1 systemd System and Service Manager
3 CHANGES WITH 243 in spe:
5 * Previously, filters defined with SystemCallFilter= would have the
6 effect that an calling an offending system call would terminate the
7 calling thread. This behaviour never made much sense, since killing
8 individual threads of unsuspecting processes is likely to create more
9 problems than it solves. With this release the default action changed
10 from killing the thread to killing the whole process. For this to
11 work correctly both a kernel version (>= 4.14) and a libseccomp
12 version (>= 2.4.0) supporting this new seccomp action is required. If
13 an older kernel or libseccomp is used the old behaviour continues to
14 be used. This change does not affect any services that have no system
15 call filters defined, or that use SystemCallErrorNumber= (and thus
16 see EPERM or another error instead of being killed when calling an
17 offending system call). Note that systemd documentation always
18 claimed that the whole process is killed. With this change behaviour
19 is thus adjusted to match the documentation.
21 * The "kernel.pid_max" sysctl is now bumped to 4194304 by default,
22 i.e. the full 22bit range the kernel allows, up from the old 16bit
23 range. This should improve security and robustness a bit, as PID
24 collisions are made less likely (though certainly still
25 possible). There are rumours this might create compatibility
26 problems, though at this moment no practical ones are known to
27 us. Downstream distributions are hence advised to undo this change in
28 their builds if they are concerned about maximum compatibility, but
29 for everybody else we recommend leaving the value bumped. Besides
30 improving security and robustness this should also simplify things as
31 the maximum number of allowed concurrent tasks was previously bounded
32 by both "kernel.pid_max" and "kernel.threads-max" and now only a
33 single knob is left ("kernel.threads-max"). There have been concerns
34 that usability is affected by this change because larger PID numbers
35 are harder to type, but we believe the change from 5 digit PIDs to 7
36 digit PIDs is not too hampering for usability.
38 * MemoryLow= and MemoryMin= gained hierarchy-aware counterparts,
39 DefaultMemoryLow= and DefaultMemoryMin=, which can be used to
40 hierarchically set default memory protection values for a particular
41 subtree of the unit hierarchy.
43 * Memory protection directives can now take a value of zero, allowing
44 explicit opting out of a default value propagated by an ancestor.
46 * A new setting DisableControllers= has been added that may be used to
47 explicitly disable one or more cgroups controllers for a unit and all
50 * systemd now defaults to the "unified" cgroup hierarchy setup during
51 build-time, i.e. -Ddefault-hierarchy=unified is now the build-time
52 default. Previously, -Ddefault-hierarchy=hybrid was the default. This
53 change reflects the fact that cgroupsv2 support has matured
54 substantially in both systemd and in the kernel, and is clearly the
55 way forward. Downstream production distributions might want to
56 continue to use -Ddefault-hierarchy=hybrid (or even =legacy) for
57 their builds as unfortunately the popular container managers have not
58 caught up with the kernel API changes.
60 * Man pages are not built by default anymore (html pages were already
61 disabled by default), to make development builds quicker. When
62 building systemd for a full installation with documentation, meson
63 should be called -Dman=true and/or -Dhtml=true as appropriate. The
64 default was changed based on the assumption that quick one-off or
65 repeated development builds are much more common than full optimized
66 builds for installation, and people need to pass various other
67 options to when doing "proper" builds anyway, so the gain from making
68 development builds quicker is bigger than the one time disruption for
71 Two scripts are created in the *build* directory to generate and
72 preview man and html pages on demand, e.g.:
74 build/man/man systemctl
75 build/man/html systemd.index
77 * libidn2 is used by default if both libidn2 and libidn are installed.
78 Please use -Dlibidn=true when libidn is favorable.
80 * The D-Bus "wire format" of the CPUAffinity= attribute is changed on
81 big-endian machines. Before, bytes were written and read in native
82 machine order as exposed by the native libc __cpu_mask interface.
83 Now, little-endian order is always used (CPUs 0–7 are described by
84 bits 0–7 in byte 0, CPUs 8–15 are described by byte 1, and so on).
85 This change fixes D-Bus calls that cross endianness boundary.
87 The presentation format used for CPUAffinity= by "systemctl show" and
88 "systemd-analyze dump" is changed to present CPU indices instead of
89 the raw __cpu_mask bitmask. For example, CPUAffinity=0-1 would be
90 shown as CPUAffinity=03000000000000000000000000000… (on
91 little-endian) or CPUAffinity=00000000000000300000000000000… (on
92 64-bit big-endian), and is now shown as CPUAffinity=0-1, matching the
93 input format. The maximum integer that will be printed in the new
94 format is 8191 (four digits), while the old format always used a very
95 long number (with the length varying by architecture), so they can be
96 unambiguously distinguished.
98 * /usr/sbin/halt.local is no longer supported. Implementation in
99 distributions was inconsistent and it seems this functionality was
102 To replace this functionality, users should:
103 - either define a new unit and make it a dependency of final.target
104 (systemctl add-wants final.target my-halt-local.service)
105 - or move the shutdown script to /usr/lib/systemd/system-shutdown/
106 and ensure that it accepts "halt", "poweroff", "reboot", and
107 "kexec" as an argument, see the description in systemd-shutdown(8).
109 * When a [Match] section in .link or .network file is empty (contains
110 no match patterns), a warning will be emitted. Please add any "match
111 all" pattern instead, e.g. OriginalName=* or Name=* in case all
112 interfaces should really be matched.
114 * A new setting NUMAPolicy= may be used to set process memory
115 allocation policy. Setting can be specified in system.conf and hence
116 will set the default policy for PID1. Default policy can be
117 overridden on per-service basis. Related setting NUMAMask= is used to
118 specify NUMA node mask that should be associated with the selected
121 * PID 1 will now listen to Out-Of-Memory (OOM) events the kernel
122 generates when processes it manages are reaching their memory limits,
123 and will place their units in a special state, and optionally kill or
126 * The service manager will now expose bus properties for the IO
127 resources used by units. This information is also shown in "systemctl
128 status" now (for services that have IOAccounting=yes set). Moreover,
129 the IO accounting data is included in the resource log message
130 generated whenever a unit stops.
132 * units may now configure an explicit time-out to apply to when killed
133 with SIGABRT, for example when a service watchdog is hit. Previously,
134 the regular TimeoutStopSec= time-out was applied in this case too —
135 now a separate time-out may be set using TimeoutAbortSec=.
137 * Services may now send a special WATCHDOG=trigger message with
138 sd_notify() to trigger an immediate "watchdog missed" event, and thus
139 request service take down. This is useful both for testing watchdog
140 handling, but also for defining error paths in services, that shall
141 be handled the same way as watchdog events.
143 * There are two new per-unit settings IPIngressFilterPath= and
144 IPEgressFilterPath= which allow configuration of a BPF program
145 (usually by specifying a path to a program uploaded to /sys/fs/bpf/)
146 to apply to the IP packet ingress/egress path of all processes of a
147 unit. This is useful to allow running systemd services with BPF
148 programs set up externally.
150 * systemctl gained a new "clean" verb for removing the state, cache,
151 runtime or logs directories of a service while it is terminated. The
152 new verb may also be used to remove the state maintained on disk for
153 timer units that have Persistent= configured.
155 * During the last phase of shutdown systemd will now automatically
156 increase the log level configured in the "kernel.printk" sysctl so
157 that any relevant loggable events happening during late shutdown are
158 made visible. Previously, loggable events happening so late during
159 shutdown were generally lost if the "kernel.printk" sysctl was set to
160 high thresholds, as regular logging daemons are terminated at that
161 time and thus nothing is written to disk.
163 * If processes terminated during the last phase of shutdown do not exit
164 quickly systemd will now show their names after a short time, to make
165 debugging easier. After a longer time-out they are forcibly killed,
168 * journalctl (and the other tools that display logs) will now highlight
169 warnings in yellow (previously, both LOG_NOTICE and LOG_WARNING where
170 shown in bright bold, now only LOG_NOTICE is). Moreover, audit logs
171 are now shown in blue color, to separate them visually from regular
172 logs. References to configuration files are now turned into clickable
173 links on terminals that support that.
175 * systemd-journald will now stop logging to /var/log/journal during
176 shutdown when /var/ is on a separate mount, so that it can be
177 unmounted safely during shutdown.
179 * systemd-resolved gained support for a new 'strict' DNS-over-TLS mode.
181 * systemd-resolved "Cache=" configuration option in resolved.conf has
182 been extended to also accept the 'no-negative' value. Previously,
183 only a boolean option was allowed (yes/no), having yes as the
184 default. If this option is set to 'no-negative', negative answers
185 are skipped from being cached while keeping the same cache heuristics
186 for positive answers. The default remains as "yes" (i. e. caching is
189 * The predictable naming scheme for network devices now supports
190 generating predictable names for "netdevsim" devices.
192 * systemd-networkd now supports MACsec, nlmon, IPVTAP and Xfrm
195 * systemd-networkd's bridge FDB support now allows configuration of a
196 destination address for each entry (Destination=), as well as the
197 VXLAN VNI (VNI=), as well as an option to declare what an entry is
198 associated with (AssociatedWith=).
200 * systemd-networkd's DHCPv4 support now understands a new MaxAttempts=
201 option for configuring the maximum number of attempts to request a
202 DHCP lease. It also learnt a new BlackList= option for blacklisting
203 DHCP servers (a similar setting has also been added to the IPv6 RA
204 client), as well as a SendRelease= option for configuring whether to
205 send a DHCP RELEASE message when terminating.
207 * systemd-networkd's DHCPv4 and DHCPv6 stacks can now be configured
208 separately in the [DHCPv4] and [DHCPv6] sections.
210 * systemd-networkd's DHCP support will now optionally create an
211 implicit host route to the DNS server specified in the DHCP lease, in
212 addition to the routes listed explicitly in the lease. This should
213 ensure that in multi-homed systems DNS traffic leaves the systems on
214 the interface that acquired the DNS server information even if other
215 routes such as default routes exist. This behaviour may be turned on
216 with the new RoutesToDNS= option.
218 * systemd-networkd's VXLAN support gained a new option
219 GenericProtocolExtension= for enabling VXLAN Generic Protocol
220 Extension support, as well as IPDoNotFragment= for setting the IP
221 "Don't fragment" bit on outgoing packets. A similar option has been
222 added to the GENEVE support.
224 * In systemd-networkd's [Route] section you may now configure
225 FastOpenNoCookie= for configuring per-route TCP fast-open support, as
226 well as TTLPropagate= for configuring Label Switched Path (LSP) TTL
227 propagation. The Type= setting now supports local, broadcast,
228 anycast, multicast, any, xresolve routes, too.
230 * systemd-networkd's [Network] section learnt a new option
231 DefaultRouteOnDevice= for automatically configuring a default route
232 onto the network device.
234 * systemd-networkd's bridging support gained two new options ProxyARP=
235 and ProxyARPWifi= for configuring proxy ARP behaviour as well as
236 MulticastRouter= for configuring multicast routing behaviour.
238 * systemd-networkd's FooOverUDP support gained the ability to configure
239 local and peer IP addresses via Local= and Peer=. A new option
240 PeerPort= may be used to configure the peer's IP port.
242 * systemd-networkd's TUN support gained a new setting VnetHeader= for
243 tweaking Generic Segment Offload support.
245 * networkctl gained a new "delete" command for removing virtual network
246 devices, as well as a new "--stats" switch for showing device
249 * networkd.conf gained a new setting SpeedMeter= and
250 SpeedMeterIntervalSec=, to measure bitrate of network interfaces. The
251 measured speed may be shown by 'networkctl status'.
253 * systemd-networkd's .network and .link files gained a new Property=
254 setting in the [Match] section, to match against devices with
255 specific udev properties.
257 * systemd-networkd's tunnel support gained a new option
258 AssignToLoopback= for selecting whether to use the loopback device
259 "lo" as underlying device.
261 * systemd-networkd's MACAddress= setting in the [Neighbor] section has
262 been renamed to LinkLayerAddress=, and it now allows configuration of
265 * A new tool systemd-network-generator has been added that may generate
266 .network, .netdev and .link files from IP configuration specified on
267 the kernel command line, compatible with the format Dracut expects.
269 * The CriticalConnection= setting in .network files is now deprecated,
270 and replaced by a new KeepConfiguration= setting which allows more
271 detailed configuration of the IP configuration to keep in place.
273 * systemd-analyze gained a new "timestamp" verb for parsing and
274 converting timestamps. It's similar to the existing "systemd-analyze
275 calendar" command which does the same for recurring calendar
276 events. It also gained a new "condition" verb for parsing and testing
277 ConditionXYZ= expressions.
279 * systemd-logind now exposes a per-session SetBrightness() bus call,
280 which may be used to securely change the brightness of a kernel
281 brightness device, if it belongs to the session's seat. By using this
282 call unprivileged clients can make changes to "backlight" and "leds"
283 devices securely with strict requirements on session
284 membership. Desktop environments may use this to generically make
285 brightness changes to such devices without shipping private SUID
286 binaries or specific udev rules for that purpose.
288 * "udevadm info" gained a --wait-for-initialization switch to wait for
289 a device to be initialized.
291 * systemd-hibernate-resume-generator will now look for resumeflags= on
292 the kernel command line, which is similar to rootflags= and may be
293 used to configure device timeouts for waiting for the hibernation
296 * sd-event learnt a new API call sd_event_source_disable_unref() for
297 disabling and unref'ing an event source in a single function. A
298 related call sd_event_source_disable_unrefp() has been added for use
299 with GCC's cleanup extension.
301 * The sd-id128.h public API gained a new definition
302 SD_ID128_UUID_FORMAT_STR for formatting a 128bit ID in UUID format
305 * "busctl introspect" gained a new switch --xml-interface for dumping
306 XML introspection data unmodified.
308 * PID 1 may now show the unit name instead of the unit description
309 string in its status output during boot. This may be configured in
310 the StatusUnitFormat= setting in /etc/systemd/system.conf or the
311 kernel command line option systemd.status_unit_format=.
313 * PID 1 now understands a new option KExecWatchdogSec= in
314 /etc/systemd/system.conf. It allows configuration of a watchdog
315 timeout to write to a hardware watchdog device on kexec-based
316 reboots. Previously this functionality was only available for regular
317 reboots. This option defaults to off, since it depends on drivers and
318 software setup whether the watchdog is correctly reset again after
319 the kexec completed, and thus for the general case not clear if safe
320 (since it might cause unwanted watchdog reboots after the kexec
321 completed otherwise). Moreover, the old ShutdownWatchdogSec= setting
322 has been renamed to RebootWatchdogSec= to more clearly communicate
323 what it is about. The old name of the setting is still accepted for
326 * The systemd.debug_shell kernel command line option now optionally
327 takes a tty name to spawn the debug shell on, which allows selecting
328 a different tty than the built-in default.
330 * Service units gained a new ExecCondition= setting which will run
331 before ExecStartPre= and either continue execution of the unit (for
332 clean exit codes), stop execution without marking the unit failed
333 (for exit codes 1 through 254), or stop execution and fail the unit
334 (for exit code 255 or cases of abnormal termination).
336 * A new service systemd-pstore.service has been added that pulls data
337 from from /sys/fs/pstore/ and saves it to /var/lib/pstore for later
340 * timedatectl gained new verbs for configuring per-interface NTP
341 service configuration for systemd-timesyncd.
343 * "localectl list-locales" won't list non-UTF-8 locales anymore. It's
344 2019. (You can set non-UTF-8 locales though, if you know there name.)
346 Contributions from: Aaron Barany, Adrian Bunk, Alan Jenkins, Andrej
347 Valek, Anita Zhang, Arian van Putten, Balint Reczey, Bastien Nocera,
348 Ben Boeckel, Benjamin Robin, camoz, Chen Qi, Chris Chiu, Chris Down,
349 Christian Kellner, Connor Reeder, Daniele Medri, Dan Streetman, Dave
350 Reisner, Dave Ross, David Art, David Tardon, Debarshi Ray, Dominick
351 Grift, Donald Buczek, Douglas Christman, Eric DeVolder, Evgeny
352 Vereshchagin, Feldwor, Felix Riemann, Florian Dollinger, Franck Bui,
353 Frantisek Sumsal, Franz Pletz, Hans de Goede, Insun Pyo, Ivan
354 Shapovalov, Iwan Timmer, Jack, Jakob Unterwurzacher, Jan Klötzke, Jan
355 Pokorný, Jan Synacek, Jeka Pats, Jérémy Rosen, Jiri Pirko, Joe Lin,
356 Joerg Behrmann, Joe Richey, Jóhann B. Guðmundsson, Johannes Schmitz,
357 Jonathan Rouleau, Jorge Niedbalski, Kai Lüke, Karel Zak, Kashyap
358 Chamarthy, Krayushkin Konstantin, Lennart Poettering, Lubomir Rintel,
359 Luca Boccassi, Luís Ferreira, Marc-André Lureau, Markus Felten, Martin
360 Pitt, Michael Biebl, Michael Olbrich, Michael Prokop, Michael
361 Stapelberg, Michael Zhivich, Michal Koutný, Michal Sekletar, Mike
362 Gilbert, Milan Broz, mpe85, Oliver Harley, pan93412, Paul Menzel,
363 pEJipE, Peter A. Bigot, Philip Withnall, Piotr Drąg, Rafael Fontenelle,
364 Roberto Santalla, root, RussianNeuroMancer, Sebastian Jennen, Simon
365 Schricker, Susant Sahani, Thadeu Lima de Souza Cascardo, Theo
366 Ouzhinski, Thomas Weißschuh, Tomas Mraz, Topi Miettinen, ven, Wieland
367 Hoffmann, Xi Ruoyao, Yuri Chornoivan, Yu Watanabe, Zach Smith, Zbigniew
368 Jędrzejewski-Szmek, Zhang Xianwei
370 – Somewhere, SOME-TI-ME
374 * In .link files, MACAddressPolicy=persistent (the default) is changed
375 to cover more devices. For devices like bridges, tun, tap, bond, and
376 similar interfaces that do not have other identifying information,
377 the interface name is used as the basis for persistent seed for MAC
378 and IPv4LL addresses. The way that devices that were handled
379 previously is not changed, and this change is about covering more
380 devices then previously by the "persistent" policy.
382 MACAddressPolicy=random may be used to force randomized MACs and
383 IPv4LL addresses for a device if desired.
385 Hint: the log output from udev (at debug level) was enhanced to
386 clarify what policy is followed and which attributes are used.
387 `SYSTEMD_LOG_LEVEL=debug udevadm test-builtin net_setup_link /sys/class/net/<name>`
388 may be used to view this.
390 Hint: if a bridge interface is created without any slaves, and gains
391 a slave later, then now the bridge does not inherit slave's MAC.
392 To inherit slave's MAC, for example, create the following file:
394 # /etc/systemd/network/98-bridge-inherit-mac.link
399 MACAddressPolicy=none
402 * The .device units generated by systemd-fstab-generator and other
403 generators do not automatically pull in the corresponding .mount unit
404 as a Wants= dependency. This means that simply plugging in the device
405 will not cause the mount unit to be started automatically. But please
406 note that the mount unit may be started for other reasons, in
407 particular if it is part of local-fs.target, and any unit which
408 (transitively) depends on local-fs.target is started.
410 * networkctl list/status/lldp now accept globbing wildcards for network
411 interface names to match against all existing interfaces.
413 * The $PIDFILE environment variable is set to point the absolute path
414 configured with PIDFile= for processes of that service.
416 * The fallback DNS server list was augmented with Cloudflare public DNS
417 servers. Use `-Ddns-servers=` to set a different fallback.
419 * A new special target usb-gadget.target will be started automatically
420 when a USB Device Controller is detected (which means that the system
421 is a USB peripheral).
423 * A new unit setting CPUQuotaPeriodSec= assigns the time period
424 relatively to which the CPU time quota specified by CPUQuota= is
427 * A new unit setting ProtectHostname= may be used to prevent services
428 from modifying hostname information (even if they otherwise would
429 have privileges to do so).
431 * A new unit setting NetworkNamespacePath= may be used to specify a
432 namespace for service or socket units through a path referring to a
433 Linux network namespace pseudo-file.
435 * The PrivateNetwork= setting and JoinsNamespaceOf= dependencies now
436 have an effect on .socket units: when used the listening socket is
437 created within the configured network namespace instead of the host
440 * ExecStart= command lines in unit files may now be prefixed with ':'
441 in which case environment variable substitution is
442 disabled. (Supported for the other ExecXYZ= settings, too.)
444 * .timer units gained two new boolean settings OnClockChange= and
445 OnTimezoneChange= which may be used to also trigger a unit when the
446 system clock is changed or the local timezone is
447 modified. systemd-run has been updated to make these options easily
448 accessible from the command line for transient timers.
450 * Two new conditions for units have been added: ConditionMemory= may be
451 used to conditionalize a unit based on installed system
452 RAM. ConditionCPUs= may be used to conditionalize a unit based on
455 * The @default system call filter group understood by SystemCallFilter=
456 has been updated to include the new rseq() system call introduced in
459 * A new time-set.target has been added that indicates that the system
460 time has been set from a local source (possibly imprecise). The
461 existing time-sync.target is stronger and indicates that the time has
462 been synchronized with a precise external source. Services where
463 approximate time is sufficient should use the new target.
465 * "systemctl start" (and related commands) learnt a new
466 --show-transaction option. If specified brief information about all
467 jobs queued because of the requested operation is shown.
469 * systemd-networkd recognizes a new operation state 'enslaved', used
470 (instead of 'degraded' or 'carrier') for interfaces which form a
471 bridge, bond, or similar, and an new 'degraded-carrier' operational
472 state used for the bond or bridge master interface when one of the
473 enslaved devices is not operational.
475 * .network files learnt the new IgnoreCarrierLoss= option for leaving
476 networks configured even if the carrier is lost.
478 * The RequiredForOnline= setting in .network files may now specify a
479 minimum operational state required for the interface to be considered
480 "online" by systemd-networkd-wait-online. Related to this
481 systemd-networkd-wait-online gained a new option --operational-state=
482 to configure the same, and its --interface= option was updated to
483 optionally also take an operational state specific for an interface.
485 * systemd-networkd-wait-online gained a new setting --any for waiting
486 for only one of the requested interfaces instead of all of them.
488 * systemd-networkd now implements L2TP tunnels.
490 * Two new .network settings UseAutonomousPrefix= and UseOnLinkPrefix=
491 may be used to cause autonomous and onlink prefixes received in IPv6
492 Router Advertisements to be ignored.
494 * New MulticastFlood=, NeighborSuppression=, and Learning= .network
495 file settings may be used to tweak bridge behaviour.
497 * The new TripleSampling= option in .network files may be used to
498 configure CAN triple sampling.
500 * A new .netdev settings PrivateKeyFile= and PresharedKeyFile= may be
501 used to point to private or preshared key for a WireGuard interface.
503 * /etc/crypttab now supports the same-cpu-crypt and
504 submit-from-crypt-cpus options to tweak encryption work scheduling
507 * systemd-tmpfiles will now take a BSD file lock before operating on a
508 contents of directory. This may be used to temporarily exclude
509 directories from aging by taking the same lock (useful for example
510 when extracting a tarball into /tmp or /var/tmp as a privileged user,
511 which might create files with really old timestamps, which
512 nevertheless should not be deleted). For further details, see:
514 https://systemd.io/TEMPORARY_DIRECTORIES
516 * systemd-tmpfiles' h line type gained support for the
517 FS_PROJINHERIT_FL ('P') file attribute (introduced in kernel 4.5),
518 controlling project quota inheritance.
520 * sd-boot and bootctl now implement support for an Extended Boot Loader
521 (XBOOTLDR) partition, that is intended to be mounted to /boot, in
522 addition to the ESP partition mounted to /efi or /boot/efi.
523 Configuration file fragments, kernels, initrds and other EFI images
524 to boot will be loaded from both the ESP and XBOOTLDR partitions.
525 The XBOOTLDR partition was previously described by the Boot Loader
526 Specification, but implementation was missing in sd-boot. Support for
527 this concept allows using the sd-boot boot loader in more
528 conservative scenarios where the boot loader itself is placed in the
529 ESP but the kernels to boot (and their metadata) in a separate
532 * A system may now be booted with systemd.volatile=overlay on the
533 kernel command line, which causes the root file system to be set up
534 an overlayfs mount combining the root-only root directory with a
535 writable tmpfs. In this setup, the underlying root device is not
536 modified, and any changes are lost at reboot.
538 * Similar, systemd-nspawn can now boot containers with a volatile
539 overlayfs root with the new --volatile=overlay switch.
541 * systemd-nspawn can now consume OCI runtime bundles using a new
542 --oci-bundle= option. This implementation is fully usable, with most
543 features in the specification implemented, but since this a lot of
544 new code and functionality, this feature should most likely not
545 be used in production yet.
547 * systemd-nspawn now supports various options described by the OCI
548 runtime specification on the command-line and in .nspawn files:
549 --inaccessible=/Inaccessible= may be used to mask parts of the file
550 system tree, --console=/--pipe may be used to configure how standard
551 input, output, and error are set up.
553 * busctl learned the `emit` verb to generate D-Bus signals.
555 * systemd-analyze cat-config may be used to gather and display
556 configuration spread over multiple files, for example system and user
557 presets, tmpfiles.d, sysusers.d, udev rules, etc.
559 * systemd-analyze calendar now takes an optional new parameter
560 --iterations= which may be used to show a maximum number of iterations
561 the specified expression will elapse next.
563 * The sd-bus C API gained support for naming method parameters in the
566 * systemd-logind gained D-Bus APIs to specify the "reboot parameter"
567 the reboot() system call expects.
569 * journalctl learnt a new --cursor-file= option that points to a file
570 from which a cursor should be loaded in the beginning and to which
571 the updated cursor should be stored at the end.
573 * ACRN hypervisor and Windows Subsystem for Linux (WSL) are now
574 detected by systemd-detect-virt (and may also be used in
575 ConditionVirtualization=).
577 * The behaviour of systemd-logind may now be modified with environment
578 variables $SYSTEMD_REBOOT_TO_FIRMWARE_SETUP,
579 $SYSTEMD_REBOOT_TO_BOOT_LOADER_MENU, and
580 $SYSTEMD_REBOOT_TO_BOOT_LOADER_ENTRY. They cause logind to either
581 skip the relevant operation completely (when set to false), or to
582 create a flag file in /run/systemd (when set to true), instead of
583 actually commencing the real operation when requested. The presence
584 of /run/systemd/reboot-to-firmware-setup,
585 /run/systemd/reboot-to-boot-loader-menu, and
586 /run/systemd/reboot-to-boot-loader-entry, may be used by alternative
587 boot loader implementations to replace some steps logind performs
588 during reboot with their own operations.
590 * systemctl can be used to request a reboot into the boot loader menu
591 or a specific boot loader entry with the new --boot-load-menu= and
592 --boot-loader-entry= options to a reboot command. (This requires a
593 boot loader that supports this, for example sd-boot.)
595 * kernel-install will no longer unconditionally create the output
596 directory (e.g. /efi/<machine-id>/<kernel-version>) for boot loader
597 snippets, but will do only if the machine-specific parent directory
598 (i.e. /efi/<machine-id>/) already exists. bootctl has been modified
599 to create this parent directory during sd-boot installation.
601 This makes it easier to use kernel-install with plugins which support
602 a different layout of the bootloader partitions (for example grub2).
604 * During package installation (with `ninja install`), we would create
605 symlinks for getty@tty1.service, systemd-networkd.service,
606 systemd-networkd.socket, systemd-resolved.service,
607 remote-cryptsetup.target, remote-fs.target,
608 systemd-networkd-wait-online.service, and systemd-timesyncd.service
609 in /etc, as if `systemctl enable` was called for those units, to make
610 the system usable immediately after installation. Now this is not
611 done anymore, and instead calling `systemctl preset-all` is
612 recommended after the first installation of systemd.
614 * A new boolean sandboxing option RestrictSUIDSGID= has been added that
615 is built on seccomp. When turned on creation of SUID/SGID files is
618 * The NoNewPrivileges= and the new RestrictSUIDSGID= options are now
619 implied if DynamicUser= is turned on for a service. This hardens
620 these services, so that they neither can benefit from nor create
621 SUID/SGID executables. This is a minor compatibility breakage, given
622 that when DynamicUser= was first introduced SUID/SGID behaviour was
623 unaffected. However, the security benefit of these two options is
624 substantial, and the setting is still relatively new, hence we opted
625 to make it mandatory for services with dynamic users.
627 Contributions from: Adam Jackson, Alexander Tsoy, Andrey Yashkin,
628 Andrzej Pietrasiewicz, Anita Zhang, Balint Reczey, Beniamino Galvani,
629 Ben Iofel, Benjamin Berg, Benjamin Dahlhoff, Chris, Chris Morin,
630 Christopher Wong, Claudius Ellsel, Clemens Gruber, dana, Daniel Black,
631 Davide Cavalca, David Michael, David Rheinsberg, emersion, Evgeny
632 Vereshchagin, Filipe Brandenburger, Franck Bui, Frantisek Sumsal,
633 Giacinto Cifelli, Hans de Goede, Hugo Kindel, Ignat Korchagin, Insun
634 Pyo, Jan Engelhardt, Jonas Dorel, Jonathan Lebon, Jonathon Kowalski,
635 Jörg Sommer, Jörg Thalheim, Jussi Pakkanen, Kai-Heng Feng, Lennart
636 Poettering, Lubomir Rintel, Luís Ferreira, Martin Pitt, Matthias
637 Klumpp, Michael Biebl, Michael Niewöhner, Michael Olbrich, Michal
638 Sekletar, Mike Lothian, Paul Menzel, Piotr Drąg, Riccardo Schirone,
639 Robin Elvedi, Roman Kulikov, Ronald Tschalär, Ross Burton, Ryan
640 Gonzalez, Sebastian Krzyszkowiak, Stephane Chazelas, StKob, Susant
641 Sahani, Sylvain Plantefève, Szabolcs Fruhwald, Taro Yamada, Theo
642 Ouzhinski, Thomas Haller, Tobias Jungel, Tom Yan, Tony Asleson, Topi
643 Miettinen, unixsysadmin, Van Laser, Vesa Jääskeläinen, Yu, Li-Yu,
644 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
650 * The default locale can now be configured at compile time. Otherwise,
651 a suitable default will be selected automatically (one of C.UTF-8,
654 * The version string shown by systemd and other tools now includes the
655 git commit hash when built from git. An override may be specified
656 during compilation, which is intended to be used by distributions to
657 include the package release information.
659 * systemd-cat can now filter standard input and standard error streams
660 for different syslog priorities using the new --stderr-priority=
663 * systemd-journald and systemd-journal-remote reject entries which
664 contain too many fields (CVE-2018-16865) and set limits on the
665 process' command line length (CVE-2018-16864).
667 * $DBUS_SESSION_BUS_ADDRESS environment variable is set by pam_systemd
670 * A new network device NamePolicy "keep" is implemented for link files,
671 and used by default in 99-default.link (the fallback configuration
672 provided by systemd). With this policy, if the network device name
673 was already set by userspace, the device will not be renamed again.
674 This matches the naming scheme that was implemented before
675 systemd-240. If naming-scheme < 240 is specified, the "keep" policy
676 is also enabled by default, even if not specified. Effectively, this
677 means that if naming-scheme >= 240 is specified, network devices will
678 be renamed according to the configuration, even if they have been
679 renamed already, if "keep" is not specified as the naming policy in
680 the .link file. The 99-default.link file provided by systemd includes
681 "keep" for backwards compatibility, but it is recommended for user
682 installed .link files to *not* include it.
684 The "kernel" policy, which keeps kernel names declared to be
685 "persistent", now works again as documented.
687 * kernel-install script now optionally takes the paths to one or more
688 initrd files, and passes them to all plugins.
690 * The mincore() system call has been dropped from the @system-service
691 system call filter group, as it is pretty exotic and may potentially
692 used for side-channel attacks.
694 * -fPIE is dropped from compiler and linker options. Please specify
695 -Db_pie=true option to meson to build position-independent
696 executables. Note that the meson option is supported since meson-0.49.
698 * The fs.protected_regular and fs.protected_fifos sysctls, which were
699 added in Linux 4.19 to make some data spoofing attacks harder, are
700 now enabled by default. While this will hopefully improve the
701 security of most installations, it is technically a backwards
702 incompatible change; to disable these sysctls again, place the
703 following lines in /etc/sysctl.d/60-protected.conf or a similar file:
705 fs.protected_regular = 0
706 fs.protected_fifos = 0
708 Note that the similar hardlink and symlink protection has been
709 enabled since v199, and may be disabled likewise.
711 * The files read from the EnvironmentFile= setting in unit files now
712 parse backslashes inside quotes literally, matching the behaviour of
715 * udevadm trigger, udevadm control, udevadm settle and udevadm monitor
716 now automatically become NOPs when run in a chroot() environment.
718 * The tmpfiles.d/ "C" line type will now copy directory trees not only
719 when the destination is so far missing, but also if it already exists
720 as a directory and is empty. This is useful to cater for systems
721 where directory trees are put together from multiple separate mount
722 points but otherwise empty.
724 * A new function sd_bus_close_unref() (and the associated
725 sd_bus_close_unrefp()) has been added to libsystemd, that combines
726 sd_bus_close() and sd_bus_unref() in one.
728 * udevadm control learnt a new option for --ping for testing whether a
729 systemd-udevd instance is running and reacting.
731 * udevadm trigger learnt a new option for --wait-daemon for waiting
732 systemd-udevd daemon to be initialized.
734 Contributions from: Aaron Plattner, Alberts Muktupāvels, Alex Mayer,
735 Ayman Bagabas, Beniamino Galvani, Burt P, Chris Down, Chris Lamb, Chris
736 Morin, Christian Hesse, Claudius Ellsel, dana, Daniel Axtens, Daniele
737 Medri, Dave Reisner, David Santamaría Rogado, Diego Canuhe, Dimitri
738 John Ledkov, Evgeny Vereshchagin, Fabrice Fontaine, Filipe
739 Brandenburger, Franck Bui, Frantisek Sumsal, govwin, Hans de Goede,
740 James Hilliard, Jan Engelhardt, Jani Uusitalo, Jan Janssen, Jan
741 Synacek, Jonathan McDowell, Jonathan Roemer, Jonathon Kowalski, Joost
742 Heitbrink, Jörg Thalheim, Lance, Lennart Poettering, Louis Taylor,
743 Lucas Werkmeister, Mantas Mikulėnas, Marc-Antoine Perennou,
744 marvelousblack, Michael Biebl, Michael Sloan, Michal Sekletar, Mike
745 Auty, Mike Gilbert, Mikhail Kasimov, Neil Brown, Niklas Hambüchen,
746 Patrick Williams, Paul Seyfert, Peter Hutterer, Philip Withnall, Roger
747 James, Ronnie P. Thomas, Ryan Gonzalez, Sam Morris, Stephan Edel,
748 Stephan Gerhold, Susant Sahani, Taro Yamada, Thomas Haller, Topi
749 Miettinen, YiFei Zhu, YmrDtnJu, YunQiang Su, Yu Watanabe, Zbigniew
750 Jędrzejewski-Szmek, zsergeant77, Дамјан Георгиевски
756 * NoNewPrivileges=yes has been set for all long-running services
757 implemented by systemd. Previously, this was problematic due to
758 SELinux (as this would also prohibit the transition from PID1's label
759 to the service's label). This restriction has since been lifted, but
760 an SELinux policy update is required.
761 (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.)
763 * DynamicUser=yes is dropped from systemd-networkd.service,
764 systemd-resolved.service and systemd-timesyncd.service, which was
765 enabled in v239 for systemd-networkd.service and systemd-resolved.service,
766 and since v236 for systemd-timesyncd.service. The users and groups
767 systemd-network, systemd-resolve and systemd-timesync are created
768 by systemd-sysusers again. Distributors or system administrators
769 may need to create these users and groups if they not exist (or need
770 to re-enable DynamicUser= for those units) while upgrading systemd.
771 Also, the clock file for systemd-timesyncd may need to move from
772 /var/lib/private/systemd/timesync/clock to /var/lib/systemd/timesync/clock.
774 * When unit files are loaded from disk, previously systemd would
775 sometimes (depending on the unit loading order) load units from the
776 target path of symlinks in .wants/ or .requires/ directories of other
777 units. This meant that unit could be loaded from different paths
778 depending on whether the unit was requested explicitly or as a
779 dependency of another unit, not honouring the priority of directories
780 in search path. It also meant that it was possible to successfully
781 load and start units which are not found in the unit search path, as
782 long as they were requested as a dependency and linked to from
783 .wants/ or .requires/. The target paths of those symlinks are not
784 used for loading units anymore and the unit file must be found in
787 * A new service type has been added: Type=exec. It's very similar to
788 Type=simple but ensures the service manager will wait for both fork()
789 and execve() of the main service binary to complete before proceeding
790 with follow-up units. This is primarily useful so that the manager
791 propagates any errors in the preparation phase of service execution
792 back to the job that requested the unit to be started. For example,
793 consider a service that has ExecStart= set to a file system binary
794 that doesn't exist. With Type=simple starting the unit would be
795 considered instantly successful, as only fork() has to complete
796 successfully and the manager does not wait for execve(), and hence
797 its failure is seen "too late". With the new Type=exec service type
798 starting the unit will fail, as the manager will wait for the
799 execve() and notice its failure, which is then propagated back to the
802 NOTE: with the next release 241 of systemd we intend to change the
803 systemd-run tool to default to Type=exec for transient services
804 started by it. This should be mostly safe, but in specific corner
805 cases might result in problems, as the systemd-run tool will then
806 block on NSS calls (such as user name look-ups due to User=) done
807 between the fork() and execve(), which under specific circumstances
808 might cause problems. It is recommended to specify "-p Type=simple"
809 explicitly in the few cases where this applies. For regular,
810 non-transient services (i.e. those defined with unit files on disk)
811 we will continue to default to Type=simple.
813 * The Linux kernel's current default RLIMIT_NOFILE resource limit for
814 userspace processes is set to 1024 (soft) and 4096
815 (hard). Previously, systemd passed this on unmodified to all
816 processes it forked off. With this systemd release the hard limit
817 systemd passes on is increased to 512K, overriding the kernel's
818 defaults and substantially increasing the number of simultaneous file
819 descriptors unprivileged userspace processes can allocate. Note that
820 the soft limit remains at 1024 for compatibility reasons: the
821 traditional UNIX select() call cannot deal with file descriptors >=
822 1024 and increasing the soft limit globally might thus result in
823 programs unexpectedly allocating a high file descriptor and thus
824 failing abnormally when attempting to use it with select() (of
825 course, programs shouldn't use select() anymore, and prefer
826 poll()/epoll, but the call unfortunately remains undeservedly popular
827 at this time). This change reflects the fact that file descriptor
828 handling in the Linux kernel has been optimized in more recent
829 kernels and allocating large numbers of them should be much cheaper
830 both in memory and in performance than it used to be. Programs that
831 want to take benefit of the increased limit have to "opt-in" into
832 high file descriptors explicitly by raising their soft limit. Of
833 course, when they do that they must acknowledge that they cannot use
834 select() anymore (and neither can any shared library they use — or
835 any shared library used by any shared library they use and so on).
836 Which default hard limit is most appropriate is of course hard to
837 decide. However, given reports that ~300K file descriptors are used
838 in real-life applications we believe 512K is sufficiently high as new
839 default for now. Note that there are also reports that using very
840 high hard limits (e.g. 1G) is problematic: some software allocates
841 large arrays with one element for each potential file descriptor
842 (Java, …) — a high hard limit thus triggers excessively large memory
843 allocations in these applications. Hopefully, the new default of 512K
844 is a good middle ground: higher than what real-life applications
845 currently need, and low enough for avoid triggering excessively large
846 allocations in problematic software. (And yes, somebody should fix
849 * The fs.nr_open and fs.file-max sysctls are now automatically bumped
850 to the highest possible values, as separate accounting of file
851 descriptors is no longer necessary, as memcg tracks them correctly as
852 part of the memory accounting anyway. Thus, from the four limits on
853 file descriptors currently enforced (fs.file-max, fs.nr_open,
854 RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two,
855 and keep only the latter two. A set of build-time options
856 (-Dbump-proc-sys-fs-file-max=false and -Dbump-proc-sys-fs-nr-open=false)
857 has been added to revert this change in behaviour, which might be
858 an option for systems that turn off memcg in the kernel.
860 * When no /etc/locale.conf file exists (and hence no locale settings
861 are in place), systemd will now use the "C.UTF-8" locale by default,
862 and set LANG= to it. This locale is supported by various
863 distributions including Fedora, with clear indications that upstream
864 glibc is going to make it available too. This locale enables UTF-8
865 mode by default, which appears appropriate for 2018.
867 * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by
868 default. This effectively switches the RFC3704 Reverse Path filtering
869 from Strict mode to Loose mode. This is more appropriate for hosts
870 that have multiple links with routes to the same networks (e.g.
871 a client with a Wi-Fi and Ethernet both connected to the internet).
873 Consult the kernel documentation for details on this sysctl:
874 https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt
876 * CPUAccounting=yes no longer enables the CPU controller when using
877 kernel 4.15+ and the unified cgroup hierarchy, as required accounting
878 statistics are now provided independently from the CPU controller.
880 * Support for disabling a particular cgroup controller within a sub-tree
881 has been added through the DisableControllers= directive.
883 * cgroup_no_v1=all on the kernel command line now also implies
884 using the unified cgroup hierarchy, unless one explicitly passes
885 systemd.unified_cgroup_hierarchy=0 on the kernel command line.
887 * The new "MemoryMin=" unit file property may now be used to set the
888 memory usage protection limit of processes invoked by the unit. This
889 controls the cgroup v2 memory.min attribute. Similarly, the new
890 "IODeviceLatencyTargetSec=" property has been added, wrapping the new
891 cgroup v2 io.latency cgroup property for configuring per-service I/O
894 * systemd now supports the cgroup v2 devices BPF logic, as counterpart
895 to the cgroup v1 "devices" cgroup controller.
897 * systemd-escape now is able to combine --unescape with --template. It
898 also learnt a new option --instance for extracting and unescaping the
899 instance part of a unit name.
901 * sd-bus now provides the sd_bus_message_readv() which is similar to
902 sd_bus_message_read() but takes a va_list object. The pair
903 sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout()
904 has been added for configuring the default method call timeout to
905 use. sd_bus_error_move() may be used to efficiently move the contents
906 from one sd_bus_error structure to another, invalidating the
907 source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may
908 be used to control whether a bus connection object is automatically
909 flushed when an sd-event loop is exited.
911 * When processing classic BSD syslog log messages, journald will now
912 save the original time-stamp string supplied in the new
913 SYSLOG_TIMESTAMP= journal field. This permits consumers to
914 reconstruct the original BSD syslog message more correctly.
916 * StandardOutput=/StandardError= in service files gained support for
917 new "append:…" parameters, for connecting STDOUT/STDERR of a service
918 to a file, and appending to it.
920 * The signal to use as last step of killing of unit processes is now
921 configurable. Previously it was hard-coded to SIGKILL, which may now
922 be overridden with the new KillSignal= setting. Note that this is the
923 signal used when regular termination (i.e. SIGTERM) does not suffice.
924 Similarly, the signal used when aborting a program in case of a
925 watchdog timeout may now be configured too (WatchdogSignal=).
927 * The XDG_SESSION_DESKTOP environment variable may now be configured in
928 the pam_systemd argument line, using the new desktop= switch. This is
929 useful to initialize it properly from a display manager without
930 having to touch C code.
932 * Most configuration options that previously accepted percentage values
933 now also accept permille values with the '‰' suffix (instead of '%').
935 * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for
938 * systemd-resolved's configuration file resolved.conf gained a new
939 option ReadEtcHosts= which may be used to turn off processing and
940 honoring /etc/hosts entries.
942 * The "--wait" switch may now be passed to "systemctl
943 is-system-running", in which case the tool will synchronously wait
944 until the system finished start-up.
946 * hostnamed gained a new bus call to determine the DMI product UUID.
948 * On x86-64 systemd will now prefer using the RDRAND processor
949 instruction over /dev/urandom whenever it requires randomness that
950 neither has to be crypto-grade nor should be reproducible. This
951 should substantially reduce the amount of entropy systemd requests
952 from the kernel during initialization on such systems, though not
953 reduce it to zero. (Why not zero? systemd still needs to allocate
954 UUIDs and such uniquely, which require high-quality randomness.)
956 * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP
957 tunnels. It also gained a new option ForceDHCPv6PDOtherInformation=
958 for forcing the "Other Information" bit in IPv6 RA messages. The
959 bonding logic gained four new options AdActorSystemPriority=,
960 AdUserPortKey=, AdActorSystem= for configuring various 802.3ad
961 aspects, and DynamicTransmitLoadBalancing= for enabling dynamic
962 shuffling of flows. The tunnel logic gained a new
963 IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid
964 Deployment. The policy rule logic gained four new options IPProtocol=,
965 SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained
966 support for the MulticastToUnicast= option. networkd also gained
967 support for configuring static IPv4 ARP or IPv6 neighbor entries.
969 * .preset files (as read by 'systemctl preset') may now be used to
970 instantiate services.
972 * /etc/crypttab now understands the sector-size= option to configure
973 the sector size for an encrypted partition.
975 * Key material for encrypted disks may now be placed on a formatted
976 medium, and referenced from /etc/crypttab by the UUID of the file
977 system, followed by "=" suffixed by the path to the key file.
979 * The "collect" udev component has been removed without replacement, as
980 it is neither used nor maintained.
982 * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=,
983 LogsDirectory=, ConfigurationDirectory= settings are used in a
984 service the executed processes will now receive a set of environment
985 variables containing the full paths of these directories.
