seteuid.2, setresuid.2, setreuid.2, setuid.2: spfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Add notes on F_SETLKW deadlock detection and its limitations

Reported-by: Jeff Layton <jlayton@poochiereds.net>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: wfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: The kernel does not perform deadlock detection for OFD locks

See kernel commit 57b65325fe34ec4c917bc4e555144b4a94d9e1f7
And this mail thread:

    http://thread.gmane.org/gmane.linux.file-systems/81318/focus=81327
    From: Jeff Layton <jlayton <at> redhat.com>
    Subject: [PATCH v5 13/14] locks: skip deadlock detection on FL_FILE_PVT locks
    Date: 2014-01-09 14:19:46 GMT

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

open.2: Rework and extend the discussion of open file descriptions

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

lseek.2: Add reference to open(2) for discussion of file descriptors and OFDs

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Under OFD locks add a reference to open(2) for explanation of OFDs

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fork.2: Add notes on inheritance of flock() and OFD locks across fork()

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Explain POSIX background to EACCES/EAGAIN error for F_GETLK

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: F_OFD_GETLK fails with EAGAIN when there is a lock conflict

For F_GETLK, POSIX allows EACCES or EAGAIN because different
implementations of traditional record locks returned one of
these errors, and portable applications had to handle either
possibility. However, that argument doesn't apply for OFD
locks, since Linux is the only implementation and it returns
EAGAIN.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Describe how to check whether the kernel supports a particular command

Reported-by: Jeff Layton <jlayton@poochiereds.net>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: ERRORS: add EINVAL for invalid 'cmd'

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Minor wording improvement

Reported-by: Jeff Layton <jlayton@poochiereds.net>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Add an explicit note that mandatory locking is not in POSIX

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Make the warning that mandatory locks are unreliable more prominent

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: wfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Rewrite introductory paragraphs on mandatory locking

Make the structure more logical, and also explicitly mention
OFD locks.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Add EINVAL error for OFD locks where 'l_pid' is not 0

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: CONFORMING TO: note that OFD locks are Linux-specific

But they make their way into the next POSIX release

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Describe semantics for compatible and conflicting OFD locks

And describe how threads can use OFD locks to ensure
exclusive access to a file.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Detail the limitations of traditional (process-associated) locks

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Minor fixes to introductory par on OFD locks

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: OFD locks are Linux-specific and available since Linux 3.15

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Tweaks and rewrites of pieces of Jeff Layton's patch

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Document open file description locks

As provided by the fcntl() operations F_OFD_SETLK,
F_OFD_SETLKW, and F_OFD_GETLK

Signed-off-by: Jeff Layton <jlayton@poochiereds.net>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Add para introducing advisory locks and noting existence of OFD locks

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Minor change: move some NOTES text on record locking

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: wfix + ffix: fine-tuning on NFSv4 client locking text

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Add NOTES subhead for record locking and NFS

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: nfsv4leasetime controls the "contact lost" interval for NFSv4

Jeff Layton:
   The difference here is subtle. The gracetime is how long after a reboot
   should knfsd allow clients to reclaim state (and deny the creation of
   new locks and opens). The leasetime is how long the NFSv4 lease period
   is. There is a relationship between the two that's illustrated in the
   comments above write_gracetime:

     /**
      * write_gracetime - Set or report current NFSv4 grace period time
      *
      * As above, but sets the time of the NFSv4 grace period.
      *
      * Note this should never be set to less than the *previous*
      * lease-period time, but we don't try to enforce this.  (In the common
      * case (a new boot), we don't know what the previous lease time was
      * anyway.)
      */

   The value you're interested in here is the nfsv4leasetime. If the
   client doesn't renew its lease within that period, then it's subject to
   the server giving up on it and dropping any state that it holds on that
   clients' behalf.

