ°Õ¤Î¥Õ¥¡¥¤¥ë¤ËÂФ·¤Æ
¥Õ¥¡¥¤¥ë¥ê¡¼¥¹¤òÀßÄꤹ¤ë
.RB ( fcntl (2)
»²¾È)¡£
.TP
.B CAP_LINUX_IMMUTABLE
.\"O Set the
.\"O .B FS_APPEND_FL
.\"O and
.\"O .B FS_IMMUTABLE_FL
.\"O .\" These attributes are now available on ext2, ext3, Reiserfs, XFS, JFS
.\"O i-node flags (see
.\"O .BR chattr (1)).
³ÈÄ¥¥Õ¥¡¥¤¥ë°À­
.B FS_APPEND_FL
¤È
.B FS_IMMUTABLE_FL
¤òÀßÄꤹ¤ë
.RB ( chattr (1)
»²¾È)¡£
.\" ¤³¤ì¤é¤Î°À­¤Ï ext2, ext3, Reiserfs, XFS, JFS ¤ÇÍøÍѲÄǽ¤Ç¤¢¤ë¡£
.TP
.\"O .BR CAP_MAC_ADMIN " (since Linux 2.6.25)"
.BR CAP_MAC_ADMIN " (Linux 2.6.25 °Ê¹ß)"
.\"O Override Mandatory Access Control (MAC).
.\"O Implemented for the Smack Linux Security Module (LSM).
¶¯À©¥¢¥¯¥»¥¹À©¸æ (MAC) ¤ò¾å½ñ¤­¤¹¤ë¡£
Smack Linux Security Module (LSM) ÍѤ˼ÂÁõ¤µ¤ì¤Æ¤¤¤ë¡£
.TP
.\"O .BR CAP_MAC_OVERRIDE " (since Linux 2.6.25)"
.BR CAP_MAC_OVERRIDE " (Linux 2.6.25 °Ê¹ß)"
.\"O Allow MAC configuration or state changes.
.\"O Implemented for the Smack LSM.
MAC ¤ÎÀßÄê¤ä¾õÂÖ¤òÊѹ¹¤¹¤ë¡£
Smack LSM ÍѤ˼ÂÁõ¤µ¤ì¤Æ¤¤¤ë¡£
.TP
.\"O .BR CAP_MKNOD " (since Linux 2.4)"
.BR CAP_MKNOD " (Linux 2.4 °Ê¹ß)"
.\"O Create special files using
.\"O .BR mknod (2).
(Linux 2.4 °Ê¹ß)
.BR mknod (2)
¤ò»ÈÍѤ·¤Æ¥¹¥Ú¥·¥ã¥ë¡¦¥Õ¥¡¥¤¥ë¤òºîÀ®¤¹¤ë¡£
.TP
.B CAP_NET_ADMIN
.\"O Perform various network-related operations
.\"O (e.g., setting privileged socket options,
.\"O enabling multicasting, interface configuration,
.\"O modifying routing tables).
³Æ¼ï¤Î¥Í¥Ã¥È¥ï¡¼¥¯´ØÏ¢¤ÎÁàºî¤ò¼Â¹Ô¤¹¤ë¡£
(Î㤨¤Ð¡¢Æø¢¤¬É¬Íפʥ½¥±¥Ã¥È¥ª¥×¥·¥ç¥ó¤òÀßÄꤹ¤ë¡¢¥Þ¥ë¥Á¥­¥ã¥¹¥È¤òÍ­¸ú¤Ë¤¹¤ë¡¢
¥¤¥ó¥¿¡¼¥Õ¥§¡¼¥¹¤òÀßÄꤹ¤ë¡¢¥ë¡¼¥Æ¥£¥ó¥°¥Æ¡¼¥Ö¥ë¤òÊѹ¹¤¹¤ë¤Ê¤É)
.TP
.B CAP_NET_BIND_SERVICE
.\"O Bind a socket to Internet domain privileged ports
.\"O (port numbers less than 1024).
¥¤¥ó¥¿¡¼¥Í¥Ã¥È¥É¥á¥¤¥ó¤ÎÆø¢¥Ý¡¼¥È (¥Ý¡¼¥ÈÈֹ椬 1024 ÈÖ̤Ëþ)
¤ò¥Ð¥¤¥ó¥É¤Ç¤­¤ë¡£
.TP
.B CAP_NET_BROADCAST
.\"O (Unused)  Make socket broadcasts, and listen to multicasts.
(̤»ÈÍÑ) ¥½¥±¥Ã¥È¤Î¥Ö¥í¡¼¥É¥­¥ã¥¹¥È¤È¡¢¥Þ¥ë¥Á¥­¥ã¥¹¥È¤ÎÂÔ¤Á¼õ¤±¤ò¹Ô¤¦¡£
.TP
.B CAP_NET_RAW
.\"O Use RAW and PACKET sockets.
.\"O .\" Also various IP options and setsockopt(SO_BINDTODEVICE)
RAW ¥½¥±¥Ã¥È¤È PACKET ¥½¥±¥Ã¥È¤ò»ÈÍѤ¹¤ë¡£
.\" ¤Þ¤¿¡¢³Æ¼ï¤Î IP ¥ª¥×¥·¥ç¥ó¤È SO_BINDTODEVICE ¥½¥±¥Ã¥È¥ª¥×¥·¥ç¥ó¤ò»ÈÍѤǤ­¤ë¡£
.TP
.B CAP_SETGID
.\"O Make arbitrary manipulations of process GIDs and supplementary GID list;
.\"O forge GID when passing socket credentials via UNIX domain sockets.
¥×¥í¥»¥¹¤Î GID ¤ÈÄɲäΠGID ¥ê¥¹¥È¤ËÂФ¹¤ëǤ°Õ¤ÎÁàºî¤ò¹Ô¤¦¡£
UNIX ¥É¥á¥¤¥ó¥½¥±¥Ã¥È·Ðͳ¤Ç¥½¥±¥Ã¥È¤Î»ñ³Ê¾ðÊó (credential) ¤òÅϤ¹ºÝ¤Ë
µ¶¤Î GID ¤òÅϤ¹¤³¤È¤¬¤Ç¤­¤ë¡£
.TP
.\"O .BR CAP_SETFCAP " (since Linux 2.6.24)"
.BR CAP_SETFCAP " (Linux 2.6.24 °Ê¹ß)"
.\"O Set file capabilities.
¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤òÀßÄꤹ¤ë¡£
.TP
.B CAP_SETPCAP
.\"O If file capabilities are not supported:
.\"O grant or remove any capability in the
.\"O caller's permitted capability set to or from any other process.
.\"O (This property of
.\"O .B CAP_SETPCAP
.\"O is not available when the kernel is configured to support
.\"O file capabilities, since
.\"O .B CAP_SETPCAP
.\"O has entirely different semantics for such kernels.)
¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬¥µ¥Ý¡¼¥È¤µ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç:
¸Æ¤Ó½Ð¤·¸µ¤¬µö²Ä¤µ¤ì¤Æ¤¤¤ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Ë´Þ¤Þ¤ì¤ëǤ°Õ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò¡¢
¾¤Î¥×¥í¥»¥¹¤ËÉÕÍ¿¤·¤¿¤ê¡¢ºï½ü¤·¤¿¤ê¤Ç¤­¤ë¡£
(¥«¡¼¥Í¥ë¤¬¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò¥µ¥Ý¡¼¥È¤·¤Æ¤¤¤ë¾ì¹ç¡¢
.B CAP_SETPCAP
¤Ï¤³¤ÎÌò³ä¤ò»ý¤¿¤Ê¤¤¡£
¤Ê¤¼¤Ê¤é¡¢¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò¥µ¥Ý¡¼¥È¤·¤Æ¤¤¤ë¥«¡¼¥Í¥ë¤Ç¤Ï
.B CAP_SETPCAP
¤ÏÁ´¤¯Ê̤ΰÕÌ£¤ò»ý¤Ä¤«¤é¤Ç¤¢¤ë¡£)