986 Specifically, RUNTIME_DIRECTORY=, STATE_DIRECTORY, CACHE_DIRECTORY,
987 LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options
988 are used. Note that these options may be used multiple times per
989 service in which case the resulting paths will be concatenated and
992 * Predictable interface naming has been extended to cover InfiniBand
993 NICs. They will be exposed with an "ib" prefix.
995 * tmpfiles.d/ line types may now be suffixed with a '-' character, in
996 which case the respective line failing is ignored.
998 * .link files may now be used to configure the equivalent to the
999 "ethtool advertise" commands.
1001 * The sd-device.h and sd-hwdb.h APIs are now exported, as an
1002 alternative to libudev.h. Previously, the latter was just an internal
1003 wrapper around the former, but now these two APIs are exposed
1006 * sd-id128.h gained a new function sd_id128_get_boot_app_specific()
1007 which calculates an app-specific boot ID similar to how
1008 sd_id128_get_machine_app_specific() generates an app-specific machine
1011 * A new tool systemd-id128 has been added that can be used to determine
1012 and generate various 128bit IDs.
1014 * /etc/os-release gained two new standardized fields DOCUMENTATION_URL=
1017 * systemd-hibernate-resume-generator will now honor the "noresume"
1018 kernel command line option, in which case it will bypass resuming
1019 from any hibernated image.
1021 * The systemd-sleep.conf configuration file gained new options
1022 AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=,
1023 AllowHybridSleep= for prohibiting specific sleep modes even if the
1024 kernel exports them.
1026 * portablectl is now officially supported and has thus moved to
1029 * bootctl learnt the two new commands "set-default" and "set-oneshot"
1030 for setting the default boot loader item to boot to (either
1031 persistently or only for the next boot). This is currently only
1032 compatible with sd-boot, but may be implemented on other boot loaders
1033 too, that follow the boot loader interface. The updated interface is
1034 now documented here:
1036 https://systemd.io/BOOT_LOADER_INTERFACE
1038 * A new kernel command line option systemd.early_core_pattern= is now
1039 understood which may be used to influence the core_pattern PID 1
1040 installs during early boot.
1042 * busctl learnt two new options -j and --json= for outputting method
1043 call replies, properties and monitoring output in JSON.
1045 * journalctl's JSON output now supports simple ANSI coloring as well as
1046 a new "json-seq" mode for generating RFC7464 output.
1048 * Unit files now support the %g/%G specifiers that resolve to the UNIX
1049 group/GID of the service manager runs as, similar to the existing
1050 %u/%U specifiers that resolve to the UNIX user/UID.
1052 * systemd-logind learnt a new global configuration option
1053 UserStopDelaySec= that may be set in logind.conf. It specifies how
1054 long the systemd --user instance shall remain started after a user
1055 logs out. This is useful to speed up repetitive re-connections of the
1056 same user, as it means the user's service manager doesn't have to be
1057 stopped/restarted on each iteration, but can be reused between
1058 subsequent options. This setting defaults to 10s. systemd-logind also
1059 exports two new properties on its Manager D-Bus objects indicating
1060 whether the system's lid is currently closed, and whether the system
1063 * systemd gained support for a generic boot counting logic, which
1064 generically permits automatic reverting to older boot loader entries
1065 if newer updated ones don't work. The boot loader side is implemented
1066 in sd-boot, but is kept open for other boot loaders too. For details
1069 https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT
1071 * The SuccessAction=/FailureAction= unit file settings now learnt two
1072 new parameters: "exit" and "exit-force", which result in immediate
1073 exiting of the service manager, and are only useful in systemd --user
1074 and container environments.
1076 * Unit files gained support for a pair of options
1077 FailureActionExitStatus=/SuccessActionExitStatus= for configuring the
1078 exit status to use as service manager exit status when
1079 SuccessAction=/FailureAction= is set to exit or exit-force.
1081 * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service
1082 options may now be used to configure the log rate limiting applied by
1083 journald per-service.
1085 * systemd-analyze gained a new verb "timespan" for parsing and
1086 normalizing time span values (i.e. strings like "5min 7s 8us").
1088 * systemd-analyze also gained a new verb "security" for analyzing the
1089 security and sand-boxing settings of services in order to determine an
1090 "exposure level" for them, indicating whether a service would benefit
1091 from more sand-boxing options turned on for them.
1093 * "systemd-analyze syscall-filter" will now also show system calls
1094 supported by the local kernel but not included in any of the defined
1097 * .nspawn files now understand the Ephemeral= setting, matching the
1098 --ephemeral command line switch.
1100 * sd-event gained the new APIs sd_event_source_get_floating() and
1101 sd_event_source_set_floating() for controlling whether a specific
1102 event source is "floating", i.e. destroyed along with the even loop
1105 * Unit objects on D-Bus gained a new "Refs" property that lists all
1106 clients that currently have a reference on the unit (to ensure it is
1109 * The JoinControllers= option in system.conf is no longer supported, as
1110 it didn't work correctly, is hard to support properly, is legacy (as
1111 the concept only exists on cgroup v1) and apparently wasn't used.
1113 * Journal messages that are generated whenever a unit enters the failed
1114 state are now tagged with a unique MESSAGE_ID. Similarly, messages
1115 generated whenever a service process exits are now made recognizable,
1116 too. A tagged message is also emitted whenever a unit enters the
1117 "dead" state on success.
1119 * systemd-run gained a new switch --working-directory= for configuring
1120 the working directory of the service to start. A shortcut -d is
1121 equivalent, setting the working directory of the service to the
1122 current working directory of the invoking program. The new --shell
1123 (or just -S) option has been added for invoking the $SHELL of the
1124 caller as a service, and implies --pty --same-dir --wait --collect
1125 --service-type=exec. Or in other words, "systemd-run -S" is now the
1126 quickest way to quickly get an interactive in a fully clean and
1127 well-defined system service context.
1129 * machinectl gained a new verb "import-fs" for importing an OS tree
1130 from a directory. Moreover, when a directory or tarball is imported
1131 and single top-level directory found with the OS itself below the OS
1132 tree is automatically mangled and moved one level up.
1134 * systemd-importd will no longer set up an implicit btrfs loop-back
1135 file system on /var/lib/machines. If one is already set up, it will
1136 continue to be used.
1138 * A new generator "systemd-run-generator" has been added. It will
1139 synthesize a unit from one or more program command lines included in
1140 the kernel command line. This is very useful in container managers
1143 # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"'
1145 This will run "systemd-nspawn" on an image, invoke the specified
1146 command line and immediately shut down the container again, returning
1147 the command line's exit code.
1149 * The block device locking logic is now documented:
1151 https://systemd.io/BLOCK_DEVICE_LOCKING
1153 * loginctl and machinectl now optionally output the various tables in
1154 JSON using the --output= switch. It is our intention to add similar
1155 support to systemctl and all other commands.
1157 * udevadm's query and trigger verb now optionally take a .device unit
1160 * systemd-udevd's network naming logic now understands a new
1161 net.naming-scheme= kernel command line switch, which may be used to
1162 pick a specific version of the naming scheme. This helps stabilizing
1163 interface names even as systemd/udev are updated and the naming logic
1166 * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and
1167 SD_ID128_ALLF to test if a 128bit ID is set to all 0xFF bytes, and to
1168 initialize one to all 0xFF.
1170 * After loading the SELinux policy systemd will now recursively relabel
1171 all files and directories listed in
1172 /run/systemd/relabel-extra.d/*.relabel (which should be simple
1173 newline separated lists of paths) in addition to the ones it already
1174 implicitly relabels in /run, /dev and /sys. After the relabelling is
1175 completed the *.relabel files (and /run/systemd/relabel-extra.d/) are
1176 removed. This is useful to permit initrds (i.e. code running before
1177 the SELinux policy is in effect) to generate files in the host
1178 filesystem safely and ensure that the correct label is applied during
1179 the transition to the host OS.
1181 * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding
1182 mknod() handling in user namespaces. Previously mknod() would always
1183 fail with EPERM in user namespaces. Since 4.18 mknod() will succeed
1184 but device nodes generated that way cannot be opened, and attempts to
1185 open them result in EPERM. This breaks the "graceful fallback" logic
1186 in systemd's PrivateDevices= sand-boxing option. This option is
1187 implemented defensively, so that when systemd detects it runs in a
1188 restricted environment (such as a user namespace, or an environment
1189 where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD)
1190 where device nodes cannot be created the effect of PrivateDevices= is
1191 bypassed (following the logic that 2nd-level sand-boxing is not
1192 essential if the system systemd runs in is itself already sand-boxed
1193 as a whole). This logic breaks with 4.18 in container managers where
1194 user namespacing is used: suddenly PrivateDevices= succeeds setting
1195 up a private /dev/ file system containing devices nodes — but when
1196 these are opened they don't work.
1198 At this point it is recommended that container managers utilizing
1199 user namespaces that intend to run systemd in the payload explicitly
1200 block mknod() with seccomp or similar, so that the graceful fallback
1203 We are very sorry for the breakage and the requirement to change
1204 container configurations for newer kernels. It's purely caused by an
1205 incompatible kernel change. The relevant kernel developers have been
1206 notified about this userspace breakage quickly, but they chose to
1209 * PermissionsStartOnly= setting is deprecated (but is still supported
1210 for backwards compatibility). The same functionality is provided by
1211 the more flexible "+", "!", and "!!" prefixes to ExecStart= and other
1214 * $DBUS_SESSION_BUS_ADDRESS environment variable is not set by
1215 pam_systemd anymore.
1217 * The naming scheme for network devices was changed to always rename
1218 devices, even if they were already renamed by userspace. The "kernel"
1219 policy was changed to only apply as a fallback, if no other naming
1222 * The requirements to build systemd is bumped to meson-0.46 and
1225 Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander
1226 Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson,
1227 Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov,
1228 asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt
1229 Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen
1230 Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius
1231 Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn
1232 Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner,
1233 David Anderson, Davide Cavalca, David Leeds, David Malcolm, David
1234 Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus,
1235 Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters,
1236 Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad,
1237 Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank
1238 Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe
1239 Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit
1240 Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan
1241 Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld,
1242 javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi
1243 Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens,
1244 Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi,
1245 Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry
1246 Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz
1247 Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier,
1248 Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin
1249 Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko
1250 Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck,
1251 Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich,
1252 Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal
1253 Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal
1254 Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby,
1255 Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł
1256 Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller,
1257 Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez,
1258 Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam
1259 Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher,
1260 Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee
1261 (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen
1262 Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim,
1263 Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas
1264 Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias
1265 Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore
1266 Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech
1267 Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward,
1268 Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe,
1269 Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein
1271 — Warsaw, 2018-12-21
1275 * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
1276 builtin will name network interfaces differently than in previous
1277 versions for virtual network interfaces created with SR-IOV and NPAR
1278 and for devices where the PCI network controller device does not have
1279 a slot number associated.
1281 SR-IOV virtual devices are now named based on the name of the parent
1282 interface, with a suffix of "v<N>", where <N> is the virtual device
1283 number. Previously those virtual devices were named as if completely
1286 The ninth and later NPAR virtual devices will be named following the
1287 scheme used for the first eight NPAR partitions. Previously those
1288 devices were not renamed and the kernel default (eth<n>) was used.
1290 "net_id" will also generate names for PCI devices where the PCI
1291 network controller device does not have an associated slot number
1292 itself, but one of its parents does. Previously those devices were
1293 not renamed and the kernel default (eth<n>) was used.
1295 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
1296 systemd-logind.service. Since v235, IPAddressDeny=any has been set to
1297 the unit. So, it is expected that the default behavior of
1298 systemd-logind is not changed. However, if distribution packagers or
1299 administrators disabled or modified IPAddressDeny= setting by a
1300 drop-in config file, then it may be necessary to update the file to
1301 re-enable AF_INET and AF_INET6 to support network user name services,
1304 * When the RestrictNamespaces= unit property is specified multiple
1305 times, then the specified types are merged now. Previously, only the
1306 last assignment was used. So, if distribution packagers or
1307 administrators modified the setting by a drop-in config file, then it
1308 may be necessary to update the file.
1310 * When OnFailure= is used in combination with Restart= on a service
1311 unit, then the specified units will no longer be triggered on
1312 failures that result in restarting. Previously, the specified units
1313 would be activated each time the unit failed, even when the unit was
1314 going to be restarted automatically. This behaviour contradicted the
1315 documentation. With this release the code is adjusted to match the
1318 * systemd-tmpfiles will now print a notice whenever it encounters
1319 tmpfiles.d/ lines referencing the /var/run/ directory. It will
1320 recommend reworking them to use the /run/ directory instead (for
1321 which /var/run/ is simply a symlinked compatibility alias). This way
1322 systemd-tmpfiles can properly detect line conflicts and merge lines
1323 referencing the same file by two paths, without having to access
1326 * systemctl disable/unmask/preset/preset-all cannot be used with
1327 --runtime. Previously this was allowed, but resulted in unintuitive
1328 behaviour that wasn't useful. systemctl disable/unmask will now undo
1329 both runtime and persistent enablement/masking, i.e. it will remove
1330 any relevant symlinks both in /run and /etc.
1332 * Note that all long-running system services shipped with systemd will
1333 now default to a system call whitelist (rather than a blacklist, as
1334 before). In particular, systemd-udevd will now enforce one too. For
1335 most cases this should be safe, however downstream distributions
1336 which disabled sandboxing of systemd-udevd (specifically the
1337 MountFlags= setting), might want to disable this security feature
1338 too, as the default whitelisting will prohibit all mount, swap,
1339 reboot and clock changing operations from udev rules.
1341 * sd-boot acquired new loader configuration settings to optionally turn
1342 off Windows and MacOS boot partition discovery as well as
1343 reboot-into-firmware menu items. It is also able to pick a better
1344 screen resolution for HiDPI systems, and now provides loader
1345 configuration settings to change the resolution explicitly.
1347 * systemd-resolved now supports DNS-over-TLS. It's still
1348 turned off by default, use DNSOverTLS=opportunistic to turn it on in
1349 resolved.conf. We intend to make this the default as soon as couple
1350 of additional techniques for optimizing the initial latency caused by
1351 establishing a TLS/TCP connection are implemented.
1353 * systemd-resolved.service and systemd-networkd.service now set
1354 DynamicUser=yes. The users systemd-resolve and systemd-network are
1355 not created by systemd-sysusers anymore.
1357 NOTE: This has a chance of breaking nss-ldap and similar NSS modules
1358 that embed a network facing module into any process using getpwuid()
1359 or related call: the dynamic allocation of the user ID for
1360 systemd-resolved.service means the service manager has to check NSS
1361 if the user name is already taken when forking off the service. Since
1362 the user in the common case won't be defined in /etc/passwd the
1363 lookup is likely to trigger nss-ldap which in turn might use NSS to
1364 ask systemd-resolved for hostname lookups. This will hence result in
1365 a deadlock: a user name lookup in order to start
1366 systemd-resolved.service will result in a host name lookup for which
1367 systemd-resolved.service needs to be started already. There are
1368 multiple ways to work around this problem: pre-allocate the
1369 "systemd-resolve" user on such systems, so that nss-ldap won't be
1370 triggered; or use a different NSS package that doesn't do networking
1371 in-process but provides a local asynchronous name cache; or configure
1372 the NSS package to avoid lookups for UIDs in the range `pkg-config
1373 systemd --variable=dynamicuidmin` … `pkg-config systemd
1374 --variable=dynamicuidmax`, so that it does not consider itself
1375 authoritative for the same UID range systemd allocates dynamic users
1378 * The systemd-resolve tool has been renamed to resolvectl (it also
1379 remains available under the old name, for compatibility), and its
1380 interface is now verb-based, similar in style to the other <xyz>ctl
1381 tools, such as systemctl or loginctl.
1383 * The resolvectl/systemd-resolve tool also provides 'resolvconf'
1384 compatibility. It may be symlinked under the 'resolvconf' name, in
1385 which case it will take arguments and input compatible with the
1386 Debian and FreeBSD resolvconf tool.
1388 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
1389 where the system initially suspends, and after a timeout resumes and
1392 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
1393 set the client will only send a DUID as client identifier.
1395 * The nss-systemd glibc NSS module will now enumerate dynamic users and
1396 groups in effect. Previously, it could resolve UIDs/GIDs to user
1397 names/groups and vice versa, but did not support enumeration.
1399 * journald's Compress= configuration setting now optionally accepts a
1400 byte threshold value. All journal objects larger than this threshold
1401 will be compressed, smaller ones will not. Previously this threshold
1402 was not configurable and set to 512.
1404 * A new system.conf setting NoNewPrivileges= is now available which may
1405 be used to turn off acquisition of new privileges system-wide
1406 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
1407 for all its children). Note that turning this option on means setuid
1408 binaries and file system capabilities lose their special powers.
1409 While turning on this option is a big step towards a more secure
1410 system, doing so is likely to break numerous pre-existing UNIX tools,
1411 in particular su and sudo.
1413 * A new service systemd-time-sync-wait.service has been added. If
1414 enabled it will delay the time-sync.target unit at boot until time
1415 synchronization has been received from the network. This
1416 functionality is useful on systems lacking a local RTC or where it is
1417 acceptable that the boot process shall be delayed by external network
1420 * When hibernating, systemd will now inform the kernel of the image
1421 write offset, on kernels new enough to support this. This means swap
1422 files should work for hibernation now.
1424 * When loading unit files, systemd will now look for drop-in unit files
1425 extensions in additional places. Previously, for a unit file name
1426 "foo-bar-baz.service" it would look for dropin files in
1427 "foo-bar-baz.service.d/*.conf". Now, it will also look in
1428 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
1429 service name truncated after all inner dashes. This scheme allows
1430 writing drop-ins easily that apply to a whole set of unit files at
1431 once. It's particularly useful for mount and slice units (as their
1432 naming is prefix based), but is also useful for service and other
1433 units, for packages that install multiple unit files at once,
1434 following a strict naming regime of beginning the unit file name with
1435 the package's name. Two new specifiers are now supported in unit
1436 files to match this: %j and %J are replaced by the part of the unit
1437 name following the last dash.
1439 * Unit files and other configuration files that support specifier
1440 expansion now understand another three new specifiers: %T and %V will
1441 resolve to /tmp and /var/tmp respectively, or whatever temporary
1442 directory has been set for the calling user. %E will expand to either
1443 /etc (for system units) or $XDG_CONFIG_HOME (for user units).
1445 * The ExecStart= lines of unit files are no longer required to
1446 reference absolute paths. If non-absolute paths are specified the
1447 specified binary name is searched within the service manager's
1448 built-in $PATH, which may be queried with 'systemd-path
1449 search-binaries-default'. It's generally recommended to continue to
1450 use absolute paths for all binaries specified in unit files.
1452 * Units gained a new load state "bad-setting", which is used when a
1453 unit file was loaded, but contained fatal errors which prevent it
1454 from being started (for example, a service unit has been defined
1455 lacking both ExecStart= and ExecStop= lines).
1457 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
1458 support alternative debuggers, for example lldb. The old name
1459 continues to be available however, for compatibility reasons. Use the
1460 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
1461 to pick an alternative debugger instead of the default gdb.
1463 * systemctl and the other tools will now output escape sequences that
1464 generate proper clickable hyperlinks in various terminal emulators
1465 where useful (for example, in the "systemctl status" output you can
1466 now click on the unit file name to quickly open it in the
1467 editor/viewer of your choice). Note that not all terminal emulators
1468 support this functionality yet, but many do. Unfortunately, the
1469 "less" pager doesn't support this yet, hence this functionality is
1470 currently automatically turned off when a pager is started (which
1471 happens quite often due to auto-paging). We hope to remove this
1472 limitation as soon as "less" learns these escape sequences. This new
1473 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
1474 environment variable. For details on these escape sequences see:
1475 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
1477 * networkd's .network files now support a new IPv6MTUBytes= option for
1478 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
1479 option in the [Route] section to configure the MTU to use for
1480 specific routes. It also gained support for configuration of the DHCP
1481 "UserClass" option through the new UserClass= setting. It gained
1482 three new options in the new [CAN] section for configuring CAN
1483 networks. The MULTICAST and ALLMULTI interface flags may now be
1484 controlled explicitly with the new Multicast= and AllMulticast=
1487 * networkd will now automatically make use of the kernel's route
1488 expiration feature, if it is available.
1490 * udevd's .link files now support setting the number of receive and
1491 transmit channels, using the RxChannels=, TxChannels=,
1492 OtherChannels=, CombinedChannels= settings.
1494 * Support for UDPSegmentationOffload= has been removed, given its
1495 limited support in hardware, and waning software support.
1497 * networkd's .netdev files now support creating "netdevsim" interfaces.
1499 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
1500 to query the unit belonging to a specific kernel control group.
1502 * systemd-analyze gained a new verb "cat-config", which may be used to
1503 dump the contents of any configuration file, with all its matching
1504 drop-in files added in, and honouring the usual search and masking
1505 logic applied to systemd configuration files. For example use
1506 "systemd-analyze cat-config systemd/system.conf" to get the complete
1507 system configuration file of systemd how it would be loaded by PID 1
1508 itself. Similar to this, various tools such as systemd-tmpfiles or
1509 systemd-sysusers, gained a new option "--cat-config", which does the
1510 corresponding operation for their own configuration settings. For
1511 example, "systemd-tmpfiles --cat-config" will now output the full
1512 list of tmpfiles.d/ lines in place.
1514 * timedatectl gained three new verbs: "show" shows bus properties of
1515 systemd-timedated, "timesync-status" shows the current NTP
1516 synchronization state of systemd-timesyncd, and "show-timesync"
1517 shows bus properties of systemd-timesyncd.
1519 * systemd-timesyncd gained a bus interface on which it exposes details
1522 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
1523 understood by systemd-timedated. It takes a colon-separated list of
1524 unit names of NTP client services. The list is used by
1525 "timedatectl set-ntp".
1527 * systemd-nspawn gained a new --rlimit= switch for setting initial
1528 resource limits for the container payload. There's a new switch
1529 --hostname= to explicitly override the container's hostname. A new
1530 --no-new-privileges= switch may be used to control the
1531 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
1532 --oom-score-adjust= switch controls the OOM scoring adjustment value
1533 for the payload. The new --cpu-affinity= switch controls the CPU
1534 affinity of the container payload. The new --resolv-conf= switch
1535 allows more detailed control of /etc/resolv.conf handling of the
1536 container. Similarly, the new --timezone= switch allows more detailed
1537 control of /etc/localtime handling of the container.
1539 * systemd-detect-virt gained a new --list switch, which will print a
1540 list of all currently known VM and container environments.
1542 * Support for "Portable Services" has been added, see
1543 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
1544 experimental, but this is expected to change soon. Reflecting this
1545 experimental state, the "portablectl" binary is not installed into
1546 /usr/bin yet. The binary has to be called with the full path
1547 /usr/lib/systemd/portablectl instead.
1549 * journalctl's and systemctl's -o switch now knows a new log output
1550 mode "with-unit". The output it generates is very similar to the
1551 regular "short" mode, but displays the unit name instead of the
1552 syslog tag for each log line. Also, the date is shown with timezone
1553 information. This mode is probably more useful than the classic
1554 "short" output mode for most purposes, except where pixel-perfect
1555 compatibility with classic /var/log/messages formatting is required.
1557 * A new --dump-bus-properties switch has been added to the systemd
1558 binary, which may be used to dump all supported D-Bus properties.
1559 (Options which are still supported, but are deprecated, are *not*
1562 * sd-bus gained a set of new calls:
1563 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
1564 enable/disable the "floating" state of a bus slot object,
1565 i.e. whether the slot object pins the bus it is allocated for into
1566 memory or if the bus slot object gets disconnected when the bus goes
1567 away. sd_bus_open_with_description(),
1568 sd_bus_open_user_with_description(),
1569 sd_bus_open_system_with_description() may be used to allocate bus
1570 objects and set their description string already during allocation.
1572 * sd-event gained support for watching inotify events from the event
1573 loop, in an efficient way, sharing inotify handles between multiple
1574 users. For this a new function sd_event_add_inotify() has been added.
1576 * sd-event and sd-bus gained support for calling special user-supplied
1577 destructor functions for userdata pointers associated with
1578 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
1579 functions sd_bus_slot_set_destroy_callback,
1580 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
1581 sd_bus_track_get_destroy_callback,
1582 sd_event_source_set_destroy_callback,
1583 sd_event_source_get_destroy_callback have been added.
1585 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
1587 * PID 1 will now automatically reschedule .timer units whenever the
1588 local timezone changes. (They previously got rescheduled
1589 automatically when the system clock changed.)
1591 * New documentation has been added to document cgroups delegation,
1592 portable services and the various code quality tools we have set up:
1594 https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md
1595 https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md
1596 https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md
1598 * The Boot Loader Specification has been added to the source tree.
1600 https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md
1602 While moving it into our source tree we have updated it and further
1603 changes are now accepted through the usual github PR workflow.
1605 * pam_systemd will now look for PAM userdata fields systemd.memory_max,
1606 systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
1607 earlier PAM modules. The data in these fields is used to initialize
1608 the session scope's resource properties. Thus external PAM modules
1609 may now configure per-session limits, for example sourced from
1610 external user databases.
1612 * socket units with Accept=yes will now maintain a "refused" counter in
1613 addition to the existing "accepted" counter, counting connections
1614 refused due to the enforced limits.
1616 * The "systemd-path search-binaries-default" command may now be use to
1617 query the default, built-in $PATH PID 1 will pass to the services it
1620 * A new unit file setting PrivateMounts= has been added. It's a boolean
1621 option. If enabled the unit's processes are invoked in their own file
1622 system namespace. Note that this behaviour is also implied if any
1623 other file system namespacing options (such as PrivateTmp=,
1624 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
1625 primarily useful for services that do not use any of the other file
1626 system namespacing options. One such service is systemd-udevd.service
1627 where this is now used by default.
1629 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
1630 when the system is booted in UEFI "secure mode".
1632 * A new unit "system-update-pre.target" is added, which defines an
1633 optional synchronization point for offline system updates, as
1634 implemented by the pre-existing "system-update.target" unit. It
1635 allows ordering services before the service that executes the actual
1636 update process in a generic way.
1638 * Systemd now emits warnings whenever .include syntax is used.
1640 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
1641 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
1642 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
1643 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
1644 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
1645 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
1646 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
1647 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
1648 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
1649 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
1650 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
1651 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
1652 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
1653 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
1654 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
1655 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
1656 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
1657 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
1658 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
1659 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
1660 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
1661 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
1662 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
1663 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
1664 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
1665 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
1666 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
1667 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
1668 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
1670 — Berlin, 2018-06-22
1674 * The MemoryAccounting= unit property now defaults to on. After
1675 discussions with the upstream control group maintainers we learnt
1676 that the negative impact of cgroup memory accounting on current
1677 kernels is finally relatively minimal, so that it should be safe to
1678 enable this by default without affecting system performance. Besides
1679 memory accounting only task accounting is turned on by default, all
1680 other forms of resource accounting (CPU, IO, IP) remain off for now,
1681 because it's not clear yet that their impact is small enough to move
1682 from opt-in to opt-out. We recommend downstreams to leave memory
1683 accounting on by default if kernel 4.14 or higher is primarily
1684 used. On very resource constrained systems or when support for old
1685 kernels is a necessity, -Dmemory-accounting-default=false can be used
1686 to revert this change.
1688 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
1689 %udev_rules_update) and the journal catalog (%journal_catalog_update)
1690 from the upgrade scriptlets of individual packages now do nothing.
1691 Transfiletriggers have been added which will perform those updates
1692 once at the end of the transaction.
1694 Similar transfiletriggers have been added to execute any sysctl.d
1695 and binfmt.d rules. Thus, it should be unnecessary to provide any
1696 scriptlets to execute this configuration from package installation
1699 * systemd-sysusers gained a mode where the configuration to execute is
1700 specified on the command line, but this configuration is not executed
1701 directly, but instead it is merged with the configuration on disk,
1702 and the result is executed. This is useful for package installation
1703 scripts which want to create the user before installing any files on
1704 disk (in case some of those files are owned by that user), while
1705 still allowing local admin overrides.
1707 This functionality is exposed to rpm scriptlets through a new
1708 %sysusers_create_package macro. Old %sysusers_create and
1709 %sysusers_create_inline macros are deprecated.
1711 A transfiletrigger for sysusers.d configuration is now installed,
1712 which means that it should be unnecessary to call systemd-sysusers from
1713 package installation scripts, unless the package installs any files
1714 owned by those newly-created users, in which case
1715 %sysusers_create_package should be used.
1717 * Analogous change has been done for systemd-tmpfiles: it gained a mode
1718 where the command-line configuration is merged with the configuration
1719 on disk. This is exposed as the new %tmpfiles_create_package macro,
1720 and %tmpfiles_create is deprecated. A transfiletrigger is installed
1721 for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
1722 from package installation scripts.
1724 * sysusers.d configuration for a user may now also specify the group
1725 number, in addition to the user number ("u username 123:456"), or
1726 without the user number ("u username -:456").
1728 * Configution items for systemd-sysusers can now be specified as
1729 positional arguments when the new --inline switch is used.
1731 * The login shell of users created through sysusers.d may now be
1732 specified (previously, it was always /bin/sh for root and
1733 /sbin/nologin for other users).
1735 * systemd-analyze gained a new --global switch to look at global user
1736 configuration. It also gained a unit-paths verb to list the unit load
1737 paths that are compiled into systemd (which can be used with
1738 --systemd, --user, or --global).
1740 * udevadm trigger gained a new --settle/-w option to wait for any
1741 triggered events to finish (but just those, and not any other events
1742 which are triggered meanwhile).
1744 * The action that systemd-logind takes when the lid is closed and the
1745 machine is connected to external power can now be configured using
1746 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
1747 was determined by HandleLidSwitch=, and, for backwards compatibility,
1748 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
1750 * journalctl will periodically call sd_journal_process() to make it
1751 resilient against inotify queue overruns when journal files are
1752 rotated very quickly.
1754 * Two new functions in libsystemd — sd_bus_get_n_queued_read and
1755 sd_bus_get_n_queued_write — may be used to check the number of
1756 pending bus messages.
1758 * systemd gained a new
1759 org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
1760 which can be used to migrate foreign processes to scope and service
1761 units. The primary user for this new API is systemd itself: the
1762 systemd --user instance uses this call of the systemd --system
1763 instance to migrate processes if it itself gets the request to
1764 migrate processes and the kernel refuses this due to access
1765 restrictions. Thanks to this "systemd-run --scope --user …" works
1766 again in pure cgroup v2 environments when invoked from the user
1769 * A new TemporaryFileSystem= setting can be used to mask out part of
1770 the real file system tree with tmpfs mounts. This may be combined
1771 with BindPaths= and BindReadOnlyPaths= to hide files or directories
1772 not relevant to the unit, while still allowing some paths lower in
1773 the tree to be accessed.
1775 ProtectHome=tmpfs may now be used to hide user home and runtime
1776 directories from units, in a way that is mostly equivalent to
1777 "TemporaryFileSystem=/home /run/user /root".
1779 * Non-service units are now started with KeyringMode=shared by default.
1780 This means that mount and swapon and other mount tools have access
1781 to keys in the main keyring.
1783 * /sys/fs/bpf is now mounted automatically.
1785 * QNX virtualization is now detected by systemd-detect-virt and may
1786 be used in ConditionVirtualization=.
1788 * IPAccounting= may now be enabled also for slice units.
1790 * A new -Dsplit-bin= build configuration switch may be used to specify
1791 whether bin and sbin directories are merged, or if they should be
1792 included separately in $PATH and various listings of executable
1793 directories. The build configuration scripts will try to autodetect
1794 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
1795 system, but distributions are encouraged to configure this
1798 * A new -Dok-color= build configuration switch may be used to change
1799 the colour of "OK" status messages.
1801 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
1802 PrivateNetwork=yes was buggy in previous versions of systemd. This
1803 means that after the upgrade and daemon-reexec, any such units must
1806 * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
1807 will not exclude read-only files owned by root from cleanup.
1809 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
1810 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
1811 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
1812 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
1813 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
1814 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
1815 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
1816 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
1817 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
1818 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
1819 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
1820 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
1821 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
1822 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
1823 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
1824 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
1826 — Warsaw, 2018-03-05
1830 * Some keyboards come with a zoom see-saw or rocker which until now got
1831 mapped to the Linux "zoomin/out" keys in hwdb. However, these
1832 keycodes are not recognized by any major desktop. They now produce
1833 Up/Down key events so that they can be used for scrolling.
1835 * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
1836 slightly: previously, if an argument was specified for lines of this
1837 type (i.e. the right-most column was set) this string was appended to
1838 existing files each time systemd-tmpfiles was run. This behaviour was
1839 different from what the documentation said, and not particularly
1840 useful, as repeated systemd-tmpfiles invocations would not be
1841 idempotent and grow such files without bounds. With this release
1842 behaviour has been altered to match what the documentation says:
1843 lines of this type only have an effect if the indicated files don't
1844 exist yet, and only then the argument string is written to the file.
1846 * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
1847 systemd-tmpfiles behaviour: previously, read-only files owned by root
1848 were always excluded from the file "aging" algorithm (i.e. the
1849 automatic clean-up of directories like /tmp based on
1850 atime/mtime/ctime). We intend to drop this restriction, and age files
1851 by default even when owned by root and read-only. This behaviour was
1852 inherited from older tools, but there have been requests to remove
1853 it, and it's not obvious why this restriction was made in the first
1854 place. Please speak up now, if you are aware of software that reqires
1855 this behaviour, otherwise we'll remove the restriction in v238.
1857 * A new environment variable $SYSTEMD_OFFLINE is now understood by
1858 systemctl. It takes a boolean argument. If on, systemctl assumes it
1859 operates on an "offline" OS tree, and will not attempt to talk to the
1860 service manager. Previously, this mode was implicitly enabled if a
1861 chroot() environment was detected, and this new environment variable
1862 now provides explicit control.
1864 * .path and .socket units may now be created transiently, too.
1865 Previously only service, mount, automount and timer units were
1866 supported as transient units. The systemd-run tool has been updated
1867 to expose this new functionality, you may hence use it now to bind
1868 arbitrary commands to path or socket activation on-the-fly from the
1869 command line. Moreover, almost all properties are now exposed for the
1870 unit types that already supported transient operation.
1872 * The systemd-mount command gained support for a new --owner= parameter
1873 which takes a user name, which is then resolved and included in uid=
1874 and gid= mount options string of the file system to mount.
1876 * A new unit condition ConditionControlGroupController= has been added
1877 that checks whether a specific cgroup controller is available.
1879 * Unit files, udev's .link files, and systemd-networkd's .netdev and
1880 .network files all gained support for a new condition
1881 ConditionKernelVersion= for checking against specific kernel
1884 * In systemd-networkd, the [IPVLAN] section in .netdev files gained
1885 support for configuring device flags in the Flags= setting. In the
1886 same files, the [Tunnel] section gained support for configuring
1887 AllowLocalRemote=. The [Route] section in .network files gained
1888 support for configuring InitialCongestionWindow=,
1889 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
1890 understands RapidCommit=.
1892 * systemd-networkd's DHCPv6 support gained support for Prefix
1895 * sd-bus gained support for a new "watch-bind" feature. When this
1896 feature is enabled, an sd_bus connection may be set up to connect to
1897 an AF_UNIX socket in the file system as soon as it is created. This
1898 functionality is useful for writing early-boot services that
1899 automatically connect to the system bus as soon as it is started,
1900 without ugly time-based polling. systemd-networkd and
1901 systemd-resolved have been updated to make use of this
1902 functionality. busctl exposes this functionality in a new
1903 --watch-bind= command line switch.
1905 * sd-bus will now optionally synthesize a local "Connected" signal as
1906 soon as a D-Bus connection is set up fully. This message mirrors the
1907 already existing "Disconnected" signal which is synthesized when the
1908 connection is terminated. This signal is generally useful but
1909 particularly handy in combination with the "watch-bind" feature
1910 described above. Synthesizing of this message has to be requested
1911 explicitly through the new API call sd_bus_set_connected_signal(). In
1912 addition a new call sd_bus_is_ready() has been added that checks
1913 whether a connection is fully set up (i.e. between the "Connected" and
1914 "Disconnected" signals).
1916 * sd-bus gained two new calls sd_bus_request_name_async() and
1917 sd_bus_release_name_async() for asynchronously registering bus
1918 names. Similar, there is now sd_bus_add_match_async() for installing
1919 a signal match asynchronously. All of systemd's own services have
1920 been updated to make use of these calls. Doing these operations
1921 asynchronously has two benefits: it reduces the risk of deadlocks in
1922 case of cyclic dependencies between bus services, and it speeds up
1923 service initialization since synchronization points for bus
1924 round-trips are removed.
1926 * sd-bus gained two new calls sd_bus_match_signal() and
1927 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
1928 and sd_bus_add_match_async() but instead of taking a D-Bus match
1929 string take match fields as normal function parameters.
1931 * sd-bus gained two new calls sd_bus_set_sender() and
1932 sd_bus_message_set_sender() for setting the sender name of outgoing
1933 messages (either for all outgoing messages or for just one specific
1934 one). These calls are only useful in direct connections as on
1935 brokered connections the broker fills in the sender anyway,
1936 overwriting whatever the client filled in.
1938 * sd-event gained a new pseudo-handle that may be specified on all API
1939 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
1940 used this refers to the default event loop object of the calling
1941 thread. Note however that this does not implicitly allocate one —
1942 which has to be done prior by using sd_event_default(). Similarly
1943 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
1944 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
1945 to the default bus of the specified type of the calling thread. Here
1946 too this does not implicitly allocate bus connection objects, this
1947 has to be done prior with sd_bus_default() and friends.
1949 * sd-event gained a new call pair
1950 sd_event_source_{get|set}_io_fd_own(). This may be used to request
1951 automatic closure of the file descriptor an IO event source watches
1952 when the event source is destroyed.
1954 * systemd-networkd gained support for natively configuring WireGuard
1957 * In previous versions systemd synthesized user records both for the
1958 "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
1959 internally. In order to simplify distribution-wide renames of the
1960 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
1961 new transitional flag file has been added: if
1962 /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
1963 user and group record within the systemd codebase is disabled.
1965 * systemd-notify gained a new --uid= option for selecting the source
1966 user/UID to use for notification messages sent to the service
1969 * journalctl gained a new --grep= option to list only entries in which
1970 the message matches a certain pattern. By default matching is case
1971 insensitive if the pattern is lowercase, and case sensitive
1972 otherwise. Option --case-sensitive=yes|no can be used to override
1973 this an specify case sensitivity or case insensitivity.
1975 * There's now a "systemd-analyze service-watchdogs" command for printing
1976 the current state of the service runtime watchdog, and optionally
1977 enabling or disabling the per-service watchdogs system-wide if given a
1978 boolean argument (i.e. the concept you configure in WatchdogSec=), for
1979 debugging purposes. There's also a kernel command line option
1980 systemd.service_watchdogs= for controlling the same.
1982 * Two new "log-level" and "log-target" options for systemd-analyze were
1983 added that merge the now deprecated get-log-level, set-log-level and
1984 get-log-target, set-log-target pairs. The deprecated options are still
1985 understood for backwards compatibility. The two new options print the
1986 current value when no arguments are given, and set them when a
1987 level/target is given as an argument.
1989 * sysusers.d's "u" lines now optionally accept both a UID and a GID
1990 specification, separated by a ":" character, in order to create users
1991 where UID and GID do not match.
1993 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
1994 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
1995 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
1996 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
1997 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
1998 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
1999 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
2000 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
2001 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
2002 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
2003 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
2004 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
2005 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
2006 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
2007 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
2008 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
2009 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
2010 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
2011 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
2012 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
2019 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
2020 in v235 has been extended to also set the dummy.ko module option
2021 numdummies=0, preventing the kernel from automatically creating
2022 dummy0. All dummy interfaces must now be explicitly created.
2024 * Unknown '%' specifiers in configuration files are now rejected. This
2025 applies to units and tmpfiles.d configuration. Any percent characters
2026 that are followed by a letter or digit that are not supposed to be
2027 interpreted as the beginning of a specifier should be escaped by
2028 doubling ("%%"). (So "size=5%" is still accepted, as well as
2029 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
2030 valid specifiers today.)
2032 * systemd-resolved now maintains a new dynamic
2033 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
2034 recommended to make /etc/resolv.conf a symlink to it. This file
2035 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
2036 includes dynamically acquired search domains, achieving more correct
2037 DNS resolution by software that bypasses local DNS APIs such as NSS.
2039 * The "uaccess" udev tag has been dropped from /dev/kvm and
2040 /dev/dri/renderD*. These devices now have the 0666 permissions by
2041 default (but this may be changed at build-time). /dev/dri/renderD*
2042 will now be owned by the "render" group along with /dev/kfd.
2044 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
2045 systemd-journal-gatewayd.service and
2046 systemd-journal-upload.service. This means "nss-systemd" must be
2047 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
2048 services are resolved properly.
2050 * In /etc/fstab two new mount options are now understood:
2051 x-systemd.makefs and x-systemd.growfs. The former has the effect that
2052 the configured file system is formatted before it is mounted, the
2053 latter that the file system is resized to the full block device size
2054 after it is mounted (i.e. if the file system is smaller than the
2055 partition it resides on, it's grown). This is similar to the fsck
2056 logic in /etc/fstab, and pulls in systemd-makefs@.service and
2057 systemd-growfs@.service as necessary, similar to
2058 systemd-fsck@.service. Resizing is currently only supported on ext4
2061 * In systemd-networkd, the IPv6 RA logic now optionally may announce
2062 DNS server and domain information.
2064 * Support for the LUKS2 on-disk format for encrypted partitions has
2065 been added. This requires libcryptsetup2 during compilation and
2068 * The systemd --user instance will now signal "readiness" when its
2069 basic.target unit has been reached, instead of when the run queue ran
2070 empty for the first time.
2072 * Tmpfiles.d with user configuration are now also supported.
2073 systemd-tmpfiles gained a new --user switch, and snippets placed in
2074 ~/.config/user-tmpfiles.d/ and corresponding directories will be
2075 executed by systemd-tmpfiles --user running in the new
2076 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
2077 running in the user session.
2079 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
2080 %S resolves to the top-level state directory (/var/lib for the system
2081 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
2082 top-level cache directory (/var/cache for the system instance,
2083 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
2084 logs directory (/var/log for the system instance,
2085 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
2086 existing %t specifier, that resolves to the top-level runtime
2087 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
2090 * journalctl learnt a new parameter --output-fields= for limiting the
2091 set of journal fields to output in verbose and JSON output modes.
2093 * systemd-timesyncd's configuration file gained a new option
2094 RootDistanceMaxSec= for setting the maximum root distance of servers
2095 it'll use, as well as the new options PollIntervalMinSec= and
2096 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
2098 * bootctl gained a new command "list" for listing all available boot
2099 menu items on systems that follow the boot loader specification.
2101 * systemctl gained a new --dry-run switch that shows what would be done
2102 instead of doing it, and is currently supported by the shutdown and
2105 * ConditionSecurity= can now detect the TOMOYO security module.
2107 * Unit file [Install] sections are now also respected in unit drop-in
2108 files. This is intended to be used by drop-ins under /usr/lib/.
2110 * systemd-firstboot may now also set the initial keyboard mapping.
2112 * Udev "changed" events for devices which are exposed as systemd
2113 .device units are now propagated to units specified in
2114 ReloadPropagatedFrom= as reload requests.
2116 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
2117 unit template name (i.e. a name in the form of 'foobar@.service',
2118 without the instance component between the '@' and - the '.'), then
2119 the escaped sysfs path of the device is automatically used as the
2122 * SystemCallFilter= in unit files has been extended so that an "errno"
2123 can be specified individually for each system call. Example:
2124 SystemCallFilter=~uname:EILSEQ.
2126 * The cgroup delegation logic has been substantially updated. Delegate=
2127 now optionally takes a list of controllers (instead of a boolean, as
2128 before), which lists the controllers to delegate at least.
2130 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
2132 * A new LogLevelMax= setting configures the maximum log level any
2133 process of the service may log at (i.e. anything with a lesser
2134 priority than what is specified is automatically dropped). A new
2135 LogExtraFields= setting allows configuration of additional journal
2136 fields to attach to all log records generated by any of the unit's
2139 * New StandardInputData= and StandardInputText= settings along with the
2140 new option StandardInput=data may be used to configure textual or
2141 binary data that shall be passed to the executed service process via
2142 standard input, encoded in-line in the unit file.
2144 * StandardInput=, StandardOutput= and StandardError= may now be used to
2145 connect stdin/stdout/stderr of executed processes directly with a
2146 file or AF_UNIX socket in the file system, using the new "file:" option.
2148 * A new unit file option CollectMode= has been added, that allows
2149 tweaking the garbage collection logic for units. It may be used to
2150 tell systemd to garbage collect units that have failed automatically
2151 (normally it only GCs units that exited successfully). systemd-run
2152 and systemd-mount expose this new functionality with a new -G option.
2154 * "machinectl bind" may now be used to bind mount non-directories
2155 (i.e. regularfiles, devices, fifos, sockets).
2157 * systemd-analyze gained a new verb "calendar" for validating and
2158 testing calendar time specifications to use for OnCalendar= in timer
2159 units. Besides validating the expression it will calculate the next
2160 time the specified expression would elapse.
2162 * In addition to the pre-existing FailureAction= unit file setting
2163 there's now SuccessAction=, for configuring a shutdown action to
2164 execute when a unit completes successfully. This is useful in
2165 particular inside containers that shall terminate after some workload
2166 has been completed. Also, both options are now supported for all unit
2167 types, not just services.
2169 * networkds's IP rule support gained two new options
2170 IncomingInterface= and OutgoingInterface= for configuring the incoming
2171 and outgoing interfaces of configured rules. systemd-networkd also
2172 gained support for "vxcan" network devices.
2174 * networkd gained a new setting RequiredForOnline=, taking a
2175 boolean. If set, systemd-wait-online will take it into consideration
2176 when determining that the system is up, otherwise it will ignore the
2177 interface for this purpose.
