1 systemd System and Service Manager
5 * Timer units now support calendar time events in addition to
6 monotonic time events. That means you can now trigger a unit
7 based on a calendar time specification such as "Thu,Fri
8 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
9 or fifth day of any month of the year 2013, given that it is
10 a thursday or friday. This brings timer event support
11 considerably closer to cron's capabilities. For details on
12 the supported calendar time specification language see
15 * udev now supports a number of different naming policies for
16 network interfaces for predictable names, and a combination
17 of these policies is now the default. Please see this wiki
20 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
22 * Auke Kok's bootchart implementation has been added to the
23 systemd tree. It's an optional component that can graph the
24 boot in quite some detail. It's one of the best bootchart
25 implementations around and minimal in its code and
28 * nss-myhostname has been integrated into the systemd source
29 tree. nss-myhostname guarantees that the local hostname
30 always stays resolvable via NSS. It has been a weak
31 requirement of systemd-hostnamed since a long time, and
32 since its code is actually trivial we decided to just
33 include it in systemd's source tree. It can be turned off
34 with a configure switch.
36 * The read-ahead logic is now capable of properly detecting
37 whether a btrfs file system is on SSD or rotating media, in
38 order to optimize the read-ahead scheme. Previously, it was
39 only capable of detecting this on traditional file systems
42 * In udev, additional device properties are now read from the
43 IAB in addition to the OUI database. Also, Bluetooth company
44 identities are attached to the devices as well.
46 * In service files %U may be used as specifier that is
47 replaced by the configured user name of the service.
49 * nspawn may now be invoked without a controlling TTY. This
50 makes it suitable for invocation as its own service. This
51 may be used to set up a simple containerized server system
52 using only core OS tools.
54 * systemd and nspawn can now accept socket file descriptors
55 when they are started for socket activation. This enables
56 implementation of socket activated nspawn
57 containers. i.e. think about autospawning an entire OS image
58 when the first SSH or HTTP connection is received. We expect
59 that similar functionality will also be added to libvirt-lxc
62 * journalctl will now suppress ANSI color codes when
65 * systemctl will no longer show control group information for
66 a unit if a the control group is empty anyway.
68 * logind can now automatically suspend/hibernate/shutdown the
71 * /etc/machine-info and hostnamed now also expose the chassis
72 type of the system. This can be used to determine whether
73 the local system is a laptop, desktop, handset or
74 tablet. This information may either be configured by the
75 user/vendor or is automatically determined from ACPI and DMI
76 information if possible.
78 * A number of PolicyKit actions are now bound together with
79 "imply" rules. This should simplify creating UIs because
80 many actions will now authenticate similar ones as well.
82 * Unit files learnt a new condition ConditionACPower= which
83 may be used to conditionalize a unit depending on whether an
84 AC power source is connected or not, of whether the system
85 is running on battery power.
87 * systemctl gained a new "is-failed" verb that may be used in
88 shell scripts and suchlike to check whether a specific unit
89 is in the "failed" state.
91 * The EnvironmentFile= setting in unit files now supports file
92 globbing, and can hence be used to easily read a number of
93 environment files at once.
95 * systemd will no longer detect and recognize specific
96 distributions. All distribution-specific #ifdeffery has been
97 removed, systemd is now fully generic and
98 distribution-agnostic. Effectively, not too much is lost as
99 a lot of the code is still accessible via explicit configure
100 switches. However, support for some distribution specific
101 legacy configuration file formats has been dropped. We
102 recommend distributions to simply adopt the configuration
103 files everybody else uses now and convert the old
104 configuration from packaging scripts. Most distributions
105 already did that. If that's not possible or desirable,
106 distributions are welcome to forward port the specific
107 pieces of code locally from the git history.
109 * When logging a message about a unit systemd will now always
110 log the unit name in the message meta data.
112 * localectl will now also discover system locale data that is
113 not stored in locale archives, but directly unpacked.
115 * logind will no longer unconditionally use framebuffer
116 devices as seat masters, i.e. as devices that are required
117 to be existing before a seat is considered preset. Instead,
118 it will now look for all devices that are tagged as
119 "seat-master" in udev. By default framebuffer devices will
120 be marked as such, but depending on local systems other
121 devices might be marked as well. This may be used to
122 integrate graphics cards using closed source drivers (such
123 as NVidia ones) more nicely into logind. Note however, that
124 we recommend using the open source NVidia drivers instead,
125 and no udev rules for the closed-source drivers will be
126 shipped from us upstream.
128 Contributions from: Adam Williamson, Alessandro Crismani, Auke
129 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
130 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
131 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
132 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
133 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
134 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
135 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
136 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
137 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
138 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
139 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
144 * udev gained support for loading additional device properties
145 from an indexed database that is keyed by vendor/product IDs
146 and similar device identifiers. For the beginning this
147 "hwdb" is populated with data from the well-known PCI and
148 USB database, but also includes PNP, ACPI and OID data. In
149 the longer run this indexed database shall grow into
150 becoming the one central database for non-essential
151 userspace device metadata. Previously, data from the PCI/USB
152 database was only attached to select devices, since the
153 lookup was a relatively expensive operation due to O(n) time
154 complexity (with n being the number of entries in the
155 database). Since this is now O(1), we decided to add in this
156 data for all devices where this is available, by
157 default. Note that the indexed database needs to be rebuilt
158 when new data files are installed. To achieve this you need
159 to update your packaging scripts to invoke "udevadm hwdb
160 --update" after installation of hwdb data files. For
161 RPM-based distributions we introduced the new
162 %udev_hwdb_update macro for this purpose.