.\"O If file capabilities are supported:
.\"O add any capability from the calling thread's bounding set
.\"O to its inheritable set;
.\"O drop capabilities from the bounding set (via
.\"O .BR prctl (2)
.\"O .BR PR_CAPBSET_DROP );
.\"O make changes to the
.\"O .I securebits
.\"O flags.
¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬¥µ¥Ý¡¼¥È¤µ¤ì¤Æ¤¤¤ë¾ì¹ç:
¸Æ¤Ó½Ð¤·¸µ¥¹¥ì¥Ã¥É¤Î¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤ÎǤ°Õ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò
¼«¿È¤Î·Ñ¾µ²Äǽ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ËÄɲäǤ­¤ë¡£
.RB ( prctl (2)
.BR PR_CAPBSET_DROP
¤ò»È¤Ã¤Æ)
¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤«¤é¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤òºï½ü¤Ç¤­¤ë¡£
.I securebits
¥Õ¥é¥°¤òÊѹ¹¤Ç¤­¤ë¡£
.TP
.B CAP_SETUID
.\"O Make arbitrary manipulations of process UIDs
.\"O .RB ( setuid (2),
.\"O .BR setreuid (2),
.\"O .BR setresuid (2),
.\"O .BR setfsuid (2));
.\"O make forged UID when passing socket credentials via UNIX domain sockets.
¥×¥í¥»¥¹¤Î UID ¤ËÂФ¹¤ëǤ°Õ¤ÎÁàºî
.RB ( setuid (2),
.BR setreuid (2),
.BR setresuid (2),
.BR setfsuid (2))
¤ò¹Ô¤¦¡£
UNIX ¥É¥á¥¤¥ó¥½¥±¥Ã¥È·Ðͳ¤Ç¥½¥±¥Ã¥È¤Î»ñ³Ê¾ðÊó (credential) ¤òÅϤ¹ºÝ¤Ë
µ¶¤Î UID ¤òÅϤ¹¤³¤È¤¬¤Ç¤­¤ë¡£
.\" FIXME CAP_SETUID also an effect in exec(); document this.
.TP
.B CAP_SYS_ADMIN
.PD 0
.RS
.IP * 2
.\"O Perform a range of system administration operations including:
.\"O .BR quotactl (2),
.\"O .BR mount (2),
.\"O .BR umount (2),
.\"O .BR swapon (2),
.\"O .BR swapoff (2),
.\"O .BR sethostname (2),
.\"O and
.\"O .BR setdomainname (2);
°Ê²¼¤Î¥·¥¹¥Æ¥à´ÉÍýÍѤÎÁàºî¤ò¼Â¹Ô¤¹¤ë:
.BR quotactl (2),
.BR mount (2),
.BR umount (2),
.BR swapon (2),
.BR swapoff (2),
.BR sethostname (2),
.BR setdomainname (2).
.IP *
.\"O perform
.\"O .B IPC_SET
.\"O and
.\"O .B IPC_RMID
.\"O operations on arbitrary System V IPC objects;
Ǥ°Õ¤Î System V IPC ¥ª¥Ö¥¸¥§¥¯¥È¤ËÂФ¹¤ë
.B IPC_SET
¤È
.B IPC_RMID
Áàºî¤ò¼Â¹Ô¤¹¤ë¡£
.IP *
.\"O perform operations on
.\"O .I trusted
.\"O and
.\"O .I security
.\"O Extended Attributes (see
.\"O .BR attr (5));
³Èĥ°À­
.I trusted
¤È
.I security
¤ËÂФ¹¤ëÁàºî¤ò¼Â¹Ô¤¹¤ë
.RB ( attr (5)
»²¾È)¡£
.IP *
.\"O use
.\"O .BR lookup_dcookie (2);
.BR lookup_dcookie (2)
¤ò¸Æ¤Ó½Ð¤¹¡£
.IP *
.\"O use
.\"O .BR ioprio_set (2)
.\"O to assign
.\"O .B IOPRIO_CLASS_RT
.\"O and (before Linux 2.6.25)
.\"O .B IOPRIO_CLASS_IDLE
.\"O I/O scheduling classes;
.BR ioprio_set (2)
¤ò»È¤Ã¤Æ I/O ¥¹¥±¥¸¥å¡¼¥ê¥ó¥°¥¯¥é¥¹
.BR IOPRIO_CLASS_RT ,
.B IOPRIO_CLASS_IDLE
¤ò³ä¤êÅö¤Æ¤ë
.RB ( IOPRIO_CLASS_IDLE
¤Ï Linux 2.6.25 ¤è¤êÁ°¤Î¥Ð¡¼¥¸¥ç¥ó¤Î¤ß)¡£
.IP *
.\"O forge UID when passing socket credentials;
¥½¥±¥Ã¥È¤Î»ñ³Ê¾ðÊó (credential) ¤òÅϤ¹ºÝ¤Ëµ¶¤Î UID ¤òÅϤ¹¡£
.IP *
.\"O exceed
.\"O .IR /proc/sys/fs/file-max ,
.\"O the system-wide limit on the number of open files,
.\"O in system calls that open files (e.g.,
.\"O .BR accept (2),
.\"O .BR execve (2),
.\"O .BR open (2),
.\"O .BR pipe (2));
¥Õ¥¡¥¤¥ë¤ò¥ª¡¼¥×¥ó¤¹¤ë¥·¥¹¥Æ¥à¥³¡¼¥ë (Î㤨¤Ð
.BR accept (2),
.BR execve (2),
.BR open (2),
.BR pipe (2))
¤Ç¥·¥¹¥Æ¥àÁ´ÂΤǥª¡¼¥×¥ó¤Ç¤­¤ë¥Õ¥¡¥¤¥ë¿ô¤Î¾å¸Â
.I /proc/sys/fs/file-max
¤òĶ²á¤¹¤ë¡£
.IP *
.\"O employ
.\"O .B CLONE_NEWNS
.\"O flag with
.\"O .BR clone (2)
.\"O and
.\"O .BR unshare (2);
.BR clone (2)
¤È
.BR unshare (2)
¤Ç
.B CLONE_NEWNS
¥Õ¥é¥°¤òÍøÍѤ¹¤ë¡£
.IP *
.\"O perform
.\"O .B KEYCTL_CHOWN
.\"O and
.\"O .B KEYCTL_SETPERM
.\"O .BR keyctl (2)
.\"O operations;
.BR keyctl (2)
¤Î
.B KEYCTL_CHOWN
¤È
.B KEYCTL_SETPERM
Áàºî¤ò¼Â¹Ô¤¹¤ë¡£
.IP *
.\"O perform
.\"O .BR madvise (2)
.\"O .B MADV_HWPOISON
.\"O operation.
.BR madvise (2)
¤Î
.B MADV_HWPOISON
Áàºî¤ò¼Â¹Ô¤¹¤ë¡£
.RE
.PD
.TP
.B CAP_SYS_BOOT
.\"O Use
.\"O .BR reboot (2)
.\"O and
.\"O .BR kexec_load (2).
.BR reboot (2)
¤È
.BR kexec_load (2)
¤ò¸Æ¤Ó½Ð¤¹¡£
.TP
.B CAP_SYS_CHROOT
.\"O Use
.\"O .BR chroot (2).
.BR chroot (2).
¤ò¸Æ¤Ó½Ð¤¹¡£
.TP
.B CAP_SYS_MODULE
.\"O Load and unload kernel modules
.\"O (see
.\"O .BR init_module (2)
.\"O and
.\"O .BR delete_module (2));
.\"O in kernels before 2.6.25:
.\"O drop capabilities from the system-wide capability bounding set.
¥«¡¼¥Í¥ë¥â¥¸¥å¡¼¥ë¤Î¥í¡¼¥É¡¢¥¢¥ó¥í¡¼¥É¤ò¹Ô¤¦
.RB ( init_module (2)
¤È
.BR delete_module (2)
¤ò»²¾È¤Î¤³¤È)¡£
¥Ð¡¼¥¸¥ç¥ó 2.6.25 ¤è¤êÁ°¤Î¥«¡¼¥Í¥ë¤Ç¡¢
¥·¥¹¥Æ¥àÁ´ÂΤΥ±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È (capability bounding set)
¤«¤é¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò³°¤¹¡£
.TP
.B CAP_SYS_NICE
.PD 0
.RS
.IP * 2
.\"O Raise process nice value
.\"O .RB ( nice (2),
.\"O .BR setpriority (2))
.\"O and change the nice value for arbitrary processes;
¥×¥í¥»¥¹¤Î nice Ãͤΰú¤­¾å¤²
.RB ( nice (2),
.BR setpriority (2))
¤ä¡¢Ç¤°Õ¤Î¥×¥í¥»¥¹¤Î nice ÃͤÎÊѹ¹¤ò¹Ô¤¦¡£
.IP *
.\"O set real-time scheduling policies for calling process,
.\"O and set scheduling policies and priorities for arbitrary processes
.\"O .RB ( sched_setscheduler (2),
.\"O .BR sched_setparam (2));
¸Æ¤Ó½Ð¤·¸µ¥×¥í¥»¥¹¤ËÂФ¹¤ë¥ê¥¢¥ë¥¿¥¤¥à¡¦¥¹¥±¥¸¥å¡¼¥ê¥ó¥°¥Ý¥ê¥·¡¼¤È¡¢
Ǥ°Õ¤Î¥×¥í¥»¥¹¤ËÂФ¹¤ë¥¹¥±¥¸¥å¡¼¥ê¥ó¥°¥Ý¥ê¥·¡¼¤ÈÍ¥ÀèÅÙ¤òÀßÄꤹ¤ë
.RB ( sched_setscheduler (2),
.BR sched_setparam (2))¡£
.IP *
.\"O set CPU affinity for arbitrary processes
.\"O .RB ( sched_setaffinity (2));
Ǥ°Õ¤Î¥×¥í¥»¥¹¤ËÂФ¹¤ë CPU affinity ¤òÀßÄê¤Ç¤­¤ë
.RB ( sched_setaffinity (2))¡£
.IP *
.\"O set I/O scheduling class and priority for arbitrary processes
.\"O .RB ( ioprio_set (2));
Ǥ°Õ¤Î¥×¥í¥»¥¹¤ËÂФ·¤Æ I/O ¥¹¥±¥¸¥å¡¼¥ê¥ó¥°¥¯¥é¥¹¤ÈÍ¥ÀèÅÙ¤òÀßÄê¤Ç¤­¤ë
.RB ( ioprio_set (2))¡£
.IP *
.\"O apply
.\"O .BR migrate_pages (2)
.\"O to arbitrary processes and allow processes
.\"O to be migrated to arbitrary nodes;
.BR migrate_pages (2)
¤òǤ°Õ¤Î¥×¥í¥»¥¹¤ËŬÍѤ·¡¢¥×¥í¥»¥¹¤òǤ°Õ¤Î¥Î¡¼¥É¤Ë°ÜÆ°¤¹¤ë¡£
.\" FIXME CAP_SYS_NICE also has the following effect for
.\" migrate_pages(2):
.\"     do_migrate_pages(mm, &old, &new,
.\"         capable(CAP_SYS_NICE) ? MPOL_MF_MOVE_ALL : MPOL_MF_MOVE);
.IP *
.\"O apply
.\"O .BR move_pages (2)
.\"O to arbitrary processes;
.BR move_pages (2)
¤òǤ°Õ¤Î¥×¥í¥»¥¹¤ËÂФ·¤Æ¹Ô¤¦¡£
.IP *
.\"O use the
.\"O .B MPOL_MF_MOVE_ALL
.\"O flag with
.\"O .BR mbind (2)
.\"O and
.\"O .BR move_pages (2).
.BR mbind (2)
¤È
.BR move_pages (2)
¤Ç
.B MPOL_MF_MOVE_ALL
¥Õ¥é¥°¤ò»ÈÍѤ¹¤ë¡£
.RE
.PD
.TP
.B CAP_SYS_PACCT
.\"O Use
.\"O .BR acct (2).
.BR acct (2)
¤ò¸Æ¤Ó½Ð¤¹¡£
.TP
.B CAP_SYS_PTRACE
.\"O Trace arbitrary processes using
.\"O .BR ptrace (2);
.\"O apply
.\"O .BR get_robust_list (2)
.\"O to arbitrary processes.
.BR ptrace (2)
¤ò»È¤Ã¤ÆǤ°Õ¤Î¥×¥í¥»¥¹¤ò¥È¥ì¡¼¥¹¤¹¤ë¡£
Ǥ°Õ¤Î¥×¥í¥»¥¹¤Ë
.BR get_robust_list (2)
¤òŬÍѤ¹¤ë¡£
.TP
.B CAP_SYS_RAWIO
.\"O Perform I/O port operations
.\"O .RB ( iopl (2)
.\"O and
.\"O .BR ioperm (2));
.\"O access
.\"O .IR /proc/kcore .
I/O ¥Ý¡¼¥ÈÁàºî¤ò¼Â¹Ô¤¹¤ë
.RB ( iopl (2)
¡¢
.BR ioperm (2))¡£
.I /proc/kcore
¤Ë¥¢¥¯¥»¥¹¤Ç¤­¤ë¡£
.TP
.B CAP_SYS_RESOURCE
.PD 0
.RS
.IP * 2
.\"O Use reserved space on ext2 file systems;
ext2 ¥Õ¥¡¥¤¥ë¥·¥¹¥Æ¥à¾å¤ÎͽÌ󤵤ì¤Æ¤¤¤ëÎΰè¤ò»ÈÍѤ¹¤ë¡£
.IP *
.\"O make
.\"O .BR ioctl (2)
.\"O calls controlling ext3 journaling;
ext3 ¤Î¥¸¥ã¡¼¥Ê¥ëµ¡Ç½¤òÀ©¸æ¤¹¤ë
.BR ioctl (2)
¤ò»ÈÍѤ¹¤ë¡£
.IP *
.\"O override disk quota limits;
¥Ç¥£¥¹¥¯ quota ¤Î¾å¸Â¤ò¾å½ñ¤­¤¹¤ë¡£
.IP *
.\"O increase resource limits (see
.\"O .BR setrlimit (2));
¥ê¥½¡¼¥¹¾å¸Â¤òÁý¤ä¤¹
.RB ( setrlimit (2))¡£
.IP *
.\"O override
.\"O .B RLIMIT_NPROC
.\"O resource limit;
.B RLIMIT_NPROC
¥ê¥½¡¼¥¹À©¸Â¤ò¾å½ñ¤­¤¹¤ë¡£
.IP *
.\"O raise
.\"O .I msg_qbytes
.\"O limit for a System V message queue above the limit in
.\"O .I /proc/sys/kernel/msgmnb
.\"O (see
.\"O .BR msgop (2)
.\"O and
.\"O .BR msgctl (2)).
¥á¥Ã¥»¡¼¥¸¥­¥å¡¼¤Ë´Ø¤¹¤ë¾å¸Â
.I msg_qbytes
¤ò
.I /proc/sys/kernel/msgmnb
¤Ë»ØÄꤵ¤ì¤Æ¤¤¤ë¾å¸Â¤è¤ê¤âÂ礭¤¯ÀßÄꤹ¤ë
.RB ( msgop (2)
¤È
.BR msgctl (2)
»²¾È)¡£
.IP *
.\"O use
.\"O .BR F_SETPIPE_SZ
.\"O to increase the capacity of a pipe above the limit specified by
.\"O .IR /proc/sys/fs/pipe-max-size .
.I /proc/sys/fs/pipe-max-size
¤Ë»ØÄꤵ¤ì¤Æ¤¤¤ë¾å¸Â¤òĶ¤¨¤Æ¥Ñ¥¤¥×¤ÎÍÆÎ̤òÁý¤ä¤¹¤Î¤Ë
.B F_SETPIPE_SZ
¤ò»ÈÍѤ¹¤ë¡£
.RE
.PD
.TP
.B CAP_SYS_TIME
.\"O Set system clock
.\"O .RB ( settimeofday (2),
.\"O .BR stime (2),
.\"O .BR adjtimex (2));
.\"O set real-time (hardware) clock.
¥·¥¹¥Æ¥à¥¯¥í¥Ã¥¯¤òÊѹ¹¤¹¤ë
.RB ( settimeofday (2),
.BR stime (2),
.BR adjtimex (2))¡£
¥ê¥¢¥ë¥¿¥¤¥à (¥Ï¡¼¥É¥¦¥§¥¢) ¥¯¥í¥Ã¥¯¤òÊѹ¹¤¹¤ë¡£
.TP
.B CAP_SYS_TTY_CONFIG
.\"O Use
.\"O .BR vhangup (2).
.BR vhangup (2)
¤ò¸Æ¤Ó½Ð¤¹¡£
.\"
.\"O .SS Past and Current Implementation
.SS ²áµî¤È¸½ºß¤Î¼ÂÁõ
.\"O A full implementation of capabilities requires that:
´°Á´¤Ê·Á¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò¼ÂÁõ¤¹¤ë¤Ë¤Ï¡¢°Ê²¼¤ÎÍ×·ï¤òËþ¤¿¤¹É¬Íפ¬¤¢¤ë¡§
.IP 1. 3
.\"O For all privileged operations,
.\"O the kernel must check whether the thread has the required
.\"O capability in its effective set.
Á´¤Æ¤ÎÆø¢Áàºî¤Ë¤Ä¤¤¤Æ¡¢¥«¡¼¥Í¥ë¤Ï¤½¤Î¥¹¥ì¥Ã¥É¤Î¼Â¸ú¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Ë
ɬÍפʥ±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬¤¢¤ë¤«¤ò³Îǧ¤¹¤ë¡£
.IP 2.
.\"O The kernel must provide system calls allowing a thread's capability sets to
.\"O be changed and retrieved.
¥«¡¼¥Í¥ë¤Ç¡¢¤¢¤ë¥¹¥ì¥Ã¥É¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤òÊѹ¹¤·¤¿¤ê¡¢
¼èÆÀ¤·¤¿¤ê¤Ç¤­¤ë¥·¥¹¥Æ¥à¥³¡¼¥ë¤¬Ä󶡤µ¤ì¤ë¡£
.IP 3.
.\"O The file system must support attaching capabilities to an executable file,
.\"O so that a process gains those capabilities when the file is executed.
¥Õ¥¡¥¤¥ë¥·¥¹¥Æ¥à¤¬¡¢¼Â¹Ô²Äǽ¥Õ¥¡¥¤¥ë¤Ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤òÉÕÍ¿¤Ç¤­¡¢¥Õ¥¡¥¤¥ë
¼Â¹Ô»þ¤Ë¤½¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò¥×¥í¥»¥¹¤¬¼èÆÀ¤Ç¤­¤ë¤è¤¦¤Êµ¡Ç½¤ò¥µ¥Ý¡¼¥È¤¹¤ë¡£
.PP
.\"O Before kernel 2.6.24, only the first two of these requirements are met;
.\"O since kernel 2.6.24, all three requirements are met.
¥«¡¼¥Í¥ë 2.6.24 ¤è¤êÁ°¤Ç¤Ï¡¢ºÇ½é¤Î 2¤Ä¤ÎÍ×·ï¤Î¤ß¤¬Ëþ¤¿¤µ¤ì¤Æ¤¤¤ë¡£
¥«¡¼¥Í¥ë 2.6.24 °Ê¹ß¤Ç¤Ï¡¢3¤Ä¤ÎÍ׷魯¤Ù¤Æ¤¬Ëþ¤¿¤µ¤ì¤Æ¤¤¤ë¡£
.\"
.\"O .SS Thread Capability Sets
.SS ¥¹¥ì¥Ã¥É¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È
.\"O Each thread has three capability sets containing zero or more
.\"O of the above capabilities:
³Æ¥¹¥ì¥Ã¥É¤Ï°Ê²¼¤Î 3¼ïÎà¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ò»ý¤Ä¡£³Æ¡¹¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Ï
¾åµ­¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ÎÁȤ߹ç¤ï¤»¤Ç¤¢¤ë (Á´¤Æ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬Ìµ¸ú¤Ç¤â¤è¤¤)¡£
.TP
.\"O .IR Permitted :
.\"O This is a limiting superset for the effective
.\"O capabilities that the thread may assume.
.\"O It is also a limiting superset for the capabilities that
.\"O may be added to the inheritable set by a thread that does not have the
.\"O .B CAP_SETPCAP
.\"O capability in its effective set.
.IR "µö²Ä (permitted)" :
¤½¤Î¥¹¥ì¥Ã¥É¤¬»ý¤Ä¤³¤È¤Ë¤Ê¤Ã¤Æ¤¤¤ë¼Â¸ú¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Î
¸ÂÄêŪ¤Ê¥¹¡¼¥Ñ¡¼¥»¥Ã¥È¤Ç¤¢¤ë¡£
¤³¤ì¤Ï¡¢¼Â¸ú¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Ë
.B CAP_SETPCAP
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò»ý¤Ã¤Æ¤¤¤Ê¤¤¥¹¥ì¥Ã¥É¤¬·Ñ¾µ²Äǽ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Ë
ÄɲòÄǽ¤Ê¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Î¸ÂÄêŪ¤Ê¥¹¡¼¥Ñ¡¼¥»¥Ã¥È¤Ç¤â¤¢¤ë¡£

