In recent times, a number of other namespace flags have been
added to clone(2). As such, it is no longer clear to use
the generic term "namespace" to refer to the particular
namespace controlled by CLONE_NEWNS; instead, use the
term "mount-point namespace".
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
I submitted a patch to fix this. See the LKML thread
"[patch] Fix type errors in inotify interfaces", 18 Nov 2008
If/when these patches are accepted, the pages need to be updated.
It seems that inotify(7) is wrong here:
"/proc/sys/fs/inotify/max_user_watches
This specifies a limit on the number of watches that can be
associated with each inotify instance."
On my system, the default value for this variable is 8192. But I
cannot create more than 8192 watches in total for the same UID
even when they are on different inotify instances. So I suggest
to rephrase this as: "This specifies an upper limit on the
number of watches that can be created per real user ID."
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
signal(7) provides some further details on the use of real-time
signals by the two Linux threading implementations.
Signed-off-by: Michael Kerrisk <michael.kerrisk@gmail.com>
Integrate the changes that occurred in POSIX.1-2008 into the
main list (to be consistent with the list, elsewhere on this
page, of functions that are cancellation points).
Also, fix an error that said that strerror() was added to
the list in POSIX.1-2008. It was strsignal() that was
added. (strerror() was already in the list in POSIX.1-2001.)
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
In particular, note that in each pthreads function that takes
a thread ID argument, that ID by definition refers to a thread
in the same process as the caller.
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Simply continuing after an error is in most cases wrong,
and can lead to infinite loops (e.g., for EMFILE).
So handle an error by terminating.
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=504202
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Reported-by: Olaf van der Spek <olafvdspek@gmail.com>
Fill in some gaps in example code (variable declarations,
adding listening socket to epoll set).
Give variables more meaningful names.
Other minor changes.
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=504202
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Reported-by: Olaf van der Spek <OlafvdSpek@gmail.com>
Lines for these two characters were added in the previous patch,
but the actual characters were not included in the 4th column
of the table. This fixes that.
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Describe the usual success (0) and failure (non-zero) returns,
and note that POSIX.1-2001 specifies that pthreads functions
can never fail with the error EINTR.
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
If setsockopt() is used to set a timeout on a socket(),
then the various socket interfaces are not automatically
restarted, even if SA_RESTART is specified when
establishing the signal handler. Analogous behavior occurs
for the "stop signals" case.
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Reported-by: Didier <did447@gmail.com>
Include text describing semantics of fork() and execve() for
signal dispositions, signal mask, and pending signal set.
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>