164 * The Journal gained support for the "Message Catalog", an
165 indexed database to link up additional information with
166 journal entries. For further details please check:
168 http://www.freedesktop.org/wiki/Software/systemd/catalog
170 The indexed message catalog database also needs to be
171 rebuilt after installation of message catalog files. Use
172 "journalctl --update-catalog" for this. For RPM-based
173 distributions we introduced the %journal_catalog_update
174 macro for this purpose.
176 * The Python Journal bindings gained support for the standard
177 Python logging framework.
179 * The Journal API gained new functions for checking whether
180 the underlying file system of a journal file is capable of
181 properly reporting file change notifications, or whether
182 applications that want to reflect journal changes "live"
183 need to recheck journal files continously in appropriate
186 * It is now possible to set the "age" field for tmpfiles
187 entries to 0, indicating that files matching this entry
188 shall always be removed when the directories are cleaned up.
190 * coredumpctl gained a new "gdb" verb which invokes gdb
191 right-away on the selected coredump.
193 * There's now support for "hybrid sleep" on kernels that
194 support this, in addition to "suspend" and "hibernate". Use
195 "systemctl hybrid-sleep" to make use of this.
197 * logind's HandleSuspendKey= setting (and related settings)
198 now gained support for a new "lock" setting to simply
199 request the screen lock on all local sessions, instead of
200 actually executing a suspend or hibernation.
202 * systemd will now mount the EFI variables file system by
205 * Socket units now gained support for configuration of the
206 SMACK security label.
208 * timedatectl will now output the time of the last and next
209 daylight saving change.
211 * We dropped support for various legacy and distro-specific
212 concepts, such as insserv, early-boot SysV services
213 (i.e. those for non-standard runlevels such as 'b' or 'S')
214 or ArchLinux /etc/rc.conf support. We recommend the
215 distributions who still need support this to either continue
216 to maintain the necessary patches downstream, or find a
217 different solution. (Talk to us if you have questions!)
219 * Various systemd components will now bypass PolicyKit checks
220 for root and otherwise handle properly if PolicyKit is not
221 found to be around. This should fix most issues for
222 PolicyKit-less systems. Quite frankly this should have been
223 this way since day one. It is absolutely our intention to
224 make systemd work fine on PolicyKit-less systems, and we
225 consider it a bug if something doesn't work as it should if
226 PolicyKit is not around.
228 * For embedded systems it is now possible to build udev and
229 systemd without blkid and/or kmod support.
231 * "systemctl switch-root" is now capable of switching root
232 more than once. I.e. in addition to transitions from the
233 initrd to the host OS it is now possible to transition to
234 further OS images from the host. This is useful to implement
235 offline updating tools.
237 * Various other additions have been made to the RPM macros
238 shipped with systemd. Use %udev_rules_update() after
239 installing new udev rules files. %_udevhwdbdir,
240 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
241 %_sysctldir are now available which resolve to the right
242 directories for packages to place various data files in.
244 * journalctl gained the new --full switch (in addition to
245 --all, to disable ellipsation for long messages.
247 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
248 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
249 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
250 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
251 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
252 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
253 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
254 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
255 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
259 * journalctl gained new --since= and --until= switches to
260 filter by time. It also now supports nice filtering for
261 units via --unit=/-u.
263 * Type=oneshot services may use ExecReload= and do the
266 * The journal daemon now supports time-based rotation and
267 vacuuming, in addition to the usual disk-space based
270 * The journal will now index the available field values for
271 each field name. This enables clients to show pretty drop
272 downs of available match values when filtering. The bash
273 completion of journalctl has been updated
274 accordingly. journalctl gained a new switch -F to list all
275 values a certain field takes in the journal database.
277 * More service events are now written as structured messages
278 to the journal, and made recognizable via message IDs.
280 * The timedated, localed and hostnamed mini-services which
281 previously only provided support for changing time, locale
282 and hostname settings from graphical DEs such as GNOME now
283 also have a minimal (but very useful) text-based client
284 utility each. This is probably the nicest way to changing
285 these settings from the command line now, especially since
286 it lists available options and is fully integrated with bash
289 * There's now a new tool "systemd-coredumpctl" to list and
290 extract coredumps from the journal.
292 * We now install a README each in /var/log/ and
293 /etc/rc.d/init.d explaining where the system logs and init
294 scripts went. This hopefully should help folks who go to
295 that dirs and look into the otherwise now empty void and
298 * When user-services are invoked (by systemd --user) the
299 $MANAGERPID env var is set to the PID of systemd.
301 * SIGRTMIN+24 when sent to a --user instance will now result
302 in immediate termination of systemd.
304 * gatewayd received numerous feature additions such as a
305 "follow" mode, for live syncing and filtering.
307 * browse.html now allows filtering and showing detailed
308 information on specific entries. Keyboard navigation and
309 mouse screen support has been added.