.\"O If a thread drops a capability from its permitted set,
.\"O it can never reacquire that capability (unless it
.\"O .BR execve (2)s
.\"O either a set-user-ID-root program, or
.\"O a program whose associated file capabilities grant that capability).
µö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤«¤éºï½ü¤·¤Æ¤·¤Þ¤Ã¤¿¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Ï¡¢
(set-user-ID-root ¥×¥í¥°¥é¥à¤«¡¢
¤½¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Çµö²Ä¤·¤Æ¤¤¤ë¥×¥í¥°¥é¥à¤ò
.BR execve (2)
¤·¤Ê¤¤¸Â¤ê¤Ï) ¤â¤¦°ìÅÙ³ÍÆÀ¤¹¤ë¤³¤È¤Ï¤Ç¤­¤Ê¤¤¡£
.TP
.\"O .IR Inheritable :
.\"O This is a set of capabilities preserved across an
.\"O .BR execve (2).
.\"O It provides a mechanism for a process to assign capabilities
.\"O to the permitted set of the new program during an
.\"O .BR execve (2).
.IR "·Ñ¾µ²Äǽ (inheritable)" :
.BR execve (2)
¤òÁ°¸å¤ÇÊÝ»ý¤µ¤ì¤ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Ç¤¢¤ë¡£
¤³¤Î»ÅÁȤߤò»È¤¦¤³¤È¤Ç¡¢¤¢¤ë¥×¥í¥»¥¹¤¬
.BR execve (2)
¤ò¹Ô¤¦ºÝ¤Ë¿·¤·¤¤¥×¥í¥°¥é¥à¤Îµö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤È¤·¤Æ
³ä¤êÅö¤Æ¤ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò»ØÄꤹ¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
.TP
.\"O .IR Effective :
.\"O This is the set of capabilities used by the kernel to
.\"O perform permission checks for the thread.
.IR "¼Â¸ú (effective)" :
¥«¡¼¥Í¥ë¤¬¥¹¥ì¥Ã¥É¤Î¸¢¸Â (permission) ¤ò¥Á¥§¥Ã¥¯¤¹¤ë¤È¤­¤Ë
»ÈÍѤ¹¤ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Ç¤¢¤ë¡£
.PP
.\"O A child created via
.\"O .BR fork (2)
.\"O inherits copies of its parent's capability sets.
.\"O See below for a discussion of the treatment of capabilities during
.\"O .BR execve (2).
.BR fork (2)
¤ÇºîÀ®¤µ¤ì¤ë»Ò¥×¥í¥»¥¹¤Ï¡¢¿Æ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Î¥³¥Ô¡¼¤ò·Ñ¾µ¤¹¤ë¡£
.BR execve (2)
Ãæ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Î°·¤¤¤Ë¤Ä¤¤¤Æ¤Ï²¼µ­¤ò»²¾È¤Î¤³¤È¡£
.PP
.\"O Using
.\"O .BR capset (2),
.\"O a thread may manipulate its own capability sets (see below).
.BR capset (2)
¤ò»È¤¦¤È¡¢¥×¥í¥»¥¹¤Ï¼«Ê¬¼«¿È¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È
¤òÁàºî¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë (²¼µ­»²¾È)¡£
.\"
.\"O .SS File Capabilities
.SS ¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£
.\"O Since kernel 2.6.24, the kernel supports
.\"O associating capability sets with an executable file using
.\"O .BR setcap (8).
.\"O The file capability sets are stored in an extended attribute (see
.\"O .BR setxattr (2))
.\"O named
.\"O .IR "security.capability" .
.\"O Writing to this extended attribute requires the
.\"O .BR CAP_SETFCAP
.\"O capability.
¥«¡¼¥Í¥ë 2.6.24 °Ê¹ß¤Ç¤Ï¡¢
.BR setcap (8)
¤ò»È¤Ã¤Æ¼Â¹Ô¥Õ¥¡¥¤¥ë¤Ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤òÂбþÉÕ¤±¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Ï
.I "security.capability"
¤È¤¤¤¦Ì¾Á°¤Î³Èĥ°À­¤ËÊݸ¤µ¤ì¤ë
.RB ( setxattr (2)
»²¾È)¡£¤³¤Î³Èĥ°À­¤Ø¤Î½ñ¤­¹þ¤ß¤Ë¤Ï
.B CAP_SETFCAP
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬É¬ÍפǤ¢¤ë¡£
.\"O The file capability sets,
.\"O in conjunction with the capability sets of the thread,
.\"O determine the capabilities of a thread after an
.\"O .BR execve (2).
¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤È¥¹¥ì¥Ã¥É¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ÎξÊý¤¬
¹Í褵¤ì¡¢
.BR execve (2)
¸å¤Î¥¹¥ì¥Ã¥É¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤¬·èÄꤵ¤ì¤ë¡£