2179 * The sd_notify() protocol gained support for a new operation: with
2180 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
2181 store again, ahead of POLLHUP or POLLERR when they are removed
2184 * A new document doc/UIDS-GIDS.md has been added to the source tree,
2185 that documents the UID/GID range and assignment assumptions and
2186 requirements of systemd.
2188 * The watchdog device PID 1 will ping may now be configured through the
2189 WatchdogDevice= configuration file setting, or by setting the
2190 systemd.watchdog_service= kernel commandline option.
2192 * systemd-resolved's gained support for registering DNS-SD services on
2193 the local network using MulticastDNS. Services may either be
2194 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
2195 the same dir below /run, /usr/lib), or through its D-Bus API.
2197 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
2198 extend the effective start, runtime, and stop time. The service must
2199 continue to send EXTEND_TIMEOUT_USEC within the period specified to
2200 prevent the service manager from making the service as timedout.
2202 * systemd-resolved's DNSSEC support gained support for RFC 8080
2203 (Ed25519 keys and signatures).
2205 * The systemd-resolve command line tool gained a new set of options
2206 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
2207 --set-nta= and --revert to configure per-interface DNS configuration
2208 dynamically during runtime. It's useful for pushing DNS information
2209 into systemd-resolved from DNS hook scripts that various interface
2210 managing software supports (such as pppd).
2212 * systemd-nspawn gained a new --network-namespace-path= command line
2213 option, which may be used to make a container join an existing
2214 network namespace, by specifying a path to a "netns" file.
2216 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
2217 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
2218 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
2219 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
2220 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
2221 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
2222 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
2223 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
2224 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
2225 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
2226 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
2227 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
2228 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
2229 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
2230 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
2231 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
2232 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
2233 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
2234 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
2235 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
2236 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
2237 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
2238 Jędrzejewski-Szmek, Zeal Jagannatha
2240 — Berlin, 2017-12-14
2244 * INCOMPATIBILITY: systemd-logind.service and other long-running
2245 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
2246 communication with the outside. This generally improves security of
2247 the system, and is in almost all cases a safe and good choice, as
2248 these services do not and should not provide any network-facing
2249 functionality. However, systemd-logind uses the glibc NSS API to
2250 query the user database. This creates problems on systems where NSS
2251 is set up to directly consult network services for user database
2252 lookups. In particular, this creates incompatibilities with the
2253 "nss-nis" module, which attempts to directly contact the NIS/YP
2254 network servers it is configured for, and will now consistently
2255 fail. In such cases, it is possible to turn off IP sandboxing for
2256 systemd-logind.service (set IPAddressDeny= in its [Service] section
2257 to the empty string, via a .d/ unit file drop-in). Downstream
2258 distributions might want to update their nss-nis packaging to include
2259 such a drop-in snippet, accordingly, to hide this incompatibility
2260 from the user. Another option is to make use of glibc's nscd service
2261 to proxy such network requests through a privilege-separated, minimal
2262 local caching daemon, or to switch to more modern technologies such
2263 sssd, whose NSS hook-ups generally do not involve direct network
2264 access. In general, we think it's definitely time to question the
2265 implementation choices of nss-nis, i.e. whether it's a good idea
2266 today to embed a network-facing loadable module into all local
2267 processes that need to query the user database, including the most
2268 trivial and benign ones, such as "ls". For more details about
2269 IPAddressDeny= see below.
2271 * A new modprobe.d drop-in is now shipped by default that sets the
2272 bonding module option max_bonds=0. This overrides the kernel default,
2273 to avoid conflicts and ambiguity as to whether or not bond0 should be
2274 managed by systemd-networkd or not. This resolves multiple issues
2275 with bond0 properties not being applied, when bond0 is configured
2276 with systemd-networkd. Distributors may choose to not package this,
2277 however in that case users will be prevented from correctly managing
2278 bond0 interface using systemd-networkd.
2280 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
2281 which print the logging level and target of the system manager. They
2282 complement the existing "set-log-level" and "set-log-target" verbs
2283 used to change those values.
2285 * journald.conf gained a new boolean setting ReadKMsg= which defaults
2286 to on. If turned off kernel log messages will not be read by
2287 systemd-journald or included in the logs. It also gained a new
2288 setting LineMax= for configuring the maximum line length in
2289 STDOUT/STDERR log streams. The new default for this value is 48K, up
2290 from the previous hardcoded 2048.
2292 * A new unit setting RuntimeDirectoryPreserve= has been added, which
2293 allows more detailed control of what to do with a runtime directory
2294 configured with RuntimeDirectory= (i.e. a directory below /run or
2295 $XDG_RUNTIME_DIR) after a unit is stopped.
2297 * The RuntimeDirectory= setting for units gained support for creating
2298 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
2299 one top-level directory.
2301 * Units gained new options StateDirectory=, CacheDirectory=,
2302 LogsDirectory= and ConfigurationDirectory= which are closely related
2303 to RuntimeDirectory= but manage per-service directories below
2304 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
2305 possible to write unit files which when activated automatically gain
2306 properly owned service specific directories in these locations, thus
2307 making unit files self-contained and increasing compatibility with
2308 stateless systems and factory reset where /etc or /var are
2309 unpopulated at boot. Matching these new settings there's also
2310 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
2311 ConfigurationDirectoryMode= for configuring the access mode of these
2312 directories. These settings are particularly useful in combination
2313 with DynamicUser=yes as they provide secure, properly-owned,
2314 writable, and stateful locations for storage, excluded from the
2315 sandbox that such services live in otherwise.
2317 * Automake support has been removed from this release. systemd is now
2320 * systemd-journald will now aggressively cache client metadata during
2321 runtime, speeding up log write performance under pressure. This comes
2322 at a small price though: as much of the metadata is read
2323 asynchronously from /proc/ (and isn't implicitly attached to log
2324 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
2325 metadata stored alongside a log entry might be slightly
2326 out-of-date. Previously it could only be slightly newer than the log
2327 message. The time window is small however, and given that the kernel
2328 is unlikely to be improved anytime soon in this regard, this appears
2331 * nss-myhostname/systemd-resolved will now by default synthesize an
2332 A/AAAA resource record for the "_gateway" hostname, pointing to the
2333 current default IP gateway. Previously it did that for the "gateway"
2334 name, hampering adoption, as some distributions wanted to leave that
2335 host name open for local use. The old behaviour may still be
2336 requested at build time.
2338 * systemd-networkd's [Address] section in .network files gained a new
2339 Scope= setting for configuring the IP address scope. The [Network]
2340 section gained a new boolean setting ConfigureWithoutCarrier= that
2341 tells systemd-networkd to ignore link sensing when configuring the
2342 device. The [DHCP] section gained a new Anonymize= boolean option for
2343 turning on a number of options suggested in RFC 7844. A new
2344 [RoutingPolicyRule] section has been added for configuring the IP
2345 routing policy. The [Route] section has gained support for a new
2346 Type= setting which permits configuring
2347 blackhole/unreachable/prohibit routes.
2349 * The [VRF] section in .netdev files gained a new Table= setting for
2350 configuring the routing table to use. The [Tunnel] section gained a
2351 new Independent= boolean field for configuring tunnels independent of
2352 an underlying network interface. The [Bridge] section gained a new
2353 GroupForwardMask= option for configuration of propagation of link
2354 local frames between bridge ports.
2356 * The WakeOnLan= setting in .link files gained support for a number of
2357 new modes. A new TCP6SegmentationOffload= setting has been added for
2358 configuring TCP/IPv6 hardware segmentation offload.
2360 * The IPv6 RA sender implementation may now optionally send out RDNSS
2361 and RDNSSL records to supply DNS configuration to peers.
2363 * systemd-nspawn gained support for a new --system-call-filter= command
2364 line option for adding and removing entries in the default system
2365 call filter it applies. Moreover systemd-nspawn has been changed to
2366 implement a system call whitelist instead of a blacklist.
2368 * systemd-run gained support for a new --pipe command line option. If
2369 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
2370 are directly passed on to the activated transient service
2371 executable. This allows invoking arbitrary processes as systemd
2372 services (for example to take benefit of dependency management,
2373 accounting management, resource management or log management that is
2374 done automatically for services) — while still allowing them to be
2375 integrated in a classic UNIX shell pipeline.
2377 * When a service sends RELOAD=1 via sd_notify() and reload propagation
2378 using ReloadPropagationTo= is configured, a reload is now propagated
2379 to configured units. (Previously this was only done on explicitly
2380 requested reloads, using "systemctl reload" or an equivalent
2383 * For each service unit a restart counter is now kept: it is increased
2384 each time the service is restarted due to Restart=, and may be
2385 queried using "systemctl show -p NRestarts …".
2387 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
2388 @signal and @timer have been added, for usage with SystemCallFilter=
2389 in unit files and the new --system-call-filter= command line option
2390 of systemd-nspawn (see above).
2392 * ExecStart= lines in unit files gained two new modifiers: when a
2393 command line is prefixed with "!" the command will be executed as
2394 configured, except for the credentials applied by
2395 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
2396 "+", but does still apply namespacing options unlike "+". There's
2397 also "!!" now, which is mostly identical, but becomes a NOP on
2398 systems that support ambient capabilities. This is useful to write
2399 unit files that work with ambient capabilities where possible but
2400 automatically fall back to traditional privilege dropping mechanisms
2401 on systems where this is not supported.
2403 * ListenNetlink= settings in socket units now support RDMA netlink
2406 * A new unit file setting LockPersonality= has been added which permits
2407 locking down the chosen execution domain ("personality") of a service
2410 * A new special target "getty-pre.target" has been added, which is
2411 ordered before all text logins, and may be used to order services
2412 before textual logins acquire access to the console.
2414 * systemd will now attempt to load the virtio-rng.ko kernel module very
2415 early on if a VM environment supporting this is detected. This should
2416 improve entropy during early boot in virtualized environments.
2418 * A _netdev option is now supported in /etc/crypttab that operates in a
2419 similar way as the same option in /etc/fstab: it permits configuring
2420 encrypted devices that need to be ordered after the network is up.
2421 Following this logic, two new special targets
2422 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
2423 added that are to cryptsetup.target what remote-fs.target and
2424 remote-fs-pre.target are to local-fs.target.
2426 * Service units gained a new UnsetEnvironment= setting which permits
2427 unsetting specific environment variables for services that are
2428 normally passed to it (for example in order to mask out locale
2429 settings for specific services that can't deal with it).
2431 * Units acquired a new boolean option IPAccounting=. When turned on, IP
2432 traffic accounting (packet count as well as byte count) is done for
2433 the service, and shown as part of "systemctl status" or "systemd-run
2436 * Service units acquired two new options IPAddressAllow= and
2437 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
2438 for configuring a simple IP access control list for all sockets of
2439 the unit. These options are available also on .slice and .socket
2440 units, permitting flexible access list configuration for individual
2441 services as well as groups of services (as defined by a slice unit),
2442 including system-wide. Note that IP ACLs configured this way are
2443 enforced on every single IPv4 and IPv6 socket created by any process
2444 of the service unit, and apply to ingress as well as egress traffic.
2446 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
2447 structured log message is generated each time the unit is stopped,
2448 containing information about the consumed resources of this
2451 * A new setting KeyringMode= has been added to unit files, which may be
2452 used to control how the kernel keyring is set up for executed
2455 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
2456 "systemctl kexec" and "systemctl exit" are now always asynchronous in
2457 behaviour (that is: these commands return immediately after the
2458 operation was enqueued instead of waiting for the operation to
2459 complete). Previously, "systemctl poweroff" and "systemctl reboot"
2460 were asynchronous on systems using systemd-logind (i.e. almost
2461 always, and like they were on sysvinit), and the other three commands
2462 were unconditionally synchronous. With this release this is cleaned
2463 up, and callers will see the same asynchronous behaviour on all
2464 systems for all five operations.
2466 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
2469 * .timer units now accept calendar specifications in other timezones
2470 than UTC or the local timezone.
2472 * The tmpfiles snippet var.conf has been changed to create
2473 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
2474 the "utmp" group already, and it appears to be generally understood
2475 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
2476 databases. Previously this was implemented correctly for all these
2477 databases excepts btmp, which has been opened up like this now
2478 too. Note that while the other databases are world-readable
2479 (i.e. 0644), btmp is not and remains more restrictive.
2481 * The systemd-resolve tool gained a new --reset-server-features
2482 switch. When invoked like this systemd-resolved will forget
2483 everything it learnt about the features supported by the configured
2484 upstream DNS servers, and restarts the feature probing logic on the
2485 next resolver look-up for them at the highest feature level
2488 * The status dump systemd-resolved sends to the logs upon receiving
2489 SIGUSR1 now also includes information about all DNS servers it is
2490 configured to use, and the features levels it probed for them.
2492 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
2493 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
2494 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
2495 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
2496 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
2497 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
2498 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
2499 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
2500 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
2501 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
2502 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
2503 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
2504 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
2505 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
2506 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
2507 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
2508 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
2509 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
2510 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
2511 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
2513 — Berlin, 2017-10-06
2517 * Meson is now supported as build system in addition to Automake. It is
2518 our plan to remove Automake in one of our next releases, so that
2519 Meson becomes our exclusive build system. Hence, please start using
2520 the Meson build system in your downstream packaging. There's plenty
2521 of documentation around how to use Meson, the extremely brief
2524 ./autogen.sh && ./configure && make && sudo make install
2528 meson build && ninja -C build && sudo ninja -C build install
2530 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
2531 which permits configuring a timeout on the time a job is
2532 running. This is particularly useful for setting timeouts on jobs for
2535 * Unit files gained two new options ConditionUser= and ConditionGroup=
2536 for conditionalizing units based on the identity of the user/group
2537 running a systemd user instance.
2539 * systemd-networkd now understands a new FlowLabel= setting in the
2540 [VXLAN] section of .network files, as well as a Priority= in
2541 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
2542 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
2543 gained support for configuration of GENEVE links, and IPv6 address
2544 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
2546 * .link files now understand a new Port= setting.
2548 * systemd-networkd's DHCP support gained support for DHCP option 119
2549 (domain search list).
2551 * systemd-networkd gained support for serving IPv6 address ranges using
2552 the Router Advertisement protocol. The new .network configuration
2553 section [IPv6Prefix] may be used to configure the ranges to
2554 serve. This is implemented based on a new, minimal, native server
2555 implementation of RA.
2557 * journalctl's --output= switch gained support for a new parameter
2558 "short-iso-precise" for a mode where timestamps are shown as precise
2561 * systemd-udevd's "net_id" builtin may now generate stable network
2562 interface names from IBM PowerVM VIO devices as well as ACPI platform
2565 * MulticastDNS support in systemd-resolved may now be explicitly
2566 enabled/disabled using the new MulticastDNS= configuration file
2569 * systemd-resolved may now optionally use libidn2 instead of the libidn
2570 for processing internationalized domain names. Support for libidn2
2571 should be considered experimental and should not be enabled by
2574 * "machinectl pull-tar" and related call may now do verification of
2575 downloaded images using SUSE-style .sha256 checksum files in addition
2576 to the already existing support for validating using Ubuntu-style
2579 * sd-bus gained support for a new sd_bus_message_appendv() call which
2580 is va_list equivalent of sd_bus_message_append().
2582 * sd-boot gained support for validating images using SHIM/MOK.
2584 * The SMACK code learnt support for "onlycap".
2586 * systemd-mount --umount is now much smarter in figuring out how to
2587 properly unmount a device given its mount or device path.
2589 * The code to call libnss_dns as a fallback from libnss_resolve when
2590 the communication with systemd-resolved fails was removed. This
2591 fallback was redundant and interfered with the [!UNAVAIL=return]
2592 suffix. See nss-resolve(8) for the recommended configuration.
2594 * systemd-logind may now be restarted without losing state. It stores
2595 the file descriptors for devices it manages in the system manager
2596 using the FDSTORE= mechanism. Please note that further changes in
2597 other components may be required to make use of this (for example
2598 Xorg has code to listen for stops of systemd-logind and terminate
2599 itself when logind is stopped or restarted, in order to avoid using
2600 stale file descriptors for graphical devices, which is now
2601 counterproductive and must be reverted in order for restarts of
2602 systemd-logind to be safe. See
2603 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
2605 * All kernel-install plugins are called with the environment variable
2606 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
2607 /etc/machine-id. If the machine ID could not be determined,
2608 $KERNEL_INSTALL_MACHINE_ID will be empty. Plugins should not put
2609 anything in the entry directory (passed as the second argument) if
2610 $KERNEL_INSTALL_MACHINE_ID is empty. For backwards compatibility, a
2611 temporary directory is passed as the entry directory and removed
2612 after all the plugins exit.
2614 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
2615 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
2616 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
2617 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
2618 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
2619 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
2620 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
2621 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
2622 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
2623 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
2624 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
2625 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
2626 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
2627 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
2628 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
2629 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
2630 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
2631 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
2632 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
2633 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
2634 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
2635 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
2636 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
2637 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
2638 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
2639 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
2640 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
2641 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
2642 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
2645 — Berlin, 2017-07-12
2649 * The "hybrid" control group mode has been modified to improve
2650 compatibility with "legacy" cgroups-v1 setups. Specifically, the
2651 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
2652 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
2653 cgroups-v1 hierarchy), the only externally visible change being that
2654 the cgroups-v2 hierarchy is also mounted, to
2655 /sys/fs/cgroup/unified. This should provide a large degree of
2656 compatibility with "legacy" cgroups-v1, while taking benefit of the
2657 better management capabilities of cgroups-v2.
2659 * The default control group setup mode may be selected both a boot-time
2660 via a set of kernel command line parameters (specifically:
2661 systemd.unified_cgroup_hierarchy= and
2662 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
2663 default selected on the configure command line
2664 (--with-default-hierarchy=). The upstream default is "hybrid"
2665 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
2666 this will change in a future systemd version to be "unified" (pure
2667 cgroups-v2 mode). The third option for the compile time option is
2668 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
2669 distributions to default to "hybrid" mode for release distributions,
2670 starting with v233. We recommend "unified" for development
2671 distributions (specifically: distributions such as Fedora's rawhide)
2672 as that's where things are headed in the long run. Use "legacy" for
2673 greatest stability and compatibility only.
2675 * Note one current limitation of "unified" and "hybrid" control group
2676 setup modes: the kernel currently does not permit the systemd --user
2677 instance (i.e. unprivileged code) to migrate processes between two
2678 disconnected cgroup subtrees, even if both are managed and owned by
2679 the user. This effectively means "systemd-run --user --scope" doesn't
2680 work when invoked from outside of any "systemd --user" service or
2681 scope. Specifically, it is not supported from session scopes. We are
2682 working on fixing this in a future systemd version. (See #3388 for
2683 further details about this.)
2685 * DBus policy files are now installed into /usr rather than /etc. Make
2686 sure your system has dbus >= 1.9.18 running before upgrading to this
2687 version, or override the install path with --with-dbuspolicydir= .
2689 * All python scripts shipped with systemd (specifically: the various
2690 tests written in Python) now require Python 3.
2692 * systemd unit tests can now run standalone (without the source or
2693 build directories), and can be installed into /usr/lib/systemd/tests/
2694 with 'make install-tests'.
2696 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
2697 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
2700 * Support for the %c, %r, %R specifiers in unit files has been
2701 removed. Specifiers are not supposed to be dependent on configuration
2702 in the unit file itself (so that they resolve the same regardless
2703 where used in the unit files), but these specifiers were influenced
2704 by the Slice= option.
2706 * The shell invoked by debug-shell.service now defaults to /bin/sh in
2707 all cases. If distributions want to use a different shell for this
2708 purpose (for example Fedora's /sbin/sushell) they need to specify
2709 this explicitly at configure time using --with-debug-shell=.
2711 * The confirmation spawn prompt has been reworked to offer the
2714 (c)ontinue, proceed without asking anymore
2715 (D)ump, show the state of the unit
2716 (f)ail, don't execute the command and pretend it failed
2718 (i)nfo, show a short summary of the unit
2719 (j)obs, show jobs that are in progress
2720 (s)kip, don't execute the command and pretend it succeeded
2721 (y)es, execute the command
2723 The 'n' choice for the confirmation spawn prompt has been removed,
2724 because its meaning was confusing.
2726 The prompt may now also be redirected to an alternative console by
2727 specifying the console as parameter to systemd.confirm_spawn=.
2729 * Services of Type=notify require a READY=1 notification to be sent
2730 during startup. If no such message is sent, the service now fails,
2731 even if the main process exited with a successful exit code.
2733 * Services that fail to start up correctly now always have their
2734 ExecStopPost= commands executed. Previously, they'd enter "failed"
2735 state directly, without executing these commands.
2737 * The option MulticastDNS= of network configuration files has acquired
2738 an actual implementation. With MulticastDNS=yes a host can resolve
2739 names of remote hosts and reply to mDNS A and AAAA requests.
2741 * When units are about to be started an additional check is now done to
2742 ensure that all dependencies of type BindsTo= (when used in
2743 combination with After=) have been started.
2745 * systemd-analyze gained a new verb "syscall-filter" which shows which
2746 system call groups are defined for the SystemCallFilter= unit file
2747 setting, and which system calls they contain.
2749 * A new system call filter group "@filesystem" has been added,
2750 consisting of various file system related system calls. Group
2751 "@reboot" has been added, covering reboot, kexec and shutdown related
2752 calls. Finally, group "@swap" has been added covering swap
2753 configuration related calls.
2755 * A new unit file option RestrictNamespaces= has been added that may be
2756 used to restrict access to the various process namespace types the
2757 Linux kernel provides. Specifically, it may be used to take away the
2758 right for a service unit to create additional file system, network,
2759 user, and other namespaces. This sandboxing option is particularly
2760 relevant due to the high amount of recently discovered namespacing
2761 related vulnerabilities in the kernel.
2763 * systemd-udev's .link files gained support for a new AutoNegotiation=
2764 setting for configuring Ethernet auto-negotiation.
2766 * systemd-networkd's .network files gained support for a new
2767 ListenPort= setting in the [DHCP] section to explicitly configure the
2768 UDP client port the DHCP client shall listen on.
2770 * .network files gained a new Unmanaged= boolean setting for explicitly
2771 excluding one or more interfaces from management by systemd-networkd.
2773 * The systemd-networkd ProxyARP= option has been renamed to
2774 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
2775 renamed to ReduceARPProxy=. The old names continue to be available
2778 * systemd-networkd gained support for configuring IPv6 Proxy NDP
2779 addresses via the new IPv6ProxyNDPAddress= .network file setting.
2781 * systemd-networkd's bonding device support gained support for two new
2782 configuration options ActiveSlave= and PrimarySlave=.
2784 * The various options in the [Match] section of .network files gained
2785 support for negative matching.
2787 * New systemd-specific mount options are now understood in /etc/fstab:
2789 x-systemd.mount-timeout= may be used to configure the maximum
2790 permitted runtime of the mount command.
2792 x-systemd.device-bound may be set to bind a mount point to its
2793 backing device unit, in order to automatically remove a mount point
2794 if its backing device is unplugged. This option may also be
2795 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
2796 on the block device, which is now automatically set for all CDROM
2797 drives, so that mounted CDs are automatically unmounted when they are
2798 removed from the drive.
2800 x-systemd.after= and x-systemd.before= may be used to explicitly
2801 order a mount after or before another unit or mount point.
2803 * Enqueued start jobs for device units are now automatically garbage
2804 collected if there are no jobs waiting for them anymore.
2806 * systemctl list-jobs gained two new switches: with --after, for every
2807 queued job the jobs it's waiting for are shown; with --before the
2808 jobs which it's blocking are shown.
2810 * systemd-nspawn gained support for ephemeral boots from disk images
2811 (or in other words: --ephemeral and --image= may now be
2812 combined). Moreover, ephemeral boots are now supported for normal
2813 directories, even if the backing file system is not btrfs. Of course,
2814 if the file system does not support file system snapshots or
2815 reflinks, the initial copy operation will be relatively expensive, but
2816 this should still be suitable for many use cases.
2818 * Calendar time specifications in .timer units now support
2819 specifications relative to the end of a month by using "~" instead of
2820 "-" as separator between month and day. For example, "*-02~03" means
2821 "the third last day in February". In addition a new syntax for
2822 repeated events has been added using the "/" character. For example,
2823 "9..17/2:00" means "every two hours from 9am to 5pm".
2825 * systemd-socket-proxyd gained a new parameter --connections-max= for
2826 configuring the maximum number of concurrent connections.
2828 * sd-id128 gained a new API for generating unique IDs for the host in a
2829 way that does not leak the machine ID. Specifically,
2830 sd_id128_get_machine_app_specific() derives an ID based on the
2831 machine ID a in well-defined, non-reversible, stable way. This is
2832 useful whenever an identifier for the host is needed but where the
2833 identifier shall not be useful to identify the system beyond the
2834 scope of the application itself. (Internally this uses HMAC-SHA256 as
2835 keyed hash function using the machine ID as input.)
2837 * NotifyAccess= gained a new supported value "exec". When set
2838 notifications are accepted from all processes systemd itself invoked,
2839 including all control processes.
2841 * .nspawn files gained support for defining overlay mounts using the
2842 Overlay= and OverlayReadOnly= options. Previously this functionality
2843 was only available on the systemd-nspawn command line.
2845 * systemd-nspawn's --bind= and --overlay= options gained support for
2846 bind/overlay mounts whose source lies within the container tree by
2847 prefixing the source path with "+".
2849 * systemd-nspawn's --bind= and --overlay= options gained support for
2850 automatically allocating a temporary source directory in /var/tmp
2851 that is removed when the container dies. Specifically, if the source
2852 directory is specified as empty string this mechanism is selected. An
2853 example usage is --overlay=+/var::/var, which creates an overlay
2854 mount based on the original /var contained in the image, overlayed
2855 with a temporary directory in the host's /var/tmp. This way changes
2856 to /var are automatically flushed when the container shuts down.
2858 * systemd-nspawn --image= option does now permit raw file system block
2859 devices (in addition to images containing partition tables, as
2862 * The disk image dissection logic in systemd-nspawn gained support for
2863 automatically setting up LUKS encrypted as well as Verity protected
2864 partitions. When a container is booted from an encrypted image the
2865 passphrase is queried at start-up time. When a container with Verity
2866 data is started, the root hash is search in a ".roothash" file
2867 accompanying the disk image (alternatively, pass the root hash via
2868 the new --root-hash= command line option).
2870 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
2871 be used to dissect disk images the same way as systemd-nspawn does
2872 it, following the Bootable Partition Specification. It may even be
2873 used to mount disk images with complex partition setups (including
2874 LUKS and Verity partitions) to a local host directory, in order to
2875 inspect them. This tool is not considered public API (yet), and is
2876 thus not installed into /usr/bin. Please do not rely on its
2877 existence, since it might go away or be changed in later systemd
2880 * A new generator "systemd-verity-generator" has been added, similar in
2881 style to "systemd-cryptsetup-generator", permitting automatic setup of
2882 Verity root partitions when systemd boots up. In order to make use of
2883 this your partition setup should follow the Discoverable Partitions
2884 Specification, and the GPT partition ID of the root file system
2885 partition should be identical to the upper 128bit of the Verity root
2886 hash. The GPT partition ID of the Verity partition protecting it
2887 should be the lower 128bit of the Verity root hash. If the partition
2888 image follows this model it is sufficient to specify a single
2889 "roothash=" kernel command line argument to both configure which root
2890 image and verity partition to use as well as the root hash for
2891 it. Note that systemd-nspawn's Verity support follows the same
2892 semantics, meaning that disk images with proper Verity data in place
2893 may be booted in containers with systemd-nspawn as well as on
2894 physical systems via the verity generator. Also note that the "mkosi"
2895 tool available at https://github.com/systemd/mkosi has been updated
2896 to generate Verity protected disk images following this scheme. In
2897 fact, it has been updated to generate disk images that optionally
2898 implement a complete UEFI SecureBoot trust chain, involving a signed
2899 kernel and initrd image that incorporates such a root hash as well as
2900 a Verity-enabled root partition.
2902 * The hardware database (hwdb) udev supports has been updated to carry
2903 accelerometer quirks.
2905 * All system services are now run with a fresh kernel keyring set up
2906 for them. The invocation ID is stored by default in it, thus
2907 providing a safe, non-overridable way to determine the invocation
2910 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
2911 options for bind mounting arbitrary paths in a service-specific
2912 way. When these options are used, arbitrary host or service files and
2913 directories may be mounted to arbitrary locations in the service's
2916 * Documentation has been added that lists all of systemd's low-level
2917 environment variables:
2919 https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md
2921 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
2922 whether a specific socket file descriptor matches a specified socket
2925 * systemd-firstboot has been updated to check for the
2926 systemd.firstboot= kernel command line option. It accepts a boolean
2927 and when set to false the first boot questions are skipped.
2929 * systemd-fstab-generator has been updated to check for the
2930 systemd.volatile= kernel command line option, which either takes an
2931 optional boolean parameter or the special value "state". If used the
2932 system may be booted in a "volatile" boot mode. Specifically,
2933 "systemd.volatile" is used, the root directory will be mounted as
2934 tmpfs, and only /usr is mounted from the actual root file system. If
2935 "systemd.volatile=state" is used, the root directory will be mounted
2936 as usual, but /var is mounted as tmpfs. This concept provides similar
2937 functionality as systemd-nspawn's --volatile= option, but provides it
2938 on physical boots. Use this option for implementing stateless
2939 systems, or testing systems with all state and/or configuration reset
2940 to the defaults. (Note though that many distributions are not
2941 prepared to boot up without a populated /etc or /var, though.)
2943 * systemd-gpt-auto-generator gained support for LUKS encrypted root
2944 partitions. Previously it only supported LUKS encrypted partitions
2945 for all other uses, except for the root partition itself.
2947 * Socket units gained support for listening on AF_VSOCK sockets for
2948 communication in virtualized QEMU environments.
2950 * The "configure" script gained a new option --with-fallback-hostname=
2951 for specifying the fallback hostname to use if none is configured in
2952 /etc/hostname. For example, by specifying
2953 --with-fallback-hostname=fedora it is possible to default to a
2954 hostname of "fedora" on pristine installations.
2956 * systemd-cgls gained support for a new --unit= switch for listing only
2957 the control groups of a specific unit. Similar --user-unit= has been
2958 added for listing only the control groups of a specific user unit.
2960 * systemd-mount gained a new --umount switch for unmounting a mount or
2961 automount point (and all mount/automount points below it).
2963 * systemd will now refuse full configuration reloads (via systemctl
2964 daemon-reload and related calls) unless at least 16MiB of free space
2965 are available in /run. This is a safety precaution in order to ensure
2966 that generators can safely operate after the reload completed.
2968 * A new unit file option RootImage= has been added, which has a similar
2969 effect as RootDirectory= but mounts the service's root directory from
2970 a disk image instead of plain directory. This logic reuses the same
2971 image dissection and mount logic that systemd-nspawn already uses,
2972 and hence supports any disk images systemd-nspawn supports, including
2973 those following the Discoverable Partition Specification, as well as
2974 Verity enabled images. This option enables systemd to run system
2975 services directly off disk images acting as resource bundles,
2976 possibly even including full integrity data.
2978 * A new MountAPIVFS= unit file option has been added, taking a boolean
2979 argument. If enabled /proc, /sys and /dev (collectively called the
2980 "API VFS") will be mounted for the service. This is only relevant if
2981 RootDirectory= or RootImage= is used for the service, as these mounts
2982 are of course in place in the host mount namespace anyway.
2984 * systemd-nspawn gained support for a new --pivot-root= switch. If
2985 specified the root directory within the container image is pivoted to
2986 the specified mount point, while the original root disk is moved to a
2987 different place. This option enables booting of ostree images
2988 directly with systemd-nspawn.
2990 * The systemd build scripts will no longer complain if the NTP server
2991 addresses are not changed from the defaults. Google now supports
2992 these NTP servers officially. We still recommend downstreams to
2993 properly register an NTP pool with the NTP pool project though.
2995 * coredumpctl gained a new "--reverse" option for printing the list
2996 of coredumps in reverse order.
2998 * coredumpctl will now show additional information about truncated and
2999 inaccessible coredumps, as well as coredumps that are still being
3000 processed. It also gained a new --quiet switch for suppressing
3001 additional informational message in its output.
3003 * coredumpctl gained support for only showing coredumps newer and/or
3004 older than specific timestamps, using the new --since= and --until=
3005 options, reminiscent of journalctl's options by the same name.
3007 * The systemd-coredump logic has been improved so that it may be reused
3008 to collect backtraces in non-compiled languages, for example in
3009 scripting languages such as Python.
3011 * machinectl will now show the UID shift of local containers, if user
3012 namespacing is enabled for them.
3014 * systemd will now optionally run "environment generator" binaries at
3015 configuration load time. They may be used to add environment
3016 variables to the environment block passed to services invoked. One
3017 user environment generator is shipped by default that sets up
3018 environment variables based on files dropped into /etc/environment.d
3019 and ~/.config/environment.d/.
3021 * systemd-resolved now includes the new, recently published 2017 DNSSEC
3024 * hostnamed has been updated to report a new chassis type of
3025 "convertible" to cover "foldable" laptops that can both act as a
3026 tablet and as a laptop, such as various Lenovo Yoga devices.
3028 Contributions from: Adrián López, Alexander Galanin, Alexander
3029 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
3030 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
3031 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
3032 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
3033 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
3034 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
3035 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
3036 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
3037 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
3038 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
3039 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
3040 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
3041 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
3042 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
3043 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
3044 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
3045 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
3046 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
3047 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
3048 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
3049 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
3050 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
3051 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
3052 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
3053 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
3054 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
3055 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
3058 — Berlin, 2017-03-01
3062 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
3063 RestrictAddressFamilies= enabled. These sandboxing options should
3064 generally be compatible with the various external udev call-out
3065 binaries we are aware of, however there may be exceptions, in
3066 particular when exotic languages for these call-outs are used. In
3067 this case, consider turning off these settings locally.
3069 * The new RemoveIPC= option can be used to remove IPC objects owned by
3070 the user or group of a service when that service exits.
3072 * The new ProtectKernelModules= option can be used to disable explicit
3073 load and unload operations of kernel modules by a service. In
3074 addition access to /usr/lib/modules is removed if this option is set.
3076 * ProtectSystem= option gained a new value "strict", which causes the
3077 whole file system tree with the exception of /dev, /proc, and /sys,
3078 to be remounted read-only for a service.
3080 * The new ProtectKernelTunables= option can be used to disable
3081 modification of configuration files in /sys and /proc by a service.
3082 Various directories and files are remounted read-only, so access is
3083 restricted even if the file permissions would allow it.
3085 * The new ProtectControlGroups= option can be used to disable write
3086 access by a service to /sys/fs/cgroup.
3088 * Various systemd services have been hardened with
3089 ProtectKernelTunables=yes, ProtectControlGroups=yes,
3090 RestrictAddressFamilies=.
3092 * Support for dynamically creating users for the lifetime of a service
3093 has been added. If DynamicUser=yes is specified, user and group IDs
3094 will be allocated from the range 61184..65519 for the lifetime of the
3095 service. They can be resolved using the new nss-systemd.so NSS
3096 module. The module must be enabled in /etc/nsswitch.conf. Services
3097 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
3098 any resources allocated by the service will be cleaned up when the
3099 service exits. They also have ProtectHome=read-only and
3100 ProtectSystem=strict enabled, so they are not able to make any
3101 permanent modifications to the system.
3103 * The nss-systemd module also always resolves root and nobody, making
3104 it possible to have no /etc/passwd or /etc/group files in minimal
3105 container or chroot environments.
3107 * Services may be started with their own user namespace using the new
3108 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
3109 under which the service is running are mapped. All other users are
3112 * Support for the cgroup namespace has been added to systemd-nspawn. If
3113 supported by kernel, the container system started by systemd-nspawn
3114 will have its own view of the cgroup hierarchy. This new behaviour
3115 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
3117 * The new MemorySwapMax= option can be used to limit the maximum swap
3118 usage under the unified cgroup hierarchy.
3120 * Support for the CPU controller in the unified cgroup hierarchy has
3121 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
3122 options. This controller requires out-of-tree patches for the kernel
3123 and the support is provisional.
3125 * Mount and automount units may now be created transiently
3126 (i.e. dynamically at runtime via the bus API, instead of requiring
3127 unit files in the file system).
3129 * systemd-mount is a new tool which may mount file systems – much like
3130 mount(8), optionally pulling in additional dependencies through
3131 transient .mount and .automount units. For example, this tool
3132 automatically runs fsck on a backing block device before mounting,
3133 and allows the automount logic to be used dynamically from the
3134 command line for establishing mount points. This tool is particularly
3135 useful when dealing with removable media, as it will ensure fsck is
3136 run – if necessary – before the first access and that the file system
3137 is quickly unmounted after each access by utilizing the automount
3138 logic. This maximizes the chance that the file system on the
3139 removable media stays in a clean state, and if it isn't in a clean
3140 state is fixed automatically.
3142 * LazyUnmount=yes option for mount units has been added to expose the
3143 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
3146 * /efi will be used as the mount point of the EFI boot partition, if
3147 the directory is present, and the mount point was not configured
3148 through other means (e.g. fstab). If /efi directory does not exist,
3149 /boot will be used as before. This makes it easier to automatically
3150 mount the EFI partition on systems where /boot is used for something
3153 * When operating on GPT disk images for containers, systemd-nspawn will
3154 now mount the ESP to /boot or /efi according to the same rules as PID
3155 1 running on a host. This allows tools like "bootctl" to operate
3156 correctly within such containers, in order to make container images
3157 bootable on physical systems.
3159 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
3161 * Two new user session targets have been added to support running
3162 graphical sessions under the systemd --user instance:
3163 graphical-session.target and graphical-session-pre.target. See
3164 systemd.special(7) for a description of how those targets should be
3167 * The vconsole initialization code has been significantly reworked to
3168 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
3169 support unicode keymaps. Font and keymap configuration will now be
3170 copied to all allocated virtual consoles.
3172 * FreeBSD's bhyve virtualization is now detected.
3174 * Information recorded in the journal for core dumps now includes the
3175 contents of /proc/mountinfo and the command line of the process at
3176 the top of the process hierarchy (which is usually the init process
3179 * systemd-journal-gatewayd learned the --directory= option to serve
3180 files from the specified location.
3182 * journalctl --root=… can be used to peruse the journal in the
3183 /var/log/ directories inside of a container tree. This is similar to
3184 the existing --machine= option, but does not require the container to
3187 * The hardware database has been extended to support
3188 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
3191 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
3192 specify the click rate for mice which include a horizontal wheel with
3193 a click rate that is different than the one for the vertical wheel.
3195 * systemd-run gained a new --wait option that makes service execution
3196 synchronous. (Specifically, the command will not return until the
3197 specified service binary exited.)
3199 * systemctl gained a new --wait option that causes the start command to
3200 wait until the units being started have terminated again.
3202 * A new journal output mode "short-full" has been added which displays
3203 timestamps with abbreviated English day names and adds a timezone
3204 suffix. Those timestamps include more information than the default
3205 "short" output mode, and can be passed directly to journalctl's
3206 --since= and --until= options.
3208 * /etc/resolv.conf will be bind-mounted into containers started by
3209 systemd-nspawn, if possible, so any changes to resolv.conf contents
3210 are automatically propagated to the container.
3212 * The number of instances for socket-activated services originating
3213 from a single IP address can be limited with
3214 MaxConnectionsPerSource=, extending the existing setting of
3217 * systemd-networkd gained support for vcan ("Virtual CAN") interface
3220 * .netdev and .network configuration can now be extended through
3223 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
3224 Segmentation Offload, Generic Receive Offload, Large Receive Offload
3225 can be enabled and disabled using the new UDPSegmentationOffload=,
3226 TCPSegmentationOffload=, GenericSegmentationOffload=,
3227 GenericReceiveOffload=, LargeReceiveOffload= options in the
3228 [Link] section of .link files.
3230 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
3231 Port VLAN ID can be configured for bridge devices using the new STP=,
3232 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
3233 section of .netdev files.
3235 * The route table to which routes received over DHCP or RA should be
3236 added can be configured with the new RouteTable= option in the [DHCP]
3237 and [IPv6AcceptRA] sections of .network files.
3239 * The Address Resolution Protocol can be disabled on links managed by
3240 systemd-networkd using the ARP=no setting in the [Link] section of
3243 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
3244 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
3245 encode information about the result and exit codes of the current
3246 service runtime cycle.
3248 * systemd-sysctl will now configure kernel parameters in the order
3249 they occur in the configuration files. This matches what sysctl
3250 has been traditionally doing.
3252 * kernel-install "plugins" that are executed to perform various
3253 tasks after a new kernel is added and before an old one is removed
3254 can now return a special value to terminate the procedure and
3255 prevent any later plugins from running.
3257 * Journald's SplitMode=login setting has been deprecated. It has been
3258 removed from documentation, and its use is discouraged. In a future
3259 release it will be completely removed, and made equivalent to current
3260 default of SplitMode=uid.
3262 * Storage=both option setting in /etc/systemd/coredump.conf has been
3263 removed. With fast LZ4 compression storing the core dump twice is not
3266 * The --share-system systemd-nspawn option has been replaced with an
3267 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
3268 this functionality is discouraged. In addition the variables
3269 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
3270 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
3271 individual namespaces.
3273 * "machinectl list" now shows the IP address of running containers in
3274 the output, as well as OS release information.
3276 * "loginctl list" now shows the TTY of each session in the output.
3278 * sd-bus gained new API calls sd_bus_track_set_recursive(),
3279 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
3280 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
3281 tracking objects in a "recursive" mode, where a single client can be
3282 counted multiple times, if it takes multiple references.
3284 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
3285 sd_bus_get_exit_on_disconnect(). They may be used to make a
3286 process using sd-bus automatically exit if the bus connection is
3289 * Bus clients of the service manager may now "pin" loaded units into
3290 memory, by taking an explicit reference on them. This is useful to
3291 ensure the client can retrieve runtime data about the service even
3292 after the service completed execution. Taking such a reference is
3293 available only for privileged clients and should be helpful to watch
3294 running services in a race-free manner, and in particular collect
3295 information about exit statuses and results.
3297 * The nss-resolve module has been changed to strictly return UNAVAIL
3298 when communication via D-Bus with resolved failed, and NOTFOUND when
3299 a lookup completed but was negative. This means it is now possible to
3300 neatly configure fallbacks using nsswitch.conf result checking
3301 expressions. Taking benefit of this, the new recommended
3302 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
3304 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
3306 * A new setting CtrlAltDelBurstAction= has been added to
3307 /etc/systemd/system.conf which may be used to configure the precise
3308 behaviour if the user on the console presses Ctrl-Alt-Del more often
3309 than 7 times in 2s. Previously this would unconditionally result in
3310 an expedited, immediate reboot. With this new setting the precise
3311 operation may be configured in more detail, and also turned off
3314 * In .netdev files two new settings RemoteChecksumTx= and
3315 RemoteChecksumRx= are now understood that permit configuring the
3316 remote checksumming logic for VXLAN networks.
3318 * The service manager learnt a new "invocation ID" concept for invoked
3319 services. Each runtime cycle of a service will get a new invocation
3320 ID (a 128bit random UUID) assigned that identifies the current
3321 run of the service uniquely and globally. A new invocation ID
3322 is generated each time a service starts up. The journal will store
3323 the invocation ID of a service along with any logged messages, thus
3324 making the invocation ID useful for matching the online runtime of a
3325 service with the offline log data it generated in a safe way without
3326 relying on synchronized timestamps. In many ways this new service
3327 invocation ID concept is similar to the kernel's boot ID concept that
3328 uniquely and globally identifies the runtime of each boot. The
3329 invocation ID of a service is passed to the service itself via an
3330 environment variable ($INVOCATION_ID). A new bus call
3331 GetUnitByInvocationID() has been added that is similar to GetUnit()
3332 but instead of retrieving the bus path for a unit by its name
3333 retrieves it by its invocation ID. The returned path is valid only as
3334 long as the passed invocation ID is current.
3336 * systemd-resolved gained a new "DNSStubListener" setting in
3337 resolved.conf. It either takes a boolean value or the special values
3338 "udp" and "tcp", and configures whether to enable the stub DNS
3339 listener on 127.0.0.53:53.
3341 * IP addresses configured via networkd may now carry additional
3342 configuration settings supported by the kernel. New options include:
3343 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
3344 PrefixRoute=, AutoJoin=.
3346 * The PAM configuration fragment file for "user@.service" shipped with
3347 systemd (i.e. the --user instance of systemd) has been stripped to
3348 the minimum necessary to make the system boot. Previously, it
3349 contained Fedora-specific stanzas that did not apply to other
3350 distributions. It is expected that downstream distributions add
3351 additional configuration lines, matching their needs to this file,
3352 using it only as rough template of what systemd itself needs. Note
3353 that this reduced fragment does not even include an invocation of
3354 pam_limits which most distributions probably want to add, even though
3355 systemd itself does not need it. (There's also the new build time
3356 option --with-pamconfdir=no to disable installation of the PAM
3359 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
3360 capability is now also dropped from its set (in addition to
3361 CAP_SYS_MKNOD as before).
3363 * In service unit files it is now possible to connect a specific named
3364 file descriptor with stdin/stdout/stdout of an executed service. The
3365 name may be specified in matching .socket units using the
3366 FileDescriptorName= setting.
3368 * A number of journal settings may now be configured on the kernel
3369 command line. Specifically, the following options are now understood:
3370 systemd.journald.max_level_console=,
3371 systemd.journald.max_level_store=,
3372 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
3373 systemd.journald.max_level_wall=.
3375 * "systemctl is-enabled --full" will now show by which symlinks a unit
3376 file is enabled in the unit dependency tree.