311 * gatewayd/journalctl now supports HTML5/JSON
312 Server-Sent-Events as output.
314 * The SysV init script compatibility logic will now
315 heuristically determine whether a script supports the
316 "reload" verb, and only then make this available as
319 * "systemctl status --follow" has been removed, use "journalctl
322 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
323 have been removed since they are hardly useful to be
326 * And I'd like to take the opportunity to specifically mention
327 Zbigniew for his great contributions. Zbigniew, you rock!
329 Contributions from: Andrew Eikum, Christian Hesse, Colin
330 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
331 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
332 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
333 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
334 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
335 Jędrzejewski-Szmek, Сковорода Никита Андреевич
339 * If /etc/vconsole.conf is non-existent or empty we will no
340 longer load any console font or key map at boot by
341 default. Instead the kernel defaults will be left
342 intact. This is definitely the right thing to do, as no
343 configuration should mean no configuration, and hard-coding
344 font names that are different on all archs is probably a bad
345 idea. Also, the kernel default key map and font should be
346 good enough for most cases anyway, and mostly identical to
347 the userspace fonts/key maps we previously overloaded them
348 with. If distributions want to continue to default to a
349 non-kernel font or key map they should ship a default
350 /etc/vconsole.conf with the appropriate contents.
352 Contributions from: Colin Walters, Daniel J Walsh, Dave
353 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
354 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
358 * journalctl gained a new --cursor= switch to show entries
359 starting from the specified location in the journal.
361 * We now enforce a size limit on journal entry fields exported
362 with "-o json" in journalctl. Fields larger than 4K will be
363 assigned null. This can be turned off with --all.
365 * An (optional) journal gateway daemon is now available as
366 "systemd-journal-gatewayd.service". This service provides
367 access to the journal via HTTP and JSON. This functionality
368 will be used to implement live log synchronization in both
369 pull and push modes, but has various other users too, such
370 as easy log access for debugging of embedded devices. Right
371 now it is already useful to retrieve the journal via HTTP:
373 # systemctl start systemd-journal-gatewayd.service
374 # wget http://localhost:19531/entries
376 This will download the journal contents in a
377 /var/log/messages compatible format. The same as JSON:
379 # curl -H"Accept: application/json" http://localhost:19531/entries
381 This service is also accessible via a web browser where a
382 single static HTML5 app is served that uses the JSON logic
383 to enable the user to do some basic browsing of the
384 journal. This will be extended later on. Here's an example
385 screenshot of this app in its current state:
387 http://0pointer.de/public/journal-gatewayd
389 Contributions from: Kay Sievers, Lennart Poettering, Robert
390 Milasan, Tom Gundersen
394 * The bash completion logic is now available for journalctl
397 * We don't mount the "cpuset" controller anymore together with
398 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
399 started if no parameters are assigned to it. "cpuset" hence
400 broke code that assumed it it could create "cpu" groups and
403 * journalctl -f will now subscribe to terminal size changes,
404 and line break accordingly.
406 Contributions from: Dave Reisner, Kay Sievers, Lennart
407 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
411 * nspawn will now create a symlink /etc/localtime in the
412 container environment, copying the host's timezone
413 setting. Previously this has been done via a bind mount, but
414 since symlinks cannot be bind mounted this has now been
415 changed to create/update the appropriate symlink.
417 * journalctl -n's line number argument is now optional, and
418 will default to 10 if omitted.
420 * journald will now log the maximum size the journal files may
421 take up on disk. This is particularly useful if the default
422 built-in logic of determining this parameter from the file
423 system size is used. Use "systemctl status
424 systemd-journald.service" to see this information.
426 * The multi-seat X wrapper tool has been stripped down. As X
427 is now capable of enumerating graphics devices via udev in a
428 seat-aware way the wrapper is not strictly necessary
429 anymore. A stripped down temporary stop-gap is still shipped
430 until the upstream display managers have been updated to
431 fully support the new X logic. Expect this wrapper to be
432 removed entirely in one of the next releases.
434 * HandleSleepKey= in logind.conf has been split up into
435 HandleSuspendKey= and HandleHibernateKey=. The old setting
436 is not available anymore. X11 and the kernel are
437 distuingishing between these keys and we should too. This
438 also means the inhibition lock for these keys has been split
441 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
442 Poettering, Lukas Nykryn, Václav Pavlín
446 * Whenever a unit changes state we'll now log this to the
447 journal and show along the unit's own log output in
450 * ConditionPathIsMountPoint= can now properly detect bind
451 mount points too. (Previously, a bind mount of one file
452 system to another place in the same file system could not be
453 detected as mount, since they shared struct stat's st_dev
456 * We will now mount the cgroup controllers cpu, cpuacct,
457 cpuset and the controllers net_cls, net_prio together by
460 * nspawn containers will now have a virtualized boot
461 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
462 over with a randomized ID at container initialization). This
463 has the effect of making "journalctl -b" do the right thing
466 * The JSON output journal serialization has been updated not
467 to generate "endless" list objects anymore, but rather one
468 JSON object per line. This is more in line how most JSON
469 parsers expect JSON objects. The new output mode
470 "json-pretty" has been added to provide similar output, but
471 neatly aligned for readability by humans.
473 * We dropped all explicit sync() invocations in the shutdown
474 code. The kernel does this implicitly anyway in the kernel
475 reboot() syscall. halt(8)'s -n option is now a compatibility
478 * We now support virtualized reboot() in containers, as
479 supported by newer kernels. We will fall back to exit() if
480 CAP_SYS_REBOOT is not available to the container. Also,
481 nspawn makes use of this now and will actually reboot the
482 container if the containerized OS asks for that.
484 * journalctl will only show local log output by default
485 now. Use --merge (-m) to show remote log output, too.