   Note that this is not a firm timeout. The server runs a job
   periodically to clean out expired stateful objects, and it's likely
   that there is some time (maybe even up to another whole lease period)
   between when the timeout expires and the job actually runs. If the
   client gets a RENEW in there within that window, its lease will be
   renewed and its state preserved.

Reported-by: Jeff Layton <jlayton@poochiereds.net>
Cowritten-by: Jeff Layton <jlayton@poochiereds.net>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Refine discussion of locks when NFSv4 client loses contact with server

Reported-by: NeilBrown <neilb@suse.de>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Note treatment of locks when an NFS client loses contact with the server

Based on text sent by Neil Brown.

Reported-by: NeilBrown <neilb@suse.de>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Minor rewording

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: srcfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fcntl.2: Reword discussion of mandatory lock bug a litte

Jeff Layton confirmed that the bug remains even in modern kernels.

Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

flock.2: In some modern BSDs, fcntl() and flock() locks do interact

So, reword and extend the discussion of this topic in NOTES.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

flock.2: Move NOTES text describing implementation of flock()

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

flock.2: Add more details on NFS, including Linux 2.6.37 changes

Also: move NOTES text describing interaction of fcntl()
and flock() locks.

Reviewed-by: NeilBrown <neilb@suse.de>
Reported-by: NeilBrown <neilb@suse.de>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

mkdtemp.3: ATTRIBUTES: Note function that is thread-safe

The function mkdtemp() is thread safe.

Signed-off-by: Peng Haitao <penght@cn.fujitsu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

basename.3: ATTRIBUTES: Note functions that are thread-safe

The functions basename() and dirname() are thread safe.

Signed-off-by: Peng Haitao <penght@cn.fujitsu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

system.3: ATTRIBUTES: Note function that is thread-safe

The function system() is thread safe.

Signed-off-by: Peng Haitao <penght@cn.fujitsu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

catgets.3: wfix

Reported-by: Peng Haitao <penght@cn.fujitsu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

catgets.3: ATTRIBUTES: Note function that is thread-safe

The function catgets() is thread safe.

Signed-off-by: Peng Haitao <penght@cn.fujitsu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

environ.7: SEE ALSO: add env(1), printenv(1), ld.so(8)

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

fanotify_mark.2: tfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

send.2: tfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

charsets.7: Drop unneeded text on UTF-8 for Biblical Hebrew

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

localedef.1: Align with recent charmap(5) / repertoiremap(5) changes

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

charmap.5: Update to match current glibc

charmap(5) was outdated, bring it to closer to reality by fixing
syntax descriptions to match current glibc code and practices,
adding missing options, removing obsolete comments and references,
and removing now incorrect examples.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

repertoiremap.5: New page for repertoiremap(5)

Rather obsolete feature but localedef(1) refers to repertoiremaps.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

charsets.7: Minor tweaks

And restore a piece about Biblical Hebrew that was
inadvertently deleted by Marko Myllynen's patch.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

charsets.7: Update to reflect past developments

Rewrite the introduction to make Unicode's prominence more obvious.
Reformulate parts of the text to reflect current Unicode world.
Minor clarification for ASCII/ISO sections, some other minor fixes.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

mq_send.3: ATTRIBUTES: Note functions that are thread-safe

The functions mq_send() and mq_timedsend() are thread safe.

Signed-off-by: Peng Haitao <penght@cn.fujitsu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

mq_receive.3: ATTRIBUTES: Note functions that are thread-safe

The functions mq_receive() and mq_timedreceive() are thread safe.

Signed-off-by: Peng Haitao <penght@cn.fujitsu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

mq_open.3: ATTRIBUTES: Note function that is thread-safe

The function mq_open() is thread safe.

Signed-off-by: Peng Haitao <penght@cn.fujitsu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

posix_openpt.3: ATTRIBUTES: Note function that is thread-safe

The function posix_openpt() is thread safe.

Signed-off-by: Peng Haitao <penght@cn.fujitsu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

posix_fallocate.3: ATTRIBUTES: Note function that is thread-safe

The function posix_fallocate() is thread safe.