.\"O The three file capability sets are:
3 ¤Ä¤Î¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤¬ÄêµÁ¤µ¤ì¤Æ¤¤¤ë¡£
.TP
.\"O .IR Permitted " (formerly known as " forced ):
.\"O These capabilities are automatically permitted to the thread,
.\"O regardless of the thread's inheritable capabilities.
.IR "µö²Ä (Permitted)" " (°ÊÁ°¤Î" "¶¯À© (Forced)" "):"
¥¹¥ì¥Ã¥É¤Î·Ñ¾µ²Äǽ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Ë´Ø¤ï¤é¤º¡¢¤½¤Î¥¹¥ì¥Ã¥É¤Ë¼«Æ°Åª¤Ë
ǧ¤á¤é¤ì¤ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡£
.TP
.\"O .IR Inheritable " (formerly known as " allowed ):
.\"O This set is ANDed with the thread's inheritable set to determine which
.\"O inheritable capabilities are enabled in the permitted set of
.\"O the thread after the
.\"O .BR execve (2).
.IR "·Ñ¾µ²Äǽ (Inheritable)" " (°ÊÁ°¤Î " "µöÍÆ (Allowed)" "):"
¤³¤Î¥»¥Ã¥È¤È¡¢¥¹¥ì¥Ã¥É¤Î·Ñ¾µ²Äǽ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤È¤Î
ÏÀÍýÀÑ (AND) ¤¬¤È¤é¤ì¡¢
.BR execve (2)
¤Î¸å¤Ë¤½¤Î¥¹¥ì¥Ã¥É¤Îµö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ÇÍ­¸ú¤È¤Ê¤ë
·Ñ¾µ²Äǽ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬·èÄꤵ¤ì¤ë¡£
.TP
.\"O .IR Effective :
.IR "¼Â¸ú (Effective)" :
.\"O This is not a set, but rather just a single bit.
.\"O If this bit is set, then during an
.\"O .BR execve (2)
.\"O all of the new permitted capabilities for the thread are
.\"O also raised in the effective set.
.\"O If this bit is not set, then after an
.\"O .BR execve (2),
.\"O none of the new permitted capabilities is in the new effective set.
¤³¤ì¤Ï½¸¹ç¤Ç¤Ï¤Ê¤¯¡¢1 ¥Ó¥Ã¥È¤Î¾ðÊó¤Ç¤¢¤ë¡£
¤³¤Î¥Ó¥Ã¥È¤¬¥»¥Ã¥È¤µ¤ì¤Æ¤¤¤ë¤È¡¢
.BR execve (2)
¼Â¹ÔÃæ¤Ë¡¢¤½¤Î¥¹¥ì¥Ã¥É¤Î¿·¤·¤¤µö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬Á´¤Æ
¼Â¸ú¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£½¸¹ç¤Ë¤ª¤¤¤Æ¤â¥»¥Ã¥È¤µ¤ì¤ë¡£
¤³¤Î¥Ó¥Ã¥È¤¬¥»¥Ã¥È¤µ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç¡¢
.BR execve (2)
¸å¤Ë¤Ï¿·¤·¤¤µö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Î¤É¤ì¤â¿·¤·¤¤¼Â¸ú¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£½¸¹ç
¤Ë¥»¥Ã¥È¤µ¤ì¤Ê¤¤¡£

.\"O Enabling the file effective capability bit implies
.\"O that any file permitted or inheritable capability that causes a
.\"O thread to acquire the corresponding permitted capability during an
.\"O .BR execve (2)
.\"O (see the transformation rules described below) will also acquire that
.\"O capability in its effective set.
¥Õ¥¡¥¤¥ë¤Î¼Â¸ú¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥Ó¥Ã¥È¤òÍ­¸ú¤Ë¤¹¤ë¤È¤¤¤¦¤Î¤Ï¡¢
.BR execve (2)
¼Â¹Ô»þ¤Ë¡¢¥Õ¥¡¥¤¥ë¤Îµö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤È·Ñ¾µ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ËÂбþ¤¹¤ë¤â¤Î¤¬
¥¹¥ì¥Ã¥É¤Îµö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤È¤·¤Æ¥»¥Ã¥È¤µ¤ì¤ë¤¬¡¢
¤³¤ì¤¬¼Â¸ú¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Ë¤â¥»¥Ã¥È¤µ¤ì¤ë¤È¤¤¤¦¤³¤È¤Ç¤¢¤ë
(¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ÎÊÑ´¹¥ë¡¼¥ë¤Ï²¼µ­»²¾È)¡£
.\"O Therefore, when assigning capabilities to a file
.\"O .RB ( setcap (8),
.\"O .BR cap_set_file (3),
.\"O .BR cap_set_fd (3)),
.\"O if we specify the effective flag as being enabled for any capability,
.\"O then the effective flag must also be specified as enabled
.\"O for all other capabilities for which the corresponding permitted or
.\"O inheritable flags is enabled.
¤·¤¿¤¬¤Ã¤Æ¡¢¥Õ¥¡¥¤¥ë¤Ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò³ä¤êÅö¤Æ¤ëºÝ
.RB ( setcap (8),
.BR cap_set_file (3),
.BR cap_set_fd (3))¡¢
¤¤¤º¤ì¤«¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ËÂФ·¤Æ¼Â¸ú¥Õ¥é¥°¤òÍ­¸ú¤È»ØÄꤹ¤ë¾ì¹ç¡¢
µö²Ä¥Õ¥é¥°¤ä·Ñ¾µ²Äǽ¥Õ¥é¥°¤òÍ­¸ú¤Ë¤·¤¿Â¾¤ÎÁ´¤Æ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£
¤Ë¤Ä¤¤¤Æ¤â¼Â¸ú¥Õ¥é¥°¤òÍ­¸ú¤È»ØÄꤷ¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£
.\"
.\"O .SS Transformation of Capabilities During execve()
.SS "execve() Ãæ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ÎÊÑ´¹"
.PP
.\"O During an
.\"O .BR execve (2),
.\"O the kernel calculates the new capabilities of
.\"O the process using the following algorithm:
.BR execve (2)
¼Â¹Ô»þ¤Ë¡¢¥«¡¼¥Í¥ë¤Ï¥×¥í¥»¥¹¤Î¿·¤·¤¤¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò¼¡¤Î
¥¢¥ë¥´¥ê¥º¥à¤òÍѤ¤¤Æ·×»»¤¹¤ë¡§
.in +4n
.nf

P'(permitted) = (P(inheritable) & F(inheritable)) |
                (F(permitted) & cap_bset)

P'(effective) = F(effective) ? P'(permitted) : 0

.\"O P'(inheritable) = P(inheritable)    [i.e., unchanged]
P'(inheritable) = P(inheritable)    [¤Ä¤Þ¤ê¡¢Êѹ¹¤µ¤ì¤Ê¤¤]