487 * libsystemd-journal gained the new sd_journal_get_usage()
488 call to determine the current disk usage of all journal
489 files. This is exposed in the new "journalctl --disk-usage"
492 * journald gained a new configuration setting SplitMode= in
493 journald.conf which may be used to control how user journals
494 are split off. See journald.conf(5) for details.
496 * A new condition type ConditionFileNotEmpty= has been added.
498 * tmpfiles' "w" lines now support file globbing, to write
499 multiple files at once.
501 * We added Python bindings for the journal submission
502 APIs. More Python APIs for a number of selected APIs will
503 likely follow. Note that we intend to add native bindings
504 only for the Python language, as we consider it common
505 enough to deserve bindings shipped within systemd. There are
506 various projects outside of systemd that provide bindings
507 for languages such as PHP or Lua.
509 * Many conditions will now resolve specifiers such as %i. In
510 addition, PathChanged= and related directives of .path units
511 now support specifiers as well.
513 * There's now a new RPM macro definition for the system preset
516 * journald will now warn if it can't foward a message to the
517 syslog daemon because it's socket is full.
519 * timedated will no longer write or process /etc/timezone,
520 except on Debian. As we do not support late mounted /usr
521 anymore /etc/localtime always being a symlink is now safe,
522 and hence the information in /etc/timezone is not necessary
525 * logind will now always reserve one VT for a text getty (VT6
526 by default). Previously if more than 6 X sessions where
527 started they took up all the VTs with auto-spawned gettys,
528 so that no text gettys were available anymore.
530 * udev will now automatically inform the btrfs kernel logic
531 about btrfs RAID components showing up. This should make
532 simple hotplug based btrfs RAID assembly work.
534 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
535 (but not for its children which will stay at the kernel
536 default). This should allow setups with a lot more listening
539 * systemd will now always pass the configured timezone to the
540 kernel at boot. timedated will do the same when the timezone
543 * logind's inhibition logic has been updated. By default,
544 logind will now handle the lid switch, the power and sleep
545 keys all the time, even in graphical sessions. If DEs want
546 to handle these events on their own they should take the new
547 handle-power-key, handle-sleep-key and handle-lid-switch
548 inhibitors during their runtime. A simple way to achiveve
549 that is to invoke the DE wrapped in an invocation of:
551 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
553 * Access to unit operations is now checked via SELinux taking
554 the unit file label and client process label into account.
556 * systemd will now notify the administrator in the journal
557 when he over-mounts a non-empty directory.
559 * There are new specifiers that are resolved in unit files,
560 for the host name (%H), the machine ID (%m) and the boot ID
563 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
564 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
565 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
566 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
567 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
568 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
569 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
573 * Support for reading structured kernel messages from
574 /dev/kmsg has now been added and is enabled by default.
576 * Support for reading kernel messages from /proc/kmsg has now
577 been removed. If you want kernel messages in the journal
578 make sure to run a recent kernel (>= 3.5) that supports
579 reading structured messages from /dev/kmsg (see
580 above). /proc/kmsg is now exclusive property of classic
581 syslog daemons again.
583 * The libudev API gained the new
584 udev_device_new_from_device_id() call.
586 * The logic for file system namespace (ReadOnlyDirectory=,
587 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
588 require pivot_root() anymore. This means fewer temporary
589 directories are created below /tmp for this feature.
591 * nspawn containers will now see and receive all submounts
592 made on the host OS below the root file system of the
595 * Forward Secure Sealing is now supported for Journal files,
596 which provide cryptographical sealing of journal files so
597 that attackers cannot alter log history anymore without this
598 being detectable. Lennart will soon post a blog story about
599 this explaining it in more detail.
601 * There are two new service settings RestartPreventExitStatus=
602 and SuccessExitStatus= which allow configuration of exit
603 status (exit code or signal) which will be excepted from the
604 restart logic, resp. consider successful.
606 * journalctl gained the new --verify switch that can be used
607 to check the integrity of the structure of journal files and
608 (if Forward Secure Sealing is enabled) the contents of
611 * nspawn containers will now be run with /dev/stdin, /dev/fd/
612 and similar symlinks pre-created. This makes running shells
613 as container init process a lot more fun.
615 * The fstab support can now handle PARTUUID= and PARTLABEL=
618 * A new ConditionHost= condition has been added to match
619 against the hostname (with globs) and machine ID. This is
620 useful for clusters where a single OS image is used to
621 provision a large number of hosts which shall run slightly
622 different sets of services.
624 * Services which hit the restart limit will now be placed in a
627 Contributions from: Bertram Poettering, Dave Reisner, Huang
628 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
629 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
633 * When running in --user mode systemd will now become a
634 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
635 tree a lot more organized.
637 * A new PartOf= unit dependency type has been introduced that
638 may be used to group services in a natural way.
640 * "systemctl enable" may now be used to enable instances of
643 * journalctl now prints error log levels in red, and
644 warning/notice log levels in bright white. It also supports
645 filtering by log level now.
647 * cgtop gained a new -n switch (similar to top), to configure
648 the maximum number of iterations to run for. It also gained
649 -b, to run in batch mode (accepting no input).
651 * The suffix ".service" may now be ommited on most systemctl
652 command lines involving service unit names.
654 * There's a new bus call in logind to lock all sessions, as
655 well as a loginctl verb for it "lock-sessions".
657 * libsystemd-logind.so gained a new call sd_journal_perror()
658 that works similar to libc perror() but logs to the journal
659 and encodes structured information about the error number.