3378 * Support for VeraCrypt encrypted partitions has been added to the
3379 "cryptsetup" logic and /etc/crypttab.
3381 * systemd-detect-virt gained support for a new --private-users switch
3382 that checks whether the invoking processes are running inside a user
3383 namespace. Similar, a new special value "private-users" for the
3384 existing ConditionVirtualization= setting has been added, permitting
3385 skipping of specific units in user namespace environments.
3387 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
3388 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
3389 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
3390 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
3391 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
3392 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
3393 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
3394 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
3395 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
3396 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
3397 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
3398 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
3399 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
3400 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
3401 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
3402 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
3403 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
3404 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
3405 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
3406 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
3407 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
3408 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
3409 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
3410 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
3411 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
3412 Jędrzejewski-Szmek, Zeal Jagannatha
3414 — Santa Fe, 2016-11-03
3418 * In service units the various ExecXYZ= settings have been extended
3419 with an additional special character as first argument of the
3420 assigned value: if the character '+' is used the specified command
3421 line it will be run with full privileges, regardless of User=,
3422 Group=, CapabilityBoundingSet= and similar options. The effect is
3423 similar to the existing PermissionsStartOnly= option, but allows
3424 configuration of this concept for each executed command line
3427 * Services may now alter the service watchdog timeout at runtime by
3428 sending a WATCHDOG_USEC= message via sd_notify().
3430 * MemoryLimit= and related unit settings now optionally take percentage
3431 specifications. The percentage is taken relative to the amount of
3432 physical memory in the system (or in case of containers, the assigned
3433 amount of memory). This allows scaling service resources neatly with
3434 the amount of RAM available on the system. Similarly, systemd-logind's
3435 RuntimeDirectorySize= option now also optionally takes percentage
3438 * In similar fashion TasksMax= takes percentage values now, too. The
3439 value is taken relative to the configured maximum number of processes
3440 on the system. The per-service task maximum has been changed to 15%
3441 using this functionality. (Effectively this is an increase of 512 →
3442 4915 for service units, given the kernel's default pid_max setting.)
3444 * Calendar time specifications in .timer units now understand a ".."
3445 syntax for time ranges. Example: "4..7:10" may now be used for
3446 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
3449 * The InaccessableDirectories=, ReadOnlyDirectories= and
3450 ReadWriteDirectories= unit file settings have been renamed to
3451 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
3452 applied to all kinds of file nodes, and not just directories, with
3453 the exception of symlinks. Specifically these settings may now be
3454 used on block and character device nodes, UNIX sockets and FIFOS as
3455 well as regular files. The old names of these settings remain
3456 available for compatibility.
3458 * systemd will now log about all service processes it kills forcibly
3459 (using SIGKILL) because they remained after the clean shutdown phase
3460 of the service completed. This should help identifying services that
3461 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
3462 systemd-logind's configuration a similar log message is generated for
3463 processes killed at the end of each session due to this setting.
3465 * systemd will now set the $JOURNAL_STREAM environment variable for all
3466 services whose stdout/stderr are connected to the Journal (which
3467 effectively means by default: all services). The variable contains
3468 the device and inode number of the file descriptor used for
3469 stdout/stderr. This may be used by invoked programs to detect whether
3470 their stdout/stderr is connected to the Journal, in which case they
3471 can switch over to direct Journal communication, thus being able to
3472 pass extended, structured metadata along with their log messages. As
3473 one example, this is now used by glib's logging primitives.
3475 * When using systemd's default tmp.mount unit for /tmp, the mount point
3476 will now be established with the "nosuid" and "nodev" options. This
3477 avoids privilege escalation attacks that put traps and exploits into
3478 /tmp. However, this might cause problems if you e. g. put container
3479 images or overlays into /tmp; if you need this, override tmp.mount's
3480 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
3483 * systemd now supports the "memory" cgroup controller also on
3486 * The systemd-cgtop tool now optionally takes a control group path as
3487 command line argument. If specified, the control group list shown is
3488 limited to subgroups of that group.
3490 * The SystemCallFilter= unit file setting gained support for
3491 pre-defined, named system call filter sets. For example
3492 SystemCallFilter=@clock is now an effective way to make all clock
3493 changing-related system calls unavailable to a service. A number of
3494 similar pre-defined groups are defined. Writing system call filters
3495 for system services is simplified substantially with this new
3496 concept. Accordingly, all of systemd's own, long-running services now
3497 enable system call filtering based on this, by default.
3499 * A new service setting MemoryDenyWriteExecute= has been added, taking
3500 a boolean value. If turned on, a service may no longer create memory
3501 mappings that are writable and executable at the same time. This
3502 enhances security for services where this is enabled as it becomes
3503 harder to dynamically write and then execute memory in exploited
3504 service processes. This option has been enabled for all of systemd's
3505 own long-running services.
3507 * A new RestrictRealtime= service setting has been added, taking a
3508 boolean argument. If set the service's processes may no longer
3509 acquire realtime scheduling. This improves security as realtime
3510 scheduling may otherwise be used to easily freeze the system.
3512 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
3513 value. This may be used for requesting that the system manager inside
3514 of the container reports start-up completion to nspawn which then
3515 propagates this notification further to the service manager
3516 supervising nspawn itself. A related option NotifyReady= in .nspawn
3517 files has been added too. This functionality allows ordering of the
3518 start-up of multiple containers using the usual systemd ordering
3521 * machinectl gained a new command "stop" that is an alias for
3524 * systemd-resolved gained support for contacting DNS servers on
3525 link-local IPv6 addresses.
3527 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
3528 its caches. A method call for requesting the same operation has been
3529 added to the bus API too, and is made available via "systemd-resolve
3532 * systemd-resolve gained a new --status switch. If passed a brief
3533 summary of the used DNS configuration with per-interface information
3536 * resolved.conf gained a new Cache= boolean option, defaulting to
3537 on. If turned off local DNS caching is disabled. This comes with a
3538 performance penalty in particular when DNSSEC is enabled. Note that
3539 resolved disables its internal caching implicitly anyway, when the
3540 configured DNS server is on a host-local IP address such as ::1 or
3541 127.0.0.1, thus automatically avoiding double local caching.
3543 * systemd-resolved now listens on the local IP address 127.0.0.53:53
3544 for DNS requests. This improves compatibility with local programs
3545 that do not use the libc NSS or systemd-resolved's bus APIs for name
3546 resolution. This minimal DNS service is only available to local
3547 programs and does not implement the full DNS protocol, but enough to
3548 cover local DNS clients. A new, static resolv.conf file, listing just
3549 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
3550 now recommended to make /etc/resolv.conf a symlink to this file in
3551 order to route all DNS lookups to systemd-resolved, regardless if
3552 done via NSS, the bus API or raw DNS packets. Note that this local
3553 DNS service is not as fully featured as the libc NSS or
3554 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
3555 used to deliver link-local address information (as this implies
3556 sending a local interface index along), LLMNR/mDNS support via this
3557 interface is severely restricted. It is thus strongly recommended for
3558 all applications to use the libc NSS API or native systemd-resolved
3561 * systemd-networkd's bridge support learned a new setting
3562 VLANFiltering= for controlling VLAN filtering. Moreover a new section
3563 in .network files has been added for configuring VLAN bridging in
3564 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
3566 * systemd-networkd's IPv6 Router Advertisement code now makes use of
3567 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
3568 now be acquired without relying on DHCPv6. Two new options
3569 UseDomains= and UseDNS= have been added to configure this behaviour.
3571 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
3572 renamed IPv6AcceptRA=, without altering its behaviour. The old
3573 setting name remains available for compatibility reasons.
3575 * The systemd-networkd VTI/VTI6 tunneling support gained new options
3576 Key=, InputKey= and OutputKey=.
3578 * systemd-networkd gained support for VRF ("Virtual Routing Function")
3579 interface configuration.
3581 * "systemctl edit" may now be used to create new unit files by
3582 specifying the --force switch.
3584 * sd-event gained a new function sd_event_get_iteration() for
3585 requesting the current iteration counter of the event loop. It starts
3586 at zero and is increased by one with each event loop iteration.
3588 * A new rpm macro %systemd_ordering is provided by the macros.systemd
3589 file. It can be used in lieu of %systemd_requires in packages which
3590 don't use any systemd functionality and are intended to be installed
3591 in minimal containers without systemd present. This macro provides
3592 ordering dependencies to ensure that if the package is installed in
3593 the same rpm transaction as systemd, systemd will be installed before
3594 the scriptlets for the package are executed, allowing unit presets
3597 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
3598 been added to simplify packaging of generators.
3600 * The os-release file gained VERSION_CODENAME field for the
3601 distribution nickname (e.g. VERSION_CODENAME=woody).
3603 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
3604 can be set to disable parsing of metadata and the creation
3605 of persistent symlinks for that device.
3607 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
3608 to make them available to logged-in users has been reverted.
3610 * Much of the common code of the various systemd components is now
3611 built into an internal shared library libsystemd-shared-231.so
3612 (incorporating the systemd version number in the name, to be updated
3613 with future releases) that the components link to. This should
3614 decrease systemd footprint both in memory during runtime and on
3615 disk. Note that the shared library is not for public use, and is
3616 neither API nor ABI stable, but is likely to change with every new
3617 released update. Packagers need to make sure that binaries
3618 linking to libsystemd-shared.so are updated in step with the
3621 * Configuration for "mkosi" is now part of the systemd
3622 repository. mkosi is a tool to easily build legacy-free OS images,
3623 and is available on github: https://github.com/systemd/mkosi. If
3624 "mkosi" is invoked in the build tree a new raw OS image is generated
3625 incorporating the systemd sources currently being worked on and a
3626 clean, fresh distribution installation. The generated OS image may be
3627 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
3628 UEFI PC. This functionality is particularly useful to easily test
3629 local changes made to systemd in a pristine, defined environment. See
3630 doc/HACKING for details.
3632 * configure learned the --with-support-url= option to specify the
3633 distribution's bugtracker.
3635 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
3636 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
3637 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
3638 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
3639 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
3640 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
3641 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
3642 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
3643 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
3644 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
3645 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
3646 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
3647 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
3648 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
3649 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
3650 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
3651 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
3652 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
3653 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
3655 — Berlin, 2016-07-25
3659 * DNSSEC is now turned on by default in systemd-resolved (in
3660 "allow-downgrade" mode), but may be turned off during compile time by
3661 passing "--with-default-dnssec=no" to "configure" (and of course,
3662 during runtime with DNSSEC= in resolved.conf). We recommend
3663 downstreams to leave this on at least during development cycles and
3664 report any issues with the DNSSEC logic upstream. We are very
3665 interested in collecting feedback about the DNSSEC validator and its
3666 limitations in the wild. Note however, that DNSSEC support is
3667 probably nothing downstreams should turn on in stable distros just
3668 yet, as it might create incompatibilities with a few DNS servers and
3669 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
3670 automatically whenever we detect such incompatible setups, but there
3671 might be systems we do not cover yet. Hence: please help us testing
3672 the DNSSEC code, leave this on where you can, report back, but then
3673 again don't consider turning this on in your stable, LTS or
3674 production release just yet. (Note that you have to enable
3675 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
3676 and its DNSSEC mode for host name resolution from local
3679 * systemd-resolve conveniently resolves DANE records with the --tlsa
3680 option and OPENPGPKEY records with the --openpgp option. It also
3681 supports dumping raw DNS record data via the new --raw= switch.
3683 * systemd-logind will now by default terminate user processes that are
3684 part of the user session scope unit (session-XX.scope) when the user
3685 logs out. This behavior is controlled by the KillUserProcesses=
3686 setting in logind.conf, and the previous default of "no" is now
3687 changed to "yes". This means that user sessions will be properly
3688 cleaned up after, but additional steps are necessary to allow
3689 intentionally long-running processes to survive logout.
3691 While the user is logged in at least once, user@.service is running,
3692 and any service that should survive the end of any individual login
3693 session can be started at a user service or scope using systemd-run.
3694 systemd-run(1) man page has been extended with an example which shows
3695 how to run screen in a scope unit underneath user@.service. The same
3696 command works for tmux.
3698 After the user logs out of all sessions, user@.service will be
3699 terminated too, by default, unless the user has "lingering" enabled.
3700 To effectively allow users to run long-term tasks even if they are
3701 logged out, lingering must be enabled for them. See loginctl(1) for
3702 details. The default polkit policy was modified to allow users to
3703 set lingering for themselves without authentication.
3705 Previous defaults can be restored at compile time by the
3706 --without-kill-user-processes option to "configure".
3708 * systemd-logind gained new configuration settings SessionsMax= and
3709 InhibitorsMax=, both with a default of 8192. It will not register new
3710 user sessions or inhibitors above this limit.
3712 * systemd-logind will now reload configuration on SIGHUP.
3714 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
3715 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
3716 enable. Also, support for the "io" cgroup controller in the unified
3717 hierarchy has been added, so that the "memory", "pids" and "io" are
3718 now the controllers that are supported on the unified hierarchy.
3720 WARNING: it is not possible to use previous systemd versions with
3721 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
3722 is necessary to also update systemd in the initramfs if using the
3723 unified hierarchy. An updated SELinux policy is also required.
3725 * LLDP support has been extended, and both passive (receive-only) and
3726 active (sender) modes are supported. Passive mode ("routers-only") is
3727 enabled by default in systemd-networkd. Active LLDP mode is enabled
3728 by default for containers on the internal network. The "networkctl
3729 lldp" command may be used to list information gathered. "networkctl
3730 status" will also show basic LLDP information on connected peers now.
3732 * The IAID and DUID unique identifier sent in DHCP requests may now be
3733 configured for the system and each .network file managed by
3734 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
3736 * systemd-networkd gained support for configuring proxy ARP support for
3737 each interface, via the ProxyArp= setting in .network files. It also
3738 gained support for configuring the multicast querier feature of
3739 bridge devices, via the new MulticastQuerier= setting in .netdev
3740 files. Similarly, snooping on the IGMP traffic can be controlled
3741 via the new setting MulticastSnooping=.
3743 A new setting PreferredLifetime= has been added for addresses
3744 configured in .network file to configure the lifetime intended for an
3747 The systemd-networkd DHCP server gained the option EmitRouter=, which
3748 defaults to yes, to configure whether the DHCP Option 3 (Router)
3751 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
3752 systemd-socket-activate and installed into /usr/bin. It is now fully
3755 * systemd-journald now uses separate threads to flush changes to disk
3756 when closing journal files, thus reducing impact of slow disk I/O on
3757 logging performance.
3759 * The sd-journal API gained two new calls
3760 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
3761 can be used to open journal files using file descriptors instead of
3762 file or directory paths. sd_journal_open_container() has been
3763 deprecated, sd_journal_open_directory_fd() should be used instead
3764 with the flag SD_JOURNAL_OS_ROOT.
3766 * journalctl learned a new output mode "-o short-unix" that outputs log
3767 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
3768 UTC). It also gained support for a new --no-hostname setting to
3769 suppress the hostname column in the family of "short" output modes.
3771 * systemd-ask-password now optionally skips printing of the password to
3772 stdout with --no-output which can be useful in scripts.
3774 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
3775 (devices tagged with ID_MAKER_TOOL) are now tagged with
3776 "uaccess" and are available to logged in users.
3778 * The DeviceAllow= unit setting now supports specifiers (with "%").
3780 * "systemctl show" gained a new --value switch, which allows print a
3781 only the contents of a specific unit property, without also printing
3782 the property's name. Similar support was added to "show*" verbs
3783 of loginctl and machinectl that output "key=value" lists.
3785 * A new unit type "generated" was added for files dynamically generated
3786 by generator tools. Similarly, a new unit type "transient" is used
3787 for unit files created using the runtime API. "systemctl enable" will
3788 refuse to operate on such files.
3790 * A new command "systemctl revert" has been added that may be used to
3791 revert to the vendor version of a unit file, in case local changes
3792 have been made by adding drop-ins or overriding the unit file.
3794 * "machinectl clean" gained a new verb to automatically remove all or
3795 just hidden container images.
3797 * systemd-tmpfiles gained support for a new line type "e" for emptying
3798 directories, if they exist, without creating them if they don't.
3800 * systemd-nspawn gained support for automatically patching the UID/GIDs
3801 of the owners and the ACLs of all files and directories in a
3802 container tree to match the UID/GID user namespacing range selected
3803 for the container invocation. This mode is enabled via the new
3804 --private-users-chown switch. It also gained support for
3805 automatically choosing a free, previously unused UID/GID range when
3806 starting a container, via the new --private-users=pick setting (which
3807 implies --private-users-chown). Together, these options for the first
3808 time make user namespacing for nspawn containers fully automatic and
3809 thus deployable. The systemd-nspawn@.service template unit file has
3810 been changed to use this functionality by default.
3812 * systemd-nspawn gained a new --network-zone= switch, that allows
3813 creating ad-hoc virtual Ethernet links between multiple containers,
3814 that only exist as long as at least one container referencing them is
3815 running. This allows easy connecting of multiple containers with a
3816 common link that implements an Ethernet broadcast domain. Each of
3817 these network "zones" may be named relatively freely by the user, and
3818 may be referenced by any number of containers, but each container may
3819 only reference one of these "zones". On the lower level, this is
3820 implemented by an automatically managed bridge network interface for
3821 each zone, that is created when the first container referencing its
3822 zone is created and removed when the last one referencing its zone
3825 * The default start timeout may now be configured on the kernel command
3826 line via systemd.default_timeout_start_sec=. It was already
3827 configurable via the DefaultTimeoutStartSec= option in
3828 /etc/systemd/system.conf.
3830 * Socket units gained a new TriggerLimitIntervalSec= and
3831 TriggerLimitBurst= setting to configure a limit on the activation
3832 rate of the socket unit.
3834 * The LimitNICE= setting now optionally takes normal UNIX nice values
3835 in addition to the raw integer limit value. If the specified
3836 parameter is prefixed with "+" or "-" and is in the range -20..19 the
3837 value is understood as UNIX nice value. If not prefixed like this it
3838 is understood as raw RLIMIT_NICE limit.
3840 * Note that the effect of the PrivateDevices= unit file setting changed
3841 slightly with this release: the per-device /dev file system will be
3842 mounted read-only from this version on, and will have "noexec"
3843 set. This (minor) change of behavior might cause some (exceptional)
3844 legacy software to break, when PrivateDevices=yes is set for its
3845 service. Please leave PrivateDevices= off if you run into problems
3848 * systemd-bootchart has been split out to a separate repository:
3849 https://github.com/systemd/systemd-bootchart
3851 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
3852 merged into the kernel in its current form.
3854 * The compatibility libraries libsystemd-daemon.so,
3855 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
3856 which have been deprecated since systemd-209 have been removed along
3857 with the corresponding pkg-config files. All symbols provided by
3858 those libraries are provided by libsystemd.so.
3860 * The Capabilities= unit file setting has been removed (it is ignored
3861 for backwards compatibility). AmbientCapabilities= and
3862 CapabilityBoundingSet= should be used instead.
3864 * A new special target has been added, initrd-root-device.target,
3865 which creates a synchronization point for dependencies of the root
3866 device in early userspace. Initramfs builders must ensure that this
3867 target is now included in early userspace.
3869 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
3870 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
3871 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
3872 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
3873 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
3874 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
3875 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
3876 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
3877 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
3878 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
3879 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
3880 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
3881 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
3882 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
3883 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
3884 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
3885 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
3886 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
3887 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
3888 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
3889 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
3890 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
3891 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
3892 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
3893 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
3896 — Fairfax, 2016-05-21
3900 * The systemd-resolved DNS resolver service has gained a substantial
3901 set of new features, most prominently it may now act as a DNSSEC
3902 validating stub resolver. DNSSEC mode is currently turned off by
3903 default, but is expected to be turned on by default in one of the
3904 next releases. For now, we invite everybody to test the DNSSEC logic
3905 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
3906 service also gained a full set of D-Bus interfaces, including calls
3907 to configure DNS and DNSSEC settings per link (for use by external
3908 network management software). systemd-resolved and systemd-networkd
3909 now distinguish between "search" and "routing" domains. The former
3910 are used to qualify single-label names, the latter are used purely
3911 for routing lookups within certain domains to specific links.
3912 resolved now also synthesizes RRs for all entries from /etc/hosts.
3914 * The systemd-resolve tool (which is a client utility for
3915 systemd-resolved) has been improved considerably and is now fully
3916 supported and documented. Hence it has moved from /usr/lib/systemd to
3919 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
3922 * The coredump collection logic has been reworked: when a coredump is
3923 collected it is now written to disk, compressed and processed
3924 (including stacktrace extraction) from a new instantiated service
3925 systemd-coredump@.service, instead of directly from the
3926 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
3927 processing large coredumps can take up a substantial amount of
3928 resources and time, and this previously happened entirely outside of
3929 systemd's service supervision. With the new logic the core_pattern
3930 hook only does minimal metadata collection before passing off control
3931 to the new instantiated service, which is configured with a time
3932 limit, a nice level and other settings to minimize negative impact on
3933 the rest of the system. Also note that the new logic will honour the
3934 RLIMIT_CORE setting of the crashed process, which now allows users
3935 and processes to turn off coredumping for their processes by setting
3938 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
3939 and all forked processes by default. Previously, PID 1 would leave
3940 the setting at "0" for all processes, as set by the kernel. Note that
3941 the resource limit traditionally has no effect on the generated
3942 coredumps on the system if the /proc/sys/kernel/core_pattern hook
3943 logic is used. Since the limit is now honoured (see above) its
3944 default has been changed so that the coredumping logic is enabled by
3945 default for all processes, while allowing specific opt-out.
3947 * When the stacktrace is extracted from processes of system users, this
3948 is now done as "systemd-coredump" user, in order to sandbox this
3949 potentially security sensitive parsing operation. (Note that when
3950 processing coredumps of normal users this is done under the user ID
3951 of process that crashed, as before.) Packagers should take notice
3952 that it is now necessary to create the "systemd-coredump" system user
3953 and group at package installation time.
3955 * The systemd-activate socket activation testing tool gained support
3956 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
3957 and --seqpacket switches. It also has been extended to support both
3958 new-style and inetd-style file descriptor passing. Use the new
3959 --inetd switch to request inetd-style file descriptor passing.
3961 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
3962 variable, which takes a boolean value. If set to false, ANSI color
3963 output is disabled in the tools even when run on a terminal that
3966 * The VXLAN support in networkd now supports two new settings
3967 DestinationPort= and PortRange=.
3969 * A new systemd.machine_id= kernel command line switch has been added,
3970 that may be used to set the machine ID in /etc/machine-id if it is
3971 not initialized yet. This command line option has no effect if the
3972 file is already initialized.
3974 * systemd-nspawn gained a new --as-pid2 switch that invokes any
3975 specified command line as PID 2 rather than PID 1 in the
3976 container. In this mode PID 1 is a minimal stub init process that
3977 implements the special POSIX and Linux semantics of PID 1 regarding
3978 signal and child process management. Note that this stub init process
3979 is implemented in nspawn itself and requires no support from the
3980 container image. This new logic is useful to support running
3981 arbitrary commands in the container, as normal processes are
3982 generally not prepared to run as PID 1.
3984 * systemd-nspawn gained a new --chdir= switch for setting the current
3985 working directory for the process started in the container.
3987 * "journalctl /dev/sda" will now output all kernel log messages for
3988 specified device from the current boot, in addition to all devices
3989 that are parents of it. This should make log output about devices
3990 pretty useful, as long as kernel drivers attach enough metadata to
3991 the log messages. (The usual SATA drivers do.)
3993 * The sd-journal API gained two new calls
3994 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
3995 that report whether log data from /run or /var has been found.
3997 * journalctl gained a new switch "--fields" that prints all journal
3998 record field names currently in use in the journal. This is backed
3999 by two new sd-journal API calls sd_journal_enumerate_fields() and
4000 sd_journal_restart_fields().
4002 * Most configurable timeouts in systemd now expect an argument of
4003 "infinity" to turn them off, instead of "0" as before. The semantics
4004 from now on is that a timeout of "0" means "now", and "infinity"
4005 means "never". To maintain backwards compatibility, "0" continues to
4006 turn off previously existing timeout settings.
4008 * "systemctl reload-or-try-restart" has been renamed to "systemctl
4009 try-reload-or-restart" to clarify what it actually does: the "try"
4010 logic applies to both reloading and restarting, not just restarting.
4011 The old name continues to be accepted for compatibility.
4013 * On boot-up, when PID 1 detects that the system clock is behind the
4014 release date of the systemd version in use, the clock is now set
4015 to the latter. Previously, this was already done in timesyncd, in order
4016 to avoid running with clocks set to the various clock epochs such as
4017 1902, 1938 or 1970. With this change the logic is now done in PID 1
4018 in addition to timesyncd during early boot-up, so that it is enforced
4019 before the first process is spawned by systemd. Note that the logic
4020 in timesyncd remains, as it is more comprehensive and ensures
4021 clock monotonicity by maintaining a persistent timestamp file in
4022 /var. Since /var is generally not available in earliest boot or the
4023 initrd, this part of the logic remains in timesyncd, and is not done
4026 * Support for tweaking details in net_cls.class_id through the
4027 NetClass= configuration directive has been removed, as the kernel
4028 people have decided to deprecate that controller in cgroup v2.
4029 Userspace tools such as nftables are moving over to setting rules
4030 that are specific to the full cgroup path of a task, which obsoletes
4031 these controllers anyway. The NetClass= directive is kept around for
4032 legacy compatibility reasons. For a more in-depth description of the
4033 kernel change, please refer to the respective upstream commit:
4035 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
4037 * A new service setting RuntimeMaxSec= has been added that may be used
4038 to specify a maximum runtime for a service. If the timeout is hit, the
4039 service is terminated and put into a failure state.
4041 * A new service setting AmbientCapabilities= has been added. It allows
4042 configuration of additional Linux process capabilities that are
4043 passed to the activated processes. This is only available on very
4046 * The process resource limit settings in service units may now be used
4047 to configure hard and soft limits individually.
4049 * The various libsystemd APIs such as sd-bus or sd-event now publicly
4050 expose support for gcc's __attribute__((cleanup())) C extension.
4051 Specifically, for many object destructor functions alternative
4052 versions have been added that have names suffixed with "p" and take a
4053 pointer to a pointer to the object to destroy, instead of just a
4054 pointer to the object itself. This is useful because these destructor
4055 functions may be used directly as parameters to the cleanup
4056 construct. Internally, systemd has been a heavy user of this GCC
4057 extension for a long time, and with this change similar support is
4058 now available to consumers of the library outside of systemd. Note
4059 that by using this extension in your sources compatibility with old
4060 and strictly ANSI compatible C compilers is lost. However, all gcc or
4061 LLVM versions of recent years support this extension.
4063 * Timer units gained support for a new setting RandomizedDelaySec= that
4064 allows configuring some additional randomized delay to the configured
4065 time. This is useful to spread out timer events to avoid load peaks in
4066 clusters or larger setups.
4068 * Calendar time specifications now support sub-second accuracy.
4070 * Socket units now support listening on SCTP and UDP-lite protocol
4073 * The sd-event API now comes with a full set of man pages.
4075 * Older versions of systemd contained experimental support for
4076 compressing journal files and coredumps with the LZ4 compressor that
4077 was not compatible with the lz4 binary (due to API limitations of the
4078 lz4 library). This support has been removed; only support for files
4079 compatible with the lz4 binary remains. This LZ4 logic is now
4080 officially supported and no longer considered experimental.
4082 * The dkr image import logic has been removed again from importd. dkr's
4083 micro-services focus doesn't fit into the machine image focus of
4084 importd, and quickly got out of date with the upstream dkr API.
4086 * Creation of the /run/lock/lockdev/ directory was dropped from
4087 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
4088 been available for many years. If you still need this, you need to
4089 create your own tmpfiles.d config file with:
4091 d /run/lock/lockdev 0775 root lock -
4093 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
4094 and RebootArgument= have been moved from the [Service] section of
4095 unit files to [Unit], and they are now supported on all unit types,
4096 not just service units. Of course, systemd will continue to
4097 understand these settings also at the old location, in order to
4098 maintain compatibility.
4100 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
4101 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
4102 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
4103 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
4104 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
4105 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
4106 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
4107 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
4108 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
4109 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
4110 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
4111 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
4112 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
4113 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
4114 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
4115 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
4116 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
4117 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
4118 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
4120 — Berlin, 2016-02-11
4124 * A number of properties previously only settable in unit
4125 files are now also available as properties to set when
4126 creating transient units programmatically via the bus, as it
4127 is exposed with systemd-run's --property=
4128 setting. Specifically, these are: SyslogIdentifier=,
4129 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
4130 EnvironmentFile=, ReadWriteDirectories=,
4131 ReadOnlyDirectories=, InaccessibleDirectories=,
4132 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
4134 * When creating transient services via the bus API it is now
4135 possible to pass in a set of file descriptors to use as
4136 STDIN/STDOUT/STDERR for the invoked process.
4138 * Slice units may now be created transiently via the bus APIs,
4139 similar to the way service and scope units may already be
4140 created transiently.
4142 * Wherever systemd expects a calendar timestamp specification
4143 (like in journalctl's --since= and --until= switches) UTC
4144 timestamps are now supported. Timestamps suffixed with "UTC"
4145 are now considered to be in Universal Time Coordinated
4146 instead of the local timezone. Also, timestamps may now
4147 optionally be specified with sub-second accuracy. Both of
4148 these additions also apply to recurring calendar event
4149 specification, such as OnCalendar= in timer units.
4151 * journalctl gained a new "--sync" switch that asks the
4152 journal daemon to write all so far unwritten log messages to
4153 disk and sync the files, before returning.
4155 * systemd-tmpfiles learned two new line types "q" and "Q" that
4156 operate like "v", but also set up a basic btrfs quota
4157 hierarchy when used on a btrfs file system with quota
4160 * tmpfiles' "v", "q" and "Q" will now create a plain directory
4161 instead of a subvolume (even on a btrfs file system) if the
4162 root directory is a plain directory, and not a
4163 subvolume. This should simplify things with certain chroot()
4164 environments which are not aware of the concept of btrfs
4167 * systemd-detect-virt gained a new --chroot switch to detect
4168 whether execution takes place in a chroot() environment.
4170 * CPUAffinity= now takes CPU index ranges in addition to
4173 * The various memory-related resource limit settings (such as
4174 LimitAS=) now understand the usual K, M, G, ... suffixes to
4175 the base of 1024 (IEC). Similar, the time-related resource
4176 limit settings understand the usual min, h, day, ...
4179 * There's a new system.conf setting DefaultTasksMax= to
4180 control the default TasksMax= setting for services and
4181 scopes running on the system. (TasksMax= is the primary
4182 setting that exposes the "pids" cgroup controller on systemd
4183 and was introduced in the previous systemd release.) The
4184 setting now defaults to 512, which means services that are
4185 not explicitly configured otherwise will only be able to
4186 create 512 processes or threads at maximum, from this
4187 version on. Note that this means that thread- or
4188 process-heavy services might need to be reconfigured to set
4189 TasksMax= to a higher value. It is sufficient to set
4190 TasksMax= in these specific unit files to a higher value, or
4191 even "infinity". Similar, there's now a logind.conf setting
4192 UserTasksMax= that defaults to 4096 and limits the total
4193 number of processes or tasks each user may own
4194 concurrently. nspawn containers also have the TasksMax=
4195 value set by default now, to 8192. Note that all of this
4196 only has an effect if the "pids" cgroup controller is
4197 enabled in the kernel. The general benefit of these changes
4198 should be a more robust and safer system, that provides a
4199 certain amount of per-service fork() bomb protection.
4201 * systemd-nspawn gained the new --network-veth-extra= switch
4202 to define additional and arbitrarily-named virtual Ethernet
4203 links between the host and the container.
4205 * A new service execution setting PassEnvironment= has been
4206 added that allows importing select environment variables
4207 from PID1's environment block into the environment block of
4210 * Timer units gained support for a new RemainAfterElapse=
4211 setting which takes a boolean argument. It defaults to on,
4212 exposing behaviour unchanged to previous releases. If set to
4213 off, timer units are unloaded after they elapsed if they
4214 cannot elapse again. This is particularly useful for
4215 transient timer units, which shall not stay around longer
4216 than until they first elapse.
4218 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
4219 default now (the kernel default is 16). This is beneficial
4220 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
4221 allows substantially larger numbers of queued
4222 datagrams. This should increase the capability of systemd to
4223 parallelize boot-up, as logging and sd_notify() are unlikely
4224 to stall execution anymore. If you need to change the value
4225 from the new defaults, use the usual sysctl.d/ snippets.
4227 * The compression framing format used by the journal or
4228 coredump processing has changed to be in line with what the
4229 official LZ4 tools generate. LZ4 compression support in
4230 systemd was considered unsupported previously, as the format
4231 was not compatible with the normal tools. With this release
4232 this has changed now, and it is hence safe for downstream
4233 distributions to turn it on. While not compressing as well
4234 as the XZ, LZ4 is substantially faster, which makes
4235 it a good default choice for the compression logic in the
4236 journal and in coredump handling.
4238 * Any reference to /etc/mtab has been dropped from
4239 systemd. The file has been obsolete since a while, but
4240 systemd refused to work on systems where it was incorrectly
4241 set up (it should be a symlink or non-existent). Please make
4242 sure to update to util-linux 2.27.1 or newer in conjunction
4243 with this systemd release, which also drops any reference to
4244 /etc/mtab. If you maintain a distribution make sure that no
4245 software you package still references it, as this is a
4246 likely source of bugs. There's also a glibc bug pending,
4247 asking for removal of any reference to this obsolete file:
4249 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
4251 Note that only util-linux versions built with
4252 --enable-libmount-force-mountinfo are supported.
4254 * Support for the ".snapshot" unit type has been removed. This
4255 feature turned out to be little useful and little used, and
4256 has now been removed from the core and from systemctl.
4258 * The dependency types RequiresOverridable= and
4259 RequisiteOverridable= have been removed from systemd. They
4260 have been used only very sparingly to our knowledge and
4261 other options that provide a similar effect (such as
4262 systemctl --mode=ignore-dependencies) are much more useful
4263 and commonly used. Moreover, they were only half-way
4264 implemented as the option to control behaviour regarding
4265 these dependencies was never added to systemctl. By removing
4266 these dependency types the execution engine becomes a bit
4267 simpler. Unit files that use these dependencies should be
4268 changed to use the non-Overridable dependency types
4269 instead. In fact, when parsing unit files with these
4270 options, that's what systemd will automatically convert them
4271 too, but it will also warn, asking users to fix the unit
4272 files accordingly. Removal of these dependency types should
4273 only affect a negligible number of unit files in the wild.
4275 * Behaviour of networkd's IPForward= option changed
4276 (again). It will no longer maintain a per-interface setting,
4277 but propagate one way from interfaces where this is enabled
4278 to the global kernel setting. The global setting will be
4279 enabled when requested by a network that is set up, but
4280 never be disabled again. This change was made to make sure
4281 IPv4 and IPv6 behaviour regarding packet forwarding is
4282 similar (as the Linux IPv6 stack does not support
4283 per-interface control of this setting) and to minimize
4286 * In unit files the behaviour of %u, %U, %h, %s has
4287 changed. These specifiers will now unconditionally resolve
4288 to the various user database fields of the user that the
4289 systemd instance is running as, instead of the user
4290 configured in the specific unit via User=. Note that this
4291 effectively doesn't change much, as resolving of these
4292 specifiers was already turned off in the --system instance
4293 of systemd, as we cannot do NSS lookups from PID 1. In the
4294 --user instance of systemd these specifiers where correctly
4295 resolved, but hardly made any sense, since the user instance
4296 lacks privileges to do user switches anyway, and User= is
4297 hence useless. Moreover, even in the --user instance of
4298 systemd behaviour was awkward as it would only take settings
4299 from User= assignment placed before the specifier into
4300 account. In order to unify and simplify the logic around
4301 this the specifiers will now always resolve to the
4302 credentials of the user invoking the manager (which in case
4303 of PID 1 is the root user).
4305 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
4306 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
4307 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
4308 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
4309 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
4310 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
4311 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
4312 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
4313 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
4314 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
4315 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
4316 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
4317 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
4318 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
4321 — Berlin, 2015-11-18
4325 * systemd now depends on util-linux v2.27. More specifically,
4326 the newly added mount monitor feature in libmount now
4327 replaces systemd's former own implementation.
4329 * libmount mandates /etc/mtab not to be regular file, and
4330 systemd now enforces this condition at early boot.
4331 /etc/mtab has been deprecated and warned about for a very
4332 long time, so systems running systemd should already have
4333 stopped having this file around as anything else than a
4334 symlink to /proc/self/mounts.
4336 * Support for the "pids" cgroup controller has been added. It
4337 allows accounting the number of tasks in a cgroup and
4338 enforcing limits on it. This adds two new setting
4339 TasksAccounting= and TasksMax= to each unit, as well as a
4340 global option DefaultTasksAccounting=.
4342 * Support for the "net_cls" cgroup controller has been added.
4343 It allows assigning a net class ID to each task in the
4344 cgroup, which can then be used in firewall rules and traffic
4345 shaping configurations. Note that the kernel netfilter net
4346 class code does not currently work reliably for ingress
4347 packets on unestablished sockets.
4349 This adds a new config directive called NetClass= to CGroup
4350 enabled units. Allowed values are positive numbers for fixed
4351 assignments and "auto" for picking a free value
4354 * 'systemctl is-system-running' now returns 'offline' if the
4355 system is not booted with systemd. This command can now be
4356 used as a substitute for 'systemd-notify --booted'.
4358 * Watchdog timeouts have been increased to 3 minutes for all
4359 in-tree service files. Apparently, disk IO issues are more
4360 frequent than we hoped, and user reported >1 minute waiting
4363 * 'machine-id-commit' functionality has been merged into
4364 'machine-id-setup --commit'. The separate binary has been
4367 * The WorkingDirectory= directive in unit files may now be set
4368 to the special value '~'. In this case, the working
4369 directory is set to the home directory of the user
4370 configured in User=.
4372 * "machinectl shell" will now open the shell in the home
4373 directory of the selected user by default.
4375 * The CrashChVT= configuration file setting is renamed to
4376 CrashChangeVT=, following our usual logic of not
4377 abbreviating unnecessarily. The old directive is still
4378 supported for compat reasons. Also, this directive now takes
4379 an integer value between 1 and 63, or a boolean value. The
4380 formerly supported '-1' value for disabling stays around for
4383 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
4384 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
4385 RootDirectory= properties can now be set for transient
4388 * The systemd-analyze tool gained a new "set-log-target" verb
4389 to change the logging target the system manager logs to
4390 dynamically during runtime. This is similar to how
4391 "systemd-analyze set-log-level" already changes the log
4394 * In nspawn /sys is now mounted as tmpfs, with only a selected
4395 set of subdirectories mounted in from the real sysfs. This
4396 enhances security slightly, and is useful for ensuring user
4397 namespaces work correctly.
4399 * Support for USB FunctionFS activation has been added. This
4400 allows implementation of USB gadget services that are
4401 activated as soon as they are requested, so that they don't
4402 have to run continuously, similar to classic socket
4405 * The "systemctl exit" command now optionally takes an
4406 additional parameter that sets the exit code to return from
4407 the systemd manager when exiting. This is only relevant when
4408 running the systemd user instance, or when running the
4409 system instance in a container.
4411 * sd-bus gained the new API calls sd_bus_path_encode_many()
4412 and sd_bus_path_decode_many() that allow easy encoding and
4413 decoding of multiple identifier strings inside a D-Bus
4414 object path. Another new call sd_bus_default_flush_close()
4415 has been added to flush and close per-thread default
4418 * systemd-cgtop gained support for a -M/--machine= switch to
4419 show the control groups within a certain container only.
4421 * "systemctl kill" gained support for an optional --fail
4422 switch. If specified the requested operation will fail of no
4423 processes have been killed, because the unit had no
4424 processes attached, or similar.
4426 * A new systemd.crash_reboot=1 kernel command line option has
4427 been added that triggers a reboot after crashing. This can
4428 also be set through CrashReboot= in systemd.conf.
4430 * The RuntimeDirectory= setting now understands unit
4431 specifiers like %i or %f.
4433 * A new (still internal) library API sd-ipv4acd has been added,
4434 that implements address conflict detection for IPv4. It's
4435 based on code from sd-ipv4ll, and will be useful for
4436 detecting DHCP address conflicts.
4438 * File descriptors passed during socket activation may now be
4439 named. A new API sd_listen_fds_with_names() is added to
4440 access the names. The default names may be overridden,
4441 either in the .socket file using the FileDescriptorName=
4442 parameter, or by passing FDNAME= when storing the file
4443 descriptors using sd_notify().
4445 * systemd-networkd gained support for:
4447 - Setting the IPv6 Router Advertisement settings via
4448 IPv6AcceptRouterAdvertisements= in .network files.
4450 - Configuring the HelloTimeSec=, MaxAgeSec= and
4451 ForwardDelaySec= bridge parameters in .netdev files.
4453 - Configuring PreferredSource= for static routes in
4456 * The "ask-password" framework used to query for LUKS harddisk
4457 passwords or SSL passwords during boot gained support for
4458 caching passwords in the kernel keyring, if it is
4459 available. This makes sure that the user only has to type in
4460 a passphrase once if there are multiple objects to unlock
4461 with the same one. Previously, such password caching was
4462 available only when Plymouth was used; this moves the
4463 caching logic into the systemd codebase itself. The
4464 "systemd-ask-password" utility gained a new --keyname=
4465 switch to control which kernel keyring key to use for
4466 caching a password in. This functionality is also useful for
4467 enabling display managers such as gdm to automatically
4468 unlock the user's GNOME keyring if its passphrase, the
4469 user's password and the harddisk password are the same, if
4470 gdm-autologin is used.
4472 * When downloading tar or raw images using "machinectl
4473 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
4474 file is now also downloaded, if it is available and stored
4475 next to the image file.
4477 * Units of type ".socket" gained a new boolean setting
4478 Writable= which is only useful in conjunction with
4479 ListenSpecial=. If true, enables opening the specified
4480 special file in O_RDWR mode rather than O_RDONLY mode.
4482 * systemd-rfkill has been reworked to become a singleton
4483 service that is activated through /dev/rfkill on each rfkill
4484 state change and saves the settings to disk. This way,
4485 systemd-rfkill is now compatible with devices that exist
4486 only intermittendly, and even restores state if the previous
4487 system shutdown was abrupt rather than clean.
4489 * The journal daemon gained support for vacuuming old journal
4490 files controlled by the number of files that shall remain,
4491 in addition to the already existing control by size and by
4492 date. This is useful as journal interleaving performance
4493 degrades with too many separate journal files, and allows
4494 putting an effective limit on them. The new setting defaults
4495 to 100, but this may be changed by setting SystemMaxFiles=
4496 and RuntimeMaxFiles= in journald.conf. Also, the
4497 "journalctl" tool gained the new --vacuum-files= switch to
4498 manually vacuum journal files to leave only the specified
4499 number of files in place.
4501 * udev will now create /dev/disk/by-path links for ATA devices
4502 on kernels where that is supported.
4504 * Galician, Serbian, Turkish and Korean translations were added.
4506 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
4507 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
4508 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
4509 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
4510 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
4511 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
4512 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
4513 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
4514 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
4515 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
4516 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
4517 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
4518 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
4519 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
4520 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
4521 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
4522 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
4523 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
4525 — Berlin, 2015-10-07
4529 * The DHCP implementation of systemd-networkd gained a set of
4532 - The DHCP server now supports emitting DNS and NTP
4533 information. It may be enabled and configured via
4534 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
4535 and NTP information is enabled, but no servers are
4536 configured, the corresponding uplink information (if there
4537 is any) is propagated.
4539 - Server and client now support transmission and reception
4540 of timezone information. It can be configured via the
4541 newly introduced network options UseTimezone=,
4542 EmitTimezone=, and Timezone=. Transmission of timezone
4543 information is enabled between host and containers by
4544 default now: the container will change its local timezone
4545 to what the host has set.
4547 - Lease timeouts can now be configured via
4548 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
4550 - The DHCP server improved on the stability of
4551 leases. Clients are more likely to get the same lease
4552 information back, even if the server loses state.
4554 - The DHCP server supports two new configuration options to
4555 control the lease address pool metrics, PoolOffset= and
4558 * The encapsulation limit of tunnels in systemd-networkd may
4559 now be configured via 'EncapsulationLimit='. It allows
4560 modifying the maximum additional levels of encapsulation
4561 that are permitted to be prepended to a packet.
4563 * systemd now supports the concept of user buses replacing
4564 session buses, if used with dbus-1.10 (and enabled via dbus
4565 --enable-user-session). It previously only supported this on
4566 kdbus-enabled systems, and this release expands this to
4567 'dbus-daemon' systems.
4569 * systemd-networkd now supports predictable interface names
4572 * systemd now optionally supports the new Linux kernel
4573 "unified" control group hierarchy. If enabled via the kernel
4574 command-line option 'systemd.unified_cgroup_hierarchy=1',
4575 systemd will try to mount the unified cgroup hierarchy
4576 directly on /sys/fs/cgroup. If not enabled, or not
4577 available, systemd will fall back to the legacy cgroup
4578 hierarchy setup, as before. Host system and containers can
4579 mix and match legacy and unified hierarchies as they
4580 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
4581 environment variable to individually select the hierarchy to
4582 use for executed containers. By default, nspawn will use the
4583 unified hierarchy for the containers if the host uses the
4584 unified hierarchy, and the legacy hierarchy otherwise.