Signed-off-by: Peng Haitao <penght@cn.fujitsu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

locale.5: Clarify AM/PM settings a bit

localedef(1) complains if really undefined, should be empty instead.
Also: add some SEE ALSO references.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

locale.1: Add FILES section, add charmap(5) reference

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

locale.7: Add some SEE ALSO references

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

malloc_get_state.3: SYNOPSIS: use correct header

The nonstandard functions malloc_set_state() and
malloc_get_state() are provided by <malloc.h> not <stdlib.h>.

Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

sendmmsg.2: Add reference to feature_test_macros(7)

Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

recvmmsg.2: Add reference to feature_test_macros(7)

Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

open_by_handle_at.2: Add reference to feature_test_macros(7)

Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

getdate.3: Use blank definition of _GNU_SOURCE

Only the definedness of _GNU_SOURCE matters, so eliminate a possible
source of confusion by not using a specific replacement text in the
example program.

Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

proc.5: srcfix: Remove obsolete FIXMEs

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

Changes.old: Fix to 3.41 changelog

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

bootparam.7: Minor tweaks to Petr Gajdos's patch

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

bootparam.7: Describe 'rootflags' and 'rootdelay' kernel parameters

Patch based on text from Documentation/kernel-parameters.txt

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

man-pages.7: Note that .TH 'date' field is nowadays automatically updated by scripts

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

signal.7: Clarify text describing EINTR semantics for socket interfaces

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

signal.7: srcfix: Add FIXMEs

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

signal.7: Describe EINTR semantics for recvmmsg(2)

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

locale.5: Minor NOTES clarification

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

execve.2: RLIMIT_NPROC is checked only if preceded by set*uid()

Reported-by: Vasiliy Kulikov <segoon@openwall.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

execve.2: Before kernel 2.6.0, RLIMIT_NPROC had no effect for set*uid()

Reported-by: NeilBrown <neilb@suse.de>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

execve.2: Document EAGAIN error

See also https://bugzilla.kernel.org/show_bug.cgi?id=42704

Reviewed-by: Vasiliy Kulikov <segoon@openwall.com>
Reviewed-by: NeilBrown <neilb@suse.de>
Reported-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

setreuid.2: Error checking should always be performed, even when caller is UID 0

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

setresuid.2: Error checking should always be performed, even when caller is UID 0

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

seteuid.2: Error checking should always be performed, even when caller is UID 0

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

setuid.2: Error checking should always be performed, even when caller is UID 0

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

setuid.2: Since Linux 3.1, the EAGAIN case for RLIMIT_NPROC no longer occurs

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

setreuid.2: Since Linux 3.1, the EAGAIN case for RLIMIT_NPROC no longer occurs

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

setresuid.2: Since Linux 3.1, the EAGAIN case for RLIMIT_NPROC no longer occurs

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

setuid.2: Document EAGAIN error that can occur after kernel alloc_uid() failure

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

setreuid.2: Document EAGAIN error that can occur after kernel alloc_uid() failure

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

setresuid.2: Document EAGAIN error that can occur after kernel alloc_uid() failure

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

setreuid.2: Add EAGAIN error for hitting RLIMIT_NPROC limit

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

setuid.2: Correct the description of the EAGAIN error

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

setresuid.2: Correct the description of the EAGAIN error

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

seteuid.2: seteuid() and setegid() are implemented as library functions

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

credentials.7: Minor layout fixes to make list format more consistent

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

locale.5: ffix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

locale.5: Drop a garbled sentence

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

locale.5: srcfix

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

locale.5: Document glibc conventions regarding days and week

Based on existing practice and glibc community wiki page at
https://sourceware.org/glibc/wiki/Locales

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

Changes.old: tfix in 3.68 changelog

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

Changes.old: Fix to 3.68 changelog

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

Changes.old: Fixes to 3.68 changelog

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>

Start of man-pages-3.69: updating Changes and Changes.old