661 * /etc/crypttab entries now understand the new keyfile-size=
664 * shutdown(8) now can send a (configurable) wall message when
665 a shutdown is cancelled.
667 * The mount propagation mode for the root file system will now
668 default to "shared", which is useful to make containers work
669 nicely out-of-the-box so that they receive new mounts from
670 the host. This can be undone locally by running "mount
671 --make-rprivate /" if needed.
673 * The prefdm.service file has been removed. Distributions
674 should maintain this unit downstream if they intend to keep
675 it around. However, we recommend writing normal unit files
676 for display managers instead.
678 * Since systemd is a crucial part of the OS we will now
679 default to a number of compiler switches that improve
680 security (hardening) such as read-only relocations, stack
681 protection, and suchlike.
683 * The TimeoutSec= setting for services is now split into
684 TimeoutStartSec= and TimeoutStopSec= to allow configuration
685 of individual time outs for the start and the stop phase of
688 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
689 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
690 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
691 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
692 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
693 Gundersen, Zbigniew Jędrzejewski-Szmek
697 * The journal and id128 C APIs are now fully documented as man
700 * Extra safety checks have been added when transitioning from
701 the initial RAM disk to the main system to avoid accidental
704 * /etc/crypttab entries now understand the new keyfile-offset=
707 * systemctl -t can now be used to filter by unit load state.
709 * The journal C API gained the new sd_journal_wait() call to
710 make writing synchronous journal clients easier.
712 * journalctl gained the new -D switch to show journals from a
715 * journalctl now displays a special marker between log
716 messages of two different boots.
718 * The journal is now explicitly flushed to /var via a service
719 systemd-journal-flush.service, rather than implicitly simply
720 by seeing /var/log/journal to be writable.
722 * journalctl (and the journal C APIs) can now match for much
723 more complex expressions, with alternatives and
726 * When transitioning from the initial RAM disk to the main
727 system we will now kill all processes in a killing spree to
728 ensure no processes stay around by accident.
730 * Three new specifiers may be used in unit files: %u, %h, %s
731 resolve to the user name, user home directory resp. user
732 shell. This is useful for running systemd user instances.
734 * We now automatically rotate journal files if their data
735 object hash table gets a fill level > 75%. We also size the
736 hash table based on the configured maximum file size. This
737 together should lower hash collisions drastically and thus
738 speed things up a bit.
740 * journalctl gained the new "--header" switch to introspect
741 header data of journal files.
743 * A new setting SystemCallFilters= has been added to services
744 which may be used to apply blacklists or whitelists to
745 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
747 * nspawn gained a new --link-journal= switch (and quicker: -j)
748 to link the container journal with the host. This makes it
749 very easy to centralize log viewing on the host for all
750 guests while still keeping the journal files separated.
752 * Many bugfixes and optimizations
754 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
755 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
756 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
761 * Several tools now understand kernel command line arguments,
762 which are only read when run in an initial RAM disk. They
763 usually follow closely their normal counterparts, but are
766 * There's a new tool to analyze the readahead files that are
767 automatically generated at boot. Use:
769 /usr/lib/systemd/systemd-readahead analyze /.readahead
771 * We now provide an early debug shell on tty9 if this enabled. Use:
773 systemctl enable debug-shell.service
775 * All plymouth related units have been moved into the Plymouth
776 package. Please make sure to upgrade your Plymouth version
779 * systemd-tmpfiles now supports getting passed the basename of
780 a configuration file only, in which case it will look for it
781 in all appropriate directories automatically.
783 * udevadm info now takes a /dev or /sys path as argument, and
784 does the right thing. Example:
786 udevadm info /dev/sda
787 udevadm info /sys/class/block/sda
789 * systemctl now prints a warning if a unit is stopped but a
790 unit that might trigger it continues to run. Example: a
791 service is stopped but the socket that activates it is left
794 * "systemctl status" will now mention if the log output was
795 shortened due to rotation since a service has been started.
797 * The journal API now exposes functions to determine the
798 "cutoff" times due to rotation.
800 * journald now understands SIGUSR1 and SIGUSR2 for triggering
801 immediately flushing of runtime logs to /var if possible,
802 resp. for triggering immediate rotation of the journal
805 * It is now considered an error if a service is attempted to
806 be stopped that is not loaded.
808 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
810 * systemd-analyze now supports Python 3
812 * tmpfiles now supports cleaning up directories via aging
813 where the first level dirs are always kept around but
814 directories beneath it automatically aged. This is enabled
815 by prefixing the age field with '~'.
817 * Seat objects now expose CanGraphical, CanTTY properties
818 which is required to deal with very fast bootups where the
819 display manager might be running before the graphics drivers
820 completed initialization.
822 * Seat objects now expose a State property.
824 * We now include RPM macros for service enabling/disabling
825 based on the preset logic. We recommend RPM based
826 distributions to make use of these macros if possible. This
827 makes it simpler to reuse RPM spec files across
830 * We now make sure that the collected systemd unit name is
831 always valid when services log to the journal via
834 * There's a new man page kernel-command-line(7) detailing all
835 command line options we understand.
837 * The fstab generator may now be disabled at boot by passing
838 fstab=0 on the kernel command line.
840 * A new kernel command line option modules-load= is now understood
841 to load a specific kernel module statically, early at boot.
843 * Unit names specified on the systemctl command line are now
844 automatically escaped as needed. Also, if file system or
845 device paths are specified they are automatically turned
846 into the appropriate mount or device unit names. Example:
848 systemctl status /home
849 systemctl status /dev/sda
851 * The SysVConsole= configuration option has been removed from
854 * The SysV search path is no longer exported on the D-Bus
857 * The Names= option is been removed from unit file parsing.