4585 Please note that at this point the unified hierarchy is an
4586 experimental kernel feature and is likely to change in one
4587 of the next kernel releases. Therefore, it should not be
4588 enabled by default in downstream distributions yet. The
4589 minimum required kernel version for the unified hierarchy to
4590 work is 4.2. Note that when the unified hierarchy is used
4591 for the first time delegated access to controllers is
4592 safe. Because of this systemd-nspawn containers will get
4593 access to controllers now, as will systemd user
4594 sessions. This means containers and user sessions may now
4595 manage their own resources, partitioning up what the system
4598 * A new special scope unit "init.scope" has been introduced
4599 that encapsulates PID 1 of the system. It may be used to
4600 determine resource usage and enforce resource limits on PID
4601 1 itself. PID 1 hence moved out of the root of the control
4604 * The cgtop tool gained support for filtering out kernel
4605 threads when counting tasks in a control group. Also, the
4606 count of processes is now recursively summed up by
4607 default. Two options -k and --recursive= have been added to
4608 revert to old behaviour. The tool has also been updated to
4609 work correctly in containers now.
4611 * systemd-nspawn's --bind= and --bind-ro= options have been
4612 extended to allow creation of non-recursive bind mounts.
4614 * libsystemd gained two new calls sd_pid_get_cgroup() and
4615 sd_peer_get_cgroup() which return the control group path of
4616 a process or peer of a connected AF_UNIX socket. This
4617 function call is particularly useful when implementing
4618 delegated subtrees support in the control group hierarchy.
4620 * The "sd-event" event loop API of libsystemd now supports
4621 correct dequeuing of real-time signals, without losing
4624 * When systemd requests a polkit decision when managing units it
4625 will now add additional fields to the request, including unit
4626 name and desired operation. This enables more powerful polkit
4627 policies, that make decisions depending on these parameters.
4629 * nspawn learnt support for .nspawn settings files, that may
4630 accompany the image files or directories of containers, and
4631 may contain additional settings for the container. This is
4632 an alternative to configuring container parameters via the
4633 nspawn command line.
4635 Contributions from: Cristian Rodríguez, Daniel Mack, David
4636 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
4637 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
4638 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
4639 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
4640 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
4641 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
4642 Andersen, Tom Gundersen, Torstein Husebø
4644 — Berlin, 2015-09-08
4648 * machinectl gained a new verb 'shell' which opens a fresh
4649 shell on the target container or the host. It is similar to
4650 the existing 'login' command of machinectl, but spawns the
4651 shell directly without prompting for username or
4652 password. The pseudo machine '.host' now refers to the local
4653 host and is used by default. Hence, 'machinectl shell' can
4654 be used as replacement for 'su -' which spawns a session as
4655 a fresh systemd unit in a way that is fully isolated from
4656 the originating session.
4658 * systemd-networkd learned to cope with private-zone DHCP
4659 options and allows other programs to query the values.
4661 * SELinux access control when enabling/disabling units is no
4662 longer enforced with this release. The previous implementation
4663 was incorrect, and a new corrected implementation is not yet
4664 available. As unit file operations are still protected via
4665 polkit and D-Bus policy this is not a security problem. Yet,
4666 distributions which care about optimal SELinux support should
4667 probably not stabilize on this release.
4669 * sd-bus gained support for matches of type "arg0has=", that
4670 test for membership of strings in string arrays sent in bus
4673 * systemd-resolved now dumps the contents of its DNS and LLMNR
4674 caches to the logs on reception of the SIGUSR1 signal. This
4675 is useful to debug DNS behaviour.
4677 * The coredumpctl tool gained a new --directory= option to
4678 operate on journal files in a specific directory.
4680 * "systemctl reboot" and related commands gained a new
4681 "--message=" option which may be used to set a free-text
4682 wall message when shutting down or rebooting the
4683 system. This message is also logged, which is useful for
4684 figuring out the reason for a reboot or shutdown a
4687 * The "systemd-resolve-host" tool's -i switch now takes
4688 network interface numbers as alternative to interface names.
4690 * A new unit file setting for services has been introduced:
4691 UtmpMode= allows configuration of how precisely systemd
4692 handles utmp and wtmp entries for the service if this is
4693 enabled. This allows writing services that appear similar to
4694 user sessions in the output of the "w", "who", "last" and
4697 * systemd-resolved will now locally synthesize DNS resource
4698 records for the "localhost" and "gateway" domains as well as
4699 the local hostname. This should ensure that clients querying
4700 RRs via resolved will get similar results as those going via
4701 NSS, if nss-myhostname is enabled.
4703 Contributions from: Alastair Hughes, Alex Crawford, Daniel
4704 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
4705 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
4706 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
4707 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
4708 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
4709 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
4710 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
4711 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
4712 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
4713 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
4714 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
4716 — Berlin, 2015-08-27
4720 * The systemd-efi-boot-generator functionality was merged into
4721 systemd-gpt-auto-generator.
4723 * systemd-networkd now supports Group Policy for vxlan
4724 devices. It can be enabled via the new boolean configuration
4725 option called 'GroupPolicyExtension='.
4727 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
4728 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
4729 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
4731 — Berlin, 2015-07-31
4735 * The python-systemd code has been removed from the systemd repository.
4736 A new repository has been created which accommodates the code from
4737 now on, and we kindly ask distributions to create a separate package
4738 for this: https://github.com/systemd/python-systemd
4740 * The systemd daemon will now reload its main configuration
4741 (/etc/systemd/system.conf) on daemon-reload.
4743 * sd-dhcp now exposes vendor specific extensions via
4744 sd_dhcp_lease_get_vendor_specific().
4746 * systemd-networkd gained a number of new configuration options.
4748 - A new boolean configuration option for TAP devices called
4749 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
4750 device, thus allowing to send and receive GSO packets.
4752 - A new tunnel configuration option called 'CopyDSCP='.
4753 If enabled, the DSCP field of ip6 tunnels is copied into the
4754 decapsulated packet.
4756 - A set of boolean bridge configuration options were added.
4757 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
4758 and 'UnicastFlood=' are now parsed by networkd and applied to the
4759 respective bridge link device via the respective IFLA_BRPORT_*
4762 - A new string configuration option to override the hostname sent
4763 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
4764 is true, networkd will use the configured hostname instead of the
4765 system hostname when sending DHCP requests.
4767 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
4768 networkd will configure the IPv6 flow-label of the tunnel device
4769 according to RFC2460.
4771 - The 'macvtap' virtual network devices are now supported, similar to
4772 the already supported 'macvlan' devices.
4774 * systemd-resolved now implements RFC5452 to improve resilience against
4775 cache poisoning. Additionally, source port randomization is enabled
4776 by default to further protect against DNS spoofing attacks.
4778 * nss-mymachines now supports translating UIDs and GIDs of running
4779 containers with user-namespaces enabled. If a container 'foo'
4780 translates a host uid 'UID' to the container uid 'TUID', then
4781 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
4782 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
4783 mapped as 'vg-foo-TGID'.
4785 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
4786 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
4787 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
4788 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
4789 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
4790 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
4791 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
4792 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
4793 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
4794 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
4796 — Berlin, 2015-07-29
4800 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
4801 There are no known issues with current sysfs, and udev does not need
4802 or should be used to work around such bugs.
4804 * udev does no longer enable USB HID power management. Several reports
4805 indicate, that some devices cannot handle that setting.
4807 * The udev accelerometer helper was removed. The functionality
4808 is now fully included in iio-sensor-proxy. But this means,
4809 older iio-sensor-proxy versions will no longer provide
4810 accelerometer/orientation data with this systemd version.
4811 Please upgrade iio-sensor-proxy to version 1.0.
4813 * networkd gained a new configuration option IPv6PrivacyExtensions=
4814 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
4815 for Stateless Address") on selected networks.
4817 * For the sake of fewer build-time dependencies and less code in the
4818 main repository, the python bindings are about to be removed in the
4819 next release. A new repository has been created which accommodates
4820 the code from now on, and we kindly ask distributions to create a
4821 separate package for this. The removal will take place in v223.
4823 https://github.com/systemd/python-systemd
4825 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
4826 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
4827 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
4828 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
4829 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
4830 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
4831 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
4832 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
4833 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
4834 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
4836 — Berlin, 2015-07-07
4840 * The sd-bus.h and sd-event.h APIs have now been declared
4841 stable and have been added to the official interface of
4842 libsystemd.so. sd-bus implements an alternative D-Bus client
4843 library, that is relatively easy to use, very efficient and
4844 supports both classic D-Bus as well as kdbus as transport
4845 backend. sd-event is a generic event loop abstraction that
4846 is built around Linux epoll, but adds features such as event
4847 prioritization or efficient timer handling. Both APIs are good
4848 choices for C programs looking for a bus and/or event loop
4849 implementation that is minimal and does not have to be
4850 portable to other kernels.
4852 * kdbus support is no longer compile-time optional. It is now
4853 always built-in. However, it can still be disabled at
4854 runtime using the kdbus=0 kernel command line setting, and
4855 that setting may be changed to default to off, by specifying
4856 --disable-kdbus at build-time. Note though that the kernel
4857 command line setting has no effect if the kdbus.ko kernel
4858 module is not installed, in which case kdbus is (obviously)
4859 also disabled. We encourage all downstream distributions to
4860 begin testing kdbus by adding it to the kernel images in the
4861 development distributions, and leaving kdbus support in
4864 * The minimal required util-linux version has been bumped to
4867 * Support for chkconfig (--enable-chkconfig) was removed in
4868 favor of calling an abstraction tool
4869 /lib/systemd/systemd-sysv-install. This needs to be
4870 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
4871 in README for details.
4873 * If there's a systemd unit and a SysV init script for the
4874 same service name, and the user executes "systemctl enable"
4875 for it (or a related call), then this will now enable both
4876 (or execute the related operation on both), not just the
4879 * The libudev API documentation has been converted from gtkdoc
4882 * gudev has been removed from the systemd tree, it is now an
4885 * The systemd-cgtop tool learnt a new --raw switch to generate
4886 "raw" (machine parsable) output.
4888 * networkd's IPForwarding= .network file setting learnt the
4889 new setting "kernel", which ensures that networkd does not
4890 change the IP forwarding sysctl from the default kernel
4893 * The systemd-logind bus API now exposes a new boolean
4894 property "Docked" that reports whether logind considers the
4895 system "docked", i.e. connected to a docking station or not.
4897 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
4898 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
4899 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
4900 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
4901 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
4902 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
4903 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
4904 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
4905 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
4906 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
4907 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
4908 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
4909 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
4910 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
4911 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
4912 Fink, Zbigniew Jędrzejewski-Szmek
4914 — Berlin, 2015-06-19
4918 * The gudev library has been extracted into a separate repository
4919 available at: https://git.gnome.org/browse/libgudev/
4920 It is now managed as part of the Gnome project. Distributions
4921 are recommended to pass --disable-gudev to systemd and use
4922 gudev from the Gnome project instead. gudev is still included
4923 in systemd, for now. It will be removed soon, though. Please
4924 also see the announcement-thread on systemd-devel:
4925 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
4927 * systemd now exposes a CPUUsageNSec= property for each
4928 service unit on the bus, that contains the overall consumed
4929 CPU time of a service (the sum of what each process of the
4930 service consumed). This value is only available if
4931 CPUAccounting= is turned on for a service, and is then shown
4932 in the "systemctl status" output.
4934 * Support for configuring alternative mappings of the old SysV
4935 runlevels to systemd targets has been removed. They are now
4936 hardcoded in a way that runlevels 2, 3, 4 all map to
4937 multi-user.target and 5 to graphical.target (which
4938 previously was already the default behaviour).
4940 * The auto-mounter logic gained support for mount point
4941 expiry, using a new TimeoutIdleSec= setting in .automount
4942 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
4944 * The EFI System Partition (ESP) as mounted to /boot by
4945 systemd-efi-boot-generator will now be unmounted
4946 automatically after 2 minutes of not being used. This should
4947 minimize the risk of ESP corruptions.
4949 * New /etc/fstab options x-systemd.requires= and
4950 x-systemd.requires-mounts-for= are now supported to express
4951 additional dependencies for mounts. This is useful for
4952 journalling file systems that support external journal
4953 devices or overlay file systems that require underlying file
4954 systems to be mounted.
4956 * systemd does not support direct live-upgrades (via systemctl
4957 daemon-reexec) from versions older than v44 anymore. As no
4958 distribution we are aware of shipped such old versions in a
4959 stable release this should not be problematic.
4961 * When systemd forks off a new per-connection service instance
4962 it will now set the $REMOTE_ADDR environment variable to the
4963 remote IP address, and $REMOTE_PORT environment variable to
4964 the remote IP port. This behaviour is similar to the
4965 corresponding environment variables defined by CGI.
4967 * systemd-networkd gained support for uplink failure
4968 detection. The BindCarrier= option allows binding interface
4969 configuration dynamically to the link sense of other
4970 interfaces. This is useful to achieve behaviour like in
4973 * systemd-networkd gained support for configuring the DHCP
4974 client identifier to use when requesting leases.
4976 * systemd-networkd now has a per-network UseNTP= option to
4977 configure whether NTP server information acquired via DHCP
4978 is passed on to services like systemd-timesyncd.
4980 * systemd-networkd gained support for vti6 tunnels.
4982 * Note that systemd-networkd manages the sysctl variable
4983 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
4984 it is configured for since v219. The variable controls IP
4985 forwarding, and is a per-interface alternative to the global
4986 /proc/sys/net/ipv[46]/ip_forward. This setting is
4987 configurable in the IPForward= option, which defaults to
4988 "no". This means if networkd is used for an interface it is
4989 no longer sufficient to set the global sysctl option to turn
4990 on IP forwarding! Instead, the .network file option
4991 IPForward= needs to be turned on! Note that the
4992 implementation of this behaviour was broken in v219 and has
4995 * Many bonding and vxlan options are now configurable in
4998 * systemd-nspawn gained a new --property= setting to set unit
4999 properties for the container scope. This is useful for
5000 setting resource parameters (e.g. "CPUShares=500") on
5001 containers started from the command line.
5003 * systemd-nspawn gained a new --private-users= switch to make
5004 use of user namespacing available on recent Linux kernels.
5006 * systemd-nspawn may now be called as part of a shell pipeline
5007 in which case the pipes used for stdin and stdout are passed
5008 directly to the process invoked in the container, without
5009 indirection via a pseudo tty.
5011 * systemd-nspawn gained a new switch to control the UNIX
5012 signal to use when killing the init process of the container
5015 * systemd-nspawn gained a new --overlay= switch for mounting
5016 overlay file systems into the container using the new kernel
5019 * When a container image is imported via systemd-importd and
5020 the host file system is not btrfs, a loopback block device
5021 file is created in /var/lib/machines.raw with a btrfs file
5022 system inside. It is then mounted to /var/lib/machines to
5023 enable btrfs features for container management. The loopback
5024 file and btrfs file system is grown as needed when container
5025 images are imported via systemd-importd.
5027 * systemd-machined/systemd-importd gained support for btrfs
5028 quota, to enforce container disk space limits on disk. This
5029 is exposed in "machinectl set-limit".
5031 * systemd-importd now can import containers from local .tar,
5032 .raw and .qcow2 images, and export them to .tar and .raw. It
5033 can also import dkr v2 images now from the network (on top
5036 * systemd-importd gained support for verifying downloaded
5037 images with gpg2 (previously only gpg1 was supported).
5039 * systemd-machined, systemd-logind, systemd: most bus calls are
5040 now accessible to unprivileged processes via polkit. Also,
5041 systemd-logind will now allow users to kill their own sessions
5042 without further privileges or authorization.
5044 * systemd-shutdownd has been removed. This service was
5045 previously responsible for implementing scheduled shutdowns
5046 as exposed in /usr/bin/shutdown's time parameter. This
5047 functionality has now been moved into systemd-logind and is
5048 accessible via a bus interface.
5050 * "systemctl reboot" gained a new switch --firmware-setup that
5051 can be used to reboot into the EFI firmware setup, if that
5052 is available. systemd-logind now exposes an API on the bus
5053 to trigger such reboots, in case graphical desktop UIs want
5054 to cover this functionality.
5056 * "systemctl enable", "systemctl disable" and "systemctl mask"
5057 now support a new "--now" switch. If specified the units
5058 that are enabled will also be started, and the ones
5059 disabled/masked also stopped.
5061 * The Gummiboot EFI boot loader tool has been merged into
5062 systemd, and renamed to "systemd-boot". The bootctl tool has been
5063 updated to support systemd-boot.
5065 * An EFI kernel stub has been added that may be used to create
5066 kernel EFI binaries that contain not only the actual kernel,
5067 but also an initrd, boot splash, command line and OS release
5068 information. This combined binary can then be signed as a
5069 single image, so that the firmware can verify it all in one
5070 step. systemd-boot has special support for EFI binaries created
5071 like this and can extract OS release information from them
5072 and show them in the boot menu. This functionality is useful
5073 to implement cryptographically verified boot schemes.
5075 * Optional support has been added to systemd-fsck to pass
5076 fsck's progress report to an AF_UNIX socket in the file
5079 * udev will no longer create device symlinks for all block
5080 devices by default. A blacklist for excluding special block
5081 devices from this logic has been turned into a whitelist
5082 that requires picking block devices explicitly that require
5085 * A new (currently still internal) API sd-device.h has been
5086 added to libsystemd. This modernized API is supposed to
5087 replace libudev eventually. In fact, already much of libudev
5088 is now just a wrapper around sd-device.h.
5090 * A new hwdb database for storing metadata about pointing
5091 stick devices has been added.
5093 * systemd-tmpfiles gained support for setting file attributes
5094 similar to the "chattr" tool with new 'h' and 'H' lines.
5096 * systemd-journald will no longer unconditionally set the
5097 btrfs NOCOW flag on new journal files. This is instead done
5098 with tmpfiles snippet using the new 'h' line type. This
5099 allows easy disabling of this logic, by masking the
5100 journal-nocow.conf tmpfiles file.
5102 * systemd-journald will now translate audit message types to
5103 human readable identifiers when writing them to the
5104 journal. This should improve readability of audit messages.
5106 * The LUKS logic gained support for the offset= and skip=
5107 options in /etc/crypttab, as previously implemented by
5110 * /usr/lib/os-release gained a new optional field VARIANT= for
5111 distributions that support multiple variants (such as a
5112 desktop edition, a server edition, ...)
5114 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
5115 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
5116 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
5117 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
5118 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
5119 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
5120 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
5121 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
5122 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
5123 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
5124 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
5125 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
5126 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
5127 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
5128 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
5129 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
5130 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
5131 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
5132 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
5133 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
5134 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
5135 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
5136 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
5137 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
5138 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
5139 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
5140 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
5142 — Berlin, 2015-05-22
5146 * Introduce a new API "sd-hwdb.h" for querying the hardware
5147 metadata database. With this minimal interface one can query
5148 and enumerate the udev hwdb, decoupled from the old libudev
5149 library. libudev's interface for this is now only a wrapper
5150 around sd-hwdb. A new tool systemd-hwdb has been added to
5151 interface with and update the database.
5153 * When any of systemd's tools copies files (for example due to
5154 tmpfiles' C lines) a btrfs reflink will attempted first,
5155 before bytewise copying is done.
5157 * systemd-nspawn gained a new --ephemeral switch. When
5158 specified a btrfs snapshot is taken of the container's root
5159 directory, and immediately removed when the container
5160 terminates again. Thus, a container can be started whose
5161 changes never alter the container's root directory, and are
5162 lost on container termination. This switch can also be used
5163 for starting a container off the root file system of the
5164 host without affecting the host OS. This switch is only
5165 available on btrfs file systems.
5167 * systemd-nspawn gained a new --template= switch. It takes the
5168 path to a container tree to use as template for the tree
5169 specified via --directory=, should that directory be
5170 missing. This allows instantiating containers dynamically,
5171 on first run. This switch is only available on btrfs file
5174 * When a .mount unit refers to a mount point on which multiple
5175 mounts are stacked, and the .mount unit is stopped all of
5176 the stacked mount points will now be unmounted until no
5177 mount point remains.
5179 * systemd now has an explicit notion of supported and
5180 unsupported unit types. Jobs enqueued for unsupported unit
5181 types will now fail with an "unsupported" error code. More
5182 specifically .swap, .automount and .device units are not
5183 supported in containers, .busname units are not supported on
5184 non-kdbus systems. .swap and .automount are also not
5185 supported if their respective kernel compile time options
5188 * machinectl gained support for two new "copy-from" and
5189 "copy-to" commands for copying files from a running
5190 container to the host or vice versa.
5192 * machinectl gained support for a new "bind" command to bind
5193 mount host directories into local containers. This is
5194 currently only supported for nspawn containers.
5196 * networkd gained support for configuring bridge forwarding
5197 database entries (fdb) from .network files.
5199 * A new tiny daemon "systemd-importd" has been added that can
5200 download container images in tar, raw, qcow2 or dkr formats,
5201 and make them available locally in /var/lib/machines, so
5202 that they can run as nspawn containers. The daemon can GPG
5203 verify the downloads (not supported for dkr, since it has no
5204 provisions for verifying downloads). It will transparently
5205 decompress bz2, xz, gzip compressed downloads if necessary,
5206 and restore sparse files on disk. The daemon uses privilege
5207 separation to ensure the actual download logic runs with
5208 fewer privileges than the daemon itself. machinectl has
5209 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
5210 make the functionality of importd available to the
5211 user. With this in place the Fedora and Ubuntu "Cloud"
5212 images can be downloaded and booted as containers unmodified
5213 (the Fedora images lack the appropriate GPG signature files
5214 currently, so they cannot be verified, but this will change
5215 soon, hopefully). Note that downloading images is currently
5216 only fully supported on btrfs.
5218 * machinectl is now able to list container images found in
5219 /var/lib/machines, along with some metadata about sizes of
5220 disk and similar. If the directory is located on btrfs and
5221 quota is enabled, this includes quota display. A new command
5222 "image-status" has been added that shows additional
5223 information about images.
5225 * machinectl is now able to clone container images
5226 efficiently, if the underlying file system (btrfs) supports
5227 it, with the new "machinectl clone" command. It also
5228 gained commands for renaming and removing images, as well as
5229 marking them read-only or read-write (supported also on
5230 legacy file systems).
5232 * networkd gained support for collecting LLDP network
5233 announcements, from hardware that supports this. This is
5234 shown in networkctl output.
5236 * systemd-run gained support for a new -t (--pty) switch for
5237 invoking a binary on a pty whose input and output is
5238 connected to the invoking terminal. This allows executing
5239 processes as system services while interactively
5240 communicating with them via the terminal. Most interestingly
5241 this is supported across container boundaries. Invoking
5242 "systemd-run -t /bin/bash" is an alternative to running a
5243 full login session, the difference being that the former
5244 will not register a session, nor go through the PAM session
5247 * tmpfiles gained support for a new "v" line type for creating
5248 btrfs subvolumes. If the underlying file system is a legacy
5249 file system, this automatically degrades to creating a
5250 normal directory. Among others /var/lib/machines is now
5251 created like this at boot, should it be missing.
5253 * The directory /var/lib/containers/ has been deprecated and
5254 been replaced by /var/lib/machines. The term "machines" has
5255 been used in the systemd context as generic term for both
5256 VMs and containers, and hence appears more appropriate for
5257 this, as the directory can also contain raw images bootable
5260 * systemd-nspawn when invoked with -M but without --directory=
5261 or --image= is now capable of searching for the container
5262 root directory, subvolume or disk image automatically, in
5263 /var/lib/machines. systemd-nspawn@.service has been updated
5264 to make use of this, thus allowing it to be used for raw
5267 * A new machines.target unit has been introduced that is
5268 supposed to group all containers/VMs invoked as services on
5269 the system. systemd-nspawn@.service has been updated to
5270 integrate with that.
5272 * machinectl gained a new "start" command, for invoking a
5273 container as a service. "machinectl start foo" is mostly
5274 equivalent to "systemctl start systemd-nspawn@foo.service",
5275 but handles escaping in a nicer way.
5277 * systemd-nspawn will now mount most of the cgroupfs tree
5278 read-only into each container, with the exception of the
5279 container's own subtree in the name=systemd hierarchy.
5281 * journald now sets the special FS_NOCOW file flag for its
5282 journal files. This should improve performance on btrfs, by
5283 avoiding heavy fragmentation when journald's write-pattern
5284 is used on COW file systems. It degrades btrfs' data
5285 integrity guarantees for the files to the same levels as for
5286 ext3/ext4 however. This should be OK though as journald does
5287 its own data integrity checks and all its objects are
5288 checksummed on disk. Also, journald should handle btrfs disk
5289 full events a lot more gracefully now, by processing SIGBUS
5290 errors, and not relying on fallocate() anymore.
5292 * When journald detects that journal files it is writing to
5293 have been deleted it will immediately start new journal
5296 * systemd now provides a way to store file descriptors
5297 per-service in PID 1. This is useful for daemons to ensure
5298 that fds they require are not lost during a daemon
5299 restart. The fds are passed to the daemon on the next
5300 invocation in the same way socket activation fds are
5301 passed. This is now used by journald to ensure that the
5302 various sockets connected to all the system's stdout/stderr
5303 are not lost when journald is restarted. File descriptors
5304 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
5305 an extension to sd_notify(). Note that a limit is enforced
5306 on the number of fds a service can store in PID 1, and it
5307 defaults to 0, so that no fds may be stored, unless this is
5308 explicitly turned on.
5310 * The default TERM variable to use for units connected to a
5311 terminal, when no other value is explicitly is set is now
5312 vt220 rather than vt102. This should be fairly safe still,
5313 but allows PgUp/PgDn work.
5315 * The /etc/crypttab option header= as known from Debian is now
5318 * "loginctl user-status" and "loginctl session-status" will
5319 now show the last 10 lines of log messages of the
5320 user/session following the status output. Similar,
5321 "machinectl status" will show the last 10 log lines
5322 associated with a virtual machine or container
5323 service. (Note that this is usually not the log messages
5324 done in the VM/container itself, but simply what the
5325 container manager logs. For nspawn this includes all console
5328 * "loginctl session-status" without further argument will now
5329 show the status of the session of the caller. Similar,
5330 "lock-session", "unlock-session", "activate",
5331 "enable-linger", "disable-linger" may now be called without
5332 session/user parameter in which case they apply to the
5333 caller's session/user.
5335 * An X11 session scriptlet is now shipped that uploads
5336 $DISPLAY and $XAUTHORITY into the environment of the systemd
5337 --user daemon if a session begins. This should improve
5338 compatibility with X11 enabled applications run as systemd
5341 * Generators are now subject to masking via /etc and /run, the
5342 same way as unit files.
5344 * networkd .network files gained support for configuring
5345 per-link IPv4/IPv6 packet forwarding as well as IPv4
5346 masquerading. This is by default turned on for veth links to
5347 containers, as registered by systemd-nspawn. This means that
5348 nspawn containers run with --network-veth will now get
5349 automatic routed access to the host's networks without any
5350 further configuration or setup, as long as networkd runs on
5353 * systemd-nspawn gained the --port= (-p) switch to expose TCP
5354 or UDP posts of a container on the host. With this in place
5355 it is possible to run containers with private veth links
5356 (--network-veth), and have their functionality exposed on
5357 the host as if their services were running directly on the
5360 * systemd-nspawn's --network-veth switch now gained a short
5361 version "-n", since with the changes above it is now truly
5362 useful out-of-the-box. The systemd-nspawn@.service has been
5363 updated to make use of it too by default.
5365 * systemd-nspawn will now maintain a per-image R/W lock, to
5366 ensure that the same image is not started more than once
5367 writable. (It's OK to run an image multiple times
5368 simultaneously in read-only mode.)
5370 * systemd-nspawn's --image= option is now capable of
5371 dissecting and booting MBR and GPT disk images that contain
5372 only a single active Linux partition. Previously it
5373 supported only GPT disk images with proper GPT type
5374 IDs. This allows running cloud images from major
5375 distributions directly with systemd-nspawn, without
5378 * In addition to collecting mouse dpi data in the udev
5379 hardware database, there's now support for collecting angle
5380 information for mouse scroll wheels. The database is
5381 supposed to guarantee similar scrolling behavior on mice
5382 that it knows about. There's also support for collecting
5383 information about Touchpad types.
5385 * udev's input_id built-in will now also collect touch screen
5386 dimension data and attach it to probed devices.
5388 * /etc/os-release gained support for a Distribution Privacy
5391 * networkd gained support for creating "ipvlan", "gretap",
5392 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
5394 * systemd-tmpfiles gained support for "a" lines for setting
5397 * systemd-nspawn will now mount /tmp in the container to
5398 tmpfs, automatically.
5400 * systemd now exposes the memory.usage_in_bytes cgroup
5401 attribute and shows it for each service in the "systemctl
5402 status" output, if available.
5404 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
5405 immediate reboot is triggered. This useful if shutdown is
5406 hung and is unable to complete, to expedite the
5407 operation. Note that this kind of reboot will still unmount
5408 all file systems, and hence should not result in fsck being
5411 * A .device unit for an optical block device will now be
5412 considered active only when a medium is in the drive. Also,
5413 mount units are now bound to their backing devices thus
5414 triggering automatic unmounting when devices become
5415 unavailable. With this in place systemd will now
5416 automatically unmount left-over mounts when a CD-ROM is
5417 ejected or an USB stick is yanked from the system.
5419 * networkd-wait-online now has support for waiting for
5420 specific interfaces only (with globbing), and for giving up
5421 after a configurable timeout.
5423 * networkd now exits when idle. It will be automatically
5424 restarted as soon as interfaces show up, are removed or
5425 change state. networkd will stay around as long as there is
5426 at least one DHCP state machine or similar around, that keep
5429 * networkd may now configure IPv6 link-local addressing in
5430 addition to IPv4 link-local addressing.
5432 * The IPv6 "token" for use in SLAAC may now be configured for
5433 each .network interface in networkd.
5435 * Routes configured with networkd may now be assigned a scope
5438 * networkd's [Match] sections now support globbing and lists
5439 of multiple space-separated matches per item.
5441 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
5442 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
5443 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
5444 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
5445 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
5446 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
5447 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
5448 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
5449 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
5450 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
5451 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
5452 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
5453 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
5454 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
5455 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
5456 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
5457 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
5458 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
5459 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
5460 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
5461 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
5462 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
5463 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
5464 Hoffmann, Zbigniew Jędrzejewski-Szmek
5466 — Berlin, 2015-02-16
5470 * When querying unit file enablement status (for example via
5471 "systemctl is-enabled"), a new state "indirect" is now known
5472 which indicates that a unit might not be enabled itself, but
5473 another unit listed in its Also= setting might be.
5475 * Similar to the various existing ConditionXYZ= settings for
5476 units, there are now matching AssertXYZ= settings. While
5477 failing conditions cause a unit to be skipped, but its job
5478 to succeed, failing assertions declared like this will cause
5479 a unit start operation and its job to fail.
5481 * hostnamed now knows a new chassis type "embedded".
5483 * systemctl gained a new "edit" command. When used on a unit
5484 file, this allows extending unit files with .d/ drop-in
5485 configuration snippets or editing the full file (after
5486 copying it from /usr/lib to /etc). This will invoke the
5487 user's editor (as configured with $EDITOR), and reload the
5488 modified configuration after editing.
5490 * "systemctl status" now shows the suggested enablement state
5491 for a unit, as declared in the (usually vendor-supplied)
5492 system preset files.
5494 * nss-myhostname will now resolve the single-label host name
5495 "gateway" to the locally configured default IP routing
5496 gateways, ordered by their metrics. This assigns a stable
5497 name to the used gateways, regardless which ones are
5498 currently configured. Note that the name will only be
5499 resolved after all other name sources (if nss-myhostname is
5500 configured properly) and should hence not negatively impact
5501 systems that use the single-label host name "gateway" in
5504 * systemd-inhibit now allows filtering by mode when listing
5507 * Scope and service units gained a new "Delegate" boolean
5508 property, which, when set, allows processes running inside the
5509 unit to further partition resources. This is primarily
5510 useful for systemd user instances as well as container
5513 * journald will now pick up audit messages directly from
5514 the kernel, and log them like any other log message. The
5515 audit fields are split up and fully indexed. This means that
5516 journalctl in many ways is now a (nicer!) alternative to
5517 ausearch, the traditional audit client. Note that this
5518 implements only a minimal audit client. If you want the
5519 special audit modes like reboot-on-log-overflow, please use
5520 the traditional auditd instead, which can be used in
5521 parallel to journald.
5523 * The ConditionSecurity= unit file option now understands the
5524 special string "audit" to check whether auditing is
5527 * journalctl gained two new commands --vacuum-size= and
5528 --vacuum-time= to delete old journal files until the
5529 remaining ones take up no more than the specified size on disk,
5530 or are not older than the specified time.
5532 * A new, native PPPoE library has been added to sd-network,
5533 systemd's library of light-weight networking protocols. This
5534 library will be used in a future version of networkd to
5535 enable PPPoE communication without an external pppd daemon.
5537 * The busctl tool now understands a new "capture" verb that
5538 works similar to "monitor", but writes a packet capture
5539 trace to STDOUT that can be redirected to a file which is
5540 compatible with libcap's capture file format. This can then
5541 be loaded in Wireshark and similar tools to inspect bus
5544 * The busctl tool now understands a new "tree" verb that shows
5545 the object trees of a specific service on the bus, or of all
5548 * The busctl tool now understands a new "introspect" verb that
5549 shows all interfaces and members of objects on the bus,
5550 including their signature and values. This is particularly
5551 useful to get more information about bus objects shown by
5552 the new "busctl tree" command.
5554 * The busctl tool now understands new verbs "call",
5555 "set-property" and "get-property" for invoking bus method
5556 calls, setting and getting bus object properties in a
5559 * busctl gained a new --augment-creds= argument that controls
5560 whether the tool shall augment credential information it
5561 gets from the bus with data from /proc, in a possibly
5564 * nspawn's --link-journal= switch gained two new values
5565 "try-guest" and "try-host" that work like "guest" and
5566 "host", but do not fail if the host has no persistent
5567 journalling enabled. -j is now equivalent to
5568 --link-journal=try-guest.
5570 * macvlan network devices created by nspawn will now have
5571 stable MAC addresses.
5573 * A new SmackProcessLabel= unit setting has been added, which
5574 controls the SMACK security label processes forked off by
5575 the respective unit shall use.
5577 * If compiled with --enable-xkbcommon, systemd-localed will
5578 verify x11 keymap settings by compiling the given keymap. It
5579 will spew out warnings if the compilation fails. This
5580 requires libxkbcommon to be installed.
5582 * When a coredump is collected, a larger number of metadata
5583 fields is now collected and included in the journal records
5584 created for it. More specifically, control group membership,
5585 environment variables, memory maps, working directory,
5586 chroot directory, /proc/$PID/status, and a list of open file
5587 descriptors is now stored in the log entry.
5589 * The udev hwdb now contains DPI information for mice. For
5592 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
5594 * All systemd programs that read standalone configuration
5595 files in /etc now also support a corresponding series of
5596 .conf.d configuration directories in /etc/, /run/,
5597 /usr/local/lib/, /usr/lib/, and (if configured with
5598 --enable-split-usr) /lib/. In particular, the following
5599 configuration files now have corresponding configuration
5600 directories: system.conf user.conf, logind.conf,
5601 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
5602 resolved.conf, timesyncd.conf, journal-remote.conf, and
5603 journal-upload.conf. Note that distributions should use the
5604 configuration directories in /usr/lib/; the directories in
5605 /etc/ are reserved for the system administrator.
5607 * systemd-rfkill will no longer take the rfkill device name
5608 into account when storing rfkill state on disk, as the name
5609 might be dynamically assigned and not stable. Instead, the
5610 ID_PATH udev variable combined with the rfkill type (wlan,
5611 bluetooth, ...) is used.
5613 * A new service systemd-machine-id-commit.service has been
5614 added. When used on systems where /etc is read-only during
5615 boot, and /etc/machine-id is not initialized (but an empty
5616 file), this service will copy the temporary machine ID
5617 created as replacement into /etc after the system is fully
5618 booted up. This is useful for systems that are freshly
5619 installed with a non-initialized machine ID, but should get
5620 a fixed machine ID for subsequent boots.
5622 * networkd's .netdev files now provide a large set of
5623 configuration parameters for VXLAN devices. Similarly, the
5624 bridge port cost parameter is now configurable in .network
5625 files. There's also new support for configuring IP source
5626 routing. networkd .link files gained support for a new
5627 OriginalName= match that is useful to match against the
5628 original interface name the kernel assigned. .network files
5629 may include MTU= and MACAddress= fields for altering the MTU
5630 and MAC address while being connected to a specific network
5633 * The LUKS logic gained supported for configuring
5634 UUID-specific key files. There's also new support for naming
5635 LUKS device from the kernel command line, using the new
5636 luks.name= argument.
5638 * Timer units may now be transiently created via the bus API
5639 (this was previously already available for scope and service
5640 units). In addition it is now possible to create multiple
5641 transient units at the same time with a single bus call. The
5642 "systemd-run" tool has been updated to make use of this for
5643 running commands on a specified time, in at(1)-style.
5645 * tmpfiles gained support for "t" lines, for assigning
5646 extended attributes to files. Among other uses this may be
5647 used to assign SMACK labels to files.
5649 Contributions from: Alin Rauta, Alison Chaiken, Andrej
5650 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
5651 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
5652 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
5653 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
5654 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
5655 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
5656 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
5657 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
5658 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
5659 Michael Biebl, Michael Chapman, Michael Marineau, Michal
5660 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
5661 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
5662 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
5663 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
5664 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
5665 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
5666 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
5668 — Berlin, 2014-12-10
5672 * journalctl gained the new options -t/--identifier= to match
5673 on the syslog identifier (aka "tag"), as well as --utc to
5674 show log timestamps in the UTC timezone. journalctl now also
5675 accepts -n/--lines=all to disable line capping in a pager.
5677 * journalctl gained a new switch, --flush, that synchronously
5678 flushes logs from /run/log/journal to /var/log/journal if
5679 persistent storage is enabled. systemd-journal-flush.service
5680 now waits until the operation is complete.
5682 * Services can notify the manager before they start a reload
5683 (by sending RELOADING=1) or shutdown (by sending
5684 STOPPING=1). This allows the manager to track and show the
5685 internal state of daemons and closes a race condition when
5686 the process is still running but has closed its D-Bus
5689 * Services with Type=oneshot do not have to have any ExecStart
5692 * User units are now loaded also from
5693 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
5694 /run/systemd/user directory that was already previously
5695 supported, but is under the control of the user.
5697 * Job timeouts (i.e. timeouts on the time a job that is
5698 queued stays in the run queue) can now optionally result in
5699 immediate reboot or power-off actions (JobTimeoutAction= and
5700 JobTimeoutRebootArgument=). This is useful on ".target"
5701 units, to limit the maximum time a target remains
5702 undispatched in the run queue, and to trigger an emergency
5703 operation in such a case. This is now used by default to
5704 turn off the system if boot-up (as defined by everything in
5705 basic.target) hangs and does not complete for at least
5706 15min. Also, if power-off or reboot hang for at least 30min
5707 an immediate power-off/reboot operation is triggered. This
5708 functionality is particularly useful to increase reliability
5709 on embedded devices, but also on laptops which might
5710 accidentally get powered on when carried in a backpack and
5711 whose boot stays stuck in a hard disk encryption passphrase
5714 * systemd-logind can be configured to also handle lid switch
5715 events even when the machine is docked or multiple displays
5716 are attached (HandleLidSwitchDocked= option).
5718 * A helper binary and a service have been added which can be
5719 used to resume from hibernation in the initramfs. A
5720 generator will parse the resume= option on the kernel
5721 command line to trigger resume.
5723 * A user console daemon systemd-consoled has been
5724 added. Currently, it is a preview, and will so far open a
5725 single terminal on each session of the user marked as
5726 Desktop=systemd-console.
5728 * Route metrics can be specified for DHCP routes added by
5731 * The SELinux context of socket-activated services can be set
5732 from the information provided by the networking stack
5733 (SELinuxContextFromNet= option).
5735 * Userspace firmware loading support has been removed and
5736 the minimum supported kernel version is thus bumped to 3.7.
5738 * Timeout for udev workers has been increased from 1 to 3
5739 minutes, but a warning will be printed after 1 minute to
5740 help diagnose kernel modules that take a long time to load.
5742 * Udev rules can now remove tags on devices with TAG-="foobar".
5744 * systemd's readahead implementation has been removed. In many
5745 circumstances it didn't give expected benefits even for
5746 rotational disk drives and was becoming less relevant in the
5747 age of SSDs. As none of the developers has been using
5748 rotating media anymore, and nobody stepped up to actively
5749 maintain this component of systemd it has now been removed.
5751 * Swap units can use Options= to specify discard options.
5752 Discard options specified for swaps in /etc/fstab are now
5755 * Docker containers are now detected as a separate type of
5758 * The Password Agent protocol gained support for queries where
5759 the user input is shown, useful e.g. for user names.
5760 systemd-ask-password gained a new --echo option to turn that
5763 * The default sysctl.d/ snippets will now set:
5765 net.core.default_qdisc = fq_codel
5767 This selects Fair Queuing Controlled Delay as the default
5768 queuing discipline for network interfaces. fq_codel helps
5769 fight the network bufferbloat problem. It is believed to be
5770 a good default with no tuning required for most workloads.
5771 Downstream distributions may override this choice. On 10Gbit
5772 servers that do not do forwarding, "fq" may perform better.
5773 Systems without a good clocksource should use "pfifo_fast".
5775 * If kdbus is enabled during build a new option BusPolicy= is
5776 available for service units, that allows locking all service
5777 processes into a stricter bus policy, in order to limit
5778 access to various bus services, or even hide most of them
5779 from the service's view entirely.
5781 * networkctl will now show the .network and .link file
5782 networkd has applied to a specific interface.
5784 * sd-login gained a new API call sd_session_get_desktop() to
5785 query which desktop environment has been selected for a
5788 * UNIX utmp support is now compile-time optional to support
5789 legacy-free systems.
5791 * systemctl gained two new commands "add-wants" and
5792 "add-requires" for pulling in units from specific targets
5795 * If the word "rescue" is specified on the kernel command line
5796 the system will now boot into rescue mode (aka
5797 rescue.target), which was previously available only by
5798 specifying "1" or "systemd.unit=rescue.target" on the kernel
5799 command line. This new kernel command line option nicely
5800 mirrors the already existing "emergency" kernel command line
5803 * New kernel command line options mount.usr=, mount.usrflags=,
5804 mount.usrfstype= have been added that match root=, rootflags=,
5805 rootfstype= but allow mounting a specific file system to
5808 * The $NOTIFY_SOCKET is now also passed to control processes of
5809 services, not only the main process.
5811 * This version reenables support for fsck's -l switch. This
5812 means at least version v2.25 of util-linux is required for
5813 operation, otherwise dead-locks on device nodes may
5814 occur. Again: you need to update util-linux to at least
5815 v2.25 when updating systemd to v217.
5817 * The "multi-seat-x" tool has been removed from systemd, as
5818 its functionality has been integrated into X servers 1.16,
5819 and the tool is hence redundant. It is recommended to update
5820 display managers invoking this tool to simply invoke X
5821 directly from now on, again.
5823 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
5824 message flag has been added for all of systemd's polkit
5825 authenticated method calls has been added. In particular this
5826 now allows optional interactive authorization via polkit for
5827 many of PID1's privileged operations such as unit file
5828 enabling and disabling.
5830 * "udevadm hwdb --update" learnt a new switch "--usr" for
5831 placing the rebuilt hardware database in /usr instead of
5832 /etc. When used only hardware database entries stored in
5833 /usr will be used, and any user database entries in /etc are
5834 ignored. This functionality is useful for vendors to ship a
5835 pre-built database on systems where local configuration is
5836 unnecessary or unlikely.
5838 * Calendar time specifications in .timer units now also
5839 understand the strings "semi-annually", "quarterly" and
5840 "minutely" as shortcuts (in addition to the preexisting
5841 "anually", "hourly", ...).
5843 * systemd-tmpfiles will now correctly create files in /dev
5844 at boot which are marked for creation only at boot. It is
5845 recommended to always create static device nodes with 'c!'
5846 and 'b!', so that they are created only at boot and not
5847 overwritten at runtime.
5849 * When the watchdog logic is used for a service (WatchdogSec=)
5850 and the watchdog timeout is hit the service will now be
5851 terminated with SIGABRT (instead of just SIGTERM), in order
5852 to make sure a proper coredump and backtrace is
5853 generated. This ensures that hanging services will result in
5854 similar coredump/backtrace behaviour as services that hit a
5857 Contributions from: Andreas Henriksson, Andrei Borzenkov,
5858 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
5859 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
5860 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
5861 Herrmann, David Sommerseth, David Strauss, Emil Renner
5862 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
5863 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
5864 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
5865 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
5866 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
5867 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
5868 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
5869 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
5870 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
5871 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
5872 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
5873 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
5874 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
5875 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
5876 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
5877 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
5880 — Berlin, 2014-10-28
5884 * timedated no longer reads NTP implementation unit names from
5885 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
5886 implementations should add a
5888 Conflicts=systemd-timesyncd.service
5890 to their unit files to take over and replace systemd's NTP
5891 default functionality.
5893 * systemd-sysusers gained a new line type "r" for configuring
5894 which UID/GID ranges to allocate system users/groups
5895 from. Lines of type "u" may now add an additional column
5896 that specifies the home directory for the system user to be
5897 created. Also, systemd-sysusers may now optionally read user
5898 information from STDIN instead of a file. This is useful for
5899 invoking it from RPM preinst scriptlets that need to create
5900 users before the first RPM file is installed since these
5901 files might need to be owned by them. A new
5902 %sysusers_create_inline RPM macro has been introduced to do
5903 just that. systemd-sysusers now updates the shadow files as
5904 well as the user/group databases, which should enhance
5905 compatibility with certain tools like grpck.
5907 * A number of bus APIs of PID 1 now optionally consult polkit to
5908 permit access for otherwise unprivileged clients under certain
5909 conditions. Note that this currently doesn't support
5910 interactive authentication yet, but this is expected to be
5911 added eventually, too.
