Martin Pitt [Tue, 17 Feb 2015 06:50:22 +0000 (07:50 +0100)]
New upstream release 219
- Fix spelling mistake in systemd.unit(5). (Closes: #773302)
- Fix timeouts with D-Bus, leading to SIGFPE. (Closes: #774012)
- Fix load/save of multiple rfkill states. (Closes: #759489)
- Non-persistant journal (/run/log/journal) is now readable by group adm.
(Closes: #771980)
- Read netdev user mount option to correctly order network mounts after
network.target. (Closes: #769186)
- Fix 60-keyboard.hwdb documentation and whitespace handling.
(Closes: #757367)
- Fix ThinkPad X1 Carbon 20BT trackpad buttons (LP: #1414930)
Drop all backported patches and port the others to new upstream release.
Martin Pitt [Mon, 16 Feb 2015 13:49:32 +0000 (14:49 +0100)]
Clean up leftover symlinks on purge
debian/systemd.postrm: Clean getty@tty1.service and remote-fs.target enablement
symlinks on purge. These are created in the postinst with systemctl enable.
Martin Pitt [Fri, 13 Feb 2015 10:11:27 +0000 (11:11 +0100)]
rules: simplify mmc RPMB handling
Simplifiy commit 82c7a9 following upstream commit 0c13be38.
We don't actually want a by-path/ symlink for MMC RPMB devices, so just add
them to the blacklist. This will prevent creating wrong by-path links and
blkid'ing those.
This is not a good long-term approach to fix this bug, only as a workaround for
a stable release update. See discussion in https://launchpad.net/bugs/1420544,
as long as that didn't come to a conclusion let's not ship this.
Martin Pitt [Tue, 10 Feb 2015 09:29:30 +0000 (10:29 +0100)]
Fix debian/tests/logind test under sysvinit/upstart
With dropped systemd-logind-launch we don't have a visible
/sys/fs/cgroup/systemd/ any more under cgmanager. So adjust the test to check
/proc/self/cgroup instead.
Martin Pitt [Mon, 9 Feb 2015 09:40:40 +0000 (10:40 +0100)]
update-rc.d patch cleanup
Merge Run-update-rc.d-defaults-before-update-rc.d-enable-d.patch into
Make-systemctl-enable-disable-call-update-rc.d-for-s.patch as the former is
fixing the latter and is not an independent change.
Martin Pitt [Thu, 5 Feb 2015 09:26:58 +0000 (10:26 +0100)]
/lib/lsb/init-functions.d/40-systemd: Avoid deadlocks during bootup and shutdown
DHCP/ifupdown and similar hooks which call "/etc/init.d/foo reload" can easily
cause deadlocks, since the synchronous wait plus systemd's normal behaviour of
transactionally processing all dependencies first easily causes dependency
loops. Thus during boot/shutdown operate only on the unit and not on its
dependencies, just like SysV behaves.
For now we want this for shutdown order, to bring down ifup@.service after
network.target. Until we properly integrate ifupdown into
network-online.target, network-online.target and network.target are effectively
the same (on ifupdown only systems), so let's ignore the unnecessary boot
dependencies for now.
Martin Pitt [Tue, 3 Feb 2015 12:45:25 +0000 (13:45 +0100)]
SysV generator patch cleanup
Merge two patches into Add-support-for-rcS.d-init-scripts-to-the-sysv-gener.patch:
- Do-not-order-rcS.d-services-after-local-fs.target-if.patch, as it
partially reverts the above, and is just fixing it.
- Map-rcS.d-init-script-dependencies-to-their-systemd-.patch as it's just
adding some missing functionality for the same purpose.
Martin Pitt [Tue, 3 Feb 2015 07:36:47 +0000 (08:36 +0100)]
ifup@.service: Drop Before=network.target ordering
network.target is an early-boot target which runs before bringing up any
interface,, and this unit can happen at any time on a hotplug event. This
avoids dependency loops during boot.
Martin Pitt [Sun, 1 Feb 2015 17:41:13 +0000 (18:41 +0100)]
ifup@.service: Don't fail if the interface is not configured in /etc/network/interfaces at all
We first need to ask ifquery if it knows about the interface in the first
place; only then check its --state. ifup@.service is also being triggered by
dynamic interface bringup from networkd/NetworkManager.
Michael Biebl [Thu, 29 Jan 2015 08:58:07 +0000 (09:58 +0100)]
Map $x-display-manager LSB facility to display-manager.service
...instead of making it a target. Using a target had the downside that
multiple display managers could hook into it at the same time which could
lead to several failed start attempts for the non-default display
manager.
Martin Pitt [Mon, 26 Jan 2015 12:03:38 +0000 (13:03 +0100)]
ifup@.service: Lessen dependencies
Drop dependency on networking.service (i. e. /etc/init.d/networking), and
merely ensure that /run/network exists. This avoids unnecessary
dependencies/waiting during boot, and dependency cycles if hooks wait for other
interfaces to come up (like ifenslave with bonding interfaces).
Martin Pitt [Mon, 26 Jan 2015 11:03:04 +0000 (12:03 +0100)]
Generate POT file during package build
This eases translator's work and enables automatic import into Launchpad
translations. It's not totally obvious how to build it as a simple
"make -C build-deb/po systemd.pot" does not work due to upstream commit 6fc00209d5.
Michael Biebl [Sat, 24 Jan 2015 22:42:52 +0000 (23:42 +0100)]
Remove leftover bits from sd-readahead
The readahead functionality was removed in v217. Remove leftover bits
from ca3e157bc2302c6ea46abe12e0553f6b10c1c2d4 which was installing the
man pages as example files.
Martin Pitt [Thu, 22 Jan 2015 13:49:33 +0000 (14:49 +0100)]
Enable EFI support
This mostly auto-mounts /sys/firmware/efi/efivars, but also provides a
generator for auto-detecting the root and the /boot/efi partition if they
aren't in /etc/fstab.