859 * There's a new man page bootup(7) detailing the boot process.
861 * Every unit and every generator we ship with systemd now
862 comes with full documentation. The self-explanatory boot is
865 * A couple of services gained "systemd-" prefixes in their
866 name if they wrap systemd code, rather than only external
867 code. Among them fsck@.service which is now
868 systemd-fsck@.service.
870 * The HaveWatchdog property has been removed from the D-Bus
873 * systemd.confirm_spawn= on the kernel command line should now
876 * There's a new man page crypttab(5) which details all options
877 we actually understand.
879 * systemd-nspawn gained a new --capability= switch to pass
880 additional capabilities to the container.
882 * timedated will now read known NTP implementation unit names
883 from /usr/lib/systemd/ntp-units.d/*.list,
884 systemd-timedated-ntp.target has been removed.
886 * journalctl gained a new switch "-b" that lists log data of
887 the current boot only.
889 * The notify socket is in the abstract namespace again, in
890 order to support daemons which chroot() at start-up.
892 * There is a new Storage= configuration option for journald
893 which allows configuration of where log data should go. This
894 also provides a way to disable journal logging entirely, so
895 that data collected is only forwarded to the console, the
896 kernel log buffer or another syslog implementation.
898 * Many bugfixes and optimizations
900 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
901 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
902 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
903 Shawn Landden, Tom Gundersen
907 * "systemctl help <unit>" now shows the man page if one is
910 * Several new man pages have been added.
912 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
913 MaxLevelConsole= can now be specified in
914 journald.conf. These options allow reducing the amount of
915 data stored on disk or forwarded by the log level.
917 * TimerSlackNSec= can now be specified in system.conf for
918 PID1. This allows system-wide power savings.
920 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
921 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
926 * logind is now capable of (optionally) handling power and
927 sleep keys as well as the lid switch.
929 * journalctl now understands the syntax "journalctl
930 /usr/bin/avahi-daemon" to get all log output of a specific
933 * CapabilityBoundingSet= in system.conf now also influences
934 the capability bound set of usermode helpers of the kernel.
936 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
937 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
938 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
939 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
943 * Note that we skipped 139 releases here in order to set the
944 new version to something that is greater than both udev's
945 and systemd's most recent version number.
947 * udev: all udev sources are merged into the systemd source tree now.
948 All future udev development will happen in the systemd tree. It
949 is still fully supported to use the udev daemon and tools without
950 systemd running, like in initramfs or other init systems. Building
951 udev though, will require the *build* of the systemd tree, but
952 udev can be properly *run* without systemd.
954 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
955 should be used to create dead device nodes as workarounds for broken
958 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
959 no longer supported. udev_monitor_new_from_netlink() needs to be
960 used to subscribe to events.
962 * udev: when udevd is started by systemd, processes which are left
963 behind by forking them off of udev rules, are unconditionally cleaned
964 up and killed now after the event handling has finished. Services or
965 daemons must be started as systemd services. Services can be
966 pulled-in by udev to get started, but they can no longer be directly
967 forked by udev rules.
969 * udev: the daemon binary is called systemd-udevd now and installed
970 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
971 to adapt to that, create symlink, or rename the binary after building
974 * libudev no longer provides these symbols:
975 udev_monitor_from_socket()
976 udev_queue_get_failed_list_entry()
977 udev_get_{dev,sys,run}_path()
978 The versions number was bumped and symbol versioning introduced.
980 * systemd-loginctl and systemd-journalctl have been renamed
981 to loginctl and journalctl to match systemctl.
983 * The config files: /etc/systemd/systemd-logind.conf and
984 /etc/systemd/systemd-journald.conf have been renamed to
985 logind.conf and journald.conf. Package updates should rename
986 the files to the new names on upgrade.
988 * For almost all files the license is now LGPL2.1+, changed
989 from the previous GPL2.0+. Exceptions are some minor stuff
990 of udev (which will be changed to LGPL2.1 eventually, too),
991 and the MIT licensed sd-daemon.[ch] library that is suitable
992 to be used as drop-in files.
994 * systemd and logind now handle system sleep states, in
995 particular suspending and hibernating.
997 * logind now implements a sleep/shutdown/idle inhibiting logic
998 suitable for a variety of uses. Soonishly Lennart will blog
999 about this in more detail.
1001 * var-run.mount and var-lock.mount are no longer provided
1002 (which prevously bind mounted these directories to their new
1003 places). Distributions which have not converted these
1004 directories to symlinks should consider stealing these files
1005 from git history and add them downstream.
1007 * We introduced the Documentation= field for units and added
1008 this to all our shipped units. This is useful to make it
1009 easier to explore the boot and the purpose of the various
1012 * All smaller setup units (such as
1013 systemd-vconsole-setup.service) now detect properly if they
1014 are run in a container and are skipped when
1015 appropriate. This guarantees an entirely noise-free boot in
1016 Linux container environments such as systemd-nspawn.
1018 * A framework for implementing offline system updates is now
1019 integrated, for details see:
1020 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
1022 * A new service type Type=idle is available now which helps us
1023 avoiding ugly interleaving of getty output and boot status
1026 * There's now a system-wide CapabilityBoundingSet= option to
1027 globally reduce the set of capabilities for the
1028 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
1029 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
1030 even CAP_NET_ADMIN system-wide for secure systems.