.fi
.in
.\"O where:
³ÆÊÑ¿ô¤Î°ÕÌ£¤Ï°Ê²¼¤ÎÄ̤ê:
.RS 4
.IP P 10
.\"O denotes the value of a thread capability set before the
.\"O .BR execve (2)
.BR execve (2)
Á°¤Î¥¹¥ì¥Ã¥É¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ÎÃÍ
.IP P'
.\"O denotes the value of a capability set after the
.\"O .BR execve (2)
.BR execve (2)
¸å¤Î¥¹¥ì¥Ã¥É¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ÎÃÍ
.IP F
.\"O denotes a file capability set
¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ÎÃÍ
.IP cap_bset
.\"O is the value of the capability bounding set (described below).
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤ÎÃÍ (²¼µ­»²¾È)
.RE
.\"
.\"O .SS Capabilities and execution of programs by root
.SS ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤È¡¢¥ë¡¼¥È¤Ë¤è¤ë¥×¥í¥°¥é¥à¤Î¼Â¹Ô
.\"O In order to provide an all-powerful
.\"O .I root
.\"O using capability sets, during an
.\"O .BR execve (2):
.BR execve (2)
»þ¤Ë¡¢¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ò»È¤Ã¤Æ¡¢Á´¤Æ¤Î¸¢¸Â¤ò»ý¤Ã¤¿
.I root
¤ò¼Â¸½¤¹¤ë¤Ë¤Ï¡¢°Ê²¼¤Î¤è¤¦¤Ë¤¹¤ë¡£
.IP 1. 3
.\"O If a set-user-ID-root program is being executed,
.\"O or the real user ID of the process is 0 (root)
.\"O then the file inheritable and permitted sets are defined to be all ones
.\"O (i.e., all capabilities enabled).
set-user-ID-root ¥×¥í¥°¥é¥à¤¬¼Â¹Ô¤µ¤ì¤ë¾ì¹ç¡¢
¤Þ¤¿¤Ï¥×¥í¥»¥¹¤Î¼Â¥æ¡¼¥¶ ID ¤¬ 0 (root) ¤Î¾ì¹ç¡¢
¥Õ¥¡¥¤¥ë¤Î·Ñ¾µ²Äǽ¥»¥Ã¥È¤Èµö²Ä¥»¥Ã¥È¤òÁ´¤Æ 1
(Á´¤Æ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬Í­¸ú) ¤ËÄêµÁ¤¹¤ë¡£
.IP 2.
.\"O If a set-user-ID-root program is being executed,
.\"O then the file effective bit is defined to be one (enabled).
set-user-ID-root ¥×¥í¥°¥é¥à¤¬¼Â¹Ô¤µ¤ì¤ë¾ì¹ç¡¢
¥Õ¥¡¥¤¥ë¤Î¼Â¸ú¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥Ó¥Ã¥È¤ò 1 (enabled) ¤ËÄêµÁ¤¹¤ë¡£
.PP
.\"O The upshot of the above rules,
.\"O combined with the capabilities transformations described above,
.\"O is that when a process
.\"O .BR execve (2)s
.\"O a set-user-ID-root program, or when a process with an effective UID of 0
.\"O .BR execve (2)s
.\"O a program,
.\"O it gains all capabilities in its permitted and effective capability sets,
.\"O except those masked out by the capability bounding set.
.\"O .\" If a process with real UID 0, and nonzero effective UID does an
.\"O .\" exec(), then it gets all capabilities in its
.\"O .\" permitted set, and no effective capabilities
.\"O This provides semantics that are the same as those provided by
.\"O traditional UNIX systems.
¾åµ­¤Î¥ë¡¼¥ë¤Ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£ÊÑ´¹¤òŬÍѤ·¤¿·ë²Ì¤ò¤Þ¤È¤á¤ë¤È¡¢
¥×¥í¥»¥¹¤¬ set-user-ID-root ¥×¥í¥°¥é¥à¤ò
.BR execve (2)
¤¹¤ë¾ì¹ç¡¢¤Þ¤¿¤Ï¼Â¸ú UID ¤¬ 0 ¤Î¥×¥í¥»¥¹¤¬¥×¥í¥°¥é¥à¤ò
.BR execve (2)
¤¹¤ë¾ì¹ç¡¢µö²Ä¤È¼Â¸ú¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ÎÁ´¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£
(Àµ³Î¤Ë¤Ï¡¢¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Ë¤è¤ë¥Þ¥¹¥¯¤Ç½ü³°¤µ¤ì¤ë¤â¤Î
°Ê³°¤ÎÁ´¤Æ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£) ¤ò¼èÆÀ¤¹¤ë¤È¤¤¤¦¤³¤È¤Ç¤¢¤ë¡£
.\" ¼Â UID ¤¬ 0 ¤Ç¼Â¸ú UID ¤¬ 0 °Ê³°¤Î¥×¥í¥»¥¹¤¬ exec () ¤ò¹Ô¤¦¤È¡¢
.\" µö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Ë´Þ¤Þ¤ì¤ëÁ´¤Æ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£
.\" ¤¬¼èÆÀ¤µ¤ì¡¢¼Â¸ú¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Ï¼èÆÀ¤µ¤ì¤Ê¤¤¡£
¤³¤ì¤Ë¤è¤ê¡¢ÅÁÅýŪ¤Ê UNIX ¥·¥¹¥Æ¥à¤ÈƱ¤¸¿¶¤ëÉñ¤¤¤¬¤Ç¤­¤ë¤è¤¦¤Ë¤Ê¤Ã¤Æ¤¤¤ë¡£
.\"O .SS Capability bounding set
.SS ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È
.\"O The capability bounding set is a security mechanism that can be used
.\"O to limit the capabilities that can be gained during an
.\"O .BR execve (2).
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È (capability bounding set) ¤Ï¡¢
.BR execve (2)
»þ¤Ë³ÍÆÀ¤Ç¤­¤ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤òÀ©¸Â¤¹¤ë¤¿¤á¤Ë»È¤ï¤ì¤ë
¥»¥­¥å¥ê¥Æ¥£µ¡¹½¤Ç¤¢¤ë¡£
.\"O The bounding set is used in the following ways:
¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Ï°Ê²¼¤Î¤è¤¦¤Ë»ÈÍѤµ¤ì¤ë¡£
.IP * 2
.\"O During an
.\"O .BR execve (2),
.\"O the capability bounding set is ANDed with the file permitted
.\"O capability set, and the result of this operation is assigned to the
.\"O thread's permitted capability set.
.\"O The capability bounding set thus places a limit on the permitted
.\"O capabilities that may be granted by an executable file.
.BR execve (2)
¼Â¹Ô»þ¤Ë¡¢¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤È
¥Õ¥¡¥¤¥ë¤Îµö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ÎÏÀÍýÏ (AND) ¤ò¼è¤Ã¤¿¤â¤Î¤¬¡¢
¤½¤Î¥¹¥ì¥Ã¥É¤Îµö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Ë³ä¤êÅö¤Æ¤é¤ì¤ë¡£
¤Ä¤Þ¤ê¡¢¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Ï¡¢
¼Â¹Ô¥Õ¥¡¥¤¥ë¤¬Ç§¤á¤Æ¤¤¤ëµö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ËÂФ·¤Æ
À©¸Â¤ò²Ý¤¹Æ¯¤­¤ò¤¹¤ë¡£
.IP *
.\"O (Since Linux 2.6.25)
.\"O The capability bounding set acts as a limiting superset for
.\"O the capabilities that a thread can add to its inheritable set using
.\"O .BR capset (2).
.\"O This means that if a capability is not in the bounding set,
.\"O then a thread can't add this capability to its
.\"O inheritable set, even if it was in its permitted capabilities,
.\"O and thereby cannot have this capability preserved in its
.\"O permitted set when it
.\"O .BR execve (2)s
.\"O a file that has the capability in its inheritable set.
(Linux 2.6.25 °Ê¹ß)
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Ï¡¢¥¹¥ì¥Ã¥É¤¬
.BR capset (2)
¤Ë¤è¤ê¼«¿È¤Î·Ñ¾µ²Äǽ¥»¥Ã¥È¤ËÄɲòÄǽ¤Ê¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ÎÊ콸ÃĤò
À©¸Â¤¹¤ëÌò³ä¤ò»ý¤Ä¡£
¥¹¥ì¥Ã¥É¤Ëµö²Ä¤µ¤ì¤¿¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Ç¤¢¤Ã¤Æ¤â¡¢¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Ë
´Þ¤Þ¤ì¤Æ¤¤¤Ê¤±¤ì¤Ð¡¢¥¹¥ì¥Ã¥É¤Ï¤½¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Ï¼«¿È¤Î·Ñ¾µ²Äǽ¥»¥Ã¥È¤Ë
ÄɲäǤ­¤º¡¢¤½¤Î·ë²Ì¡¢·Ñ¾µ²Äǽ¥»¥Ã¥È¤Ë¤½¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò´Þ¤à¥Õ¥¡¥¤¥ë¤ò
.BR execve (2)
¤¹¤ë¾ì¹ç¡¢¤½¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤òµö²Ä¥»¥Ã¥È¤Ë»ý¤Á³¤±¤ë¤³¤È¤¬¤Ç¤­¤Ê¤¤¡¢
¤È¤¤¤¦¤³¤È¤Ç¤¢¤ë¡£
.PP
.\"O Note that the bounding set masks the file permitted capabilities,
.\"O but not the inherited capabilities.
.\"O If a thread maintains a capability in its inherited set
.\"O that is not in its bounding set,
.\"O then it can still gain that capability in its permitted set
.\"O by executing a file that has the capability in its inherited set.
¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤¬¥Þ¥¹¥¯¤ò¹Ô¤¦¤Î¤Ï¡¢·Ñ¾µ²Äǽ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Ç¤Ï¤Ê¤¯¡¢
¥Õ¥¡¥¤¥ë¤Îµö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Î¥Þ¥¹¥¯¤ò¹Ô¤¦ÅÀ¤ËÃí°Õ¤¹¤ë¤³¤È¡£
¤¢¤ë¥¹¥ì¥Ã¥É¤Î·Ñ¾µ²Äǽ¥»¥Ã¥È¤Ë¤½¤Î¥¹¥ì¥Ã¥É¤Î¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Ë
¸ºß¤·¤Ê¤¤¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬´Þ¤Þ¤ì¤Æ¤¤¤ë¾ì¹ç¡¢¤½¤Î¥¹¥ì¥Ã¥É¤Ï¡¢
·Ñ¾µ²Äǽ¥»¥Ã¥È¤Ë´Þ¤Þ¤ì¤ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò»ý¤Ä¥Õ¥¡¥¤¥ë¤ò¼Â¹Ô¤¹¤ë¤³¤È¤Ë¤è¤ê¡¢
µö²Ä¥»¥Ã¥È¤Ë´Þ¤Þ¤ì¤ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤â³ÍÆÀ¤Ç¤­¤ë¤È¤¤¤¦¤³¤È¤Ç¤¢¤ë¡£
.PP
.\"O Depending on the kernel version, the capability bounding set is either
.\"O a system-wide attribute, or a per-process attribute.
¥«¡¼¥Í¥ë¤Î¥Ð¡¼¥¸¥ç¥ó¤Ë¤è¤ê¡¢¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Ï
¥·¥¹¥Æ¥à¶¦Ä̤ΰÀ­¤Î¾ì¹ç¤È¡¢¥×¥í¥»¥¹Ã±°Ì¤Î°À­¤Î¾ì¹ç¤¬¤¢¤ë¡£
.PP
.\"O .B "Capability bounding set prior to Linux 2.6.25"
.B "Linux 2.6.25 ¤è¤êÁ°¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È"
.PP
.\"O In kernels before 2.6.25, the capability bounding set is a system-wide
.\"O attribute that affects all threads on the system.
2.6.25 ¤è¤êÁ°¤Î¥«¡¼¥Í¥ë¤Ç¤Ï¡¢¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Ï
¥·¥¹¥Æ¥à¶¦Ä̤ΰÀ­¤Ç¡¢¥·¥¹¥Æ¥à¾å¤ÎÁ´¤Æ¤Î¥¹¥ì¥Ã¥É¤ËŬÍѤµ¤ì¤ë¡£
.\"O The bounding set is accessible via the file
.\"O .IR /proc/sys/kernel/cap-bound .
.\"O motoki: accessible = ¡Ö»²¾È²Äǽ¡×¤Ç¤è¤¤¤«¡¢Ê¸Ì®¤òÍ׳Îǧ
¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Ï
.I /proc/sys/kernel/cap-bound
¥Õ¥¡¥¤¥ë·Ðͳ¤Ç»²¾È¤Ç¤­¤ë¡£
.\"O (Confusingly, this bit mask parameter is expressed as a
.\"O signed decimal number in
.\"O .IR /proc/sys/kernel/cap-bound .)
(´Ö°ã¤¨¤ä¤¹¤¤¤¬¡¢¤³¤Î¥Ó¥Ã¥È¥Þ¥¹¥¯·Á¼°¤Î¥Ñ¥é¥á¡¼¥¿¤Ï¡¢
.I /proc/sys/kernel/cap-bound
¤Ç¤ÏÉä¹æÉÕ¤­¤Î½½¿Ê¿ô¤Çɽ¸½¤µ¤ì¤ë¡£)

.\"O Only the
.\"O .B init
.\"O process may set capabilities in the capability bounding set;
.\"O other than that, the superuser (more precisely: programs with the
.\"O .B CAP_SYS_MODULE
.\"O capability) may only clear capabilities from this set.
.B init
¥×¥í¥»¥¹¤À¤±¤¬¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Ç
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò¥»¥Ã¥È¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
¤½¤ì°Ê³°¤Ç¤Ï¡¢¥¹¡¼¥Ñ¡¼¥æ¡¼¥¶ (¤è¤êÀµ³Î¤Ë¤Ï¡¢
.B CAP_SYS_MODULE
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò»ý¤Ã¤¿¥×¥í¥°¥é¥à) ¤¬¡¢
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Î¥¯¥ê¥¢¤¬
¤Ç¤­¤ë¤À¤±¤Ç¤¢¤ë¡£

.\"O On a standard system the capability bounding set always masks out the
.\"O .B CAP_SETPCAP
.\"O capability.
.\"O To remove this restriction (dangerous!), modify the definition of
.\"O .B CAP_INIT_EFF_SET
.\"O in
.\"O .I include/linux/capability.h
.\"O and rebuild the kernel.
Ä̾ï¤Î¥·¥¹¥Æ¥à¤Ç¤Ï¡¢¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Ï¡¢
.B CAP_SETPCAP
¤¬Ìµ¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤ë¡£
¤³¤ÎÀ©¸Â¤ò¼è¤êµî¤ë¤Ë¤Ï (¼è¤êµî¤ë¤Î¤Ï´í¸±!)¡¢
.I include/linux/capability.h
Æâ¤Î
.B CAP_INIT_EFF_SET
¤ÎÄêµÁ¤ò½¤Àµ¤·¡¢¥«¡¼¥Í¥ë¤òºÆ¹½ÃÛ¤¹¤ëɬÍפ¬¤¢¤ë¡£

.\"O The system-wide capability bounding set feature was added
.\"O to Linux starting with kernel version 2.2.11.
¥·¥¹¥Æ¥à¶¦Ä̤Υ±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥Èµ¡Ç½¤Ï¡¢
¥«¡¼¥Í¥ë 2.2.11 °Ê¹ß¤Ç Linux ¤ËÄɲ䵤줿¡£
.\"
.PP
.\"O .B "Capability bounding set from Linux 2.6.25 onward"
.B "Linux 2.6.25 °Ê¹ß¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È"
.PP
.\"O From Linux 2.6.25, the
.\"O .I "capability bounding set"
.\"O is a per-thread attribute.
.\"O (There is no longer a system-wide capability bounding set.)
Linux 2.6.25 °Ê¹ß¤Ç¤Ï¡¢
¡Ö¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¡×¤Ï¥¹¥ì¥Ã¥Éñ°Ì¤Î°À­¤Ç¤¢¤ë
(¥·¥¹¥Æ¥à¶¦Ä̤Υ±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Ï¤â¤Ï¤ä¸ºß¤·¤Ê¤¤)¡£

.\"O The bounding set is inherited at
.\"O .BR fork (2)
.\"O from the thread's parent, and is preserved across an
.\"O .BR execve (2).
¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Ï
.BR fork (2)
»þ¤Ë¤Ï¥¹¥ì¥Ã¥É¤Î¿Æ¥×¥í¥»¥¹¤«¤é·Ñ¾µ¤µ¤ì¡¢
.BR execve (2)
¤ÎÁ°¸å¤Ç¤ÏÊÝ»ý¤µ¤ì¤ë¡£