5913 * /etc/machine-info now has new fields for configuring the
5914 deployment environment of the machine, as well as the
5915 location of the machine. hostnamectl has been updated with
5916 new command to update these fields.
5918 * systemd-timesyncd has been updated to automatically acquire
5919 NTP server information from systemd-networkd, which might
5920 have been discovered via DHCP.
5922 * systemd-resolved now includes a caching DNS stub resolver
5923 and a complete LLMNR name resolution implementation. A new
5924 NSS module "nss-resolve" has been added which can be used
5925 instead of glibc's own "nss-dns" to resolve hostnames via
5926 systemd-resolved. Hostnames, addresses and arbitrary RRs may
5927 be resolved via systemd-resolved D-Bus APIs. In contrast to
5928 the glibc internal resolver systemd-resolved is aware of
5929 multi-homed system, and keeps DNS server and caches separate
5930 and per-interface. Queries are sent simultaneously on all
5931 interfaces that have DNS servers configured, in order to
5932 properly handle VPNs and local LANs which might resolve
5933 separate sets of domain names. systemd-resolved may acquire
5934 DNS server information from systemd-networkd automatically,
5935 which in turn might have discovered them via DHCP. A tool
5936 "systemd-resolve-host" has been added that may be used to
5937 query the DNS logic in resolved. systemd-resolved implements
5938 IDNA and automatically uses IDNA or UTF-8 encoding depending
5939 on whether classic DNS or LLMNR is used as transport. In the
5940 next releases we intend to add a DNSSEC and mDNS/DNS-SD
5941 implementation to systemd-resolved.
5943 * A new NSS module nss-mymachines has been added, that
5944 automatically resolves the names of all local registered
5945 containers to their respective IP addresses.
5947 * A new client tool "networkctl" for systemd-networkd has been
5948 added. It currently is entirely passive and will query
5949 networking configuration from udev, rtnetlink and networkd,
5950 and present it to the user in a very friendly
5951 way. Eventually, we hope to extend it to become a full
5952 control utility for networkd.
5954 * .socket units gained a new DeferAcceptSec= setting that
5955 controls the kernels' TCP_DEFER_ACCEPT sockopt for
5956 TCP. Similarly, support for controlling TCP keep-alive
5957 settings has been added (KeepAliveTimeSec=,
5958 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
5959 turning off Nagle's algorithm on TCP has been added
5962 * logind learned a new session type "web", for use in projects
5963 like Cockpit which register web clients as PAM sessions.
5965 * timer units with at least one OnCalendar= setting will now
5966 be started only after time-sync.target has been
5967 reached. This way they will not elapse before the system
5968 clock has been corrected by a local NTP client or
5969 similar. This is particular useful on RTC-less embedded
5970 machines, that come up with an invalid system clock.
5972 * systemd-nspawn's --network-veth= switch should now result in
5973 stable MAC addresses for both the outer and the inner side
5976 * systemd-nspawn gained a new --volatile= switch for running
5977 container instances with /etc or /var unpopulated.
5979 * The kdbus client code has been updated to use the new Linux
5980 3.17 memfd subsystem instead of the old kdbus-specific one.
5982 * systemd-networkd's DHCP client and server now support
5983 FORCERENEW. There are also new configuration options to
5984 configure the vendor client identifier and broadcast mode
5987 * systemd will no longer inform the kernel about the current
5988 timezone, as this is necessarily incorrect and racy as the
5989 kernel has no understanding of DST and similar
5990 concepts. This hence means FAT timestamps will be always
5991 considered UTC, similar to what Android is already
5992 doing. Also, when the RTC is configured to the local time
5993 (rather than UTC) systemd will never synchronize back to it,
5994 as this might confuse Windows at a later boot.
5996 * systemd-analyze gained a new command "verify" for offline
5997 validation of unit files.
5999 * systemd-networkd gained support for a couple of additional
6000 settings for bonding networking setups. Also, the metric for
6001 statically configured routes may now be configured. For
6002 network interfaces where this is appropriate the peer IP
6003 address may now be configured.
6005 * systemd-networkd's DHCP client will no longer request
6006 broadcasting by default, as this tripped up some networks.
6007 For hardware where broadcast is required the feature should
6008 be switched back on using RequestBroadcast=yes.
6010 * systemd-networkd will now set up IPv4LL addresses (when
6011 enabled) even if DHCP is configured successfully.
6013 * udev will now default to respect network device names given
6014 by the kernel when the kernel indicates that these are
6015 predictable. This behavior can be tweaked by changing
6016 NamePolicy= in the relevant .link file.
6018 * A new library systemd-terminal has been added that
6019 implements full TTY stream parsing and rendering. This
6020 library is supposed to be used later on for implementing a
6021 full userspace VT subsystem, replacing the current kernel
6024 * A new tool systemd-journal-upload has been added to push
6025 journal data to a remote system running
6026 systemd-journal-remote.
6028 * journald will no longer forward all local data to another
6029 running syslog daemon. This change has been made because
6030 rsyslog (which appears to be the most commonly used syslog
6031 implementation these days) no longer makes use of this, and
6032 instead pulls the data out of the journal on its own. Since
6033 forwarding the messages to a non-existent syslog server is
6034 more expensive than we assumed we have now turned this
6035 off. If you run a syslog server that is not a recent rsyslog
6036 version, you have to turn this option on again
6037 (ForwardToSyslog= in journald.conf).
6039 * journald now optionally supports the LZ4 compressor for
6040 larger journal fields. This compressor should perform much
6041 better than XZ which was the previous default.
6043 * machinectl now shows the IP addresses of local containers,
6044 if it knows them, plus the interface name of the container.
6046 * A new tool "systemd-escape" has been added that makes it
6047 easy to escape strings to build unit names and similar.
6049 * sd_notify() messages may now include a new ERRNO= field
6050 which is parsed and collected by systemd and shown among the
6051 "systemctl status" output for a service.
6053 * A new component "systemd-firstboot" has been added that
6054 queries the most basic systemd information (timezone,
6055 hostname, root password) interactively on first
6056 boot. Alternatively it may also be used to provision these
6057 things offline on OS images installed into directories.
6059 * The default sysctl.d/ snippets will now set
6061 net.ipv4.conf.default.promote_secondaries=1
6063 This has the benefit of no flushing secondary IP addresses
6064 when primary addresses are removed.
6066 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
6067 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
6068 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
6069 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
6070 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
6071 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
6072 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
6073 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
6074 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
6075 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
6076 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
6077 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
6078 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
6079 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
6080 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
6082 — Berlin, 2014-08-19
6086 * A new tool systemd-sysusers has been added. This tool
6087 creates system users and groups in /etc/passwd and
6088 /etc/group, based on static declarative system user/group
6089 definitions in /usr/lib/sysusers.d/. This is useful to
6090 enable factory resets and volatile systems that boot up with
6091 an empty /etc directory, and thus need system users and
6092 groups created during early boot. systemd now also ships
6093 with two default sysusers.d/ files for the most basic
6094 users and groups systemd and the core operating system
6097 * A new tmpfiles snippet has been added that rebuilds the
6098 essential files in /etc on boot, should they be missing.
6100 * A directive for ensuring automatic clean-up of
6101 /var/cache/man/ has been removed from the default
6102 configuration. This line should now be shipped by the man
6103 implementation. The necessary change has been made to the
6104 man-db implementation. Note that you need to update your man
6105 implementation to one that ships this line, otherwise no
6106 automatic clean-up of /var/cache/man will take place.
6108 * A new condition ConditionNeedsUpdate= has been added that
6109 may conditionalize services to only run when /etc or /var
6110 are "older" than the vendor operating system resources in
6111 /usr. This is useful for reconstructing or updating /etc
6112 after an offline update of /usr or a factory reset, on the
6113 next reboot. Services that want to run once after such an
6114 update or reset should use this condition and order
6115 themselves before the new systemd-update-done.service, which
6116 will mark the two directories as fully updated. A number of
6117 service files have been added making use of this, to rebuild
6118 the udev hardware database, the journald message catalog and
6119 dynamic loader cache (ldconfig). The systemd-sysusers tool
6120 described above also makes use of this now. With this in
6121 place it is now possible to start up a minimal operating
6122 system with /etc empty cleanly. For more information on the
6123 concepts involved see this recent blog story:
6125 http://0pointer.de/blog/projects/stateless.html
6127 * A new system group "input" has been introduced, and all
6128 input device nodes get this group assigned. This is useful
6129 for system-level software to get access to input devices. It
6130 complements what is already done for "audio" and "video".
6132 * systemd-networkd learnt minimal DHCPv4 server support in
6133 addition to the existing DHCPv4 client support. It also
6134 learnt DHCPv6 client and IPv6 Router Solicitation client
6135 support. The DHCPv4 client gained support for static routes
6136 passed in from the server. Note that the [DHCPv4] section
6137 known in older systemd-networkd versions has been renamed to
6138 [DHCP] and is now also used by the DHCPv6 client. Existing
6139 .network files using settings of this section should be
6140 updated, though compatibility is maintained. Optionally, the
6141 client hostname may now be sent to the DHCP server.
6143 * networkd gained support for vxlan virtual networks as well
6144 as tun/tap and dummy devices.
6146 * networkd gained support for automatic allocation of address
6147 ranges for interfaces from a system-wide pool of
6148 addresses. This is useful for dynamically managing a large
6149 number of interfaces with a single network configuration
6150 file. In particular this is useful to easily assign
6151 appropriate IP addresses to the veth links of a large number
6152 of nspawn instances.
6154 * RPM macros for processing sysusers, sysctl and binfmt
6155 drop-in snippets at package installation time have been
6158 * The /etc/os-release file should now be placed in
6159 /usr/lib/os-release. The old location is automatically
6160 created as symlink. /usr/lib is the more appropriate
6161 location of this file, since it shall actually describe the
6162 vendor operating system shipped in /usr, and not the
6163 configuration stored in /etc.
6165 * .mount units gained a new boolean SloppyOptions= setting
6166 that maps to mount(8)'s -s option which enables permissive
6167 parsing of unknown mount options.
6169 * tmpfiles learnt a new "L+" directive which creates a symlink
6170 but (unlike "L") deletes a pre-existing file first, should
6171 it already exist and not already be the correct
6172 symlink. Similarly, "b+", "c+" and "p+" directives have been
6173 added as well, which create block and character devices, as
6174 well as fifos in the filesystem, possibly removing any
6175 pre-existing files of different types.
6177 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
6178 'argument' field (which so far specified the source to
6179 symlink/copy the files from) is now optional. If omitted the
6180 same file os copied from /usr/share/factory/ suffixed by the
6181 full destination path. This is useful for populating /etc
6182 with essential files, by copying them from vendor defaults
6183 shipped in /usr/share/factory/etc.
6185 * A new command "systemctl preset-all" has been added that
6186 applies the service preset settings to all installed unit
6187 files. A new switch --preset-mode= has been added that
6188 controls whether only enable or only disable operations
6191 * A new command "systemctl is-system-running" has been added
6192 that allows checking the overall state of the system, for
6193 example whether it is fully up and running.
6195 * When the system boots up with an empty /etc, the equivalent
6196 to "systemctl preset-all" is executed during early boot, to
6197 make sure all default services are enabled after a factory
6200 * systemd now contains a minimal preset file that enables the
6201 most basic services systemd ships by default.
6203 * Unit files' [Install] section gained a new DefaultInstance=
6204 field for defining the default instance to create if a
6205 template unit is enabled with no instance specified.
6207 * A new passive target cryptsetup-pre.target has been added
6208 that may be used by services that need to make they run and
6209 finish before the first LUKS cryptographic device is set up.
6211 * The /dev/loop-control and /dev/btrfs-control device nodes
6212 are now owned by the "disk" group by default, opening up
6213 access to this group.
6215 * systemd-coredump will now automatically generate a
6216 stack trace of all core dumps taking place on the system,
6217 based on elfutils' libdw library. This stack trace is logged
6220 * systemd-coredump may now optionally store coredumps directly
6221 on disk (in /var/lib/systemd/coredump, possibly compressed),
6222 instead of storing them unconditionally in the journal. This
6223 mode is the new default. A new configuration file
6224 /etc/systemd/coredump.conf has been added to configure this
6225 and other parameters of systemd-coredump.
6227 * coredumpctl gained a new "info" verb to show details about a
6228 specific coredump. A new switch "-1" has also been added
6229 that makes sure to only show information about the most
6230 recent entry instead of all entries. Also, as the tool is
6231 generally useful now the "systemd-" prefix of the binary
6232 name has been removed. Distributions that want to maintain
6233 compatibility with the old name should add a symlink from
6234 the old name to the new name.
6236 * journald's SplitMode= now defaults to "uid". This makes sure
6237 that unprivileged users can access their own coredumps with
6238 coredumpctl without restrictions.
6240 * New kernel command line options "systemd.wants=" (for
6241 pulling an additional unit during boot), "systemd.mask="
6242 (for masking a specific unit for the boot), and
6243 "systemd.debug-shell" (for enabling the debug shell on tty9)
6244 have been added. This is implemented in the new generator
6245 "systemd-debug-generator".
6247 * systemd-nspawn will now by default filter a couple of
6248 syscalls for containers, among them those required for
6249 kernel module loading, direct x86 IO port access, swap
6250 management, and kexec. Most importantly though
6251 open_by_handle_at() is now prohibited for containers,
6252 closing a hole similar to a recently discussed vulnerability
6253 in docker regarding access to files on file hierarchies the
6254 container should normally not have access to. Note that, for
6255 nspawn, we generally make no security claims anyway (and
6256 this is explicitly documented in the man page), so this is
6257 just a fix for one of the most obvious problems.
6259 * A new man page file-hierarchy(7) has been added that
6260 contains a minimized, modernized version of the file system
6261 layout systemd expects, similar in style to the FHS
6262 specification or hier(5). A new tool systemd-path(1) has
6263 been added to query many of these paths for the local
6266 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
6267 longer done. Since the directory now has a per-user size
6268 limit, and is cleaned on logout this appears unnecessary,
6269 in particular since this now brings the lifecycle of this
6270 directory closer in line with how IPC objects are handled.
6272 * systemd.pc now exports a number of additional directories,
6273 including $libdir (which is useful to identify the library
6274 path for the primary architecture of the system), and a
6275 couple of drop-in directories.
6277 * udev's predictable network interface names now use the dev_port
6278 sysfs attribute, introduced in linux 3.15 instead of dev_id to
6279 distinguish between ports of the same PCI function. dev_id should
6280 only be used for ports using the same HW address, hence the need
6283 * machined has been updated to export the OS version of a
6284 container (read from /etc/os-release and
6285 /usr/lib/os-release) on the bus. This is now shown in
6286 "machinectl status" for a machine.
6288 * A new service setting RestartForceExitStatus= has been
6289 added. If configured to a set of exit signals or process
6290 return values, the service will be restarted when the main
6291 daemon process exits with any of them, regardless of the
6294 * systemctl's -H switch for connecting to remote systemd
6295 machines has been extended so that it may be used to
6296 directly connect to a specific container on the
6297 host. "systemctl -H root@foobar:waldi" will now connect as
6298 user "root" to host "foobar", and then proceed directly to
6299 the container named "waldi". Note that currently you have to
6300 authenticate as user "root" for this to work, as entering
6301 containers is a privileged operation.
6303 Contributions from: Andreas Henriksson, Benjamin Steinwender,
6304 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
6305 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
6306 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
6307 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
6308 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
6309 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
6310 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
6311 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
6312 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
6313 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
6314 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
6316 — Berlin, 2014-07-03
6320 * As an experimental feature, udev now tries to lock the
6321 disk device node (flock(LOCK_SH|LOCK_NB)) while it
6322 executes events for the disk or any of its partitions.
6323 Applications like partitioning programs can lock the
6324 disk device node (flock(LOCK_EX)) and claim temporary
6325 device ownership that way; udev will entirely skip all event
6326 handling for this disk and its partitions. If the disk
6327 was opened for writing, the close will trigger a partition
6328 table rescan in udev's "watch" facility, and if needed
6329 synthesize "change" events for the disk and all its partitions.
6330 This is now unconditionally enabled, and if it turns out to
6331 cause major problems, we might turn it on only for specific
6332 devices, or might need to disable it entirely. Device Mapper
6333 devices are excluded from this logic.
6335 * We temporarily dropped the "-l" switch for fsck invocations,
6336 since they collide with the flock() logic above. util-linux
6337 upstream has been changed already to avoid this conflict,
6338 and we will re-add "-l" as soon as util-linux with this
6339 change has been released.
6341 * The dependency on libattr has been removed. Since a long
6342 time, the extended attribute calls have moved to glibc, and
6343 libattr is thus unnecessary.
6345 * Virtualization detection works without privileges now. This
6346 means the systemd-detect-virt binary no longer requires
6347 CAP_SYS_PTRACE file capabilities, and our daemons can run
6348 with fewer privileges.
6350 * systemd-networkd now runs under its own "systemd-network"
6351 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
6352 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
6353 loses the ability to write to files owned by root this way.
6355 * Similarly, systemd-resolved now runs under its own
6356 "systemd-resolve" user with no capabilities remaining.
6358 * Similarly, systemd-bus-proxyd now runs under its own
6359 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
6361 * systemd-networkd gained support for setting up "veth"
6362 virtual Ethernet devices for container connectivity, as well
6363 as GRE and VTI tunnels.
6365 * systemd-networkd will no longer automatically attempt to
6366 manually load kernel modules necessary for certain tunnel
6367 transports. Instead, it is assumed the kernel loads them
6368 automatically when required. This only works correctly on
6369 very new kernels. On older kernels, please consider adding
6370 the kernel modules to /etc/modules-load.d/ as a work-around.
6372 * The resolv.conf file systemd-resolved generates has been
6373 moved to /run/systemd/resolve/. If you have a symlink from
6374 /etc/resolv.conf, it might be necessary to correct it.
6376 * Two new service settings, ProtectHome= and ProtectSystem=,
6377 have been added. When enabled, they will make the user data
6378 (such as /home) inaccessible or read-only and the system
6379 (such as /usr) read-only, for specific services. This allows
6380 very light-weight per-service sandboxing to avoid
6381 modifications of user data or system files from
6382 services. These two new switches have been enabled for all
6383 of systemd's long-running services, where appropriate.
6385 * Socket units gained new SocketUser= and SocketGroup=
6386 settings to set the owner user and group of AF_UNIX sockets
6387 and FIFOs in the file system.
6389 * Socket units gained a new RemoveOnStop= setting. If enabled,
6390 all FIFOS and sockets in the file system will be removed
6391 when the specific socket unit is stopped.
6393 * Socket units gained a new Symlinks= setting. It takes a list
6394 of symlinks to create to file system sockets or FIFOs
6395 created by the specific Unix sockets. This is useful to
6396 manage symlinks to socket nodes with the same lifecycle as
6399 * The /dev/log socket and /dev/initctl FIFO have been moved to
6400 /run, and have been replaced by symlinks. This allows
6401 connecting to these facilities even if PrivateDevices=yes is
6402 used for a service (which makes /dev/log itself unavailable,
6403 but /run is left). This also has the benefit of ensuring
6404 that /dev only contains device nodes, directories and
6405 symlinks, and nothing else.
6407 * sd-daemon gained two new calls sd_pid_notify() and
6408 sd_pid_notifyf(). They are similar to sd_notify() and
6409 sd_notifyf(), but allow overriding of the source PID of
6410 notification messages if permissions permit this. This is
6411 useful to send notify messages on behalf of a different
6412 process (for example, the parent process). The
6413 systemd-notify tool has been updated to make use of this
6414 when sending messages (so that notification messages now
6415 originate from the shell script invoking systemd-notify and
6416 not the systemd-notify process itself. This should minimize
6417 a race where systemd fails to associate notification
6418 messages to services when the originating process already
6421 * A new "on-abnormal" setting for Restart= has been added. If
6422 set, it will result in automatic restarts on all "abnormal"
6423 reasons for a process to exit, which includes unclean
6424 signals, core dumps, timeouts and watchdog timeouts, but
6425 does not include clean and unclean exit codes or clean
6426 signals. Restart=on-abnormal is an alternative for
6427 Restart=on-failure for services that shall be able to
6428 terminate and avoid restarts on certain errors, by
6429 indicating so with an unclean exit code. Restart=on-failure
6430 or Restart=on-abnormal is now the recommended setting for
6431 all long-running services.
6433 * If the InaccessibleDirectories= service setting points to a
6434 mount point (or if there are any submounts contained within
6435 it), it is now attempted to completely unmount it, to make
6436 the file systems truly unavailable for the respective
6439 * The ReadOnlyDirectories= service setting and
6440 systemd-nspawn's --read-only parameter are now recursively
6441 applied to all submounts, too.
6443 * Mount units may now be created transiently via the bus APIs.
6445 * The support for SysV and LSB init scripts has been removed
6446 from the systemd daemon itself. Instead, it is now
6447 implemented as a generator that creates native systemd units
6448 from these scripts when needed. This enables us to remove a
6449 substantial amount of legacy code from PID 1, following the
6450 fact that many distributions only ship a very small number
6451 of LSB/SysV init scripts nowadays.
6453 * Privileged Xen (dom0) domains are not considered
6454 virtualization anymore by the virtualization detection
6455 logic. After all, they generally have unrestricted access to
6456 the hardware and usually are used to manage the unprivileged
6459 * systemd-tmpfiles gained a new "C" line type, for copying
6460 files or entire directories.
6462 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
6463 lines. So far, they have been non-globbing versions of the
6464 latter, and have thus been redundant. In future, it is
6465 recommended to only use "z". "m" has hence been removed
6466 from the documentation, even though it stays supported.
6468 * A tmpfiles snippet to recreate the most basic structure in
6469 /var has been added. This is enough to create the /var/run →
6470 /run symlink and create a couple of structural
6471 directories. This allows systems to boot up with an empty or
6472 volatile /var. Of course, while with this change, the core OS
6473 now is capable with dealing with a volatile /var, not all
6474 user services are ready for it. However, we hope that sooner
6475 or later, many service daemons will be changed upstream so
6476 that they are able to automatically create their necessary
6477 directories in /var at boot, should they be missing. This is
6478 the first step to allow state-less systems that only require
6479 the vendor image for /usr to boot.
6481 * systemd-nspawn has gained a new --tmpfs= switch to mount an
6482 empty tmpfs instance to a specific directory. This is
6483 particularly useful for making use of the automatic
6484 reconstruction of /var (see above), by passing --tmpfs=/var.
6486 * Access modes specified in tmpfiles snippets may now be
6487 prefixed with "~", which indicates that they shall be masked
6488 by whether the existing file or directory is currently
6489 writable, readable or executable at all. Also, if specified,
6490 the sgid/suid/sticky bits will be masked for all
6493 * A new passive target unit "network-pre.target" has been
6494 added which is useful for services that shall run before any
6495 network is configured, for example firewall scripts.
6497 * The "floppy" group that previously owned the /dev/fd*
6498 devices is no longer used. The "disk" group is now used
6499 instead. Distributions should probably deprecate usage of
6502 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
6503 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
6504 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
6505 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
6506 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
6507 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
6510 — Berlin, 2014-06-11
6514 * A new "systemd-timesyncd" daemon has been added for
6515 synchronizing the system clock across the network. It
6516 implements an SNTP client. In contrast to NTP
6517 implementations such as chrony or the NTP reference server,
6518 this only implements a client side, and does not bother with
6519 the full NTP complexity, focusing only on querying time from
6520 one remote server and synchronizing the local clock to
6521 it. Unless you intend to serve NTP to networked clients or
6522 want to connect to local hardware clocks, this simple NTP
6523 client should be more than appropriate for most
6524 installations. The daemon runs with minimal privileges, and
6525 has been hooked up with networkd to only operate when
6526 network connectivity is available. The daemon saves the
6527 current clock to disk every time a new NTP sync has been
6528 acquired, and uses this to possibly correct the system clock
6529 early at bootup, in order to accommodate for systems that
6530 lack an RTC such as the Raspberry Pi and embedded devices,
6531 and to make sure that time monotonically progresses on these
6532 systems, even if it is not always correct. To make use of
6533 this daemon, a new system user and group "systemd-timesync"
6534 needs to be created on installation of systemd.
6536 * The queue "seqnum" interface of libudev has been disabled, as
6537 it was generally incompatible with device namespacing as
6538 sequence numbers of devices go "missing" if the devices are
6539 part of a different namespace.
6541 * "systemctl list-timers" and "systemctl list-sockets" gained
6542 a --recursive switch for showing units of these types also
6543 for all local containers, similar in style to the already
6544 supported --recursive switch for "systemctl list-units".
6546 * A new RebootArgument= setting has been added for service
6547 units, which may be used to specify a kernel reboot argument
6548 to use when triggering reboots with StartLimitAction=.
6550 * A new FailureAction= setting has been added for service
6551 units which may be used to specify an operation to trigger
6552 when a service fails. This works similarly to
6553 StartLimitAction=, but unlike it, controls what is done
6554 immediately rather than only after several attempts to
6555 restart the service in question.
6557 * hostnamed got updated to also expose the kernel name,
6558 release, and version on the bus. This is useful for
6559 executing commands like hostnamectl with the -H switch.
6560 systemd-analyze makes use of this to properly display
6561 details when running non-locally.
6563 * The bootchart tool can now show cgroup information in the
6564 graphs it generates.
6566 * The CFS CPU quota cgroup attribute is now exposed for
6567 services. The new CPUQuota= switch has been added for this
6568 which takes a percentage value. Setting this will have the
6569 result that a service may never get more CPU time than the
6570 specified percentage, even if the machine is otherwise idle.
6572 * systemd-networkd learned IPIP and SIT tunnel support.
6574 * LSB init scripts exposing a dependency on $network will now
6575 get a dependency on network-online.target rather than simply
6576 network.target. This should bring LSB handling closer to
6577 what it was on SysV systems.
6579 * A new fsck.repair= kernel option has been added to control
6580 how fsck shall deal with unclean file systems at boot.
6582 * The (.ini) configuration file parser will now silently
6583 ignore sections whose name begins with "X-". This may be
6584 used to maintain application-specific extension sections in unit
6587 * machined gained a new API to query the IP addresses of
6588 registered containers. "machinectl status" has been updated
6589 to show these addresses in its output.
6591 * A new call sd_uid_get_display() has been added to the
6592 sd-login APIs for querying the "primary" session of a
6593 user. The "primary" session of the user is elected from the
6594 user's sessions and generally a graphical session is
6595 preferred over a text one.
6597 * A minimal systemd-resolved daemon has been added. It
6598 currently simply acts as a companion to systemd-networkd and
6599 manages resolv.conf based on per-interface DNS
6600 configuration, possibly supplied via DHCP. In the long run
6601 we hope to extend this into a local DNSSEC enabled DNS and
6604 * The systemd-networkd-wait-online tool is now enabled by
6605 default. It will delay network-online.target until a network
6606 connection has been configured. The tool primarily integrates
6607 with networkd, but will also make a best effort to make sense
6608 of network configuration performed in some other way.
6610 * Two new service options StartupCPUShares= and
6611 StartupBlockIOWeight= have been added that work similarly to
6612 CPUShares= and BlockIOWeight= however only apply during
6613 system startup. This is useful to prioritize certain services
6614 differently during bootup than during normal runtime.
6616 * hostnamed has been changed to prefer the statically
6617 configured hostname in /etc/hostname (unless set to
6618 'localhost' or empty) over any dynamic one supplied by
6619 dhcp. With this change, the rules for picking the hostname
6620 match more closely the rules of other configuration settings
6621 where the local administrator's configuration in /etc always
6622 overrides any other settings.
6624 Contributions from: Ali H. Caliskan, Alison Chaiken, Bas van
6625 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
6626 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
6627 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
6628 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
6629 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
6630 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
6631 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
6632 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
6633 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
6634 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
6635 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
6636 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
6637 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
6638 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
6639 Lindskog, WaLyong Cho, Will Woods, Zbigniew
6642 — Beijing, 2014-05-28
6646 * When restoring the screen brightness at boot, stay away from
6647 the darkest setting or from the lowest 5% of the available
6648 range, depending on which is the larger value of both. This
6649 should effectively protect the user from rebooting into a
6650 black screen, should the brightness have been set to minimum
6653 * sd-login gained a new sd_machine_get_class() call to
6654 determine the class ("vm" or "container") of a machine
6655 registered with machined.
6657 * sd-login gained new calls
6658 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
6659 to query the identity of the peer of a local AF_UNIX
6660 connection. They operate similarly to their sd_pid_get_xyz()
6663 * PID 1 will now maintain a system-wide system state engine
6664 with the states "starting", "running", "degraded",
6665 "maintenance", "stopping". These states are bound to system
6666 startup, normal runtime, runtime with at least one failed
6667 service, rescue/emergency mode and system shutdown. This
6668 state is shown in the "systemctl status" output when no unit
6669 name is passed. It is useful to determine system state, in
6670 particularly when doing so for many systems or containers at
6673 * A new command "list-machines" has been added to "systemctl"
6674 that lists all local OS containers and shows their system
6675 state (see above), if systemd runs inside of them.
6677 * systemctl gained a new "-r" switch to recursively enumerate
6678 units on all local containers, when used with the
6679 "list-unit" command (which is the default one that is
6680 executed when no parameters are specified).
6682 * The GPT automatic partition discovery logic will now honour
6683 two GPT partition flags: one may be set on a partition to
6684 cause it to be mounted read-only, and the other may be set
6685 on a partition to ignore it during automatic discovery.
6687 * Two new GPT type UUIDs have been added for automatic root
6688 partition discovery, for 32-bit and 64-bit ARM. This is not
6689 particularly useful for discovering the root directory on
6690 these architectures during bare-metal boots (since UEFI is
6691 not common there), but still very useful to allow booting of
6692 ARM disk images in nspawn with the -i option.
6694 * MAC addresses of interfaces created with nspawn's
6695 --network-interface= switch will now be generated from the
6696 machine name, and thus be stable between multiple invocations
6699 * logind will now automatically remove all IPC objects owned
6700 by a user if she or he fully logs out. This makes sure that
6701 users who are logged out cannot continue to consume IPC
6702 resources. This covers SysV memory, semaphores and message
6703 queues as well as POSIX shared memory and message
6704 queues. Traditionally, SysV and POSIX IPC had no lifecycle
6705 limits. With this functionality, that is corrected. This may
6706 be turned off by using the RemoveIPC= switch of logind.conf.
6708 * The systemd-machine-id-setup and tmpfiles tools gained a
6709 --root= switch to operate on a specific root directory,
6712 * journald can now forward logged messages to the TTYs of all
6713 logged in users ("wall"). This is the default for all
6714 emergency messages now.
6716 * A new tool systemd-journal-remote has been added to stream
6717 journal log messages across the network.
6719 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
6720 controller trees are mounted into it. Note that the
6721 directories mounted beneath it are not read-only. This is a
6722 security measure and is particularly useful because glibc
6723 actually includes a search logic to pick any tmpfs it can
6724 find to implement shm_open() if /dev/shm is not available
6725 (which it might very well be in namespaced setups).
6727 * machinectl gained a new "poweroff" command to cleanly power
6728 down a local OS container.
6730 * The PrivateDevices= unit file setting will now also drop the
6731 CAP_MKNOD capability from the capability bound set, and
6732 imply DevicePolicy=closed.
6734 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
6735 comprehensively on all long-running systemd services where
6736 this is appropriate.
6738 * systemd-udevd will now run in a disassociated mount
6739 namespace. To mount directories from udev rules, make sure to
6740 pull in mount units via SYSTEMD_WANTS properties.
6742 * The kdbus support gained support for uploading policy into
6743 the kernel. sd-bus gained support for creating "monitoring"
6744 connections that can eavesdrop into all bus communication
6745 for debugging purposes.
6747 * Timestamps may now be specified in seconds since the UNIX
6748 epoch Jan 1st, 1970 by specifying "@" followed by the value
6751 * Native tcpwrap support in systemd has been removed. tcpwrap
6752 is old code, not really maintained anymore and has serious
6753 shortcomings, and better options such as firewalls
6754 exist. For setups that require tcpwrap usage, please
6755 consider invoking your socket-activated service via tcpd,
6756 like on traditional inetd.
6758 * A new system.conf configuration option
6759 DefaultTimerAccuracySec= has been added that controls the
6760 default AccuracySec= setting of .timer units.
6762 * Timer units gained a new WakeSystem= switch. If enabled,
6763 timers configured this way will cause the system to resume
6764 from system suspend (if the system supports that, which most
6767 * Timer units gained a new Persistent= switch. If enabled,
6768 timers configured this way will save to disk when they have
6769 been last triggered. This information is then used on next
6770 reboot to possible execute overdue timer events, that
6771 could not take place because the system was powered off.
6772 This enables simple anacron-like behaviour for timer units.
6774 * systemctl's "list-timers" will now also list the time a
6775 timer unit was last triggered in addition to the next time
6776 it will be triggered.
6778 * systemd-networkd will now assign predictable IPv4LL
6779 addresses to its local interfaces.
6781 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
6782 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
6783 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
6784 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
6785 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
6786 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
6787 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
6788 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
6791 — Berlin, 2014-03-25
6795 * A new unit file setting RestrictAddressFamilies= has been
6796 added to restrict which socket address families unit
6797 processes gain access to. This takes address family names
6798 like "AF_INET" or "AF_UNIX", and is useful to minimize the
6799 attack surface of services via exotic protocol stacks. This
6800 is built on seccomp system call filters.
6802 * Two new unit file settings RuntimeDirectory= and
6803 RuntimeDirectoryMode= have been added that may be used to
6804 manage a per-daemon runtime directories below /run. This is
6805 an alternative for setting up directory permissions with
6806 tmpfiles snippets, and has the advantage that the runtime
6807 directory's lifetime is bound to the daemon runtime and that
6808 the daemon starts up with an empty directory each time. This
6809 is particularly useful when writing services that drop
6810 privileges using the User= or Group= setting.
6812 * The DeviceAllow= unit setting now supports globbing for
6813 matching against device group names.
6815 * The systemd configuration file system.conf gained new
6816 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
6817 DefaultMemoryAccounting= to globally turn on/off accounting
6818 for specific resources (cgroups) for all units. These
6819 settings may still be overridden individually in each unit
6822 * systemd-gpt-auto-generator is now able to discover /srv and
6823 root partitions in addition to /home and swap partitions. It
6824 also supports LUKS-encrypted partitions now. With this in
6825 place, automatic discovery of partitions to mount following
6826 the Discoverable Partitions Specification
6827 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
6828 is now a lot more complete. This allows booting without
6829 /etc/fstab and without root= on the kernel command line on
6830 systems prepared appropriately.
6832 * systemd-nspawn gained a new --image= switch which allows
6833 booting up disk images and Linux installations on any block
6834 device that follow the Discoverable Partitions Specification
6835 (see above). This means that installations made with
6836 appropriately updated installers may now be started and
6837 deployed using container managers, completely
6838 unmodified. (We hope that libvirt-lxc will add support for
6839 this feature soon, too.)
6841 * systemd-nspawn gained a new --network-macvlan= setting to
6842 set up a private macvlan interface for the
6843 container. Similarly, systemd-networkd gained a new
6844 Kind=macvlan setting in .netdev files.
6846 * systemd-networkd now supports configuring local addresses
6849 * A new tool systemd-network-wait-online has been added to
6850 synchronously wait for network connectivity using
6853 * The sd-bus.h bus API gained a new sd_bus_track object for
6854 tracking the lifecycle of bus peers. Note that sd-bus.h is
6855 still not a public API though (unless you specify
6856 --enable-kdbus on the configure command line, which however
6857 voids your warranty and you get no API stability guarantee).
6859 * The $XDG_RUNTIME_DIR runtime directories for each user are
6860 now individual tmpfs instances, which has the benefit of
6861 introducing separate pools for each user, with individual
6862 size limits, and thus making sure that unprivileged clients
6863 can no longer negatively impact the system or other users by
6864 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
6865 RuntimeDirectorySize= has been introduced that allows
6866 controlling the default size limit for all users. It
6867 defaults to 10% of the available physical memory. This is no
6868 replacement for quotas on tmpfs though (which the kernel
6869 still does not support), as /dev/shm and /tmp are still
6870 shared resources used by both the system and unprivileged
6873 * logind will now automatically turn off automatic suspending
6874 on laptop lid close when more than one display is
6875 connected. This was previously expected to be implemented
6876 individually in desktop environments (such as GNOME),
6877 however has been added to logind now, in order to fix a
6878 boot-time race where a desktop environment might not have
6879 been started yet and thus not been able to take an inhibitor
6880 lock at the time where logind already suspends the system
6881 due to a closed lid.
6883 * logind will now wait at least 30s after each system
6884 suspend/resume cycle, and 3min after system boot before
6885 suspending the system due to a closed laptop lid. This
6886 should give USB docking stations and similar enough time to
6887 be probed and configured after system resume and boot in
6888 order to then act as suspend blocker.
6890 * systemd-run gained a new --property= setting which allows
6891 initialization of resource control properties (and others)
6892 for the created scope or service unit. Example: "systemd-run
6893 --property=BlockIOWeight=10 updatedb" may be used to run
6894 updatedb at a low block IO scheduling weight.
6896 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
6897 now also work in --scope mode.
6899 * When systemd is compiled with kdbus support, basic support
6900 for enforced policies is now in place. (Note that enabling
6901 kdbus still voids your warranty and no API compatibility
6904 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
6905 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
6906 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
6907 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
6908 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
6909 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
6910 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
6911 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
6912 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
6913 Zbigniew Jędrzejewski-Szmek
6915 — Berlin, 2014-03-12
6919 * systemd will now relabel /dev after loading the SMACK policy
6920 according to SMACK rules.
6922 * A new unit file option AppArmorProfile= has been added to
6923 set the AppArmor profile for the processes of a unit.
6925 * A new condition check ConditionArchitecture= has been added
6926 to conditionalize units based on the system architecture, as
6927 reported by uname()'s "machine" field.
6929 * systemd-networkd now supports matching on the system
6930 virtualization, architecture, kernel command line, host name
6933 * logind is now a lot more aggressive when suspending the
6934 machine due to a closed laptop lid. Instead of acting only
6935 on the lid close action, it will continuously watch the lid
6936 status and act on it. This is useful for laptops where the
6937 power button is on the outside of the chassis so that it can
6938 be reached without opening the lid (such as the Lenovo
6939 Yoga). On those machines, logind will now immediately
6940 re-suspend the machine if the power button has been
6941 accidentally pressed while the laptop was suspended and in a
6942 backpack or similar.
6944 * logind will now watch SW_DOCK switches and inhibit reaction
6945 to the lid switch if it is pressed. This means that logind
6946 will not suspend the machine anymore if the lid is closed
6947 and the system is docked, if the laptop supports SW_DOCK
6948 notifications via the input layer. Note that ACPI docking
6949 stations do not generate this currently. Also note that this
6950 logic is usually not fully sufficient and Desktop
6951 Environments should take a lid switch inhibitor lock when an
6952 external display is connected, as systemd will not watch
6955 * nspawn will now make use of the devices cgroup controller by
6956 default, and only permit creation of and access to the usual
6957 API device nodes like /dev/null or /dev/random, as well as
6958 access to (but not creation of) the pty devices.
6960 * We will now ship a default .network file for
6961 systemd-networkd that automatically configures DHCP for
6962 network interfaces created by nspawn's --network-veth or
6963 --network-bridge= switches.
6965 * systemd will now understand the usual M, K, G, T suffixes
6966 according to SI conventions (i.e. to the base 1000) when
6967 referring to throughput and hardware metrics. It will stay
6968 with IEC conventions (i.e. to the base 1024) for software
6969 metrics, according to what is customary according to
6970 Wikipedia. We explicitly document which base applies for
6971 each configuration option.
6973 * The DeviceAllow= setting in unit files now supports a syntax
6974 to whitelist an entire group of devices node majors at once,
6975 based on the /proc/devices listing. For example, with the
6976 string "char-pts", it is now possible to whitelist all
6977 current and future pseudo-TTYs at once.
6979 * sd-event learned a new "post" event source. Event sources of
6980 this type are triggered by the dispatching of any event
6981 source of a type that is not "post". This is useful for
6982 implementing clean-up and check event sources that are
6983 triggered by other work being done in the program.
6985 * systemd-networkd is no longer statically enabled, but uses
6986 the usual [Install] sections so that it can be
6987 enabled/disabled using systemctl. It still is enabled by
6990 * When creating a veth interface pair with systemd-nspawn, the
6991 host side will now be prefixed with "vb-" if
6992 --network-bridge= is used, and with "ve-" if --network-veth
6993 is used. This way, it is easy to distinguish these cases on
6994 the host, for example to apply different configuration to
6995 them with systemd-networkd.
6997 * The compatibility libraries for libsystemd-journal.so,
6998 libsystem-id128.so, libsystemd-login.so and
6999 libsystemd-daemon.so do not make use of IFUNC
7000 anymore. Instead, we now build libsystemd.so multiple times
7001 under these alternative names. This means that the footprint
7002 is drastically increased, but given that these are
7003 transitional compatibility libraries, this should not matter
7004 much. This change has been made necessary to support the ARM
7005 platform for these compatibility libraries, as the ARM
7006 toolchain is not really at the same level as the toolchain
7007 for other architectures like x86 and does not support
7008 IFUNC. Please make sure to use --enable-compat-libs only
7009 during a transitional period!
7011 * The .include syntax has been deprecated and is not documented
7012 anymore. Drop-in files in .d directories should be used instead.
7014 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
7015 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
7016 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
7017 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
7018 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
7019 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
7020 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
7021 Zbigniew Jędrzejewski-Szmek
7023 — Berlin, 2014-02-24
7027 * A new component "systemd-networkd" has been added that can
7028 be used to configure local network interfaces statically or
7029 via DHCP. It is capable of bringing up bridges, VLANs, and
7030 bonding. Currently, no hook-ups for interactive network
7031 configuration are provided. Use this for your initrd,
7032 container, embedded, or server setup if you need a simple,
7033 yet powerful, network configuration solution. This
7034 configuration subsystem is quite nifty, as it allows wildcard
7035 hotplug matching in interfaces. For example, with a single
7036 configuration snippet, you can configure that all Ethernet
7037 interfaces showing up are automatically added to a bridge,
7038 or similar. It supports link-sensing and more.
7040 * A new tool "systemd-socket-proxyd" has been added which can
7041 act as a bidirectional proxy for TCP sockets. This is
7042 useful for adding socket activation support to services that
7043 do not actually support socket activation, including virtual
7044 machines and the like.
7046 * Add a new tool to save/restore rfkill state on
7049 * Save/restore state of keyboard backlights in addition to
7050 display backlights on shutdown/boot.
7052 * udev learned a new SECLABEL{} construct to label device
7053 nodes with a specific security label when they appear. For
7054 now, only SECLABEL{selinux} is supported, but the syntax is
7055 prepared for additional security frameworks.
7057 * udev gained a new scheme to configure link-level attributes
7058 from files in /etc/systemd/network/*.link. These files can
7059 match against MAC address, device path, driver name and type,
7060 and will apply attributes like the naming policy, link speed,
7061 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
7062 address assignment policy (randomized, ...).
7064 * The configuration of network interface naming rules for
7065 "permanent interface names" has changed: a new NamePolicy=
7066 setting in the [Link] section of .link files determines the
7067 priority of possible naming schemes (onboard, slot, MAC,
7068 path). The default value of this setting is determined by
7069 /usr/lib/net/links/99-default.link. Old
7070 80-net-name-slot.rules udev configuration file has been
7071 removed, so local configuration overriding this file should
7072 be adapted to override 99-default.link instead.
7074 * When the User= switch is used in a unit file, also
7075 initialize $SHELL= based on the user database entry.
7077 * systemd no longer depends on libdbus. All communication is
7078 now done with sd-bus, systemd's low-level bus library
7081 * kdbus support has been added to PID 1 itself. When kdbus is
7082 enabled, this causes PID 1 to set up the system bus and
7083 enable support for a new ".busname" unit type that
7084 encapsulates bus name activation on kdbus. It works a little
7085 bit like ".socket" units, except for bus names. A new
7086 generator has been added that converts classic dbus1 service
7087 activation files automatically into native systemd .busname
7090 * sd-bus: add a light-weight vtable implementation that allows
7091 defining objects on the bus with a simple static const
7092 vtable array of its methods, signals and properties.
7094 * systemd will not generate or install static dbus
7095 introspection data anymore to /usr/share/dbus-1/interfaces,
7096 as the precise format of these files is unclear, and
7097 nothing makes use of it.
7099 * A proxy daemon is now provided to proxy clients connecting
7100 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
7101 compatibility with classic D-Bus.
7103 * A bus driver implementation has been added that supports the
7104 classic D-Bus bus driver calls on kdbus, also for
7105 compatibility purposes.
7107 * A new API "sd-event.h" has been added that implements a
7108 minimal event loop API built around epoll. It provides a
7109 couple of features that direct epoll usage is lacking:
7110 prioritization of events, scales to large numbers of timer
7111 events, per-event timer slack (accuracy), system-wide
7112 coalescing of timer events, exit handlers, watchdog
7113 supervision support using systemd's sd_notify() API, child
7116 * A new API "sd-rntl.h" has been added that provides an API
7117 around the route netlink interface of the kernel, similar in
7118 style to "sd-bus.h".
7120 * A new API "sd-dhcp-client.h" has been added that provides a
7121 small DHCPv4 client-side implementation. This is used by
7124 * There is a new kernel command line option
7125 "systemd.restore_state=0|1". When set to "0", none of the
7126 systemd tools will restore saved runtime state to hardware
7127 devices. More specifically, the rfkill and backlight states
7130 * The FsckPassNo= compatibility option in mount/service units
7131 has been removed. The fstab generator will now add the
7132 necessary dependencies automatically, and does not require
7133 PID1's support for that anymore.
7135 * journalctl gained a new switch, --list-boots, that lists
7136 recent boots with their times and boot IDs.