1032 * There are now system-wide DefaultLimitXXX= options to
1033 globally change the defaults of the various resource limits
1034 for all units started by PID 1.
1036 * Harald Hoyer's systemd test suite has been integrated into
1037 systemd which allows easy testing of systemd builds in qemu
1038 and nspawn. (This is really awesome! Ask us for details!)
1040 * The fstab parser is now implemented as generator, not inside
1043 * systemctl will now warn you if .mount units generated from
1044 /etc/fstab are out of date due to changes in fstab that
1045 haven't been read by systemd yet.
1047 * systemd is now suitable for usage in initrds. Dracut has
1048 already been updated to make use of this. With this in place
1049 initrds get a slight bit faster but primarily are much
1050 easier to introspect and debug since "systemctl status" in
1051 the host system can be used to introspect initrd services,
1052 and the journal from the initrd is kept around too.
1054 * systemd-delta has been added, a tool to explore differences
1055 between user/admin configuration and vendor defaults.
1057 * PrivateTmp= now affects both /tmp and /var/tmp.
1059 * Boot time status messages are now much prettier and feature
1060 proper english language. Booting up systemd has never been
1063 * Read-ahead pack files now include the inode number of all
1064 files to pre-cache. When the inode changes the pre-caching
1065 is not attempted. This should be nicer to deal with updated
1066 packages which might result in changes of read-ahead
1069 * We now temporaritly lower the kernel's read_ahead_kb variable
1070 when collecting read-ahead data to ensure the kernel's
1071 built-in read-ahead does not add noise to our measurements
1072 of necessary blocks to pre-cache.
1074 * There's now RequiresMountsFor= to add automatic dependencies
1075 for all mounts necessary for a specific file system path.
1077 * MountAuto= and SwapAuto= have been removed from
1078 system.conf. Mounting file systems at boot has to take place
1081 * nspawn now learned a new switch --uuid= to set the machine
1082 ID on the command line.
1084 * nspawn now learned the -b switch to automatically search
1087 * vt102 is now the default TERM for serial TTYs, upgraded from
1090 * systemd-logind now works on VT-less systems.
1092 * The build tree has been reorganized. The individual
1093 components now have directories of their own.
1095 * A new condition type ConditionPathIsReadWrite= is now available.
1097 * nspawn learned the new -C switch to create cgroups for the
1098 container in other hierarchies.
1100 * We now have support for hardware watchdogs, configurable in
1103 * The scheduled shutdown logic now has a public API.
1105 * We now mount /tmp as tmpfs by default, but this can be
1106 masked and /etc/fstab can override it.
1108 * Since udisks doesn't make use of /media anymore we are not
1109 mounting a tmpfs on it anymore.
1111 * journalctl gained a new --local switch to only interleave
1112 locally generated journal files.
1114 * We can now load the IMA policy at boot automatically.
1116 * The GTK tools have been split off into a systemd-ui.
1118 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
1119 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
1120 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
1121 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
1122 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
1123 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
1124 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
1125 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
1126 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
1131 * This is mostly a bugfix release
1133 * Support optional initialization of the machine ID from the
1134 KVM or container configured UUID.
1136 * Support immediate reboots with "systemctl reboot -ff"
1138 * Show /etc/os-release data in systemd-analyze output
1140 * Many bugfixes for the journal, including endianess fixes and
1141 ensuring that disk space enforcement works
1143 * sd-login.h is C++ comptaible again
1145 * Extend the /etc/os-release format on request of the Debian
1148 * We now refuse non-UTF8 strings used in various configuration
1149 and unit files. This is done to ensure we don't pass invalid
1150 data over D-Bus or expose it elsewhere.
1152 * Register Mimo USB Screens as suitable for automatic seat
1155 * Read SELinux client context from journal clients in a race
1158 * Reorder configuration file lookup order. /etc now always
1159 overrides /run in order to allow the administrator to always
1160 and unconditionally override vendor supplied or
1161 automatically generated data.
1163 * The various user visible bits of the journal now have man
1164 pages. We still lack man pages for the journal API calls
1167 * We now ship all man pages in HTML format again in the
1170 Contributions from: Dave Reisner, Dirk Eibach, Frederic
1171 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
1172 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
1177 * This is mostly a bugfix release
1179 * systems lacking /etc/os-release are no longer supported.
1181 * Various functionality updates to libsystemd-login.so
1183 * Track class of PAM logins to distuingish greeters from
1186 Contributions from: Kay Sievers, Lennart Poettering, Michael
1191 * This is an important bugfix release for v41.
1193 * Building man pages is now optional which should be useful
1194 for those building systemd from git but unwilling to install
1197 * Watchdog support for supervising services is now usable. In
1198 a future release support for hardware watchdogs
1199 (i.e. /dev/watchdog) will be added building on this.
1201 * Service start rate limiting is now configurable and can be
1202 turned off per service. When a start rate limit is hit a
1203 reboot can automatically be triggered.
1205 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
1207 Contributions from: Benjamin Franzke, Bill Nottingham,
1208 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
1209 Schmidt, Michał Górny, Piotr Drąg
1213 * The systemd binary is installed /usr/lib/systemd/systemd now;
1214 An existing /sbin/init symlink needs to be adapted with the
1217 * The code that loads kernel modules has been ported to invoke
1218 libkmod directly, instead of modprobe. This means we do not
1219 support systems with module-init-tools anymore.