.\"O A thread may remove capabilities from its capability bounding set using the
.\"O .BR prctl (2)
.\"O .B PR_CAPBSET_DROP
.\"O operation, provided it has the
.\"O .B CAP_SETPCAP
.\"O capability.
.\"O Once a capability has been dropped from the bounding set,
.\"O it cannot be restored to that set.
.\"O A thread can determine if a capability is in its bounding set using the
.\"O .BR prctl (2)
.\"O .B PR_CAPBSET_READ
.\"O operation.
¥¹¥ì¥Ã¥É¤¬
.B CAP_SETPCAP
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò»ý¤Ã¤Æ¤¤¤ë¾ì¹ç¡¢¤½¤Î¥¹¥ì¥Ã¥É¤Ï
.BR prctl (2)
¤Î
.BR PR_CAPBSET_DROP
Áàºî¤ò»È¤Ã¤Æ¼«¿È¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤«¤é
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤òºï½ü¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
¤¤¤Ã¤¿¤ó¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤«¤éºï½ü¤·¤Æ¤·¤Þ¤¦¤È¡¢
¥¹¥ì¥Ã¥É¤Ï¤½¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤òºÆÅÙ¥»¥Ã¥È¤¹¤ë¤³¤È¤Ï¤Ç¤­¤Ê¤¤¡£
.BR prctl (2)
¤Î
.B PR_CAPBSET_READ
Áàºî¤ò»È¤¦¤³¤È¤Ç¡¢¥¹¥ì¥Ã¥É¤¬¤¢¤ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬¼«¿È¤Î¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È
¤Ë´Þ¤Þ¤ì¤Æ¤¤¤ë¤«¤òÃΤ뤳¤È¤¬¤Ç¤­¤ë¡£

.\"O Removing capabilities from the bounding set is only supported if file
.\"O capabilities are compiled into the kernel
.\"O (CONFIG_SECURITY_FILE_CAPABILITIES).
.\"O In that case, the
.\"O .B init
.\"O process (the ancestor of all processes) begins with a full bounding set.
.\"O If file capabilities are not compiled into the kernel, then
.\"O .B init
.\"O begins with a full bounding set minus
.\"O .BR CAP_SETPCAP ,
.\"O because this capability has a different meaning when there are
.\"O no file capabilities.
¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤«¤é¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Îºï½ü¤¬¥µ¥Ý¡¼¥È¤µ¤ì¤ë¤Î¤Ï¡¢
¥«¡¼¥Í¥ë¤Î¥³¥ó¥Ñ¥¤¥ë»þ¤Ë¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬Í­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤ë¾ì¹ç
(CONFIG_SECURITY_FILE_CAPABILITIES) ¤À¤±¤Ç¤¢¤ë¡£
¤³¤Î¾ì¹ç¤Ë¤Ï¡¢ (Á´¤Æ¤Î¥×¥í¥»¥¹¤ÎÀèÁĤǤ¢¤ë) 
.I init
¥×¥í¥»¥¹¤Ï¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤ÇÁ´¤Æ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬
¥»¥Ã¥È¤µ¤ì¤¿¾õÂ֤dz«»Ï¤¹¤ë¡£
¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬Í­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤Ê¤¤¾ì¹ç¤Ë¤Ï¡¢
.I init
¤Ï¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤Ç
.B CAP_SETPCAP
°Ê³°¤ÎÁ´¤Æ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬¥»¥Ã¥È¤µ¤ì¤¿¾õÂ֤dz«»Ï¤¹¤ë¡£
¤³¤Î¤è¤¦¤Ë¤Ê¤Ã¤Æ¤¤¤ë¤Î¤Ï¡¢
.B CAP_SETPCAP
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬¥µ¥Ý¡¼¥È¤µ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç¤Ë¤Ï
°ã¤Ã¤¿°ÕÌ£¤ò»ý¤Ä¤«¤é¤Ç¤¢¤ë¡£

