We protect most of the API from use accross forks, but we still allow both
sd_event and sd_event_source objects to be unref'ed. This would cause
problems as it would unregister sources from the underlying eventfd, hence
also affecting the original instance in the parent process.
This fixes the issue by not touching the fds on unref when done accross a fork,
but still free the memory.
This fixes a regression introduced by
"udevd: move main-loop to sd-event": 693d371d30
where the worker processes were disabling the inotify event source in the
main daemon.
Skip /proc/mountinfo entries for which libmount returns a NULL pointer
for 'source' or 'target'. This happened on Semaphore CI's build servers
when the test suite is run.
This way, development builds will not rely on gc-sections to
paper over cyclic link dependencies. Newly introduced broken
link requirements will immediatley fail.
Since we introduced AX_NORMALIZE_PATH, using --with-rootprefix=/ does
produce an empty string, but using --with-rootprefix= (empty) now
produces "." instead which is wrong.
Work around it until we can find a better solution for AX_NORMALIZE_PATH
upstream at autoconf-archive.
Bug: https://github.com/systemd/systemd/issues/54
Since we started using AX_NORMALIZE_PATH, that is a valid supported
setup and is more explicit than --with-rootprefix= (empty) which is
actually currently broken.
Let's advocate for it in the ./configure suggestion from autogen.sh.
Stop to pretend that we can split selinux related code from other.
We have too many cross-references and it breaks all the time and
I am no longer willing to maintain that mess for no real benefit.
We currently have cyclic dependencies which are only resolved on
machines with gc-sections toolchains. We need a simpler and at the
same time more strict model to manage our convenienc libraries and
linking.
The first thing to give up is the "optimization" of not linking
libselinux for a very few tools. If that is an issue, please fix
the mess that libselinux creates in selinux itself, and do not ask
consumers to work around it.
If the EDITOR environment variable is not set, the Debian policy
recommends to use the /usr/bin/editor program as default editor.
This file is managed via the dpkg alternatives mechanism and typically
used in Debian/Ubuntu and derivatives to configure the default editor.
See section 11.4 of the Debian policy [1].
Therefor prefer /usr/bin/editor over specific editors if available.
[1] https://www.debian.org/doc/debian-policy/ch-customized-programs.html
Fix a regression caused by 4034a06d ("util: rework word parsing and c
unescaping code") which broke octal escape sequences.
The reason for this breakage is that cunescape_one() expects 4 characters
in an octal encoding, which is a stray left-over from the old code which
operated on different variables to make the length check.
While at it, add a test case to prevent the same thing from happening
again.
Now that listen_fds() have been split out, we can safely move the allocation
of the manager object after doing the forking (the fork is done to notify legcay
init-systems that the fds are ready).
Subsequently, we can merge manager_listen() back into managre_new().
This entails a minor behaviour change: the application of permissions to
static device nodes now happens after the fork (but still before notifying
systemd about being ready).
This will simply silently fail on non-systemd systems, so there is no reason
to make it conditional.
Also make it clear that we notify systemd about being ready as the last step
before starting the event loop, whereas the forking might need to happen
earlier.
This allows a fd to be created and configured as part of one monitor, to be passed in
to create a second monitor without having to redo any of the configuration.
This should have no behavioural change, but it is odd to tie the cgroup cleaning to
whether or not we are passed sockets.
The point really is if we are guaranteed to be in a dedicated cgroup, so instead
check for our parent being PID1 (we already implicitly only do this on systemd
systems).
We used to block all signals, and restore the original signal mask before exec'ing
external processes.
Now we just block the signals we care about and unconditionally unblock all signals
before exec'ing.
In 5a8bcb674f, IPForwarding was introduced
to set forwarding flags on interfaces in .network files. networkd sets
forwarding options regardless of the previous setting, even if it was
set by e.g. sysctl. This commit creates a new option for IPForwarding,
"kernel", that preserves the sysctl settings rather than always setting
them.
See https://bugs.freedesktop.org/show_bug.cgi?id=89509 for the initial
bug report.
David Herrmann