1221 * Watchdog support is now already useful, but still not
1224 * A new kernel command line option systemd.setenv= is
1225 understood to set system wide environment variables
1226 dynamically at boot.
1228 * We now limit the set of capabilities of systemd-journald.
1230 * We now set SIGPIPE to ignore by default, since it only is
1231 useful in shell pipelines, and has little use in general
1232 code. This can be disabled with IgnoreSIPIPE=no in unit
1235 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
1236 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
1241 * This is mostly a bugfix release
1243 * We now expose the reason why a service failed in the
1244 "Result" D-Bus property.
1246 * Rudimentary service watchdog support (will be completed over
1247 the next few releases.)
1249 * When systemd forks off in order execute some service we will
1250 now immediately changes its argv[0] to reflect which process
1251 it will execute. This is useful to minimize the time window
1252 with a generic argv[0], which makes bootcharts more useful
1254 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
1255 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
1256 Mike Kazantsev, Ray Strode
1260 * This is mostly a test release, but incorporates many
1263 * New systemd-cgtop tool to show control groups by their
1266 * Linking against libacl for ACLs is optional again. If
1267 disabled, support tracking device access for active logins
1268 goes becomes unavailable, and so does access to the user
1269 journals by the respective users.
1271 * If a group "adm" exists, journal files are automatically
1272 owned by them, thus allow members of this group full access
1273 to the system journal as well as all user journals.
1275 * The journal now stores the SELinux context of the logging
1276 client for all entries.
1278 * Add C++ inclusion guards to all public headers
1280 * New output mode "cat" in the journal to print only text
1281 messages, without any meta data like date or time.
1283 * Include tiny X server wrapper as a temporary stop-gap to
1284 teach XOrg udev display enumeration. This is used by display
1285 managers such as gdm, and will go away as soon as XOrg
1286 learned native udev hotplugging for display devices.
1288 * Add new systemd-cat tool for executing arbitrary programs
1289 with STDERR/STDOUT connected to the journal. Can also act as
1290 BSD logger replacement, and does so by default.
1292 * Optionally store all locally generated coredumps in the
1293 journal along with meta data.
1295 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
1296 writing short strings to files (for usage for /sys), and for
1297 creating symlinks, character and block device nodes.
1299 * New unit file option ControlGroupPersistent= to make cgroups
1300 persistent, following the mechanisms outlined in
1301 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
1303 * Support multiple local RTCs in a sane way
1305 * No longer monopolize IO when replaying readahead data on
1306 rotating disks, since we might starve non-file-system IO to
1307 death, since fanotify() will not see accesses done by blkid,
1310 * Don't show kernel threads in systemd-cgls anymore, unless
1311 requested with new -k switch.
1313 Contributions from: Dan Horák, Kay Sievers, Lennart
1314 Poettering, Michal Schmidt
1318 * This is mostly a test release, but incorporates many
1321 * The git repository moved to:
1322 git://anongit.freedesktop.org/systemd/systemd
1323 ssh://git.freedesktop.org/git/systemd/systemd
1325 * First release with the journal
1326 http://0pointer.de/blog/projects/the-journal.html
1328 * The journal replaces both systemd-kmsg-syslogd and
1329 systemd-stdout-bridge.
1331 * New sd_pid_get_unit() API call in libsystemd-logind
1333 * Many systemadm clean-ups
1335 * Introduce remote-fs-pre.target which is ordered before all
1336 remote mounts and may be used to start services before all
1339 * Added Mageia support
1341 * Add bash completion for systemd-loginctl
1343 * Actively monitor PID file creation for daemons which exit in
1344 the parent process before having finished writing the PID
1345 file in the daemon process. Daemons which do this need to be
1346 fixed (i.e. PID file creation must have finished before the
1347 parent exits), but we now react a bit more gracefully to them.
1349 * Add colourful boot output, mimicking the well-known output
1350 of existing distributions.
1352 * New option PassCredentials= for socket units, for
1353 compatibility with a recent kernel ABI breakage.
1355 * /etc/rc.local is now hooked in via a generator binary, and
1356 thus will no longer act as synchronization point during
1359 * systemctl list-unit-files now supports --root=.
1361 * systemd-tmpfiles now understands two new commands: z, Z for
1362 relabelling files according to the SELinux database. This is
1363 useful to apply SELinux labels to specific files in /sys,
1366 * Output of SysV services is now forwarded to both the console
1367 and the journal by default, not only just the console.
1369 * New man pages for all APIs from libsystemd-login.
1371 * The build tree got reorganized and a the build system is a
1372 lot more modular allowing embedded setups to specifically
1373 select the components of systemd they are interested in.
1375 * Support for Linux systems lacking the kernel VT subsystem is
1378 * configure's --with-rootdir= got renamed to
1379 --with-rootprefix= to follow the naming used by udev and
1382 * Unless specified otherwise we'll now install to /usr instead
1383 of /usr/local by default.
1385 * Processes with '@' in argv[0][0] are now excluded from the
1386 final shut-down killing spree, following the logic explained
1388 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
1390 * All processes remaining in a service cgroup when we enter
1391 the START or START_PRE states are now killed with
1392 SIGKILL. That means it is no longer possible to spawn
1393 background processes from ExecStart= lines (which was never
1394 supported anyway, and bad style).
1396 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
1397 reloading of units together.
1399 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
1400 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
1401 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
1402 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
1403 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
~ianmdlvl / elogind.git / blob