.\"O Removing a capability from the bounding set does not remove it
.\"O from the thread's inherited set.
.\"O However it does prevent the capability from being added
.\"O back into the thread's inherited set in the future.
¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤«¤é¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤òºï½ü¤·¤Æ¤â¡¢
¥¹¥ì¥Ã¥É¤Î·Ñ¾µ²Äǽ¥»¥Ã¥È¤«¤é¤Ï¤½¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Ïºï½ü¤µ¤ì¤Ê¤¤¡£
¤·¤«¤·¤Ê¤¬¤é¡¢¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤«¤é¤Îºï½ü¤Ë¤è¤ê¡¢
¤³¤ÎÀ褽¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò¥¹¥ì¥Ã¥É¤Î·Ñ¾µ²Äǽ¥»¥Ã¥È¤ËÄɲ乤뤳¤È
¤Ï¤Ç¤­¤Ê¤¯¤Ê¤ë¡£
.\"
.\"
.\"O .SS Effect of User ID Changes on Capabilities
.SS "¥æ¡¼¥¶ ID Êѹ¹¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Ø¤Î±Æ¶Á"
.\"O To preserve the traditional semantics for transitions between
.\"O 0 and nonzero user IDs,
.\"O the kernel makes the following changes to a thread's capability
.\"O sets on changes to the thread's real, effective, saved set,
.\"O and file system user IDs (using
.\"O .BR setuid (2),
.\"O .BR setresuid (2),
.\"O or similar):
¥æ¡¼¥¶ ID ¤¬ 0 ¤È 0 °Ê³°¤Î´Ö¤ÇÊѲ½¤¹¤ëºÝ¤Î¿¶¤ëÉñ¤¤¤ò½¾Íè¤ÈƱ¤¸¤Ë¤¹¤ë¤¿¤á¡¢
¥¹¥ì¥Ã¥É¤Î¼Â UID¡¢¼Â¸ú UID¡¢Êݸ set-user-ID¡¢¥Õ¥¡¥¤¥ë¥·¥¹¥Æ¥à UID ¤¬
.RB ( setuid (2),
.BR setresuid (2)
¤Ê¤É¤ò»È¤Ã¤Æ) Êѹ¹¤µ¤ì¤¿ºÝ¤Ë¡¢¥«¡¼¥Í¥ë¤Ï¤½¤Î¥¹¥ì¥Ã¥É¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Ë
°Ê²¼¤ÎÊѹ¹¤ò¹Ô¤¦:
.IP 1. 3
.\"O If one or more of the real, effective or saved set user IDs
.\"O was previously 0, and as a result of the UID changes all of these IDs
.\"O have a nonzero value,
.\"O then all capabilities are cleared from the permitted and effective
.\"O capability sets.
UID ¤ÎÊѹ¹Á°¤Ë¤Ï¼Â UID¡¢¼Â¸ú UID¡¢Êݸ set-user-ID ¤Î¤¦¤Á
¾¯¤Ê¤¯¤È¤â°ì¤Ä¤¬ 0 ¤Ç¡¢Êѹ¹¸å¤Ë¼Â UID¡¢¼Â¸ú UID¡¢Êݸ set-user-ID ¤¬
¤¹¤Ù¤Æ 0 °Ê³°¤ÎÃͤˤʤ俾ì¹ç¡¢µö²Ä¤È¼Â¸ú¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Î
Á´¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò¥¯¥ê¥¢¤¹¤ë¡£
.IP 2.
.\"O If the effective user ID is changed from 0 to nonzero,
.\"O then all capabilities are cleared from the effective set.
¼Â¸ú UID ¤¬ 0 ¤«¤é 0 °Ê³°¤ËÊѹ¹¤µ¤ì¤¿¾ì¹ç¡¢
¼Â¸ú¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ÎÁ´¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò¥¯¥ê¥¢¤¹¤ë¡£
.IP 3.
.\"O If the effective user ID is changed from nonzero to 0,
.\"O then the permitted set is copied to the effective set.
¼Â¸ú UID ¤¬ 0 °Ê³°¤«¤é 0 ¤ËÊѹ¹¤µ¤ì¤¿¾ì¹ç¡¢
µö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ÎÆâÍƤò¼Â¸ú¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Ë¥³¥Ô¡¼¤¹¤ë¡£
.IP 4.
.\"O If the file system user ID is changed from 0 to nonzero (see
.\"O .BR setfsuid (2))
.\"O then the following capabilities are cleared from the effective set:
.\"O .BR CAP_CHOWN ,
.\"O .BR CAP_DAC_OVERRIDE ,
.\"O .BR CAP_DAC_READ_SEARCH ,
.\"O .BR CAP_FOWNER ,
.\"O .BR CAP_FSETID ,
.\"O .B CAP_LINUX_IMMUTABLE
.\"O (since Linux 2.2.30),
.\"O .BR CAP_MAC_OVERRIDE ,
.\"O and
.\"O .B CAP_MKNOD
.\"O (since Linux 2.2.30).
¥Õ¥¡¥¤¥ë¥·¥¹¥Æ¥à UID ¤¬ 0 ¤«¤é 0 °Ê³°¤ËÊѹ¹¤µ¤ì¤¿¾ì¹ç
.RB ( setfsuid (2)
»²¾È)¡¢¼Â¸ú¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Î°Ê²¼¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬¥¯¥ê¥¢¤µ¤ì¤ë:
.BR CAP_CHOWN ,
.BR CAP_DAC_OVERRIDE ,
.BR CAP_DAC_READ_SEARCH ,
.BR CAP_FOWNER ,
.BR CAP_FSETID ,
.B CAP_LINUX_IMMUTABLE
(Linux 2.2.30 °Ê¹ß),
.BR CAP_MAC_OVERRIDE ,
.B CAP_MKNOD
(Linux 2.2.30 °Ê¹ß)¡£
.\"O If the file system UID is changed from nonzero to 0,
.\"O then any of these capabilities that are enabled in the permitted set
.\"O are enabled in the effective set.
¥Õ¥¡¥¤¥ë¥·¥¹¥Æ¥à UID ¤¬ 0 °Ê³°¤«¤é 0 ¤ËÊѹ¹¤µ¤ì¤¿¾ì¹ç¡¢
¾åµ­¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Î¤¦¤Áµö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ÇÍ­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤ë¤â¤Î¤¬
¼Â¸ú¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ÇÍ­¸ú¤Ë¤µ¤ì¤ë¡£
.PP
.\"O If a thread that has a 0 value for one or more of its user IDs wants
.\"O to prevent its permitted capability set being cleared when it resets
.\"O all of its user IDs to nonzero values, it can do so using the
.\"O .BR prctl (2)
.\"O .B PR_SET_KEEPCAPS
.\"O operation.
³Æ¼ï UID ¤Î¤¦¤Á¾¯¤Ê¤¯¤È¤â°ì¤Ä¤¬ 0 ¤Ç¤¢¤ë¥¹¥ì¥Ã¥É¤¬¡¢
¤½¤Î UID ¤ÎÁ´¤Æ¤¬ 0 °Ê³°¤Ë¤Ê¤Ã¤¿¤È¤­¤Ëµö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤¬
¥¯¥ê¥¢¤µ¤ì¤Ê¤¤¤è¤¦¤Ë¤·¤¿¤¤¾ì¹ç¤Ë¤Ï¡¢
.BR prctl (2)
¤Î
.B PR_SET_KEEPCAPS
Áàºî¤ò»È¤¨¤Ð¤è¤¤¡£
.\"
.\"O .SS Programmatically adjusting capability sets
.SS ¥×¥í¥°¥é¥à¤Ç¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤òÄ´À°¤¹¤ë
.\"O A thread can retrieve and change its capability sets using the
.\"O .BR capget (2)
.\"O and
.\"O .BR capset (2)
.\"O system calls.
.\"O However, the use of
.\"O .BR cap_get_proc (3)
.\"O and
.\"O .BR cap_set_proc (3),
.\"O both provided in the
.\"O .I libcap
.\"O package,
.\"O is preferred for this purpose.
³Æ¥¹¥ì¥Ã¥É¤Ï¡¢
.BR capget (2)
¤ä
.BR capset (2)
¤ò»È¤Ã¤Æ¡¢¼«¿È¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ò¼èÆÀ¤·¤¿¤êÊѹ¹¤·¤¿¤ê¤Ç¤­¤ë¡£
¤¿¤À¤·¡¢¤³¤ì¤ò¹Ô¤¦¤Ë¤Ï¡¢
.I libcap
¥Ñ¥Ã¥±¡¼¥¸¤ÇÄ󶡤µ¤ì¤Æ¤¤¤ë
.BR cap_get_proc (3)
¤ä
.BR cap_set_proc (3)
¤ò»È¤¦¤Î¤¬Ë¾¤Þ¤·¤¤¡£
.\"O The following rules govern changes to the thread capability sets:
¥¹¥ì¥Ã¥É¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ÎÊѹ¹¤Ë¤Ï°Ê²¼¤Î¥ë¡¼¥ë¤¬Å¬ÍѤµ¤ì¤ë¡£
.IP 1. 3
.\"O If the caller does not have the
.\"O .B CAP_SETPCAP
.\"O capability,
.\"O the new inheritable set must be a subset of the combination
.\"O of the existing inheritable and permitted sets.
.\"O [XXX] motoki: combination ¤Ã¤Æ AND ? OR ?
¸Æ¤Ó½Ð¤·Â¦¤¬
.B CAP_SETPCAP
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò»ý¤Ã¤Æ¤¤¤Ê¤¤¾ì¹ç¡¢¿·¤·¤¤·Ñ¾µ²Äǽ¥»¥Ã¥È¤Ï¡¢
´û¸¤Î·Ñ¾µ²Äǽ¥»¥Ã¥È¤Èµö²Ä¥»¥Ã¥È¤ÎÀѽ¸¹ç (AND) ¤ÎÉôʬ½¸¹ç¤Ç
¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£
.IP 2.
.\"O (Since kernel 2.6.25)
.\"O The new inheritable set must be a subset of the combination of the
.\"O existing inheritable set and the capability bounding set.
.\"O [XXX] motoki: combination ¤Ã¤Æ AND ? OR ?
(¥«¡¼¥Í¥ë 2.6.25 °Ê¹ß)
¿·¤·¤¤·Ñ¾µ²Äǽ¥»¥Ã¥È¤Ï¡¢´û¸¤Î·Ñ¾µ²Äǽ¥»¥Ã¥È¤È¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦
¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤ÎÀѽ¸¹ç (AND) ¤ÎÉôʬ½¸¹ç¤Ç¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£
.IP 3.
.\"O The new permitted set must be a subset of the existing permitted set
.\"O (i.e., it is not possible to acquire permitted capabilities
.\"O that the thread does not currently have).
¿·¤·¤¤µö²Ä¥»¥Ã¥È¤Ï¡¢´û¸¤Îµö²Ä¥»¥Ã¥È¤ÎÉôʬ½¸¹ç¤Ç¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤
(¤Ä¤Þ¤ê¡¢¤½¤Î¥¹¥ì¥Ã¥É¤¬¸½ºß»ý¤Ã¤Æ¤¤¤Ê¤¤µö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò
³ÍÆÀ¤¹¤ë¤³¤È¤Ï¤Ç¤­¤Ê¤¤)¡£
.IP 4.
.\"O The new effective set must be a subset of the new permitted set.
¿·¤·¤¤¼Â¸ú¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Ï¿·¤·¤¤µö²Ä¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤Î
Éôʬ½¸¹ç¤Ë¤Ê¤Ã¤Æ¤¤¤Ê¤±¤ì¤Ð¤Ê¤é¤Ê¤¤¡£
.\"O .SS The """securebits"" flags: establishing a capabilities-only environment
.SS securebits ¥Õ¥é¥°: ¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤À¤±¤Î´Ä¶­¤ò¹½ÃÛ¤¹¤ë
.\" For some background:
.\"       see http://lwn.net/Articles/280279/ and
.\"       http://article.gmane.org/gmane.linux.kernel.lsm/5476/
.\"O Starting with kernel 2.6.26,
.\"O and with a kernel in which file capabilities are enabled,
.\"O Linux implements a set of per-thread
.\"O .I securebits
.\"O flags that can be used to disable special handling of capabilities for UID 0
.\"O .RI ( root ).
¥«¡¼¥Í¥ë 2.6.26 °Ê¹ß¤Ç¡¢
¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬Í­¸ú¤Ë¤Ê¤Ã¤¿¥«¡¼¥Í¥ë¤Ç¤Ï¡¢
¥¹¥ì¥Ã¥Éñ°Ì¤Î
.I securebits
¥Õ¥é¥°¤¬¼ÂÁõ¤µ¤ì¤Æ¤ª¤ê¡¢¤³¤Î¥Õ¥é¥°¤ò»È¤¦¤È UID 0
.RI ( root )
¤ËÂФ¹¤ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ÎÆÃÊÌ°·¤¤¤ò̵¸ú¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
.\"O These flags are as follows:
°Ê²¼¤Î¤è¤¦¤Ê¥Õ¥é¥°¤¬¤¢¤ë¡£
.TP
.B SECBIT_KEEP_CAPS
.\"O Setting this flag allows a thread that has one or more 0 UIDs to retain
.\"O its capabilities when it switches all of its UIDs to a nonzero value.
.\"O If this flag is not set,
.\"O then such a UID switch causes the thread to lose all capabilities.
.\"O This flag is always cleared on an
.\"O .BR execve (2).
.\"O (This flag provides the same functionality as the older
.\"O .BR prctl (2)
.\"O .B PR_SET_KEEPCAPS
.\"O operation.)
¤³¤Î¥Õ¥é¥°¤ò¥»¥Ã¥È¤µ¤ì¤Æ¤¤¤ë¾ì¹ç¡¢UID ¤¬ 0 ¤Î¥¹¥ì¥Ã¥É¤Î UID ¤¬ 0 °Ê³°¤ÎÃͤË
ÀÚ¤êÂؤï¤ëºÝ¤Ë¡¢¤½¤Î¥¹¥ì¥Ã¥É¤Ï¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò°Ý»ý¤¹¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
¤³¤Î¥Õ¥é¥°¤¬¥»¥Ã¥È¤µ¤ì¤Æ¤¤¤Ê¤¤¾ì¹ç¤Ë¤Ï¡¢UID ¤¬ 0 ¤«¤é 0 °Ê³°¤ÎÃͤË
ÀÚ¤êÂؤï¤ë¤È¡¢¤½¤Î¥¹¥ì¥Ã¥É¤ÏÁ´¤Æ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò¼º¤¦¡£
¤³¤Î¥Õ¥é¥°¤Ï
.BR execve (2)
»þ¤Ë¤ÏÁ´¤Æ¥¯¥ê¥¢¤µ¤ì¤ë
(¤³¤Î¥Õ¥é¥°¤Ï¡¢°ÊÁ°¤Î
.BR prctl (2)
¤Î
.B PR_SET_KEEPCAPS
Áàºî¤ÈƱ¤¸µ¡Ç½¤òÄ󶡤¹¤ë¤â¤Î¤Ç¤¢¤ë)¡£
.TP
.B SECBIT_NO_SETUID_FIXUP
.\"O Setting this flag stops the kernel from adjusting  capability sets when
.\"O the threads's effective and file system UIDs are switched between
.\"O zero and nonzero values.
.\"O (See the subsection
.\"O .IR "Effect of User ID Changes on Capabilities" .)
¤³¤Î¥Õ¥é¥°¤ò¥»¥Ã¥È¤¹¤ë¤È¡¢¥¹¥ì¥Ã¥É¤Î¼Â¸ú UID ¤È¥Õ¥¡¥¤¥ë¥·¥¹¥Æ¥à UID ¤¬
0 ¤È 0 °Ê³°¤Î´Ö¤ÇÀÚ¤êÂؤï¤Ã¤¿¾ì¹ç¤Ë¡¢
¥«¡¼¥Í¥ë¤Ï¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ÎÄ´À°¤ò¹Ô¤ï¤Ê¤¯¤Ê¤ë
(¡Ö¥æ¡¼¥¶ ID Êѹ¹¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Ø¤Î±Æ¶Á¡×¤ÎÀá¤ò»²¾È)¡£
.TP
.B SECBIT_NOROOT
.\"O If this bit is set, then the kernel does not grant capabilities
.\"O when a set-user-ID-root program is executed, or when a process with
.\"O an effective or real UID of 0 calls
.\"O .BR execve (2).
.\"O (See the subsection
.\"O .IR "Capabilities and execution of programs by root" .)
¤³¤Î¥Ó¥Ã¥È¤¬¥»¥Ã¥È¤µ¤ì¤Æ¤¤¤ë¾ì¹ç¡¢
set-user-ID-root ¥×¥í¥°¥é¥à¤Î¼Â¹Ô»þ¤ä¡¢
¼Â¸ú UID ¤« ¼Â UID ¤¬ 0 ¤Î¥×¥í¥»¥¹¤¬
.BR execve (2)
¤ò¸Æ¤Ó½Ð¤·¤¿»þ¤Ë¡¢¥«¡¼¥Í¥ë¤Ï¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤òµö²Ä¤·¤Ê¤¤
(¡Ö¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤È¡¢¥ë¡¼¥È¤Ë¤è¤ë¥×¥í¥°¥é¥à¤Î¼Â¹Ô¡×¤ÎÀá¤ò»²¾È)¡£
.PP
.\"O Each of the above "base" flags has a companion "locked" flag.
.\"O Setting any of the "locked" flags is irreversible,
.\"O and has the effect of preventing further changes to the
.\"O corresponding "base" flag.
.\"O The locked flags are:
.\"O .BR SECBIT_KEEP_CAPS_LOCKED ,
.\"O .BR SECBIT_NO_SETUID_FIXUP_LOCKED ,
.\"O and
.\"O .BR SECBIT_NOROOT_LOCKED .
¾åµ­¤Î "base" ¥Õ¥é¥°¤Î³Æ¡¹¤Ë¤ÏÂбþ¤¹¤ë "locked" ¥Õ¥é¥°¤¬Â¸ºß¤¹¤ë¡£
¤¤¤º¤ì¤Î "locked" ¥Õ¥é¥°¤â°ìÅÙ¥»¥Ã¥È¤µ¤ì¤ë¤ÈÌ᤹¤³¤È¤Ï¤Ç¤­¤º¡¢
¤½¤ì°Ê¹ß¤ÏÂбþ¤¹¤ë "base" ¥Õ¥é¥°¤òÊѹ¹¤¹¤ë¤³¤È¤¬¤Ç¤­¤Ê¤¯¤Ê¤ë¡£
"locked" ¥Õ¥é¥°¤Ï
.BR SECBIT_KEEP_CAPS_LOCKED ,
.BR SECBIT_NO_SETUID_FIXUP_LOCKED ,
.BR SECBIT_NOROOT_LOCKED
¤È¤¤¤¦Ì¾Á°¤Ç¤¢¤ë¡£
.PP
.\"O The
.\"O .I securebits
.\"O flags can be modified and retrieved using the
.\"O .BR prctl (2)
.\"O .B PR_SET_SECUREBITS
.\"O and
.\"O .B PR_GET_SECUREBITS
.\"O operations.
.\"O The
.\"O .B CAP_SETPCAP
.\"O capability is required to modify the flags.
.I securebits
¥Õ¥é¥°¤Ï¡¢
.BR prctl (2)
¤ÎÁàºî
.B PR_SET_SECUREBITS
¤ä
.B PR_GET_SECUREBITS
¤ò»È¤¦¤³¤È¤ÇÊѹ¹¤·¤¿¤ê¼èÆÀ¤·¤¿¤ê¤Ç¤­¤ë¡£
¥Õ¥é¥°¤òÊѹ¹¤¹¤ë¤Ë¤Ï
.B CAP_SETPCAP
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬É¬ÍפǤ¢¤ë¡£