7138 * The various tools like systemctl, loginctl, timedatectl,
7139 busctl, systemd-run, ... have gained a new switch "-M" to
7140 connect to a specific, local OS container (as direct
7141 connection, without requiring SSH). This works on any
7142 container that is registered with machined, such as those
7143 created by libvirt-lxc or nspawn.
7145 * systemd-run and systemd-analyze also gained support for "-H"
7146 to connect to remote hosts via SSH. This is particularly
7147 useful for systemd-run because it enables queuing of jobs
7148 onto remote systems.
7150 * machinectl gained a new command "login" to open a getty
7151 login in any local container. This works with any container
7152 that is registered with machined (such as those created by
7153 libvirt-lxc or nspawn), and which runs systemd inside.
7155 * machinectl gained a new "reboot" command that may be used to
7156 trigger a reboot on a specific container that is registered
7157 with machined. This works on any container that runs an init
7158 system of some kind.
7160 * systemctl gained a new "list-timers" command to print a nice
7161 listing of installed timer units with the times they elapse
7164 * Alternative reboot() parameters may now be specified on the
7165 "systemctl reboot" command line and are passed to the
7166 reboot() system call.
7168 * systemctl gained a new --job-mode= switch to configure the
7169 mode to queue a job with. This is a more generic version of
7170 --fail, --irreversible, and --ignore-dependencies, which are
7171 still available but not advertised anymore.
7173 * /etc/systemd/system.conf gained new settings to configure
7174 various default timeouts of units, as well as the default
7175 start limit interval and burst. These may still be overridden
7178 * PID1 will now export on the bus profile data of the security
7179 policy upload process (such as the SELinux policy upload to
7182 * journald: when forwarding logs to the console, include
7183 timestamps (following the setting in
7184 /sys/module/printk/parameters/time).
7186 * OnCalendar= in timer units now understands the special
7187 strings "yearly" and "annually". (Both are equivalent)
7189 * The accuracy of timer units is now configurable with the new
7190 AccuracySec= setting. It defaults to 1min.
7192 * A new dependency type JoinsNamespaceOf= has been added that
7193 allows running two services within the same /tmp and network
7194 namespace, if PrivateNetwork= or PrivateTmp= are used.
7196 * A new command "cat" has been added to systemctl. It outputs
7197 the original unit file of a unit, and concatenates the
7198 contents of additional "drop-in" unit file snippets, so that
7199 the full configuration is shown.
7201 * systemctl now supports globbing on the various "list-xyz"
7202 commands, like "list-units" or "list-sockets", as well as on
7203 those commands which take multiple unit names.
7205 * journalctl's --unit= switch gained support for globbing.
7207 * All systemd daemons now make use of the watchdog logic so
7208 that systemd automatically notices when they hang.
7210 * If the $container_ttys environment variable is set,
7211 getty-generator will automatically spawn a getty for each
7212 listed tty. This is useful for container managers to request
7213 login gettys to be spawned on as many ttys as needed.
7215 * %h, %s, %U specifier support is not available anymore when
7216 used in unit files for PID 1. This is because NSS calls are
7217 not safe from PID 1. They stay available for --user
7218 instances of systemd, and as special case for the root user.
7220 * loginctl gained a new "--no-legend" switch to turn off output
7223 * The "sd-login.h" API gained three new calls:
7224 sd_session_is_remote(), sd_session_get_remote_user(),
7225 sd_session_get_remote_host() to query information about
7228 * The udev hardware database now also carries vendor/product
7229 information of SDIO devices.
7231 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
7232 determine whether watchdog notifications are requested by
7235 * Socket-activated per-connection services now include a
7236 short description of the connection parameters in the
7239 * tmpfiles gained a new "--boot" option. When this is not used,
7240 only lines where the command character is not suffixed with
7241 "!" are executed. When this option is specified, those
7242 options are executed too. This partitions tmpfiles
7243 directives into those that can be safely executed at any
7244 time, and those which should be run only at boot (for
7245 example, a line that creates /run/nologin).
7247 * A new API "sd-resolve.h" has been added which provides a simple
7248 asynchronous wrapper around glibc NSS host name resolution
7249 calls, such as getaddrinfo(). In contrast to glibc's
7250 getaddrinfo_a(), it does not use signals. In contrast to most
7251 other asynchronous name resolution libraries, this one does
7252 not reimplement DNS, but reuses NSS, so that alternate
7253 host name resolution systems continue to work, such as mDNS,
7254 LDAP, etc. This API is based on libasyncns, but it has been
7255 cleaned up for inclusion in systemd.
7257 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
7258 "sd-daemon.h" are no longer found in individual libraries
7259 libsystemd-journal.so, libsystemd-login.so,
7260 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
7261 merged them into a single library, libsystemd.so, which
7262 provides all symbols. The reason for this is cyclic
7263 dependencies, as these libraries tend to use each other's
7264 symbols. So far, we have managed to workaround that by linking
7265 a copy of a good part of our code into each of these
7266 libraries again and again, which, however, makes certain
7267 things hard to do, like sharing static variables. Also, it
7268 substantially increases footprint. With this change, there
7269 is only one library for the basic APIs systemd
7270 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
7271 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
7272 library as well, however are subject to the --enable-kdbus
7273 switch (see below). Note that "sd-dhcp-client.h" is not part
7274 of this library (this is because it only consumes, never
7275 provides, services of/to other APIs). To make the transition
7276 easy from the separate libraries to the unified one, we
7277 provide the --enable-compat-libs compile-time switch which
7278 will generate stub libraries that are compatible with the
7279 old ones but redirect all calls to the new one.
7281 * All of the kdbus logic and the new APIs "sd-bus.h",
7282 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
7283 and "sd-utf8.h" are compile-time optional via the
7284 "--enable-kdbus" switch, and they are not compiled in by
7285 default. To make use of kdbus, you have to explicitly enable
7286 the switch. Note however, that neither the kernel nor the
7287 userspace API for all of this is considered stable yet. We
7288 want to maintain the freedom to still change the APIs for
7289 now. By specifying this build-time switch, you acknowledge
7290 that you are aware of the instability of the current
7293 * Also, note that while kdbus is pretty much complete,
7294 it lacks one thing: proper policy support. This means you
7295 can build a fully working system with all features; however,
7296 it will be highly insecure. Policy support will be added in
7297 one of the next releases, at the same time that we will
7298 declare the APIs stable.
7300 * When the kernel command line argument "kdbus" is specified,
7301 systemd will automatically load the kdbus.ko kernel module. At
7302 this stage of development, it is only useful for testing kdbus
7303 and should not be used in production. Note: if "--enable-kdbus"
7304 is specified, and the kdbus.ko kernel module is available, and
7305 "kdbus" is added to the kernel command line, the entire system
7306 runs with kdbus instead of dbus-daemon, with the above mentioned
7307 problem of missing the system policy enforcement. Also a future
7308 version of kdbus.ko or a newer systemd will not be compatible with
7309 each other, and will unlikely be able to boot the machine if only
7310 one of them is updated.
7312 * systemctl gained a new "import-environment" command which
7313 uploads the caller's environment (or parts thereof) into the
7314 service manager so that it is inherited by services started
7315 by the manager. This is useful to upload variables like
7316 $DISPLAY into the user service manager.
7318 * A new PrivateDevices= switch has been added to service units
7319 which allows running a service with a namespaced /dev
7320 directory that does not contain any device nodes for
7321 physical devices. More specifically, it only includes devices
7322 such as /dev/null, /dev/urandom, and /dev/zero which are API
7325 * logind has been extended to support behaviour like VT
7326 switching on seats that do not support a VT. This makes
7327 multi-session available on seats that are not the first seat
7328 (seat0), and on systems where kernel support for VTs has
7329 been disabled at compile-time.
7331 * If a process holds a delay lock for system sleep or shutdown
7332 and fails to release it in time, we will now log its
7333 identity. This makes it easier to identify processes that
7334 cause slow suspends or power-offs.
7336 * When parsing /etc/crypttab, support for a new key-slot=
7337 option as supported by Debian is added. It allows indicating
7338 which LUKS slot to use on disk, speeding up key loading.
7340 * The sd_journal_sendv() API call has been checked and
7341 officially declared to be async-signal-safe so that it may
7342 be invoked from signal handlers for logging purposes.
7344 * Boot-time status output is now enabled automatically after a
7345 short timeout if boot does not progress, in order to give
7346 the user an indication what she or he is waiting for.
7348 * The boot-time output has been improved to show how much time
7349 remains until jobs expire.
7351 * The KillMode= switch in service units gained a new possible
7352 value "mixed". If set, and the unit is shut down, then the
7353 initial SIGTERM signal is sent only to the main daemon
7354 process, while the following SIGKILL signal is sent to
7355 all remaining processes of the service.
7357 * When a scope unit is registered, a new property "Controller"
7358 may be set. If set to a valid bus name, systemd will send a
7359 RequestStop() signal to this name when it would like to shut
7360 down the scope. This may be used to hook manager logic into
7361 the shutdown logic of scope units. Also, scope units may now
7362 be put in a special "abandoned" state, in which case the
7363 manager process which created them takes no further
7364 responsibilities for it.
7366 * When reading unit files, systemd will now verify
7367 the access mode of these files, and warn about certain
7368 suspicious combinations. This has been added to make it
7369 easier to track down packaging bugs where unit files are
7370 marked executable or world-writable.
7372 * systemd-nspawn gained a new "--setenv=" switch to set
7373 container-wide environment variables. The similar option in
7374 systemd-activate was renamed from "--environment=" to
7375 "--setenv=" for consistency.
7377 * systemd-nspawn has been updated to create a new kdbus domain
7378 for each container that is invoked, thus allowing each
7379 container to have its own set of system and user buses,
7380 independent of the host.
7382 * systemd-nspawn gained a new --drop-capability= switch to run
7383 the container with less capabilities than the default. Both
7384 --drop-capability= and --capability= now take the special
7385 string "all" for dropping or keeping all capabilities.
7387 * systemd-nspawn gained new switches for executing containers
7388 with specific SELinux labels set.
7390 * systemd-nspawn gained a new --quiet switch to not generate
7391 any additional output but the container's own console
7394 * systemd-nspawn gained a new --share-system switch to run a
7395 container without PID namespacing enabled.
7397 * systemd-nspawn gained a new --register= switch to control
7398 whether the container is registered with systemd-machined or
7399 not. This is useful for containers that do not run full
7400 OS images, but only specific apps.
7402 * systemd-nspawn gained a new --keep-unit which may be used
7403 when invoked as the only program from a service unit, and
7404 results in registration of the unit service itself in
7405 systemd-machined, instead of a newly opened scope unit.
7407 * systemd-nspawn gained a new --network-interface= switch for
7408 moving arbitrary interfaces to the container. The new
7409 --network-veth switch creates a virtual Ethernet connection
7410 between host and container. The new --network-bridge=
7411 switch then allows assigning the host side of this virtual
7412 Ethernet connection to a bridge device.
7414 * systemd-nspawn gained a new --personality= switch for
7415 setting the kernel personality for the container. This is
7416 useful when running a 32-bit container on a 64-bit host. A
7417 similar option Personality= is now also available for service
7420 * logind will now also track a "Desktop" identifier for each
7421 session which encodes the desktop environment of it. This is
7422 useful for desktop environments that want to identify
7423 multiple running sessions of itself easily.
7425 * A new SELinuxContext= setting for service units has been
7426 added that allows setting a specific SELinux execution
7427 context for a service.
7429 * Most systemd client tools will now honour $SYSTEMD_LESS for
7430 settings of the "less" pager. By default, these tools will
7431 override $LESS to allow certain operations to work, such as
7432 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
7433 influence this logic.
7435 * systemd's "seccomp" hook-up has been changed to make use of
7436 the libseccomp library instead of using its own
7437 implementation. This has benefits for portability among
7440 * For usage together with SystemCallFilter=, a new
7441 SystemCallErrorNumber= setting has been introduced that
7442 allows configuration of a system error number to be returned
7443 on filtered system calls, instead of immediately killing the
7444 process. Also, SystemCallArchitectures= has been added to
7445 limit access to system calls of a particular architecture
7446 (in order to turn off support for unused secondary
7447 architectures). There is also a global
7448 SystemCallArchitectures= setting in system.conf now to turn
7449 off support for non-native system calls system-wide.
7451 * systemd requires a kernel with a working name_to_handle_at(),
7452 please see the kernel config requirements in the README file.
7454 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
7455 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
7456 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
7457 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
7458 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
7459 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
7460 Elia Pinto, Florian Weimer, George McCollister, Goffredo
7461 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
7462 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
7463 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
7464 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
7465 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
7466 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
7467 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
7468 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
7469 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
7470 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
7471 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
7472 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
7473 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
7474 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
7475 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
7476 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
7477 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
7479 — Berlin, 2014-02-20
7483 * logind has gained support for facilitating privileged input
7484 and drm device access for unprivileged clients. This work is
7485 useful to allow Wayland display servers (and similar
7486 programs, such as kmscon) to run under the user's ID and
7487 access input and drm devices which are normally
7488 protected. When this is used (and the kernel is new enough)
7489 logind will "mute" IO on the file descriptors passed to
7490 Wayland as long as it is in the background and "unmute" it
7491 if it returns into the foreground. This allows secure
7492 session switching without allowing background sessions to
7493 eavesdrop on input and display data. This also introduces
7494 session switching support if VT support is turned off in the
7495 kernel, and on seats that are not seat0.
7497 * A new kernel command line option luks.options= is understood
7498 now which allows specifying LUKS options for usage for LUKS
7499 encrypted partitions specified with luks.uuid=.
7501 * tmpfiles.d(5) snippets may now use specifier expansion in
7502 path names. More specifically %m, %b, %H, %v, are now
7503 replaced by the local machine id, boot id, hostname, and
7504 kernel version number.
7506 * A new tmpfiles.d(5) command "m" has been introduced which
7507 may be used to change the owner/group/access mode of a file
7508 or directory if it exists, but do nothing if it does not.
7510 * This release removes high-level support for the
7511 MemorySoftLimit= cgroup setting. The underlying kernel
7512 cgroup attribute memory.soft_limit= is currently badly
7513 designed and likely to be removed from the kernel API in its
7514 current form, hence we should not expose it for now.
7516 * The memory.use_hierarchy cgroup attribute is now enabled for
7517 all cgroups systemd creates in the memory cgroup
7518 hierarchy. This option is likely to be come the built-in
7519 default in the kernel anyway, and the non-hierarchical mode
7520 never made much sense in the intrinsically hierarchical
7523 * A new field _SYSTEMD_SLICE= is logged along with all journal
7524 messages containing the slice a message was generated
7525 from. This is useful to allow easy per-customer filtering of
7526 logs among other things.
7528 * systemd-journald will no longer adjust the group of journal
7529 files it creates to the "systemd-journal" group. Instead we
7530 rely on the journal directory to be owned by the
7531 "systemd-journal" group, and its setgid bit set, so that the
7532 kernel file system layer will automatically enforce that
7533 journal files inherit this group assignment. The reason for
7534 this change is that we cannot allow NSS look-ups from
7535 journald which would be necessary to resolve
7536 "systemd-journal" to a numeric GID, because this might
7537 create deadlocks if NSS involves synchronous queries to
7538 other daemons (such as nscd, or sssd) which in turn are
7539 logging clients of journald and might block on it, which
7540 would then dead lock. A tmpfiles.d(5) snippet included in
7541 systemd will make sure the setgid bit and group are
7542 properly set on the journal directory if it exists on every
7543 boot. However, we recommend adjusting it manually after
7544 upgrades too (or from RPM scriptlets), so that the change is
7545 not delayed until next reboot.
7547 * Backlight and random seed files in /var/lib/ have moved into
7548 the /var/lib/systemd/ directory, in order to centralize all
7549 systemd generated files in one directory.
7551 * Boot time performance measurements (as displayed by
7552 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
7553 performance information if that's available to determine how
7554 much time BIOS and boot loader initialization required. With
7555 a sufficiently new BIOS you hence no longer need to boot
7556 with Gummiboot to get access to such information.
7558 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
7559 Cristian Rodríguez, Dave Reisner, David Herrmann, David
7560 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
7561 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
7562 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
7563 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
7564 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
7566 — Berlin, 2013-10-02
7570 * The Restart= option for services now understands a new
7571 on-watchdog setting, which will restart the service
7572 automatically if the service stops sending out watchdog keep
7573 alive messages (as configured with WatchdogSec=).
7575 * The getty generator (which is responsible for bringing up a
7576 getty on configured serial consoles) will no longer only
7577 start a getty on the primary kernel console but on all
7578 others, too. This makes the order in which console= is
7579 specified on the kernel command line less important.
7581 * libsystemd-logind gained a new sd_session_get_vt() call to
7582 retrieve the VT number of a session.
7584 * If the option "tries=0" is set for an entry of /etc/crypttab
7585 its passphrase is queried indefinitely instead of any
7586 maximum number of tries.
7588 * If a service with a configure PID file terminates its PID
7589 file will now be removed automatically if it still exists
7590 afterwards. This should put an end to stale PID files.
7592 * systemd-run will now also take relative binary path names
7593 for execution and no longer insists on absolute paths.
7595 * InaccessibleDirectories= and ReadOnlyDirectories= now take
7596 paths that are optionally prefixed with "-" to indicate that
7597 it should not be considered a failure if they do not exist.
7599 * journalctl -o (and similar commands) now understands a new
7600 output mode "short-precise", it is similar to "short" but
7601 shows timestamps with usec accuracy.
7603 * The option "discard" (as known from Debian) is now
7604 synonymous to "allow-discards" in /etc/crypttab. In fact,
7605 "discard" is preferred now (since it is easier to remember
7608 * Some licensing clean-ups were made, so that more code is now
7609 LGPL-2.1 licensed than before.
7611 * A minimal tool to save/restore the display backlight
7612 brightness across reboots has been added. It will store the
7613 backlight setting as late as possible at shutdown, and
7614 restore it as early as possible during reboot.
7616 * A logic to automatically discover and enable home and swap
7617 partitions on GPT disks has been added. With this in place
7618 /etc/fstab becomes optional for many setups as systemd can
7619 discover certain partitions located on the root disk
7620 automatically. Home partitions are recognized under their
7621 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
7622 partitions are recognized under their GPT type ID
7623 0657fd6da4ab43c484e50933c84b4f4f.
7625 * systemd will no longer pass any environment from the kernel
7626 or initrd to system services. If you want to set an
7627 environment for all services, do so via the kernel command
7628 line systemd.setenv= assignment.
7630 * The systemd-sysctl tool no longer natively reads the file
7631 /etc/sysctl.conf. If desired, the file should be symlinked
7632 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
7633 legacy support by a symlink rather than built-in code, it
7634 also makes the otherwise hidden order of application of the
7635 different files visible. (Note that this partly reverts to a
7636 pre-198 application order of sysctl knobs!)
7638 * The "systemctl set-log-level" and "systemctl dump" commands
7639 have been moved to systemd-analyze.
7641 * systemd-run learned the new --remain-after-exit switch,
7642 which causes the scope unit not to be cleaned up
7643 automatically after the process terminated.
7645 * tmpfiles learned a new --exclude-prefix= switch to exclude
7646 certain paths from operation.
7648 * journald will now automatically flush all messages to disk
7649 as soon as a message at the log level CRIT, ALERT or EMERG
7652 Contributions from: Andrew Cook, Brandon Philips, Christian
7653 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
7654 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
7655 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
7656 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
7657 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
7658 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
7659 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
7660 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
7661 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
7662 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
7663 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
7664 William Giokas, Zbigniew Jędrzejewski-Szmek
7666 — Berlin, 2013-09-13
7670 * The documentation has been updated to cover the various new
7671 concepts introduced with 205.
7673 * Unit files now understand the new %v specifier which
7674 resolves to the kernel version string as returned by "uname
7677 * systemctl now supports filtering the unit list output by
7678 load state, active state and sub state, using the new
7681 * "systemctl status" will now show the results of the
7682 condition checks (like ConditionPathExists= and similar) of
7683 the last start attempts of the unit. They are also logged to
7686 * "journalctl -b" may now be used to look for boot output of a
7687 specific boot. Try "journalctl -b -1" for the previous boot,
7688 but the syntax is substantially more powerful.
7690 * "journalctl --show-cursor" has been added which prints the
7691 cursor string the last shown log line. This may then be used
7692 with the new "journalctl --after-cursor=" switch to continue
7693 browsing logs from that point on.
7695 * "journalctl --force" may now be used to force regeneration
7698 * Creation of "dead" device nodes has been moved from udev
7699 into kmod and tmpfiles. Previously, udev would read the kmod
7700 databases to pre-generate dead device nodes based on meta
7701 information contained in kernel modules, so that these would
7702 be auto-loaded on access rather then at boot. As this
7703 does not really have much to do with the exposing actual
7704 kernel devices to userspace this has always been slightly
7705 alien in the udev codebase. Following the new scheme kmod
7706 will now generate a runtime snippet for tmpfiles from the
7707 module meta information and it now is tmpfiles' job to the
7708 create the nodes. This also allows overriding access and
7709 other parameters for the nodes using the usual tmpfiles
7710 facilities. As side effect this allows us to remove the
7711 CAP_SYS_MKNOD capability bit from udevd entirely.
7713 * logind's device ACLs may now be applied to these "dead"
7714 devices nodes too, thus finally allowing managed access to
7715 devices such as /dev/snd/sequencer without loading the
7716 backing module right-away.
7718 * A new RPM macro has been added that may be used to apply
7719 tmpfiles configuration during package installation.
7721 * systemd-detect-virt and ConditionVirtualization= now can
7722 detect User-Mode-Linux machines (UML).
7724 * journald will now implicitly log the effective capabilities
7725 set of processes in the message metadata.
7727 * systemd-cryptsetup has gained support for TrueCrypt volumes.
7729 * The initrd interface has been simplified (more specifically,
7730 support for passing performance data via environment
7731 variables and fsck results via files in /run has been
7732 removed). These features were non-essential, and are
7733 nowadays available in a much nicer way by having systemd in
7734 the initrd serialize its state and have the hosts systemd
7735 deserialize it again.
7737 * The udev "keymap" data files and tools to apply keyboard
7738 specific mappings of scan to key codes, and force-release
7739 scan code lists have been entirely replaced by a udev
7740 "keyboard" builtin and a hwdb data file.
7742 * systemd will now honour the kernel's "quiet" command line
7743 argument also during late shutdown, resulting in a
7744 completely silent shutdown when used.
7746 * There's now an option to control the SO_REUSEPORT socket
7747 option in .socket units.
7749 * Instance units will now automatically get a per-template
7750 subslice of system.slice unless something else is explicitly
7751 configured. For example, instances of sshd@.service will now
7752 implicitly be placed in system-sshd.slice rather than
7753 system.slice as before.
7755 * Test coverage support may now be enabled at build time.
7757 Contributions from: Dave Reisner, Frederic Crozat, Harald
7758 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
7759 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
7760 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
7761 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
7762 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
7763 Giokas, Zbigniew Jędrzejewski-Szmek
7765 — Berlin, 2013-07-23
7769 * Two new unit types have been introduced:
7771 Scope units are very similar to service units, however, are
7772 created out of pre-existing processes — instead of PID 1
7773 forking off the processes. By using scope units it is
7774 possible for system services and applications to group their
7775 own child processes (worker processes) in a powerful way
7776 which then maybe used to organize them, or kill them
7777 together, or apply resource limits on them.
7779 Slice units may be used to partition system resources in an
7780 hierarchical fashion and then assign other units to them. By
7781 default there are now three slices: system.slice (for all
7782 system services), user.slice (for all user sessions),
7783 machine.slice (for VMs and containers).
7785 Slices and scopes have been introduced primarily in
7786 context of the work to move cgroup handling to a
7787 single-writer scheme, where only PID 1
7788 creates/removes/manages cgroups.
7790 * There's a new concept of "transient" units. In contrast to
7791 normal units these units are created via an API at runtime,
7792 not from configuration from disk. More specifically this
7793 means it is now possible to run arbitrary programs as
7794 independent services, with all execution parameters passed
7795 in via bus APIs rather than read from disk. Transient units
7796 make systemd substantially more dynamic then it ever was,
7797 and useful as a general batch manager.
7799 * logind has been updated to make use of scope and slice units
7800 for managing user sessions. As a user logs in he will get
7801 his own private slice unit, to which all sessions are added
7802 as scope units. We also added support for automatically
7803 adding an instance of user@.service for the user into the
7804 slice. Effectively logind will no longer create cgroup
7805 hierarchies on its own now, it will defer entirely to PID 1
7806 for this by means of scope, service and slice units. Since
7807 user sessions this way become entities managed by PID 1
7808 the output of "systemctl" is now a lot more comprehensive.
7810 * A new mini-daemon "systemd-machined" has been added which
7811 may be used by virtualization managers to register local
7812 VMs/containers. nspawn has been updated accordingly, and
7813 libvirt will be updated shortly. machined will collect a bit
7814 of meta information about the VMs/containers, and assign
7815 them their own scope unit (see above). The collected
7816 meta-data is then made available via the "machinectl" tool,
7817 and exposed in "ps" and similar tools. machined/machinectl
7818 is compile-time optional.
7820 * As discussed earlier, the low-level cgroup configuration
7821 options ControlGroup=, ControlGroupModify=,
7822 ControlGroupPersistent=, ControlGroupAttribute= have been
7823 removed. Please use high-level attribute settings instead as
7824 well as slice units.
7826 * A new bus call SetUnitProperties() has been added to alter
7827 various runtime parameters of a unit. This is primarily
7828 useful to alter cgroup parameters dynamically in a nice way,
7829 but will be extended later on to make more properties
7830 modifiable at runtime. systemctl gained a new set-properties
7831 command that wraps this call.
7833 * A new tool "systemd-run" has been added which can be used to
7834 run arbitrary command lines as transient services or scopes,
7835 while configuring a number of settings via the command
7836 line. This tool is currently very basic, however already
7837 very useful. We plan to extend this tool to even allow
7838 queuing of execution jobs with time triggers from the
7839 command line, similar in fashion to "at".
7841 * nspawn will now inform the user explicitly that kernels with
7842 audit enabled break containers, and suggest the user to turn
7845 * Support for detecting the IMA and AppArmor security
7846 frameworks with ConditionSecurity= has been added.
7848 * journalctl gained a new "-k" switch for showing only kernel
7849 messages, mimicking dmesg output; in addition to "--user"
7850 and "--system" switches for showing only user's own logs
7853 * systemd-delta can now show information about drop-in
7854 snippets extending unit files.
7856 * libsystemd-bus has been substantially updated but is still
7857 not available as public API.
7859 * systemd will now look for the "debug" argument on the kernel
7860 command line and enable debug logging, similar to what
7861 "systemd.log_level=debug" already did before.
7863 * "systemctl set-default", "systemctl get-default" has been
7864 added to configure the default.target symlink, which
7865 controls what to boot into by default.
7867 * "systemctl set-log-level" has been added as a convenient
7868 way to raise and lower systemd logging threshold.
7870 * "systemd-analyze plot" will now show the time the various
7871 generators needed for execution, as well as information
7872 about the unit file loading.
7874 * libsystemd-journal gained a new sd_journal_open_files() call
7875 for opening specific journal files. journactl also gained a
7876 new switch to expose this new functionality. Previously we
7877 only supported opening all files from a directory, or all
7878 files from the system, as opening individual files only is
7879 racy due to journal file rotation.
7881 * systemd gained the new DefaultEnvironment= setting in
7882 /etc/systemd/system.conf to set environment variables for
7885 * If a privileged process logs a journal message with the
7886 OBJECT_PID= field set, then journald will automatically
7887 augment this with additional OBJECT_UID=, OBJECT_GID=,
7888 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
7889 system services want to log events about specific client
7890 processes. journactl/systemctl has been updated to make use
7891 of this information if all log messages regarding a specific
7894 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
7895 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
7896 Reisner, David Coppa, David King, David Strauss, Eelco
7897 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
7898 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
7899 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
7900 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
7901 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
7902 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
7903 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
7904 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
7905 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
7906 Łukasz Stelmach, 장동준
7910 * The Python bindings gained some minimal support for the APIs
7911 exposed by libsystemd-logind.
7913 * ConditionSecurity= gained support for detecting SMACK. Since
7914 this condition already supports SELinux and AppArmor we only
7915 miss IMA for this. Patches welcome!
7917 Contributions from: Karol Lewandowski, Lennart Poettering,
7918 Zbigniew Jędrzejewski-Szmek
7922 * systemd-nspawn will now create /etc/resolv.conf if
7923 necessary, before bind-mounting the host's file onto it.
7925 * systemd-nspawn will now store meta information about a
7926 container on the container's cgroup as extended attribute
7927 fields, including the root directory.
7929 * The cgroup hierarchy has been reworked in many ways. All
7930 objects any of the components systemd creates in the cgroup
7931 tree are now suffixed. More specifically, user sessions are
7932 now placed in cgroups suffixed with ".session", users in
7933 cgroups suffixed with ".user", and nspawn containers in
7934 cgroups suffixed with ".nspawn". Furthermore, all cgroup
7935 names are now escaped in a simple scheme to avoid collision
7936 of userspace object names with kernel filenames. This work
7937 is preparation for making these objects relocatable in the
7938 cgroup tree, in order to allow easy resource partitioning of
7939 these objects without causing naming conflicts.
7941 * systemctl list-dependencies gained the new switches
7942 --plain, --reverse, --after and --before.
7944 * systemd-inhibit now shows the process name of processes that
7945 have taken an inhibitor lock.
7947 * nss-myhostname will now also resolve "localhost"
7948 implicitly. This makes /etc/hosts an optional file and
7949 nicely handles that on IPv6 ::1 maps to both "localhost" and
7952 * libsystemd-logind.so gained a new call
7953 sd_get_machine_names() to enumerate running containers and
7954 VMs (currently only supported by very new libvirt and
7955 nspawn). sd_login_monitor can now be used to watch
7956 VMs/containers coming and going.
7958 * .include is not allowed recursively anymore, and only in
7959 unit files. Usually it is better to use drop-in snippets in
7960 .d/*.conf anyway, as introduced with systemd 198.
7962 * systemd-analyze gained a new "critical-chain" command that
7963 determines the slowest chain of units run during system
7964 boot-up. It is very useful for tracking down where
7965 optimizing boot time is the most beneficial.
7967 * systemd will no longer allow manipulating service paths in
7968 the name=systemd:/system cgroup tree using ControlGroup= in
7969 units. (But is still fine with it in all other dirs.)
7971 * There's a new systemd-nspawn@.service service file that may
7972 be used to easily run nspawn containers as system
7973 services. With the container's root directory in
7974 /var/lib/container/foobar it is now sufficient to run
7975 "systemctl start systemd-nspawn@foobar.service" to boot it.
7977 * systemd-cgls gained a new parameter "--machine" to list only
7978 the processes within a certain container.
7980 * ConditionSecurity= now can check for "apparmor". We still
7981 are lacking checks for SMACK and IMA for this condition
7982 check though. Patches welcome!
7984 * A new configuration file /etc/systemd/sleep.conf has been
7985 added that may be used to configure which kernel operation
7986 systemd is supposed to execute when "suspend", "hibernate"
7987 or "hybrid-sleep" is requested. This makes the new kernel
7988 "freeze" state accessible to the user.
7990 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
7991 the passed argument if applicable.
7993 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
7994 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
7995 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
7996 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
7997 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
7998 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
7999 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
8004 * The output of 'systemctl list-jobs' got some polishing. The
8005 '--type=' argument may now be passed more than once. A new
8006 command 'systemctl list-sockets' has been added which shows
8007 a list of kernel sockets systemd is listening on with the
8008 socket units they belong to, plus the units these socket
8011 * The experimental libsystemd-bus library got substantial
8012 updates to work in conjunction with the (also experimental)
8013 kdbus kernel project. It works well enough to exchange
8014 messages with some sophistication. Note that kdbus is not
8015 ready yet, and the library is mostly an elaborate test case
8016 for now, and not installable.
8018 * systemd gained a new unit 'systemd-static-nodes.service'
8019 that generates static device nodes earlier during boot, and
8020 can run in conjunction with udev.
8022 * libsystemd-login gained a new call sd_pid_get_user_unit()
8023 to retrieve the user systemd unit a process is running
8024 in. This is useful for systems where systemd is used as
8027 * systemd-nspawn now places all containers in the new /machine
8028 top-level cgroup directory in the name=systemd
8029 hierarchy. libvirt will soon do the same, so that we get a
8030 uniform separation of /system, /user and /machine for system
8031 services, user processes and containers/virtual
8032 machines. This new cgroup hierarchy is also useful to stick
8033 stable names to specific container instances, which can be
8034 recognized later this way (this name may be controlled
8035 via systemd-nspawn's new -M switch). libsystemd-login also
8036 gained a new call sd_pid_get_machine_name() to retrieve the
8037 name of the container/VM a specific process belongs to.
8039 * bootchart can now store its data in the journal.
8041 * libsystemd-journal gained a new call
8042 sd_journal_add_conjunction() for AND expressions to the
8043 matching logic. This can be used to express more complex
8044 logical expressions.
8046 * journactl can now take multiple --unit= and --user-unit=
8049 * The cryptsetup logic now understands the "luks.key=" kernel
8050 command line switch for specifying a file to read the
8051 decryption key from. Also, if a configured key file is not
8052 found the tool will now automatically fall back to prompting
8055 * Python systemd.journal module was updated to wrap recently
8056 added functions from libsystemd-journal. The interface was
8057 changed to bring the low level interface in s.j._Reader
8058 closer to the C API, and the high level interface in
8059 s.j.Reader was updated to wrap and convert all data about
8062 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
8063 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
8064 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
8065 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
8066 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
8067 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
8071 * journalctl --update-catalog now understands a new --root=
8072 option to operate on catalogs found in a different root
8075 * During shutdown after systemd has terminated all running
8076 services a final killing loop kills all remaining left-over
8077 processes. We will now print the name of these processes
8078 when we send SIGKILL to them, since this usually indicates a
8081 * If /etc/crypttab refers to password files stored on
8082 configured mount points automatic dependencies will now be
8083 generated to ensure the specific mount is established first
8084 before the key file is attempted to be read.
8086 * 'systemctl status' will now show information about the
8087 network sockets a socket unit is listening on.
8089 * 'systemctl status' will also shown information about any
8090 drop-in configuration file for units. (Drop-In configuration
8091 files in this context are files such as
8092 /etc/systemd/systemd/foobar.service.d/*.conf)
8094 * systemd-cgtop now optionally shows summed up CPU times of
8095 cgroups. Press '%' while running cgtop to switch between
8096 percentage and absolute mode. This is useful to determine
8097 which cgroups use up the most CPU time over the entire
8098 runtime of the system. systemd-cgtop has also been updated
8099 to be 'pipeable' for processing with further shell tools.
8101 * 'hostnamectl set-hostname' will now allow setting of FQDN
8104 * The formatting and parsing of time span values has been
8105 changed. The parser now understands fractional expressions
8106 such as "5.5h". The formatter will now output fractional
8107 expressions for all time spans under 1min, i.e. "5.123456s"
8108 rather than "5s 123ms 456us". For time spans under 1s
8109 millisecond values are shown, for those under 1ms
8110 microsecond values are shown. This should greatly improve
8111 all time-related output of systemd.
8113 * libsystemd-login and libsystemd-journal gained new
8114 functions for querying the poll() events mask and poll()
8115 timeout value for integration into arbitrary event
8118 * localectl gained the ability to list available X11 keymaps
8119 (models, layouts, variants, options).
8121 * 'systemd-analyze dot' gained the ability to filter for
8122 specific units via shell-style globs, to create smaller,
8123 more useful graphs. I.e. it is now possible to create simple
8124 graphs of all the dependencies between only target units, or
8125 of all units that Avahi has dependencies with.
8127 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
8128 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
8129 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
8130 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
8131 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
8132 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
8133 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
8137 * The boot-time readahead implementation for rotating media
8138 will now read the read-ahead data in multiple passes which
8139 consist of all read requests made in equidistant time
8140 intervals. This means instead of strictly reading read-ahead
8141 data in its physical order on disk we now try to find a
8142 middle ground between physical and access time order.
8144 * /etc/os-release files gained a new BUILD_ID= field for usage
8145 on operating systems that provide continuous builds of OS
8148 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
8149 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
8150 William Douglas, Zbigniew Jędrzejewski-Szmek
8154 * systemd-python gained an API exposing libsystemd-daemon.
8156 * The SMACK setup logic gained support for uploading CIPSO
8159 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
8160 ReadOnlyDirectories= and InaccessibleDirectories= has
8161 changed. The private /tmp and /var/tmp directories are now
8162 shared by all processes of a service (which means
8163 ExecStartPre= may now leave data in /tmp that ExecStart= of
8164 the same service can still access). When a service is
8165 stopped its temporary directories are immediately deleted
8166 (normal clean-up with tmpfiles is still done in addition to
8169 * By default, systemd will now set a couple of sysctl
8170 variables in the kernel: the safe sysrq options are turned
8171 on, IP route verification is turned on, and source routing
8172 disabled. The recently added hardlink and softlink
8173 protection of the kernel is turned on. These settings should
8174 be reasonably safe, and good defaults for all new systems.
8176 * The predictable network naming logic may now be turned off
8177 with a new kernel command line switch: net.ifnames=0.
8179 * A new libsystemd-bus module has been added that implements a
8180 pretty complete D-Bus client library. For details see:
8182 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
8184 * journald will now explicitly flush the journal files to disk
8185 at the latest 5min after each write. The file will then also
8186 be marked offline until the next write. This should increase
8187 reliability in case of a crash. The synchronization delay
8188 can be configured via SyncIntervalSec= in journald.conf.
8190 * There's a new remote-fs-setup.target unit that can be used
8191 to pull in specific services when at least one remote file
8192 system is to be mounted.
8194 * There are new targets timers.target and paths.target as
8195 canonical targets to pull user timer and path units in
8196 from. This complements sockets.target with a similar
8197 purpose for socket units.
8199 * libudev gained a new call udev_device_set_attribute_value()
8200 to set sysfs attributes of a device.
8202 * The udev daemon now sets the default number of worker
8203 processes executed in parallel based on the number of available
8204 CPUs instead of the amount of available RAM. This is supposed
8205 to provide a more reliable default and limit a too aggressive
8206 parallelism for setups with 1000s of devices connected.
8208 Contributions from: Auke Kok, Colin Walters, Cristian
8209 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
8210 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
8211 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
8212 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
8213 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
8214 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
8215 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
8216 Zbigniew Jędrzejewski-Szmek
8220 * Configuration of unit files may now be extended via drop-in
8221 files without having to edit/override the unit files
8222 themselves. More specifically, if the administrator wants to
8223 change one value for a service file foobar.service he can
8224 now do so by dropping in a configuration snippet into
8225 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
8226 will load all these snippets and apply them on top of the
8227 main unit configuration file, possibly extending or
8228 overriding its settings. Using these drop-in snippets is
8229 generally nicer than the two earlier options for changing
8230 unit files locally: copying the files from
8231 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
8232 them there; or creating a new file in /etc/systemd/system/
8233 that incorporates the original one via ".include". Drop-in
8234 snippets into these .d/ directories can be placed in any
8235 directory systemd looks for units in, and the usual
8236 overriding semantics between /usr/lib, /etc and /run apply
8239 * Most unit file settings which take lists of items can now be
8240 reset by assigning the empty string to them. For example,
8241 normally, settings such as Environment=FOO=BAR append a new
8242 environment variable assignment to the environment block,
8243 each time they are used. By assigning Environment= the empty
8244 string the environment block can be reset to empty. This is
8245 particularly useful with the .d/*.conf drop-in snippets
8246 mentioned above, since this adds the ability to reset list
8247 settings from vendor unit files via these drop-ins.
8249 * systemctl gained a new "list-dependencies" command for
8250 listing the dependencies of a unit recursively.
8252 * Inhibitors are now honored and listed by "systemctl
8253 suspend", "systemctl poweroff" (and similar) too, not only
8254 GNOME. These commands will also list active sessions by
8257 * Resource limits (as exposed by the various control group
8258 controllers) can now be controlled dynamically at runtime
8259 for all units. More specifically, you can now use a command
8260 like "systemctl set-cgroup-attr foobar.service cpu.shares
8261 2000" to alter the CPU shares a specific service gets. These
8262 settings are stored persistently on disk, and thus allow the
8263 administrator to easily adjust the resource usage of
8264 services with a few simple commands. This dynamic resource
8265 management logic is also available to other programs via the
8266 bus. Almost any kernel cgroup attribute and controller is
8269 * systemd-vconsole-setup will now copy all font settings to
8270 all allocated VTs, where it previously applied them only to
8273 * libsystemd-login gained the new sd_session_get_tty() API
8276 * This release drops support for a few legacy or
8277 distribution-specific LSB facility names when parsing init
8278 scripts: $x-display-manager, $mail-transfer-agent,
8279 $mail-transport-agent, $mail-transfer-agent, $smtp,
8280 $null. Also, the mail-transfer-agent.target unit backing
8281 this has been removed. Distributions which want to retain
8282 compatibility with this should carry the burden for
8283 supporting this themselves and patch support for these back
8284 in, if they really need to. Also, the facilities $syslog and
8285 $local_fs are now ignored, since systemd does not support
8286 early-boot LSB init scripts anymore, and these facilities
8287 are implied anyway for normal services. syslog.target has
8290 * There are new bus calls on PID1's Manager object for
8291 cancelling jobs, and removing snapshot units. Previously,
8292 both calls were only available on the Job and Snapshot
8295 * systemd-journal-gatewayd gained SSL support.
8297 * The various "environment" files, such as /etc/locale.conf
8298 now support continuation lines with a backslash ("\") as
8299 last character in the line, similarly in style (but different)
8300 to how this is supported in shells.
8302 * For normal user processes the _SYSTEMD_USER_UNIT= field is
8303 now implicitly appended to every log entry logged. systemctl
8304 has been updated to filter by this field when operating on a
8305 user systemd instance.
8307 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
8308 CAP_AUDIT_CONTROL capabilities to the capabilities set for
8309 the container. This makes it easier to boot unmodified
8310 Fedora systems in a container, which however still requires
8311 audit=0 to be passed on the kernel command line. Auditing in
8312 kernel and userspace is unfortunately still too broken in
8313 context of containers, hence we recommend compiling it out
8314 of the kernel or using audit=0. Hopefully this will be fixed
8315 one day for good in the kernel.
8317 * nspawn gained the new --bind= and --bind-ro= parameters to
8318 bind mount specific directories from the host into the
8321 * nspawn will now mount its own devpts file system instance
8322 into the container, in order not to leak pty devices from
8323 the host into the container.
8325 * systemd will now read the firmware boot time performance
8326 information from the EFI variables, if the used boot loader
8327 supports this, and takes it into account for boot performance
8328 analysis via "systemd-analyze". This is currently supported
8329 only in conjunction with Gummiboot, but could be supported
8330 by other boot loaders too. For details see:
8332 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
8334 * A new generator has been added that automatically mounts the
8335 EFI System Partition (ESP) to /boot, if that directory
8336 exists, is empty, and no other file system has been
8337 configured to be mounted there.
8339 * logind will now send out PrepareForSleep(false) out
8340 unconditionally, after coming back from suspend. This may be
8341 used by applications as asynchronous notification for
8342 system resume events.
8344 * "systemctl unlock-sessions" has been added, that allows
8345 unlocking the screens of all user sessions at once, similar
8346 to how "systemctl lock-sessions" already locked all users
8347 sessions. This is backed by a new D-Bus call UnlockSessions().
8349 * "loginctl seat-status" will now show the master device of a
8350 seat. (i.e. the device of a seat that needs to be around for
8351 the seat to be considered available, usually the graphics
8354 * tmpfiles gained a new "X" line type, that allows
8355 configuration of files and directories (with wildcards) that
8356 shall be excluded from automatic cleanup ("aging").
8358 * udev default rules set the device node permissions now only
8359 at "add" events, and do not change them any longer with a
8360 later "change" event.
8362 * The log messages for lid events and power/sleep keypresses
8363 now carry a message ID.
8365 * We now have a substantially larger unit test suite, but this
8366 continues to be work in progress.
8368 * udevadm hwdb gained a new --root= parameter to change the
8369 root directory to operate relative to.
8371 * logind will now issue a background sync() request to the kernel
8372 early at shutdown, so that dirty buffers are flushed to disk early
8373 instead of at the last moment, in order to optimize shutdown
8376 * A new bootctl tool has been added that is an interface for
8377 certain boot loader operations. This is currently a preview
8378 and is likely to be extended into a small mechanism daemon
8379 like timedated, localed, hostnamed, and can be used by
8380 graphical UIs to enumerate available boot options, and
8381 request boot into firmware operations.
8383 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
8384 the rest of the package. It also has been updated to work
8385 correctly in initrds.
8387 * polkit previously has been runtime optional, and is now also
8388 compile time optional via a configure switch.
8390 * systemd-analyze has been reimplemented in C. Also "systemctl
8391 dot" has moved into systemd-analyze.
8393 * "systemctl status" with no further parameters will now print
8394 the status of all active or failed units.
8396 * Operations such as "systemctl start" can now be executed
8397 with a new mode "--irreversible" which may be used to queue
8398 operations that cannot accidentally be reversed by a later
8399 job queuing. This is by default used to make shutdown
8400 requests more robust.
8402 * The Python API of systemd now gained a new module for
8403 reading journal files.
8405 * A new tool kernel-install has been added that can install
8406 kernel images according to the Boot Loader Specification:
8408 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
8410 * Boot time console output has been improved to provide
8411 animated boot time output for hanging jobs.
