GitLab

Arch does not use proper SysV runlevels. Instead /etc/rc.conf contains
an array of daemons to load in a multi-user runlevel, and some
dependencies between them.

This patch enables this behaviour in SystemD (only when compiled with
Arch support), which allows SystemD to be a drop in replacement on Arch
(modulo some minor bugs).

v2: simplified after suggestion from Lennart

This patch introduces socket_ipv6_is_supported() call that checks for
IPv6 availability. Code then check for it before using specific calls.

In order to be less intrusive, this patch avoids IPv6 entries being
parsed at all, this way we don't get such entries in the system and
all other code paths are automatically ignored. However an extra check
is done at socket_address_listen() to make sure of that.

As the number of Netlink messages is not know upfront anymore,
loopback-setup.c was refactored to dynamically calculate the sequence
number and count.

Lennart's suggestions were fixed and squashed with the original patch,
that was sent by Gustavo Sverzut Barbieri (barbieri@profusion.mobi).

Use locale variables from /etc/profile.env as Gentoo's openrc
runscript does (tool that is used by /etc/init.d/* services).

This patch is a bit bigger than expected since Gentoo being
non-standard in some places.

 1. it is installing binaries at /usr/bin instead of /bin.

 2. it is using CamelCase names for consolefonts.

 3. /etc/rc.conf:unicode=(yes|no) just forbids loadkeys and setfont
    "-u" options, but do not disable the actual kernel default_utf8
    from vt module.

Garbage value could have been used if there was an error creating the
password file.

"condrestart" should be equivalent to "try-restart", i.e. it should be
translated into "TryRestartUnit". See
http://fedoraproject.org/wiki/Packaging:SysVInitScript#condrestart_and_try-restart

Fixes https://bugzilla.redhat.com/show_bug.cgi?id=635780

This variable is defined in /etc/rc.conf

These variables are defined in /etc/rc.conf

it may happen that parent directories of RANDOM_SEED do not exist, in
that case we must create it otherwise save will not work.

On Tue, Sep 14, 2010 at 10:23:28PM +0200, Lennart Poettering wrote:
> On Tue, 07.09.10 17:02, Zbigniew Jędrzejewski-Szmek (zbyszek@in.waw.pl) wrote:
>
> Urks, now I commited the wrong one... Can you prep an updated patch against
> current git?
Once more sorry about that.

Fix below.

Zbyszek

https://bugzilla.redhat.com/show_bug.cgi?id=630156

rhbz 629268

old
UNIT                      LOAD   ACTIVE       SUB                JOB
dev-hugepages.automount   loaded active       running                            Huge Page
dev-mqueue.automount      loaded active       waiting                            POSIX Mes
proc-sys...misc.automount loaded active       waiting                            Arbitrary
sys-kern...ebug.automount loaded active       running                            Debug Fil

new
UNIT                      LOAD   ACTIVE SUB       JOB
dev-hugepages.automount   loaded active running       Huge Pages File System Automount Poi
dev-mqueue.automount      loaded active waiting       POSIX Message Queue File System Auto
systemd-shutdownd.socket  loaded active listening     systemd Delayed Shutdown Socket