.\"O The
.\"O .I securebits
.\"O flags are inherited by child processes.
.\"O During an
.\"O .BR execve (2),
.\"O all of the flags are preserved, except
.\"O .B SECURE_KEEP_CAPS
.\"O which is always cleared.
.I securebits
¥Õ¥é¥°¤Ï»Ò¥×¥í¥»¥¹¤Ë·Ñ¾µ¤µ¤ì¤ë¡£
.BR execve (2)
¤Ë¤ª¤¤¤Æ¤Ï¡¢
.B SECURE_KEEP_CAPS
¤¬¾ï¤Ë¥¯¥ê¥¢¤µ¤ì¤ë°Ê³°¤Ï¡¢Á´¤Æ¤Î¥Õ¥é¥°¤¬ÊÝ»ý¤µ¤ì¤ë¡£

.\"O An application can use the following call to lock itself,
.\"O and all of its descendants,
.\"O into an environment where the only way of gaining capabilities
.\"O is by executing a program with associated file capabilities:
¥¢¥×¥ê¥±¡¼¥·¥ç¥ó¤Ï¡¢°Ê²¼¤Î¸Æ¤Ó½Ð¤·¤ò¹Ô¤¦¤³¤È¤Ë¤è¤ê¡¢
¼«Ê¬¼«¿È¤ª¤è¤Ó»Ò¹¤È¤Ê¤ë¥×¥í¥»¥¹Á´¤Æ¤ËÂФ·¤Æ¡¢
ɬÍפʥե¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò»ý¤Ã¤¿¥×¥í¥°¥é¥à¤ò¼Â¹Ô¤·¤Ê¤¤¸Â¤ê¡¢
Âбþ¤¹¤ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò³ÍÆÀ¤Ç¤­¤Ê¤¤¤è¤¦¤Ê¾õ¶·¤ËÊĤ¸¤³¤á¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
.in +4n
.nf

prctl(PR_SET_SECUREBITS,
        SECBIT_KEEP_CAPS_LOCKED |
        SECBIT_NO_SETUID_FIXUP |
        SECBIT_NO_SETUID_FIXUP_LOCKED |
        SECBIT_NOROOT |
        SECBIT_NOROOT_LOCKED);
.fi
.in
.\"O .SH "CONFORMING TO"
.SH ½àµò
.PP
.\"O No standards govern capabilities, but the Linux capability implementation
.\"O is based on the withdrawn POSIX.1e draft standard; see
.\"O .IR http://wt.xpilot.org/publications/posix.1e/ .
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Ë´Ø¤¹¤ëɸ½à¤Ï¤Ê¤¤¤¬¡¢ Linux ¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ÏÇѰƤˤʤä¿
POSIX.1e Áð°Æ¤Ë´ð¤Å¤¤¤Æ¼ÂÁõ¤µ¤ì¤Æ¤¤¤ë¡£
.I http://wt.xpilot.org/publications/posix.1e/
¤ò»²¾È¡£
.\"O .SH NOTES
.SH Ãí°Õ
.\"O Since kernel 2.5.27, capabilities are an optional kernel component,
.\"O and can be enabled/disabled via the CONFIG_SECURITY_CAPABILITIES
.\"O kernel configuration option.
¥«¡¼¥Í¥ë 2.5.27 °Ê¹ß¡¢¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ÏÁªÂò¼°¤Î¥«¡¼¥Í¥ë¥³¥ó¥Ý¡¼¥Í¥ó¥È
¤È¤Ê¤Ã¤Æ¤ª¤ê¡¢¥«¡¼¥Í¥ëÀßÄꥪ¥×¥·¥ç¥ó CONFIG_SECURITY_CAPABILITIES
¤Ë¤è¤êÍ­¸ú/̵¸ú¤òÀÚ¤êÂؤ¨¤ë¤³¤È¤¬¤Ç¤­¤ë¡£

.\"O The
.\"O .I /proc/PID/task/TID/status
.\"O file can be used to view the capability sets of a thread.
.\"O The
.\"O .I /proc/PID/status
.\"O file shows the capability sets of a process's main thread.
.I /proc/PID/task/TID/status
¥Õ¥¡¥¤¥ë¤ò»È¤¦¤È¡¢¥¹¥ì¥Ã¥É¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤ò¸«¤ë¤³¤È¤¬¤Ç¤­¤ë¡£
.I /proc/PID/status
¥Õ¥¡¥¤¥ë¤Ë¤Ï¡¢¥×¥í¥»¥¹¤Î¥á¥¤¥ó¥¹¥ì¥Ã¥É¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¥»¥Ã¥È¤¬É½¼¨¤µ¤ì¤ë¡£

.\"O The
.\"O .I libcap
.\"O package provides a suite of routines for setting and
.\"O getting capabilities that is more comfortable and less likely
.\"O to change than the interface provided by
.\"O .BR capset (2)
.\"O and
.\"O .BR capget (2).
.I libcap
¥Ñ¥Ã¥±¡¼¥¸¤Ï¡¢¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤òÀßÄꡦ¼èÆÀ¤¹¤ë¤¿¤á¤Î
¥ë¡¼¥Á¥ó·²¤òÄ󶡤·¤Æ¤¤¤ë¡£¤³¤ì¤é¤Î¥¤¥ó¥¿¥Õ¥§¡¼¥¹¤Ï¡¢
.BR capset (2)
¤È
.BR capget (2)
¤¬Ä󶡤¹¤ë¥¤¥ó¥¿¡¼¥Õ¥§¡¼¥¹¤ÈÈæ¤Ù¤Æ¡¢¤è¤ê»È¤¤¤ä¤¹¤¯¡¢Êѹ¹¤µ¤ì¤ë²ÄǽÀ­¤¬¾¯¤Ê¤¤¡£
.\"O This package also provides the
.\"O .BR setcap (8)
.\"O and
.\"O .BR getcap (8)
.\"O programs.
.\"O It can be found at
.\"O .br
.\"O .IR http://www.kernel.org/pub/linux/libs/security/linux-privs .
¤³¤Î¥Ñ¥Ã¥±¡¼¥¸¤Ç¤Ï¡¢
.BR setcap (8),
.BR getcap (8)
¤È¤¤¤¦¥×¥í¥°¥é¥à¤âÄ󶡤µ¤ì¤Æ¤¤¤ë¡£
¥Ñ¥Ã¥±¡¼¥¸¤Ï
.I http://www.kernel.org/pub/linux/libs/security/linux-privs
¤ÇÆþ¼ê¤Ç¤­¤ë¡£

.\"O Before kernel 2.6.24, and since kernel 2.6.24 if
.\"O file capabilities are not enabled, a thread with the
.\"O .B CAP_SETPCAP
.\"O capability can manipulate the capabilities of threads other than itself.
.\"O However, this is only theoretically possible,
.\"O since no thread ever has
.\"O .BR CAP_SETPCAP
.\"O in either of these cases:
¥Ð¡¼¥¸¥ç¥ó 2.6.24 ¤è¤êÁ°¡¢¤ª¤è¤Ó¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬
Í­¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤Ê¤¤2.6.24 °Ê¹ß¤Î¥«¡¼¥Í¥ë¤Ç¤Ï¡¢
.B CAP_SETPCAP
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò»ý¤Ã¤¿¥¹¥ì¥Ã¥É¤Ï¼«Ê¬°Ê³°¤Î¥¹¥ì¥Ã¥É¤Î
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤òÁàºî¤Ç¤­¤ë¡£
¤·¤«¤·¤Ê¤¬¤é¡¢¤³¤ì¤ÏÍýÏÀŪ¤Ë²Äǽ¤È¤¤¤¦¤À¤±¤Ç¤¢¤ë¡£
°Ê²¼¤Î¤¤¤º¤ì¤«¤Î¾ì¹ç¤Ë¤ª¤¤¤Æ¤â¡¢¤É¤Î¥¹¥ì¥Ã¥É¤â
.BR CAP_SETPCAP
¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤ò»ý¤Ä¤³¤È¤Ï¤Ê¤¤¤«¤é¤Ç¤¢¤ë¡£
.IP * 2
.\"O In the pre-2.6.25 implementation the system-wide capability bounding set,
.\"O .IR /proc/sys/kernel/cap-bound ,
.\"O always masks out this capability, and this can not be changed
.\"O without modifying the kernel source and rebuilding.
2.6.25 ¤è¤êÁ°¤Î¼ÂÁõ¤Ç¤Ï¡¢¥·¥¹¥Æ¥à¶¦Ä̤Υ±¡¼¥Ñ¥Ó¥ê¥Æ¥£¡¦¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È
.I /proc/sys/kernel/cap-bound
¤Ç¤Ï¤³¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤Ï¾ï¤Ë̵¸ú¤Ë¤Ê¤Ã¤Æ¤ª¤ê¡¢
¥½¡¼¥¹¤òÊѹ¹¤·¤Æ¥«¡¼¥Í¥ë¤òºÆ¥³¥ó¥Ñ¥¤¥ë¤·¤Ê¤¤¸Â¤ê¡¢
¤³¤ì¤òÊѹ¹¤¹¤ë¤³¤È¤Ï¤Ç¤­¤Ê¤¤¡£
.IP *
.\"O If file capabilities are disabled in the current implementation, then
.\"O .B init
.\"O starts out with this capability removed from its per-process bounding
.\"O set, and that bounding set is inherited by all other processes
.\"O created on the system.
¸½ºß¤Î¼ÂÁõ¤Ç¤Ï¥Õ¥¡¥¤¥ë¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤¬Ìµ¸ú¤Ë¤Ê¤Ã¤Æ¤¤¤ë¾ì¹ç¡¢
¥×¥í¥»¥¹Ëè¤Î¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤«¤é¤³¤Î¥±¡¼¥Ñ¥Ó¥ê¥Æ¥£¤òÈ´¤¤¤Æ
.B init
¤Ï³«»Ï¤µ¤ì¡¢
¥·¥¹¥Æ¥à¾å¤ÇÀ¸À®¤µ¤ì¤ë¾¤ÎÁ´¤Æ¤Î¥×¥í¥»¥¹¤Ç¤³¤Î¥Ð¥¦¥ó¥Ç¥£¥ó¥°¥»¥Ã¥È¤¬
·Ñ¾µ¤µ¤ì¤ë¡£
.\"O .SH "SEE ALSO"
.SH ´ØÏ¢¹àÌÜ
.BR capget (2),
.BR prctl (2),
.BR setfsuid (2),
.BR cap_clear (3),
.BR cap_copy_ext (3),
.BR cap_from_text (3),
.BR cap_get_file (3),
.BR cap_get_proc (3),
.BR cap_init (3),
.BR capgetp (3),
.BR capsetp (3),
.BR credentials (7),
.BR pthreads (7),
.BR getcap (8),
.BR setcap (8)
.PP
.\"O .I include/linux/capability.h
.\"O in the kernel source
¥«¡¼¥Í¥ë¥½¡¼¥¹Æâ¤Î
.I include/linux/capability.h