8413 * A new tool systemd-activate has been added which can be used
8414 to test socket activation with, directly from the command
8415 line. This should make it much easier to test and debug
8416 socket activation in daemons.
8418 * journalctl gained a new "--reverse" (or -r) option to show
8419 journal output in reverse order (i.e. newest line first).
8421 * journalctl gained a new "--pager-end" (or -e) option to jump
8422 to immediately jump to the end of the journal in the
8423 pager. This is only supported in conjunction with "less".
8425 * journalctl gained a new "--user-unit=" option, that works
8426 similarly to "--unit=" but filters for user units rather than
8429 * A number of unit files to ease adoption of systemd in
8430 initrds has been added. This moves some minimal logic from
8431 the various initrd implementations into systemd proper.
8433 * The journal files are now owned by a new group
8434 "systemd-journal", which exists specifically to allow access
8435 to the journal, and nothing else. Previously, we used the
8436 "adm" group for that, which however possibly covers more
8437 than just journal/log file access. This new group is now
8438 already used by systemd-journal-gatewayd to ensure this
8439 daemon gets access to the journal files and as little else
8440 as possible. Note that "make install" will also set FS ACLs
8441 up for /var/log/journal to give "adm" and "wheel" read
8442 access to it, in addition to "systemd-journal" which owns
8443 the journal files. We recommend that packaging scripts also
8444 add read access to "adm" + "wheel" to /var/log/journal, and
8445 all existing/future journal files. To normal users and
8446 administrators little changes, however packagers need to
8447 ensure to create the "systemd-journal" system group at
8448 package installation time.
8450 * The systemd-journal-gatewayd now runs as unprivileged user
8451 systemd-journal-gateway:systemd-journal-gateway. Packaging
8452 scripts need to create these system user/group at
8455 * timedated now exposes a new boolean property CanNTP that
8456 indicates whether a local NTP service is available or not.
8458 * systemd-detect-virt will now also detect xen PVs
8460 * The pstore file system is now mounted by default, if it is
8463 * In addition to the SELinux and IMA policies we will now also
8464 load SMACK policies at early boot.
8466 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
8467 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
8468 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
8469 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
8470 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
8471 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
8472 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
8473 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
8474 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
8475 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
8476 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
8477 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
8478 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
8479 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
8483 * Timer units now support calendar time events in addition to
8484 monotonic time events. That means you can now trigger a unit
8485 based on a calendar time specification such as "Thu,Fri
8486 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
8487 or fifth day of any month of the year 2013, given that it is
8488 a thursday or friday. This brings timer event support
8489 considerably closer to cron's capabilities. For details on
8490 the supported calendar time specification language see
8493 * udev now supports a number of different naming policies for
8494 network interfaces for predictable names, and a combination
8495 of these policies is now the default. Please see this wiki
8496 document for details:
8498 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
8500 * Auke Kok's bootchart implementation has been added to the
8501 systemd tree. It is an optional component that can graph the
8502 boot in quite some detail. It is one of the best bootchart
8503 implementations around and minimal in its code and
8506 * nss-myhostname has been integrated into the systemd source
8507 tree. nss-myhostname guarantees that the local hostname
8508 always stays resolvable via NSS. It has been a weak
8509 requirement of systemd-hostnamed since a long time, and
8510 since its code is actually trivial we decided to just
8511 include it in systemd's source tree. It can be turned off
8512 with a configure switch.
8514 * The read-ahead logic is now capable of properly detecting
8515 whether a btrfs file system is on SSD or rotating media, in
8516 order to optimize the read-ahead scheme. Previously, it was
8517 only capable of detecting this on traditional file systems
8520 * In udev, additional device properties are now read from the
8521 IAB in addition to the OUI database. Also, Bluetooth company
8522 identities are attached to the devices as well.
8524 * In service files %U may be used as specifier that is
8525 replaced by the configured user name of the service.
8527 * nspawn may now be invoked without a controlling TTY. This
8528 makes it suitable for invocation as its own service. This
8529 may be used to set up a simple containerized server system
8530 using only core OS tools.
8532 * systemd and nspawn can now accept socket file descriptors
8533 when they are started for socket activation. This enables
8534 implementation of socket activated nspawn
8535 containers. i.e. think about autospawning an entire OS image
8536 when the first SSH or HTTP connection is received. We expect
8537 that similar functionality will also be added to libvirt-lxc
8540 * journalctl will now suppress ANSI color codes when
8541 presenting log data.
8543 * systemctl will no longer show control group information for
8544 a unit if the control group is empty anyway.
8546 * logind can now automatically suspend/hibernate/shutdown the
8549 * /etc/machine-info and hostnamed now also expose the chassis
8550 type of the system. This can be used to determine whether
8551 the local system is a laptop, desktop, handset or
8552 tablet. This information may either be configured by the
8553 user/vendor or is automatically determined from ACPI and DMI
8554 information if possible.
8556 * A number of polkit actions are now bound together with "imply"
8557 rules. This should simplify creating UIs because many actions
8558 will now authenticate similar ones as well.
8560 * Unit files learnt a new condition ConditionACPower= which
8561 may be used to conditionalize a unit depending on whether an
8562 AC power source is connected or not, of whether the system
8563 is running on battery power.
8565 * systemctl gained a new "is-failed" verb that may be used in
8566 shell scripts and suchlike to check whether a specific unit
8567 is in the "failed" state.
8569 * The EnvironmentFile= setting in unit files now supports file
8570 globbing, and can hence be used to easily read a number of
8571 environment files at once.
8573 * systemd will no longer detect and recognize specific
8574 distributions. All distribution-specific #ifdeffery has been
8575 removed, systemd is now fully generic and
8576 distribution-agnostic. Effectively, not too much is lost as
8577 a lot of the code is still accessible via explicit configure
8578 switches. However, support for some distribution specific
8579 legacy configuration file formats has been dropped. We
8580 recommend distributions to simply adopt the configuration
8581 files everybody else uses now and convert the old
8582 configuration from packaging scripts. Most distributions
8583 already did that. If that's not possible or desirable,
8584 distributions are welcome to forward port the specific
8585 pieces of code locally from the git history.
8587 * When logging a message about a unit systemd will now always
8588 log the unit name in the message meta data.
8590 * localectl will now also discover system locale data that is
8591 not stored in locale archives, but directly unpacked.
8593 * logind will no longer unconditionally use framebuffer
8594 devices as seat masters, i.e. as devices that are required
8595 to be existing before a seat is considered preset. Instead,
8596 it will now look for all devices that are tagged as
8597 "seat-master" in udev. By default, framebuffer devices will
8598 be marked as such, but depending on local systems, other
8599 devices might be marked as well. This may be used to
8600 integrate graphics cards using closed source drivers (such
8601 as NVidia ones) more nicely into logind. Note however, that
8602 we recommend using the open source NVidia drivers instead,
8603 and no udev rules for the closed-source drivers will be
8604 shipped from us upstream.
8606 Contributions from: Adam Williamson, Alessandro Crismani, Auke
8607 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
8608 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
8609 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
8610 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
8611 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
8612 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
8613 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
8614 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
8615 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
8616 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
8617 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
8622 * udev gained support for loading additional device properties
8623 from an indexed database that is keyed by vendor/product IDs
8624 and similar device identifiers. For the beginning this
8625 "hwdb" is populated with data from the well-known PCI and
8626 USB database, but also includes PNP, ACPI and OID data. In
8627 the longer run this indexed database shall grow into
8628 becoming the one central database for non-essential
8629 userspace device metadata. Previously, data from the PCI/USB
8630 database was only attached to select devices, since the
8631 lookup was a relatively expensive operation due to O(n) time
8632 complexity (with n being the number of entries in the
8633 database). Since this is now O(1), we decided to add in this
8634 data for all devices where this is available, by
8635 default. Note that the indexed database needs to be rebuilt
8636 when new data files are installed. To achieve this you need
8637 to update your packaging scripts to invoke "udevadm hwdb
8638 --update" after installation of hwdb data files. For
8639 RPM-based distributions we introduced the new
8640 %udev_hwdb_update macro for this purpose.
8642 * The Journal gained support for the "Message Catalog", an
8643 indexed database to link up additional information with
8644 journal entries. For further details please check:
8646 https://www.freedesktop.org/wiki/Software/systemd/catalog
8648 The indexed message catalog database also needs to be
8649 rebuilt after installation of message catalog files. Use
8650 "journalctl --update-catalog" for this. For RPM-based
8651 distributions we introduced the %journal_catalog_update
8652 macro for this purpose.
8654 * The Python Journal bindings gained support for the standard
8655 Python logging framework.
8657 * The Journal API gained new functions for checking whether
8658 the underlying file system of a journal file is capable of
8659 properly reporting file change notifications, or whether
8660 applications that want to reflect journal changes "live"
8661 need to recheck journal files continuously in appropriate
8664 * It is now possible to set the "age" field for tmpfiles
8665 entries to 0, indicating that files matching this entry
8666 shall always be removed when the directories are cleaned up.
8668 * coredumpctl gained a new "gdb" verb which invokes gdb
8669 right-away on the selected coredump.
8671 * There's now support for "hybrid sleep" on kernels that
8672 support this, in addition to "suspend" and "hibernate". Use
8673 "systemctl hybrid-sleep" to make use of this.
8675 * logind's HandleSuspendKey= setting (and related settings)
8676 now gained support for a new "lock" setting to simply
8677 request the screen lock on all local sessions, instead of
8678 actually executing a suspend or hibernation.
8680 * systemd will now mount the EFI variables file system by
8683 * Socket units now gained support for configuration of the
8684 SMACK security label.
8686 * timedatectl will now output the time of the last and next
8687 daylight saving change.
8689 * We dropped support for various legacy and distro-specific
8690 concepts, such as insserv, early-boot SysV services
8691 (i.e. those for non-standard runlevels such as 'b' or 'S')
8692 or ArchLinux /etc/rc.conf support. We recommend the
8693 distributions who still need support this to either continue
8694 to maintain the necessary patches downstream, or find a
8695 different solution. (Talk to us if you have questions!)
8697 * Various systemd components will now bypass polkit checks for
8698 root and otherwise handle properly if polkit is not found to
8699 be around. This should fix most issues for polkit-less
8700 systems. Quite frankly this should have been this way since
8701 day one. It is absolutely our intention to make systemd work
8702 fine on polkit-less systems, and we consider it a bug if
8703 something does not work as it should if polkit is not around.
8705 * For embedded systems it is now possible to build udev and
8706 systemd without blkid and/or kmod support.
8708 * "systemctl switch-root" is now capable of switching root
8709 more than once. I.e. in addition to transitions from the
8710 initrd to the host OS it is now possible to transition to
8711 further OS images from the host. This is useful to implement
8712 offline updating tools.
8714 * Various other additions have been made to the RPM macros
8715 shipped with systemd. Use %udev_rules_update() after
8716 installing new udev rules files. %_udevhwdbdir,
8717 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
8718 %_sysctldir are now available which resolve to the right
8719 directories for packages to place various data files in.
8721 * journalctl gained the new --full switch (in addition to
8722 --all, to disable ellipsation for long messages.
8724 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
8725 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
8726 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
8727 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
8728 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
8729 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
8730 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
8731 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
8732 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
8736 * journalctl gained new --since= and --until= switches to
8737 filter by time. It also now supports nice filtering for
8738 units via --unit=/-u.
8740 * Type=oneshot services may use ExecReload= and do the
8743 * The journal daemon now supports time-based rotation and
8744 vacuuming, in addition to the usual disk-space based
8747 * The journal will now index the available field values for
8748 each field name. This enables clients to show pretty drop
8749 downs of available match values when filtering. The bash
8750 completion of journalctl has been updated
8751 accordingly. journalctl gained a new switch -F to list all
8752 values a certain field takes in the journal database.
8754 * More service events are now written as structured messages
8755 to the journal, and made recognizable via message IDs.
8757 * The timedated, localed and hostnamed mini-services which
8758 previously only provided support for changing time, locale
8759 and hostname settings from graphical DEs such as GNOME now
8760 also have a minimal (but very useful) text-based client
8761 utility each. This is probably the nicest way to changing
8762 these settings from the command line now, especially since
8763 it lists available options and is fully integrated with bash
8766 * There's now a new tool "systemd-coredumpctl" to list and
8767 extract coredumps from the journal.
8769 * We now install a README each in /var/log/ and
8770 /etc/rc.d/init.d explaining where the system logs and init
8771 scripts went. This hopefully should help folks who go to
8772 that dirs and look into the otherwise now empty void and
8773 scratch their heads.
8775 * When user-services are invoked (by systemd --user) the
8776 $MANAGERPID env var is set to the PID of systemd.
8778 * SIGRTMIN+24 when sent to a --user instance will now result
8779 in immediate termination of systemd.
8781 * gatewayd received numerous feature additions such as a
8782 "follow" mode, for live syncing and filtering.
8784 * browse.html now allows filtering and showing detailed
8785 information on specific entries. Keyboard navigation and
8786 mouse screen support has been added.
8788 * gatewayd/journalctl now supports HTML5/JSON
8789 Server-Sent-Events as output.
8791 * The SysV init script compatibility logic will now
8792 heuristically determine whether a script supports the
8793 "reload" verb, and only then make this available as
8796 * "systemctl status --follow" has been removed, use "journalctl
8799 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
8800 have been removed since they are hardly useful to be
8803 * And I'd like to take the opportunity to specifically mention
8804 Zbigniew for his great contributions. Zbigniew, you rock!
8806 Contributions from: Andrew Eikum, Christian Hesse, Colin
8807 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
8808 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
8809 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
8810 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
8811 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
8812 Jędrzejewski-Szmek, Сковорода Никита Андреевич
8816 * If /etc/vconsole.conf is non-existent or empty we will no
8817 longer load any console font or key map at boot by
8818 default. Instead the kernel defaults will be left
8819 intact. This is definitely the right thing to do, as no
8820 configuration should mean no configuration, and hard-coding
8821 font names that are different on all archs is probably a bad
8822 idea. Also, the kernel default key map and font should be
8823 good enough for most cases anyway, and mostly identical to
8824 the userspace fonts/key maps we previously overloaded them
8825 with. If distributions want to continue to default to a
8826 non-kernel font or key map they should ship a default
8827 /etc/vconsole.conf with the appropriate contents.
8829 Contributions from: Colin Walters, Daniel J Walsh, Dave
8830 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
8831 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
8835 * journalctl gained a new --cursor= switch to show entries
8836 starting from the specified location in the journal.
8838 * We now enforce a size limit on journal entry fields exported
8839 with "-o json" in journalctl. Fields larger than 4K will be
8840 assigned null. This can be turned off with --all.
8842 * An (optional) journal gateway daemon is now available as
8843 "systemd-journal-gatewayd.service". This service provides
8844 access to the journal via HTTP and JSON. This functionality
8845 will be used to implement live log synchronization in both
8846 pull and push modes, but has various other users too, such
8847 as easy log access for debugging of embedded devices. Right
8848 now it is already useful to retrieve the journal via HTTP:
8850 # systemctl start systemd-journal-gatewayd.service
8851 # wget http://localhost:19531/entries
8853 This will download the journal contents in a
8854 /var/log/messages compatible format. The same as JSON:
8856 # curl -H"Accept: application/json" http://localhost:19531/entries
8858 This service is also accessible via a web browser where a
8859 single static HTML5 app is served that uses the JSON logic
8860 to enable the user to do some basic browsing of the
8861 journal. This will be extended later on. Here's an example
8862 screenshot of this app in its current state:
8864 http://0pointer.de/public/journal-gatewayd
8866 Contributions from: Kay Sievers, Lennart Poettering, Robert
8867 Milasan, Tom Gundersen
8871 * The bash completion logic is now available for journalctl
8874 * We do not mount the "cpuset" controller anymore together with
8875 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
8876 started if no parameters are assigned to it. "cpuset" hence
8877 broke code that assumed it could create "cpu" groups and
8880 * journalctl -f will now subscribe to terminal size changes,
8881 and line break accordingly.
8883 Contributions from: Dave Reisner, Kay Sievers, Lennart
8884 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
8888 * nspawn will now create a symlink /etc/localtime in the
8889 container environment, copying the host's timezone
8890 setting. Previously this has been done via a bind mount, but
8891 since symlinks cannot be bind mounted this has now been
8892 changed to create/update the appropriate symlink.
8894 * journalctl -n's line number argument is now optional, and
8895 will default to 10 if omitted.
8897 * journald will now log the maximum size the journal files may
8898 take up on disk. This is particularly useful if the default
8899 built-in logic of determining this parameter from the file
8900 system size is used. Use "systemctl status
8901 systemd-journald.service" to see this information.
8903 * The multi-seat X wrapper tool has been stripped down. As X
8904 is now capable of enumerating graphics devices via udev in a
8905 seat-aware way the wrapper is not strictly necessary
8906 anymore. A stripped down temporary stop-gap is still shipped
8907 until the upstream display managers have been updated to
8908 fully support the new X logic. Expect this wrapper to be
8909 removed entirely in one of the next releases.
8911 * HandleSleepKey= in logind.conf has been split up into
8912 HandleSuspendKey= and HandleHibernateKey=. The old setting
8913 is not available anymore. X11 and the kernel are
8914 distinguishing between these keys and we should too. This
8915 also means the inhibition lock for these keys has been split
8918 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
8919 Poettering, Lukas Nykryn, Václav Pavlín
8923 * Whenever a unit changes state we will now log this to the
8924 journal and show along the unit's own log output in
8927 * ConditionPathIsMountPoint= can now properly detect bind
8928 mount points too. (Previously, a bind mount of one file
8929 system to another place in the same file system could not be
8930 detected as mount, since they shared struct stat's st_dev
8933 * We will now mount the cgroup controllers cpu, cpuacct,
8934 cpuset and the controllers net_cls, net_prio together by
8937 * nspawn containers will now have a virtualized boot
8938 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
8939 over with a randomized ID at container initialization). This
8940 has the effect of making "journalctl -b" do the right thing
8943 * The JSON output journal serialization has been updated not
8944 to generate "endless" list objects anymore, but rather one
8945 JSON object per line. This is more in line how most JSON
8946 parsers expect JSON objects. The new output mode
8947 "json-pretty" has been added to provide similar output, but
8948 neatly aligned for readability by humans.
8950 * We dropped all explicit sync() invocations in the shutdown
8951 code. The kernel does this implicitly anyway in the kernel
8952 reboot() syscall. halt(8)'s -n option is now a compatibility
8955 * We now support virtualized reboot() in containers, as
8956 supported by newer kernels. We will fall back to exit() if
8957 CAP_SYS_REBOOT is not available to the container. Also,
8958 nspawn makes use of this now and will actually reboot the
8959 container if the containerized OS asks for that.
8961 * journalctl will only show local log output by default
8962 now. Use --merge (-m) to show remote log output, too.
8964 * libsystemd-journal gained the new sd_journal_get_usage()
8965 call to determine the current disk usage of all journal
8966 files. This is exposed in the new "journalctl --disk-usage"
8969 * journald gained a new configuration setting SplitMode= in
8970 journald.conf which may be used to control how user journals
8971 are split off. See journald.conf(5) for details.
8973 * A new condition type ConditionFileNotEmpty= has been added.
8975 * tmpfiles' "w" lines now support file globbing, to write
8976 multiple files at once.
8978 * We added Python bindings for the journal submission
8979 APIs. More Python APIs for a number of selected APIs will
8980 likely follow. Note that we intend to add native bindings
8981 only for the Python language, as we consider it common
8982 enough to deserve bindings shipped within systemd. There are
8983 various projects outside of systemd that provide bindings
8984 for languages such as PHP or Lua.
8986 * Many conditions will now resolve specifiers such as %i. In
8987 addition, PathChanged= and related directives of .path units
8988 now support specifiers as well.
8990 * There's now a new RPM macro definition for the system preset
8993 * journald will now warn if it ca not forward a message to the
8994 syslog daemon because its socket is full.
8996 * timedated will no longer write or process /etc/timezone,
8997 except on Debian. As we do not support late mounted /usr
8998 anymore /etc/localtime always being a symlink is now safe,
8999 and hence the information in /etc/timezone is not necessary
9002 * logind will now always reserve one VT for a text getty (VT6
9003 by default). Previously if more than 6 X sessions where
9004 started they took up all the VTs with auto-spawned gettys,
9005 so that no text gettys were available anymore.
9007 * udev will now automatically inform the btrfs kernel logic
9008 about btrfs RAID components showing up. This should make
9009 simple hotplug based btrfs RAID assembly work.
9011 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
9012 (but not for its children which will stay at the kernel
9013 default). This should allow setups with a lot more listening
9016 * systemd will now always pass the configured timezone to the
9017 kernel at boot. timedated will do the same when the timezone
9020 * logind's inhibition logic has been updated. By default,
9021 logind will now handle the lid switch, the power and sleep
9022 keys all the time, even in graphical sessions. If DEs want
9023 to handle these events on their own they should take the new
9024 handle-power-key, handle-sleep-key and handle-lid-switch
9025 inhibitors during their runtime. A simple way to achieve
9026 that is to invoke the DE wrapped in an invocation of:
9028 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
9030 * Access to unit operations is now checked via SELinux taking
9031 the unit file label and client process label into account.
9033 * systemd will now notify the administrator in the journal
9034 when he over-mounts a non-empty directory.
9036 * There are new specifiers that are resolved in unit files,
9037 for the host name (%H), the machine ID (%m) and the boot ID
9040 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
9041 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
9042 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
9043 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
9044 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
9045 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
9046 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
9050 * Support for reading structured kernel messages from
9051 /dev/kmsg has now been added and is enabled by default.
9053 * Support for reading kernel messages from /proc/kmsg has now
9054 been removed. If you want kernel messages in the journal
9055 make sure to run a recent kernel (>= 3.5) that supports
9056 reading structured messages from /dev/kmsg (see
9057 above). /proc/kmsg is now exclusive property of classic
9058 syslog daemons again.
9060 * The libudev API gained the new
9061 udev_device_new_from_device_id() call.
9063 * The logic for file system namespace (ReadOnlyDirectory=,
9064 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
9065 require pivot_root() anymore. This means fewer temporary
9066 directories are created below /tmp for this feature.
9068 * nspawn containers will now see and receive all submounts
9069 made on the host OS below the root file system of the
9072 * Forward Secure Sealing is now supported for Journal files,
9073 which provide cryptographical sealing of journal files so
9074 that attackers cannot alter log history anymore without this
9075 being detectable. Lennart will soon post a blog story about
9076 this explaining it in more detail.
9078 * There are two new service settings RestartPreventExitStatus=
9079 and SuccessExitStatus= which allow configuration of exit
9080 status (exit code or signal) which will be excepted from the
9081 restart logic, resp. consider successful.
9083 * journalctl gained the new --verify switch that can be used
9084 to check the integrity of the structure of journal files and
9085 (if Forward Secure Sealing is enabled) the contents of
9088 * nspawn containers will now be run with /dev/stdin, /dev/fd/
9089 and similar symlinks pre-created. This makes running shells
9090 as container init process a lot more fun.
9092 * The fstab support can now handle PARTUUID= and PARTLABEL=
9095 * A new ConditionHost= condition has been added to match
9096 against the hostname (with globs) and machine ID. This is
9097 useful for clusters where a single OS image is used to
9098 provision a large number of hosts which shall run slightly
9099 different sets of services.
9101 * Services which hit the restart limit will now be placed in a
9104 Contributions from: Bertram Poettering, Dave Reisner, Huang
9105 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
9106 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
9110 * When running in --user mode systemd will now become a
9111 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
9112 tree a lot more organized.
9114 * A new PartOf= unit dependency type has been introduced that
9115 may be used to group services in a natural way.
9117 * "systemctl enable" may now be used to enable instances of
9120 * journalctl now prints error log levels in red, and
9121 warning/notice log levels in bright white. It also supports
9122 filtering by log level now.
9124 * cgtop gained a new -n switch (similar to top), to configure
9125 the maximum number of iterations to run for. It also gained
9126 -b, to run in batch mode (accepting no input).
9128 * The suffix ".service" may now be omitted on most systemctl
9129 command lines involving service unit names.
9131 * There's a new bus call in logind to lock all sessions, as
9132 well as a loginctl verb for it "lock-sessions".
9134 * libsystemd-logind.so gained a new call sd_journal_perror()
9135 that works similar to libc perror() but logs to the journal
9136 and encodes structured information about the error number.
9138 * /etc/crypttab entries now understand the new keyfile-size=
9141 * shutdown(8) now can send a (configurable) wall message when
9142 a shutdown is cancelled.
9144 * The mount propagation mode for the root file system will now
9145 default to "shared", which is useful to make containers work
9146 nicely out-of-the-box so that they receive new mounts from
9147 the host. This can be undone locally by running "mount
9148 --make-rprivate /" if needed.
9150 * The prefdm.service file has been removed. Distributions
9151 should maintain this unit downstream if they intend to keep
9152 it around. However, we recommend writing normal unit files
9153 for display managers instead.
9155 * Since systemd is a crucial part of the OS we will now
9156 default to a number of compiler switches that improve
9157 security (hardening) such as read-only relocations, stack
9158 protection, and suchlike.
9160 * The TimeoutSec= setting for services is now split into
9161 TimeoutStartSec= and TimeoutStopSec= to allow configuration
9162 of individual time outs for the start and the stop phase of
9165 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
9166 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
9167 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
9168 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
9169 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
9170 Gundersen, Zbigniew Jędrzejewski-Szmek
9174 * The journal and id128 C APIs are now fully documented as man
9177 * Extra safety checks have been added when transitioning from
9178 the initial RAM disk to the main system to avoid accidental
9181 * /etc/crypttab entries now understand the new keyfile-offset=
9184 * systemctl -t can now be used to filter by unit load state.
9186 * The journal C API gained the new sd_journal_wait() call to
9187 make writing synchronous journal clients easier.
9189 * journalctl gained the new -D switch to show journals from a
9192 * journalctl now displays a special marker between log
9193 messages of two different boots.
9195 * The journal is now explicitly flushed to /var via a service
9196 systemd-journal-flush.service, rather than implicitly simply
9197 by seeing /var/log/journal to be writable.
9199 * journalctl (and the journal C APIs) can now match for much
9200 more complex expressions, with alternatives and
9203 * When transitioning from the initial RAM disk to the main
9204 system we will now kill all processes in a killing spree to
9205 ensure no processes stay around by accident.
9207 * Three new specifiers may be used in unit files: %u, %h, %s
9208 resolve to the user name, user home directory resp. user
9209 shell. This is useful for running systemd user instances.
9211 * We now automatically rotate journal files if their data
9212 object hash table gets a fill level > 75%. We also size the
9213 hash table based on the configured maximum file size. This
9214 together should lower hash collisions drastically and thus
9215 speed things up a bit.
9217 * journalctl gained the new "--header" switch to introspect
9218 header data of journal files.
9220 * A new setting SystemCallFilters= has been added to services
9221 which may be used to apply blacklists or whitelists to
9222 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
9224 * nspawn gained a new --link-journal= switch (and quicker: -j)
9225 to link the container journal with the host. This makes it
9226 very easy to centralize log viewing on the host for all
9227 guests while still keeping the journal files separated.
9229 * Many bugfixes and optimizations
9231 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
9232 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
9233 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
9238 * Several tools now understand kernel command line arguments,
9239 which are only read when run in an initial RAM disk. They
9240 usually follow closely their normal counterparts, but are
9243 * There's a new tool to analyze the readahead files that are
9244 automatically generated at boot. Use:
9246 /usr/lib/systemd/systemd-readahead analyze /.readahead
9248 * We now provide an early debug shell on tty9 if this enabled. Use:
9250 systemctl enable debug-shell.service
9252 * All plymouth related units have been moved into the Plymouth
9253 package. Please make sure to upgrade your Plymouth version
9256 * systemd-tmpfiles now supports getting passed the basename of
9257 a configuration file only, in which case it will look for it
9258 in all appropriate directories automatically.
9260 * udevadm info now takes a /dev or /sys path as argument, and
9261 does the right thing. Example:
9263 udevadm info /dev/sda
9264 udevadm info /sys/class/block/sda
9266 * systemctl now prints a warning if a unit is stopped but a
9267 unit that might trigger it continues to run. Example: a
9268 service is stopped but the socket that activates it is left
9271 * "systemctl status" will now mention if the log output was
9272 shortened due to rotation since a service has been started.
9274 * The journal API now exposes functions to determine the
9275 "cutoff" times due to rotation.
9277 * journald now understands SIGUSR1 and SIGUSR2 for triggering
9278 immediately flushing of runtime logs to /var if possible,
9279 resp. for triggering immediate rotation of the journal
9282 * It is now considered an error if a service is attempted to
9283 be stopped that is not loaded.
9285 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
9287 * systemd-analyze now supports Python 3
9289 * tmpfiles now supports cleaning up directories via aging
9290 where the first level dirs are always kept around but
9291 directories beneath it automatically aged. This is enabled
9292 by prefixing the age field with '~'.
9294 * Seat objects now expose CanGraphical, CanTTY properties
9295 which is required to deal with very fast bootups where the
9296 display manager might be running before the graphics drivers
9297 completed initialization.
9299 * Seat objects now expose a State property.
9301 * We now include RPM macros for service enabling/disabling
9302 based on the preset logic. We recommend RPM based
9303 distributions to make use of these macros if possible. This
9304 makes it simpler to reuse RPM spec files across
9307 * We now make sure that the collected systemd unit name is
9308 always valid when services log to the journal via
9311 * There's a new man page kernel-command-line(7) detailing all
9312 command line options we understand.
9314 * The fstab generator may now be disabled at boot by passing
9315 fstab=0 on the kernel command line.
9317 * A new kernel command line option modules-load= is now understood
9318 to load a specific kernel module statically, early at boot.
9320 * Unit names specified on the systemctl command line are now
9321 automatically escaped as needed. Also, if file system or
9322 device paths are specified they are automatically turned
9323 into the appropriate mount or device unit names. Example:
9325 systemctl status /home
9326 systemctl status /dev/sda
9328 * The SysVConsole= configuration option has been removed from
9329 system.conf parsing.
9331 * The SysV search path is no longer exported on the D-Bus
9334 * The Names= option has been removed from unit file parsing.
9336 * There's a new man page bootup(7) detailing the boot process.
9338 * Every unit and every generator we ship with systemd now
9339 comes with full documentation. The self-explanatory boot is
9342 * A couple of services gained "systemd-" prefixes in their
9343 name if they wrap systemd code, rather than only external
9344 code. Among them fsck@.service which is now
9345 systemd-fsck@.service.
9347 * The HaveWatchdog property has been removed from the D-Bus
9350 * systemd.confirm_spawn= on the kernel command line should now
9353 * There's a new man page crypttab(5) which details all options
9354 we actually understand.
9356 * systemd-nspawn gained a new --capability= switch to pass
9357 additional capabilities to the container.
9359 * timedated will now read known NTP implementation unit names
9360 from /usr/lib/systemd/ntp-units.d/*.list,
9361 systemd-timedated-ntp.target has been removed.
9363 * journalctl gained a new switch "-b" that lists log data of
9364 the current boot only.
9366 * The notify socket is in the abstract namespace again, in
9367 order to support daemons which chroot() at start-up.
9369 * There is a new Storage= configuration option for journald
9370 which allows configuration of where log data should go. This
9371 also provides a way to disable journal logging entirely, so
9372 that data collected is only forwarded to the console, the
9373 kernel log buffer or another syslog implementation.
9375 * Many bugfixes and optimizations
9377 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
9378 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
9379 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
9380 Shawn Landden, Tom Gundersen
9384 * "systemctl help <unit>" now shows the man page if one is
9387 * Several new man pages have been added.
9389 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
9390 MaxLevelConsole= can now be specified in
9391 journald.conf. These options allow reducing the amount of
9392 data stored on disk or forwarded by the log level.
9394 * TimerSlackNSec= can now be specified in system.conf for
9395 PID1. This allows system-wide power savings.
9397 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
9398 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
9403 * logind is now capable of (optionally) handling power and
9404 sleep keys as well as the lid switch.
9406 * journalctl now understands the syntax "journalctl
9407 /usr/bin/avahi-daemon" to get all log output of a specific
9410 * CapabilityBoundingSet= in system.conf now also influences
9411 the capability bound set of usermode helpers of the kernel.
9413 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
9414 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
9415 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
9416 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
9420 * Note that we skipped 139 releases here in order to set the
9421 new version to something that is greater than both udev's
9422 and systemd's most recent version number.
9424 * udev: all udev sources are merged into the systemd source tree now.
9425 All future udev development will happen in the systemd tree. It
9426 is still fully supported to use the udev daemon and tools without
9427 systemd running, like in initramfs or other init systems. Building
9428 udev though, will require the *build* of the systemd tree, but
9429 udev can be properly *run* without systemd.
9431 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
9432 should be used to create dead device nodes as workarounds for broken
9435 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
9436 no longer supported. udev_monitor_new_from_netlink() needs to be
9437 used to subscribe to events.
9439 * udev: when udevd is started by systemd, processes which are left
9440 behind by forking them off of udev rules, are unconditionally cleaned
9441 up and killed now after the event handling has finished. Services or
9442 daemons must be started as systemd services. Services can be
9443 pulled-in by udev to get started, but they can no longer be directly
9444 forked by udev rules.
9446 * udev: the daemon binary is called systemd-udevd now and installed
9447 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
9448 to adapt to that, create symlink, or rename the binary after building
9451 * libudev no longer provides these symbols:
9452 udev_monitor_from_socket()
9453 udev_queue_get_failed_list_entry()
9454 udev_get_{dev,sys,run}_path()
9455 The versions number was bumped and symbol versioning introduced.
9457 * systemd-loginctl and systemd-journalctl have been renamed
9458 to loginctl and journalctl to match systemctl.
9460 * The config files: /etc/systemd/systemd-logind.conf and
9461 /etc/systemd/systemd-journald.conf have been renamed to
9462 logind.conf and journald.conf. Package updates should rename
9463 the files to the new names on upgrade.
9465 * For almost all files the license is now LGPL2.1+, changed
9466 from the previous GPL2.0+. Exceptions are some minor stuff
9467 of udev (which will be changed to LGPL2.1 eventually, too),
9468 and the MIT licensed sd-daemon.[ch] library that is suitable
9469 to be used as drop-in files.
9471 * systemd and logind now handle system sleep states, in
9472 particular suspending and hibernating.
9474 * logind now implements a sleep/shutdown/idle inhibiting logic
9475 suitable for a variety of uses. Soonishly Lennart will blog
9476 about this in more detail.
9478 * var-run.mount and var-lock.mount are no longer provided
9479 (which previously bind mounted these directories to their new
9480 places). Distributions which have not converted these
9481 directories to symlinks should consider stealing these files
9482 from git history and add them downstream.
9484 * We introduced the Documentation= field for units and added
9485 this to all our shipped units. This is useful to make it
9486 easier to explore the boot and the purpose of the various
9489 * All smaller setup units (such as
9490 systemd-vconsole-setup.service) now detect properly if they
9491 are run in a container and are skipped when
9492 appropriate. This guarantees an entirely noise-free boot in
9493 Linux container environments such as systemd-nspawn.
9495 * A framework for implementing offline system updates is now
9496 integrated, for details see:
9497 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
9499 * A new service type Type=idle is available now which helps us
9500 avoiding ugly interleaving of getty output and boot status
9503 * There's now a system-wide CapabilityBoundingSet= option to
9504 globally reduce the set of capabilities for the
9505 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
9506 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
9507 even CAP_NET_ADMIN system-wide for secure systems.
9509 * There are now system-wide DefaultLimitXXX= options to
9510 globally change the defaults of the various resource limits
9511 for all units started by PID 1.
9513 * Harald Hoyer's systemd test suite has been integrated into
9514 systemd which allows easy testing of systemd builds in qemu
9515 and nspawn. (This is really awesome! Ask us for details!)
9517 * The fstab parser is now implemented as generator, not inside
9520 * systemctl will now warn you if .mount units generated from
9521 /etc/fstab are out of date due to changes in fstab that
9522 have not been read by systemd yet.
9524 * systemd is now suitable for usage in initrds. Dracut has
9525 already been updated to make use of this. With this in place
9526 initrds get a slight bit faster but primarily are much
9527 easier to introspect and debug since "systemctl status" in
9528 the host system can be used to introspect initrd services,
9529 and the journal from the initrd is kept around too.
9531 * systemd-delta has been added, a tool to explore differences
9532 between user/admin configuration and vendor defaults.
9534 * PrivateTmp= now affects both /tmp and /var/tmp.
9536 * Boot time status messages are now much prettier and feature
9537 proper english language. Booting up systemd has never been
9540 * Read-ahead pack files now include the inode number of all
9541 files to pre-cache. When the inode changes the pre-caching
9542 is not attempted. This should be nicer to deal with updated
9543 packages which might result in changes of read-ahead
9546 * We now temporaritly lower the kernel's read_ahead_kb variable
9547 when collecting read-ahead data to ensure the kernel's
9548 built-in read-ahead does not add noise to our measurements
9549 of necessary blocks to pre-cache.
9551 * There's now RequiresMountsFor= to add automatic dependencies
9552 for all mounts necessary for a specific file system path.
9554 * MountAuto= and SwapAuto= have been removed from
9555 system.conf. Mounting file systems at boot has to take place
9558 * nspawn now learned a new switch --uuid= to set the machine
9559 ID on the command line.
9561 * nspawn now learned the -b switch to automatically search
9564 * vt102 is now the default TERM for serial TTYs, upgraded from
9567 * systemd-logind now works on VT-less systems.
9569 * The build tree has been reorganized. The individual
9570 components now have directories of their own.
9572 * A new condition type ConditionPathIsReadWrite= is now available.
9574 * nspawn learned the new -C switch to create cgroups for the
9575 container in other hierarchies.
9577 * We now have support for hardware watchdogs, configurable in
9580 * The scheduled shutdown logic now has a public API.
9582 * We now mount /tmp as tmpfs by default, but this can be
9583 masked and /etc/fstab can override it.
9585 * Since udisks does not make use of /media anymore we are not
9586 mounting a tmpfs on it anymore.
9588 * journalctl gained a new --local switch to only interleave
9589 locally generated journal files.
9591 * We can now load the IMA policy at boot automatically.
9593 * The GTK tools have been split off into a systemd-ui.
9595 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
9596 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
9597 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
9598 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
9599 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
9600 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
9601 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
9602 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
9603 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
9608 * This is mostly a bugfix release
9610 * Support optional initialization of the machine ID from the
9611 KVM or container configured UUID.
9613 * Support immediate reboots with "systemctl reboot -ff"
9615 * Show /etc/os-release data in systemd-analyze output
9617 * Many bugfixes for the journal, including endianness fixes and
9618 ensuring that disk space enforcement works
9620 * sd-login.h is C++ compatible again
9622 * Extend the /etc/os-release format on request of the Debian
9625 * We now refuse non-UTF8 strings used in various configuration
9626 and unit files. This is done to ensure we do not pass invalid
9627 data over D-Bus or expose it elsewhere.
9629 * Register Mimo USB Screens as suitable for automatic seat
9632 * Read SELinux client context from journal clients in a race
9635 * Reorder configuration file lookup order. /etc now always
9636 overrides /run in order to allow the administrator to always
9637 and unconditionally override vendor-supplied or
9638 automatically generated data.
9640 * The various user visible bits of the journal now have man
9641 pages. We still lack man pages for the journal API calls
9644 * We now ship all man pages in HTML format again in the
9647 Contributions from: Dave Reisner, Dirk Eibach, Frederic
9648 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
9649 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
9654 * This is mostly a bugfix release
9656 * systems lacking /etc/os-release are no longer supported.
9658 * Various functionality updates to libsystemd-login.so
9660 * Track class of PAM logins to distinguish greeters from
9663 Contributions from: Kay Sievers, Lennart Poettering, Michael
9668 * This is an important bugfix release for v41.
9670 * Building man pages is now optional which should be useful
9671 for those building systemd from git but unwilling to install
9674 * Watchdog support for supervising services is now usable. In
9675 a future release support for hardware watchdogs
9676 (i.e. /dev/watchdog) will be added building on this.
9678 * Service start rate limiting is now configurable and can be
9679 turned off per service. When a start rate limit is hit a
9680 reboot can automatically be triggered.
9682 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
9684 Contributions from: Benjamin Franzke, Bill Nottingham,
9685 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
9686 Schmidt, Michał Górny, Piotr Drąg
9690 * The systemd binary is installed /usr/lib/systemd/systemd now;
9691 An existing /sbin/init symlink needs to be adapted with the
9694 * The code that loads kernel modules has been ported to invoke
9695 libkmod directly, instead of modprobe. This means we do not
9696 support systems with module-init-tools anymore.
9698 * Watchdog support is now already useful, but still not
9701 * A new kernel command line option systemd.setenv= is
9702 understood to set system wide environment variables
9703 dynamically at boot.
9705 * We now limit the set of capabilities of systemd-journald.
9707 * We now set SIGPIPE to ignore by default, since it only is
9708 useful in shell pipelines, and has little use in general
9709 code. This can be disabled with IgnoreSIPIPE=no in unit
9712 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
9713 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
9718 * This is mostly a bugfix release
9720 * We now expose the reason why a service failed in the
9721 "Result" D-Bus property.
9723 * Rudimentary service watchdog support (will be completed over
9724 the next few releases.)
9726 * When systemd forks off in order execute some service we will
9727 now immediately changes its argv[0] to reflect which process
9728 it will execute. This is useful to minimize the time window
9729 with a generic argv[0], which makes bootcharts more useful
9731 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
9732 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
9733 Mike Kazantsev, Ray Strode
9737 * This is mostly a test release, but incorporates many
9740 * New systemd-cgtop tool to show control groups by their
9743 * Linking against libacl for ACLs is optional again. If
9744 disabled, support tracking device access for active logins
9745 goes becomes unavailable, and so does access to the user
9746 journals by the respective users.
9748 * If a group "adm" exists, journal files are automatically
9749 owned by them, thus allow members of this group full access
9750 to the system journal as well as all user journals.
9752 * The journal now stores the SELinux context of the logging
9753 client for all entries.
9755 * Add C++ inclusion guards to all public headers
9757 * New output mode "cat" in the journal to print only text
9758 messages, without any meta data like date or time.
9760 * Include tiny X server wrapper as a temporary stop-gap to
9761 teach XOrg udev display enumeration. This is used by display
9762 managers such as gdm, and will go away as soon as XOrg
9763 learned native udev hotplugging for display devices.
9765 * Add new systemd-cat tool for executing arbitrary programs
9766 with STDERR/STDOUT connected to the journal. Can also act as
9767 BSD logger replacement, and does so by default.
9769 * Optionally store all locally generated coredumps in the
9770 journal along with meta data.
9772 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
9773 writing short strings to files (for usage for /sys), and for
9774 creating symlinks, character and block device nodes.
9776 * New unit file option ControlGroupPersistent= to make cgroups
9777 persistent, following the mechanisms outlined in
9778 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
9780 * Support multiple local RTCs in a sane way
9782 * No longer monopolize IO when replaying readahead data on
9783 rotating disks, since we might starve non-file-system IO to
9784 death, since fanotify() will not see accesses done by blkid,
9787 * Do not show kernel threads in systemd-cgls anymore, unless
9788 requested with new -k switch.
9790 Contributions from: Dan Horák, Kay Sievers, Lennart
9791 Poettering, Michal Schmidt
9795 * This is mostly a test release, but incorporates many
9798 * The git repository moved to:
9799 git://anongit.freedesktop.org/systemd/systemd
9800 ssh://git.freedesktop.org/git/systemd/systemd
9802 * First release with the journal
9803 http://0pointer.de/blog/projects/the-journal.html
9805 * The journal replaces both systemd-kmsg-syslogd and
9806 systemd-stdout-bridge.
9808 * New sd_pid_get_unit() API call in libsystemd-logind
9810 * Many systemadm clean-ups
9812 * Introduce remote-fs-pre.target which is ordered before all
9813 remote mounts and may be used to start services before all
9816 * Added Mageia support
9818 * Add bash completion for systemd-loginctl
9820 * Actively monitor PID file creation for daemons which exit in
9821 the parent process before having finished writing the PID
9822 file in the daemon process. Daemons which do this need to be
9823 fixed (i.e. PID file creation must have finished before the
9824 parent exits), but we now react a bit more gracefully to them.
9826 * Add colourful boot output, mimicking the well-known output
9827 of existing distributions.
9829 * New option PassCredentials= for socket units, for
9830 compatibility with a recent kernel ABI breakage.
9832 * /etc/rc.local is now hooked in via a generator binary, and
9833 thus will no longer act as synchronization point during
9836 * systemctl list-unit-files now supports --root=.
9838 * systemd-tmpfiles now understands two new commands: z, Z for
9839 relabelling files according to the SELinux database. This is
9840 useful to apply SELinux labels to specific files in /sys,
9843 * Output of SysV services is now forwarded to both the console
9844 and the journal by default, not only just the console.
9846 * New man pages for all APIs from libsystemd-login.
9848 * The build tree got reorganized and the build system is a
9849 lot more modular allowing embedded setups to specifically
9850 select the components of systemd they are interested in.
9852 * Support for Linux systems lacking the kernel VT subsystem is
9855 * configure's --with-rootdir= got renamed to
9856 --with-rootprefix= to follow the naming used by udev and
9859 * Unless specified otherwise we will now install to /usr instead
9860 of /usr/local by default.
9862 * Processes with '@' in argv[0][0] are now excluded from the
9863 final shut-down killing spree, following the logic explained
9865 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
9867 * All processes remaining in a service cgroup when we enter
9868 the START or START_PRE states are now killed with
9869 SIGKILL. That means it is no longer possible to spawn
9870 background processes from ExecStart= lines (which was never
9871 supported anyway, and bad style).
9873 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
9874 reloading of units together.
9876 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
9877 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
9878 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
9879 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
9880 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek