1 systemd System and Service Manager
3 CHANGES WITH 243 in spe:
5 * This release enables unprivileged programs (i.e. requiring neither
6 setuid nor file capabilities) to send ICMP Echo (i.e. ping) requests
7 by turning on the "net.ipv4.ping_group_range" sysctl of the Linux
8 kernel for the whole UNIX group range, i.e. all processes. This
9 change should be reasonably safe, as the kernel support for it was
10 specifically implemented to allow safe access to ICMP Echo for
11 processes lacking any privileges. If this is not desirable, it can be
12 disabled again by setting the parameter to "1 0".
14 * Previously, filters defined with SystemCallFilter= would have the
15 effect that any calling of an offending system call would terminate
16 the calling thread. This behaviour never made much sense, since
17 killing individual threads of unsuspecting processes is likely to
18 create more problems than it solves. With this release the default
19 action changed from killing the thread to killing the whole
20 process. For this to work correctly both a kernel version (>= 4.14)
21 and a libseccomp version (>= 2.4.0) supporting this new seccomp
22 action is required. If an older kernel or libseccomp is used the old
23 behaviour continues to be used. This change does not affect any
24 services that have no system call filters defined, or that use
25 SystemCallErrorNumber= (and thus see EPERM or another error instead
26 of being killed when calling an offending system call). Note that
27 systemd documentation always claimed that the whole process is
28 killed. With this change behaviour is thus adjusted to match the
31 * On 64 bit systems, the "kernel.pid_max" sysctl is now bumped to
32 4194304 by default, i.e. the full 22bit range the kernel allows, up
33 from the old 16bit range. This should improve security and
34 robustness, as PID collisions are made less likely (though certainly
35 still possible). There are rumours this might create compatibility
36 problems, though at this moment no practical ones are known to
37 us. Downstream distributions are hence advised to undo this change in
38 their builds if they are concerned about maximum compatibility, but
39 for everybody else we recommend leaving the value bumped. Besides
40 improving security and robustness this should also simplify things as
41 the maximum number of allowed concurrent tasks was previously bounded
42 by both "kernel.pid_max" and "kernel.threads-max" and now effectively
43 only a single knob is left ("kernel.threads-max"). There have been
44 concerns that usability is affected by this change because larger PID
45 numbers are harder to type, but we believe the change from 5 digits
46 to 7 digits doesn't hamper usability.
48 * MemoryLow= and MemoryMin= gained hierarchy-aware counterparts,
49 DefaultMemoryLow= and DefaultMemoryMin=, which can be used to
50 hierarchically set default memory protection values for a particular
51 subtree of the unit hierarchy.
53 * Memory protection directives can now take a value of zero, allowing
54 explicit opting out of a default value propagated by an ancestor.
56 * systemd now defaults to the "unified" cgroup hierarchy setup during
57 build-time, i.e. -Ddefault-hierarchy=unified is now the build-time
58 default. Previously, -Ddefault-hierarchy=hybrid was the default. This
59 change reflects the fact that cgroupsv2 support has matured
60 substantially in both systemd and in the kernel, and is clearly the
61 way forward. Downstream production distributions might want to
62 continue to use -Ddefault-hierarchy=hybrid (or even =legacy) for
63 their builds as unfortunately the popular container managers have not
64 caught up with the kernel API changes.
66 * Man pages are not built by default anymore (html pages were already
67 disabled by default), to make development builds quicker. When
68 building systemd for a full installation with documentation, meson
69 should be called with -Dman=true and/or -Dhtml=true as appropriate.
70 The default was changed based on the assumption that quick one-off or
71 repeated development builds are much more common than full optimized
72 builds for installation, and people need to pass various other
73 options to when doing "proper" builds anyway, so the gain from making
74 development builds quicker is bigger than the one time disruption for
77 Two scripts are created in the *build* directory to generate and
78 preview man and html pages on demand, e.g.:
80 build/man/man systemctl
81 build/man/html systemd.index
83 * libidn2 is used by default if both libidn2 and libidn are installed.
84 Please use -Dlibidn=true if libidn is preferred.
86 * The D-Bus "wire format" of the CPUAffinity= attribute is changed on
87 big-endian machines. Before, bytes were written and read in native
88 machine order as exposed by the native libc __cpu_mask interface.
89 Now, little-endian order is always used (CPUs 0–7 are described by
90 bits 0–7 in byte 0, CPUs 8–15 are described by byte 1, and so on).
91 This change fixes D-Bus calls that cross endianness boundary.
93 The presentation format used for CPUAffinity= by "systemctl show" and
94 "systemd-analyze dump" is changed to present CPU indices instead of
95 the raw __cpu_mask bitmask. For example, CPUAffinity=0-1 would be
96 shown as CPUAffinity=03000000000000000000000000000… (on
97 little-endian) or CPUAffinity=00000000000000300000000000000… (on
98 64-bit big-endian), and is now shown as CPUAffinity=0-1, matching the
99 input format. The maximum integer that will be printed in the new
100 format is 8191 (four digits), while the old format always used a very
101 long number (with the length varying by architecture), so they can be
102 unambiguously distinguished.
104 * /usr/sbin/halt.local is no longer supported. Implementation in
105 distributions was inconsistent and it seems this functionality was
108 To replace this functionality, users should:
109 - either define a new unit and make it a dependency of final.target
110 (systemctl add-wants final.target my-halt-local.service)
111 - or move the shutdown script to /usr/lib/systemd/system-shutdown/
112 and ensure that it accepts "halt", "poweroff", "reboot", and
113 "kexec" as an argument, see the description in systemd-shutdown(8).
115 * When a [Match] section in .link or .network file is empty (contains
116 no match patterns), a warning will be emitted. Please add any "match
117 all" pattern instead, e.g. OriginalName=* or Name=* in case all
118 interfaces should really be matched.
120 * A new setting NUMAPolicy= may be used to set process memory
121 allocation policy. This setting can be specified in
122 /etc/systemd/system.conf and hence will set the default policy for
123 PID1. The default policy can be overridden on a per-service
124 basis. The related setting NUMAMask= is used to specify NUMA node
125 mask that should be associated with the selected policy.
127 * PID 1 will now listen to Out-Of-Memory (OOM) events the kernel
128 generates when processes it manages are reaching their memory limits,
129 and will place their units in a special state, and optionally kill or
132 * The service manager will now expose bus properties for the IO
133 resources used by units. This information is also shown in "systemctl
134 status" now (for services that have IOAccounting=yes set). Moreover,
135 the IO accounting data is included in the resource log message
136 generated whenever a unit stops.
138 * Units may now configure an explicit time-out to wait for when killed
139 with SIGABRT, for example when a service watchdog is hit. Previously,
140 the regular TimeoutStopSec= time-out was applied in this case too —
141 now a separate time-out may be set using TimeoutAbortSec=.
143 * Services may now send a special WATCHDOG=trigger message with
144 sd_notify() to trigger an immediate "watchdog missed" event, and thus
145 trigger service termination. This is useful both for testing watchdog
146 handling, but also for defining error paths in services, that shall
147 be handled the same way as watchdog events.
149 * There are two new per-unit settings IPIngressFilterPath= and
150 IPEgressFilterPath= which allow configuration of a BPF program
151 (usually by specifying a path to a program uploaded to /sys/fs/bpf/)
152 to apply to the IP packet ingress/egress path of all processes of a
153 unit. This is useful to allow running systemd services with BPF
154 programs set up externally.
156 * systemctl gained a new "clean" verb for removing the state, cache,
157 runtime or logs directories of a service while it is terminated. The
158 new verb may also be used to remove the state maintained on disk for
159 timer units that have Persistent= configured.
161 * During the last phase of shutdown systemd will now automatically
162 increase the log level configured in the "kernel.printk" sysctl so
163 that any relevant loggable events happening during late shutdown are
164 made visible. Previously, loggable events happening so late during
165 shutdown were generally lost if the "kernel.printk" sysctl was set to
166 high thresholds, as regular logging daemons are terminated at that
167 time and thus nothing is written to disk.
169 * If processes terminated during the last phase of shutdown do not exit
170 quickly systemd will now show their names after a short time, to make
171 debugging easier. After a longer time-out they are forcibly killed,
174 * journalctl (and the other tools that display logs) will now highlight
175 warnings in yellow (previously, both LOG_NOTICE and LOG_WARNING where
176 shown in bright bold, now only LOG_NOTICE is). Moreover, audit logs
177 are now shown in blue color, to separate them visually from regular
178 logs. References to configuration files are now turned into clickable
179 links on terminals that support that.
181 * systemd-journald will now stop logging to /var/log/journal during
182 shutdown when /var/ is on a separate mount, so that it can be
183 unmounted safely during shutdown.
185 * systemd-resolved gained support for a new 'strict' DNS-over-TLS mode.
187 * systemd-resolved "Cache=" configuration option in resolved.conf has
188 been extended to also accept the 'no-negative' value. Previously,
189 only a boolean option was allowed (yes/no), having yes as the
190 default. If this option is set to 'no-negative', negative answers are
191 not cached while the old cache heuristics are used positive answers.
192 The default remains unchanged.
194 * The predictable naming scheme for network devices now supports
195 generating predictable names for "netdevsim" devices.
197 Moreover, the "en" prefix was dropped from the ID_NET_NAME_ONBOARD
200 Those two changes form a new net.naming-policy-scheme= entry.
201 Distributions which want to preserve naming stability may want to set
202 the -Ddefault-net-naming-scheme= configuration option.
204 * systemd-networkd now supports MACsec, nlmon, IPVTAP and Xfrm
207 * systemd-networkd's bridge FDB support now allows configuration of a
208 destination address for each entry (Destination=), as well as the
209 VXLAN VNI (VNI=), as well as an option to declare what an entry is
210 associated with (AssociatedWith=).
212 * systemd-networkd's DHCPv4 support now understands a new MaxAttempts=
213 option for configuring the maximum number of DHCP lease requests. It
214 also learnt a new BlackList= option for blacklisting DHCP servers (a
215 similar setting has also been added to the IPv6 RA client), as well
216 as a SendRelease= option for configuring whether to send a DHCP
217 RELEASE message when terminating.
219 * systemd-networkd's DHCPv4 and DHCPv6 stacks can now be configured
220 separately in the [DHCPv4] and [DHCPv6] sections.
222 * systemd-networkd's DHCP support will now optionally create an
223 implicit host route to the DNS server specified in the DHCP lease, in
224 addition to the routes listed explicitly in the lease. This should
225 ensure that in multi-homed systems DNS traffic leaves the systems on
226 the interface that acquired the DNS server information even if other
227 routes such as default routes exist. This behaviour may be turned on
228 with the new RoutesToDNS= option.
230 * systemd-networkd's VXLAN support gained a new option
231 GenericProtocolExtension= for enabling VXLAN Generic Protocol
232 Extension support, as well as IPDoNotFragment= for setting the IP
233 "Don't fragment" bit on outgoing packets. A similar option has been
234 added to the GENEVE support.
236 * In systemd-networkd's [Route] section you may now configure
237 FastOpenNoCookie= for configuring per-route TCP fast-open support, as
238 well as TTLPropagate= for configuring Label Switched Path (LSP) TTL
239 propagation. The Type= setting now supports local, broadcast,
240 anycast, multicast, any, xresolve routes, too.
242 * systemd-networkd's [Network] section learnt a new option
243 DefaultRouteOnDevice= for automatically configuring a default route
244 onto the network device.
246 * systemd-networkd's bridging support gained two new options ProxyARP=
247 and ProxyARPWifi= for configuring proxy ARP behaviour as well as
248 MulticastRouter= for configuring multicast routing behaviour. A new
249 option MulticastIGMPVersion= may be used to change bridge's multicast
250 Internet Group Management Protocol (IGMP) version.
252 * systemd-networkd's FooOverUDP support gained the ability to configure
253 local and peer IP addresses via Local= and Peer=. A new option
254 PeerPort= may be used to configure the peer's IP port.
256 * systemd-networkd's TUN support gained a new setting VnetHeader= for
257 tweaking Generic Segment Offload support.
259 * networkctl gained a new "delete" command for removing virtual network
260 devices, as well as a new "--stats" switch for showing device
263 * networkd.conf gained a new setting SpeedMeter= and
264 SpeedMeterIntervalSec=, to measure bitrate of network interfaces. The
265 measured speed may be shown by 'networkctl status'.
267 * "networkctl status" now displays MTU and queue lengths, and more
268 detailed information about VXLAN and bridge devices.
270 * systemd-networkd's .network and .link files gained a new Property=
271 setting in the [Match] section, to match against devices with
272 specific udev properties.
274 * systemd-networkd's tunnel support gained a new option
275 AssignToLoopback= for selecting whether to use the loopback device
276 "lo" as underlying device.
278 * systemd-networkd's MACAddress= setting in the [Neighbor] section has
279 been renamed to LinkLayerAddress=, and it now allows configuration of
282 * A new tool systemd-network-generator has been added that may generate
283 .network, .netdev and .link files from IP configuration specified on
284 the kernel command line in the format used by Dracut.
286 * The CriticalConnection= setting in .network files is now deprecated,
287 and replaced by a new KeepConfiguration= setting which allows more
288 detailed configuration of the IP configuration to keep in place.
290 * systemd-analyze gained a few new verbs:
292 - "systemd-analyze timestamp" parses and converts timestamps. This is
293 similar to the existing "systemd-analyze calendar" command which
294 does the same for recurring calendar events.
296 - "systemd-analyze timespan" parses and converts timespans (i.e.
297 durations as opposed to points in time).
299 - "systemd-analyze condition" will parse and test ConditionXYZ=
302 - "systemd-analyze exit-status" will parse and convert exit status
303 codes to their names and back.
305 - "systemd-analyze unit-files" will print a list of all unit
306 file paths and unit aliases.
308 * SuccessExitStatus=, RestartPreventExitStatus=, and
309 RestartForceExitStatus= now accept exit status names (e.g. "DATAERR"
310 is equivalent to "65"). Those exit status name mappings may be
311 displayed with the sytemd-analyze exit-status verb describe above.
313 * systemd-logind now exposes a per-session SetBrightness() bus call,
314 which may be used to securely change the brightness of a kernel
315 brightness device, if it belongs to the session's seat. By using this
316 call unprivileged clients can make changes to "backlight" and "leds"
317 devices securely with strict requirements on session membership.
318 Desktop environments may use this to generically make brightness
319 changes to such devices without shipping private SUID binaries or
320 udev rules for that purpose.
322 * "udevadm info" gained a --wait-for-initialization switch to wait for
323 a device to be initialized.
325 * systemd-hibernate-resume-generator will now look for resumeflags= on
326 the kernel command line, which is similar to rootflags= and may be
327 used to configure device timeout for the hibernation device.
329 * sd-event learnt a new API call sd_event_source_disable_unref() for
330 disabling and unref'ing an event source in a single function. A
331 related call sd_event_source_disable_unrefp() has been added for use
332 with gcc's cleanup extension.
334 * The sd-id128.h public API gained a new definition
335 SD_ID128_UUID_FORMAT_STR for formatting a 128bit ID in UUID format
338 * "busctl introspect" gained a new switch --xml-interface for dumping
339 XML introspection data unmodified.
341 * PID 1 may now show the unit name instead of the unit description
342 string in its status output during boot. This may be configured in
343 the StatusUnitFormat= setting in /etc/systemd/system.conf or the
344 kernel command line option systemd.status_unit_format=.
346 * PID 1 now understands a new option KExecWatchdogSec= in
347 /etc/systemd/system.conf to set a watchdog timeout for kexec reboots.
348 Previously watchdog functionality was only available for regular
349 reboots. The new setting defaults to off, because we don't know in
350 the general case if the watchdog will be reset after kexec (some
351 drivers do reset it, but not all), and the new userspace might not be
352 configured to handle the watchdog.
354 Moreover, the old ShutdownWatchdogSec= setting has been renamed to
355 RebootWatchdogSec= to more clearly communicate what it is about. The
356 old name is still accepted for compatibility.
358 * The systemd.debug_shell kernel command line option now optionally
359 takes a tty name to spawn the debug shell on, which allows a
360 different tty to be selected than the built-in default.
362 * Service units gained a new ExecCondition= setting which will run
363 before ExecStartPre= and either continue execution of the unit (for
364 clean exit codes), stop execution without marking the unit failed
365 (for exit codes 1 through 254), or stop execution and fail the unit
366 (for exit code 255 or abnormal termination).
368 * A new service systemd-pstore.service has been added that pulls data
369 from /sys/fs/pstore/ and saves it to /var/lib/pstore for later
372 * timedatectl gained new verbs for configuring per-interface NTP
373 service configuration for systemd-timesyncd.
375 * "localectl list-locales" won't list non-UTF-8 locales anymore. It's
376 2019. (You can set non-UTF-8 locales though, if you know their name.)
378 * If variable assignments in sysctl.d/ files are prefixed with "-" any
379 failures to apply them are now ignored.
381 * systemd-random-seed.service now optionally credits entropy when
382 applying the seed to the system. Set $SYSTEMD_RANDOM_SEED_CREDIT to
383 true for the service to enable this behaviour, but please consult the
384 documentation first, since this comes with a couple of caveats.
386 * systemd-random-seed.service is now a synchronization point for full
387 initialization of the kernel's entropy pool. Services that require
388 /dev/urandom to be correctly initialized should be ordered after this
391 * The systemd-boot boot loader has been updated to optionally maintain
392 a random seed file in the EFI System Partition (ESP). During the boot
393 phase, this random seed is read and updated with a new seed
394 cryptographically derived from it. Another derived seed is passed to
395 the OS. The latter seed is then credited to the kernel's entropy pool
396 very early during userspace initialization (from PID 1). This allows
397 systems to boot up with a fully initialized kernel entropy pool from
398 earliest boot on, and thus entirely removes all entropy pool
399 initialization delays from systems using systemd-boot. Special care
400 is taken to ensure different seeds are derived on system images
401 replicated to multiple systems. "bootctl status" will show whether
402 a seed was received from the boot loader.
404 * bootctl gained two new verbs:
406 - "bootctl random-seed" will generate the file in ESP and an EFI
407 variable to allow a random seed to be passed to the OS as described
410 - "bootctl is-installed" checks whether systemd-boot is currently
413 * bootctl will warn if it detects that boot entries are misconfigured
414 (for example if the kernel image was removed without purging the
417 * A new document has been added describing systemd's use and support
418 for the kernel's entropy pool subsystem:
420 https://systemd.io/RANDOM_SEEDS
422 * When the system is hibernated the swap device to write the
423 hibernation image to is now automatically picked from all available
424 swap devices, preferring the swap device with the highest configured
425 priority over all others, and picking the device with the most free
426 space if there are multiple devices with the highest priority.
428 * /etc/crypttab support has learnt a new keyfile-timeout= per-device
429 option that permits selecting the timout how long to wait for a
430 device with an encryption key before asking for the password.
432 * IOWeight= has learnt to properly set the IO weight when using the
433 BFQ scheduler officially found in kernels 5.0+.
435 * A new mailing list has been created for reporting of security issues:
436 systemd-security@redhat.com. For mode details, see
437 https://systemd.io/CONTRIBUTING#security-vulnerability-reports.
439 Contributions from: Aaron Barany, Adrian Bunk, Alan Jenkins, Albrecht
440 Lohofener, Andrej Valek, Anita Zhang, Arian van Putten, Balint Reczey,
441 Bastien Nocera, Ben Boeckel, Benjamin Robin, camoz, Chen Qi, Chris
442 Chiu, Chris Down, Christian Kellner, Clinton Roy, Connor Reeder, Daniel
443 Black, Daniele Medri, Dan Streetman, Dave Reisner, Dave Ross, David
444 Art, David Tardon, Debarshi Ray, Dominick Grift, Donald Buczek, Douglas
445 Christman, Eric DeVolder, EtherGraf, Evgeny Vereshchagin, Feldwor,
446 Felix Riemann, Florian Dollinger, Francesco Pennica, Franck Bui,
447 Frantisek Sumsal, Franz Pletz, frederik, Hans de Goede, Iago López
448 Galeiras, Insun Pyo, Ivan Shapovalov, Iwan Timmer, Jack, Jakob
449 Unterwurzacher, Jan Klötzke, Jan Pokorný, Jan Synacek, Jeka Pats,
450 Jérémy Rosen, Jiri Pirko, Joe Lin, Joerg Behrmann, Joe Richey, Jóhann
451 B. Guðmundsson, Johannes Christ, Johannes Schmitz, Jonathan Rouleau,
452 Jorge Niedbalski, Kai Krakow, Kai Lüke, Karel Zak, Kashyap Chamarthy,
453 Krayushkin Konstantin, Lennart Poettering, Lubomir Rintel, Luca
454 Boccassi, Luís Ferreira, Marc-André Lureau, Markus Felten, Martin Pitt,
455 Matthew Leeds, Mattias Jernberg, Michael Biebl, Michael Olbrich,
456 Michael Prokop, Michael Stapelberg, Michael Zhivich, Michal Koutný,
457 Michal Sekletar, Mike Gilbert, Milan Broz, Miroslav Lichvar, mpe85,
458 Mr-Foo, Network Silence, Oliver Harley, pan93412, Paul Menzel, pEJipE,
459 Peter A. Bigot, Philip Withnall, Piotr Drąg, Rafael Fontenelle, Roberto
460 Santalla, Ronan Pigott, root, RussianNeuroMancer, Sebastian Jennen,
461 shinygold, Shreyas Behera, Simon Schricker, Susant Sahani, Thadeu Lima
462 de Souza Cascardo, Theo Ouzhinski, Thiebaud Weksteen, Thomas Haller,
463 Thomas Weißschuh, Tomas Mraz, Tommi Rantala, Topi Miettinen, ven,
464 Wieland Hoffmann, William A. Kennington III, William Wold, Xi Ruoyao,
465 Yuri Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek,
468 – Somewhere, SOME-TI-ME
472 * In .link files, MACAddressPolicy=persistent (the default) is changed
473 to cover more devices. For devices like bridges, tun, tap, bond, and
474 similar interfaces that do not have other identifying information,
475 the interface name is used as the basis for persistent seed for MAC
476 and IPv4LL addresses. The way that devices that were handled
477 previously is not changed, and this change is about covering more
478 devices then previously by the "persistent" policy.
480 MACAddressPolicy=random may be used to force randomized MACs and
481 IPv4LL addresses for a device if desired.
483 Hint: the log output from udev (at debug level) was enhanced to
484 clarify what policy is followed and which attributes are used.
485 `SYSTEMD_LOG_LEVEL=debug udevadm test-builtin net_setup_link /sys/class/net/<name>`
486 may be used to view this.
488 Hint: if a bridge interface is created without any slaves, and gains
489 a slave later, then now the bridge does not inherit slave's MAC.
490 To inherit slave's MAC, for example, create the following file:
492 # /etc/systemd/network/98-bridge-inherit-mac.link
497 MACAddressPolicy=none
500 * The .device units generated by systemd-fstab-generator and other
501 generators do not automatically pull in the corresponding .mount unit
502 as a Wants= dependency. This means that simply plugging in the device
503 will not cause the mount unit to be started automatically. But please
504 note that the mount unit may be started for other reasons, in
505 particular if it is part of local-fs.target, and any unit which
506 (transitively) depends on local-fs.target is started.
508 * networkctl list/status/lldp now accept globbing wildcards for network
509 interface names to match against all existing interfaces.
511 * The $PIDFILE environment variable is set to point the absolute path
512 configured with PIDFile= for processes of that service.
514 * The fallback DNS server list was augmented with Cloudflare public DNS
515 servers. Use `-Ddns-servers=` to set a different fallback.
517 * A new special target usb-gadget.target will be started automatically
518 when a USB Device Controller is detected (which means that the system
519 is a USB peripheral).
521 * A new unit setting CPUQuotaPeriodSec= assigns the time period
522 relatively to which the CPU time quota specified by CPUQuota= is
525 * A new unit setting ProtectHostname= may be used to prevent services
526 from modifying hostname information (even if they otherwise would
527 have privileges to do so).
529 * A new unit setting NetworkNamespacePath= may be used to specify a
530 namespace for service or socket units through a path referring to a
531 Linux network namespace pseudo-file.
533 * The PrivateNetwork= setting and JoinsNamespaceOf= dependencies now
534 have an effect on .socket units: when used the listening socket is
535 created within the configured network namespace instead of the host
538 * ExecStart= command lines in unit files may now be prefixed with ':'
539 in which case environment variable substitution is
540 disabled. (Supported for the other ExecXYZ= settings, too.)
542 * .timer units gained two new boolean settings OnClockChange= and
543 OnTimezoneChange= which may be used to also trigger a unit when the
544 system clock is changed or the local timezone is
545 modified. systemd-run has been updated to make these options easily
546 accessible from the command line for transient timers.
548 * Two new conditions for units have been added: ConditionMemory= may be
549 used to conditionalize a unit based on installed system
550 RAM. ConditionCPUs= may be used to conditionalize a unit based on
553 * The @default system call filter group understood by SystemCallFilter=
554 has been updated to include the new rseq() system call introduced in
557 * A new time-set.target has been added that indicates that the system
558 time has been set from a local source (possibly imprecise). The
559 existing time-sync.target is stronger and indicates that the time has
560 been synchronized with a precise external source. Services where
561 approximate time is sufficient should use the new target.
563 * "systemctl start" (and related commands) learnt a new
564 --show-transaction option. If specified brief information about all
565 jobs queued because of the requested operation is shown.
567 * systemd-networkd recognizes a new operation state 'enslaved', used
568 (instead of 'degraded' or 'carrier') for interfaces which form a
569 bridge, bond, or similar, and an new 'degraded-carrier' operational
570 state used for the bond or bridge master interface when one of the
571 enslaved devices is not operational.
573 * .network files learnt the new IgnoreCarrierLoss= option for leaving
574 networks configured even if the carrier is lost.
576 * The RequiredForOnline= setting in .network files may now specify a
577 minimum operational state required for the interface to be considered
578 "online" by systemd-networkd-wait-online. Related to this
579 systemd-networkd-wait-online gained a new option --operational-state=
580 to configure the same, and its --interface= option was updated to
581 optionally also take an operational state specific for an interface.
583 * systemd-networkd-wait-online gained a new setting --any for waiting
584 for only one of the requested interfaces instead of all of them.
586 * systemd-networkd now implements L2TP tunnels.
588 * Two new .network settings UseAutonomousPrefix= and UseOnLinkPrefix=
589 may be used to cause autonomous and onlink prefixes received in IPv6
590 Router Advertisements to be ignored.
592 * New MulticastFlood=, NeighborSuppression=, and Learning= .network
593 file settings may be used to tweak bridge behaviour.
595 * The new TripleSampling= option in .network files may be used to
596 configure CAN triple sampling.
598 * A new .netdev settings PrivateKeyFile= and PresharedKeyFile= may be
599 used to point to private or preshared key for a WireGuard interface.
601 * /etc/crypttab now supports the same-cpu-crypt and
602 submit-from-crypt-cpus options to tweak encryption work scheduling
605 * systemd-tmpfiles will now take a BSD file lock before operating on a
606 contents of directory. This may be used to temporarily exclude
607 directories from aging by taking the same lock (useful for example
608 when extracting a tarball into /tmp or /var/tmp as a privileged user,
609 which might create files with really old timestamps, which
610 nevertheless should not be deleted). For further details, see:
612 https://systemd.io/TEMPORARY_DIRECTORIES
614 * systemd-tmpfiles' h line type gained support for the
615 FS_PROJINHERIT_FL ('P') file attribute (introduced in kernel 4.5),
616 controlling project quota inheritance.
618 * sd-boot and bootctl now implement support for an Extended Boot Loader
619 (XBOOTLDR) partition, that is intended to be mounted to /boot, in
620 addition to the ESP partition mounted to /efi or /boot/efi.
621 Configuration file fragments, kernels, initrds and other EFI images
622 to boot will be loaded from both the ESP and XBOOTLDR partitions.
623 The XBOOTLDR partition was previously described by the Boot Loader
624 Specification, but implementation was missing in sd-boot. Support for
625 this concept allows using the sd-boot boot loader in more
626 conservative scenarios where the boot loader itself is placed in the
627 ESP but the kernels to boot (and their metadata) in a separate
630 * A system may now be booted with systemd.volatile=overlay on the
631 kernel command line, which causes the root file system to be set up
632 an overlayfs mount combining the root-only root directory with a
633 writable tmpfs. In this setup, the underlying root device is not
634 modified, and any changes are lost at reboot.
636 * Similar, systemd-nspawn can now boot containers with a volatile
637 overlayfs root with the new --volatile=overlay switch.
639 * systemd-nspawn can now consume OCI runtime bundles using a new
640 --oci-bundle= option. This implementation is fully usable, with most
641 features in the specification implemented, but since this a lot of
642 new code and functionality, this feature should most likely not
643 be used in production yet.
645 * systemd-nspawn now supports various options described by the OCI
646 runtime specification on the command-line and in .nspawn files:
647 --inaccessible=/Inaccessible= may be used to mask parts of the file
648 system tree, --console=/--pipe may be used to configure how standard
649 input, output, and error are set up.
651 * busctl learned the `emit` verb to generate D-Bus signals.
653 * systemd-analyze cat-config may be used to gather and display
654 configuration spread over multiple files, for example system and user
655 presets, tmpfiles.d, sysusers.d, udev rules, etc.
657 * systemd-analyze calendar now takes an optional new parameter
658 --iterations= which may be used to show a maximum number of iterations
659 the specified expression will elapse next.
661 * The sd-bus C API gained support for naming method parameters in the
664 * systemd-logind gained D-Bus APIs to specify the "reboot parameter"
665 the reboot() system call expects.
667 * journalctl learnt a new --cursor-file= option that points to a file
668 from which a cursor should be loaded in the beginning and to which
669 the updated cursor should be stored at the end.
671 * ACRN hypervisor and Windows Subsystem for Linux (WSL) are now
672 detected by systemd-detect-virt (and may also be used in
673 ConditionVirtualization=).
675 * The behaviour of systemd-logind may now be modified with environment
676 variables $SYSTEMD_REBOOT_TO_FIRMWARE_SETUP,
677 $SYSTEMD_REBOOT_TO_BOOT_LOADER_MENU, and
678 $SYSTEMD_REBOOT_TO_BOOT_LOADER_ENTRY. They cause logind to either
679 skip the relevant operation completely (when set to false), or to
680 create a flag file in /run/systemd (when set to true), instead of
681 actually commencing the real operation when requested. The presence
682 of /run/systemd/reboot-to-firmware-setup,
683 /run/systemd/reboot-to-boot-loader-menu, and
684 /run/systemd/reboot-to-boot-loader-entry, may be used by alternative
685 boot loader implementations to replace some steps logind performs
686 during reboot with their own operations.
688 * systemctl can be used to request a reboot into the boot loader menu
689 or a specific boot loader entry with the new --boot-load-menu= and
690 --boot-loader-entry= options to a reboot command. (This requires a
691 boot loader that supports this, for example sd-boot.)
693 * kernel-install will no longer unconditionally create the output
694 directory (e.g. /efi/<machine-id>/<kernel-version>) for boot loader
695 snippets, but will do only if the machine-specific parent directory
696 (i.e. /efi/<machine-id>/) already exists. bootctl has been modified
697 to create this parent directory during sd-boot installation.
699 This makes it easier to use kernel-install with plugins which support
700 a different layout of the bootloader partitions (for example grub2).
702 * During package installation (with `ninja install`), we would create
703 symlinks for getty@tty1.service, systemd-networkd.service,
704 systemd-networkd.socket, systemd-resolved.service,
705 remote-cryptsetup.target, remote-fs.target,
706 systemd-networkd-wait-online.service, and systemd-timesyncd.service
707 in /etc, as if `systemctl enable` was called for those units, to make
708 the system usable immediately after installation. Now this is not
709 done anymore, and instead calling `systemctl preset-all` is
710 recommended after the first installation of systemd.
712 * A new boolean sandboxing option RestrictSUIDSGID= has been added that
713 is built on seccomp. When turned on creation of SUID/SGID files is
716 * The NoNewPrivileges= and the new RestrictSUIDSGID= options are now
717 implied if DynamicUser= is turned on for a service. This hardens
718 these services, so that they neither can benefit from nor create
719 SUID/SGID executables. This is a minor compatibility breakage, given
720 that when DynamicUser= was first introduced SUID/SGID behaviour was
721 unaffected. However, the security benefit of these two options is
722 substantial, and the setting is still relatively new, hence we opted
723 to make it mandatory for services with dynamic users.
725 Contributions from: Adam Jackson, Alexander Tsoy, Andrey Yashkin,
726 Andrzej Pietrasiewicz, Anita Zhang, Balint Reczey, Beniamino Galvani,
727 Ben Iofel, Benjamin Berg, Benjamin Dahlhoff, Chris, Chris Morin,
728 Christopher Wong, Claudius Ellsel, Clemens Gruber, dana, Daniel Black,
729 Davide Cavalca, David Michael, David Rheinsberg, emersion, Evgeny
730 Vereshchagin, Filipe Brandenburger, Franck Bui, Frantisek Sumsal,
731 Giacinto Cifelli, Hans de Goede, Hugo Kindel, Ignat Korchagin, Insun
732 Pyo, Jan Engelhardt, Jonas Dorel, Jonathan Lebon, Jonathon Kowalski,
733 Jörg Sommer, Jörg Thalheim, Jussi Pakkanen, Kai-Heng Feng, Lennart
734 Poettering, Lubomir Rintel, Luís Ferreira, Martin Pitt, Matthias
735 Klumpp, Michael Biebl, Michael Niewöhner, Michael Olbrich, Michal
736 Sekletar, Mike Lothian, Paul Menzel, Piotr Drąg, Riccardo Schirone,
737 Robin Elvedi, Roman Kulikov, Ronald Tschalär, Ross Burton, Ryan
738 Gonzalez, Sebastian Krzyszkowiak, Stephane Chazelas, StKob, Susant
739 Sahani, Sylvain Plantefève, Szabolcs Fruhwald, Taro Yamada, Theo
740 Ouzhinski, Thomas Haller, Tobias Jungel, Tom Yan, Tony Asleson, Topi
741 Miettinen, unixsysadmin, Van Laser, Vesa Jääskeläinen, Yu, Li-Yu,
742 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
748 * The default locale can now be configured at compile time. Otherwise,
749 a suitable default will be selected automatically (one of C.UTF-8,
752 * The version string shown by systemd and other tools now includes the
753 git commit hash when built from git. An override may be specified
754 during compilation, which is intended to be used by distributions to
755 include the package release information.
757 * systemd-cat can now filter standard input and standard error streams
758 for different syslog priorities using the new --stderr-priority=
761 * systemd-journald and systemd-journal-remote reject entries which
762 contain too many fields (CVE-2018-16865) and set limits on the
763 process' command line length (CVE-2018-16864).
765 * $DBUS_SESSION_BUS_ADDRESS environment variable is set by pam_systemd
768 * A new network device NamePolicy "keep" is implemented for link files,
769 and used by default in 99-default.link (the fallback configuration
770 provided by systemd). With this policy, if the network device name
771 was already set by userspace, the device will not be renamed again.
772 This matches the naming scheme that was implemented before
773 systemd-240. If naming-scheme < 240 is specified, the "keep" policy
774 is also enabled by default, even if not specified. Effectively, this
775 means that if naming-scheme >= 240 is specified, network devices will
776 be renamed according to the configuration, even if they have been
777 renamed already, if "keep" is not specified as the naming policy in
778 the .link file. The 99-default.link file provided by systemd includes
779 "keep" for backwards compatibility, but it is recommended for user
780 installed .link files to *not* include it.
782 The "kernel" policy, which keeps kernel names declared to be
783 "persistent", now works again as documented.
785 * kernel-install script now optionally takes the paths to one or more
786 initrd files, and passes them to all plugins.
788 * The mincore() system call has been dropped from the @system-service
789 system call filter group, as it is pretty exotic and may potentially
790 used for side-channel attacks.
792 * -fPIE is dropped from compiler and linker options. Please specify
793 -Db_pie=true option to meson to build position-independent
794 executables. Note that the meson option is supported since meson-0.49.
796 * The fs.protected_regular and fs.protected_fifos sysctls, which were
797 added in Linux 4.19 to make some data spoofing attacks harder, are
798 now enabled by default. While this will hopefully improve the
799 security of most installations, it is technically a backwards
800 incompatible change; to disable these sysctls again, place the
801 following lines in /etc/sysctl.d/60-protected.conf or a similar file:
803 fs.protected_regular = 0
804 fs.protected_fifos = 0
806 Note that the similar hardlink and symlink protection has been
807 enabled since v199, and may be disabled likewise.
809 * The files read from the EnvironmentFile= setting in unit files now
810 parse backslashes inside quotes literally, matching the behaviour of
813 * udevadm trigger, udevadm control, udevadm settle and udevadm monitor
814 now automatically become NOPs when run in a chroot() environment.
816 * The tmpfiles.d/ "C" line type will now copy directory trees not only
817 when the destination is so far missing, but also if it already exists
818 as a directory and is empty. This is useful to cater for systems
819 where directory trees are put together from multiple separate mount
820 points but otherwise empty.
822 * A new function sd_bus_close_unref() (and the associated
823 sd_bus_close_unrefp()) has been added to libsystemd, that combines
824 sd_bus_close() and sd_bus_unref() in one.
826 * udevadm control learnt a new option for --ping for testing whether a
827 systemd-udevd instance is running and reacting.
829 * udevadm trigger learnt a new option for --wait-daemon for waiting
830 systemd-udevd daemon to be initialized.
832 Contributions from: Aaron Plattner, Alberts Muktupāvels, Alex Mayer,
833 Ayman Bagabas, Beniamino Galvani, Burt P, Chris Down, Chris Lamb, Chris
834 Morin, Christian Hesse, Claudius Ellsel, dana, Daniel Axtens, Daniele
835 Medri, Dave Reisner, David Santamaría Rogado, Diego Canuhe, Dimitri
836 John Ledkov, Evgeny Vereshchagin, Fabrice Fontaine, Filipe
837 Brandenburger, Franck Bui, Frantisek Sumsal, govwin, Hans de Goede,
838 James Hilliard, Jan Engelhardt, Jani Uusitalo, Jan Janssen, Jan
839 Synacek, Jonathan McDowell, Jonathan Roemer, Jonathon Kowalski, Joost
840 Heitbrink, Jörg Thalheim, Lance, Lennart Poettering, Louis Taylor,
841 Lucas Werkmeister, Mantas Mikulėnas, Marc-Antoine Perennou,
842 marvelousblack, Michael Biebl, Michael Sloan, Michal Sekletar, Mike
843 Auty, Mike Gilbert, Mikhail Kasimov, Neil Brown, Niklas Hambüchen,
844 Patrick Williams, Paul Seyfert, Peter Hutterer, Philip Withnall, Roger
845 James, Ronnie P. Thomas, Ryan Gonzalez, Sam Morris, Stephan Edel,
846 Stephan Gerhold, Susant Sahani, Taro Yamada, Thomas Haller, Topi
847 Miettinen, YiFei Zhu, YmrDtnJu, YunQiang Su, Yu Watanabe, Zbigniew
848 Jędrzejewski-Szmek, zsergeant77, Дамјан Георгиевски
854 * NoNewPrivileges=yes has been set for all long-running services
855 implemented by systemd. Previously, this was problematic due to
856 SELinux (as this would also prohibit the transition from PID1's label
857 to the service's label). This restriction has since been lifted, but
858 an SELinux policy update is required.
859 (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.)
861 * DynamicUser=yes is dropped from systemd-networkd.service,
862 systemd-resolved.service and systemd-timesyncd.service, which was
863 enabled in v239 for systemd-networkd.service and systemd-resolved.service,
864 and since v236 for systemd-timesyncd.service. The users and groups
865 systemd-network, systemd-resolve and systemd-timesync are created
866 by systemd-sysusers again. Distributors or system administrators
867 may need to create these users and groups if they not exist (or need
868 to re-enable DynamicUser= for those units) while upgrading systemd.
869 Also, the clock file for systemd-timesyncd may need to move from
870 /var/lib/private/systemd/timesync/clock to /var/lib/systemd/timesync/clock.
872 * When unit files are loaded from disk, previously systemd would
873 sometimes (depending on the unit loading order) load units from the
874 target path of symlinks in .wants/ or .requires/ directories of other
875 units. This meant that unit could be loaded from different paths
876 depending on whether the unit was requested explicitly or as a
877 dependency of another unit, not honouring the priority of directories
878 in search path. It also meant that it was possible to successfully
879 load and start units which are not found in the unit search path, as
880 long as they were requested as a dependency and linked to from
881 .wants/ or .requires/. The target paths of those symlinks are not
882 used for loading units anymore and the unit file must be found in
885 * A new service type has been added: Type=exec. It's very similar to
886 Type=simple but ensures the service manager will wait for both fork()
887 and execve() of the main service binary to complete before proceeding
888 with follow-up units. This is primarily useful so that the manager
889 propagates any errors in the preparation phase of service execution
890 back to the job that requested the unit to be started. For example,
891 consider a service that has ExecStart= set to a file system binary
892 that doesn't exist. With Type=simple starting the unit would be
893 considered instantly successful, as only fork() has to complete
894 successfully and the manager does not wait for execve(), and hence
895 its failure is seen "too late". With the new Type=exec service type
896 starting the unit will fail, as the manager will wait for the
897 execve() and notice its failure, which is then propagated back to the
900 NOTE: with the next release 241 of systemd we intend to change the
901 systemd-run tool to default to Type=exec for transient services
902 started by it. This should be mostly safe, but in specific corner
903 cases might result in problems, as the systemd-run tool will then
904 block on NSS calls (such as user name look-ups due to User=) done
905 between the fork() and execve(), which under specific circumstances
906 might cause problems. It is recommended to specify "-p Type=simple"
907 explicitly in the few cases where this applies. For regular,
908 non-transient services (i.e. those defined with unit files on disk)
909 we will continue to default to Type=simple.
911 * The Linux kernel's current default RLIMIT_NOFILE resource limit for
912 userspace processes is set to 1024 (soft) and 4096
913 (hard). Previously, systemd passed this on unmodified to all
914 processes it forked off. With this systemd release the hard limit
915 systemd passes on is increased to 512K, overriding the kernel's
916 defaults and substantially increasing the number of simultaneous file
917 descriptors unprivileged userspace processes can allocate. Note that
918 the soft limit remains at 1024 for compatibility reasons: the
919 traditional UNIX select() call cannot deal with file descriptors >=
920 1024 and increasing the soft limit globally might thus result in
921 programs unexpectedly allocating a high file descriptor and thus
922 failing abnormally when attempting to use it with select() (of
923 course, programs shouldn't use select() anymore, and prefer
924 poll()/epoll, but the call unfortunately remains undeservedly popular
925 at this time). This change reflects the fact that file descriptor
926 handling in the Linux kernel has been optimized in more recent
927 kernels and allocating large numbers of them should be much cheaper
928 both in memory and in performance than it used to be. Programs that
929 want to take benefit of the increased limit have to "opt-in" into
930 high file descriptors explicitly by raising their soft limit. Of
931 course, when they do that they must acknowledge that they cannot use
932 select() anymore (and neither can any shared library they use — or
933 any shared library used by any shared library they use and so on).
934 Which default hard limit is most appropriate is of course hard to
935 decide. However, given reports that ~300K file descriptors are used
936 in real-life applications we believe 512K is sufficiently high as new
937 default for now. Note that there are also reports that using very
938 high hard limits (e.g. 1G) is problematic: some software allocates
939 large arrays with one element for each potential file descriptor
940 (Java, …) — a high hard limit thus triggers excessively large memory
941 allocations in these applications. Hopefully, the new default of 512K
942 is a good middle ground: higher than what real-life applications
943 currently need, and low enough for avoid triggering excessively large
944 allocations in problematic software. (And yes, somebody should fix
947 * The fs.nr_open and fs.file-max sysctls are now automatically bumped
948 to the highest possible values, as separate accounting of file
949 descriptors is no longer necessary, as memcg tracks them correctly as
950 part of the memory accounting anyway. Thus, from the four limits on
951 file descriptors currently enforced (fs.file-max, fs.nr_open,
952 RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two,
953 and keep only the latter two. A set of build-time options
954 (-Dbump-proc-sys-fs-file-max=false and -Dbump-proc-sys-fs-nr-open=false)
955 has been added to revert this change in behaviour, which might be
956 an option for systems that turn off memcg in the kernel.
958 * When no /etc/locale.conf file exists (and hence no locale settings
959 are in place), systemd will now use the "C.UTF-8" locale by default,
960 and set LANG= to it. This locale is supported by various
961 distributions including Fedora, with clear indications that upstream
962 glibc is going to make it available too. This locale enables UTF-8
963 mode by default, which appears appropriate for 2018.
965 * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by
966 default. This effectively switches the RFC3704 Reverse Path filtering
967 from Strict mode to Loose mode. This is more appropriate for hosts
968 that have multiple links with routes to the same networks (e.g.
969 a client with a Wi-Fi and Ethernet both connected to the internet).
971 Consult the kernel documentation for details on this sysctl:
972 https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt
974 * CPUAccounting=yes no longer enables the CPU controller when using
975 kernel 4.15+ and the unified cgroup hierarchy, as required accounting
976 statistics are now provided independently from the CPU controller.
978 * Support for disabling a particular cgroup controller within a sub-tree
979 has been added through the DisableControllers= directive.
981 * cgroup_no_v1=all on the kernel command line now also implies
982 using the unified cgroup hierarchy, unless one explicitly passes
983 systemd.unified_cgroup_hierarchy=0 on the kernel command line.
985 * The new "MemoryMin=" unit file property may now be used to set the
986 memory usage protection limit of processes invoked by the unit. This
987 controls the cgroup v2 memory.min attribute. Similarly, the new
988 "IODeviceLatencyTargetSec=" property has been added, wrapping the new
989 cgroup v2 io.latency cgroup property for configuring per-service I/O
992 * systemd now supports the cgroup v2 devices BPF logic, as counterpart
993 to the cgroup v1 "devices" cgroup controller.
995 * systemd-escape now is able to combine --unescape with --template. It
996 also learnt a new option --instance for extracting and unescaping the
997 instance part of a unit name.
999 * sd-bus now provides the sd_bus_message_readv() which is similar to
1000 sd_bus_message_read() but takes a va_list object. The pair
1001 sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout()
1002 has been added for configuring the default method call timeout to
1003 use. sd_bus_error_move() may be used to efficiently move the contents
1004 from one sd_bus_error structure to another, invalidating the
1005 source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may
1006 be used to control whether a bus connection object is automatically
1007 flushed when an sd-event loop is exited.
1009 * When processing classic BSD syslog log messages, journald will now
1010 save the original time-stamp string supplied in the new
1011 SYSLOG_TIMESTAMP= journal field. This permits consumers to
1012 reconstruct the original BSD syslog message more correctly.
1014 * StandardOutput=/StandardError= in service files gained support for
1015 new "append:…" parameters, for connecting STDOUT/STDERR of a service
1016 to a file, and appending to it.
1018 * The signal to use as last step of killing of unit processes is now
1019 configurable. Previously it was hard-coded to SIGKILL, which may now
1020 be overridden with the new KillSignal= setting. Note that this is the
1021 signal used when regular termination (i.e. SIGTERM) does not suffice.
1022 Similarly, the signal used when aborting a program in case of a
1023 watchdog timeout may now be configured too (WatchdogSignal=).
1025 * The XDG_SESSION_DESKTOP environment variable may now be configured in
1026 the pam_systemd argument line, using the new desktop= switch. This is
1027 useful to initialize it properly from a display manager without
1028 having to touch C code.
1030 * Most configuration options that previously accepted percentage values
1031 now also accept permille values with the '‰' suffix (instead of '%').
1033 * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for
1036 * systemd-resolved's configuration file resolved.conf gained a new
1037 option ReadEtcHosts= which may be used to turn off processing and
1038 honoring /etc/hosts entries.
1040 * The "--wait" switch may now be passed to "systemctl
1041 is-system-running", in which case the tool will synchronously wait
1042 until the system finished start-up.
1044 * hostnamed gained a new bus call to determine the DMI product UUID.
1046 * On x86-64 systemd will now prefer using the RDRAND processor
1047 instruction over /dev/urandom whenever it requires randomness that
1048 neither has to be crypto-grade nor should be reproducible. This
1049 should substantially reduce the amount of entropy systemd requests
1050 from the kernel during initialization on such systems, though not
1051 reduce it to zero. (Why not zero? systemd still needs to allocate
1052 UUIDs and such uniquely, which require high-quality randomness.)
1054 * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP
1055 tunnels. It also gained a new option ForceDHCPv6PDOtherInformation=
1056 for forcing the "Other Information" bit in IPv6 RA messages. The
1057 bonding logic gained four new options AdActorSystemPriority=,
1058 AdUserPortKey=, AdActorSystem= for configuring various 802.3ad
1059 aspects, and DynamicTransmitLoadBalancing= for enabling dynamic
1060 shuffling of flows. The tunnel logic gained a new
1061 IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid
1062 Deployment. The policy rule logic gained four new options IPProtocol=,
1063 SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained
1064 support for the MulticastToUnicast= option. networkd also gained
1065 support for configuring static IPv4 ARP or IPv6 neighbor entries.
1067 * .preset files (as read by 'systemctl preset') may now be used to
1068 instantiate services.
1070 * /etc/crypttab now understands the sector-size= option to configure
1071 the sector size for an encrypted partition.
1073 * Key material for encrypted disks may now be placed on a formatted
1074 medium, and referenced from /etc/crypttab by the UUID of the file
1075 system, followed by "=" suffixed by the path to the key file.
1077 * The "collect" udev component has been removed without replacement, as
1078 it is neither used nor maintained.
1080 * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=,
1081 LogsDirectory=, ConfigurationDirectory= settings are used in a
1082 service the executed processes will now receive a set of environment
1083 variables containing the full paths of these directories.
1084 Specifically, RUNTIME_DIRECTORY=, STATE_DIRECTORY, CACHE_DIRECTORY,
1085 LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options
1086 are used. Note that these options may be used multiple times per
1087 service in which case the resulting paths will be concatenated and
1088 separated by colons.
1090 * Predictable interface naming has been extended to cover InfiniBand
1091 NICs. They will be exposed with an "ib" prefix.
1093 * tmpfiles.d/ line types may now be suffixed with a '-' character, in
1094 which case the respective line failing is ignored.
1096 * .link files may now be used to configure the equivalent to the
1097 "ethtool advertise" commands.
1099 * The sd-device.h and sd-hwdb.h APIs are now exported, as an
1100 alternative to libudev.h. Previously, the latter was just an internal
1101 wrapper around the former, but now these two APIs are exposed
1104 * sd-id128.h gained a new function sd_id128_get_boot_app_specific()
1105 which calculates an app-specific boot ID similar to how
1106 sd_id128_get_machine_app_specific() generates an app-specific machine
1109 * A new tool systemd-id128 has been added that can be used to determine
1110 and generate various 128bit IDs.
1112 * /etc/os-release gained two new standardized fields DOCUMENTATION_URL=
1115 * systemd-hibernate-resume-generator will now honor the "noresume"
1116 kernel command line option, in which case it will bypass resuming
1117 from any hibernated image.
1119 * The systemd-sleep.conf configuration file gained new options
1120 AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=,
1121 AllowHybridSleep= for prohibiting specific sleep modes even if the
1122 kernel exports them.
1124 * portablectl is now officially supported and has thus moved to
1127 * bootctl learnt the two new commands "set-default" and "set-oneshot"
1128 for setting the default boot loader item to boot to (either
1129 persistently or only for the next boot). This is currently only
1130 compatible with sd-boot, but may be implemented on other boot loaders
1131 too, that follow the boot loader interface. The updated interface is
1132 now documented here:
1134 https://systemd.io/BOOT_LOADER_INTERFACE
1136 * A new kernel command line option systemd.early_core_pattern= is now
1137 understood which may be used to influence the core_pattern PID 1
1138 installs during early boot.
1140 * busctl learnt two new options -j and --json= for outputting method
1141 call replies, properties and monitoring output in JSON.
1143 * journalctl's JSON output now supports simple ANSI coloring as well as
1144 a new "json-seq" mode for generating RFC7464 output.
1146 * Unit files now support the %g/%G specifiers that resolve to the UNIX
1147 group/GID of the service manager runs as, similar to the existing
1148 %u/%U specifiers that resolve to the UNIX user/UID.
1150 * systemd-logind learnt a new global configuration option
1151 UserStopDelaySec= that may be set in logind.conf. It specifies how
1152 long the systemd --user instance shall remain started after a user
1153 logs out. This is useful to speed up repetitive re-connections of the
1154 same user, as it means the user's service manager doesn't have to be
1155 stopped/restarted on each iteration, but can be reused between
1156 subsequent options. This setting defaults to 10s. systemd-logind also
1157 exports two new properties on its Manager D-Bus objects indicating
1158 whether the system's lid is currently closed, and whether the system
1161 * systemd gained support for a generic boot counting logic, which
1162 generically permits automatic reverting to older boot loader entries
1163 if newer updated ones don't work. The boot loader side is implemented
1164 in sd-boot, but is kept open for other boot loaders too. For details
1167 https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT
1169 * The SuccessAction=/FailureAction= unit file settings now learnt two
1170 new parameters: "exit" and "exit-force", which result in immediate
1171 exiting of the service manager, and are only useful in systemd --user
1172 and container environments.
1174 * Unit files gained support for a pair of options
1175 FailureActionExitStatus=/SuccessActionExitStatus= for configuring the
1176 exit status to use as service manager exit status when
1177 SuccessAction=/FailureAction= is set to exit or exit-force.
1179 * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service
1180 options may now be used to configure the log rate limiting applied by
1181 journald per-service.
1183 * systemd-analyze gained a new verb "timespan" for parsing and
1184 normalizing time span values (i.e. strings like "5min 7s 8us").
1186 * systemd-analyze also gained a new verb "security" for analyzing the
1187 security and sand-boxing settings of services in order to determine an
1188 "exposure level" for them, indicating whether a service would benefit
1189 from more sand-boxing options turned on for them.
1191 * "systemd-analyze syscall-filter" will now also show system calls
1192 supported by the local kernel but not included in any of the defined
1195 * .nspawn files now understand the Ephemeral= setting, matching the
1196 --ephemeral command line switch.
1198 * sd-event gained the new APIs sd_event_source_get_floating() and
1199 sd_event_source_set_floating() for controlling whether a specific
1200 event source is "floating", i.e. destroyed along with the even loop
1203 * Unit objects on D-Bus gained a new "Refs" property that lists all
1204 clients that currently have a reference on the unit (to ensure it is
1207 * The JoinControllers= option in system.conf is no longer supported, as
1208 it didn't work correctly, is hard to support properly, is legacy (as
1209 the concept only exists on cgroup v1) and apparently wasn't used.
1211 * Journal messages that are generated whenever a unit enters the failed
1212 state are now tagged with a unique MESSAGE_ID. Similarly, messages
1213 generated whenever a service process exits are now made recognizable,
1214 too. A tagged message is also emitted whenever a unit enters the
1215 "dead" state on success.
1217 * systemd-run gained a new switch --working-directory= for configuring
1218 the working directory of the service to start. A shortcut -d is
1219 equivalent, setting the working directory of the service to the
1220 current working directory of the invoking program. The new --shell
1221 (or just -S) option has been added for invoking the $SHELL of the
1222 caller as a service, and implies --pty --same-dir --wait --collect
1223 --service-type=exec. Or in other words, "systemd-run -S" is now the
1224 quickest way to quickly get an interactive in a fully clean and
1225 well-defined system service context.
1227 * machinectl gained a new verb "import-fs" for importing an OS tree
1228 from a directory. Moreover, when a directory or tarball is imported
1229 and single top-level directory found with the OS itself below the OS
1230 tree is automatically mangled and moved one level up.
1232 * systemd-importd will no longer set up an implicit btrfs loop-back
1233 file system on /var/lib/machines. If one is already set up, it will
1234 continue to be used.
1236 * A new generator "systemd-run-generator" has been added. It will
1237 synthesize a unit from one or more program command lines included in
1238 the kernel command line. This is very useful in container managers
1241 # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"'
1243 This will run "systemd-nspawn" on an image, invoke the specified
1244 command line and immediately shut down the container again, returning
1245 the command line's exit code.
1247 * The block device locking logic is now documented:
1249 https://systemd.io/BLOCK_DEVICE_LOCKING
1251 * loginctl and machinectl now optionally output the various tables in
1252 JSON using the --output= switch. It is our intention to add similar
1253 support to systemctl and all other commands.
1255 * udevadm's query and trigger verb now optionally take a .device unit
1258 * systemd-udevd's network naming logic now understands a new
1259 net.naming-scheme= kernel command line switch, which may be used to
1260 pick a specific version of the naming scheme. This helps stabilizing
1261 interface names even as systemd/udev are updated and the naming logic
1264 * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and
1265 SD_ID128_ALLF to test if a 128bit ID is set to all 0xFF bytes, and to
1266 initialize one to all 0xFF.
1268 * After loading the SELinux policy systemd will now recursively relabel
1269 all files and directories listed in
1270 /run/systemd/relabel-extra.d/*.relabel (which should be simple
1271 newline separated lists of paths) in addition to the ones it already
1272 implicitly relabels in /run, /dev and /sys. After the relabelling is
1273 completed the *.relabel files (and /run/systemd/relabel-extra.d/) are
1274 removed. This is useful to permit initrds (i.e. code running before
1275 the SELinux policy is in effect) to generate files in the host
1276 filesystem safely and ensure that the correct label is applied during
1277 the transition to the host OS.
1279 * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding
1280 mknod() handling in user namespaces. Previously mknod() would always
1281 fail with EPERM in user namespaces. Since 4.18 mknod() will succeed
1282 but device nodes generated that way cannot be opened, and attempts to
1283 open them result in EPERM. This breaks the "graceful fallback" logic
1284 in systemd's PrivateDevices= sand-boxing option. This option is
1285 implemented defensively, so that when systemd detects it runs in a
1286 restricted environment (such as a user namespace, or an environment
1287 where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD)
1288 where device nodes cannot be created the effect of PrivateDevices= is
1289 bypassed (following the logic that 2nd-level sand-boxing is not
1290 essential if the system systemd runs in is itself already sand-boxed
1291 as a whole). This logic breaks with 4.18 in container managers where
1292 user namespacing is used: suddenly PrivateDevices= succeeds setting
1293 up a private /dev/ file system containing devices nodes — but when
1294 these are opened they don't work.
1296 At this point it is recommended that container managers utilizing
1297 user namespaces that intend to run systemd in the payload explicitly
1298 block mknod() with seccomp or similar, so that the graceful fallback
1301 We are very sorry for the breakage and the requirement to change
1302 container configurations for newer kernels. It's purely caused by an
1303 incompatible kernel change. The relevant kernel developers have been
1304 notified about this userspace breakage quickly, but they chose to
1307 * PermissionsStartOnly= setting is deprecated (but is still supported
1308 for backwards compatibility). The same functionality is provided by
1309 the more flexible "+", "!", and "!!" prefixes to ExecStart= and other
1312 * $DBUS_SESSION_BUS_ADDRESS environment variable is not set by
1313 pam_systemd anymore.
1315 * The naming scheme for network devices was changed to always rename
1316 devices, even if they were already renamed by userspace. The "kernel"
1317 policy was changed to only apply as a fallback, if no other naming
1320 * The requirements to build systemd is bumped to meson-0.46 and
1323 Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander
1324 Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson,
1325 Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov,
1326 asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt
1327 Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen
1328 Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius
1329 Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn
1330 Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner,
1331 David Anderson, Davide Cavalca, David Leeds, David Malcolm, David
1332 Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus,
1333 Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters,
1334 Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad,
1335 Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank
1336 Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe
1337 Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit
1338 Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan
1339 Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld,
1340 javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi
1341 Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens,
1342 Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi,
1343 Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry
1344 Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz
1345 Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier,
1346 Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin
1347 Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko
1348 Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck,
1349 Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich,
1350 Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal
1351 Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal
1352 Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby,
1353 Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł
1354 Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller,
1355 Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez,
1356 Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam
1357 Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher,
1358 Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee
1359 (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen
1360 Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim,
1361 Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas
1362 Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias
1363 Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore
1364 Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech
1365 Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward,
1366 Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe,
1367 Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein
1369 — Warsaw, 2018-12-21
1373 * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
1374 builtin will name network interfaces differently than in previous
1375 versions for virtual network interfaces created with SR-IOV and NPAR
1376 and for devices where the PCI network controller device does not have
1377 a slot number associated.
1379 SR-IOV virtual devices are now named based on the name of the parent
1380 interface, with a suffix of "v<N>", where <N> is the virtual device
1381 number. Previously those virtual devices were named as if completely
1384 The ninth and later NPAR virtual devices will be named following the
1385 scheme used for the first eight NPAR partitions. Previously those
1386 devices were not renamed and the kernel default (eth<n>) was used.
1388 "net_id" will also generate names for PCI devices where the PCI
1389 network controller device does not have an associated slot number
1390 itself, but one of its parents does. Previously those devices were
1391 not renamed and the kernel default (eth<n>) was used.
1393 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
1394 systemd-logind.service. Since v235, IPAddressDeny=any has been set to
1395 the unit. So, it is expected that the default behavior of
1396 systemd-logind is not changed. However, if distribution packagers or
1397 administrators disabled or modified IPAddressDeny= setting by a
1398 drop-in config file, then it may be necessary to update the file to
1399 re-enable AF_INET and AF_INET6 to support network user name services,
1402 * When the RestrictNamespaces= unit property is specified multiple
1403 times, then the specified types are merged now. Previously, only the
1404 last assignment was used. So, if distribution packagers or
1405 administrators modified the setting by a drop-in config file, then it
1406 may be necessary to update the file.
1408 * When OnFailure= is used in combination with Restart= on a service
1409 unit, then the specified units will no longer be triggered on
1410 failures that result in restarting. Previously, the specified units
1411 would be activated each time the unit failed, even when the unit was
1412 going to be restarted automatically. This behaviour contradicted the
1413 documentation. With this release the code is adjusted to match the
1416 * systemd-tmpfiles will now print a notice whenever it encounters
1417 tmpfiles.d/ lines referencing the /var/run/ directory. It will
1418 recommend reworking them to use the /run/ directory instead (for
1419 which /var/run/ is simply a symlinked compatibility alias). This way
1420 systemd-tmpfiles can properly detect line conflicts and merge lines
1421 referencing the same file by two paths, without having to access
1424 * systemctl disable/unmask/preset/preset-all cannot be used with
1425 --runtime. Previously this was allowed, but resulted in unintuitive
1426 behaviour that wasn't useful. systemctl disable/unmask will now undo
1427 both runtime and persistent enablement/masking, i.e. it will remove
1428 any relevant symlinks both in /run and /etc.
1430 * Note that all long-running system services shipped with systemd will
1431 now default to a system call whitelist (rather than a blacklist, as
1432 before). In particular, systemd-udevd will now enforce one too. For
1433 most cases this should be safe, however downstream distributions
1434 which disabled sandboxing of systemd-udevd (specifically the
1435 MountFlags= setting), might want to disable this security feature
1436 too, as the default whitelisting will prohibit all mount, swap,
1437 reboot and clock changing operations from udev rules.
1439 * sd-boot acquired new loader configuration settings to optionally turn
1440 off Windows and MacOS boot partition discovery as well as
1441 reboot-into-firmware menu items. It is also able to pick a better
1442 screen resolution for HiDPI systems, and now provides loader
1443 configuration settings to change the resolution explicitly.
1445 * systemd-resolved now supports DNS-over-TLS. It's still
1446 turned off by default, use DNSOverTLS=opportunistic to turn it on in
1447 resolved.conf. We intend to make this the default as soon as couple
1448 of additional techniques for optimizing the initial latency caused by
1449 establishing a TLS/TCP connection are implemented.
1451 * systemd-resolved.service and systemd-networkd.service now set
1452 DynamicUser=yes. The users systemd-resolve and systemd-network are
1453 not created by systemd-sysusers anymore.
1455 NOTE: This has a chance of breaking nss-ldap and similar NSS modules
1456 that embed a network facing module into any process using getpwuid()
1457 or related call: the dynamic allocation of the user ID for
1458 systemd-resolved.service means the service manager has to check NSS
1459 if the user name is already taken when forking off the service. Since
1460 the user in the common case won't be defined in /etc/passwd the
1461 lookup is likely to trigger nss-ldap which in turn might use NSS to
1462 ask systemd-resolved for hostname lookups. This will hence result in
1463 a deadlock: a user name lookup in order to start
1464 systemd-resolved.service will result in a host name lookup for which
1465 systemd-resolved.service needs to be started already. There are
1466 multiple ways to work around this problem: pre-allocate the
1467 "systemd-resolve" user on such systems, so that nss-ldap won't be
1468 triggered; or use a different NSS package that doesn't do networking
1469 in-process but provides a local asynchronous name cache; or configure
1470 the NSS package to avoid lookups for UIDs in the range `pkg-config
1471 systemd --variable=dynamicuidmin` … `pkg-config systemd
1472 --variable=dynamicuidmax`, so that it does not consider itself
1473 authoritative for the same UID range systemd allocates dynamic users
1476 * The systemd-resolve tool has been renamed to resolvectl (it also
1477 remains available under the old name, for compatibility), and its
1478 interface is now verb-based, similar in style to the other <xyz>ctl
1479 tools, such as systemctl or loginctl.
1481 * The resolvectl/systemd-resolve tool also provides 'resolvconf'
1482 compatibility. It may be symlinked under the 'resolvconf' name, in
1483 which case it will take arguments and input compatible with the
1484 Debian and FreeBSD resolvconf tool.
1486 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
1487 where the system initially suspends, and after a timeout resumes and
1490 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
1491 set the client will only send a DUID as client identifier.
1493 * The nss-systemd glibc NSS module will now enumerate dynamic users and
1494 groups in effect. Previously, it could resolve UIDs/GIDs to user
1495 names/groups and vice versa, but did not support enumeration.
1497 * journald's Compress= configuration setting now optionally accepts a
1498 byte threshold value. All journal objects larger than this threshold
1499 will be compressed, smaller ones will not. Previously this threshold
1500 was not configurable and set to 512.
1502 * A new system.conf setting NoNewPrivileges= is now available which may
1503 be used to turn off acquisition of new privileges system-wide
1504 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
1505 for all its children). Note that turning this option on means setuid
1506 binaries and file system capabilities lose their special powers.
1507 While turning on this option is a big step towards a more secure
1508 system, doing so is likely to break numerous pre-existing UNIX tools,
1509 in particular su and sudo.
1511 * A new service systemd-time-sync-wait.service has been added. If
1512 enabled it will delay the time-sync.target unit at boot until time
1513 synchronization has been received from the network. This
1514 functionality is useful on systems lacking a local RTC or where it is
1515 acceptable that the boot process shall be delayed by external network
1518 * When hibernating, systemd will now inform the kernel of the image
1519 write offset, on kernels new enough to support this. This means swap
1520 files should work for hibernation now.
1522 * When loading unit files, systemd will now look for drop-in unit files
1523 extensions in additional places. Previously, for a unit file name
1524 "foo-bar-baz.service" it would look for dropin files in
1525 "foo-bar-baz.service.d/*.conf". Now, it will also look in
1526 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
1527 service name truncated after all inner dashes. This scheme allows
1528 writing drop-ins easily that apply to a whole set of unit files at
1529 once. It's particularly useful for mount and slice units (as their
1530 naming is prefix based), but is also useful for service and other
1531 units, for packages that install multiple unit files at once,
1532 following a strict naming regime of beginning the unit file name with
1533 the package's name. Two new specifiers are now supported in unit
1534 files to match this: %j and %J are replaced by the part of the unit
1535 name following the last dash.
1537 * Unit files and other configuration files that support specifier
1538 expansion now understand another three new specifiers: %T and %V will
1539 resolve to /tmp and /var/tmp respectively, or whatever temporary
1540 directory has been set for the calling user. %E will expand to either
1541 /etc (for system units) or $XDG_CONFIG_HOME (for user units).
1543 * The ExecStart= lines of unit files are no longer required to
1544 reference absolute paths. If non-absolute paths are specified the
1545 specified binary name is searched within the service manager's
1546 built-in $PATH, which may be queried with 'systemd-path
1547 search-binaries-default'. It's generally recommended to continue to
1548 use absolute paths for all binaries specified in unit files.
1550 * Units gained a new load state "bad-setting", which is used when a
1551 unit file was loaded, but contained fatal errors which prevent it
1552 from being started (for example, a service unit has been defined
1553 lacking both ExecStart= and ExecStop= lines).
1555 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
1556 support alternative debuggers, for example lldb. The old name
1557 continues to be available however, for compatibility reasons. Use the
1558 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
1559 to pick an alternative debugger instead of the default gdb.
1561 * systemctl and the other tools will now output escape sequences that
1562 generate proper clickable hyperlinks in various terminal emulators
1563 where useful (for example, in the "systemctl status" output you can
1564 now click on the unit file name to quickly open it in the
1565 editor/viewer of your choice). Note that not all terminal emulators
1566 support this functionality yet, but many do. Unfortunately, the
1567 "less" pager doesn't support this yet, hence this functionality is
1568 currently automatically turned off when a pager is started (which
1569 happens quite often due to auto-paging). We hope to remove this
1570 limitation as soon as "less" learns these escape sequences. This new
1571 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
1572 environment variable. For details on these escape sequences see:
1573 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
1575 * networkd's .network files now support a new IPv6MTUBytes= option for
1576 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
1577 option in the [Route] section to configure the MTU to use for
1578 specific routes. It also gained support for configuration of the DHCP
1579 "UserClass" option through the new UserClass= setting. It gained
1580 three new options in the new [CAN] section for configuring CAN
1581 networks. The MULTICAST and ALLMULTI interface flags may now be
1582 controlled explicitly with the new Multicast= and AllMulticast=
1585 * networkd will now automatically make use of the kernel's route
1586 expiration feature, if it is available.
1588 * udevd's .link files now support setting the number of receive and
1589 transmit channels, using the RxChannels=, TxChannels=,
1590 OtherChannels=, CombinedChannels= settings.
1592 * Support for UDPSegmentationOffload= has been removed, given its
1593 limited support in hardware, and waning software support.
1595 * networkd's .netdev files now support creating "netdevsim" interfaces.
1597 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
1598 to query the unit belonging to a specific kernel control group.
1600 * systemd-analyze gained a new verb "cat-config", which may be used to
1601 dump the contents of any configuration file, with all its matching
1602 drop-in files added in, and honouring the usual search and masking
1603 logic applied to systemd configuration files. For example use
1604 "systemd-analyze cat-config systemd/system.conf" to get the complete
1605 system configuration file of systemd how it would be loaded by PID 1
1606 itself. Similar to this, various tools such as systemd-tmpfiles or
1607 systemd-sysusers, gained a new option "--cat-config", which does the
1608 corresponding operation for their own configuration settings. For
1609 example, "systemd-tmpfiles --cat-config" will now output the full
1610 list of tmpfiles.d/ lines in place.
1612 * timedatectl gained three new verbs: "show" shows bus properties of
1613 systemd-timedated, "timesync-status" shows the current NTP
1614 synchronization state of systemd-timesyncd, and "show-timesync"
1615 shows bus properties of systemd-timesyncd.
1617 * systemd-timesyncd gained a bus interface on which it exposes details
1620 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
1621 understood by systemd-timedated. It takes a colon-separated list of
1622 unit names of NTP client services. The list is used by
1623 "timedatectl set-ntp".
1625 * systemd-nspawn gained a new --rlimit= switch for setting initial
1626 resource limits for the container payload. There's a new switch
1627 --hostname= to explicitly override the container's hostname. A new
1628 --no-new-privileges= switch may be used to control the
1629 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
1630 --oom-score-adjust= switch controls the OOM scoring adjustment value
1631 for the payload. The new --cpu-affinity= switch controls the CPU
1632 affinity of the container payload. The new --resolv-conf= switch
1633 allows more detailed control of /etc/resolv.conf handling of the
1634 container. Similarly, the new --timezone= switch allows more detailed
1635 control of /etc/localtime handling of the container.
1637 * systemd-detect-virt gained a new --list switch, which will print a
1638 list of all currently known VM and container environments.
1640 * Support for "Portable Services" has been added, see
1641 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
1642 experimental, but this is expected to change soon. Reflecting this
1643 experimental state, the "portablectl" binary is not installed into
1644 /usr/bin yet. The binary has to be called with the full path
1645 /usr/lib/systemd/portablectl instead.
1647 * journalctl's and systemctl's -o switch now knows a new log output
1648 mode "with-unit". The output it generates is very similar to the
1649 regular "short" mode, but displays the unit name instead of the
1650 syslog tag for each log line. Also, the date is shown with timezone
1651 information. This mode is probably more useful than the classic
1652 "short" output mode for most purposes, except where pixel-perfect
1653 compatibility with classic /var/log/messages formatting is required.
1655 * A new --dump-bus-properties switch has been added to the systemd
1656 binary, which may be used to dump all supported D-Bus properties.
1657 (Options which are still supported, but are deprecated, are *not*
1660 * sd-bus gained a set of new calls:
1661 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
1662 enable/disable the "floating" state of a bus slot object,
1663 i.e. whether the slot object pins the bus it is allocated for into
1664 memory or if the bus slot object gets disconnected when the bus goes
1665 away. sd_bus_open_with_description(),
1666 sd_bus_open_user_with_description(),
1667 sd_bus_open_system_with_description() may be used to allocate bus
1668 objects and set their description string already during allocation.
1670 * sd-event gained support for watching inotify events from the event
1671 loop, in an efficient way, sharing inotify handles between multiple
1672 users. For this a new function sd_event_add_inotify() has been added.
1674 * sd-event and sd-bus gained support for calling special user-supplied
1675 destructor functions for userdata pointers associated with
1676 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
1677 functions sd_bus_slot_set_destroy_callback,
1678 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
1679 sd_bus_track_get_destroy_callback,
1680 sd_event_source_set_destroy_callback,
1681 sd_event_source_get_destroy_callback have been added.
1683 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
1685 * PID 1 will now automatically reschedule .timer units whenever the
1686 local timezone changes. (They previously got rescheduled
1687 automatically when the system clock changed.)
1689 * New documentation has been added to document cgroups delegation,
1690 portable services and the various code quality tools we have set up:
1692 https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md
1693 https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md
1694 https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md
1696 * The Boot Loader Specification has been added to the source tree.
1698 https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md
1700 While moving it into our source tree we have updated it and further
1701 changes are now accepted through the usual github PR workflow.
1703 * pam_systemd will now look for PAM userdata fields systemd.memory_max,
1704 systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
1705 earlier PAM modules. The data in these fields is used to initialize
1706 the session scope's resource properties. Thus external PAM modules
1707 may now configure per-session limits, for example sourced from
1708 external user databases.
1710 * socket units with Accept=yes will now maintain a "refused" counter in
1711 addition to the existing "accepted" counter, counting connections
1712 refused due to the enforced limits.
1714 * The "systemd-path search-binaries-default" command may now be use to
1715 query the default, built-in $PATH PID 1 will pass to the services it
1718 * A new unit file setting PrivateMounts= has been added. It's a boolean
1719 option. If enabled the unit's processes are invoked in their own file
1720 system namespace. Note that this behaviour is also implied if any
1721 other file system namespacing options (such as PrivateTmp=,
1722 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
1723 primarily useful for services that do not use any of the other file
1724 system namespacing options. One such service is systemd-udevd.service
1725 where this is now used by default.
1727 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
1728 when the system is booted in UEFI "secure mode".
1730 * A new unit "system-update-pre.target" is added, which defines an
1731 optional synchronization point for offline system updates, as
1732 implemented by the pre-existing "system-update.target" unit. It
1733 allows ordering services before the service that executes the actual
1734 update process in a generic way.
1736 * Systemd now emits warnings whenever .include syntax is used.
1738 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
1739 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
1740 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
1741 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
1742 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
1743 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
1744 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
1745 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
1746 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
1747 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
1748 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
1749 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
1750 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
1751 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
1752 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
1753 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
1754 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
1755 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
1756 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
1757 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
1758 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
1759 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
1760 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
1761 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
1762 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
1763 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
1764 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
1765 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
1766 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
1768 — Berlin, 2018-06-22
1772 * The MemoryAccounting= unit property now defaults to on. After
1773 discussions with the upstream control group maintainers we learnt
1774 that the negative impact of cgroup memory accounting on current
1775 kernels is finally relatively minimal, so that it should be safe to
1776 enable this by default without affecting system performance. Besides
1777 memory accounting only task accounting is turned on by default, all
1778 other forms of resource accounting (CPU, IO, IP) remain off for now,
1779 because it's not clear yet that their impact is small enough to move
1780 from opt-in to opt-out. We recommend downstreams to leave memory
1781 accounting on by default if kernel 4.14 or higher is primarily
1782 used. On very resource constrained systems or when support for old
1783 kernels is a necessity, -Dmemory-accounting-default=false can be used
1784 to revert this change.
1786 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
1787 %udev_rules_update) and the journal catalog (%journal_catalog_update)
1788 from the upgrade scriptlets of individual packages now do nothing.
1789 Transfiletriggers have been added which will perform those updates
1790 once at the end of the transaction.
1792 Similar transfiletriggers have been added to execute any sysctl.d
1793 and binfmt.d rules. Thus, it should be unnecessary to provide any
1794 scriptlets to execute this configuration from package installation
1797 * systemd-sysusers gained a mode where the configuration to execute is
1798 specified on the command line, but this configuration is not executed
1799 directly, but instead it is merged with the configuration on disk,
1800 and the result is executed. This is useful for package installation
1801 scripts which want to create the user before installing any files on
1802 disk (in case some of those files are owned by that user), while
1803 still allowing local admin overrides.
1805 This functionality is exposed to rpm scriptlets through a new
1806 %sysusers_create_package macro. Old %sysusers_create and
1807 %sysusers_create_inline macros are deprecated.
1809 A transfiletrigger for sysusers.d configuration is now installed,
1810 which means that it should be unnecessary to call systemd-sysusers from
1811 package installation scripts, unless the package installs any files
1812 owned by those newly-created users, in which case
1813 %sysusers_create_package should be used.
1815 * Analogous change has been done for systemd-tmpfiles: it gained a mode
1816 where the command-line configuration is merged with the configuration
1817 on disk. This is exposed as the new %tmpfiles_create_package macro,
1818 and %tmpfiles_create is deprecated. A transfiletrigger is installed
1819 for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
1820 from package installation scripts.
1822 * sysusers.d configuration for a user may now also specify the group
1823 number, in addition to the user number ("u username 123:456"), or
1824 without the user number ("u username -:456").
1826 * Configution items for systemd-sysusers can now be specified as
1827 positional arguments when the new --inline switch is used.
1829 * The login shell of users created through sysusers.d may now be
1830 specified (previously, it was always /bin/sh for root and
1831 /sbin/nologin for other users).
1833 * systemd-analyze gained a new --global switch to look at global user
1834 configuration. It also gained a unit-paths verb to list the unit load
1835 paths that are compiled into systemd (which can be used with
1836 --systemd, --user, or --global).
1838 * udevadm trigger gained a new --settle/-w option to wait for any
1839 triggered events to finish (but just those, and not any other events
1840 which are triggered meanwhile).
1842 * The action that systemd-logind takes when the lid is closed and the
1843 machine is connected to external power can now be configured using
1844 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
1845 was determined by HandleLidSwitch=, and, for backwards compatibility,
1846 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
1848 * journalctl will periodically call sd_journal_process() to make it
1849 resilient against inotify queue overruns when journal files are
1850 rotated very quickly.
1852 * Two new functions in libsystemd — sd_bus_get_n_queued_read and
1853 sd_bus_get_n_queued_write — may be used to check the number of
1854 pending bus messages.
1856 * systemd gained a new
1857 org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
1858 which can be used to migrate foreign processes to scope and service
1859 units. The primary user for this new API is systemd itself: the
1860 systemd --user instance uses this call of the systemd --system
1861 instance to migrate processes if it itself gets the request to
1862 migrate processes and the kernel refuses this due to access
1863 restrictions. Thanks to this "systemd-run --scope --user …" works
1864 again in pure cgroup v2 environments when invoked from the user
1867 * A new TemporaryFileSystem= setting can be used to mask out part of
1868 the real file system tree with tmpfs mounts. This may be combined
1869 with BindPaths= and BindReadOnlyPaths= to hide files or directories
1870 not relevant to the unit, while still allowing some paths lower in
1871 the tree to be accessed.
1873 ProtectHome=tmpfs may now be used to hide user home and runtime
1874 directories from units, in a way that is mostly equivalent to
1875 "TemporaryFileSystem=/home /run/user /root".
1877 * Non-service units are now started with KeyringMode=shared by default.
1878 This means that mount and swapon and other mount tools have access
1879 to keys in the main keyring.
1881 * /sys/fs/bpf is now mounted automatically.
1883 * QNX virtualization is now detected by systemd-detect-virt and may
1884 be used in ConditionVirtualization=.
1886 * IPAccounting= may now be enabled also for slice units.
1888 * A new -Dsplit-bin= build configuration switch may be used to specify
1889 whether bin and sbin directories are merged, or if they should be
1890 included separately in $PATH and various listings of executable
1891 directories. The build configuration scripts will try to autodetect
1892 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
1893 system, but distributions are encouraged to configure this
1896 * A new -Dok-color= build configuration switch may be used to change
1897 the colour of "OK" status messages.
1899 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
1900 PrivateNetwork=yes was buggy in previous versions of systemd. This
1901 means that after the upgrade and daemon-reexec, any such units must
1904 * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
1905 will not exclude read-only files owned by root from cleanup.
1907 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
1908 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
1909 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
1910 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
1911 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
1912 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
1913 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
1914 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
1915 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
1916 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
1917 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
1918 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
1919 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
1920 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
1921 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
1922 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
1924 — Warsaw, 2018-03-05
1928 * Some keyboards come with a zoom see-saw or rocker which until now got
1929 mapped to the Linux "zoomin/out" keys in hwdb. However, these
1930 keycodes are not recognized by any major desktop. They now produce
1931 Up/Down key events so that they can be used for scrolling.
1933 * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
1934 slightly: previously, if an argument was specified for lines of this
1935 type (i.e. the right-most column was set) this string was appended to
1936 existing files each time systemd-tmpfiles was run. This behaviour was
1937 different from what the documentation said, and not particularly
1938 useful, as repeated systemd-tmpfiles invocations would not be
1939 idempotent and grow such files without bounds. With this release
1940 behaviour has been altered to match what the documentation says:
1941 lines of this type only have an effect if the indicated files don't
1942 exist yet, and only then the argument string is written to the file.
1944 * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
1945 systemd-tmpfiles behaviour: previously, read-only files owned by root
1946 were always excluded from the file "aging" algorithm (i.e. the
1947 automatic clean-up of directories like /tmp based on
1948 atime/mtime/ctime). We intend to drop this restriction, and age files
1949 by default even when owned by root and read-only. This behaviour was
1950 inherited from older tools, but there have been requests to remove
1951 it, and it's not obvious why this restriction was made in the first
1952 place. Please speak up now, if you are aware of software that reqires
1953 this behaviour, otherwise we'll remove the restriction in v238.
1955 * A new environment variable $SYSTEMD_OFFLINE is now understood by
1956 systemctl. It takes a boolean argument. If on, systemctl assumes it
1957 operates on an "offline" OS tree, and will not attempt to talk to the
1958 service manager. Previously, this mode was implicitly enabled if a
1959 chroot() environment was detected, and this new environment variable
1960 now provides explicit control.
1962 * .path and .socket units may now be created transiently, too.
1963 Previously only service, mount, automount and timer units were
1964 supported as transient units. The systemd-run tool has been updated
1965 to expose this new functionality, you may hence use it now to bind
1966 arbitrary commands to path or socket activation on-the-fly from the
1967 command line. Moreover, almost all properties are now exposed for the
1968 unit types that already supported transient operation.
1970 * The systemd-mount command gained support for a new --owner= parameter
1971 which takes a user name, which is then resolved and included in uid=
1972 and gid= mount options string of the file system to mount.
1974 * A new unit condition ConditionControlGroupController= has been added
1975 that checks whether a specific cgroup controller is available.
1977 * Unit files, udev's .link files, and systemd-networkd's .netdev and
1978 .network files all gained support for a new condition
1979 ConditionKernelVersion= for checking against specific kernel
1982 * In systemd-networkd, the [IPVLAN] section in .netdev files gained
1983 support for configuring device flags in the Flags= setting. In the
1984 same files, the [Tunnel] section gained support for configuring
1985 AllowLocalRemote=. The [Route] section in .network files gained
1986 support for configuring InitialCongestionWindow=,
1987 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
1988 understands RapidCommit=.
1990 * systemd-networkd's DHCPv6 support gained support for Prefix
1993 * sd-bus gained support for a new "watch-bind" feature. When this
1994 feature is enabled, an sd_bus connection may be set up to connect to
1995 an AF_UNIX socket in the file system as soon as it is created. This
1996 functionality is useful for writing early-boot services that
1997 automatically connect to the system bus as soon as it is started,
1998 without ugly time-based polling. systemd-networkd and
1999 systemd-resolved have been updated to make use of this
2000 functionality. busctl exposes this functionality in a new
2001 --watch-bind= command line switch.
2003 * sd-bus will now optionally synthesize a local "Connected" signal as
2004 soon as a D-Bus connection is set up fully. This message mirrors the
2005 already existing "Disconnected" signal which is synthesized when the
2006 connection is terminated. This signal is generally useful but
2007 particularly handy in combination with the "watch-bind" feature
2008 described above. Synthesizing of this message has to be requested
2009 explicitly through the new API call sd_bus_set_connected_signal(). In
2010 addition a new call sd_bus_is_ready() has been added that checks
2011 whether a connection is fully set up (i.e. between the "Connected" and
2012 "Disconnected" signals).
2014 * sd-bus gained two new calls sd_bus_request_name_async() and
2015 sd_bus_release_name_async() for asynchronously registering bus
2016 names. Similar, there is now sd_bus_add_match_async() for installing
2017 a signal match asynchronously. All of systemd's own services have
2018 been updated to make use of these calls. Doing these operations
2019 asynchronously has two benefits: it reduces the risk of deadlocks in
2020 case of cyclic dependencies between bus services, and it speeds up
2021 service initialization since synchronization points for bus
2022 round-trips are removed.
2024 * sd-bus gained two new calls sd_bus_match_signal() and
2025 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
2026 and sd_bus_add_match_async() but instead of taking a D-Bus match
2027 string take match fields as normal function parameters.
2029 * sd-bus gained two new calls sd_bus_set_sender() and
2030 sd_bus_message_set_sender() for setting the sender name of outgoing
2031 messages (either for all outgoing messages or for just one specific
2032 one). These calls are only useful in direct connections as on
2033 brokered connections the broker fills in the sender anyway,
2034 overwriting whatever the client filled in.
2036 * sd-event gained a new pseudo-handle that may be specified on all API
2037 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
2038 used this refers to the default event loop object of the calling
2039 thread. Note however that this does not implicitly allocate one —
2040 which has to be done prior by using sd_event_default(). Similarly
2041 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
2042 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
2043 to the default bus of the specified type of the calling thread. Here
2044 too this does not implicitly allocate bus connection objects, this
2045 has to be done prior with sd_bus_default() and friends.
2047 * sd-event gained a new call pair
2048 sd_event_source_{get|set}_io_fd_own(). This may be used to request
2049 automatic closure of the file descriptor an IO event source watches
2050 when the event source is destroyed.
2052 * systemd-networkd gained support for natively configuring WireGuard
2055 * In previous versions systemd synthesized user records both for the
2056 "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
2057 internally. In order to simplify distribution-wide renames of the
2058 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
2059 new transitional flag file has been added: if
2060 /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
2061 user and group record within the systemd codebase is disabled.
2063 * systemd-notify gained a new --uid= option for selecting the source
2064 user/UID to use for notification messages sent to the service
2067 * journalctl gained a new --grep= option to list only entries in which
2068 the message matches a certain pattern. By default matching is case
2069 insensitive if the pattern is lowercase, and case sensitive
2070 otherwise. Option --case-sensitive=yes|no can be used to override
2071 this an specify case sensitivity or case insensitivity.
2073 * There's now a "systemd-analyze service-watchdogs" command for printing
2074 the current state of the service runtime watchdog, and optionally
2075 enabling or disabling the per-service watchdogs system-wide if given a
2076 boolean argument (i.e. the concept you configure in WatchdogSec=), for
2077 debugging purposes. There's also a kernel command line option
2078 systemd.service_watchdogs= for controlling the same.
2080 * Two new "log-level" and "log-target" options for systemd-analyze were
2081 added that merge the now deprecated get-log-level, set-log-level and
2082 get-log-target, set-log-target pairs. The deprecated options are still
2083 understood for backwards compatibility. The two new options print the
2084 current value when no arguments are given, and set them when a
2085 level/target is given as an argument.
2087 * sysusers.d's "u" lines now optionally accept both a UID and a GID
2088 specification, separated by a ":" character, in order to create users
2089 where UID and GID do not match.
2091 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
2092 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
2093 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
2094 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
2095 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
2096 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
2097 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
2098 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
2099 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
2100 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
2101 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
2102 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
2103 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
2104 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
2105 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
2106 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
2107 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
2108 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
2109 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
2110 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
2117 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
2118 in v235 has been extended to also set the dummy.ko module option
2119 numdummies=0, preventing the kernel from automatically creating
2120 dummy0. All dummy interfaces must now be explicitly created.
2122 * Unknown '%' specifiers in configuration files are now rejected. This
2123 applies to units and tmpfiles.d configuration. Any percent characters
2124 that are followed by a letter or digit that are not supposed to be
2125 interpreted as the beginning of a specifier should be escaped by
2126 doubling ("%%"). (So "size=5%" is still accepted, as well as
2127 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
2128 valid specifiers today.)
2130 * systemd-resolved now maintains a new dynamic
2131 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
2132 recommended to make /etc/resolv.conf a symlink to it. This file
2133 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
2134 includes dynamically acquired search domains, achieving more correct
2135 DNS resolution by software that bypasses local DNS APIs such as NSS.
2137 * The "uaccess" udev tag has been dropped from /dev/kvm and
2138 /dev/dri/renderD*. These devices now have the 0666 permissions by
2139 default (but this may be changed at build-time). /dev/dri/renderD*
2140 will now be owned by the "render" group along with /dev/kfd.
2142 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
2143 systemd-journal-gatewayd.service and
2144 systemd-journal-upload.service. This means "nss-systemd" must be
2145 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
2146 services are resolved properly.
2148 * In /etc/fstab two new mount options are now understood:
2149 x-systemd.makefs and x-systemd.growfs. The former has the effect that
2150 the configured file system is formatted before it is mounted, the
2151 latter that the file system is resized to the full block device size
2152 after it is mounted (i.e. if the file system is smaller than the
2153 partition it resides on, it's grown). This is similar to the fsck
2154 logic in /etc/fstab, and pulls in systemd-makefs@.service and
2155 systemd-growfs@.service as necessary, similar to
2156 systemd-fsck@.service. Resizing is currently only supported on ext4
2159 * In systemd-networkd, the IPv6 RA logic now optionally may announce
2160 DNS server and domain information.
2162 * Support for the LUKS2 on-disk format for encrypted partitions has
2163 been added. This requires libcryptsetup2 during compilation and
2166 * The systemd --user instance will now signal "readiness" when its
2167 basic.target unit has been reached, instead of when the run queue ran
2168 empty for the first time.
2170 * Tmpfiles.d with user configuration are now also supported.
2171 systemd-tmpfiles gained a new --user switch, and snippets placed in
2172 ~/.config/user-tmpfiles.d/ and corresponding directories will be
2173 executed by systemd-tmpfiles --user running in the new
2174 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
2175 running in the user session.
2177 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
2178 %S resolves to the top-level state directory (/var/lib for the system
2179 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
2180 top-level cache directory (/var/cache for the system instance,
2181 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
2182 logs directory (/var/log for the system instance,
2183 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
2184 existing %t specifier, that resolves to the top-level runtime
2185 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
2188 * journalctl learnt a new parameter --output-fields= for limiting the
2189 set of journal fields to output in verbose and JSON output modes.
2191 * systemd-timesyncd's configuration file gained a new option
2192 RootDistanceMaxSec= for setting the maximum root distance of servers
2193 it'll use, as well as the new options PollIntervalMinSec= and
2194 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
2196 * bootctl gained a new command "list" for listing all available boot
2197 menu items on systems that follow the boot loader specification.
2199 * systemctl gained a new --dry-run switch that shows what would be done
2200 instead of doing it, and is currently supported by the shutdown and
2203 * ConditionSecurity= can now detect the TOMOYO security module.
2205 * Unit file [Install] sections are now also respected in unit drop-in
2206 files. This is intended to be used by drop-ins under /usr/lib/.
2208 * systemd-firstboot may now also set the initial keyboard mapping.
2210 * Udev "changed" events for devices which are exposed as systemd
2211 .device units are now propagated to units specified in
2212 ReloadPropagatedFrom= as reload requests.
2214 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
2215 unit template name (i.e. a name in the form of 'foobar@.service',
2216 without the instance component between the '@' and - the '.'), then
2217 the escaped sysfs path of the device is automatically used as the
2220 * SystemCallFilter= in unit files has been extended so that an "errno"
2221 can be specified individually for each system call. Example:
2222 SystemCallFilter=~uname:EILSEQ.
2224 * The cgroup delegation logic has been substantially updated. Delegate=
2225 now optionally takes a list of controllers (instead of a boolean, as
2226 before), which lists the controllers to delegate at least.
2228 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
2230 * A new LogLevelMax= setting configures the maximum log level any
2231 process of the service may log at (i.e. anything with a lesser
2232 priority than what is specified is automatically dropped). A new
2233 LogExtraFields= setting allows configuration of additional journal
2234 fields to attach to all log records generated by any of the unit's
2237 * New StandardInputData= and StandardInputText= settings along with the
2238 new option StandardInput=data may be used to configure textual or
2239 binary data that shall be passed to the executed service process via
2240 standard input, encoded in-line in the unit file.
2242 * StandardInput=, StandardOutput= and StandardError= may now be used to
2243 connect stdin/stdout/stderr of executed processes directly with a
2244 file or AF_UNIX socket in the file system, using the new "file:" option.
2246 * A new unit file option CollectMode= has been added, that allows
2247 tweaking the garbage collection logic for units. It may be used to
2248 tell systemd to garbage collect units that have failed automatically
2249 (normally it only GCs units that exited successfully). systemd-run
2250 and systemd-mount expose this new functionality with a new -G option.
2252 * "machinectl bind" may now be used to bind mount non-directories
2253 (i.e. regularfiles, devices, fifos, sockets).
2255 * systemd-analyze gained a new verb "calendar" for validating and
2256 testing calendar time specifications to use for OnCalendar= in timer
2257 units. Besides validating the expression it will calculate the next
2258 time the specified expression would elapse.
2260 * In addition to the pre-existing FailureAction= unit file setting
2261 there's now SuccessAction=, for configuring a shutdown action to
2262 execute when a unit completes successfully. This is useful in
2263 particular inside containers that shall terminate after some workload
2264 has been completed. Also, both options are now supported for all unit
2265 types, not just services.
2267 * networkds's IP rule support gained two new options
2268 IncomingInterface= and OutgoingInterface= for configuring the incoming
2269 and outgoing interfaces of configured rules. systemd-networkd also
2270 gained support for "vxcan" network devices.
2272 * networkd gained a new setting RequiredForOnline=, taking a
2273 boolean. If set, systemd-wait-online will take it into consideration
2274 when determining that the system is up, otherwise it will ignore the
2275 interface for this purpose.
2277 * The sd_notify() protocol gained support for a new operation: with
2278 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
2279 store again, ahead of POLLHUP or POLLERR when they are removed
2282 * A new document doc/UIDS-GIDS.md has been added to the source tree,
2283 that documents the UID/GID range and assignment assumptions and
2284 requirements of systemd.
2286 * The watchdog device PID 1 will ping may now be configured through the
2287 WatchdogDevice= configuration file setting, or by setting the
2288 systemd.watchdog_service= kernel commandline option.
2290 * systemd-resolved's gained support for registering DNS-SD services on
2291 the local network using MulticastDNS. Services may either be
2292 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
2293 the same dir below /run, /usr/lib), or through its D-Bus API.
2295 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
2296 extend the effective start, runtime, and stop time. The service must
2297 continue to send EXTEND_TIMEOUT_USEC within the period specified to
2298 prevent the service manager from making the service as timedout.
2300 * systemd-resolved's DNSSEC support gained support for RFC 8080
2301 (Ed25519 keys and signatures).
2303 * The systemd-resolve command line tool gained a new set of options
2304 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
2305 --set-nta= and --revert to configure per-interface DNS configuration
2306 dynamically during runtime. It's useful for pushing DNS information
2307 into systemd-resolved from DNS hook scripts that various interface
2308 managing software supports (such as pppd).
2310 * systemd-nspawn gained a new --network-namespace-path= command line
2311 option, which may be used to make a container join an existing
2312 network namespace, by specifying a path to a "netns" file.
2314 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
2315 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
2316 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
2317 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
2318 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
2319 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
2320 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
2321 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
2322 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
2323 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
2324 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
2325 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
2326 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
2327 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
2328 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
2329 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
2330 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
2331 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
2332 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
2333 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
2334 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
2335 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
2336 Jędrzejewski-Szmek, Zeal Jagannatha
2338 — Berlin, 2017-12-14
2342 * INCOMPATIBILITY: systemd-logind.service and other long-running
2343 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
2344 communication with the outside. This generally improves security of
2345 the system, and is in almost all cases a safe and good choice, as
2346 these services do not and should not provide any network-facing
2347 functionality. However, systemd-logind uses the glibc NSS API to
2348 query the user database. This creates problems on systems where NSS
2349 is set up to directly consult network services for user database
2350 lookups. In particular, this creates incompatibilities with the
2351 "nss-nis" module, which attempts to directly contact the NIS/YP
2352 network servers it is configured for, and will now consistently
2353 fail. In such cases, it is possible to turn off IP sandboxing for
2354 systemd-logind.service (set IPAddressDeny= in its [Service] section
2355 to the empty string, via a .d/ unit file drop-in). Downstream
2356 distributions might want to update their nss-nis packaging to include
2357 such a drop-in snippet, accordingly, to hide this incompatibility
2358 from the user. Another option is to make use of glibc's nscd service
2359 to proxy such network requests through a privilege-separated, minimal
2360 local caching daemon, or to switch to more modern technologies such
2361 sssd, whose NSS hook-ups generally do not involve direct network
2362 access. In general, we think it's definitely time to question the
2363 implementation choices of nss-nis, i.e. whether it's a good idea
2364 today to embed a network-facing loadable module into all local
2365 processes that need to query the user database, including the most
2366 trivial and benign ones, such as "ls". For more details about
2367 IPAddressDeny= see below.
2369 * A new modprobe.d drop-in is now shipped by default that sets the
2370 bonding module option max_bonds=0. This overrides the kernel default,
2371 to avoid conflicts and ambiguity as to whether or not bond0 should be
2372 managed by systemd-networkd or not. This resolves multiple issues
2373 with bond0 properties not being applied, when bond0 is configured
2374 with systemd-networkd. Distributors may choose to not package this,
2375 however in that case users will be prevented from correctly managing
2376 bond0 interface using systemd-networkd.
2378 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
2379 which print the logging level and target of the system manager. They
2380 complement the existing "set-log-level" and "set-log-target" verbs
2381 used to change those values.
2383 * journald.conf gained a new boolean setting ReadKMsg= which defaults
2384 to on. If turned off kernel log messages will not be read by
2385 systemd-journald or included in the logs. It also gained a new
2386 setting LineMax= for configuring the maximum line length in
2387 STDOUT/STDERR log streams. The new default for this value is 48K, up
2388 from the previous hardcoded 2048.
2390 * A new unit setting RuntimeDirectoryPreserve= has been added, which
2391 allows more detailed control of what to do with a runtime directory
2392 configured with RuntimeDirectory= (i.e. a directory below /run or
2393 $XDG_RUNTIME_DIR) after a unit is stopped.
2395 * The RuntimeDirectory= setting for units gained support for creating
2396 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
2397 one top-level directory.
2399 * Units gained new options StateDirectory=, CacheDirectory=,
2400 LogsDirectory= and ConfigurationDirectory= which are closely related
2401 to RuntimeDirectory= but manage per-service directories below
2402 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
2403 possible to write unit files which when activated automatically gain
2404 properly owned service specific directories in these locations, thus
2405 making unit files self-contained and increasing compatibility with
2406 stateless systems and factory reset where /etc or /var are
2407 unpopulated at boot. Matching these new settings there's also
2408 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
2409 ConfigurationDirectoryMode= for configuring the access mode of these
2410 directories. These settings are particularly useful in combination
2411 with DynamicUser=yes as they provide secure, properly-owned,
2412 writable, and stateful locations for storage, excluded from the
2413 sandbox that such services live in otherwise.
2415 * Automake support has been removed from this release. systemd is now
2418 * systemd-journald will now aggressively cache client metadata during
2419 runtime, speeding up log write performance under pressure. This comes
2420 at a small price though: as much of the metadata is read
2421 asynchronously from /proc/ (and isn't implicitly attached to log
2422 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
2423 metadata stored alongside a log entry might be slightly
2424 out-of-date. Previously it could only be slightly newer than the log
2425 message. The time window is small however, and given that the kernel
2426 is unlikely to be improved anytime soon in this regard, this appears
2429 * nss-myhostname/systemd-resolved will now by default synthesize an
2430 A/AAAA resource record for the "_gateway" hostname, pointing to the
2431 current default IP gateway. Previously it did that for the "gateway"
2432 name, hampering adoption, as some distributions wanted to leave that
2433 host name open for local use. The old behaviour may still be
2434 requested at build time.
2436 * systemd-networkd's [Address] section in .network files gained a new
2437 Scope= setting for configuring the IP address scope. The [Network]
2438 section gained a new boolean setting ConfigureWithoutCarrier= that
2439 tells systemd-networkd to ignore link sensing when configuring the
2440 device. The [DHCP] section gained a new Anonymize= boolean option for
2441 turning on a number of options suggested in RFC 7844. A new
2442 [RoutingPolicyRule] section has been added for configuring the IP
2443 routing policy. The [Route] section has gained support for a new
2444 Type= setting which permits configuring
2445 blackhole/unreachable/prohibit routes.
2447 * The [VRF] section in .netdev files gained a new Table= setting for
2448 configuring the routing table to use. The [Tunnel] section gained a
2449 new Independent= boolean field for configuring tunnels independent of
2450 an underlying network interface. The [Bridge] section gained a new
2451 GroupForwardMask= option for configuration of propagation of link
2452 local frames between bridge ports.
2454 * The WakeOnLan= setting in .link files gained support for a number of
2455 new modes. A new TCP6SegmentationOffload= setting has been added for
2456 configuring TCP/IPv6 hardware segmentation offload.
2458 * The IPv6 RA sender implementation may now optionally send out RDNSS
2459 and RDNSSL records to supply DNS configuration to peers.
2461 * systemd-nspawn gained support for a new --system-call-filter= command
2462 line option for adding and removing entries in the default system
2463 call filter it applies. Moreover systemd-nspawn has been changed to
2464 implement a system call whitelist instead of a blacklist.
2466 * systemd-run gained support for a new --pipe command line option. If
2467 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
2468 are directly passed on to the activated transient service
2469 executable. This allows invoking arbitrary processes as systemd
2470 services (for example to take benefit of dependency management,
2471 accounting management, resource management or log management that is
2472 done automatically for services) — while still allowing them to be
2473 integrated in a classic UNIX shell pipeline.
2475 * When a service sends RELOAD=1 via sd_notify() and reload propagation
2476 using ReloadPropagationTo= is configured, a reload is now propagated
2477 to configured units. (Previously this was only done on explicitly
2478 requested reloads, using "systemctl reload" or an equivalent
2481 * For each service unit a restart counter is now kept: it is increased
2482 each time the service is restarted due to Restart=, and may be
2483 queried using "systemctl show -p NRestarts …".
2485 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
2486 @signal and @timer have been added, for usage with SystemCallFilter=
2487 in unit files and the new --system-call-filter= command line option
2488 of systemd-nspawn (see above).
2490 * ExecStart= lines in unit files gained two new modifiers: when a
2491 command line is prefixed with "!" the command will be executed as
2492 configured, except for the credentials applied by
2493 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
2494 "+", but does still apply namespacing options unlike "+". There's
2495 also "!!" now, which is mostly identical, but becomes a NOP on
2496 systems that support ambient capabilities. This is useful to write
2497 unit files that work with ambient capabilities where possible but
2498 automatically fall back to traditional privilege dropping mechanisms
2499 on systems where this is not supported.
2501 * ListenNetlink= settings in socket units now support RDMA netlink
2504 * A new unit file setting LockPersonality= has been added which permits
2505 locking down the chosen execution domain ("personality") of a service
2508 * A new special target "getty-pre.target" has been added, which is
2509 ordered before all text logins, and may be used to order services
2510 before textual logins acquire access to the console.
2512 * systemd will now attempt to load the virtio-rng.ko kernel module very
2513 early on if a VM environment supporting this is detected. This should
2514 improve entropy during early boot in virtualized environments.
2516 * A _netdev option is now supported in /etc/crypttab that operates in a
2517 similar way as the same option in /etc/fstab: it permits configuring
2518 encrypted devices that need to be ordered after the network is up.
2519 Following this logic, two new special targets
2520 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
2521 added that are to cryptsetup.target what remote-fs.target and
2522 remote-fs-pre.target are to local-fs.target.
2524 * Service units gained a new UnsetEnvironment= setting which permits
2525 unsetting specific environment variables for services that are
2526 normally passed to it (for example in order to mask out locale
2527 settings for specific services that can't deal with it).
2529 * Units acquired a new boolean option IPAccounting=. When turned on, IP
2530 traffic accounting (packet count as well as byte count) is done for
2531 the service, and shown as part of "systemctl status" or "systemd-run
2534 * Service units acquired two new options IPAddressAllow= and
2535 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
2536 for configuring a simple IP access control list for all sockets of
2537 the unit. These options are available also on .slice and .socket
2538 units, permitting flexible access list configuration for individual
2539 services as well as groups of services (as defined by a slice unit),
2540 including system-wide. Note that IP ACLs configured this way are
2541 enforced on every single IPv4 and IPv6 socket created by any process
2542 of the service unit, and apply to ingress as well as egress traffic.
2544 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
2545 structured log message is generated each time the unit is stopped,
2546 containing information about the consumed resources of this
2549 * A new setting KeyringMode= has been added to unit files, which may be
2550 used to control how the kernel keyring is set up for executed
2553 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
2554 "systemctl kexec" and "systemctl exit" are now always asynchronous in
2555 behaviour (that is: these commands return immediately after the
2556 operation was enqueued instead of waiting for the operation to
2557 complete). Previously, "systemctl poweroff" and "systemctl reboot"
2558 were asynchronous on systems using systemd-logind (i.e. almost
2559 always, and like they were on sysvinit), and the other three commands
2560 were unconditionally synchronous. With this release this is cleaned
2561 up, and callers will see the same asynchronous behaviour on all
2562 systems for all five operations.
2564 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
2567 * .timer units now accept calendar specifications in other timezones
2568 than UTC or the local timezone.
2570 * The tmpfiles snippet var.conf has been changed to create
2571 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
2572 the "utmp" group already, and it appears to be generally understood
2573 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
2574 databases. Previously this was implemented correctly for all these
2575 databases excepts btmp, which has been opened up like this now
2576 too. Note that while the other databases are world-readable
2577 (i.e. 0644), btmp is not and remains more restrictive.
2579 * The systemd-resolve tool gained a new --reset-server-features
2580 switch. When invoked like this systemd-resolved will forget
2581 everything it learnt about the features supported by the configured
2582 upstream DNS servers, and restarts the feature probing logic on the
2583 next resolver look-up for them at the highest feature level
2586 * The status dump systemd-resolved sends to the logs upon receiving
2587 SIGUSR1 now also includes information about all DNS servers it is
2588 configured to use, and the features levels it probed for them.
2590 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
2591 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
2592 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
2593 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
2594 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
2595 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
2596 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
2597 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
2598 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
2599 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
2600 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
2601 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
2602 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
2603 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
2604 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
2605 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
2606 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
2607 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
2608 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
2609 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
2611 — Berlin, 2017-10-06
2615 * Meson is now supported as build system in addition to Automake. It is
2616 our plan to remove Automake in one of our next releases, so that
2617 Meson becomes our exclusive build system. Hence, please start using
2618 the Meson build system in your downstream packaging. There's plenty
2619 of documentation around how to use Meson, the extremely brief
2622 ./autogen.sh && ./configure && make && sudo make install
2626 meson build && ninja -C build && sudo ninja -C build install
2628 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
2629 which permits configuring a timeout on the time a job is
2630 running. This is particularly useful for setting timeouts on jobs for
2633 * Unit files gained two new options ConditionUser= and ConditionGroup=
2634 for conditionalizing units based on the identity of the user/group
2635 running a systemd user instance.
2637 * systemd-networkd now understands a new FlowLabel= setting in the
2638 [VXLAN] section of .network files, as well as a Priority= in
2639 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
2640 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
2641 gained support for configuration of GENEVE links, and IPv6 address
2642 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
2644 * .link files now understand a new Port= setting.
2646 * systemd-networkd's DHCP support gained support for DHCP option 119
2647 (domain search list).
2649 * systemd-networkd gained support for serving IPv6 address ranges using
2650 the Router Advertisement protocol. The new .network configuration
2651 section [IPv6Prefix] may be used to configure the ranges to
2652 serve. This is implemented based on a new, minimal, native server
2653 implementation of RA.
2655 * journalctl's --output= switch gained support for a new parameter
2656 "short-iso-precise" for a mode where timestamps are shown as precise
2659 * systemd-udevd's "net_id" builtin may now generate stable network
2660 interface names from IBM PowerVM VIO devices as well as ACPI platform
2663 * MulticastDNS support in systemd-resolved may now be explicitly
2664 enabled/disabled using the new MulticastDNS= configuration file
2667 * systemd-resolved may now optionally use libidn2 instead of the libidn
2668 for processing internationalized domain names. Support for libidn2
2669 should be considered experimental and should not be enabled by
2672 * "machinectl pull-tar" and related call may now do verification of
2673 downloaded images using SUSE-style .sha256 checksum files in addition
2674 to the already existing support for validating using Ubuntu-style
2677 * sd-bus gained support for a new sd_bus_message_appendv() call which
2678 is va_list equivalent of sd_bus_message_append().
2680 * sd-boot gained support for validating images using SHIM/MOK.
2682 * The SMACK code learnt support for "onlycap".
2684 * systemd-mount --umount is now much smarter in figuring out how to
2685 properly unmount a device given its mount or device path.
2687 * The code to call libnss_dns as a fallback from libnss_resolve when
2688 the communication with systemd-resolved fails was removed. This
2689 fallback was redundant and interfered with the [!UNAVAIL=return]
2690 suffix. See nss-resolve(8) for the recommended configuration.
2692 * systemd-logind may now be restarted without losing state. It stores
2693 the file descriptors for devices it manages in the system manager
2694 using the FDSTORE= mechanism. Please note that further changes in
2695 other components may be required to make use of this (for example
2696 Xorg has code to listen for stops of systemd-logind and terminate
2697 itself when logind is stopped or restarted, in order to avoid using
2698 stale file descriptors for graphical devices, which is now
2699 counterproductive and must be reverted in order for restarts of
2700 systemd-logind to be safe. See
2701 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
2703 * All kernel-install plugins are called with the environment variable
2704 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
2705 /etc/machine-id. If the machine ID could not be determined,
2706 $KERNEL_INSTALL_MACHINE_ID will be empty. Plugins should not put
2707 anything in the entry directory (passed as the second argument) if
2708 $KERNEL_INSTALL_MACHINE_ID is empty. For backwards compatibility, a
2709 temporary directory is passed as the entry directory and removed
2710 after all the plugins exit.
2712 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
2713 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
2714 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
2715 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
2716 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
2717 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
2718 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
2719 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
2720 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
2721 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
2722 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
2723 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
2724 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
2725 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
2726 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
2727 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
2728 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
2729 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
2730 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
2731 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
2732 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
2733 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
2734 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
2735 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
2736 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
2737 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
2738 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
2739 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
2740 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
2743 — Berlin, 2017-07-12
2747 * The "hybrid" control group mode has been modified to improve
2748 compatibility with "legacy" cgroups-v1 setups. Specifically, the
2749 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
2750 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
2751 cgroups-v1 hierarchy), the only externally visible change being that
2752 the cgroups-v2 hierarchy is also mounted, to
2753 /sys/fs/cgroup/unified. This should provide a large degree of
2754 compatibility with "legacy" cgroups-v1, while taking benefit of the
2755 better management capabilities of cgroups-v2.
2757 * The default control group setup mode may be selected both a boot-time
2758 via a set of kernel command line parameters (specifically:
2759 systemd.unified_cgroup_hierarchy= and
2760 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
2761 default selected on the configure command line
2762 (--with-default-hierarchy=). The upstream default is "hybrid"
2763 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
2764 this will change in a future systemd version to be "unified" (pure
2765 cgroups-v2 mode). The third option for the compile time option is
2766 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
2767 distributions to default to "hybrid" mode for release distributions,
2768 starting with v233. We recommend "unified" for development
2769 distributions (specifically: distributions such as Fedora's rawhide)
2770 as that's where things are headed in the long run. Use "legacy" for
2771 greatest stability and compatibility only.
2773 * Note one current limitation of "unified" and "hybrid" control group
2774 setup modes: the kernel currently does not permit the systemd --user
2775 instance (i.e. unprivileged code) to migrate processes between two
2776 disconnected cgroup subtrees, even if both are managed and owned by
2777 the user. This effectively means "systemd-run --user --scope" doesn't
2778 work when invoked from outside of any "systemd --user" service or
2779 scope. Specifically, it is not supported from session scopes. We are
2780 working on fixing this in a future systemd version. (See #3388 for
2781 further details about this.)
2783 * DBus policy files are now installed into /usr rather than /etc. Make
2784 sure your system has dbus >= 1.9.18 running before upgrading to this
2785 version, or override the install path with --with-dbuspolicydir= .
2787 * All python scripts shipped with systemd (specifically: the various
2788 tests written in Python) now require Python 3.
2790 * systemd unit tests can now run standalone (without the source or
2791 build directories), and can be installed into /usr/lib/systemd/tests/
2792 with 'make install-tests'.
2794 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
2795 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
2798 * Support for the %c, %r, %R specifiers in unit files has been
2799 removed. Specifiers are not supposed to be dependent on configuration
2800 in the unit file itself (so that they resolve the same regardless
2801 where used in the unit files), but these specifiers were influenced
2802 by the Slice= option.
2804 * The shell invoked by debug-shell.service now defaults to /bin/sh in
2805 all cases. If distributions want to use a different shell for this
2806 purpose (for example Fedora's /sbin/sushell) they need to specify
2807 this explicitly at configure time using --with-debug-shell=.
2809 * The confirmation spawn prompt has been reworked to offer the
2812 (c)ontinue, proceed without asking anymore
2813 (D)ump, show the state of the unit
2814 (f)ail, don't execute the command and pretend it failed
2816 (i)nfo, show a short summary of the unit
2817 (j)obs, show jobs that are in progress
2818 (s)kip, don't execute the command and pretend it succeeded
2819 (y)es, execute the command
2821 The 'n' choice for the confirmation spawn prompt has been removed,
2822 because its meaning was confusing.
2824 The prompt may now also be redirected to an alternative console by
2825 specifying the console as parameter to systemd.confirm_spawn=.
2827 * Services of Type=notify require a READY=1 notification to be sent
2828 during startup. If no such message is sent, the service now fails,
2829 even if the main process exited with a successful exit code.
2831 * Services that fail to start up correctly now always have their
2832 ExecStopPost= commands executed. Previously, they'd enter "failed"
2833 state directly, without executing these commands.
2835 * The option MulticastDNS= of network configuration files has acquired
2836 an actual implementation. With MulticastDNS=yes a host can resolve
2837 names of remote hosts and reply to mDNS A and AAAA requests.
2839 * When units are about to be started an additional check is now done to
2840 ensure that all dependencies of type BindsTo= (when used in
2841 combination with After=) have been started.
2843 * systemd-analyze gained a new verb "syscall-filter" which shows which
2844 system call groups are defined for the SystemCallFilter= unit file
2845 setting, and which system calls they contain.
2847 * A new system call filter group "@filesystem" has been added,
2848 consisting of various file system related system calls. Group
2849 "@reboot" has been added, covering reboot, kexec and shutdown related
2850 calls. Finally, group "@swap" has been added covering swap
2851 configuration related calls.
2853 * A new unit file option RestrictNamespaces= has been added that may be
2854 used to restrict access to the various process namespace types the
2855 Linux kernel provides. Specifically, it may be used to take away the
2856 right for a service unit to create additional file system, network,
2857 user, and other namespaces. This sandboxing option is particularly
2858 relevant due to the high amount of recently discovered namespacing
2859 related vulnerabilities in the kernel.
2861 * systemd-udev's .link files gained support for a new AutoNegotiation=
2862 setting for configuring Ethernet auto-negotiation.
2864 * systemd-networkd's .network files gained support for a new
2865 ListenPort= setting in the [DHCP] section to explicitly configure the
2866 UDP client port the DHCP client shall listen on.
2868 * .network files gained a new Unmanaged= boolean setting for explicitly
2869 excluding one or more interfaces from management by systemd-networkd.
2871 * The systemd-networkd ProxyARP= option has been renamed to
2872 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
2873 renamed to ReduceARPProxy=. The old names continue to be available
2876 * systemd-networkd gained support for configuring IPv6 Proxy NDP
2877 addresses via the new IPv6ProxyNDPAddress= .network file setting.
2879 * systemd-networkd's bonding device support gained support for two new
2880 configuration options ActiveSlave= and PrimarySlave=.
2882 * The various options in the [Match] section of .network files gained
2883 support for negative matching.
2885 * New systemd-specific mount options are now understood in /etc/fstab:
2887 x-systemd.mount-timeout= may be used to configure the maximum
2888 permitted runtime of the mount command.
2890 x-systemd.device-bound may be set to bind a mount point to its
2891 backing device unit, in order to automatically remove a mount point
2892 if its backing device is unplugged. This option may also be
2893 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
2894 on the block device, which is now automatically set for all CDROM
2895 drives, so that mounted CDs are automatically unmounted when they are
2896 removed from the drive.
2898 x-systemd.after= and x-systemd.before= may be used to explicitly
2899 order a mount after or before another unit or mount point.
2901 * Enqueued start jobs for device units are now automatically garbage
2902 collected if there are no jobs waiting for them anymore.
2904 * systemctl list-jobs gained two new switches: with --after, for every
2905 queued job the jobs it's waiting for are shown; with --before the
2906 jobs which it's blocking are shown.
2908 * systemd-nspawn gained support for ephemeral boots from disk images
2909 (or in other words: --ephemeral and --image= may now be
2910 combined). Moreover, ephemeral boots are now supported for normal
2911 directories, even if the backing file system is not btrfs. Of course,
2912 if the file system does not support file system snapshots or
2913 reflinks, the initial copy operation will be relatively expensive, but
2914 this should still be suitable for many use cases.
2916 * Calendar time specifications in .timer units now support
2917 specifications relative to the end of a month by using "~" instead of
2918 "-" as separator between month and day. For example, "*-02~03" means
2919 "the third last day in February". In addition a new syntax for
2920 repeated events has been added using the "/" character. For example,
2921 "9..17/2:00" means "every two hours from 9am to 5pm".
2923 * systemd-socket-proxyd gained a new parameter --connections-max= for
2924 configuring the maximum number of concurrent connections.
2926 * sd-id128 gained a new API for generating unique IDs for the host in a
2927 way that does not leak the machine ID. Specifically,
2928 sd_id128_get_machine_app_specific() derives an ID based on the
2929 machine ID a in well-defined, non-reversible, stable way. This is
2930 useful whenever an identifier for the host is needed but where the
2931 identifier shall not be useful to identify the system beyond the
2932 scope of the application itself. (Internally this uses HMAC-SHA256 as
2933 keyed hash function using the machine ID as input.)
2935 * NotifyAccess= gained a new supported value "exec". When set
2936 notifications are accepted from all processes systemd itself invoked,
2937 including all control processes.
2939 * .nspawn files gained support for defining overlay mounts using the
2940 Overlay= and OverlayReadOnly= options. Previously this functionality
2941 was only available on the systemd-nspawn command line.
2943 * systemd-nspawn's --bind= and --overlay= options gained support for
2944 bind/overlay mounts whose source lies within the container tree by
2945 prefixing the source path with "+".
2947 * systemd-nspawn's --bind= and --overlay= options gained support for
2948 automatically allocating a temporary source directory in /var/tmp
2949 that is removed when the container dies. Specifically, if the source
2950 directory is specified as empty string this mechanism is selected. An
2951 example usage is --overlay=+/var::/var, which creates an overlay
2952 mount based on the original /var contained in the image, overlayed
2953 with a temporary directory in the host's /var/tmp. This way changes
2954 to /var are automatically flushed when the container shuts down.
2956 * systemd-nspawn --image= option does now permit raw file system block
2957 devices (in addition to images containing partition tables, as
2960 * The disk image dissection logic in systemd-nspawn gained support for
2961 automatically setting up LUKS encrypted as well as Verity protected
2962 partitions. When a container is booted from an encrypted image the
2963 passphrase is queried at start-up time. When a container with Verity
2964 data is started, the root hash is search in a ".roothash" file
2965 accompanying the disk image (alternatively, pass the root hash via
2966 the new --root-hash= command line option).
2968 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
2969 be used to dissect disk images the same way as systemd-nspawn does
2970 it, following the Bootable Partition Specification. It may even be
2971 used to mount disk images with complex partition setups (including
2972 LUKS and Verity partitions) to a local host directory, in order to
2973 inspect them. This tool is not considered public API (yet), and is
2974 thus not installed into /usr/bin. Please do not rely on its
2975 existence, since it might go away or be changed in later systemd
2978 * A new generator "systemd-verity-generator" has been added, similar in
2979 style to "systemd-cryptsetup-generator", permitting automatic setup of
2980 Verity root partitions when systemd boots up. In order to make use of
2981 this your partition setup should follow the Discoverable Partitions
2982 Specification, and the GPT partition ID of the root file system
2983 partition should be identical to the upper 128bit of the Verity root
2984 hash. The GPT partition ID of the Verity partition protecting it
2985 should be the lower 128bit of the Verity root hash. If the partition
2986 image follows this model it is sufficient to specify a single
2987 "roothash=" kernel command line argument to both configure which root
2988 image and verity partition to use as well as the root hash for
2989 it. Note that systemd-nspawn's Verity support follows the same
2990 semantics, meaning that disk images with proper Verity data in place
2991 may be booted in containers with systemd-nspawn as well as on
2992 physical systems via the verity generator. Also note that the "mkosi"
2993 tool available at https://github.com/systemd/mkosi has been updated
2994 to generate Verity protected disk images following this scheme. In
2995 fact, it has been updated to generate disk images that optionally
2996 implement a complete UEFI SecureBoot trust chain, involving a signed
2997 kernel and initrd image that incorporates such a root hash as well as
2998 a Verity-enabled root partition.
3000 * The hardware database (hwdb) udev supports has been updated to carry
3001 accelerometer quirks.
3003 * All system services are now run with a fresh kernel keyring set up
3004 for them. The invocation ID is stored by default in it, thus
3005 providing a safe, non-overridable way to determine the invocation
3008 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
3009 options for bind mounting arbitrary paths in a service-specific
3010 way. When these options are used, arbitrary host or service files and
3011 directories may be mounted to arbitrary locations in the service's
3014 * Documentation has been added that lists all of systemd's low-level
3015 environment variables:
3017 https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md
3019 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
3020 whether a specific socket file descriptor matches a specified socket
3023 * systemd-firstboot has been updated to check for the
3024 systemd.firstboot= kernel command line option. It accepts a boolean
3025 and when set to false the first boot questions are skipped.
3027 * systemd-fstab-generator has been updated to check for the
3028 systemd.volatile= kernel command line option, which either takes an
3029 optional boolean parameter or the special value "state". If used the
3030 system may be booted in a "volatile" boot mode. Specifically,
3031 "systemd.volatile" is used, the root directory will be mounted as
3032 tmpfs, and only /usr is mounted from the actual root file system. If
3033 "systemd.volatile=state" is used, the root directory will be mounted
3034 as usual, but /var is mounted as tmpfs. This concept provides similar
3035 functionality as systemd-nspawn's --volatile= option, but provides it
3036 on physical boots. Use this option for implementing stateless
3037 systems, or testing systems with all state and/or configuration reset
3038 to the defaults. (Note though that many distributions are not
3039 prepared to boot up without a populated /etc or /var, though.)
3041 * systemd-gpt-auto-generator gained support for LUKS encrypted root
3042 partitions. Previously it only supported LUKS encrypted partitions
3043 for all other uses, except for the root partition itself.
3045 * Socket units gained support for listening on AF_VSOCK sockets for
3046 communication in virtualized QEMU environments.
3048 * The "configure" script gained a new option --with-fallback-hostname=
3049 for specifying the fallback hostname to use if none is configured in
3050 /etc/hostname. For example, by specifying
3051 --with-fallback-hostname=fedora it is possible to default to a
3052 hostname of "fedora" on pristine installations.
3054 * systemd-cgls gained support for a new --unit= switch for listing only
3055 the control groups of a specific unit. Similar --user-unit= has been
3056 added for listing only the control groups of a specific user unit.
3058 * systemd-mount gained a new --umount switch for unmounting a mount or
3059 automount point (and all mount/automount points below it).
3061 * systemd will now refuse full configuration reloads (via systemctl
3062 daemon-reload and related calls) unless at least 16MiB of free space
3063 are available in /run. This is a safety precaution in order to ensure
3064 that generators can safely operate after the reload completed.
3066 * A new unit file option RootImage= has been added, which has a similar
3067 effect as RootDirectory= but mounts the service's root directory from
3068 a disk image instead of plain directory. This logic reuses the same
3069 image dissection and mount logic that systemd-nspawn already uses,
3070 and hence supports any disk images systemd-nspawn supports, including
3071 those following the Discoverable Partition Specification, as well as
3072 Verity enabled images. This option enables systemd to run system
3073 services directly off disk images acting as resource bundles,
3074 possibly even including full integrity data.
3076 * A new MountAPIVFS= unit file option has been added, taking a boolean
3077 argument. If enabled /proc, /sys and /dev (collectively called the
3078 "API VFS") will be mounted for the service. This is only relevant if
3079 RootDirectory= or RootImage= is used for the service, as these mounts
3080 are of course in place in the host mount namespace anyway.
3082 * systemd-nspawn gained support for a new --pivot-root= switch. If
3083 specified the root directory within the container image is pivoted to
3084 the specified mount point, while the original root disk is moved to a
3085 different place. This option enables booting of ostree images
3086 directly with systemd-nspawn.
3088 * The systemd build scripts will no longer complain if the NTP server
3089 addresses are not changed from the defaults. Google now supports
3090 these NTP servers officially. We still recommend downstreams to
3091 properly register an NTP pool with the NTP pool project though.
3093 * coredumpctl gained a new "--reverse" option for printing the list
3094 of coredumps in reverse order.
3096 * coredumpctl will now show additional information about truncated and
3097 inaccessible coredumps, as well as coredumps that are still being
3098 processed. It also gained a new --quiet switch for suppressing
3099 additional informational message in its output.
3101 * coredumpctl gained support for only showing coredumps newer and/or
3102 older than specific timestamps, using the new --since= and --until=
3103 options, reminiscent of journalctl's options by the same name.
3105 * The systemd-coredump logic has been improved so that it may be reused
3106 to collect backtraces in non-compiled languages, for example in
3107 scripting languages such as Python.
3109 * machinectl will now show the UID shift of local containers, if user
3110 namespacing is enabled for them.
3112 * systemd will now optionally run "environment generator" binaries at
3113 configuration load time. They may be used to add environment
3114 variables to the environment block passed to services invoked. One
3115 user environment generator is shipped by default that sets up
3116 environment variables based on files dropped into /etc/environment.d
3117 and ~/.config/environment.d/.
3119 * systemd-resolved now includes the new, recently published 2017 DNSSEC
3122 * hostnamed has been updated to report a new chassis type of
3123 "convertible" to cover "foldable" laptops that can both act as a
3124 tablet and as a laptop, such as various Lenovo Yoga devices.
3126 Contributions from: Adrián López, Alexander Galanin, Alexander
3127 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
3128 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
3129 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
3130 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
3131 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
3132 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
3133 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
3134 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
3135 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
3136 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
3137 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
3138 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
3139 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
3140 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
3141 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
3142 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
3143 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
3144 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
3145 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
3146 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
3147 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
3148 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
3149 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
3150 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
3151 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
3152 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
3153 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
3156 — Berlin, 2017-03-01
3160 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
3161 RestrictAddressFamilies= enabled. These sandboxing options should
3162 generally be compatible with the various external udev call-out
3163 binaries we are aware of, however there may be exceptions, in
3164 particular when exotic languages for these call-outs are used. In
3165 this case, consider turning off these settings locally.
3167 * The new RemoveIPC= option can be used to remove IPC objects owned by
3168 the user or group of a service when that service exits.
3170 * The new ProtectKernelModules= option can be used to disable explicit
3171 load and unload operations of kernel modules by a service. In
3172 addition access to /usr/lib/modules is removed if this option is set.
3174 * ProtectSystem= option gained a new value "strict", which causes the
3175 whole file system tree with the exception of /dev, /proc, and /sys,
3176 to be remounted read-only for a service.
3178 * The new ProtectKernelTunables= option can be used to disable
3179 modification of configuration files in /sys and /proc by a service.
3180 Various directories and files are remounted read-only, so access is
3181 restricted even if the file permissions would allow it.
3183 * The new ProtectControlGroups= option can be used to disable write
3184 access by a service to /sys/fs/cgroup.
3186 * Various systemd services have been hardened with
3187 ProtectKernelTunables=yes, ProtectControlGroups=yes,
3188 RestrictAddressFamilies=.
3190 * Support for dynamically creating users for the lifetime of a service
3191 has been added. If DynamicUser=yes is specified, user and group IDs
3192 will be allocated from the range 61184..65519 for the lifetime of the
3193 service. They can be resolved using the new nss-systemd.so NSS
3194 module. The module must be enabled in /etc/nsswitch.conf. Services
3195 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
3196 any resources allocated by the service will be cleaned up when the
3197 service exits. They also have ProtectHome=read-only and
3198 ProtectSystem=strict enabled, so they are not able to make any
3199 permanent modifications to the system.
3201 * The nss-systemd module also always resolves root and nobody, making
3202 it possible to have no /etc/passwd or /etc/group files in minimal
3203 container or chroot environments.
3205 * Services may be started with their own user namespace using the new
3206 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
3207 under which the service is running are mapped. All other users are
3210 * Support for the cgroup namespace has been added to systemd-nspawn. If
3211 supported by kernel, the container system started by systemd-nspawn
3212 will have its own view of the cgroup hierarchy. This new behaviour
3213 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
3215 * The new MemorySwapMax= option can be used to limit the maximum swap
3216 usage under the unified cgroup hierarchy.
3218 * Support for the CPU controller in the unified cgroup hierarchy has
3219 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
3220 options. This controller requires out-of-tree patches for the kernel
3221 and the support is provisional.
3223 * Mount and automount units may now be created transiently
3224 (i.e. dynamically at runtime via the bus API, instead of requiring
3225 unit files in the file system).
3227 * systemd-mount is a new tool which may mount file systems – much like
3228 mount(8), optionally pulling in additional dependencies through
3229 transient .mount and .automount units. For example, this tool
3230 automatically runs fsck on a backing block device before mounting,
3231 and allows the automount logic to be used dynamically from the
3232 command line for establishing mount points. This tool is particularly
3233 useful when dealing with removable media, as it will ensure fsck is
3234 run – if necessary – before the first access and that the file system
3235 is quickly unmounted after each access by utilizing the automount
3236 logic. This maximizes the chance that the file system on the
3237 removable media stays in a clean state, and if it isn't in a clean
3238 state is fixed automatically.
3240 * LazyUnmount=yes option for mount units has been added to expose the
3241 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
3244 * /efi will be used as the mount point of the EFI boot partition, if
3245 the directory is present, and the mount point was not configured
3246 through other means (e.g. fstab). If /efi directory does not exist,
3247 /boot will be used as before. This makes it easier to automatically
3248 mount the EFI partition on systems where /boot is used for something
3251 * When operating on GPT disk images for containers, systemd-nspawn will
3252 now mount the ESP to /boot or /efi according to the same rules as PID
3253 1 running on a host. This allows tools like "bootctl" to operate
3254 correctly within such containers, in order to make container images
3255 bootable on physical systems.
3257 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
3259 * Two new user session targets have been added to support running
3260 graphical sessions under the systemd --user instance:
3261 graphical-session.target and graphical-session-pre.target. See
3262 systemd.special(7) for a description of how those targets should be
3265 * The vconsole initialization code has been significantly reworked to
3266 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
3267 support unicode keymaps. Font and keymap configuration will now be
3268 copied to all allocated virtual consoles.
3270 * FreeBSD's bhyve virtualization is now detected.
3272 * Information recorded in the journal for core dumps now includes the
3273 contents of /proc/mountinfo and the command line of the process at
3274 the top of the process hierarchy (which is usually the init process
3277 * systemd-journal-gatewayd learned the --directory= option to serve
3278 files from the specified location.
3280 * journalctl --root=… can be used to peruse the journal in the
3281 /var/log/ directories inside of a container tree. This is similar to
3282 the existing --machine= option, but does not require the container to
3285 * The hardware database has been extended to support
3286 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
3289 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
3290 specify the click rate for mice which include a horizontal wheel with
3291 a click rate that is different than the one for the vertical wheel.
3293 * systemd-run gained a new --wait option that makes service execution
3294 synchronous. (Specifically, the command will not return until the
3295 specified service binary exited.)
3297 * systemctl gained a new --wait option that causes the start command to
3298 wait until the units being started have terminated again.
3300 * A new journal output mode "short-full" has been added which displays
3301 timestamps with abbreviated English day names and adds a timezone
3302 suffix. Those timestamps include more information than the default
3303 "short" output mode, and can be passed directly to journalctl's
3304 --since= and --until= options.
3306 * /etc/resolv.conf will be bind-mounted into containers started by
3307 systemd-nspawn, if possible, so any changes to resolv.conf contents
3308 are automatically propagated to the container.
3310 * The number of instances for socket-activated services originating
3311 from a single IP address can be limited with
3312 MaxConnectionsPerSource=, extending the existing setting of
3315 * systemd-networkd gained support for vcan ("Virtual CAN") interface
3318 * .netdev and .network configuration can now be extended through
3321 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
3322 Segmentation Offload, Generic Receive Offload, Large Receive Offload
3323 can be enabled and disabled using the new UDPSegmentationOffload=,
3324 TCPSegmentationOffload=, GenericSegmentationOffload=,
3325 GenericReceiveOffload=, LargeReceiveOffload= options in the
3326 [Link] section of .link files.
3328 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
3329 Port VLAN ID can be configured for bridge devices using the new STP=,
3330 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
3331 section of .netdev files.
3333 * The route table to which routes received over DHCP or RA should be
3334 added can be configured with the new RouteTable= option in the [DHCP]
3335 and [IPv6AcceptRA] sections of .network files.
3337 * The Address Resolution Protocol can be disabled on links managed by
3338 systemd-networkd using the ARP=no setting in the [Link] section of
3341 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
3342 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
3343 encode information about the result and exit codes of the current
3344 service runtime cycle.
3346 * systemd-sysctl will now configure kernel parameters in the order
3347 they occur in the configuration files. This matches what sysctl
3348 has been traditionally doing.
3350 * kernel-install "plugins" that are executed to perform various
3351 tasks after a new kernel is added and before an old one is removed
3352 can now return a special value to terminate the procedure and
3353 prevent any later plugins from running.
3355 * Journald's SplitMode=login setting has been deprecated. It has been
3356 removed from documentation, and its use is discouraged. In a future
3357 release it will be completely removed, and made equivalent to current
3358 default of SplitMode=uid.
3360 * Storage=both option setting in /etc/systemd/coredump.conf has been
3361 removed. With fast LZ4 compression storing the core dump twice is not
3364 * The --share-system systemd-nspawn option has been replaced with an
3365 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
3366 this functionality is discouraged. In addition the variables
3367 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
3368 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
3369 individual namespaces.
3371 * "machinectl list" now shows the IP address of running containers in
3372 the output, as well as OS release information.
3374 * "loginctl list" now shows the TTY of each session in the output.
3376 * sd-bus gained new API calls sd_bus_track_set_recursive(),
3377 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
3378 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
3379 tracking objects in a "recursive" mode, where a single client can be
3380 counted multiple times, if it takes multiple references.
3382 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
3383 sd_bus_get_exit_on_disconnect(). They may be used to make a
3384 process using sd-bus automatically exit if the bus connection is
3387 * Bus clients of the service manager may now "pin" loaded units into
3388 memory, by taking an explicit reference on them. This is useful to
3389 ensure the client can retrieve runtime data about the service even
3390 after the service completed execution. Taking such a reference is
3391 available only for privileged clients and should be helpful to watch
3392 running services in a race-free manner, and in particular collect
3393 information about exit statuses and results.
3395 * The nss-resolve module has been changed to strictly return UNAVAIL
3396 when communication via D-Bus with resolved failed, and NOTFOUND when
3397 a lookup completed but was negative. This means it is now possible to
3398 neatly configure fallbacks using nsswitch.conf result checking
3399 expressions. Taking benefit of this, the new recommended
3400 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
3402 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
3404 * A new setting CtrlAltDelBurstAction= has been added to
3405 /etc/systemd/system.conf which may be used to configure the precise
3406 behaviour if the user on the console presses Ctrl-Alt-Del more often
3407 than 7 times in 2s. Previously this would unconditionally result in
3408 an expedited, immediate reboot. With this new setting the precise
3409 operation may be configured in more detail, and also turned off
3412 * In .netdev files two new settings RemoteChecksumTx= and
3413 RemoteChecksumRx= are now understood that permit configuring the
3414 remote checksumming logic for VXLAN networks.
3416 * The service manager learnt a new "invocation ID" concept for invoked
3417 services. Each runtime cycle of a service will get a new invocation
3418 ID (a 128bit random UUID) assigned that identifies the current
3419 run of the service uniquely and globally. A new invocation ID
3420 is generated each time a service starts up. The journal will store
3421 the invocation ID of a service along with any logged messages, thus
3422 making the invocation ID useful for matching the online runtime of a
3423 service with the offline log data it generated in a safe way without
3424 relying on synchronized timestamps. In many ways this new service
3425 invocation ID concept is similar to the kernel's boot ID concept that
3426 uniquely and globally identifies the runtime of each boot. The
3427 invocation ID of a service is passed to the service itself via an
3428 environment variable ($INVOCATION_ID). A new bus call
3429 GetUnitByInvocationID() has been added that is similar to GetUnit()
3430 but instead of retrieving the bus path for a unit by its name
3431 retrieves it by its invocation ID. The returned path is valid only as
3432 long as the passed invocation ID is current.
3434 * systemd-resolved gained a new "DNSStubListener" setting in
3435 resolved.conf. It either takes a boolean value or the special values
3436 "udp" and "tcp", and configures whether to enable the stub DNS
3437 listener on 127.0.0.53:53.
3439 * IP addresses configured via networkd may now carry additional
3440 configuration settings supported by the kernel. New options include:
3441 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
3442 PrefixRoute=, AutoJoin=.
3444 * The PAM configuration fragment file for "user@.service" shipped with
3445 systemd (i.e. the --user instance of systemd) has been stripped to
3446 the minimum necessary to make the system boot. Previously, it
3447 contained Fedora-specific stanzas that did not apply to other
3448 distributions. It is expected that downstream distributions add
3449 additional configuration lines, matching their needs to this file,
3450 using it only as rough template of what systemd itself needs. Note
3451 that this reduced fragment does not even include an invocation of
3452 pam_limits which most distributions probably want to add, even though
3453 systemd itself does not need it. (There's also the new build time
3454 option --with-pamconfdir=no to disable installation of the PAM
3457 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
3458 capability is now also dropped from its set (in addition to
3459 CAP_SYS_MKNOD as before).
3461 * In service unit files it is now possible to connect a specific named
3462 file descriptor with stdin/stdout/stdout of an executed service. The
3463 name may be specified in matching .socket units using the
3464 FileDescriptorName= setting.
3466 * A number of journal settings may now be configured on the kernel
3467 command line. Specifically, the following options are now understood:
3468 systemd.journald.max_level_console=,
3469 systemd.journald.max_level_store=,
3470 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
3471 systemd.journald.max_level_wall=.
3473 * "systemctl is-enabled --full" will now show by which symlinks a unit
3474 file is enabled in the unit dependency tree.
3476 * Support for VeraCrypt encrypted partitions has been added to the
3477 "cryptsetup" logic and /etc/crypttab.
3479 * systemd-detect-virt gained support for a new --private-users switch
3480 that checks whether the invoking processes are running inside a user
3481 namespace. Similar, a new special value "private-users" for the
3482 existing ConditionVirtualization= setting has been added, permitting
3483 skipping of specific units in user namespace environments.
3485 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
3486 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
3487 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
3488 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
3489 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
3490 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
3491 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
3492 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
3493 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
3494 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
3495 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
3496 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
3497 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
3498 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
3499 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
3500 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
3501 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
3502 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
3503 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
3504 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
3505 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
3506 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
3507 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
3508 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
3509 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
3510 Jędrzejewski-Szmek, Zeal Jagannatha
3512 — Santa Fe, 2016-11-03
3516 * In service units the various ExecXYZ= settings have been extended
3517 with an additional special character as first argument of the
3518 assigned value: if the character '+' is used the specified command
3519 line it will be run with full privileges, regardless of User=,
3520 Group=, CapabilityBoundingSet= and similar options. The effect is
3521 similar to the existing PermissionsStartOnly= option, but allows
3522 configuration of this concept for each executed command line
3525 * Services may now alter the service watchdog timeout at runtime by
3526 sending a WATCHDOG_USEC= message via sd_notify().
3528 * MemoryLimit= and related unit settings now optionally take percentage
3529 specifications. The percentage is taken relative to the amount of
3530 physical memory in the system (or in case of containers, the assigned
3531 amount of memory). This allows scaling service resources neatly with
3532 the amount of RAM available on the system. Similarly, systemd-logind's
3533 RuntimeDirectorySize= option now also optionally takes percentage
3536 * In similar fashion TasksMax= takes percentage values now, too. The
3537 value is taken relative to the configured maximum number of processes
3538 on the system. The per-service task maximum has been changed to 15%
3539 using this functionality. (Effectively this is an increase of 512 →
3540 4915 for service units, given the kernel's default pid_max setting.)
3542 * Calendar time specifications in .timer units now understand a ".."
3543 syntax for time ranges. Example: "4..7:10" may now be used for
3544 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
3547 * The InaccessableDirectories=, ReadOnlyDirectories= and
3548 ReadWriteDirectories= unit file settings have been renamed to
3549 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
3550 applied to all kinds of file nodes, and not just directories, with
3551 the exception of symlinks. Specifically these settings may now be
3552 used on block and character device nodes, UNIX sockets and FIFOS as
3553 well as regular files. The old names of these settings remain
3554 available for compatibility.
3556 * systemd will now log about all service processes it kills forcibly
3557 (using SIGKILL) because they remained after the clean shutdown phase
3558 of the service completed. This should help identifying services that
3559 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
3560 systemd-logind's configuration a similar log message is generated for
3561 processes killed at the end of each session due to this setting.
3563 * systemd will now set the $JOURNAL_STREAM environment variable for all
3564 services whose stdout/stderr are connected to the Journal (which
3565 effectively means by default: all services). The variable contains
3566 the device and inode number of the file descriptor used for
3567 stdout/stderr. This may be used by invoked programs to detect whether
3568 their stdout/stderr is connected to the Journal, in which case they
3569 can switch over to direct Journal communication, thus being able to
3570 pass extended, structured metadata along with their log messages. As
3571 one example, this is now used by glib's logging primitives.
3573 * When using systemd's default tmp.mount unit for /tmp, the mount point
3574 will now be established with the "nosuid" and "nodev" options. This
3575 avoids privilege escalation attacks that put traps and exploits into
3576 /tmp. However, this might cause problems if you e. g. put container
3577 images or overlays into /tmp; if you need this, override tmp.mount's
3578 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
3581 * systemd now supports the "memory" cgroup controller also on
3584 * The systemd-cgtop tool now optionally takes a control group path as
3585 command line argument. If specified, the control group list shown is
3586 limited to subgroups of that group.
3588 * The SystemCallFilter= unit file setting gained support for
3589 pre-defined, named system call filter sets. For example
3590 SystemCallFilter=@clock is now an effective way to make all clock
3591 changing-related system calls unavailable to a service. A number of
3592 similar pre-defined groups are defined. Writing system call filters
3593 for system services is simplified substantially with this new
3594 concept. Accordingly, all of systemd's own, long-running services now
3595 enable system call filtering based on this, by default.
3597 * A new service setting MemoryDenyWriteExecute= has been added, taking
3598 a boolean value. If turned on, a service may no longer create memory
3599 mappings that are writable and executable at the same time. This
3600 enhances security for services where this is enabled as it becomes
3601 harder to dynamically write and then execute memory in exploited
3602 service processes. This option has been enabled for all of systemd's
3603 own long-running services.
3605 * A new RestrictRealtime= service setting has been added, taking a
3606 boolean argument. If set the service's processes may no longer
3607 acquire realtime scheduling. This improves security as realtime
3608 scheduling may otherwise be used to easily freeze the system.
3610 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
3611 value. This may be used for requesting that the system manager inside
3612 of the container reports start-up completion to nspawn which then
3613 propagates this notification further to the service manager
3614 supervising nspawn itself. A related option NotifyReady= in .nspawn
3615 files has been added too. This functionality allows ordering of the
3616 start-up of multiple containers using the usual systemd ordering
3619 * machinectl gained a new command "stop" that is an alias for
3622 * systemd-resolved gained support for contacting DNS servers on
3623 link-local IPv6 addresses.
3625 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
3626 its caches. A method call for requesting the same operation has been
3627 added to the bus API too, and is made available via "systemd-resolve
3630 * systemd-resolve gained a new --status switch. If passed a brief
3631 summary of the used DNS configuration with per-interface information
3634 * resolved.conf gained a new Cache= boolean option, defaulting to
3635 on. If turned off local DNS caching is disabled. This comes with a
3636 performance penalty in particular when DNSSEC is enabled. Note that
3637 resolved disables its internal caching implicitly anyway, when the
3638 configured DNS server is on a host-local IP address such as ::1 or
3639 127.0.0.1, thus automatically avoiding double local caching.
3641 * systemd-resolved now listens on the local IP address 127.0.0.53:53
3642 for DNS requests. This improves compatibility with local programs
3643 that do not use the libc NSS or systemd-resolved's bus APIs for name
3644 resolution. This minimal DNS service is only available to local
3645 programs and does not implement the full DNS protocol, but enough to
3646 cover local DNS clients. A new, static resolv.conf file, listing just
3647 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
3648 now recommended to make /etc/resolv.conf a symlink to this file in
3649 order to route all DNS lookups to systemd-resolved, regardless if
3650 done via NSS, the bus API or raw DNS packets. Note that this local
3651 DNS service is not as fully featured as the libc NSS or
3652 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
3653 used to deliver link-local address information (as this implies
3654 sending a local interface index along), LLMNR/mDNS support via this
3655 interface is severely restricted. It is thus strongly recommended for
3656 all applications to use the libc NSS API or native systemd-resolved
3659 * systemd-networkd's bridge support learned a new setting
3660 VLANFiltering= for controlling VLAN filtering. Moreover a new section
3661 in .network files has been added for configuring VLAN bridging in
3662 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
3664 * systemd-networkd's IPv6 Router Advertisement code now makes use of
3665 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
3666 now be acquired without relying on DHCPv6. Two new options
3667 UseDomains= and UseDNS= have been added to configure this behaviour.
3669 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
3670 renamed IPv6AcceptRA=, without altering its behaviour. The old
3671 setting name remains available for compatibility reasons.
3673 * The systemd-networkd VTI/VTI6 tunneling support gained new options
3674 Key=, InputKey= and OutputKey=.
3676 * systemd-networkd gained support for VRF ("Virtual Routing Function")
3677 interface configuration.
3679 * "systemctl edit" may now be used to create new unit files by
3680 specifying the --force switch.
3682 * sd-event gained a new function sd_event_get_iteration() for
3683 requesting the current iteration counter of the event loop. It starts
3684 at zero and is increased by one with each event loop iteration.
3686 * A new rpm macro %systemd_ordering is provided by the macros.systemd
3687 file. It can be used in lieu of %systemd_requires in packages which
3688 don't use any systemd functionality and are intended to be installed
3689 in minimal containers without systemd present. This macro provides
3690 ordering dependencies to ensure that if the package is installed in
3691 the same rpm transaction as systemd, systemd will be installed before
3692 the scriptlets for the package are executed, allowing unit presets
3695 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
3696 been added to simplify packaging of generators.
3698 * The os-release file gained VERSION_CODENAME field for the
3699 distribution nickname (e.g. VERSION_CODENAME=woody).
3701 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
3702 can be set to disable parsing of metadata and the creation
3703 of persistent symlinks for that device.
3705 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
3706 to make them available to logged-in users has been reverted.
3708 * Much of the common code of the various systemd components is now
3709 built into an internal shared library libsystemd-shared-231.so
3710 (incorporating the systemd version number in the name, to be updated
3711 with future releases) that the components link to. This should
3712 decrease systemd footprint both in memory during runtime and on
3713 disk. Note that the shared library is not for public use, and is
3714 neither API nor ABI stable, but is likely to change with every new
3715 released update. Packagers need to make sure that binaries
3716 linking to libsystemd-shared.so are updated in step with the
3719 * Configuration for "mkosi" is now part of the systemd
3720 repository. mkosi is a tool to easily build legacy-free OS images,
3721 and is available on github: https://github.com/systemd/mkosi. If
3722 "mkosi" is invoked in the build tree a new raw OS image is generated
3723 incorporating the systemd sources currently being worked on and a
3724 clean, fresh distribution installation. The generated OS image may be
3725 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
3726 UEFI PC. This functionality is particularly useful to easily test
3727 local changes made to systemd in a pristine, defined environment. See
3728 doc/HACKING for details.
3730 * configure learned the --with-support-url= option to specify the
3731 distribution's bugtracker.
3733 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
3734 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
3735 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
3736 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
3737 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
3738 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
3739 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
3740 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
3741 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
3742 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
3743 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
3744 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
3745 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
3746 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
3747 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
3748 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
3749 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
3750 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
3751 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
3753 — Berlin, 2016-07-25
3757 * DNSSEC is now turned on by default in systemd-resolved (in
3758 "allow-downgrade" mode), but may be turned off during compile time by
3759 passing "--with-default-dnssec=no" to "configure" (and of course,
3760 during runtime with DNSSEC= in resolved.conf). We recommend
3761 downstreams to leave this on at least during development cycles and
3762 report any issues with the DNSSEC logic upstream. We are very
3763 interested in collecting feedback about the DNSSEC validator and its
3764 limitations in the wild. Note however, that DNSSEC support is
3765 probably nothing downstreams should turn on in stable distros just
3766 yet, as it might create incompatibilities with a few DNS servers and
3767 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
3768 automatically whenever we detect such incompatible setups, but there
3769 might be systems we do not cover yet. Hence: please help us testing
3770 the DNSSEC code, leave this on where you can, report back, but then
3771 again don't consider turning this on in your stable, LTS or
3772 production release just yet. (Note that you have to enable
3773 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
3774 and its DNSSEC mode for host name resolution from local
3777 * systemd-resolve conveniently resolves DANE records with the --tlsa
3778 option and OPENPGPKEY records with the --openpgp option. It also
3779 supports dumping raw DNS record data via the new --raw= switch.
3781 * systemd-logind will now by default terminate user processes that are
3782 part of the user session scope unit (session-XX.scope) when the user
3783 logs out. This behavior is controlled by the KillUserProcesses=
3784 setting in logind.conf, and the previous default of "no" is now
3785 changed to "yes". This means that user sessions will be properly
3786 cleaned up after, but additional steps are necessary to allow
3787 intentionally long-running processes to survive logout.
3789 While the user is logged in at least once, user@.service is running,
3790 and any service that should survive the end of any individual login
3791 session can be started at a user service or scope using systemd-run.
3792 systemd-run(1) man page has been extended with an example which shows
3793 how to run screen in a scope unit underneath user@.service. The same
3794 command works for tmux.
3796 After the user logs out of all sessions, user@.service will be
3797 terminated too, by default, unless the user has "lingering" enabled.
3798 To effectively allow users to run long-term tasks even if they are
3799 logged out, lingering must be enabled for them. See loginctl(1) for
3800 details. The default polkit policy was modified to allow users to
3801 set lingering for themselves without authentication.
3803 Previous defaults can be restored at compile time by the
3804 --without-kill-user-processes option to "configure".
3806 * systemd-logind gained new configuration settings SessionsMax= and
3807 InhibitorsMax=, both with a default of 8192. It will not register new
3808 user sessions or inhibitors above this limit.
3810 * systemd-logind will now reload configuration on SIGHUP.
3812 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
3813 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
3814 enable. Also, support for the "io" cgroup controller in the unified
3815 hierarchy has been added, so that the "memory", "pids" and "io" are
3816 now the controllers that are supported on the unified hierarchy.
3818 WARNING: it is not possible to use previous systemd versions with
3819 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
3820 is necessary to also update systemd in the initramfs if using the
3821 unified hierarchy. An updated SELinux policy is also required.
3823 * LLDP support has been extended, and both passive (receive-only) and
3824 active (sender) modes are supported. Passive mode ("routers-only") is
3825 enabled by default in systemd-networkd. Active LLDP mode is enabled
3826 by default for containers on the internal network. The "networkctl
3827 lldp" command may be used to list information gathered. "networkctl
3828 status" will also show basic LLDP information on connected peers now.
3830 * The IAID and DUID unique identifier sent in DHCP requests may now be
3831 configured for the system and each .network file managed by
3832 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
3834 * systemd-networkd gained support for configuring proxy ARP support for
3835 each interface, via the ProxyArp= setting in .network files. It also
3836 gained support for configuring the multicast querier feature of
3837 bridge devices, via the new MulticastQuerier= setting in .netdev
3838 files. Similarly, snooping on the IGMP traffic can be controlled
3839 via the new setting MulticastSnooping=.
3841 A new setting PreferredLifetime= has been added for addresses
3842 configured in .network file to configure the lifetime intended for an
3845 The systemd-networkd DHCP server gained the option EmitRouter=, which
3846 defaults to yes, to configure whether the DHCP Option 3 (Router)
3849 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
3850 systemd-socket-activate and installed into /usr/bin. It is now fully
3853 * systemd-journald now uses separate threads to flush changes to disk
3854 when closing journal files, thus reducing impact of slow disk I/O on
3855 logging performance.
3857 * The sd-journal API gained two new calls
3858 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
3859 can be used to open journal files using file descriptors instead of
3860 file or directory paths. sd_journal_open_container() has been
3861 deprecated, sd_journal_open_directory_fd() should be used instead
3862 with the flag SD_JOURNAL_OS_ROOT.
3864 * journalctl learned a new output mode "-o short-unix" that outputs log
3865 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
3866 UTC). It also gained support for a new --no-hostname setting to
3867 suppress the hostname column in the family of "short" output modes.
3869 * systemd-ask-password now optionally skips printing of the password to
3870 stdout with --no-output which can be useful in scripts.
3872 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
3873 (devices tagged with ID_MAKER_TOOL) are now tagged with
3874 "uaccess" and are available to logged in users.
3876 * The DeviceAllow= unit setting now supports specifiers (with "%").
3878 * "systemctl show" gained a new --value switch, which allows print a
3879 only the contents of a specific unit property, without also printing
3880 the property's name. Similar support was added to "show*" verbs
3881 of loginctl and machinectl that output "key=value" lists.
3883 * A new unit type "generated" was added for files dynamically generated
3884 by generator tools. Similarly, a new unit type "transient" is used
3885 for unit files created using the runtime API. "systemctl enable" will
3886 refuse to operate on such files.
3888 * A new command "systemctl revert" has been added that may be used to
3889 revert to the vendor version of a unit file, in case local changes
3890 have been made by adding drop-ins or overriding the unit file.
3892 * "machinectl clean" gained a new verb to automatically remove all or
3893 just hidden container images.
3895 * systemd-tmpfiles gained support for a new line type "e" for emptying
3896 directories, if they exist, without creating them if they don't.
3898 * systemd-nspawn gained support for automatically patching the UID/GIDs
3899 of the owners and the ACLs of all files and directories in a
3900 container tree to match the UID/GID user namespacing range selected
3901 for the container invocation. This mode is enabled via the new
3902 --private-users-chown switch. It also gained support for
3903 automatically choosing a free, previously unused UID/GID range when
3904 starting a container, via the new --private-users=pick setting (which
3905 implies --private-users-chown). Together, these options for the first
3906 time make user namespacing for nspawn containers fully automatic and
3907 thus deployable. The systemd-nspawn@.service template unit file has
3908 been changed to use this functionality by default.
3910 * systemd-nspawn gained a new --network-zone= switch, that allows
3911 creating ad-hoc virtual Ethernet links between multiple containers,
3912 that only exist as long as at least one container referencing them is
3913 running. This allows easy connecting of multiple containers with a
3914 common link that implements an Ethernet broadcast domain. Each of
3915 these network "zones" may be named relatively freely by the user, and
3916 may be referenced by any number of containers, but each container may
3917 only reference one of these "zones". On the lower level, this is
3918 implemented by an automatically managed bridge network interface for
3919 each zone, that is created when the first container referencing its
3920 zone is created and removed when the last one referencing its zone
3923 * The default start timeout may now be configured on the kernel command
3924 line via systemd.default_timeout_start_sec=. It was already
3925 configurable via the DefaultTimeoutStartSec= option in
3926 /etc/systemd/system.conf.
3928 * Socket units gained a new TriggerLimitIntervalSec= and
3929 TriggerLimitBurst= setting to configure a limit on the activation
3930 rate of the socket unit.
3932 * The LimitNICE= setting now optionally takes normal UNIX nice values
3933 in addition to the raw integer limit value. If the specified
3934 parameter is prefixed with "+" or "-" and is in the range -20..19 the
3935 value is understood as UNIX nice value. If not prefixed like this it
3936 is understood as raw RLIMIT_NICE limit.
3938 * Note that the effect of the PrivateDevices= unit file setting changed
3939 slightly with this release: the per-device /dev file system will be
3940 mounted read-only from this version on, and will have "noexec"
3941 set. This (minor) change of behavior might cause some (exceptional)
3942 legacy software to break, when PrivateDevices=yes is set for its
3943 service. Please leave PrivateDevices= off if you run into problems
3946 * systemd-bootchart has been split out to a separate repository:
3947 https://github.com/systemd/systemd-bootchart
3949 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
3950 merged into the kernel in its current form.
3952 * The compatibility libraries libsystemd-daemon.so,
3953 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
3954 which have been deprecated since systemd-209 have been removed along
3955 with the corresponding pkg-config files. All symbols provided by
3956 those libraries are provided by libsystemd.so.
3958 * The Capabilities= unit file setting has been removed (it is ignored
3959 for backwards compatibility). AmbientCapabilities= and
3960 CapabilityBoundingSet= should be used instead.
3962 * A new special target has been added, initrd-root-device.target,
3963 which creates a synchronization point for dependencies of the root
3964 device in early userspace. Initramfs builders must ensure that this
3965 target is now included in early userspace.
3967 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
3968 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
3969 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
3970 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
3971 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
3972 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
3973 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
3974 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
3975 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
3976 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
3977 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
3978 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
3979 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
3980 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
3981 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
3982 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
3983 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
3984 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
3985 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
3986 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
3987 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
3988 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
3989 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
3990 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
3991 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
3994 — Fairfax, 2016-05-21
3998 * The systemd-resolved DNS resolver service has gained a substantial
3999 set of new features, most prominently it may now act as a DNSSEC
4000 validating stub resolver. DNSSEC mode is currently turned off by
4001 default, but is expected to be turned on by default in one of the
4002 next releases. For now, we invite everybody to test the DNSSEC logic
4003 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
4004 service also gained a full set of D-Bus interfaces, including calls
4005 to configure DNS and DNSSEC settings per link (for use by external
4006 network management software). systemd-resolved and systemd-networkd
4007 now distinguish between "search" and "routing" domains. The former
4008 are used to qualify single-label names, the latter are used purely
4009 for routing lookups within certain domains to specific links.
4010 resolved now also synthesizes RRs for all entries from /etc/hosts.
4012 * The systemd-resolve tool (which is a client utility for
4013 systemd-resolved) has been improved considerably and is now fully
4014 supported and documented. Hence it has moved from /usr/lib/systemd to
4017 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
4020 * The coredump collection logic has been reworked: when a coredump is
4021 collected it is now written to disk, compressed and processed
4022 (including stacktrace extraction) from a new instantiated service
4023 systemd-coredump@.service, instead of directly from the
4024 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
4025 processing large coredumps can take up a substantial amount of
4026 resources and time, and this previously happened entirely outside of
4027 systemd's service supervision. With the new logic the core_pattern
4028 hook only does minimal metadata collection before passing off control
4029 to the new instantiated service, which is configured with a time
4030 limit, a nice level and other settings to minimize negative impact on
4031 the rest of the system. Also note that the new logic will honour the
4032 RLIMIT_CORE setting of the crashed process, which now allows users
4033 and processes to turn off coredumping for their processes by setting
4036 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
4037 and all forked processes by default. Previously, PID 1 would leave
4038 the setting at "0" for all processes, as set by the kernel. Note that
4039 the resource limit traditionally has no effect on the generated
4040 coredumps on the system if the /proc/sys/kernel/core_pattern hook
4041 logic is used. Since the limit is now honoured (see above) its
4042 default has been changed so that the coredumping logic is enabled by
4043 default for all processes, while allowing specific opt-out.
4045 * When the stacktrace is extracted from processes of system users, this
4046 is now done as "systemd-coredump" user, in order to sandbox this
4047 potentially security sensitive parsing operation. (Note that when
4048 processing coredumps of normal users this is done under the user ID
4049 of process that crashed, as before.) Packagers should take notice
4050 that it is now necessary to create the "systemd-coredump" system user
4051 and group at package installation time.
4053 * The systemd-activate socket activation testing tool gained support
4054 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
4055 and --seqpacket switches. It also has been extended to support both
4056 new-style and inetd-style file descriptor passing. Use the new
4057 --inetd switch to request inetd-style file descriptor passing.
4059 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
4060 variable, which takes a boolean value. If set to false, ANSI color
4061 output is disabled in the tools even when run on a terminal that
4064 * The VXLAN support in networkd now supports two new settings
4065 DestinationPort= and PortRange=.
4067 * A new systemd.machine_id= kernel command line switch has been added,
4068 that may be used to set the machine ID in /etc/machine-id if it is
4069 not initialized yet. This command line option has no effect if the
4070 file is already initialized.
4072 * systemd-nspawn gained a new --as-pid2 switch that invokes any
4073 specified command line as PID 2 rather than PID 1 in the
4074 container. In this mode PID 1 is a minimal stub init process that
4075 implements the special POSIX and Linux semantics of PID 1 regarding
4076 signal and child process management. Note that this stub init process
4077 is implemented in nspawn itself and requires no support from the
4078 container image. This new logic is useful to support running
4079 arbitrary commands in the container, as normal processes are
4080 generally not prepared to run as PID 1.
4082 * systemd-nspawn gained a new --chdir= switch for setting the current
4083 working directory for the process started in the container.
4085 * "journalctl /dev/sda" will now output all kernel log messages for
4086 specified device from the current boot, in addition to all devices
4087 that are parents of it. This should make log output about devices
4088 pretty useful, as long as kernel drivers attach enough metadata to
4089 the log messages. (The usual SATA drivers do.)
4091 * The sd-journal API gained two new calls
4092 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
4093 that report whether log data from /run or /var has been found.
4095 * journalctl gained a new switch "--fields" that prints all journal
4096 record field names currently in use in the journal. This is backed
4097 by two new sd-journal API calls sd_journal_enumerate_fields() and
4098 sd_journal_restart_fields().
4100 * Most configurable timeouts in systemd now expect an argument of
4101 "infinity" to turn them off, instead of "0" as before. The semantics
4102 from now on is that a timeout of "0" means "now", and "infinity"
4103 means "never". To maintain backwards compatibility, "0" continues to
4104 turn off previously existing timeout settings.
4106 * "systemctl reload-or-try-restart" has been renamed to "systemctl
4107 try-reload-or-restart" to clarify what it actually does: the "try"
4108 logic applies to both reloading and restarting, not just restarting.
4109 The old name continues to be accepted for compatibility.
4111 * On boot-up, when PID 1 detects that the system clock is behind the
4112 release date of the systemd version in use, the clock is now set
4113 to the latter. Previously, this was already done in timesyncd, in order
4114 to avoid running with clocks set to the various clock epochs such as
4115 1902, 1938 or 1970. With this change the logic is now done in PID 1
4116 in addition to timesyncd during early boot-up, so that it is enforced
4117 before the first process is spawned by systemd. Note that the logic
4118 in timesyncd remains, as it is more comprehensive and ensures
4119 clock monotonicity by maintaining a persistent timestamp file in
4120 /var. Since /var is generally not available in earliest boot or the
4121 initrd, this part of the logic remains in timesyncd, and is not done
4124 * Support for tweaking details in net_cls.class_id through the
4125 NetClass= configuration directive has been removed, as the kernel
4126 people have decided to deprecate that controller in cgroup v2.
4127 Userspace tools such as nftables are moving over to setting rules
4128 that are specific to the full cgroup path of a task, which obsoletes
4129 these controllers anyway. The NetClass= directive is kept around for
4130 legacy compatibility reasons. For a more in-depth description of the
4131 kernel change, please refer to the respective upstream commit:
4133 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
4135 * A new service setting RuntimeMaxSec= has been added that may be used
4136 to specify a maximum runtime for a service. If the timeout is hit, the
4137 service is terminated and put into a failure state.
4139 * A new service setting AmbientCapabilities= has been added. It allows
4140 configuration of additional Linux process capabilities that are
4141 passed to the activated processes. This is only available on very
4144 * The process resource limit settings in service units may now be used
4145 to configure hard and soft limits individually.
4147 * The various libsystemd APIs such as sd-bus or sd-event now publicly
4148 expose support for gcc's __attribute__((cleanup())) C extension.
4149 Specifically, for many object destructor functions alternative
4150 versions have been added that have names suffixed with "p" and take a
4151 pointer to a pointer to the object to destroy, instead of just a
4152 pointer to the object itself. This is useful because these destructor
4153 functions may be used directly as parameters to the cleanup
4154 construct. Internally, systemd has been a heavy user of this GCC
4155 extension for a long time, and with this change similar support is
4156 now available to consumers of the library outside of systemd. Note
4157 that by using this extension in your sources compatibility with old
4158 and strictly ANSI compatible C compilers is lost. However, all gcc or
4159 LLVM versions of recent years support this extension.
4161 * Timer units gained support for a new setting RandomizedDelaySec= that
4162 allows configuring some additional randomized delay to the configured
4163 time. This is useful to spread out timer events to avoid load peaks in
4164 clusters or larger setups.
4166 * Calendar time specifications now support sub-second accuracy.
4168 * Socket units now support listening on SCTP and UDP-lite protocol
4171 * The sd-event API now comes with a full set of man pages.
4173 * Older versions of systemd contained experimental support for
4174 compressing journal files and coredumps with the LZ4 compressor that
4175 was not compatible with the lz4 binary (due to API limitations of the
4176 lz4 library). This support has been removed; only support for files
4177 compatible with the lz4 binary remains. This LZ4 logic is now
4178 officially supported and no longer considered experimental.
4180 * The dkr image import logic has been removed again from importd. dkr's
4181 micro-services focus doesn't fit into the machine image focus of
4182 importd, and quickly got out of date with the upstream dkr API.
4184 * Creation of the /run/lock/lockdev/ directory was dropped from
4185 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
4186 been available for many years. If you still need this, you need to
4187 create your own tmpfiles.d config file with:
4189 d /run/lock/lockdev 0775 root lock -
4191 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
4192 and RebootArgument= have been moved from the [Service] section of
4193 unit files to [Unit], and they are now supported on all unit types,
4194 not just service units. Of course, systemd will continue to
4195 understand these settings also at the old location, in order to
4196 maintain compatibility.
4198 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
4199 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
4200 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
4201 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
4202 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
4203 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
4204 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
4205 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
4206 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
4207 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
4208 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
4209 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
4210 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
4211 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
4212 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
4213 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
4214 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
4215 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
4216 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
4218 — Berlin, 2016-02-11
4222 * A number of properties previously only settable in unit
4223 files are now also available as properties to set when
4224 creating transient units programmatically via the bus, as it
4225 is exposed with systemd-run's --property=
4226 setting. Specifically, these are: SyslogIdentifier=,
4227 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
4228 EnvironmentFile=, ReadWriteDirectories=,
4229 ReadOnlyDirectories=, InaccessibleDirectories=,
4230 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
4232 * When creating transient services via the bus API it is now
4233 possible to pass in a set of file descriptors to use as
4234 STDIN/STDOUT/STDERR for the invoked process.
4236 * Slice units may now be created transiently via the bus APIs,
4237 similar to the way service and scope units may already be
4238 created transiently.
4240 * Wherever systemd expects a calendar timestamp specification
4241 (like in journalctl's --since= and --until= switches) UTC
4242 timestamps are now supported. Timestamps suffixed with "UTC"
4243 are now considered to be in Universal Time Coordinated
4244 instead of the local timezone. Also, timestamps may now
4245 optionally be specified with sub-second accuracy. Both of
4246 these additions also apply to recurring calendar event
4247 specification, such as OnCalendar= in timer units.
4249 * journalctl gained a new "--sync" switch that asks the
4250 journal daemon to write all so far unwritten log messages to
4251 disk and sync the files, before returning.
4253 * systemd-tmpfiles learned two new line types "q" and "Q" that
4254 operate like "v", but also set up a basic btrfs quota
4255 hierarchy when used on a btrfs file system with quota
4258 * tmpfiles' "v", "q" and "Q" will now create a plain directory
4259 instead of a subvolume (even on a btrfs file system) if the
4260 root directory is a plain directory, and not a
4261 subvolume. This should simplify things with certain chroot()
4262 environments which are not aware of the concept of btrfs
4265 * systemd-detect-virt gained a new --chroot switch to detect
4266 whether execution takes place in a chroot() environment.
4268 * CPUAffinity= now takes CPU index ranges in addition to
4271 * The various memory-related resource limit settings (such as
4272 LimitAS=) now understand the usual K, M, G, ... suffixes to
4273 the base of 1024 (IEC). Similar, the time-related resource
4274 limit settings understand the usual min, h, day, ...
4277 * There's a new system.conf setting DefaultTasksMax= to
4278 control the default TasksMax= setting for services and
4279 scopes running on the system. (TasksMax= is the primary
4280 setting that exposes the "pids" cgroup controller on systemd
4281 and was introduced in the previous systemd release.) The
4282 setting now defaults to 512, which means services that are
4283 not explicitly configured otherwise will only be able to
4284 create 512 processes or threads at maximum, from this
4285 version on. Note that this means that thread- or
4286 process-heavy services might need to be reconfigured to set
4287 TasksMax= to a higher value. It is sufficient to set
4288 TasksMax= in these specific unit files to a higher value, or
4289 even "infinity". Similar, there's now a logind.conf setting
4290 UserTasksMax= that defaults to 4096 and limits the total
4291 number of processes or tasks each user may own
4292 concurrently. nspawn containers also have the TasksMax=
4293 value set by default now, to 8192. Note that all of this
4294 only has an effect if the "pids" cgroup controller is
4295 enabled in the kernel. The general benefit of these changes
4296 should be a more robust and safer system, that provides a
4297 certain amount of per-service fork() bomb protection.
4299 * systemd-nspawn gained the new --network-veth-extra= switch
4300 to define additional and arbitrarily-named virtual Ethernet
4301 links between the host and the container.
4303 * A new service execution setting PassEnvironment= has been
4304 added that allows importing select environment variables
4305 from PID1's environment block into the environment block of
4308 * Timer units gained support for a new RemainAfterElapse=
4309 setting which takes a boolean argument. It defaults to on,
4310 exposing behaviour unchanged to previous releases. If set to
4311 off, timer units are unloaded after they elapsed if they
4312 cannot elapse again. This is particularly useful for
4313 transient timer units, which shall not stay around longer
4314 than until they first elapse.
4316 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
4317 default now (the kernel default is 16). This is beneficial
4318 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
4319 allows substantially larger numbers of queued
4320 datagrams. This should increase the capability of systemd to
4321 parallelize boot-up, as logging and sd_notify() are unlikely
4322 to stall execution anymore. If you need to change the value
4323 from the new defaults, use the usual sysctl.d/ snippets.
4325 * The compression framing format used by the journal or
4326 coredump processing has changed to be in line with what the
4327 official LZ4 tools generate. LZ4 compression support in
4328 systemd was considered unsupported previously, as the format
4329 was not compatible with the normal tools. With this release
4330 this has changed now, and it is hence safe for downstream
4331 distributions to turn it on. While not compressing as well
4332 as the XZ, LZ4 is substantially faster, which makes
4333 it a good default choice for the compression logic in the
4334 journal and in coredump handling.
4336 * Any reference to /etc/mtab has been dropped from
4337 systemd. The file has been obsolete since a while, but
4338 systemd refused to work on systems where it was incorrectly
4339 set up (it should be a symlink or non-existent). Please make
4340 sure to update to util-linux 2.27.1 or newer in conjunction
4341 with this systemd release, which also drops any reference to
4342 /etc/mtab. If you maintain a distribution make sure that no
4343 software you package still references it, as this is a
4344 likely source of bugs. There's also a glibc bug pending,
4345 asking for removal of any reference to this obsolete file:
4347 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
4349 Note that only util-linux versions built with
4350 --enable-libmount-force-mountinfo are supported.
4352 * Support for the ".snapshot" unit type has been removed. This
4353 feature turned out to be little useful and little used, and
4354 has now been removed from the core and from systemctl.
4356 * The dependency types RequiresOverridable= and
4357 RequisiteOverridable= have been removed from systemd. They
4358 have been used only very sparingly to our knowledge and
4359 other options that provide a similar effect (such as
4360 systemctl --mode=ignore-dependencies) are much more useful
4361 and commonly used. Moreover, they were only half-way
4362 implemented as the option to control behaviour regarding
4363 these dependencies was never added to systemctl. By removing
4364 these dependency types the execution engine becomes a bit
4365 simpler. Unit files that use these dependencies should be
4366 changed to use the non-Overridable dependency types
4367 instead. In fact, when parsing unit files with these
4368 options, that's what systemd will automatically convert them
4369 too, but it will also warn, asking users to fix the unit
4370 files accordingly. Removal of these dependency types should
4371 only affect a negligible number of unit files in the wild.
4373 * Behaviour of networkd's IPForward= option changed
4374 (again). It will no longer maintain a per-interface setting,
4375 but propagate one way from interfaces where this is enabled
4376 to the global kernel setting. The global setting will be
4377 enabled when requested by a network that is set up, but
4378 never be disabled again. This change was made to make sure
4379 IPv4 and IPv6 behaviour regarding packet forwarding is
4380 similar (as the Linux IPv6 stack does not support
4381 per-interface control of this setting) and to minimize
4384 * In unit files the behaviour of %u, %U, %h, %s has
4385 changed. These specifiers will now unconditionally resolve
4386 to the various user database fields of the user that the
4387 systemd instance is running as, instead of the user
4388 configured in the specific unit via User=. Note that this
4389 effectively doesn't change much, as resolving of these
4390 specifiers was already turned off in the --system instance
4391 of systemd, as we cannot do NSS lookups from PID 1. In the
4392 --user instance of systemd these specifiers where correctly
4393 resolved, but hardly made any sense, since the user instance
4394 lacks privileges to do user switches anyway, and User= is
4395 hence useless. Moreover, even in the --user instance of
4396 systemd behaviour was awkward as it would only take settings
4397 from User= assignment placed before the specifier into
4398 account. In order to unify and simplify the logic around
4399 this the specifiers will now always resolve to the
4400 credentials of the user invoking the manager (which in case
4401 of PID 1 is the root user).
4403 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
4404 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
4405 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
4406 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
4407 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
4408 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
4409 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
4410 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
4411 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
4412 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
4413 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
4414 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
4415 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
4416 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
4419 — Berlin, 2015-11-18
4423 * systemd now depends on util-linux v2.27. More specifically,
4424 the newly added mount monitor feature in libmount now
4425 replaces systemd's former own implementation.
4427 * libmount mandates /etc/mtab not to be regular file, and
4428 systemd now enforces this condition at early boot.
4429 /etc/mtab has been deprecated and warned about for a very
4430 long time, so systems running systemd should already have
4431 stopped having this file around as anything else than a
4432 symlink to /proc/self/mounts.
4434 * Support for the "pids" cgroup controller has been added. It
4435 allows accounting the number of tasks in a cgroup and
4436 enforcing limits on it. This adds two new setting
4437 TasksAccounting= and TasksMax= to each unit, as well as a
4438 global option DefaultTasksAccounting=.
4440 * Support for the "net_cls" cgroup controller has been added.
4441 It allows assigning a net class ID to each task in the
4442 cgroup, which can then be used in firewall rules and traffic
4443 shaping configurations. Note that the kernel netfilter net
4444 class code does not currently work reliably for ingress
4445 packets on unestablished sockets.
4447 This adds a new config directive called NetClass= to CGroup
4448 enabled units. Allowed values are positive numbers for fixed
4449 assignments and "auto" for picking a free value
4452 * 'systemctl is-system-running' now returns 'offline' if the
4453 system is not booted with systemd. This command can now be
4454 used as a substitute for 'systemd-notify --booted'.
4456 * Watchdog timeouts have been increased to 3 minutes for all
4457 in-tree service files. Apparently, disk IO issues are more
4458 frequent than we hoped, and user reported >1 minute waiting
4461 * 'machine-id-commit' functionality has been merged into
4462 'machine-id-setup --commit'. The separate binary has been
4465 * The WorkingDirectory= directive in unit files may now be set
4466 to the special value '~'. In this case, the working
4467 directory is set to the home directory of the user
4468 configured in User=.
4470 * "machinectl shell" will now open the shell in the home
4471 directory of the selected user by default.
4473 * The CrashChVT= configuration file setting is renamed to
4474 CrashChangeVT=, following our usual logic of not
4475 abbreviating unnecessarily. The old directive is still
4476 supported for compat reasons. Also, this directive now takes
4477 an integer value between 1 and 63, or a boolean value. The
4478 formerly supported '-1' value for disabling stays around for
4481 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
4482 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
4483 RootDirectory= properties can now be set for transient
4486 * The systemd-analyze tool gained a new "set-log-target" verb
4487 to change the logging target the system manager logs to
4488 dynamically during runtime. This is similar to how
4489 "systemd-analyze set-log-level" already changes the log
4492 * In nspawn /sys is now mounted as tmpfs, with only a selected
4493 set of subdirectories mounted in from the real sysfs. This
4494 enhances security slightly, and is useful for ensuring user
4495 namespaces work correctly.
4497 * Support for USB FunctionFS activation has been added. This
4498 allows implementation of USB gadget services that are
4499 activated as soon as they are requested, so that they don't
4500 have to run continuously, similar to classic socket
4503 * The "systemctl exit" command now optionally takes an
4504 additional parameter that sets the exit code to return from
4505 the systemd manager when exiting. This is only relevant when
4506 running the systemd user instance, or when running the
4507 system instance in a container.
4509 * sd-bus gained the new API calls sd_bus_path_encode_many()
4510 and sd_bus_path_decode_many() that allow easy encoding and
4511 decoding of multiple identifier strings inside a D-Bus
4512 object path. Another new call sd_bus_default_flush_close()
4513 has been added to flush and close per-thread default
4516 * systemd-cgtop gained support for a -M/--machine= switch to
4517 show the control groups within a certain container only.
4519 * "systemctl kill" gained support for an optional --fail
4520 switch. If specified the requested operation will fail of no
4521 processes have been killed, because the unit had no
4522 processes attached, or similar.
4524 * A new systemd.crash_reboot=1 kernel command line option has
4525 been added that triggers a reboot after crashing. This can
4526 also be set through CrashReboot= in systemd.conf.
4528 * The RuntimeDirectory= setting now understands unit
4529 specifiers like %i or %f.
4531 * A new (still internal) library API sd-ipv4acd has been added,
4532 that implements address conflict detection for IPv4. It's
4533 based on code from sd-ipv4ll, and will be useful for
4534 detecting DHCP address conflicts.
4536 * File descriptors passed during socket activation may now be
4537 named. A new API sd_listen_fds_with_names() is added to
4538 access the names. The default names may be overridden,
4539 either in the .socket file using the FileDescriptorName=
4540 parameter, or by passing FDNAME= when storing the file
4541 descriptors using sd_notify().
4543 * systemd-networkd gained support for:
4545 - Setting the IPv6 Router Advertisement settings via
4546 IPv6AcceptRouterAdvertisements= in .network files.
4548 - Configuring the HelloTimeSec=, MaxAgeSec= and
4549 ForwardDelaySec= bridge parameters in .netdev files.
4551 - Configuring PreferredSource= for static routes in
4554 * The "ask-password" framework used to query for LUKS harddisk
4555 passwords or SSL passwords during boot gained support for
4556 caching passwords in the kernel keyring, if it is
4557 available. This makes sure that the user only has to type in
4558 a passphrase once if there are multiple objects to unlock
4559 with the same one. Previously, such password caching was
4560 available only when Plymouth was used; this moves the
4561 caching logic into the systemd codebase itself. The
4562 "systemd-ask-password" utility gained a new --keyname=
4563 switch to control which kernel keyring key to use for
4564 caching a password in. This functionality is also useful for
4565 enabling display managers such as gdm to automatically
4566 unlock the user's GNOME keyring if its passphrase, the
4567 user's password and the harddisk password are the same, if
4568 gdm-autologin is used.
4570 * When downloading tar or raw images using "machinectl
4571 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
4572 file is now also downloaded, if it is available and stored
4573 next to the image file.
4575 * Units of type ".socket" gained a new boolean setting
4576 Writable= which is only useful in conjunction with
4577 ListenSpecial=. If true, enables opening the specified
4578 special file in O_RDWR mode rather than O_RDONLY mode.
4580 * systemd-rfkill has been reworked to become a singleton
4581 service that is activated through /dev/rfkill on each rfkill
4582 state change and saves the settings to disk. This way,
4583 systemd-rfkill is now compatible with devices that exist
4584 only intermittendly, and even restores state if the previous
4585 system shutdown was abrupt rather than clean.
4587 * The journal daemon gained support for vacuuming old journal
4588 files controlled by the number of files that shall remain,
4589 in addition to the already existing control by size and by
4590 date. This is useful as journal interleaving performance
4591 degrades with too many separate journal files, and allows
4592 putting an effective limit on them. The new setting defaults
4593 to 100, but this may be changed by setting SystemMaxFiles=
4594 and RuntimeMaxFiles= in journald.conf. Also, the
4595 "journalctl" tool gained the new --vacuum-files= switch to
4596 manually vacuum journal files to leave only the specified
4597 number of files in place.
4599 * udev will now create /dev/disk/by-path links for ATA devices
4600 on kernels where that is supported.
4602 * Galician, Serbian, Turkish and Korean translations were added.
4604 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
4605 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
4606 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
4607 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
4608 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
4609 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
4610 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
4611 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
4612 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
4613 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
4614 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
4615 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
4616 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
4617 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
4618 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
4619 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
4620 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
4621 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
4623 — Berlin, 2015-10-07
4627 * The DHCP implementation of systemd-networkd gained a set of
4630 - The DHCP server now supports emitting DNS and NTP
4631 information. It may be enabled and configured via
4632 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
4633 and NTP information is enabled, but no servers are
4634 configured, the corresponding uplink information (if there
4635 is any) is propagated.
4637 - Server and client now support transmission and reception
4638 of timezone information. It can be configured via the
4639 newly introduced network options UseTimezone=,
4640 EmitTimezone=, and Timezone=. Transmission of timezone
4641 information is enabled between host and containers by
4642 default now: the container will change its local timezone
4643 to what the host has set.
4645 - Lease timeouts can now be configured via
4646 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
4648 - The DHCP server improved on the stability of
4649 leases. Clients are more likely to get the same lease
4650 information back, even if the server loses state.
4652 - The DHCP server supports two new configuration options to
4653 control the lease address pool metrics, PoolOffset= and
4656 * The encapsulation limit of tunnels in systemd-networkd may
4657 now be configured via 'EncapsulationLimit='. It allows
4658 modifying the maximum additional levels of encapsulation
4659 that are permitted to be prepended to a packet.
4661 * systemd now supports the concept of user buses replacing
4662 session buses, if used with dbus-1.10 (and enabled via dbus
4663 --enable-user-session). It previously only supported this on
4664 kdbus-enabled systems, and this release expands this to
4665 'dbus-daemon' systems.
4667 * systemd-networkd now supports predictable interface names
4670 * systemd now optionally supports the new Linux kernel
4671 "unified" control group hierarchy. If enabled via the kernel
4672 command-line option 'systemd.unified_cgroup_hierarchy=1',
4673 systemd will try to mount the unified cgroup hierarchy
4674 directly on /sys/fs/cgroup. If not enabled, or not
4675 available, systemd will fall back to the legacy cgroup
4676 hierarchy setup, as before. Host system and containers can
4677 mix and match legacy and unified hierarchies as they
4678 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
4679 environment variable to individually select the hierarchy to
4680 use for executed containers. By default, nspawn will use the
4681 unified hierarchy for the containers if the host uses the
4682 unified hierarchy, and the legacy hierarchy otherwise.
4683 Please note that at this point the unified hierarchy is an
4684 experimental kernel feature and is likely to change in one
4685 of the next kernel releases. Therefore, it should not be
4686 enabled by default in downstream distributions yet. The
4687 minimum required kernel version for the unified hierarchy to
4688 work is 4.2. Note that when the unified hierarchy is used
4689 for the first time delegated access to controllers is
4690 safe. Because of this systemd-nspawn containers will get
4691 access to controllers now, as will systemd user
4692 sessions. This means containers and user sessions may now
4693 manage their own resources, partitioning up what the system
4696 * A new special scope unit "init.scope" has been introduced
4697 that encapsulates PID 1 of the system. It may be used to
4698 determine resource usage and enforce resource limits on PID
4699 1 itself. PID 1 hence moved out of the root of the control
4702 * The cgtop tool gained support for filtering out kernel
4703 threads when counting tasks in a control group. Also, the
4704 count of processes is now recursively summed up by
4705 default. Two options -k and --recursive= have been added to
4706 revert to old behaviour. The tool has also been updated to
4707 work correctly in containers now.
4709 * systemd-nspawn's --bind= and --bind-ro= options have been
4710 extended to allow creation of non-recursive bind mounts.
4712 * libsystemd gained two new calls sd_pid_get_cgroup() and
4713 sd_peer_get_cgroup() which return the control group path of
4714 a process or peer of a connected AF_UNIX socket. This
4715 function call is particularly useful when implementing
4716 delegated subtrees support in the control group hierarchy.
4718 * The "sd-event" event loop API of libsystemd now supports
4719 correct dequeuing of real-time signals, without losing
4722 * When systemd requests a polkit decision when managing units it
4723 will now add additional fields to the request, including unit
4724 name and desired operation. This enables more powerful polkit
4725 policies, that make decisions depending on these parameters.
4727 * nspawn learnt support for .nspawn settings files, that may
4728 accompany the image files or directories of containers, and
4729 may contain additional settings for the container. This is
4730 an alternative to configuring container parameters via the
4731 nspawn command line.
4733 Contributions from: Cristian Rodríguez, Daniel Mack, David
4734 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
4735 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
4736 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
4737 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
4738 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
4739 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
4740 Andersen, Tom Gundersen, Torstein Husebø
4742 — Berlin, 2015-09-08
4746 * machinectl gained a new verb 'shell' which opens a fresh
4747 shell on the target container or the host. It is similar to
4748 the existing 'login' command of machinectl, but spawns the
4749 shell directly without prompting for username or
4750 password. The pseudo machine '.host' now refers to the local
4751 host and is used by default. Hence, 'machinectl shell' can
4752 be used as replacement for 'su -' which spawns a session as
4753 a fresh systemd unit in a way that is fully isolated from
4754 the originating session.
4756 * systemd-networkd learned to cope with private-zone DHCP
4757 options and allows other programs to query the values.
4759 * SELinux access control when enabling/disabling units is no
4760 longer enforced with this release. The previous implementation
4761 was incorrect, and a new corrected implementation is not yet
4762 available. As unit file operations are still protected via
4763 polkit and D-Bus policy this is not a security problem. Yet,
4764 distributions which care about optimal SELinux support should
4765 probably not stabilize on this release.
4767 * sd-bus gained support for matches of type "arg0has=", that
4768 test for membership of strings in string arrays sent in bus
4771 * systemd-resolved now dumps the contents of its DNS and LLMNR
4772 caches to the logs on reception of the SIGUSR1 signal. This
4773 is useful to debug DNS behaviour.
4775 * The coredumpctl tool gained a new --directory= option to
4776 operate on journal files in a specific directory.
4778 * "systemctl reboot" and related commands gained a new
4779 "--message=" option which may be used to set a free-text
4780 wall message when shutting down or rebooting the
4781 system. This message is also logged, which is useful for
4782 figuring out the reason for a reboot or shutdown a
4785 * The "systemd-resolve-host" tool's -i switch now takes
4786 network interface numbers as alternative to interface names.
4788 * A new unit file setting for services has been introduced:
4789 UtmpMode= allows configuration of how precisely systemd
4790 handles utmp and wtmp entries for the service if this is
4791 enabled. This allows writing services that appear similar to
4792 user sessions in the output of the "w", "who", "last" and
4795 * systemd-resolved will now locally synthesize DNS resource
4796 records for the "localhost" and "gateway" domains as well as
4797 the local hostname. This should ensure that clients querying
4798 RRs via resolved will get similar results as those going via
4799 NSS, if nss-myhostname is enabled.
4801 Contributions from: Alastair Hughes, Alex Crawford, Daniel
4802 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
4803 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
4804 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
4805 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
4806 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
4807 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
4808 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
4809 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
4810 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
4811 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
4812 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
4814 — Berlin, 2015-08-27
4818 * The systemd-efi-boot-generator functionality was merged into
4819 systemd-gpt-auto-generator.
4821 * systemd-networkd now supports Group Policy for vxlan
4822 devices. It can be enabled via the new boolean configuration
4823 option called 'GroupPolicyExtension='.
4825 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
4826 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
4827 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
4829 — Berlin, 2015-07-31
4833 * The python-systemd code has been removed from the systemd repository.
4834 A new repository has been created which accommodates the code from
4835 now on, and we kindly ask distributions to create a separate package
4836 for this: https://github.com/systemd/python-systemd
4838 * The systemd daemon will now reload its main configuration
4839 (/etc/systemd/system.conf) on daemon-reload.
4841 * sd-dhcp now exposes vendor specific extensions via
4842 sd_dhcp_lease_get_vendor_specific().
4844 * systemd-networkd gained a number of new configuration options.
4846 - A new boolean configuration option for TAP devices called
4847 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
4848 device, thus allowing to send and receive GSO packets.
4850 - A new tunnel configuration option called 'CopyDSCP='.
4851 If enabled, the DSCP field of ip6 tunnels is copied into the
4852 decapsulated packet.
4854 - A set of boolean bridge configuration options were added.
4855 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
4856 and 'UnicastFlood=' are now parsed by networkd and applied to the
4857 respective bridge link device via the respective IFLA_BRPORT_*
4860 - A new string configuration option to override the hostname sent
4861 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
4862 is true, networkd will use the configured hostname instead of the
4863 system hostname when sending DHCP requests.
4865 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
4866 networkd will configure the IPv6 flow-label of the tunnel device
4867 according to RFC2460.
4869 - The 'macvtap' virtual network devices are now supported, similar to
4870 the already supported 'macvlan' devices.
4872 * systemd-resolved now implements RFC5452 to improve resilience against
4873 cache poisoning. Additionally, source port randomization is enabled
4874 by default to further protect against DNS spoofing attacks.
4876 * nss-mymachines now supports translating UIDs and GIDs of running
4877 containers with user-namespaces enabled. If a container 'foo'
4878 translates a host uid 'UID' to the container uid 'TUID', then
4879 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
4880 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
4881 mapped as 'vg-foo-TGID'.
4883 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
4884 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
4885 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
4886 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
4887 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
4888 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
4889 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
4890 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
4891 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
4892 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
4894 — Berlin, 2015-07-29
4898 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
4899 There are no known issues with current sysfs, and udev does not need
4900 or should be used to work around such bugs.
4902 * udev does no longer enable USB HID power management. Several reports
4903 indicate, that some devices cannot handle that setting.
4905 * The udev accelerometer helper was removed. The functionality
4906 is now fully included in iio-sensor-proxy. But this means,
4907 older iio-sensor-proxy versions will no longer provide
4908 accelerometer/orientation data with this systemd version.
4909 Please upgrade iio-sensor-proxy to version 1.0.
4911 * networkd gained a new configuration option IPv6PrivacyExtensions=
4912 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
4913 for Stateless Address") on selected networks.
4915 * For the sake of fewer build-time dependencies and less code in the
4916 main repository, the python bindings are about to be removed in the
4917 next release. A new repository has been created which accommodates
4918 the code from now on, and we kindly ask distributions to create a
4919 separate package for this. The removal will take place in v223.
4921 https://github.com/systemd/python-systemd
4923 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
4924 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
4925 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
4926 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
4927 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
4928 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
4929 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
4930 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
4931 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
4932 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
4934 — Berlin, 2015-07-07
4938 * The sd-bus.h and sd-event.h APIs have now been declared
4939 stable and have been added to the official interface of
4940 libsystemd.so. sd-bus implements an alternative D-Bus client
4941 library, that is relatively easy to use, very efficient and
4942 supports both classic D-Bus as well as kdbus as transport
4943 backend. sd-event is a generic event loop abstraction that
4944 is built around Linux epoll, but adds features such as event
4945 prioritization or efficient timer handling. Both APIs are good
4946 choices for C programs looking for a bus and/or event loop
4947 implementation that is minimal and does not have to be
4948 portable to other kernels.
4950 * kdbus support is no longer compile-time optional. It is now
4951 always built-in. However, it can still be disabled at
4952 runtime using the kdbus=0 kernel command line setting, and
4953 that setting may be changed to default to off, by specifying
4954 --disable-kdbus at build-time. Note though that the kernel
4955 command line setting has no effect if the kdbus.ko kernel
4956 module is not installed, in which case kdbus is (obviously)
4957 also disabled. We encourage all downstream distributions to
4958 begin testing kdbus by adding it to the kernel images in the
4959 development distributions, and leaving kdbus support in
4962 * The minimal required util-linux version has been bumped to
4965 * Support for chkconfig (--enable-chkconfig) was removed in
4966 favor of calling an abstraction tool
4967 /lib/systemd/systemd-sysv-install. This needs to be
4968 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
4969 in README for details.
4971 * If there's a systemd unit and a SysV init script for the
4972 same service name, and the user executes "systemctl enable"
4973 for it (or a related call), then this will now enable both
4974 (or execute the related operation on both), not just the
4977 * The libudev API documentation has been converted from gtkdoc
4980 * gudev has been removed from the systemd tree, it is now an
4983 * The systemd-cgtop tool learnt a new --raw switch to generate
4984 "raw" (machine parsable) output.
4986 * networkd's IPForwarding= .network file setting learnt the
4987 new setting "kernel", which ensures that networkd does not
4988 change the IP forwarding sysctl from the default kernel
4991 * The systemd-logind bus API now exposes a new boolean
4992 property "Docked" that reports whether logind considers the
4993 system "docked", i.e. connected to a docking station or not.
4995 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
4996 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
4997 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
4998 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
4999 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
5000 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
5001 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
5002 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
5003 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
5004 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
5005 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
5006 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
5007 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
5008 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
5009 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
5010 Fink, Zbigniew Jędrzejewski-Szmek
5012 — Berlin, 2015-06-19
5016 * The gudev library has been extracted into a separate repository
5017 available at: https://git.gnome.org/browse/libgudev/
5018 It is now managed as part of the Gnome project. Distributions
5019 are recommended to pass --disable-gudev to systemd and use
5020 gudev from the Gnome project instead. gudev is still included
5021 in systemd, for now. It will be removed soon, though. Please
5022 also see the announcement-thread on systemd-devel:
5023 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
5025 * systemd now exposes a CPUUsageNSec= property for each
5026 service unit on the bus, that contains the overall consumed
5027 CPU time of a service (the sum of what each process of the
5028 service consumed). This value is only available if
5029 CPUAccounting= is turned on for a service, and is then shown
5030 in the "systemctl status" output.
5032 * Support for configuring alternative mappings of the old SysV
5033 runlevels to systemd targets has been removed. They are now
5034 hardcoded in a way that runlevels 2, 3, 4 all map to
5035 multi-user.target and 5 to graphical.target (which
5036 previously was already the default behaviour).
5038 * The auto-mounter logic gained support for mount point
5039 expiry, using a new TimeoutIdleSec= setting in .automount
5040 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
5042 * The EFI System Partition (ESP) as mounted to /boot by
5043 systemd-efi-boot-generator will now be unmounted
5044 automatically after 2 minutes of not being used. This should
5045 minimize the risk of ESP corruptions.
5047 * New /etc/fstab options x-systemd.requires= and
5048 x-systemd.requires-mounts-for= are now supported to express
5049 additional dependencies for mounts. This is useful for
5050 journalling file systems that support external journal
5051 devices or overlay file systems that require underlying file
5052 systems to be mounted.
5054 * systemd does not support direct live-upgrades (via systemctl
5055 daemon-reexec) from versions older than v44 anymore. As no
5056 distribution we are aware of shipped such old versions in a
5057 stable release this should not be problematic.
5059 * When systemd forks off a new per-connection service instance
5060 it will now set the $REMOTE_ADDR environment variable to the
5061 remote IP address, and $REMOTE_PORT environment variable to
5062 the remote IP port. This behaviour is similar to the
5063 corresponding environment variables defined by CGI.
5065 * systemd-networkd gained support for uplink failure
5066 detection. The BindCarrier= option allows binding interface
5067 configuration dynamically to the link sense of other
5068 interfaces. This is useful to achieve behaviour like in
5071 * systemd-networkd gained support for configuring the DHCP
5072 client identifier to use when requesting leases.
5074 * systemd-networkd now has a per-network UseNTP= option to
5075 configure whether NTP server information acquired via DHCP
5076 is passed on to services like systemd-timesyncd.
5078 * systemd-networkd gained support for vti6 tunnels.
5080 * Note that systemd-networkd manages the sysctl variable
5081 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
5082 it is configured for since v219. The variable controls IP
5083 forwarding, and is a per-interface alternative to the global
5084 /proc/sys/net/ipv[46]/ip_forward. This setting is
5085 configurable in the IPForward= option, which defaults to
5086 "no". This means if networkd is used for an interface it is
5087 no longer sufficient to set the global sysctl option to turn
5088 on IP forwarding! Instead, the .network file option
5089 IPForward= needs to be turned on! Note that the
5090 implementation of this behaviour was broken in v219 and has
5093 * Many bonding and vxlan options are now configurable in
5096 * systemd-nspawn gained a new --property= setting to set unit
5097 properties for the container scope. This is useful for
5098 setting resource parameters (e.g. "CPUShares=500") on
5099 containers started from the command line.
5101 * systemd-nspawn gained a new --private-users= switch to make
5102 use of user namespacing available on recent Linux kernels.
5104 * systemd-nspawn may now be called as part of a shell pipeline
5105 in which case the pipes used for stdin and stdout are passed
5106 directly to the process invoked in the container, without
5107 indirection via a pseudo tty.
5109 * systemd-nspawn gained a new switch to control the UNIX
5110 signal to use when killing the init process of the container
5113 * systemd-nspawn gained a new --overlay= switch for mounting
5114 overlay file systems into the container using the new kernel
5117 * When a container image is imported via systemd-importd and
5118 the host file system is not btrfs, a loopback block device
5119 file is created in /var/lib/machines.raw with a btrfs file
5120 system inside. It is then mounted to /var/lib/machines to
5121 enable btrfs features for container management. The loopback
5122 file and btrfs file system is grown as needed when container
5123 images are imported via systemd-importd.
5125 * systemd-machined/systemd-importd gained support for btrfs
5126 quota, to enforce container disk space limits on disk. This
5127 is exposed in "machinectl set-limit".
5129 * systemd-importd now can import containers from local .tar,
5130 .raw and .qcow2 images, and export them to .tar and .raw. It
5131 can also import dkr v2 images now from the network (on top
5134 * systemd-importd gained support for verifying downloaded
5135 images with gpg2 (previously only gpg1 was supported).
5137 * systemd-machined, systemd-logind, systemd: most bus calls are
5138 now accessible to unprivileged processes via polkit. Also,
5139 systemd-logind will now allow users to kill their own sessions
5140 without further privileges or authorization.
5142 * systemd-shutdownd has been removed. This service was
5143 previously responsible for implementing scheduled shutdowns
5144 as exposed in /usr/bin/shutdown's time parameter. This
5145 functionality has now been moved into systemd-logind and is
5146 accessible via a bus interface.
5148 * "systemctl reboot" gained a new switch --firmware-setup that
5149 can be used to reboot into the EFI firmware setup, if that
5150 is available. systemd-logind now exposes an API on the bus
5151 to trigger such reboots, in case graphical desktop UIs want
5152 to cover this functionality.
5154 * "systemctl enable", "systemctl disable" and "systemctl mask"
5155 now support a new "--now" switch. If specified the units
5156 that are enabled will also be started, and the ones
5157 disabled/masked also stopped.
5159 * The Gummiboot EFI boot loader tool has been merged into
5160 systemd, and renamed to "systemd-boot". The bootctl tool has been
5161 updated to support systemd-boot.
5163 * An EFI kernel stub has been added that may be used to create
5164 kernel EFI binaries that contain not only the actual kernel,
5165 but also an initrd, boot splash, command line and OS release
5166 information. This combined binary can then be signed as a
5167 single image, so that the firmware can verify it all in one
5168 step. systemd-boot has special support for EFI binaries created
5169 like this and can extract OS release information from them
5170 and show them in the boot menu. This functionality is useful
5171 to implement cryptographically verified boot schemes.
5173 * Optional support has been added to systemd-fsck to pass
5174 fsck's progress report to an AF_UNIX socket in the file
5177 * udev will no longer create device symlinks for all block
5178 devices by default. A blacklist for excluding special block
5179 devices from this logic has been turned into a whitelist
5180 that requires picking block devices explicitly that require
5183 * A new (currently still internal) API sd-device.h has been
5184 added to libsystemd. This modernized API is supposed to
5185 replace libudev eventually. In fact, already much of libudev
5186 is now just a wrapper around sd-device.h.
5188 * A new hwdb database for storing metadata about pointing
5189 stick devices has been added.
5191 * systemd-tmpfiles gained support for setting file attributes
5192 similar to the "chattr" tool with new 'h' and 'H' lines.
5194 * systemd-journald will no longer unconditionally set the
5195 btrfs NOCOW flag on new journal files. This is instead done
5196 with tmpfiles snippet using the new 'h' line type. This
5197 allows easy disabling of this logic, by masking the
5198 journal-nocow.conf tmpfiles file.
5200 * systemd-journald will now translate audit message types to
5201 human readable identifiers when writing them to the
5202 journal. This should improve readability of audit messages.
5204 * The LUKS logic gained support for the offset= and skip=
5205 options in /etc/crypttab, as previously implemented by
5208 * /usr/lib/os-release gained a new optional field VARIANT= for
5209 distributions that support multiple variants (such as a
5210 desktop edition, a server edition, ...)
5212 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
5213 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
5214 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
5215 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
5216 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
5217 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
5218 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
5219 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
5220 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
5221 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
5222 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
5223 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
5224 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
5225 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
5226 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
5227 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
5228 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
5229 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
5230 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
5231 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
5232 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
5233 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
5234 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
5235 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
5236 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
5237 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
5238 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
5240 — Berlin, 2015-05-22
5244 * Introduce a new API "sd-hwdb.h" for querying the hardware
5245 metadata database. With this minimal interface one can query
5246 and enumerate the udev hwdb, decoupled from the old libudev
5247 library. libudev's interface for this is now only a wrapper
5248 around sd-hwdb. A new tool systemd-hwdb has been added to
5249 interface with and update the database.
5251 * When any of systemd's tools copies files (for example due to
5252 tmpfiles' C lines) a btrfs reflink will attempted first,
5253 before bytewise copying is done.
5255 * systemd-nspawn gained a new --ephemeral switch. When
5256 specified a btrfs snapshot is taken of the container's root
5257 directory, and immediately removed when the container
5258 terminates again. Thus, a container can be started whose
5259 changes never alter the container's root directory, and are
5260 lost on container termination. This switch can also be used
5261 for starting a container off the root file system of the
5262 host without affecting the host OS. This switch is only
5263 available on btrfs file systems.
5265 * systemd-nspawn gained a new --template= switch. It takes the
5266 path to a container tree to use as template for the tree
5267 specified via --directory=, should that directory be
5268 missing. This allows instantiating containers dynamically,
5269 on first run. This switch is only available on btrfs file
5272 * When a .mount unit refers to a mount point on which multiple
5273 mounts are stacked, and the .mount unit is stopped all of
5274 the stacked mount points will now be unmounted until no
5275 mount point remains.
5277 * systemd now has an explicit notion of supported and
5278 unsupported unit types. Jobs enqueued for unsupported unit
5279 types will now fail with an "unsupported" error code. More
5280 specifically .swap, .automount and .device units are not
5281 supported in containers, .busname units are not supported on
5282 non-kdbus systems. .swap and .automount are also not
5283 supported if their respective kernel compile time options
5286 * machinectl gained support for two new "copy-from" and
5287 "copy-to" commands for copying files from a running
5288 container to the host or vice versa.
5290 * machinectl gained support for a new "bind" command to bind
5291 mount host directories into local containers. This is
5292 currently only supported for nspawn containers.
5294 * networkd gained support for configuring bridge forwarding
5295 database entries (fdb) from .network files.
5297 * A new tiny daemon "systemd-importd" has been added that can
5298 download container images in tar, raw, qcow2 or dkr formats,
5299 and make them available locally in /var/lib/machines, so
5300 that they can run as nspawn containers. The daemon can GPG
5301 verify the downloads (not supported for dkr, since it has no
5302 provisions for verifying downloads). It will transparently
5303 decompress bz2, xz, gzip compressed downloads if necessary,
5304 and restore sparse files on disk. The daemon uses privilege
5305 separation to ensure the actual download logic runs with
5306 fewer privileges than the daemon itself. machinectl has
5307 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
5308 make the functionality of importd available to the
5309 user. With this in place the Fedora and Ubuntu "Cloud"
5310 images can be downloaded and booted as containers unmodified
5311 (the Fedora images lack the appropriate GPG signature files
5312 currently, so they cannot be verified, but this will change
5313 soon, hopefully). Note that downloading images is currently
5314 only fully supported on btrfs.
5316 * machinectl is now able to list container images found in
5317 /var/lib/machines, along with some metadata about sizes of
5318 disk and similar. If the directory is located on btrfs and
5319 quota is enabled, this includes quota display. A new command
5320 "image-status" has been added that shows additional
5321 information about images.
5323 * machinectl is now able to clone container images
5324 efficiently, if the underlying file system (btrfs) supports
5325 it, with the new "machinectl clone" command. It also
5326 gained commands for renaming and removing images, as well as
5327 marking them read-only or read-write (supported also on
5328 legacy file systems).
5330 * networkd gained support for collecting LLDP network
5331 announcements, from hardware that supports this. This is
5332 shown in networkctl output.
5334 * systemd-run gained support for a new -t (--pty) switch for
5335 invoking a binary on a pty whose input and output is
5336 connected to the invoking terminal. This allows executing
5337 processes as system services while interactively
5338 communicating with them via the terminal. Most interestingly
5339 this is supported across container boundaries. Invoking
5340 "systemd-run -t /bin/bash" is an alternative to running a
5341 full login session, the difference being that the former
5342 will not register a session, nor go through the PAM session
5345 * tmpfiles gained support for a new "v" line type for creating
5346 btrfs subvolumes. If the underlying file system is a legacy
5347 file system, this automatically degrades to creating a
5348 normal directory. Among others /var/lib/machines is now
5349 created like this at boot, should it be missing.
5351 * The directory /var/lib/containers/ has been deprecated and
5352 been replaced by /var/lib/machines. The term "machines" has
5353 been used in the systemd context as generic term for both
5354 VMs and containers, and hence appears more appropriate for
5355 this, as the directory can also contain raw images bootable
5358 * systemd-nspawn when invoked with -M but without --directory=
5359 or --image= is now capable of searching for the container
5360 root directory, subvolume or disk image automatically, in
5361 /var/lib/machines. systemd-nspawn@.service has been updated
5362 to make use of this, thus allowing it to be used for raw
5365 * A new machines.target unit has been introduced that is
5366 supposed to group all containers/VMs invoked as services on
5367 the system. systemd-nspawn@.service has been updated to
5368 integrate with that.
5370 * machinectl gained a new "start" command, for invoking a
5371 container as a service. "machinectl start foo" is mostly
5372 equivalent to "systemctl start systemd-nspawn@foo.service",
5373 but handles escaping in a nicer way.
5375 * systemd-nspawn will now mount most of the cgroupfs tree
5376 read-only into each container, with the exception of the
5377 container's own subtree in the name=systemd hierarchy.
5379 * journald now sets the special FS_NOCOW file flag for its
5380 journal files. This should improve performance on btrfs, by
5381 avoiding heavy fragmentation when journald's write-pattern
5382 is used on COW file systems. It degrades btrfs' data
5383 integrity guarantees for the files to the same levels as for
5384 ext3/ext4 however. This should be OK though as journald does
5385 its own data integrity checks and all its objects are
5386 checksummed on disk. Also, journald should handle btrfs disk
5387 full events a lot more gracefully now, by processing SIGBUS
5388 errors, and not relying on fallocate() anymore.
5390 * When journald detects that journal files it is writing to
5391 have been deleted it will immediately start new journal
5394 * systemd now provides a way to store file descriptors
5395 per-service in PID 1. This is useful for daemons to ensure
5396 that fds they require are not lost during a daemon
5397 restart. The fds are passed to the daemon on the next
5398 invocation in the same way socket activation fds are
5399 passed. This is now used by journald to ensure that the
5400 various sockets connected to all the system's stdout/stderr
5401 are not lost when journald is restarted. File descriptors
5402 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
5403 an extension to sd_notify(). Note that a limit is enforced
5404 on the number of fds a service can store in PID 1, and it
5405 defaults to 0, so that no fds may be stored, unless this is
5406 explicitly turned on.
5408 * The default TERM variable to use for units connected to a
5409 terminal, when no other value is explicitly is set is now
5410 vt220 rather than vt102. This should be fairly safe still,
5411 but allows PgUp/PgDn work.
5413 * The /etc/crypttab option header= as known from Debian is now
5416 * "loginctl user-status" and "loginctl session-status" will
5417 now show the last 10 lines of log messages of the
5418 user/session following the status output. Similar,
5419 "machinectl status" will show the last 10 log lines
5420 associated with a virtual machine or container
5421 service. (Note that this is usually not the log messages
5422 done in the VM/container itself, but simply what the
5423 container manager logs. For nspawn this includes all console
5426 * "loginctl session-status" without further argument will now
5427 show the status of the session of the caller. Similar,
5428 "lock-session", "unlock-session", "activate",
5429 "enable-linger", "disable-linger" may now be called without
5430 session/user parameter in which case they apply to the
5431 caller's session/user.
5433 * An X11 session scriptlet is now shipped that uploads
5434 $DISPLAY and $XAUTHORITY into the environment of the systemd
5435 --user daemon if a session begins. This should improve
5436 compatibility with X11 enabled applications run as systemd
5439 * Generators are now subject to masking via /etc and /run, the
5440 same way as unit files.
5442 * networkd .network files gained support for configuring
5443 per-link IPv4/IPv6 packet forwarding as well as IPv4
5444 masquerading. This is by default turned on for veth links to
5445 containers, as registered by systemd-nspawn. This means that
5446 nspawn containers run with --network-veth will now get
5447 automatic routed access to the host's networks without any
5448 further configuration or setup, as long as networkd runs on
5451 * systemd-nspawn gained the --port= (-p) switch to expose TCP
5452 or UDP posts of a container on the host. With this in place
5453 it is possible to run containers with private veth links
5454 (--network-veth), and have their functionality exposed on
5455 the host as if their services were running directly on the
5458 * systemd-nspawn's --network-veth switch now gained a short
5459 version "-n", since with the changes above it is now truly
5460 useful out-of-the-box. The systemd-nspawn@.service has been
5461 updated to make use of it too by default.
5463 * systemd-nspawn will now maintain a per-image R/W lock, to
5464 ensure that the same image is not started more than once
5465 writable. (It's OK to run an image multiple times
5466 simultaneously in read-only mode.)
5468 * systemd-nspawn's --image= option is now capable of
5469 dissecting and booting MBR and GPT disk images that contain
5470 only a single active Linux partition. Previously it
5471 supported only GPT disk images with proper GPT type
5472 IDs. This allows running cloud images from major
5473 distributions directly with systemd-nspawn, without
5476 * In addition to collecting mouse dpi data in the udev
5477 hardware database, there's now support for collecting angle
5478 information for mouse scroll wheels. The database is
5479 supposed to guarantee similar scrolling behavior on mice
5480 that it knows about. There's also support for collecting
5481 information about Touchpad types.
5483 * udev's input_id built-in will now also collect touch screen
5484 dimension data and attach it to probed devices.
5486 * /etc/os-release gained support for a Distribution Privacy
5489 * networkd gained support for creating "ipvlan", "gretap",
5490 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
5492 * systemd-tmpfiles gained support for "a" lines for setting
5495 * systemd-nspawn will now mount /tmp in the container to
5496 tmpfs, automatically.
5498 * systemd now exposes the memory.usage_in_bytes cgroup
5499 attribute and shows it for each service in the "systemctl
5500 status" output, if available.
5502 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
5503 immediate reboot is triggered. This useful if shutdown is
5504 hung and is unable to complete, to expedite the
5505 operation. Note that this kind of reboot will still unmount
5506 all file systems, and hence should not result in fsck being
5509 * A .device unit for an optical block device will now be
5510 considered active only when a medium is in the drive. Also,
5511 mount units are now bound to their backing devices thus
5512 triggering automatic unmounting when devices become
5513 unavailable. With this in place systemd will now
5514 automatically unmount left-over mounts when a CD-ROM is
5515 ejected or an USB stick is yanked from the system.
5517 * networkd-wait-online now has support for waiting for
5518 specific interfaces only (with globbing), and for giving up
5519 after a configurable timeout.
5521 * networkd now exits when idle. It will be automatically
5522 restarted as soon as interfaces show up, are removed or
5523 change state. networkd will stay around as long as there is
5524 at least one DHCP state machine or similar around, that keep
5527 * networkd may now configure IPv6 link-local addressing in
5528 addition to IPv4 link-local addressing.
5530 * The IPv6 "token" for use in SLAAC may now be configured for
5531 each .network interface in networkd.
5533 * Routes configured with networkd may now be assigned a scope
5536 * networkd's [Match] sections now support globbing and lists
5537 of multiple space-separated matches per item.
5539 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
5540 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
5541 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
5542 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
5543 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
5544 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
5545 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
5546 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
5547 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
5548 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
5549 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
5550 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
5551 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
5552 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
5553 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
5554 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
5555 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
5556 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
5557 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
5558 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
5559 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
5560 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
5561 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
5562 Hoffmann, Zbigniew Jędrzejewski-Szmek
5564 — Berlin, 2015-02-16
5568 * When querying unit file enablement status (for example via
5569 "systemctl is-enabled"), a new state "indirect" is now known
5570 which indicates that a unit might not be enabled itself, but
5571 another unit listed in its Also= setting might be.
5573 * Similar to the various existing ConditionXYZ= settings for
5574 units, there are now matching AssertXYZ= settings. While
5575 failing conditions cause a unit to be skipped, but its job
5576 to succeed, failing assertions declared like this will cause
5577 a unit start operation and its job to fail.
5579 * hostnamed now knows a new chassis type "embedded".
5581 * systemctl gained a new "edit" command. When used on a unit
5582 file, this allows extending unit files with .d/ drop-in
5583 configuration snippets or editing the full file (after
5584 copying it from /usr/lib to /etc). This will invoke the
5585 user's editor (as configured with $EDITOR), and reload the
5586 modified configuration after editing.
5588 * "systemctl status" now shows the suggested enablement state
5589 for a unit, as declared in the (usually vendor-supplied)
5590 system preset files.
5592 * nss-myhostname will now resolve the single-label host name
5593 "gateway" to the locally configured default IP routing
5594 gateways, ordered by their metrics. This assigns a stable
5595 name to the used gateways, regardless which ones are
5596 currently configured. Note that the name will only be
5597 resolved after all other name sources (if nss-myhostname is
5598 configured properly) and should hence not negatively impact
5599 systems that use the single-label host name "gateway" in
5602 * systemd-inhibit now allows filtering by mode when listing
5605 * Scope and service units gained a new "Delegate" boolean
5606 property, which, when set, allows processes running inside the
5607 unit to further partition resources. This is primarily
5608 useful for systemd user instances as well as container
5611 * journald will now pick up audit messages directly from
5612 the kernel, and log them like any other log message. The
5613 audit fields are split up and fully indexed. This means that
5614 journalctl in many ways is now a (nicer!) alternative to
5615 ausearch, the traditional audit client. Note that this
5616 implements only a minimal audit client. If you want the
5617 special audit modes like reboot-on-log-overflow, please use
5618 the traditional auditd instead, which can be used in
5619 parallel to journald.
5621 * The ConditionSecurity= unit file option now understands the
5622 special string "audit" to check whether auditing is
5625 * journalctl gained two new commands --vacuum-size= and
5626 --vacuum-time= to delete old journal files until the
5627 remaining ones take up no more than the specified size on disk,
5628 or are not older than the specified time.
5630 * A new, native PPPoE library has been added to sd-network,
5631 systemd's library of light-weight networking protocols. This
5632 library will be used in a future version of networkd to
5633 enable PPPoE communication without an external pppd daemon.
5635 * The busctl tool now understands a new "capture" verb that
5636 works similar to "monitor", but writes a packet capture
5637 trace to STDOUT that can be redirected to a file which is
5638 compatible with libcap's capture file format. This can then
5639 be loaded in Wireshark and similar tools to inspect bus
5642 * The busctl tool now understands a new "tree" verb that shows
5643 the object trees of a specific service on the bus, or of all
5646 * The busctl tool now understands a new "introspect" verb that
5647 shows all interfaces and members of objects on the bus,
5648 including their signature and values. This is particularly
5649 useful to get more information about bus objects shown by
5650 the new "busctl tree" command.
5652 * The busctl tool now understands new verbs "call",
5653 "set-property" and "get-property" for invoking bus method
5654 calls, setting and getting bus object properties in a
5657 * busctl gained a new --augment-creds= argument that controls
5658 whether the tool shall augment credential information it
5659 gets from the bus with data from /proc, in a possibly
5662 * nspawn's --link-journal= switch gained two new values
5663 "try-guest" and "try-host" that work like "guest" and
5664 "host", but do not fail if the host has no persistent
5665 journalling enabled. -j is now equivalent to
5666 --link-journal=try-guest.
5668 * macvlan network devices created by nspawn will now have
5669 stable MAC addresses.
5671 * A new SmackProcessLabel= unit setting has been added, which
5672 controls the SMACK security label processes forked off by
5673 the respective unit shall use.
5675 * If compiled with --enable-xkbcommon, systemd-localed will
5676 verify x11 keymap settings by compiling the given keymap. It
5677 will spew out warnings if the compilation fails. This
5678 requires libxkbcommon to be installed.
5680 * When a coredump is collected, a larger number of metadata
5681 fields is now collected and included in the journal records
5682 created for it. More specifically, control group membership,
5683 environment variables, memory maps, working directory,
5684 chroot directory, /proc/$PID/status, and a list of open file
5685 descriptors is now stored in the log entry.
5687 * The udev hwdb now contains DPI information for mice. For
5690 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
5692 * All systemd programs that read standalone configuration
5693 files in /etc now also support a corresponding series of
5694 .conf.d configuration directories in /etc/, /run/,
5695 /usr/local/lib/, /usr/lib/, and (if configured with
5696 --enable-split-usr) /lib/. In particular, the following
5697 configuration files now have corresponding configuration
5698 directories: system.conf user.conf, logind.conf,
5699 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
5700 resolved.conf, timesyncd.conf, journal-remote.conf, and
5701 journal-upload.conf. Note that distributions should use the
5702 configuration directories in /usr/lib/; the directories in
5703 /etc/ are reserved for the system administrator.
5705 * systemd-rfkill will no longer take the rfkill device name
5706 into account when storing rfkill state on disk, as the name
5707 might be dynamically assigned and not stable. Instead, the
5708 ID_PATH udev variable combined with the rfkill type (wlan,
5709 bluetooth, ...) is used.
5711 * A new service systemd-machine-id-commit.service has been
5712 added. When used on systems where /etc is read-only during
5713 boot, and /etc/machine-id is not initialized (but an empty
5714 file), this service will copy the temporary machine ID
5715 created as replacement into /etc after the system is fully
5716 booted up. This is useful for systems that are freshly
5717 installed with a non-initialized machine ID, but should get
5718 a fixed machine ID for subsequent boots.
5720 * networkd's .netdev files now provide a large set of
5721 configuration parameters for VXLAN devices. Similarly, the
5722 bridge port cost parameter is now configurable in .network
5723 files. There's also new support for configuring IP source
5724 routing. networkd .link files gained support for a new
5725 OriginalName= match that is useful to match against the
5726 original interface name the kernel assigned. .network files
5727 may include MTU= and MACAddress= fields for altering the MTU
5728 and MAC address while being connected to a specific network
5731 * The LUKS logic gained supported for configuring
5732 UUID-specific key files. There's also new support for naming
5733 LUKS device from the kernel command line, using the new
5734 luks.name= argument.
5736 * Timer units may now be transiently created via the bus API
5737 (this was previously already available for scope and service
5738 units). In addition it is now possible to create multiple
5739 transient units at the same time with a single bus call. The
5740 "systemd-run" tool has been updated to make use of this for
5741 running commands on a specified time, in at(1)-style.
5743 * tmpfiles gained support for "t" lines, for assigning
5744 extended attributes to files. Among other uses this may be
5745 used to assign SMACK labels to files.
5747 Contributions from: Alin Rauta, Alison Chaiken, Andrej
5748 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
5749 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
5750 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
5751 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
5752 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
5753 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
5754 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
5755 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
5756 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
5757 Michael Biebl, Michael Chapman, Michael Marineau, Michal
5758 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
5759 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
5760 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
5761 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
5762 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
5763 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
5764 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
5766 — Berlin, 2014-12-10
5770 * journalctl gained the new options -t/--identifier= to match
5771 on the syslog identifier (aka "tag"), as well as --utc to
5772 show log timestamps in the UTC timezone. journalctl now also
5773 accepts -n/--lines=all to disable line capping in a pager.
5775 * journalctl gained a new switch, --flush, that synchronously
5776 flushes logs from /run/log/journal to /var/log/journal if
5777 persistent storage is enabled. systemd-journal-flush.service
5778 now waits until the operation is complete.
5780 * Services can notify the manager before they start a reload
5781 (by sending RELOADING=1) or shutdown (by sending
5782 STOPPING=1). This allows the manager to track and show the
5783 internal state of daemons and closes a race condition when
5784 the process is still running but has closed its D-Bus
5787 * Services with Type=oneshot do not have to have any ExecStart
5790 * User units are now loaded also from
5791 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
5792 /run/systemd/user directory that was already previously
5793 supported, but is under the control of the user.
5795 * Job timeouts (i.e. timeouts on the time a job that is
5796 queued stays in the run queue) can now optionally result in
5797 immediate reboot or power-off actions (JobTimeoutAction= and
5798 JobTimeoutRebootArgument=). This is useful on ".target"
5799 units, to limit the maximum time a target remains
5800 undispatched in the run queue, and to trigger an emergency
5801 operation in such a case. This is now used by default to
5802 turn off the system if boot-up (as defined by everything in
5803 basic.target) hangs and does not complete for at least
5804 15min. Also, if power-off or reboot hang for at least 30min
5805 an immediate power-off/reboot operation is triggered. This
5806 functionality is particularly useful to increase reliability
5807 on embedded devices, but also on laptops which might
5808 accidentally get powered on when carried in a backpack and
5809 whose boot stays stuck in a hard disk encryption passphrase
5812 * systemd-logind can be configured to also handle lid switch
5813 events even when the machine is docked or multiple displays
5814 are attached (HandleLidSwitchDocked= option).
5816 * A helper binary and a service have been added which can be
5817 used to resume from hibernation in the initramfs. A
5818 generator will parse the resume= option on the kernel
5819 command line to trigger resume.
5821 * A user console daemon systemd-consoled has been
5822 added. Currently, it is a preview, and will so far open a
5823 single terminal on each session of the user marked as
5824 Desktop=systemd-console.
5826 * Route metrics can be specified for DHCP routes added by
5829 * The SELinux context of socket-activated services can be set
5830 from the information provided by the networking stack
5831 (SELinuxContextFromNet= option).
5833 * Userspace firmware loading support has been removed and
5834 the minimum supported kernel version is thus bumped to 3.7.
5836 * Timeout for udev workers has been increased from 1 to 3
5837 minutes, but a warning will be printed after 1 minute to
5838 help diagnose kernel modules that take a long time to load.
5840 * Udev rules can now remove tags on devices with TAG-="foobar".
5842 * systemd's readahead implementation has been removed. In many
5843 circumstances it didn't give expected benefits even for
5844 rotational disk drives and was becoming less relevant in the
5845 age of SSDs. As none of the developers has been using
5846 rotating media anymore, and nobody stepped up to actively
5847 maintain this component of systemd it has now been removed.
5849 * Swap units can use Options= to specify discard options.
5850 Discard options specified for swaps in /etc/fstab are now
5853 * Docker containers are now detected as a separate type of
5856 * The Password Agent protocol gained support for queries where
5857 the user input is shown, useful e.g. for user names.
5858 systemd-ask-password gained a new --echo option to turn that
5861 * The default sysctl.d/ snippets will now set:
5863 net.core.default_qdisc = fq_codel
5865 This selects Fair Queuing Controlled Delay as the default
5866 queuing discipline for network interfaces. fq_codel helps
5867 fight the network bufferbloat problem. It is believed to be
5868 a good default with no tuning required for most workloads.
5869 Downstream distributions may override this choice. On 10Gbit
5870 servers that do not do forwarding, "fq" may perform better.
5871 Systems without a good clocksource should use "pfifo_fast".
5873 * If kdbus is enabled during build a new option BusPolicy= is
5874 available for service units, that allows locking all service
5875 processes into a stricter bus policy, in order to limit
5876 access to various bus services, or even hide most of them
5877 from the service's view entirely.
5879 * networkctl will now show the .network and .link file
5880 networkd has applied to a specific interface.
5882 * sd-login gained a new API call sd_session_get_desktop() to
5883 query which desktop environment has been selected for a
5886 * UNIX utmp support is now compile-time optional to support
5887 legacy-free systems.
5889 * systemctl gained two new commands "add-wants" and
5890 "add-requires" for pulling in units from specific targets
5893 * If the word "rescue" is specified on the kernel command line
5894 the system will now boot into rescue mode (aka
5895 rescue.target), which was previously available only by
5896 specifying "1" or "systemd.unit=rescue.target" on the kernel
5897 command line. This new kernel command line option nicely
5898 mirrors the already existing "emergency" kernel command line
5901 * New kernel command line options mount.usr=, mount.usrflags=,
5902 mount.usrfstype= have been added that match root=, rootflags=,
5903 rootfstype= but allow mounting a specific file system to
5906 * The $NOTIFY_SOCKET is now also passed to control processes of
5907 services, not only the main process.
5909 * This version reenables support for fsck's -l switch. This
5910 means at least version v2.25 of util-linux is required for
5911 operation, otherwise dead-locks on device nodes may
5912 occur. Again: you need to update util-linux to at least
5913 v2.25 when updating systemd to v217.
5915 * The "multi-seat-x" tool has been removed from systemd, as
5916 its functionality has been integrated into X servers 1.16,
5917 and the tool is hence redundant. It is recommended to update
5918 display managers invoking this tool to simply invoke X
5919 directly from now on, again.
5921 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
5922 message flag has been added for all of systemd's polkit
5923 authenticated method calls has been added. In particular this
5924 now allows optional interactive authorization via polkit for
5925 many of PID1's privileged operations such as unit file
5926 enabling and disabling.
5928 * "udevadm hwdb --update" learnt a new switch "--usr" for
5929 placing the rebuilt hardware database in /usr instead of
5930 /etc. When used only hardware database entries stored in
5931 /usr will be used, and any user database entries in /etc are
5932 ignored. This functionality is useful for vendors to ship a
5933 pre-built database on systems where local configuration is
5934 unnecessary or unlikely.
5936 * Calendar time specifications in .timer units now also
5937 understand the strings "semi-annually", "quarterly" and
5938 "minutely" as shortcuts (in addition to the preexisting
5939 "anually", "hourly", ...).
5941 * systemd-tmpfiles will now correctly create files in /dev
5942 at boot which are marked for creation only at boot. It is
5943 recommended to always create static device nodes with 'c!'
5944 and 'b!', so that they are created only at boot and not
5945 overwritten at runtime.
5947 * When the watchdog logic is used for a service (WatchdogSec=)
5948 and the watchdog timeout is hit the service will now be
5949 terminated with SIGABRT (instead of just SIGTERM), in order
5950 to make sure a proper coredump and backtrace is
5951 generated. This ensures that hanging services will result in
5952 similar coredump/backtrace behaviour as services that hit a
5955 Contributions from: Andreas Henriksson, Andrei Borzenkov,
5956 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
5957 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
5958 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
5959 Herrmann, David Sommerseth, David Strauss, Emil Renner
5960 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
5961 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
5962 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
5963 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
5964 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
5965 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
5966 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
5967 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
5968 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
5969 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
5970 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
5971 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
5972 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
5973 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
5974 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
5975 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
5978 — Berlin, 2014-10-28
5982 * timedated no longer reads NTP implementation unit names from
5983 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
5984 implementations should add a
5986 Conflicts=systemd-timesyncd.service
5988 to their unit files to take over and replace systemd's NTP
5989 default functionality.
5991 * systemd-sysusers gained a new line type "r" for configuring
5992 which UID/GID ranges to allocate system users/groups
5993 from. Lines of type "u" may now add an additional column
5994 that specifies the home directory for the system user to be
5995 created. Also, systemd-sysusers may now optionally read user
5996 information from STDIN instead of a file. This is useful for
5997 invoking it from RPM preinst scriptlets that need to create
5998 users before the first RPM file is installed since these
5999 files might need to be owned by them. A new
6000 %sysusers_create_inline RPM macro has been introduced to do
6001 just that. systemd-sysusers now updates the shadow files as
6002 well as the user/group databases, which should enhance
6003 compatibility with certain tools like grpck.
6005 * A number of bus APIs of PID 1 now optionally consult polkit to
6006 permit access for otherwise unprivileged clients under certain
6007 conditions. Note that this currently doesn't support
6008 interactive authentication yet, but this is expected to be
6009 added eventually, too.
6011 * /etc/machine-info now has new fields for configuring the
6012 deployment environment of the machine, as well as the
6013 location of the machine. hostnamectl has been updated with
6014 new command to update these fields.
6016 * systemd-timesyncd has been updated to automatically acquire
6017 NTP server information from systemd-networkd, which might
6018 have been discovered via DHCP.
6020 * systemd-resolved now includes a caching DNS stub resolver
6021 and a complete LLMNR name resolution implementation. A new
6022 NSS module "nss-resolve" has been added which can be used
6023 instead of glibc's own "nss-dns" to resolve hostnames via
6024 systemd-resolved. Hostnames, addresses and arbitrary RRs may
6025 be resolved via systemd-resolved D-Bus APIs. In contrast to
6026 the glibc internal resolver systemd-resolved is aware of
6027 multi-homed system, and keeps DNS server and caches separate
6028 and per-interface. Queries are sent simultaneously on all
6029 interfaces that have DNS servers configured, in order to
6030 properly handle VPNs and local LANs which might resolve
6031 separate sets of domain names. systemd-resolved may acquire
6032 DNS server information from systemd-networkd automatically,
6033 which in turn might have discovered them via DHCP. A tool
6034 "systemd-resolve-host" has been added that may be used to
6035 query the DNS logic in resolved. systemd-resolved implements
6036 IDNA and automatically uses IDNA or UTF-8 encoding depending
6037 on whether classic DNS or LLMNR is used as transport. In the
6038 next releases we intend to add a DNSSEC and mDNS/DNS-SD
6039 implementation to systemd-resolved.
6041 * A new NSS module nss-mymachines has been added, that
6042 automatically resolves the names of all local registered
6043 containers to their respective IP addresses.
6045 * A new client tool "networkctl" for systemd-networkd has been
6046 added. It currently is entirely passive and will query
6047 networking configuration from udev, rtnetlink and networkd,
6048 and present it to the user in a very friendly
6049 way. Eventually, we hope to extend it to become a full
6050 control utility for networkd.
6052 * .socket units gained a new DeferAcceptSec= setting that
6053 controls the kernels' TCP_DEFER_ACCEPT sockopt for
6054 TCP. Similarly, support for controlling TCP keep-alive
6055 settings has been added (KeepAliveTimeSec=,
6056 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
6057 turning off Nagle's algorithm on TCP has been added
6060 * logind learned a new session type "web", for use in projects
6061 like Cockpit which register web clients as PAM sessions.
6063 * timer units with at least one OnCalendar= setting will now
6064 be started only after time-sync.target has been
6065 reached. This way they will not elapse before the system
6066 clock has been corrected by a local NTP client or
6067 similar. This is particular useful on RTC-less embedded
6068 machines, that come up with an invalid system clock.
6070 * systemd-nspawn's --network-veth= switch should now result in
6071 stable MAC addresses for both the outer and the inner side
6074 * systemd-nspawn gained a new --volatile= switch for running
6075 container instances with /etc or /var unpopulated.
6077 * The kdbus client code has been updated to use the new Linux
6078 3.17 memfd subsystem instead of the old kdbus-specific one.
6080 * systemd-networkd's DHCP client and server now support
6081 FORCERENEW. There are also new configuration options to
6082 configure the vendor client identifier and broadcast mode
6085 * systemd will no longer inform the kernel about the current
6086 timezone, as this is necessarily incorrect and racy as the
6087 kernel has no understanding of DST and similar
6088 concepts. This hence means FAT timestamps will be always
6089 considered UTC, similar to what Android is already
6090 doing. Also, when the RTC is configured to the local time
6091 (rather than UTC) systemd will never synchronize back to it,
6092 as this might confuse Windows at a later boot.
6094 * systemd-analyze gained a new command "verify" for offline
6095 validation of unit files.
6097 * systemd-networkd gained support for a couple of additional
6098 settings for bonding networking setups. Also, the metric for
6099 statically configured routes may now be configured. For
6100 network interfaces where this is appropriate the peer IP
6101 address may now be configured.
6103 * systemd-networkd's DHCP client will no longer request
6104 broadcasting by default, as this tripped up some networks.
6105 For hardware where broadcast is required the feature should
6106 be switched back on using RequestBroadcast=yes.
6108 * systemd-networkd will now set up IPv4LL addresses (when
6109 enabled) even if DHCP is configured successfully.
6111 * udev will now default to respect network device names given
6112 by the kernel when the kernel indicates that these are
6113 predictable. This behavior can be tweaked by changing
6114 NamePolicy= in the relevant .link file.
6116 * A new library systemd-terminal has been added that
6117 implements full TTY stream parsing and rendering. This
6118 library is supposed to be used later on for implementing a
6119 full userspace VT subsystem, replacing the current kernel
6122 * A new tool systemd-journal-upload has been added to push
6123 journal data to a remote system running
6124 systemd-journal-remote.
6126 * journald will no longer forward all local data to another
6127 running syslog daemon. This change has been made because
6128 rsyslog (which appears to be the most commonly used syslog
6129 implementation these days) no longer makes use of this, and
6130 instead pulls the data out of the journal on its own. Since
6131 forwarding the messages to a non-existent syslog server is
6132 more expensive than we assumed we have now turned this
6133 off. If you run a syslog server that is not a recent rsyslog
6134 version, you have to turn this option on again
6135 (ForwardToSyslog= in journald.conf).
6137 * journald now optionally supports the LZ4 compressor for
6138 larger journal fields. This compressor should perform much
6139 better than XZ which was the previous default.
6141 * machinectl now shows the IP addresses of local containers,
6142 if it knows them, plus the interface name of the container.
6144 * A new tool "systemd-escape" has been added that makes it
6145 easy to escape strings to build unit names and similar.
6147 * sd_notify() messages may now include a new ERRNO= field
6148 which is parsed and collected by systemd and shown among the
6149 "systemctl status" output for a service.
6151 * A new component "systemd-firstboot" has been added that
6152 queries the most basic systemd information (timezone,
6153 hostname, root password) interactively on first
6154 boot. Alternatively it may also be used to provision these
6155 things offline on OS images installed into directories.
6157 * The default sysctl.d/ snippets will now set
6159 net.ipv4.conf.default.promote_secondaries=1
6161 This has the benefit of no flushing secondary IP addresses
6162 when primary addresses are removed.
6164 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
6165 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
6166 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
6167 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
6168 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
6169 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
6170 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
6171 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
6172 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
6173 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
6174 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
6175 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
6176 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
6177 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
6178 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
6180 — Berlin, 2014-08-19
6184 * A new tool systemd-sysusers has been added. This tool
6185 creates system users and groups in /etc/passwd and
6186 /etc/group, based on static declarative system user/group
6187 definitions in /usr/lib/sysusers.d/. This is useful to
6188 enable factory resets and volatile systems that boot up with
6189 an empty /etc directory, and thus need system users and
6190 groups created during early boot. systemd now also ships
6191 with two default sysusers.d/ files for the most basic
6192 users and groups systemd and the core operating system
6195 * A new tmpfiles snippet has been added that rebuilds the
6196 essential files in /etc on boot, should they be missing.
6198 * A directive for ensuring automatic clean-up of
6199 /var/cache/man/ has been removed from the default
6200 configuration. This line should now be shipped by the man
6201 implementation. The necessary change has been made to the
6202 man-db implementation. Note that you need to update your man
6203 implementation to one that ships this line, otherwise no
6204 automatic clean-up of /var/cache/man will take place.
6206 * A new condition ConditionNeedsUpdate= has been added that
6207 may conditionalize services to only run when /etc or /var
6208 are "older" than the vendor operating system resources in
6209 /usr. This is useful for reconstructing or updating /etc
6210 after an offline update of /usr or a factory reset, on the
6211 next reboot. Services that want to run once after such an
6212 update or reset should use this condition and order
6213 themselves before the new systemd-update-done.service, which
6214 will mark the two directories as fully updated. A number of
6215 service files have been added making use of this, to rebuild
6216 the udev hardware database, the journald message catalog and
6217 dynamic loader cache (ldconfig). The systemd-sysusers tool
6218 described above also makes use of this now. With this in
6219 place it is now possible to start up a minimal operating
6220 system with /etc empty cleanly. For more information on the
6221 concepts involved see this recent blog story:
6223 http://0pointer.de/blog/projects/stateless.html
6225 * A new system group "input" has been introduced, and all
6226 input device nodes get this group assigned. This is useful
6227 for system-level software to get access to input devices. It
6228 complements what is already done for "audio" and "video".
6230 * systemd-networkd learnt minimal DHCPv4 server support in
6231 addition to the existing DHCPv4 client support. It also
6232 learnt DHCPv6 client and IPv6 Router Solicitation client
6233 support. The DHCPv4 client gained support for static routes
6234 passed in from the server. Note that the [DHCPv4] section
6235 known in older systemd-networkd versions has been renamed to
6236 [DHCP] and is now also used by the DHCPv6 client. Existing
6237 .network files using settings of this section should be
6238 updated, though compatibility is maintained. Optionally, the
6239 client hostname may now be sent to the DHCP server.
6241 * networkd gained support for vxlan virtual networks as well
6242 as tun/tap and dummy devices.
6244 * networkd gained support for automatic allocation of address
6245 ranges for interfaces from a system-wide pool of
6246 addresses. This is useful for dynamically managing a large
6247 number of interfaces with a single network configuration
6248 file. In particular this is useful to easily assign
6249 appropriate IP addresses to the veth links of a large number
6250 of nspawn instances.
6252 * RPM macros for processing sysusers, sysctl and binfmt
6253 drop-in snippets at package installation time have been
6256 * The /etc/os-release file should now be placed in
6257 /usr/lib/os-release. The old location is automatically
6258 created as symlink. /usr/lib is the more appropriate
6259 location of this file, since it shall actually describe the
6260 vendor operating system shipped in /usr, and not the
6261 configuration stored in /etc.
6263 * .mount units gained a new boolean SloppyOptions= setting
6264 that maps to mount(8)'s -s option which enables permissive
6265 parsing of unknown mount options.
6267 * tmpfiles learnt a new "L+" directive which creates a symlink
6268 but (unlike "L") deletes a pre-existing file first, should
6269 it already exist and not already be the correct
6270 symlink. Similarly, "b+", "c+" and "p+" directives have been
6271 added as well, which create block and character devices, as
6272 well as fifos in the filesystem, possibly removing any
6273 pre-existing files of different types.
6275 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
6276 'argument' field (which so far specified the source to
6277 symlink/copy the files from) is now optional. If omitted the
6278 same file os copied from /usr/share/factory/ suffixed by the
6279 full destination path. This is useful for populating /etc
6280 with essential files, by copying them from vendor defaults
6281 shipped in /usr/share/factory/etc.
6283 * A new command "systemctl preset-all" has been added that
6284 applies the service preset settings to all installed unit
6285 files. A new switch --preset-mode= has been added that
6286 controls whether only enable or only disable operations
6289 * A new command "systemctl is-system-running" has been added
6290 that allows checking the overall state of the system, for
6291 example whether it is fully up and running.
6293 * When the system boots up with an empty /etc, the equivalent
6294 to "systemctl preset-all" is executed during early boot, to
6295 make sure all default services are enabled after a factory
6298 * systemd now contains a minimal preset file that enables the
6299 most basic services systemd ships by default.
6301 * Unit files' [Install] section gained a new DefaultInstance=
6302 field for defining the default instance to create if a
6303 template unit is enabled with no instance specified.
6305 * A new passive target cryptsetup-pre.target has been added
6306 that may be used by services that need to make they run and
6307 finish before the first LUKS cryptographic device is set up.
6309 * The /dev/loop-control and /dev/btrfs-control device nodes
6310 are now owned by the "disk" group by default, opening up
6311 access to this group.
6313 * systemd-coredump will now automatically generate a
6314 stack trace of all core dumps taking place on the system,
6315 based on elfutils' libdw library. This stack trace is logged
6318 * systemd-coredump may now optionally store coredumps directly
6319 on disk (in /var/lib/systemd/coredump, possibly compressed),
6320 instead of storing them unconditionally in the journal. This
6321 mode is the new default. A new configuration file
6322 /etc/systemd/coredump.conf has been added to configure this
6323 and other parameters of systemd-coredump.
6325 * coredumpctl gained a new "info" verb to show details about a
6326 specific coredump. A new switch "-1" has also been added
6327 that makes sure to only show information about the most
6328 recent entry instead of all entries. Also, as the tool is
6329 generally useful now the "systemd-" prefix of the binary
6330 name has been removed. Distributions that want to maintain
6331 compatibility with the old name should add a symlink from
6332 the old name to the new name.
6334 * journald's SplitMode= now defaults to "uid". This makes sure
6335 that unprivileged users can access their own coredumps with
6336 coredumpctl without restrictions.
6338 * New kernel command line options "systemd.wants=" (for
6339 pulling an additional unit during boot), "systemd.mask="
6340 (for masking a specific unit for the boot), and
6341 "systemd.debug-shell" (for enabling the debug shell on tty9)
6342 have been added. This is implemented in the new generator
6343 "systemd-debug-generator".
6345 * systemd-nspawn will now by default filter a couple of
6346 syscalls for containers, among them those required for
6347 kernel module loading, direct x86 IO port access, swap
6348 management, and kexec. Most importantly though
6349 open_by_handle_at() is now prohibited for containers,
6350 closing a hole similar to a recently discussed vulnerability
6351 in docker regarding access to files on file hierarchies the
6352 container should normally not have access to. Note that, for
6353 nspawn, we generally make no security claims anyway (and
6354 this is explicitly documented in the man page), so this is
6355 just a fix for one of the most obvious problems.
6357 * A new man page file-hierarchy(7) has been added that
6358 contains a minimized, modernized version of the file system
6359 layout systemd expects, similar in style to the FHS
6360 specification or hier(5). A new tool systemd-path(1) has
6361 been added to query many of these paths for the local
6364 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
6365 longer done. Since the directory now has a per-user size
6366 limit, and is cleaned on logout this appears unnecessary,
6367 in particular since this now brings the lifecycle of this
6368 directory closer in line with how IPC objects are handled.
6370 * systemd.pc now exports a number of additional directories,
6371 including $libdir (which is useful to identify the library
6372 path for the primary architecture of the system), and a
6373 couple of drop-in directories.
6375 * udev's predictable network interface names now use the dev_port
6376 sysfs attribute, introduced in linux 3.15 instead of dev_id to
6377 distinguish between ports of the same PCI function. dev_id should
6378 only be used for ports using the same HW address, hence the need
6381 * machined has been updated to export the OS version of a
6382 container (read from /etc/os-release and
6383 /usr/lib/os-release) on the bus. This is now shown in
6384 "machinectl status" for a machine.
6386 * A new service setting RestartForceExitStatus= has been
6387 added. If configured to a set of exit signals or process
6388 return values, the service will be restarted when the main
6389 daemon process exits with any of them, regardless of the
6392 * systemctl's -H switch for connecting to remote systemd
6393 machines has been extended so that it may be used to
6394 directly connect to a specific container on the
6395 host. "systemctl -H root@foobar:waldi" will now connect as
6396 user "root" to host "foobar", and then proceed directly to
6397 the container named "waldi". Note that currently you have to
6398 authenticate as user "root" for this to work, as entering
6399 containers is a privileged operation.
6401 Contributions from: Andreas Henriksson, Benjamin Steinwender,
6402 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
6403 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
6404 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
6405 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
6406 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
6407 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
6408 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
6409 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
6410 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
6411 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
6412 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
6414 — Berlin, 2014-07-03
6418 * As an experimental feature, udev now tries to lock the
6419 disk device node (flock(LOCK_SH|LOCK_NB)) while it
6420 executes events for the disk or any of its partitions.
6421 Applications like partitioning programs can lock the
6422 disk device node (flock(LOCK_EX)) and claim temporary
6423 device ownership that way; udev will entirely skip all event
6424 handling for this disk and its partitions. If the disk
6425 was opened for writing, the close will trigger a partition
6426 table rescan in udev's "watch" facility, and if needed
6427 synthesize "change" events for the disk and all its partitions.
6428 This is now unconditionally enabled, and if it turns out to
6429 cause major problems, we might turn it on only for specific
6430 devices, or might need to disable it entirely. Device Mapper
6431 devices are excluded from this logic.
6433 * We temporarily dropped the "-l" switch for fsck invocations,
6434 since they collide with the flock() logic above. util-linux
6435 upstream has been changed already to avoid this conflict,
6436 and we will re-add "-l" as soon as util-linux with this
6437 change has been released.
6439 * The dependency on libattr has been removed. Since a long
6440 time, the extended attribute calls have moved to glibc, and
6441 libattr is thus unnecessary.
6443 * Virtualization detection works without privileges now. This
6444 means the systemd-detect-virt binary no longer requires
6445 CAP_SYS_PTRACE file capabilities, and our daemons can run
6446 with fewer privileges.
6448 * systemd-networkd now runs under its own "systemd-network"
6449 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
6450 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
6451 loses the ability to write to files owned by root this way.
6453 * Similarly, systemd-resolved now runs under its own
6454 "systemd-resolve" user with no capabilities remaining.
6456 * Similarly, systemd-bus-proxyd now runs under its own
6457 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
6459 * systemd-networkd gained support for setting up "veth"
6460 virtual Ethernet devices for container connectivity, as well
6461 as GRE and VTI tunnels.
6463 * systemd-networkd will no longer automatically attempt to
6464 manually load kernel modules necessary for certain tunnel
6465 transports. Instead, it is assumed the kernel loads them
6466 automatically when required. This only works correctly on
6467 very new kernels. On older kernels, please consider adding
6468 the kernel modules to /etc/modules-load.d/ as a work-around.
6470 * The resolv.conf file systemd-resolved generates has been
6471 moved to /run/systemd/resolve/. If you have a symlink from
6472 /etc/resolv.conf, it might be necessary to correct it.
6474 * Two new service settings, ProtectHome= and ProtectSystem=,
6475 have been added. When enabled, they will make the user data
6476 (such as /home) inaccessible or read-only and the system
6477 (such as /usr) read-only, for specific services. This allows
6478 very light-weight per-service sandboxing to avoid
6479 modifications of user data or system files from
6480 services. These two new switches have been enabled for all
6481 of systemd's long-running services, where appropriate.
6483 * Socket units gained new SocketUser= and SocketGroup=
6484 settings to set the owner user and group of AF_UNIX sockets
6485 and FIFOs in the file system.
6487 * Socket units gained a new RemoveOnStop= setting. If enabled,
6488 all FIFOS and sockets in the file system will be removed
6489 when the specific socket unit is stopped.
6491 * Socket units gained a new Symlinks= setting. It takes a list
6492 of symlinks to create to file system sockets or FIFOs
6493 created by the specific Unix sockets. This is useful to
6494 manage symlinks to socket nodes with the same lifecycle as
6497 * The /dev/log socket and /dev/initctl FIFO have been moved to
6498 /run, and have been replaced by symlinks. This allows
6499 connecting to these facilities even if PrivateDevices=yes is
6500 used for a service (which makes /dev/log itself unavailable,
6501 but /run is left). This also has the benefit of ensuring
6502 that /dev only contains device nodes, directories and
6503 symlinks, and nothing else.
6505 * sd-daemon gained two new calls sd_pid_notify() and
6506 sd_pid_notifyf(). They are similar to sd_notify() and
6507 sd_notifyf(), but allow overriding of the source PID of
6508 notification messages if permissions permit this. This is
6509 useful to send notify messages on behalf of a different
6510 process (for example, the parent process). The
6511 systemd-notify tool has been updated to make use of this
6512 when sending messages (so that notification messages now
6513 originate from the shell script invoking systemd-notify and
6514 not the systemd-notify process itself. This should minimize
6515 a race where systemd fails to associate notification
6516 messages to services when the originating process already
6519 * A new "on-abnormal" setting for Restart= has been added. If
6520 set, it will result in automatic restarts on all "abnormal"
6521 reasons for a process to exit, which includes unclean
6522 signals, core dumps, timeouts and watchdog timeouts, but
6523 does not include clean and unclean exit codes or clean
6524 signals. Restart=on-abnormal is an alternative for
6525 Restart=on-failure for services that shall be able to
6526 terminate and avoid restarts on certain errors, by
6527 indicating so with an unclean exit code. Restart=on-failure
6528 or Restart=on-abnormal is now the recommended setting for
6529 all long-running services.
6531 * If the InaccessibleDirectories= service setting points to a
6532 mount point (or if there are any submounts contained within
6533 it), it is now attempted to completely unmount it, to make
6534 the file systems truly unavailable for the respective
6537 * The ReadOnlyDirectories= service setting and
6538 systemd-nspawn's --read-only parameter are now recursively
6539 applied to all submounts, too.
6541 * Mount units may now be created transiently via the bus APIs.
6543 * The support for SysV and LSB init scripts has been removed
6544 from the systemd daemon itself. Instead, it is now
6545 implemented as a generator that creates native systemd units
6546 from these scripts when needed. This enables us to remove a
6547 substantial amount of legacy code from PID 1, following the
6548 fact that many distributions only ship a very small number
6549 of LSB/SysV init scripts nowadays.
6551 * Privileged Xen (dom0) domains are not considered
6552 virtualization anymore by the virtualization detection
6553 logic. After all, they generally have unrestricted access to
6554 the hardware and usually are used to manage the unprivileged
6557 * systemd-tmpfiles gained a new "C" line type, for copying
6558 files or entire directories.
6560 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
6561 lines. So far, they have been non-globbing versions of the
6562 latter, and have thus been redundant. In future, it is
6563 recommended to only use "z". "m" has hence been removed
6564 from the documentation, even though it stays supported.
6566 * A tmpfiles snippet to recreate the most basic structure in
6567 /var has been added. This is enough to create the /var/run →
6568 /run symlink and create a couple of structural
6569 directories. This allows systems to boot up with an empty or
6570 volatile /var. Of course, while with this change, the core OS
6571 now is capable with dealing with a volatile /var, not all
6572 user services are ready for it. However, we hope that sooner
6573 or later, many service daemons will be changed upstream so
6574 that they are able to automatically create their necessary
6575 directories in /var at boot, should they be missing. This is
6576 the first step to allow state-less systems that only require
6577 the vendor image for /usr to boot.
6579 * systemd-nspawn has gained a new --tmpfs= switch to mount an
6580 empty tmpfs instance to a specific directory. This is
6581 particularly useful for making use of the automatic
6582 reconstruction of /var (see above), by passing --tmpfs=/var.
6584 * Access modes specified in tmpfiles snippets may now be
6585 prefixed with "~", which indicates that they shall be masked
6586 by whether the existing file or directory is currently
6587 writable, readable or executable at all. Also, if specified,
6588 the sgid/suid/sticky bits will be masked for all
6591 * A new passive target unit "network-pre.target" has been
6592 added which is useful for services that shall run before any
6593 network is configured, for example firewall scripts.
6595 * The "floppy" group that previously owned the /dev/fd*
6596 devices is no longer used. The "disk" group is now used
6597 instead. Distributions should probably deprecate usage of
6600 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
6601 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
6602 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
6603 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
6604 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
6605 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
6608 — Berlin, 2014-06-11
6612 * A new "systemd-timesyncd" daemon has been added for
6613 synchronizing the system clock across the network. It
6614 implements an SNTP client. In contrast to NTP
6615 implementations such as chrony or the NTP reference server,
6616 this only implements a client side, and does not bother with
6617 the full NTP complexity, focusing only on querying time from
6618 one remote server and synchronizing the local clock to
6619 it. Unless you intend to serve NTP to networked clients or
6620 want to connect to local hardware clocks, this simple NTP
6621 client should be more than appropriate for most
6622 installations. The daemon runs with minimal privileges, and
6623 has been hooked up with networkd to only operate when
6624 network connectivity is available. The daemon saves the
6625 current clock to disk every time a new NTP sync has been
6626 acquired, and uses this to possibly correct the system clock
6627 early at bootup, in order to accommodate for systems that
6628 lack an RTC such as the Raspberry Pi and embedded devices,
6629 and to make sure that time monotonically progresses on these
6630 systems, even if it is not always correct. To make use of
6631 this daemon, a new system user and group "systemd-timesync"
6632 needs to be created on installation of systemd.
6634 * The queue "seqnum" interface of libudev has been disabled, as
6635 it was generally incompatible with device namespacing as
6636 sequence numbers of devices go "missing" if the devices are
6637 part of a different namespace.
6639 * "systemctl list-timers" and "systemctl list-sockets" gained
6640 a --recursive switch for showing units of these types also
6641 for all local containers, similar in style to the already
6642 supported --recursive switch for "systemctl list-units".
6644 * A new RebootArgument= setting has been added for service
6645 units, which may be used to specify a kernel reboot argument
6646 to use when triggering reboots with StartLimitAction=.
6648 * A new FailureAction= setting has been added for service
6649 units which may be used to specify an operation to trigger
6650 when a service fails. This works similarly to
6651 StartLimitAction=, but unlike it, controls what is done
6652 immediately rather than only after several attempts to
6653 restart the service in question.
6655 * hostnamed got updated to also expose the kernel name,
6656 release, and version on the bus. This is useful for
6657 executing commands like hostnamectl with the -H switch.
6658 systemd-analyze makes use of this to properly display
6659 details when running non-locally.
6661 * The bootchart tool can now show cgroup information in the
6662 graphs it generates.
6664 * The CFS CPU quota cgroup attribute is now exposed for
6665 services. The new CPUQuota= switch has been added for this
6666 which takes a percentage value. Setting this will have the
6667 result that a service may never get more CPU time than the
6668 specified percentage, even if the machine is otherwise idle.
6670 * systemd-networkd learned IPIP and SIT tunnel support.
6672 * LSB init scripts exposing a dependency on $network will now
6673 get a dependency on network-online.target rather than simply
6674 network.target. This should bring LSB handling closer to
6675 what it was on SysV systems.
6677 * A new fsck.repair= kernel option has been added to control
6678 how fsck shall deal with unclean file systems at boot.
6680 * The (.ini) configuration file parser will now silently
6681 ignore sections whose name begins with "X-". This may be
6682 used to maintain application-specific extension sections in unit
6685 * machined gained a new API to query the IP addresses of
6686 registered containers. "machinectl status" has been updated
6687 to show these addresses in its output.
6689 * A new call sd_uid_get_display() has been added to the
6690 sd-login APIs for querying the "primary" session of a
6691 user. The "primary" session of the user is elected from the
6692 user's sessions and generally a graphical session is
6693 preferred over a text one.
6695 * A minimal systemd-resolved daemon has been added. It
6696 currently simply acts as a companion to systemd-networkd and
6697 manages resolv.conf based on per-interface DNS
6698 configuration, possibly supplied via DHCP. In the long run
6699 we hope to extend this into a local DNSSEC enabled DNS and
6702 * The systemd-networkd-wait-online tool is now enabled by
6703 default. It will delay network-online.target until a network
6704 connection has been configured. The tool primarily integrates
6705 with networkd, but will also make a best effort to make sense
6706 of network configuration performed in some other way.
6708 * Two new service options StartupCPUShares= and
6709 StartupBlockIOWeight= have been added that work similarly to
6710 CPUShares= and BlockIOWeight= however only apply during
6711 system startup. This is useful to prioritize certain services
6712 differently during bootup than during normal runtime.
6714 * hostnamed has been changed to prefer the statically
6715 configured hostname in /etc/hostname (unless set to
6716 'localhost' or empty) over any dynamic one supplied by
6717 dhcp. With this change, the rules for picking the hostname
6718 match more closely the rules of other configuration settings
6719 where the local administrator's configuration in /etc always
6720 overrides any other settings.
6722 Contributions from: Ali H. Caliskan, Alison Chaiken, Bas van
6723 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
6724 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
6725 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
6726 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
6727 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
6728 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
6729 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
6730 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
6731 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
6732 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
6733 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
6734 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
6735 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
6736 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
6737 Lindskog, WaLyong Cho, Will Woods, Zbigniew
6740 — Beijing, 2014-05-28
6744 * When restoring the screen brightness at boot, stay away from
6745 the darkest setting or from the lowest 5% of the available
6746 range, depending on which is the larger value of both. This
6747 should effectively protect the user from rebooting into a
6748 black screen, should the brightness have been set to minimum
6751 * sd-login gained a new sd_machine_get_class() call to
6752 determine the class ("vm" or "container") of a machine
6753 registered with machined.
6755 * sd-login gained new calls
6756 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
6757 to query the identity of the peer of a local AF_UNIX
6758 connection. They operate similarly to their sd_pid_get_xyz()
6761 * PID 1 will now maintain a system-wide system state engine
6762 with the states "starting", "running", "degraded",
6763 "maintenance", "stopping". These states are bound to system
6764 startup, normal runtime, runtime with at least one failed
6765 service, rescue/emergency mode and system shutdown. This
6766 state is shown in the "systemctl status" output when no unit
6767 name is passed. It is useful to determine system state, in
6768 particularly when doing so for many systems or containers at
6771 * A new command "list-machines" has been added to "systemctl"
6772 that lists all local OS containers and shows their system
6773 state (see above), if systemd runs inside of them.
6775 * systemctl gained a new "-r" switch to recursively enumerate
6776 units on all local containers, when used with the
6777 "list-unit" command (which is the default one that is
6778 executed when no parameters are specified).
6780 * The GPT automatic partition discovery logic will now honour
6781 two GPT partition flags: one may be set on a partition to
6782 cause it to be mounted read-only, and the other may be set
6783 on a partition to ignore it during automatic discovery.
6785 * Two new GPT type UUIDs have been added for automatic root
6786 partition discovery, for 32-bit and 64-bit ARM. This is not
6787 particularly useful for discovering the root directory on
6788 these architectures during bare-metal boots (since UEFI is
6789 not common there), but still very useful to allow booting of
6790 ARM disk images in nspawn with the -i option.
6792 * MAC addresses of interfaces created with nspawn's
6793 --network-interface= switch will now be generated from the
6794 machine name, and thus be stable between multiple invocations
6797 * logind will now automatically remove all IPC objects owned
6798 by a user if she or he fully logs out. This makes sure that
6799 users who are logged out cannot continue to consume IPC
6800 resources. This covers SysV memory, semaphores and message
6801 queues as well as POSIX shared memory and message
6802 queues. Traditionally, SysV and POSIX IPC had no lifecycle
6803 limits. With this functionality, that is corrected. This may
6804 be turned off by using the RemoveIPC= switch of logind.conf.
6806 * The systemd-machine-id-setup and tmpfiles tools gained a
6807 --root= switch to operate on a specific root directory,
6810 * journald can now forward logged messages to the TTYs of all
6811 logged in users ("wall"). This is the default for all
6812 emergency messages now.
6814 * A new tool systemd-journal-remote has been added to stream
6815 journal log messages across the network.
6817 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
6818 controller trees are mounted into it. Note that the
6819 directories mounted beneath it are not read-only. This is a
6820 security measure and is particularly useful because glibc
6821 actually includes a search logic to pick any tmpfs it can
6822 find to implement shm_open() if /dev/shm is not available
6823 (which it might very well be in namespaced setups).
6825 * machinectl gained a new "poweroff" command to cleanly power
6826 down a local OS container.
6828 * The PrivateDevices= unit file setting will now also drop the
6829 CAP_MKNOD capability from the capability bound set, and
6830 imply DevicePolicy=closed.
6832 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
6833 comprehensively on all long-running systemd services where
6834 this is appropriate.
6836 * systemd-udevd will now run in a disassociated mount
6837 namespace. To mount directories from udev rules, make sure to
6838 pull in mount units via SYSTEMD_WANTS properties.
6840 * The kdbus support gained support for uploading policy into
6841 the kernel. sd-bus gained support for creating "monitoring"
6842 connections that can eavesdrop into all bus communication
6843 for debugging purposes.
6845 * Timestamps may now be specified in seconds since the UNIX
6846 epoch Jan 1st, 1970 by specifying "@" followed by the value
6849 * Native tcpwrap support in systemd has been removed. tcpwrap
6850 is old code, not really maintained anymore and has serious
6851 shortcomings, and better options such as firewalls
6852 exist. For setups that require tcpwrap usage, please
6853 consider invoking your socket-activated service via tcpd,
6854 like on traditional inetd.
6856 * A new system.conf configuration option
6857 DefaultTimerAccuracySec= has been added that controls the
6858 default AccuracySec= setting of .timer units.
6860 * Timer units gained a new WakeSystem= switch. If enabled,
6861 timers configured this way will cause the system to resume
6862 from system suspend (if the system supports that, which most
6865 * Timer units gained a new Persistent= switch. If enabled,
6866 timers configured this way will save to disk when they have
6867 been last triggered. This information is then used on next
6868 reboot to possible execute overdue timer events, that
6869 could not take place because the system was powered off.
6870 This enables simple anacron-like behaviour for timer units.
6872 * systemctl's "list-timers" will now also list the time a
6873 timer unit was last triggered in addition to the next time
6874 it will be triggered.
6876 * systemd-networkd will now assign predictable IPv4LL
6877 addresses to its local interfaces.
6879 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
6880 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
6881 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
6882 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
6883 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
6884 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
6885 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
6886 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
6889 — Berlin, 2014-03-25
6893 * A new unit file setting RestrictAddressFamilies= has been
6894 added to restrict which socket address families unit
6895 processes gain access to. This takes address family names
6896 like "AF_INET" or "AF_UNIX", and is useful to minimize the
6897 attack surface of services via exotic protocol stacks. This
6898 is built on seccomp system call filters.
6900 * Two new unit file settings RuntimeDirectory= and
6901 RuntimeDirectoryMode= have been added that may be used to
6902 manage a per-daemon runtime directories below /run. This is
6903 an alternative for setting up directory permissions with
6904 tmpfiles snippets, and has the advantage that the runtime
6905 directory's lifetime is bound to the daemon runtime and that
6906 the daemon starts up with an empty directory each time. This
6907 is particularly useful when writing services that drop
6908 privileges using the User= or Group= setting.
6910 * The DeviceAllow= unit setting now supports globbing for
6911 matching against device group names.
6913 * The systemd configuration file system.conf gained new
6914 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
6915 DefaultMemoryAccounting= to globally turn on/off accounting
6916 for specific resources (cgroups) for all units. These
6917 settings may still be overridden individually in each unit
6920 * systemd-gpt-auto-generator is now able to discover /srv and
6921 root partitions in addition to /home and swap partitions. It
6922 also supports LUKS-encrypted partitions now. With this in
6923 place, automatic discovery of partitions to mount following
6924 the Discoverable Partitions Specification
6925 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
6926 is now a lot more complete. This allows booting without
6927 /etc/fstab and without root= on the kernel command line on
6928 systems prepared appropriately.
6930 * systemd-nspawn gained a new --image= switch which allows
6931 booting up disk images and Linux installations on any block
6932 device that follow the Discoverable Partitions Specification
6933 (see above). This means that installations made with
6934 appropriately updated installers may now be started and
6935 deployed using container managers, completely
6936 unmodified. (We hope that libvirt-lxc will add support for
6937 this feature soon, too.)
6939 * systemd-nspawn gained a new --network-macvlan= setting to
6940 set up a private macvlan interface for the
6941 container. Similarly, systemd-networkd gained a new
6942 Kind=macvlan setting in .netdev files.
6944 * systemd-networkd now supports configuring local addresses
6947 * A new tool systemd-network-wait-online has been added to
6948 synchronously wait for network connectivity using
6951 * The sd-bus.h bus API gained a new sd_bus_track object for
6952 tracking the lifecycle of bus peers. Note that sd-bus.h is
6953 still not a public API though (unless you specify
6954 --enable-kdbus on the configure command line, which however
6955 voids your warranty and you get no API stability guarantee).
6957 * The $XDG_RUNTIME_DIR runtime directories for each user are
6958 now individual tmpfs instances, which has the benefit of
6959 introducing separate pools for each user, with individual
6960 size limits, and thus making sure that unprivileged clients
6961 can no longer negatively impact the system or other users by
6962 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
6963 RuntimeDirectorySize= has been introduced that allows
6964 controlling the default size limit for all users. It
6965 defaults to 10% of the available physical memory. This is no
6966 replacement for quotas on tmpfs though (which the kernel
6967 still does not support), as /dev/shm and /tmp are still
6968 shared resources used by both the system and unprivileged
6971 * logind will now automatically turn off automatic suspending
6972 on laptop lid close when more than one display is
6973 connected. This was previously expected to be implemented
6974 individually in desktop environments (such as GNOME),
6975 however has been added to logind now, in order to fix a
6976 boot-time race where a desktop environment might not have
6977 been started yet and thus not been able to take an inhibitor
6978 lock at the time where logind already suspends the system
6979 due to a closed lid.
6981 * logind will now wait at least 30s after each system
6982 suspend/resume cycle, and 3min after system boot before
6983 suspending the system due to a closed laptop lid. This
6984 should give USB docking stations and similar enough time to
6985 be probed and configured after system resume and boot in
6986 order to then act as suspend blocker.
6988 * systemd-run gained a new --property= setting which allows
6989 initialization of resource control properties (and others)
6990 for the created scope or service unit. Example: "systemd-run
6991 --property=BlockIOWeight=10 updatedb" may be used to run
6992 updatedb at a low block IO scheduling weight.
6994 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
6995 now also work in --scope mode.
6997 * When systemd is compiled with kdbus support, basic support
6998 for enforced policies is now in place. (Note that enabling
6999 kdbus still voids your warranty and no API compatibility
7002 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
7003 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
7004 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
7005 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
7006 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
7007 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
7008 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
7009 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
7010 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
7011 Zbigniew Jędrzejewski-Szmek
7013 — Berlin, 2014-03-12
7017 * systemd will now relabel /dev after loading the SMACK policy
7018 according to SMACK rules.
7020 * A new unit file option AppArmorProfile= has been added to
7021 set the AppArmor profile for the processes of a unit.
7023 * A new condition check ConditionArchitecture= has been added
7024 to conditionalize units based on the system architecture, as
7025 reported by uname()'s "machine" field.
7027 * systemd-networkd now supports matching on the system
7028 virtualization, architecture, kernel command line, host name
7031 * logind is now a lot more aggressive when suspending the
7032 machine due to a closed laptop lid. Instead of acting only
7033 on the lid close action, it will continuously watch the lid
7034 status and act on it. This is useful for laptops where the
7035 power button is on the outside of the chassis so that it can
7036 be reached without opening the lid (such as the Lenovo
7037 Yoga). On those machines, logind will now immediately
7038 re-suspend the machine if the power button has been
7039 accidentally pressed while the laptop was suspended and in a
7040 backpack or similar.
7042 * logind will now watch SW_DOCK switches and inhibit reaction
7043 to the lid switch if it is pressed. This means that logind
7044 will not suspend the machine anymore if the lid is closed
7045 and the system is docked, if the laptop supports SW_DOCK
7046 notifications via the input layer. Note that ACPI docking
7047 stations do not generate this currently. Also note that this
7048 logic is usually not fully sufficient and Desktop
7049 Environments should take a lid switch inhibitor lock when an
7050 external display is connected, as systemd will not watch
7053 * nspawn will now make use of the devices cgroup controller by
7054 default, and only permit creation of and access to the usual
7055 API device nodes like /dev/null or /dev/random, as well as
7056 access to (but not creation of) the pty devices.
7058 * We will now ship a default .network file for
7059 systemd-networkd that automatically configures DHCP for
7060 network interfaces created by nspawn's --network-veth or
7061 --network-bridge= switches.
7063 * systemd will now understand the usual M, K, G, T suffixes
7064 according to SI conventions (i.e. to the base 1000) when
7065 referring to throughput and hardware metrics. It will stay
7066 with IEC conventions (i.e. to the base 1024) for software
7067 metrics, according to what is customary according to
7068 Wikipedia. We explicitly document which base applies for
7069 each configuration option.
7071 * The DeviceAllow= setting in unit files now supports a syntax
7072 to whitelist an entire group of devices node majors at once,
7073 based on the /proc/devices listing. For example, with the
7074 string "char-pts", it is now possible to whitelist all
7075 current and future pseudo-TTYs at once.
7077 * sd-event learned a new "post" event source. Event sources of
7078 this type are triggered by the dispatching of any event
7079 source of a type that is not "post". This is useful for
7080 implementing clean-up and check event sources that are
7081 triggered by other work being done in the program.
7083 * systemd-networkd is no longer statically enabled, but uses
7084 the usual [Install] sections so that it can be
7085 enabled/disabled using systemctl. It still is enabled by
7088 * When creating a veth interface pair with systemd-nspawn, the
7089 host side will now be prefixed with "vb-" if
7090 --network-bridge= is used, and with "ve-" if --network-veth
7091 is used. This way, it is easy to distinguish these cases on
7092 the host, for example to apply different configuration to
7093 them with systemd-networkd.
7095 * The compatibility libraries for libsystemd-journal.so,
7096 libsystem-id128.so, libsystemd-login.so and
7097 libsystemd-daemon.so do not make use of IFUNC
7098 anymore. Instead, we now build libsystemd.so multiple times
7099 under these alternative names. This means that the footprint
7100 is drastically increased, but given that these are
7101 transitional compatibility libraries, this should not matter
7102 much. This change has been made necessary to support the ARM
7103 platform for these compatibility libraries, as the ARM
7104 toolchain is not really at the same level as the toolchain
7105 for other architectures like x86 and does not support
7106 IFUNC. Please make sure to use --enable-compat-libs only
7107 during a transitional period!
7109 * The .include syntax has been deprecated and is not documented
7110 anymore. Drop-in files in .d directories should be used instead.
7112 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
7113 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
7114 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
7115 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
7116 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
7117 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
7118 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
7119 Zbigniew Jędrzejewski-Szmek
7121 — Berlin, 2014-02-24
7125 * A new component "systemd-networkd" has been added that can
7126 be used to configure local network interfaces statically or
7127 via DHCP. It is capable of bringing up bridges, VLANs, and
7128 bonding. Currently, no hook-ups for interactive network
7129 configuration are provided. Use this for your initrd,
7130 container, embedded, or server setup if you need a simple,
7131 yet powerful, network configuration solution. This
7132 configuration subsystem is quite nifty, as it allows wildcard
7133 hotplug matching in interfaces. For example, with a single
7134 configuration snippet, you can configure that all Ethernet
7135 interfaces showing up are automatically added to a bridge,
7136 or similar. It supports link-sensing and more.
7138 * A new tool "systemd-socket-proxyd" has been added which can
7139 act as a bidirectional proxy for TCP sockets. This is
7140 useful for adding socket activation support to services that
7141 do not actually support socket activation, including virtual
7142 machines and the like.
7144 * Add a new tool to save/restore rfkill state on
7147 * Save/restore state of keyboard backlights in addition to
7148 display backlights on shutdown/boot.
7150 * udev learned a new SECLABEL{} construct to label device
7151 nodes with a specific security label when they appear. For
7152 now, only SECLABEL{selinux} is supported, but the syntax is
7153 prepared for additional security frameworks.
7155 * udev gained a new scheme to configure link-level attributes
7156 from files in /etc/systemd/network/*.link. These files can
7157 match against MAC address, device path, driver name and type,
7158 and will apply attributes like the naming policy, link speed,
7159 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
7160 address assignment policy (randomized, ...).
7162 * The configuration of network interface naming rules for
7163 "permanent interface names" has changed: a new NamePolicy=
7164 setting in the [Link] section of .link files determines the
7165 priority of possible naming schemes (onboard, slot, MAC,
7166 path). The default value of this setting is determined by
7167 /usr/lib/net/links/99-default.link. Old
7168 80-net-name-slot.rules udev configuration file has been
7169 removed, so local configuration overriding this file should
7170 be adapted to override 99-default.link instead.
7172 * When the User= switch is used in a unit file, also
7173 initialize $SHELL= based on the user database entry.
7175 * systemd no longer depends on libdbus. All communication is
7176 now done with sd-bus, systemd's low-level bus library
7179 * kdbus support has been added to PID 1 itself. When kdbus is
7180 enabled, this causes PID 1 to set up the system bus and
7181 enable support for a new ".busname" unit type that
7182 encapsulates bus name activation on kdbus. It works a little
7183 bit like ".socket" units, except for bus names. A new
7184 generator has been added that converts classic dbus1 service
7185 activation files automatically into native systemd .busname
7188 * sd-bus: add a light-weight vtable implementation that allows
7189 defining objects on the bus with a simple static const
7190 vtable array of its methods, signals and properties.
7192 * systemd will not generate or install static dbus
7193 introspection data anymore to /usr/share/dbus-1/interfaces,
7194 as the precise format of these files is unclear, and
7195 nothing makes use of it.
7197 * A proxy daemon is now provided to proxy clients connecting
7198 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
7199 compatibility with classic D-Bus.
7201 * A bus driver implementation has been added that supports the
7202 classic D-Bus bus driver calls on kdbus, also for
7203 compatibility purposes.
7205 * A new API "sd-event.h" has been added that implements a
7206 minimal event loop API built around epoll. It provides a
7207 couple of features that direct epoll usage is lacking:
7208 prioritization of events, scales to large numbers of timer
7209 events, per-event timer slack (accuracy), system-wide
7210 coalescing of timer events, exit handlers, watchdog
7211 supervision support using systemd's sd_notify() API, child
7214 * A new API "sd-rntl.h" has been added that provides an API
7215 around the route netlink interface of the kernel, similar in
7216 style to "sd-bus.h".
7218 * A new API "sd-dhcp-client.h" has been added that provides a
7219 small DHCPv4 client-side implementation. This is used by
7222 * There is a new kernel command line option
7223 "systemd.restore_state=0|1". When set to "0", none of the
7224 systemd tools will restore saved runtime state to hardware
7225 devices. More specifically, the rfkill and backlight states
7228 * The FsckPassNo= compatibility option in mount/service units
7229 has been removed. The fstab generator will now add the
7230 necessary dependencies automatically, and does not require
7231 PID1's support for that anymore.
7233 * journalctl gained a new switch, --list-boots, that lists
7234 recent boots with their times and boot IDs.
7236 * The various tools like systemctl, loginctl, timedatectl,
7237 busctl, systemd-run, ... have gained a new switch "-M" to
7238 connect to a specific, local OS container (as direct
7239 connection, without requiring SSH). This works on any
7240 container that is registered with machined, such as those
7241 created by libvirt-lxc or nspawn.
7243 * systemd-run and systemd-analyze also gained support for "-H"
7244 to connect to remote hosts via SSH. This is particularly
7245 useful for systemd-run because it enables queuing of jobs
7246 onto remote systems.
7248 * machinectl gained a new command "login" to open a getty
7249 login in any local container. This works with any container
7250 that is registered with machined (such as those created by
7251 libvirt-lxc or nspawn), and which runs systemd inside.
7253 * machinectl gained a new "reboot" command that may be used to
7254 trigger a reboot on a specific container that is registered
7255 with machined. This works on any container that runs an init
7256 system of some kind.
7258 * systemctl gained a new "list-timers" command to print a nice
7259 listing of installed timer units with the times they elapse
7262 * Alternative reboot() parameters may now be specified on the
7263 "systemctl reboot" command line and are passed to the
7264 reboot() system call.
7266 * systemctl gained a new --job-mode= switch to configure the
7267 mode to queue a job with. This is a more generic version of
7268 --fail, --irreversible, and --ignore-dependencies, which are
7269 still available but not advertised anymore.
7271 * /etc/systemd/system.conf gained new settings to configure
7272 various default timeouts of units, as well as the default
7273 start limit interval and burst. These may still be overridden
7276 * PID1 will now export on the bus profile data of the security
7277 policy upload process (such as the SELinux policy upload to
7280 * journald: when forwarding logs to the console, include
7281 timestamps (following the setting in
7282 /sys/module/printk/parameters/time).
7284 * OnCalendar= in timer units now understands the special
7285 strings "yearly" and "annually". (Both are equivalent)
7287 * The accuracy of timer units is now configurable with the new
7288 AccuracySec= setting. It defaults to 1min.
7290 * A new dependency type JoinsNamespaceOf= has been added that
7291 allows running two services within the same /tmp and network
7292 namespace, if PrivateNetwork= or PrivateTmp= are used.
7294 * A new command "cat" has been added to systemctl. It outputs
7295 the original unit file of a unit, and concatenates the
7296 contents of additional "drop-in" unit file snippets, so that
7297 the full configuration is shown.
7299 * systemctl now supports globbing on the various "list-xyz"
7300 commands, like "list-units" or "list-sockets", as well as on
7301 those commands which take multiple unit names.
7303 * journalctl's --unit= switch gained support for globbing.
7305 * All systemd daemons now make use of the watchdog logic so
7306 that systemd automatically notices when they hang.
7308 * If the $container_ttys environment variable is set,
7309 getty-generator will automatically spawn a getty for each
7310 listed tty. This is useful for container managers to request
7311 login gettys to be spawned on as many ttys as needed.
7313 * %h, %s, %U specifier support is not available anymore when
7314 used in unit files for PID 1. This is because NSS calls are
7315 not safe from PID 1. They stay available for --user
7316 instances of systemd, and as special case for the root user.
7318 * loginctl gained a new "--no-legend" switch to turn off output
7321 * The "sd-login.h" API gained three new calls:
7322 sd_session_is_remote(), sd_session_get_remote_user(),
7323 sd_session_get_remote_host() to query information about
7326 * The udev hardware database now also carries vendor/product
7327 information of SDIO devices.
7329 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
7330 determine whether watchdog notifications are requested by
7333 * Socket-activated per-connection services now include a
7334 short description of the connection parameters in the
7337 * tmpfiles gained a new "--boot" option. When this is not used,
7338 only lines where the command character is not suffixed with
7339 "!" are executed. When this option is specified, those
7340 options are executed too. This partitions tmpfiles
7341 directives into those that can be safely executed at any
7342 time, and those which should be run only at boot (for
7343 example, a line that creates /run/nologin).
7345 * A new API "sd-resolve.h" has been added which provides a simple
7346 asynchronous wrapper around glibc NSS host name resolution
7347 calls, such as getaddrinfo(). In contrast to glibc's
7348 getaddrinfo_a(), it does not use signals. In contrast to most
7349 other asynchronous name resolution libraries, this one does
7350 not reimplement DNS, but reuses NSS, so that alternate
7351 host name resolution systems continue to work, such as mDNS,
7352 LDAP, etc. This API is based on libasyncns, but it has been
7353 cleaned up for inclusion in systemd.
7355 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
7356 "sd-daemon.h" are no longer found in individual libraries
7357 libsystemd-journal.so, libsystemd-login.so,
7358 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
7359 merged them into a single library, libsystemd.so, which
7360 provides all symbols. The reason for this is cyclic
7361 dependencies, as these libraries tend to use each other's
7362 symbols. So far, we have managed to workaround that by linking
7363 a copy of a good part of our code into each of these
7364 libraries again and again, which, however, makes certain
7365 things hard to do, like sharing static variables. Also, it
7366 substantially increases footprint. With this change, there
7367 is only one library for the basic APIs systemd
7368 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
7369 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
7370 library as well, however are subject to the --enable-kdbus
7371 switch (see below). Note that "sd-dhcp-client.h" is not part
7372 of this library (this is because it only consumes, never
7373 provides, services of/to other APIs). To make the transition
7374 easy from the separate libraries to the unified one, we
7375 provide the --enable-compat-libs compile-time switch which
7376 will generate stub libraries that are compatible with the
7377 old ones but redirect all calls to the new one.
7379 * All of the kdbus logic and the new APIs "sd-bus.h",
7380 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
7381 and "sd-utf8.h" are compile-time optional via the
7382 "--enable-kdbus" switch, and they are not compiled in by
7383 default. To make use of kdbus, you have to explicitly enable
7384 the switch. Note however, that neither the kernel nor the
7385 userspace API for all of this is considered stable yet. We
7386 want to maintain the freedom to still change the APIs for
7387 now. By specifying this build-time switch, you acknowledge
7388 that you are aware of the instability of the current
7391 * Also, note that while kdbus is pretty much complete,
7392 it lacks one thing: proper policy support. This means you
7393 can build a fully working system with all features; however,
7394 it will be highly insecure. Policy support will be added in
7395 one of the next releases, at the same time that we will
7396 declare the APIs stable.
7398 * When the kernel command line argument "kdbus" is specified,
7399 systemd will automatically load the kdbus.ko kernel module. At
7400 this stage of development, it is only useful for testing kdbus
7401 and should not be used in production. Note: if "--enable-kdbus"
7402 is specified, and the kdbus.ko kernel module is available, and
7403 "kdbus" is added to the kernel command line, the entire system
7404 runs with kdbus instead of dbus-daemon, with the above mentioned
7405 problem of missing the system policy enforcement. Also a future
7406 version of kdbus.ko or a newer systemd will not be compatible with
7407 each other, and will unlikely be able to boot the machine if only
7408 one of them is updated.
7410 * systemctl gained a new "import-environment" command which
7411 uploads the caller's environment (or parts thereof) into the
7412 service manager so that it is inherited by services started
7413 by the manager. This is useful to upload variables like
7414 $DISPLAY into the user service manager.
7416 * A new PrivateDevices= switch has been added to service units
7417 which allows running a service with a namespaced /dev
7418 directory that does not contain any device nodes for
7419 physical devices. More specifically, it only includes devices
7420 such as /dev/null, /dev/urandom, and /dev/zero which are API
7423 * logind has been extended to support behaviour like VT
7424 switching on seats that do not support a VT. This makes
7425 multi-session available on seats that are not the first seat
7426 (seat0), and on systems where kernel support for VTs has
7427 been disabled at compile-time.
7429 * If a process holds a delay lock for system sleep or shutdown
7430 and fails to release it in time, we will now log its
7431 identity. This makes it easier to identify processes that
7432 cause slow suspends or power-offs.
7434 * When parsing /etc/crypttab, support for a new key-slot=
7435 option as supported by Debian is added. It allows indicating
7436 which LUKS slot to use on disk, speeding up key loading.
7438 * The sd_journal_sendv() API call has been checked and
7439 officially declared to be async-signal-safe so that it may
7440 be invoked from signal handlers for logging purposes.
7442 * Boot-time status output is now enabled automatically after a
7443 short timeout if boot does not progress, in order to give
7444 the user an indication what she or he is waiting for.
7446 * The boot-time output has been improved to show how much time
7447 remains until jobs expire.
7449 * The KillMode= switch in service units gained a new possible
7450 value "mixed". If set, and the unit is shut down, then the
7451 initial SIGTERM signal is sent only to the main daemon
7452 process, while the following SIGKILL signal is sent to
7453 all remaining processes of the service.
7455 * When a scope unit is registered, a new property "Controller"
7456 may be set. If set to a valid bus name, systemd will send a
7457 RequestStop() signal to this name when it would like to shut
7458 down the scope. This may be used to hook manager logic into
7459 the shutdown logic of scope units. Also, scope units may now
7460 be put in a special "abandoned" state, in which case the
7461 manager process which created them takes no further
7462 responsibilities for it.
7464 * When reading unit files, systemd will now verify
7465 the access mode of these files, and warn about certain
7466 suspicious combinations. This has been added to make it
7467 easier to track down packaging bugs where unit files are
7468 marked executable or world-writable.
7470 * systemd-nspawn gained a new "--setenv=" switch to set
7471 container-wide environment variables. The similar option in
7472 systemd-activate was renamed from "--environment=" to
7473 "--setenv=" for consistency.
7475 * systemd-nspawn has been updated to create a new kdbus domain
7476 for each container that is invoked, thus allowing each
7477 container to have its own set of system and user buses,
7478 independent of the host.
7480 * systemd-nspawn gained a new --drop-capability= switch to run
7481 the container with less capabilities than the default. Both
7482 --drop-capability= and --capability= now take the special
7483 string "all" for dropping or keeping all capabilities.
7485 * systemd-nspawn gained new switches for executing containers
7486 with specific SELinux labels set.
7488 * systemd-nspawn gained a new --quiet switch to not generate
7489 any additional output but the container's own console
7492 * systemd-nspawn gained a new --share-system switch to run a
7493 container without PID namespacing enabled.
7495 * systemd-nspawn gained a new --register= switch to control
7496 whether the container is registered with systemd-machined or
7497 not. This is useful for containers that do not run full
7498 OS images, but only specific apps.
7500 * systemd-nspawn gained a new --keep-unit which may be used
7501 when invoked as the only program from a service unit, and
7502 results in registration of the unit service itself in
7503 systemd-machined, instead of a newly opened scope unit.
7505 * systemd-nspawn gained a new --network-interface= switch for
7506 moving arbitrary interfaces to the container. The new
7507 --network-veth switch creates a virtual Ethernet connection
7508 between host and container. The new --network-bridge=
7509 switch then allows assigning the host side of this virtual
7510 Ethernet connection to a bridge device.
7512 * systemd-nspawn gained a new --personality= switch for
7513 setting the kernel personality for the container. This is
7514 useful when running a 32-bit container on a 64-bit host. A
7515 similar option Personality= is now also available for service
7518 * logind will now also track a "Desktop" identifier for each
7519 session which encodes the desktop environment of it. This is
7520 useful for desktop environments that want to identify
7521 multiple running sessions of itself easily.
7523 * A new SELinuxContext= setting for service units has been
7524 added that allows setting a specific SELinux execution
7525 context for a service.
7527 * Most systemd client tools will now honour $SYSTEMD_LESS for
7528 settings of the "less" pager. By default, these tools will
7529 override $LESS to allow certain operations to work, such as
7530 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
7531 influence this logic.
7533 * systemd's "seccomp" hook-up has been changed to make use of
7534 the libseccomp library instead of using its own
7535 implementation. This has benefits for portability among
7538 * For usage together with SystemCallFilter=, a new
7539 SystemCallErrorNumber= setting has been introduced that
7540 allows configuration of a system error number to be returned
7541 on filtered system calls, instead of immediately killing the
7542 process. Also, SystemCallArchitectures= has been added to
7543 limit access to system calls of a particular architecture
7544 (in order to turn off support for unused secondary
7545 architectures). There is also a global
7546 SystemCallArchitectures= setting in system.conf now to turn
7547 off support for non-native system calls system-wide.
7549 * systemd requires a kernel with a working name_to_handle_at(),
7550 please see the kernel config requirements in the README file.
7552 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
7553 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
7554 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
7555 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
7556 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
7557 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
7558 Elia Pinto, Florian Weimer, George McCollister, Goffredo
7559 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
7560 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
7561 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
7562 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
7563 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
7564 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
7565 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
7566 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
7567 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
7568 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
7569 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
7570 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
7571 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
7572 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
7573 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
7574 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
7575 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
7577 — Berlin, 2014-02-20
7581 * logind has gained support for facilitating privileged input
7582 and drm device access for unprivileged clients. This work is
7583 useful to allow Wayland display servers (and similar
7584 programs, such as kmscon) to run under the user's ID and
7585 access input and drm devices which are normally
7586 protected. When this is used (and the kernel is new enough)
7587 logind will "mute" IO on the file descriptors passed to
7588 Wayland as long as it is in the background and "unmute" it
7589 if it returns into the foreground. This allows secure
7590 session switching without allowing background sessions to
7591 eavesdrop on input and display data. This also introduces
7592 session switching support if VT support is turned off in the
7593 kernel, and on seats that are not seat0.
7595 * A new kernel command line option luks.options= is understood
7596 now which allows specifying LUKS options for usage for LUKS
7597 encrypted partitions specified with luks.uuid=.
7599 * tmpfiles.d(5) snippets may now use specifier expansion in
7600 path names. More specifically %m, %b, %H, %v, are now
7601 replaced by the local machine id, boot id, hostname, and
7602 kernel version number.
7604 * A new tmpfiles.d(5) command "m" has been introduced which
7605 may be used to change the owner/group/access mode of a file
7606 or directory if it exists, but do nothing if it does not.
7608 * This release removes high-level support for the
7609 MemorySoftLimit= cgroup setting. The underlying kernel
7610 cgroup attribute memory.soft_limit= is currently badly
7611 designed and likely to be removed from the kernel API in its
7612 current form, hence we should not expose it for now.
7614 * The memory.use_hierarchy cgroup attribute is now enabled for
7615 all cgroups systemd creates in the memory cgroup
7616 hierarchy. This option is likely to be come the built-in
7617 default in the kernel anyway, and the non-hierarchical mode
7618 never made much sense in the intrinsically hierarchical
7621 * A new field _SYSTEMD_SLICE= is logged along with all journal
7622 messages containing the slice a message was generated
7623 from. This is useful to allow easy per-customer filtering of
7624 logs among other things.
7626 * systemd-journald will no longer adjust the group of journal
7627 files it creates to the "systemd-journal" group. Instead we
7628 rely on the journal directory to be owned by the
7629 "systemd-journal" group, and its setgid bit set, so that the
7630 kernel file system layer will automatically enforce that
7631 journal files inherit this group assignment. The reason for
7632 this change is that we cannot allow NSS look-ups from
7633 journald which would be necessary to resolve
7634 "systemd-journal" to a numeric GID, because this might
7635 create deadlocks if NSS involves synchronous queries to
7636 other daemons (such as nscd, or sssd) which in turn are
7637 logging clients of journald and might block on it, which
7638 would then dead lock. A tmpfiles.d(5) snippet included in
7639 systemd will make sure the setgid bit and group are
7640 properly set on the journal directory if it exists on every
7641 boot. However, we recommend adjusting it manually after
7642 upgrades too (or from RPM scriptlets), so that the change is
7643 not delayed until next reboot.
7645 * Backlight and random seed files in /var/lib/ have moved into
7646 the /var/lib/systemd/ directory, in order to centralize all
7647 systemd generated files in one directory.
7649 * Boot time performance measurements (as displayed by
7650 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
7651 performance information if that's available to determine how
7652 much time BIOS and boot loader initialization required. With
7653 a sufficiently new BIOS you hence no longer need to boot
7654 with Gummiboot to get access to such information.
7656 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
7657 Cristian Rodríguez, Dave Reisner, David Herrmann, David
7658 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
7659 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
7660 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
7661 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
7662 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
7664 — Berlin, 2013-10-02
7668 * The Restart= option for services now understands a new
7669 on-watchdog setting, which will restart the service
7670 automatically if the service stops sending out watchdog keep
7671 alive messages (as configured with WatchdogSec=).
7673 * The getty generator (which is responsible for bringing up a
7674 getty on configured serial consoles) will no longer only
7675 start a getty on the primary kernel console but on all
7676 others, too. This makes the order in which console= is
7677 specified on the kernel command line less important.
7679 * libsystemd-logind gained a new sd_session_get_vt() call to
7680 retrieve the VT number of a session.
7682 * If the option "tries=0" is set for an entry of /etc/crypttab
7683 its passphrase is queried indefinitely instead of any
7684 maximum number of tries.
7686 * If a service with a configure PID file terminates its PID
7687 file will now be removed automatically if it still exists
7688 afterwards. This should put an end to stale PID files.
7690 * systemd-run will now also take relative binary path names
7691 for execution and no longer insists on absolute paths.
7693 * InaccessibleDirectories= and ReadOnlyDirectories= now take
7694 paths that are optionally prefixed with "-" to indicate that
7695 it should not be considered a failure if they do not exist.
7697 * journalctl -o (and similar commands) now understands a new
7698 output mode "short-precise", it is similar to "short" but
7699 shows timestamps with usec accuracy.
7701 * The option "discard" (as known from Debian) is now
7702 synonymous to "allow-discards" in /etc/crypttab. In fact,
7703 "discard" is preferred now (since it is easier to remember
7706 * Some licensing clean-ups were made, so that more code is now
7707 LGPL-2.1 licensed than before.
7709 * A minimal tool to save/restore the display backlight
7710 brightness across reboots has been added. It will store the
7711 backlight setting as late as possible at shutdown, and
7712 restore it as early as possible during reboot.
7714 * A logic to automatically discover and enable home and swap
7715 partitions on GPT disks has been added. With this in place
7716 /etc/fstab becomes optional for many setups as systemd can
7717 discover certain partitions located on the root disk
7718 automatically. Home partitions are recognized under their
7719 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
7720 partitions are recognized under their GPT type ID
7721 0657fd6da4ab43c484e50933c84b4f4f.
7723 * systemd will no longer pass any environment from the kernel
7724 or initrd to system services. If you want to set an
7725 environment for all services, do so via the kernel command
7726 line systemd.setenv= assignment.
7728 * The systemd-sysctl tool no longer natively reads the file
7729 /etc/sysctl.conf. If desired, the file should be symlinked
7730 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
7731 legacy support by a symlink rather than built-in code, it
7732 also makes the otherwise hidden order of application of the
7733 different files visible. (Note that this partly reverts to a
7734 pre-198 application order of sysctl knobs!)
7736 * The "systemctl set-log-level" and "systemctl dump" commands
7737 have been moved to systemd-analyze.
7739 * systemd-run learned the new --remain-after-exit switch,
7740 which causes the scope unit not to be cleaned up
7741 automatically after the process terminated.
7743 * tmpfiles learned a new --exclude-prefix= switch to exclude
7744 certain paths from operation.
7746 * journald will now automatically flush all messages to disk
7747 as soon as a message at the log level CRIT, ALERT or EMERG
7750 Contributions from: Andrew Cook, Brandon Philips, Christian
7751 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
7752 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
7753 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
7754 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
7755 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
7756 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
7757 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
7758 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
7759 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
7760 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
7761 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
7762 William Giokas, Zbigniew Jędrzejewski-Szmek
7764 — Berlin, 2013-09-13
7768 * The documentation has been updated to cover the various new
7769 concepts introduced with 205.
7771 * Unit files now understand the new %v specifier which
7772 resolves to the kernel version string as returned by "uname
7775 * systemctl now supports filtering the unit list output by
7776 load state, active state and sub state, using the new
7779 * "systemctl status" will now show the results of the
7780 condition checks (like ConditionPathExists= and similar) of
7781 the last start attempts of the unit. They are also logged to
7784 * "journalctl -b" may now be used to look for boot output of a
7785 specific boot. Try "journalctl -b -1" for the previous boot,
7786 but the syntax is substantially more powerful.
7788 * "journalctl --show-cursor" has been added which prints the
7789 cursor string the last shown log line. This may then be used
7790 with the new "journalctl --after-cursor=" switch to continue
7791 browsing logs from that point on.
7793 * "journalctl --force" may now be used to force regeneration
7796 * Creation of "dead" device nodes has been moved from udev
7797 into kmod and tmpfiles. Previously, udev would read the kmod
7798 databases to pre-generate dead device nodes based on meta
7799 information contained in kernel modules, so that these would
7800 be auto-loaded on access rather then at boot. As this
7801 does not really have much to do with the exposing actual
7802 kernel devices to userspace this has always been slightly
7803 alien in the udev codebase. Following the new scheme kmod
7804 will now generate a runtime snippet for tmpfiles from the
7805 module meta information and it now is tmpfiles' job to the
7806 create the nodes. This also allows overriding access and
7807 other parameters for the nodes using the usual tmpfiles
7808 facilities. As side effect this allows us to remove the
7809 CAP_SYS_MKNOD capability bit from udevd entirely.
7811 * logind's device ACLs may now be applied to these "dead"
7812 devices nodes too, thus finally allowing managed access to
7813 devices such as /dev/snd/sequencer without loading the
7814 backing module right-away.
7816 * A new RPM macro has been added that may be used to apply
7817 tmpfiles configuration during package installation.
7819 * systemd-detect-virt and ConditionVirtualization= now can
7820 detect User-Mode-Linux machines (UML).
7822 * journald will now implicitly log the effective capabilities
7823 set of processes in the message metadata.
7825 * systemd-cryptsetup has gained support for TrueCrypt volumes.
7827 * The initrd interface has been simplified (more specifically,
7828 support for passing performance data via environment
7829 variables and fsck results via files in /run has been
7830 removed). These features were non-essential, and are
7831 nowadays available in a much nicer way by having systemd in
7832 the initrd serialize its state and have the hosts systemd
7833 deserialize it again.
7835 * The udev "keymap" data files and tools to apply keyboard
7836 specific mappings of scan to key codes, and force-release
7837 scan code lists have been entirely replaced by a udev
7838 "keyboard" builtin and a hwdb data file.
7840 * systemd will now honour the kernel's "quiet" command line
7841 argument also during late shutdown, resulting in a
7842 completely silent shutdown when used.
7844 * There's now an option to control the SO_REUSEPORT socket
7845 option in .socket units.
7847 * Instance units will now automatically get a per-template
7848 subslice of system.slice unless something else is explicitly
7849 configured. For example, instances of sshd@.service will now
7850 implicitly be placed in system-sshd.slice rather than
7851 system.slice as before.
7853 * Test coverage support may now be enabled at build time.
7855 Contributions from: Dave Reisner, Frederic Crozat, Harald
7856 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
7857 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
7858 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
7859 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
7860 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
7861 Giokas, Zbigniew Jędrzejewski-Szmek
7863 — Berlin, 2013-07-23
7867 * Two new unit types have been introduced:
7869 Scope units are very similar to service units, however, are
7870 created out of pre-existing processes — instead of PID 1
7871 forking off the processes. By using scope units it is
7872 possible for system services and applications to group their
7873 own child processes (worker processes) in a powerful way
7874 which then maybe used to organize them, or kill them
7875 together, or apply resource limits on them.
7877 Slice units may be used to partition system resources in an
7878 hierarchical fashion and then assign other units to them. By
7879 default there are now three slices: system.slice (for all
7880 system services), user.slice (for all user sessions),
7881 machine.slice (for VMs and containers).
7883 Slices and scopes have been introduced primarily in
7884 context of the work to move cgroup handling to a
7885 single-writer scheme, where only PID 1
7886 creates/removes/manages cgroups.
7888 * There's a new concept of "transient" units. In contrast to
7889 normal units these units are created via an API at runtime,
7890 not from configuration from disk. More specifically this
7891 means it is now possible to run arbitrary programs as
7892 independent services, with all execution parameters passed
7893 in via bus APIs rather than read from disk. Transient units
7894 make systemd substantially more dynamic then it ever was,
7895 and useful as a general batch manager.
7897 * logind has been updated to make use of scope and slice units
7898 for managing user sessions. As a user logs in he will get
7899 his own private slice unit, to which all sessions are added
7900 as scope units. We also added support for automatically
7901 adding an instance of user@.service for the user into the
7902 slice. Effectively logind will no longer create cgroup
7903 hierarchies on its own now, it will defer entirely to PID 1
7904 for this by means of scope, service and slice units. Since
7905 user sessions this way become entities managed by PID 1
7906 the output of "systemctl" is now a lot more comprehensive.
7908 * A new mini-daemon "systemd-machined" has been added which
7909 may be used by virtualization managers to register local
7910 VMs/containers. nspawn has been updated accordingly, and
7911 libvirt will be updated shortly. machined will collect a bit
7912 of meta information about the VMs/containers, and assign
7913 them their own scope unit (see above). The collected
7914 meta-data is then made available via the "machinectl" tool,
7915 and exposed in "ps" and similar tools. machined/machinectl
7916 is compile-time optional.
7918 * As discussed earlier, the low-level cgroup configuration
7919 options ControlGroup=, ControlGroupModify=,
7920 ControlGroupPersistent=, ControlGroupAttribute= have been
7921 removed. Please use high-level attribute settings instead as
7922 well as slice units.
7924 * A new bus call SetUnitProperties() has been added to alter
7925 various runtime parameters of a unit. This is primarily
7926 useful to alter cgroup parameters dynamically in a nice way,
7927 but will be extended later on to make more properties
7928 modifiable at runtime. systemctl gained a new set-properties
7929 command that wraps this call.
7931 * A new tool "systemd-run" has been added which can be used to
7932 run arbitrary command lines as transient services or scopes,
7933 while configuring a number of settings via the command
7934 line. This tool is currently very basic, however already
7935 very useful. We plan to extend this tool to even allow
7936 queuing of execution jobs with time triggers from the
7937 command line, similar in fashion to "at".
7939 * nspawn will now inform the user explicitly that kernels with
7940 audit enabled break containers, and suggest the user to turn
7943 * Support for detecting the IMA and AppArmor security
7944 frameworks with ConditionSecurity= has been added.
7946 * journalctl gained a new "-k" switch for showing only kernel
7947 messages, mimicking dmesg output; in addition to "--user"
7948 and "--system" switches for showing only user's own logs
7951 * systemd-delta can now show information about drop-in
7952 snippets extending unit files.
7954 * libsystemd-bus has been substantially updated but is still
7955 not available as public API.
7957 * systemd will now look for the "debug" argument on the kernel
7958 command line and enable debug logging, similar to what
7959 "systemd.log_level=debug" already did before.
7961 * "systemctl set-default", "systemctl get-default" has been
7962 added to configure the default.target symlink, which
7963 controls what to boot into by default.
7965 * "systemctl set-log-level" has been added as a convenient
7966 way to raise and lower systemd logging threshold.
7968 * "systemd-analyze plot" will now show the time the various
7969 generators needed for execution, as well as information
7970 about the unit file loading.
7972 * libsystemd-journal gained a new sd_journal_open_files() call
7973 for opening specific journal files. journactl also gained a
7974 new switch to expose this new functionality. Previously we
7975 only supported opening all files from a directory, or all
7976 files from the system, as opening individual files only is
7977 racy due to journal file rotation.
7979 * systemd gained the new DefaultEnvironment= setting in
7980 /etc/systemd/system.conf to set environment variables for
7983 * If a privileged process logs a journal message with the
7984 OBJECT_PID= field set, then journald will automatically
7985 augment this with additional OBJECT_UID=, OBJECT_GID=,
7986 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
7987 system services want to log events about specific client
7988 processes. journactl/systemctl has been updated to make use
7989 of this information if all log messages regarding a specific
7992 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
7993 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
7994 Reisner, David Coppa, David King, David Strauss, Eelco
7995 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
7996 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
7997 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
7998 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
7999 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
8000 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
8001 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
8002 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
8003 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
8004 Łukasz Stelmach, 장동준
8008 * The Python bindings gained some minimal support for the APIs
8009 exposed by libsystemd-logind.
8011 * ConditionSecurity= gained support for detecting SMACK. Since
8012 this condition already supports SELinux and AppArmor we only
8013 miss IMA for this. Patches welcome!
8015 Contributions from: Karol Lewandowski, Lennart Poettering,
8016 Zbigniew Jędrzejewski-Szmek
8020 * systemd-nspawn will now create /etc/resolv.conf if
8021 necessary, before bind-mounting the host's file onto it.
8023 * systemd-nspawn will now store meta information about a
8024 container on the container's cgroup as extended attribute
8025 fields, including the root directory.
8027 * The cgroup hierarchy has been reworked in many ways. All
8028 objects any of the components systemd creates in the cgroup
8029 tree are now suffixed. More specifically, user sessions are
8030 now placed in cgroups suffixed with ".session", users in
8031 cgroups suffixed with ".user", and nspawn containers in
8032 cgroups suffixed with ".nspawn". Furthermore, all cgroup
8033 names are now escaped in a simple scheme to avoid collision
8034 of userspace object names with kernel filenames. This work
8035 is preparation for making these objects relocatable in the
8036 cgroup tree, in order to allow easy resource partitioning of
8037 these objects without causing naming conflicts.
8039 * systemctl list-dependencies gained the new switches
8040 --plain, --reverse, --after and --before.
8042 * systemd-inhibit now shows the process name of processes that
8043 have taken an inhibitor lock.
8045 * nss-myhostname will now also resolve "localhost"
8046 implicitly. This makes /etc/hosts an optional file and
8047 nicely handles that on IPv6 ::1 maps to both "localhost" and
8050 * libsystemd-logind.so gained a new call
8051 sd_get_machine_names() to enumerate running containers and
8052 VMs (currently only supported by very new libvirt and
8053 nspawn). sd_login_monitor can now be used to watch
8054 VMs/containers coming and going.
8056 * .include is not allowed recursively anymore, and only in
8057 unit files. Usually it is better to use drop-in snippets in
8058 .d/*.conf anyway, as introduced with systemd 198.
8060 * systemd-analyze gained a new "critical-chain" command that
8061 determines the slowest chain of units run during system
8062 boot-up. It is very useful for tracking down where
8063 optimizing boot time is the most beneficial.
8065 * systemd will no longer allow manipulating service paths in
8066 the name=systemd:/system cgroup tree using ControlGroup= in
8067 units. (But is still fine with it in all other dirs.)
8069 * There's a new systemd-nspawn@.service service file that may
8070 be used to easily run nspawn containers as system
8071 services. With the container's root directory in
8072 /var/lib/container/foobar it is now sufficient to run
8073 "systemctl start systemd-nspawn@foobar.service" to boot it.
8075 * systemd-cgls gained a new parameter "--machine" to list only
8076 the processes within a certain container.
8078 * ConditionSecurity= now can check for "apparmor". We still
8079 are lacking checks for SMACK and IMA for this condition
8080 check though. Patches welcome!
8082 * A new configuration file /etc/systemd/sleep.conf has been
8083 added that may be used to configure which kernel operation
8084 systemd is supposed to execute when "suspend", "hibernate"
8085 or "hybrid-sleep" is requested. This makes the new kernel
8086 "freeze" state accessible to the user.
8088 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
8089 the passed argument if applicable.
8091 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
8092 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
8093 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
8094 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
8095 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
8096 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
8097 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
8102 * The output of 'systemctl list-jobs' got some polishing. The
8103 '--type=' argument may now be passed more than once. A new
8104 command 'systemctl list-sockets' has been added which shows
8105 a list of kernel sockets systemd is listening on with the
8106 socket units they belong to, plus the units these socket
8109 * The experimental libsystemd-bus library got substantial
8110 updates to work in conjunction with the (also experimental)
8111 kdbus kernel project. It works well enough to exchange
8112 messages with some sophistication. Note that kdbus is not
8113 ready yet, and the library is mostly an elaborate test case
8114 for now, and not installable.
8116 * systemd gained a new unit 'systemd-static-nodes.service'
8117 that generates static device nodes earlier during boot, and
8118 can run in conjunction with udev.
8120 * libsystemd-login gained a new call sd_pid_get_user_unit()
8121 to retrieve the user systemd unit a process is running
8122 in. This is useful for systems where systemd is used as
8125 * systemd-nspawn now places all containers in the new /machine
8126 top-level cgroup directory in the name=systemd
8127 hierarchy. libvirt will soon do the same, so that we get a
8128 uniform separation of /system, /user and /machine for system
8129 services, user processes and containers/virtual
8130 machines. This new cgroup hierarchy is also useful to stick
8131 stable names to specific container instances, which can be
8132 recognized later this way (this name may be controlled
8133 via systemd-nspawn's new -M switch). libsystemd-login also
8134 gained a new call sd_pid_get_machine_name() to retrieve the
8135 name of the container/VM a specific process belongs to.
8137 * bootchart can now store its data in the journal.
8139 * libsystemd-journal gained a new call
8140 sd_journal_add_conjunction() for AND expressions to the
8141 matching logic. This can be used to express more complex
8142 logical expressions.
8144 * journactl can now take multiple --unit= and --user-unit=
8147 * The cryptsetup logic now understands the "luks.key=" kernel
8148 command line switch for specifying a file to read the
8149 decryption key from. Also, if a configured key file is not
8150 found the tool will now automatically fall back to prompting
8153 * Python systemd.journal module was updated to wrap recently
8154 added functions from libsystemd-journal. The interface was
8155 changed to bring the low level interface in s.j._Reader
8156 closer to the C API, and the high level interface in
8157 s.j.Reader was updated to wrap and convert all data about
8160 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
8161 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
8162 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
8163 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
8164 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
8165 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
8169 * journalctl --update-catalog now understands a new --root=
8170 option to operate on catalogs found in a different root
8173 * During shutdown after systemd has terminated all running
8174 services a final killing loop kills all remaining left-over
8175 processes. We will now print the name of these processes
8176 when we send SIGKILL to them, since this usually indicates a
8179 * If /etc/crypttab refers to password files stored on
8180 configured mount points automatic dependencies will now be
8181 generated to ensure the specific mount is established first
8182 before the key file is attempted to be read.
8184 * 'systemctl status' will now show information about the
8185 network sockets a socket unit is listening on.
8187 * 'systemctl status' will also shown information about any
8188 drop-in configuration file for units. (Drop-In configuration
8189 files in this context are files such as
8190 /etc/systemd/systemd/foobar.service.d/*.conf)
8192 * systemd-cgtop now optionally shows summed up CPU times of
8193 cgroups. Press '%' while running cgtop to switch between
8194 percentage and absolute mode. This is useful to determine
8195 which cgroups use up the most CPU time over the entire
8196 runtime of the system. systemd-cgtop has also been updated
8197 to be 'pipeable' for processing with further shell tools.
8199 * 'hostnamectl set-hostname' will now allow setting of FQDN
8202 * The formatting and parsing of time span values has been
8203 changed. The parser now understands fractional expressions
8204 such as "5.5h". The formatter will now output fractional
8205 expressions for all time spans under 1min, i.e. "5.123456s"
8206 rather than "5s 123ms 456us". For time spans under 1s
8207 millisecond values are shown, for those under 1ms
8208 microsecond values are shown. This should greatly improve
8209 all time-related output of systemd.
8211 * libsystemd-login and libsystemd-journal gained new
8212 functions for querying the poll() events mask and poll()
8213 timeout value for integration into arbitrary event
8216 * localectl gained the ability to list available X11 keymaps
8217 (models, layouts, variants, options).
8219 * 'systemd-analyze dot' gained the ability to filter for
8220 specific units via shell-style globs, to create smaller,
8221 more useful graphs. I.e. it is now possible to create simple
8222 graphs of all the dependencies between only target units, or
8223 of all units that Avahi has dependencies with.
8225 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
8226 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
8227 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
8228 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
8229 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
8230 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
8231 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
8235 * The boot-time readahead implementation for rotating media
8236 will now read the read-ahead data in multiple passes which
8237 consist of all read requests made in equidistant time
8238 intervals. This means instead of strictly reading read-ahead
8239 data in its physical order on disk we now try to find a
8240 middle ground between physical and access time order.
8242 * /etc/os-release files gained a new BUILD_ID= field for usage
8243 on operating systems that provide continuous builds of OS
8246 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
8247 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
8248 William Douglas, Zbigniew Jędrzejewski-Szmek
8252 * systemd-python gained an API exposing libsystemd-daemon.
8254 * The SMACK setup logic gained support for uploading CIPSO
8257 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
8258 ReadOnlyDirectories= and InaccessibleDirectories= has
8259 changed. The private /tmp and /var/tmp directories are now
8260 shared by all processes of a service (which means
8261 ExecStartPre= may now leave data in /tmp that ExecStart= of
8262 the same service can still access). When a service is
8263 stopped its temporary directories are immediately deleted
8264 (normal clean-up with tmpfiles is still done in addition to
8267 * By default, systemd will now set a couple of sysctl
8268 variables in the kernel: the safe sysrq options are turned
8269 on, IP route verification is turned on, and source routing
8270 disabled. The recently added hardlink and softlink
8271 protection of the kernel is turned on. These settings should
8272 be reasonably safe, and good defaults for all new systems.
8274 * The predictable network naming logic may now be turned off
8275 with a new kernel command line switch: net.ifnames=0.
8277 * A new libsystemd-bus module has been added that implements a
8278 pretty complete D-Bus client library. For details see:
8280 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
8282 * journald will now explicitly flush the journal files to disk
8283 at the latest 5min after each write. The file will then also
8284 be marked offline until the next write. This should increase
8285 reliability in case of a crash. The synchronization delay
8286 can be configured via SyncIntervalSec= in journald.conf.
8288 * There's a new remote-fs-setup.target unit that can be used
8289 to pull in specific services when at least one remote file
8290 system is to be mounted.
8292 * There are new targets timers.target and paths.target as
8293 canonical targets to pull user timer and path units in
8294 from. This complements sockets.target with a similar
8295 purpose for socket units.
8297 * libudev gained a new call udev_device_set_attribute_value()
8298 to set sysfs attributes of a device.
8300 * The udev daemon now sets the default number of worker
8301 processes executed in parallel based on the number of available
8302 CPUs instead of the amount of available RAM. This is supposed
8303 to provide a more reliable default and limit a too aggressive
8304 parallelism for setups with 1000s of devices connected.
8306 Contributions from: Auke Kok, Colin Walters, Cristian
8307 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
8308 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
8309 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
8310 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
8311 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
8312 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
8313 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
8314 Zbigniew Jędrzejewski-Szmek
8318 * Configuration of unit files may now be extended via drop-in
8319 files without having to edit/override the unit files
8320 themselves. More specifically, if the administrator wants to
8321 change one value for a service file foobar.service he can
8322 now do so by dropping in a configuration snippet into
8323 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
8324 will load all these snippets and apply them on top of the
8325 main unit configuration file, possibly extending or
8326 overriding its settings. Using these drop-in snippets is
8327 generally nicer than the two earlier options for changing
8328 unit files locally: copying the files from
8329 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
8330 them there; or creating a new file in /etc/systemd/system/
8331 that incorporates the original one via ".include". Drop-in
8332 snippets into these .d/ directories can be placed in any
8333 directory systemd looks for units in, and the usual
8334 overriding semantics between /usr/lib, /etc and /run apply
8337 * Most unit file settings which take lists of items can now be
8338 reset by assigning the empty string to them. For example,
8339 normally, settings such as Environment=FOO=BAR append a new
8340 environment variable assignment to the environment block,
8341 each time they are used. By assigning Environment= the empty
8342 string the environment block can be reset to empty. This is
8343 particularly useful with the .d/*.conf drop-in snippets
8344 mentioned above, since this adds the ability to reset list
8345 settings from vendor unit files via these drop-ins.
8347 * systemctl gained a new "list-dependencies" command for
8348 listing the dependencies of a unit recursively.
8350 * Inhibitors are now honored and listed by "systemctl
8351 suspend", "systemctl poweroff" (and similar) too, not only
8352 GNOME. These commands will also list active sessions by
8355 * Resource limits (as exposed by the various control group
8356 controllers) can now be controlled dynamically at runtime
8357 for all units. More specifically, you can now use a command
8358 like "systemctl set-cgroup-attr foobar.service cpu.shares
8359 2000" to alter the CPU shares a specific service gets. These
8360 settings are stored persistently on disk, and thus allow the
8361 administrator to easily adjust the resource usage of
8362 services with a few simple commands. This dynamic resource
8363 management logic is also available to other programs via the
8364 bus. Almost any kernel cgroup attribute and controller is
8367 * systemd-vconsole-setup will now copy all font settings to
8368 all allocated VTs, where it previously applied them only to
8371 * libsystemd-login gained the new sd_session_get_tty() API
8374 * This release drops support for a few legacy or
8375 distribution-specific LSB facility names when parsing init
8376 scripts: $x-display-manager, $mail-transfer-agent,
8377 $mail-transport-agent, $mail-transfer-agent, $smtp,
8378 $null. Also, the mail-transfer-agent.target unit backing
8379 this has been removed. Distributions which want to retain
8380 compatibility with this should carry the burden for
8381 supporting this themselves and patch support for these back
8382 in, if they really need to. Also, the facilities $syslog and
8383 $local_fs are now ignored, since systemd does not support
8384 early-boot LSB init scripts anymore, and these facilities
8385 are implied anyway for normal services. syslog.target has
8388 * There are new bus calls on PID1's Manager object for
8389 cancelling jobs, and removing snapshot units. Previously,
8390 both calls were only available on the Job and Snapshot
8393 * systemd-journal-gatewayd gained SSL support.
8395 * The various "environment" files, such as /etc/locale.conf
8396 now support continuation lines with a backslash ("\") as
8397 last character in the line, similarly in style (but different)
8398 to how this is supported in shells.
8400 * For normal user processes the _SYSTEMD_USER_UNIT= field is
8401 now implicitly appended to every log entry logged. systemctl
8402 has been updated to filter by this field when operating on a
8403 user systemd instance.
8405 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
8406 CAP_AUDIT_CONTROL capabilities to the capabilities set for
8407 the container. This makes it easier to boot unmodified
8408 Fedora systems in a container, which however still requires
8409 audit=0 to be passed on the kernel command line. Auditing in
8410 kernel and userspace is unfortunately still too broken in
8411 context of containers, hence we recommend compiling it out
8412 of the kernel or using audit=0. Hopefully this will be fixed
8413 one day for good in the kernel.
8415 * nspawn gained the new --bind= and --bind-ro= parameters to
8416 bind mount specific directories from the host into the
8419 * nspawn will now mount its own devpts file system instance
8420 into the container, in order not to leak pty devices from
8421 the host into the container.
8423 * systemd will now read the firmware boot time performance
8424 information from the EFI variables, if the used boot loader
8425 supports this, and takes it into account for boot performance
8426 analysis via "systemd-analyze". This is currently supported
8427 only in conjunction with Gummiboot, but could be supported
8428 by other boot loaders too. For details see:
8430 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
8432 * A new generator has been added that automatically mounts the
8433 EFI System Partition (ESP) to /boot, if that directory
8434 exists, is empty, and no other file system has been
8435 configured to be mounted there.
8437 * logind will now send out PrepareForSleep(false) out
8438 unconditionally, after coming back from suspend. This may be
8439 used by applications as asynchronous notification for
8440 system resume events.
8442 * "systemctl unlock-sessions" has been added, that allows
8443 unlocking the screens of all user sessions at once, similar
8444 to how "systemctl lock-sessions" already locked all users
8445 sessions. This is backed by a new D-Bus call UnlockSessions().
8447 * "loginctl seat-status" will now show the master device of a
8448 seat. (i.e. the device of a seat that needs to be around for
8449 the seat to be considered available, usually the graphics
8452 * tmpfiles gained a new "X" line type, that allows
8453 configuration of files and directories (with wildcards) that
8454 shall be excluded from automatic cleanup ("aging").
8456 * udev default rules set the device node permissions now only
8457 at "add" events, and do not change them any longer with a
8458 later "change" event.
8460 * The log messages for lid events and power/sleep keypresses
8461 now carry a message ID.
8463 * We now have a substantially larger unit test suite, but this
8464 continues to be work in progress.
8466 * udevadm hwdb gained a new --root= parameter to change the
8467 root directory to operate relative to.
8469 * logind will now issue a background sync() request to the kernel
8470 early at shutdown, so that dirty buffers are flushed to disk early
8471 instead of at the last moment, in order to optimize shutdown
8474 * A new bootctl tool has been added that is an interface for
8475 certain boot loader operations. This is currently a preview
8476 and is likely to be extended into a small mechanism daemon
8477 like timedated, localed, hostnamed, and can be used by
8478 graphical UIs to enumerate available boot options, and
8479 request boot into firmware operations.
8481 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
8482 the rest of the package. It also has been updated to work
8483 correctly in initrds.
8485 * polkit previously has been runtime optional, and is now also
8486 compile time optional via a configure switch.
8488 * systemd-analyze has been reimplemented in C. Also "systemctl
8489 dot" has moved into systemd-analyze.
8491 * "systemctl status" with no further parameters will now print
8492 the status of all active or failed units.
8494 * Operations such as "systemctl start" can now be executed
8495 with a new mode "--irreversible" which may be used to queue
8496 operations that cannot accidentally be reversed by a later
8497 job queuing. This is by default used to make shutdown
8498 requests more robust.
8500 * The Python API of systemd now gained a new module for
8501 reading journal files.
8503 * A new tool kernel-install has been added that can install
8504 kernel images according to the Boot Loader Specification:
8506 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
8508 * Boot time console output has been improved to provide
8509 animated boot time output for hanging jobs.
8511 * A new tool systemd-activate has been added which can be used
8512 to test socket activation with, directly from the command
8513 line. This should make it much easier to test and debug
8514 socket activation in daemons.
8516 * journalctl gained a new "--reverse" (or -r) option to show
8517 journal output in reverse order (i.e. newest line first).
8519 * journalctl gained a new "--pager-end" (or -e) option to jump
8520 to immediately jump to the end of the journal in the
8521 pager. This is only supported in conjunction with "less".
8523 * journalctl gained a new "--user-unit=" option, that works
8524 similarly to "--unit=" but filters for user units rather than
8527 * A number of unit files to ease adoption of systemd in
8528 initrds has been added. This moves some minimal logic from
8529 the various initrd implementations into systemd proper.
8531 * The journal files are now owned by a new group
8532 "systemd-journal", which exists specifically to allow access
8533 to the journal, and nothing else. Previously, we used the
8534 "adm" group for that, which however possibly covers more
8535 than just journal/log file access. This new group is now
8536 already used by systemd-journal-gatewayd to ensure this
8537 daemon gets access to the journal files and as little else
8538 as possible. Note that "make install" will also set FS ACLs
8539 up for /var/log/journal to give "adm" and "wheel" read
8540 access to it, in addition to "systemd-journal" which owns
8541 the journal files. We recommend that packaging scripts also
8542 add read access to "adm" + "wheel" to /var/log/journal, and
8543 all existing/future journal files. To normal users and
8544 administrators little changes, however packagers need to
8545 ensure to create the "systemd-journal" system group at
8546 package installation time.
8548 * The systemd-journal-gatewayd now runs as unprivileged user
8549 systemd-journal-gateway:systemd-journal-gateway. Packaging
8550 scripts need to create these system user/group at
8553 * timedated now exposes a new boolean property CanNTP that
8554 indicates whether a local NTP service is available or not.
8556 * systemd-detect-virt will now also detect xen PVs
8558 * The pstore file system is now mounted by default, if it is
8561 * In addition to the SELinux and IMA policies we will now also
8562 load SMACK policies at early boot.
8564 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
8565 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
8566 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
8567 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
8568 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
8569 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
8570 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
8571 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
8572 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
8573 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
8574 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
8575 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
8576 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
8577 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
8581 * Timer units now support calendar time events in addition to
8582 monotonic time events. That means you can now trigger a unit
8583 based on a calendar time specification such as "Thu,Fri
8584 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
8585 or fifth day of any month of the year 2013, given that it is
8586 a thursday or friday. This brings timer event support
8587 considerably closer to cron's capabilities. For details on
8588 the supported calendar time specification language see
8591 * udev now supports a number of different naming policies for
8592 network interfaces for predictable names, and a combination
8593 of these policies is now the default. Please see this wiki
8594 document for details:
8596 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
8598 * Auke Kok's bootchart implementation has been added to the
8599 systemd tree. It is an optional component that can graph the
8600 boot in quite some detail. It is one of the best bootchart
8601 implementations around and minimal in its code and
8604 * nss-myhostname has been integrated into the systemd source
8605 tree. nss-myhostname guarantees that the local hostname
8606 always stays resolvable via NSS. It has been a weak
8607 requirement of systemd-hostnamed since a long time, and
8608 since its code is actually trivial we decided to just
8609 include it in systemd's source tree. It can be turned off
8610 with a configure switch.
8612 * The read-ahead logic is now capable of properly detecting
8613 whether a btrfs file system is on SSD or rotating media, in
8614 order to optimize the read-ahead scheme. Previously, it was
8615 only capable of detecting this on traditional file systems
8618 * In udev, additional device properties are now read from the
8619 IAB in addition to the OUI database. Also, Bluetooth company
8620 identities are attached to the devices as well.
8622 * In service files %U may be used as specifier that is
8623 replaced by the configured user name of the service.
8625 * nspawn may now be invoked without a controlling TTY. This
8626 makes it suitable for invocation as its own service. This
8627 may be used to set up a simple containerized server system
8628 using only core OS tools.
8630 * systemd and nspawn can now accept socket file descriptors
8631 when they are started for socket activation. This enables
8632 implementation of socket activated nspawn
8633 containers. i.e. think about autospawning an entire OS image
8634 when the first SSH or HTTP connection is received. We expect
8635 that similar functionality will also be added to libvirt-lxc
8638 * journalctl will now suppress ANSI color codes when
8639 presenting log data.
8641 * systemctl will no longer show control group information for
8642 a unit if the control group is empty anyway.
8644 * logind can now automatically suspend/hibernate/shutdown the
8647 * /etc/machine-info and hostnamed now also expose the chassis
8648 type of the system. This can be used to determine whether
8649 the local system is a laptop, desktop, handset or
8650 tablet. This information may either be configured by the
8651 user/vendor or is automatically determined from ACPI and DMI
8652 information if possible.
8654 * A number of polkit actions are now bound together with "imply"
8655 rules. This should simplify creating UIs because many actions
8656 will now authenticate similar ones as well.
8658 * Unit files learnt a new condition ConditionACPower= which
8659 may be used to conditionalize a unit depending on whether an
8660 AC power source is connected or not, of whether the system
8661 is running on battery power.
8663 * systemctl gained a new "is-failed" verb that may be used in
8664 shell scripts and suchlike to check whether a specific unit
8665 is in the "failed" state.
8667 * The EnvironmentFile= setting in unit files now supports file
8668 globbing, and can hence be used to easily read a number of
8669 environment files at once.
8671 * systemd will no longer detect and recognize specific
8672 distributions. All distribution-specific #ifdeffery has been
8673 removed, systemd is now fully generic and
8674 distribution-agnostic. Effectively, not too much is lost as
8675 a lot of the code is still accessible via explicit configure
8676 switches. However, support for some distribution specific
8677 legacy configuration file formats has been dropped. We
8678 recommend distributions to simply adopt the configuration
8679 files everybody else uses now and convert the old
8680 configuration from packaging scripts. Most distributions
8681 already did that. If that's not possible or desirable,
8682 distributions are welcome to forward port the specific
8683 pieces of code locally from the git history.
8685 * When logging a message about a unit systemd will now always
8686 log the unit name in the message meta data.
8688 * localectl will now also discover system locale data that is
8689 not stored in locale archives, but directly unpacked.
8691 * logind will no longer unconditionally use framebuffer
8692 devices as seat masters, i.e. as devices that are required
8693 to be existing before a seat is considered preset. Instead,
8694 it will now look for all devices that are tagged as
8695 "seat-master" in udev. By default, framebuffer devices will
8696 be marked as such, but depending on local systems, other
8697 devices might be marked as well. This may be used to
8698 integrate graphics cards using closed source drivers (such
8699 as NVidia ones) more nicely into logind. Note however, that
8700 we recommend using the open source NVidia drivers instead,
8701 and no udev rules for the closed-source drivers will be
8702 shipped from us upstream.
8704 Contributions from: Adam Williamson, Alessandro Crismani, Auke
8705 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
8706 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
8707 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
8708 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
8709 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
8710 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
8711 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
8712 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
8713 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
8714 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
8715 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
8720 * udev gained support for loading additional device properties
8721 from an indexed database that is keyed by vendor/product IDs
8722 and similar device identifiers. For the beginning this
8723 "hwdb" is populated with data from the well-known PCI and
8724 USB database, but also includes PNP, ACPI and OID data. In
8725 the longer run this indexed database shall grow into
8726 becoming the one central database for non-essential
8727 userspace device metadata. Previously, data from the PCI/USB
8728 database was only attached to select devices, since the
8729 lookup was a relatively expensive operation due to O(n) time
8730 complexity (with n being the number of entries in the
8731 database). Since this is now O(1), we decided to add in this
8732 data for all devices where this is available, by
8733 default. Note that the indexed database needs to be rebuilt
8734 when new data files are installed. To achieve this you need
8735 to update your packaging scripts to invoke "udevadm hwdb
8736 --update" after installation of hwdb data files. For
8737 RPM-based distributions we introduced the new
8738 %udev_hwdb_update macro for this purpose.
8740 * The Journal gained support for the "Message Catalog", an
8741 indexed database to link up additional information with
8742 journal entries. For further details please check:
8744 https://www.freedesktop.org/wiki/Software/systemd/catalog
8746 The indexed message catalog database also needs to be
8747 rebuilt after installation of message catalog files. Use
8748 "journalctl --update-catalog" for this. For RPM-based
8749 distributions we introduced the %journal_catalog_update
8750 macro for this purpose.
8752 * The Python Journal bindings gained support for the standard
8753 Python logging framework.
8755 * The Journal API gained new functions for checking whether
8756 the underlying file system of a journal file is capable of
8757 properly reporting file change notifications, or whether
8758 applications that want to reflect journal changes "live"
8759 need to recheck journal files continuously in appropriate
8762 * It is now possible to set the "age" field for tmpfiles
8763 entries to 0, indicating that files matching this entry
8764 shall always be removed when the directories are cleaned up.
8766 * coredumpctl gained a new "gdb" verb which invokes gdb
8767 right-away on the selected coredump.
8769 * There's now support for "hybrid sleep" on kernels that
8770 support this, in addition to "suspend" and "hibernate". Use
8771 "systemctl hybrid-sleep" to make use of this.
8773 * logind's HandleSuspendKey= setting (and related settings)
8774 now gained support for a new "lock" setting to simply
8775 request the screen lock on all local sessions, instead of
8776 actually executing a suspend or hibernation.
8778 * systemd will now mount the EFI variables file system by
8781 * Socket units now gained support for configuration of the
8782 SMACK security label.
8784 * timedatectl will now output the time of the last and next
8785 daylight saving change.
8787 * We dropped support for various legacy and distro-specific
8788 concepts, such as insserv, early-boot SysV services
8789 (i.e. those for non-standard runlevels such as 'b' or 'S')
8790 or ArchLinux /etc/rc.conf support. We recommend the
8791 distributions who still need support this to either continue
8792 to maintain the necessary patches downstream, or find a
8793 different solution. (Talk to us if you have questions!)
8795 * Various systemd components will now bypass polkit checks for
8796 root and otherwise handle properly if polkit is not found to
8797 be around. This should fix most issues for polkit-less
8798 systems. Quite frankly this should have been this way since
8799 day one. It is absolutely our intention to make systemd work
8800 fine on polkit-less systems, and we consider it a bug if
8801 something does not work as it should if polkit is not around.
8803 * For embedded systems it is now possible to build udev and
8804 systemd without blkid and/or kmod support.
8806 * "systemctl switch-root" is now capable of switching root
8807 more than once. I.e. in addition to transitions from the
8808 initrd to the host OS it is now possible to transition to
8809 further OS images from the host. This is useful to implement
8810 offline updating tools.
8812 * Various other additions have been made to the RPM macros
8813 shipped with systemd. Use %udev_rules_update() after
8814 installing new udev rules files. %_udevhwdbdir,
8815 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
8816 %_sysctldir are now available which resolve to the right
8817 directories for packages to place various data files in.
8819 * journalctl gained the new --full switch (in addition to
8820 --all, to disable ellipsation for long messages.
8822 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
8823 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
8824 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
8825 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
8826 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
8827 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
8828 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
8829 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
8830 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
8834 * journalctl gained new --since= and --until= switches to
8835 filter by time. It also now supports nice filtering for
8836 units via --unit=/-u.
8838 * Type=oneshot services may use ExecReload= and do the
8841 * The journal daemon now supports time-based rotation and
8842 vacuuming, in addition to the usual disk-space based
8845 * The journal will now index the available field values for
8846 each field name. This enables clients to show pretty drop
8847 downs of available match values when filtering. The bash
8848 completion of journalctl has been updated
8849 accordingly. journalctl gained a new switch -F to list all
8850 values a certain field takes in the journal database.
8852 * More service events are now written as structured messages
8853 to the journal, and made recognizable via message IDs.
8855 * The timedated, localed and hostnamed mini-services which
8856 previously only provided support for changing time, locale
8857 and hostname settings from graphical DEs such as GNOME now
8858 also have a minimal (but very useful) text-based client
8859 utility each. This is probably the nicest way to changing
8860 these settings from the command line now, especially since
8861 it lists available options and is fully integrated with bash
8864 * There's now a new tool "systemd-coredumpctl" to list and
8865 extract coredumps from the journal.
8867 * We now install a README each in /var/log/ and
8868 /etc/rc.d/init.d explaining where the system logs and init
8869 scripts went. This hopefully should help folks who go to
8870 that dirs and look into the otherwise now empty void and
8871 scratch their heads.
8873 * When user-services are invoked (by systemd --user) the
8874 $MANAGERPID env var is set to the PID of systemd.
8876 * SIGRTMIN+24 when sent to a --user instance will now result
8877 in immediate termination of systemd.
8879 * gatewayd received numerous feature additions such as a
8880 "follow" mode, for live syncing and filtering.
8882 * browse.html now allows filtering and showing detailed
8883 information on specific entries. Keyboard navigation and
8884 mouse screen support has been added.
8886 * gatewayd/journalctl now supports HTML5/JSON
8887 Server-Sent-Events as output.
8889 * The SysV init script compatibility logic will now
8890 heuristically determine whether a script supports the
8891 "reload" verb, and only then make this available as
8894 * "systemctl status --follow" has been removed, use "journalctl
8897 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
8898 have been removed since they are hardly useful to be
8901 * And I'd like to take the opportunity to specifically mention
8902 Zbigniew for his great contributions. Zbigniew, you rock!
8904 Contributions from: Andrew Eikum, Christian Hesse, Colin
8905 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
8906 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
8907 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
8908 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
8909 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
8910 Jędrzejewski-Szmek, Сковорода Никита Андреевич
8914 * If /etc/vconsole.conf is non-existent or empty we will no
8915 longer load any console font or key map at boot by
8916 default. Instead the kernel defaults will be left
8917 intact. This is definitely the right thing to do, as no
8918 configuration should mean no configuration, and hard-coding
8919 font names that are different on all archs is probably a bad
8920 idea. Also, the kernel default key map and font should be
8921 good enough for most cases anyway, and mostly identical to
8922 the userspace fonts/key maps we previously overloaded them
8923 with. If distributions want to continue to default to a
8924 non-kernel font or key map they should ship a default
8925 /etc/vconsole.conf with the appropriate contents.
8927 Contributions from: Colin Walters, Daniel J Walsh, Dave
8928 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
8929 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
8933 * journalctl gained a new --cursor= switch to show entries
8934 starting from the specified location in the journal.
8936 * We now enforce a size limit on journal entry fields exported
8937 with "-o json" in journalctl. Fields larger than 4K will be
8938 assigned null. This can be turned off with --all.
8940 * An (optional) journal gateway daemon is now available as
8941 "systemd-journal-gatewayd.service". This service provides
8942 access to the journal via HTTP and JSON. This functionality
8943 will be used to implement live log synchronization in both
8944 pull and push modes, but has various other users too, such
8945 as easy log access for debugging of embedded devices. Right
8946 now it is already useful to retrieve the journal via HTTP:
8948 # systemctl start systemd-journal-gatewayd.service
8949 # wget http://localhost:19531/entries
8951 This will download the journal contents in a
8952 /var/log/messages compatible format. The same as JSON:
8954 # curl -H"Accept: application/json" http://localhost:19531/entries
8956 This service is also accessible via a web browser where a
8957 single static HTML5 app is served that uses the JSON logic
8958 to enable the user to do some basic browsing of the
8959 journal. This will be extended later on. Here's an example
8960 screenshot of this app in its current state:
8962 http://0pointer.de/public/journal-gatewayd
8964 Contributions from: Kay Sievers, Lennart Poettering, Robert
8965 Milasan, Tom Gundersen
8969 * The bash completion logic is now available for journalctl
8972 * We do not mount the "cpuset" controller anymore together with
8973 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
8974 started if no parameters are assigned to it. "cpuset" hence
8975 broke code that assumed it could create "cpu" groups and
8978 * journalctl -f will now subscribe to terminal size changes,
8979 and line break accordingly.
8981 Contributions from: Dave Reisner, Kay Sievers, Lennart
8982 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
8986 * nspawn will now create a symlink /etc/localtime in the
8987 container environment, copying the host's timezone
8988 setting. Previously this has been done via a bind mount, but
8989 since symlinks cannot be bind mounted this has now been
8990 changed to create/update the appropriate symlink.
8992 * journalctl -n's line number argument is now optional, and
8993 will default to 10 if omitted.
8995 * journald will now log the maximum size the journal files may
8996 take up on disk. This is particularly useful if the default
8997 built-in logic of determining this parameter from the file
8998 system size is used. Use "systemctl status
8999 systemd-journald.service" to see this information.
9001 * The multi-seat X wrapper tool has been stripped down. As X
9002 is now capable of enumerating graphics devices via udev in a
9003 seat-aware way the wrapper is not strictly necessary
9004 anymore. A stripped down temporary stop-gap is still shipped
9005 until the upstream display managers have been updated to
9006 fully support the new X logic. Expect this wrapper to be
9007 removed entirely in one of the next releases.
9009 * HandleSleepKey= in logind.conf has been split up into
9010 HandleSuspendKey= and HandleHibernateKey=. The old setting
9011 is not available anymore. X11 and the kernel are
9012 distinguishing between these keys and we should too. This
9013 also means the inhibition lock for these keys has been split
9016 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
9017 Poettering, Lukas Nykryn, Václav Pavlín
9021 * Whenever a unit changes state we will now log this to the
9022 journal and show along the unit's own log output in
9025 * ConditionPathIsMountPoint= can now properly detect bind
9026 mount points too. (Previously, a bind mount of one file
9027 system to another place in the same file system could not be
9028 detected as mount, since they shared struct stat's st_dev
9031 * We will now mount the cgroup controllers cpu, cpuacct,
9032 cpuset and the controllers net_cls, net_prio together by
9035 * nspawn containers will now have a virtualized boot
9036 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
9037 over with a randomized ID at container initialization). This
9038 has the effect of making "journalctl -b" do the right thing
9041 * The JSON output journal serialization has been updated not
9042 to generate "endless" list objects anymore, but rather one
9043 JSON object per line. This is more in line how most JSON
9044 parsers expect JSON objects. The new output mode
9045 "json-pretty" has been added to provide similar output, but
9046 neatly aligned for readability by humans.
9048 * We dropped all explicit sync() invocations in the shutdown
9049 code. The kernel does this implicitly anyway in the kernel
9050 reboot() syscall. halt(8)'s -n option is now a compatibility
9053 * We now support virtualized reboot() in containers, as
9054 supported by newer kernels. We will fall back to exit() if
9055 CAP_SYS_REBOOT is not available to the container. Also,
9056 nspawn makes use of this now and will actually reboot the
9057 container if the containerized OS asks for that.
9059 * journalctl will only show local log output by default
9060 now. Use --merge (-m) to show remote log output, too.
9062 * libsystemd-journal gained the new sd_journal_get_usage()
9063 call to determine the current disk usage of all journal
9064 files. This is exposed in the new "journalctl --disk-usage"
9067 * journald gained a new configuration setting SplitMode= in
9068 journald.conf which may be used to control how user journals
9069 are split off. See journald.conf(5) for details.
9071 * A new condition type ConditionFileNotEmpty= has been added.
9073 * tmpfiles' "w" lines now support file globbing, to write
9074 multiple files at once.
9076 * We added Python bindings for the journal submission
9077 APIs. More Python APIs for a number of selected APIs will
9078 likely follow. Note that we intend to add native bindings
9079 only for the Python language, as we consider it common
9080 enough to deserve bindings shipped within systemd. There are
9081 various projects outside of systemd that provide bindings
9082 for languages such as PHP or Lua.
9084 * Many conditions will now resolve specifiers such as %i. In
9085 addition, PathChanged= and related directives of .path units
9086 now support specifiers as well.
9088 * There's now a new RPM macro definition for the system preset
9091 * journald will now warn if it ca not forward a message to the
9092 syslog daemon because its socket is full.
9094 * timedated will no longer write or process /etc/timezone,
9095 except on Debian. As we do not support late mounted /usr
9096 anymore /etc/localtime always being a symlink is now safe,
9097 and hence the information in /etc/timezone is not necessary
9100 * logind will now always reserve one VT for a text getty (VT6
9101 by default). Previously if more than 6 X sessions where
9102 started they took up all the VTs with auto-spawned gettys,
9103 so that no text gettys were available anymore.
9105 * udev will now automatically inform the btrfs kernel logic
9106 about btrfs RAID components showing up. This should make
9107 simple hotplug based btrfs RAID assembly work.
9109 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
9110 (but not for its children which will stay at the kernel
9111 default). This should allow setups with a lot more listening
9114 * systemd will now always pass the configured timezone to the
9115 kernel at boot. timedated will do the same when the timezone
9118 * logind's inhibition logic has been updated. By default,
9119 logind will now handle the lid switch, the power and sleep
9120 keys all the time, even in graphical sessions. If DEs want
9121 to handle these events on their own they should take the new
9122 handle-power-key, handle-sleep-key and handle-lid-switch
9123 inhibitors during their runtime. A simple way to achieve
9124 that is to invoke the DE wrapped in an invocation of:
9126 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
9128 * Access to unit operations is now checked via SELinux taking
9129 the unit file label and client process label into account.
9131 * systemd will now notify the administrator in the journal
9132 when he over-mounts a non-empty directory.
9134 * There are new specifiers that are resolved in unit files,
9135 for the host name (%H), the machine ID (%m) and the boot ID
9138 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
9139 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
9140 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
9141 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
9142 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
9143 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
9144 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
9148 * Support for reading structured kernel messages from
9149 /dev/kmsg has now been added and is enabled by default.
9151 * Support for reading kernel messages from /proc/kmsg has now
9152 been removed. If you want kernel messages in the journal
9153 make sure to run a recent kernel (>= 3.5) that supports
9154 reading structured messages from /dev/kmsg (see
9155 above). /proc/kmsg is now exclusive property of classic
9156 syslog daemons again.
9158 * The libudev API gained the new
9159 udev_device_new_from_device_id() call.
9161 * The logic for file system namespace (ReadOnlyDirectory=,
9162 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
9163 require pivot_root() anymore. This means fewer temporary
9164 directories are created below /tmp for this feature.
9166 * nspawn containers will now see and receive all submounts
9167 made on the host OS below the root file system of the
9170 * Forward Secure Sealing is now supported for Journal files,
9171 which provide cryptographical sealing of journal files so
9172 that attackers cannot alter log history anymore without this
9173 being detectable. Lennart will soon post a blog story about
9174 this explaining it in more detail.
9176 * There are two new service settings RestartPreventExitStatus=
9177 and SuccessExitStatus= which allow configuration of exit
9178 status (exit code or signal) which will be excepted from the
9179 restart logic, resp. consider successful.
9181 * journalctl gained the new --verify switch that can be used
9182 to check the integrity of the structure of journal files and
9183 (if Forward Secure Sealing is enabled) the contents of
9186 * nspawn containers will now be run with /dev/stdin, /dev/fd/
9187 and similar symlinks pre-created. This makes running shells
9188 as container init process a lot more fun.
9190 * The fstab support can now handle PARTUUID= and PARTLABEL=
9193 * A new ConditionHost= condition has been added to match
9194 against the hostname (with globs) and machine ID. This is
9195 useful for clusters where a single OS image is used to
9196 provision a large number of hosts which shall run slightly
9197 different sets of services.
9199 * Services which hit the restart limit will now be placed in a
9202 Contributions from: Bertram Poettering, Dave Reisner, Huang
9203 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
9204 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
9208 * When running in --user mode systemd will now become a
9209 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
9210 tree a lot more organized.
9212 * A new PartOf= unit dependency type has been introduced that
9213 may be used to group services in a natural way.
9215 * "systemctl enable" may now be used to enable instances of
9218 * journalctl now prints error log levels in red, and
9219 warning/notice log levels in bright white. It also supports
9220 filtering by log level now.
9222 * cgtop gained a new -n switch (similar to top), to configure
9223 the maximum number of iterations to run for. It also gained
9224 -b, to run in batch mode (accepting no input).
9226 * The suffix ".service" may now be omitted on most systemctl
9227 command lines involving service unit names.
9229 * There's a new bus call in logind to lock all sessions, as
9230 well as a loginctl verb for it "lock-sessions".
9232 * libsystemd-logind.so gained a new call sd_journal_perror()
9233 that works similar to libc perror() but logs to the journal
9234 and encodes structured information about the error number.
9236 * /etc/crypttab entries now understand the new keyfile-size=
9239 * shutdown(8) now can send a (configurable) wall message when
9240 a shutdown is cancelled.
9242 * The mount propagation mode for the root file system will now
9243 default to "shared", which is useful to make containers work
9244 nicely out-of-the-box so that they receive new mounts from
9245 the host. This can be undone locally by running "mount
9246 --make-rprivate /" if needed.
9248 * The prefdm.service file has been removed. Distributions
9249 should maintain this unit downstream if they intend to keep
9250 it around. However, we recommend writing normal unit files
9251 for display managers instead.
9253 * Since systemd is a crucial part of the OS we will now
9254 default to a number of compiler switches that improve
9255 security (hardening) such as read-only relocations, stack
9256 protection, and suchlike.
9258 * The TimeoutSec= setting for services is now split into
9259 TimeoutStartSec= and TimeoutStopSec= to allow configuration
9260 of individual time outs for the start and the stop phase of
9263 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
9264 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
9265 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
9266 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
9267 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
9268 Gundersen, Zbigniew Jędrzejewski-Szmek
9272 * The journal and id128 C APIs are now fully documented as man
9275 * Extra safety checks have been added when transitioning from
9276 the initial RAM disk to the main system to avoid accidental
9279 * /etc/crypttab entries now understand the new keyfile-offset=
9282 * systemctl -t can now be used to filter by unit load state.
9284 * The journal C API gained the new sd_journal_wait() call to
9285 make writing synchronous journal clients easier.
9287 * journalctl gained the new -D switch to show journals from a
9290 * journalctl now displays a special marker between log
9291 messages of two different boots.
9293 * The journal is now explicitly flushed to /var via a service
9294 systemd-journal-flush.service, rather than implicitly simply
9295 by seeing /var/log/journal to be writable.
9297 * journalctl (and the journal C APIs) can now match for much
9298 more complex expressions, with alternatives and
9301 * When transitioning from the initial RAM disk to the main
9302 system we will now kill all processes in a killing spree to
9303 ensure no processes stay around by accident.
9305 * Three new specifiers may be used in unit files: %u, %h, %s
9306 resolve to the user name, user home directory resp. user
9307 shell. This is useful for running systemd user instances.
9309 * We now automatically rotate journal files if their data
9310 object hash table gets a fill level > 75%. We also size the
9311 hash table based on the configured maximum file size. This
9312 together should lower hash collisions drastically and thus
9313 speed things up a bit.
9315 * journalctl gained the new "--header" switch to introspect
9316 header data of journal files.
9318 * A new setting SystemCallFilters= has been added to services
9319 which may be used to apply blacklists or whitelists to
9320 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
9322 * nspawn gained a new --link-journal= switch (and quicker: -j)
9323 to link the container journal with the host. This makes it
9324 very easy to centralize log viewing on the host for all
9325 guests while still keeping the journal files separated.
9327 * Many bugfixes and optimizations
9329 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
9330 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
9331 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
9336 * Several tools now understand kernel command line arguments,
9337 which are only read when run in an initial RAM disk. They
9338 usually follow closely their normal counterparts, but are
9341 * There's a new tool to analyze the readahead files that are
9342 automatically generated at boot. Use:
9344 /usr/lib/systemd/systemd-readahead analyze /.readahead
9346 * We now provide an early debug shell on tty9 if this enabled. Use:
9348 systemctl enable debug-shell.service
9350 * All plymouth related units have been moved into the Plymouth
9351 package. Please make sure to upgrade your Plymouth version
9354 * systemd-tmpfiles now supports getting passed the basename of
9355 a configuration file only, in which case it will look for it
9356 in all appropriate directories automatically.
9358 * udevadm info now takes a /dev or /sys path as argument, and
9359 does the right thing. Example:
9361 udevadm info /dev/sda
9362 udevadm info /sys/class/block/sda
9364 * systemctl now prints a warning if a unit is stopped but a
9365 unit that might trigger it continues to run. Example: a
9366 service is stopped but the socket that activates it is left
9369 * "systemctl status" will now mention if the log output was
9370 shortened due to rotation since a service has been started.
9372 * The journal API now exposes functions to determine the
9373 "cutoff" times due to rotation.
9375 * journald now understands SIGUSR1 and SIGUSR2 for triggering
9376 immediately flushing of runtime logs to /var if possible,
9377 resp. for triggering immediate rotation of the journal
9380 * It is now considered an error if a service is attempted to
9381 be stopped that is not loaded.
9383 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
9385 * systemd-analyze now supports Python 3
9387 * tmpfiles now supports cleaning up directories via aging
9388 where the first level dirs are always kept around but
9389 directories beneath it automatically aged. This is enabled
9390 by prefixing the age field with '~'.
9392 * Seat objects now expose CanGraphical, CanTTY properties
9393 which is required to deal with very fast bootups where the
9394 display manager might be running before the graphics drivers
9395 completed initialization.
9397 * Seat objects now expose a State property.
9399 * We now include RPM macros for service enabling/disabling
9400 based on the preset logic. We recommend RPM based
9401 distributions to make use of these macros if possible. This
9402 makes it simpler to reuse RPM spec files across
9405 * We now make sure that the collected systemd unit name is
9406 always valid when services log to the journal via
9409 * There's a new man page kernel-command-line(7) detailing all
9410 command line options we understand.
9412 * The fstab generator may now be disabled at boot by passing
9413 fstab=0 on the kernel command line.
9415 * A new kernel command line option modules-load= is now understood
9416 to load a specific kernel module statically, early at boot.
9418 * Unit names specified on the systemctl command line are now
9419 automatically escaped as needed. Also, if file system or
9420 device paths are specified they are automatically turned
9421 into the appropriate mount or device unit names. Example:
9423 systemctl status /home
9424 systemctl status /dev/sda
9426 * The SysVConsole= configuration option has been removed from
9427 system.conf parsing.
9429 * The SysV search path is no longer exported on the D-Bus
9432 * The Names= option has been removed from unit file parsing.
9434 * There's a new man page bootup(7) detailing the boot process.
9436 * Every unit and every generator we ship with systemd now
9437 comes with full documentation. The self-explanatory boot is
9440 * A couple of services gained "systemd-" prefixes in their
9441 name if they wrap systemd code, rather than only external
9442 code. Among them fsck@.service which is now
9443 systemd-fsck@.service.
9445 * The HaveWatchdog property has been removed from the D-Bus
9448 * systemd.confirm_spawn= on the kernel command line should now
9451 * There's a new man page crypttab(5) which details all options
9452 we actually understand.
9454 * systemd-nspawn gained a new --capability= switch to pass
9455 additional capabilities to the container.
9457 * timedated will now read known NTP implementation unit names
9458 from /usr/lib/systemd/ntp-units.d/*.list,
9459 systemd-timedated-ntp.target has been removed.
9461 * journalctl gained a new switch "-b" that lists log data of
9462 the current boot only.
9464 * The notify socket is in the abstract namespace again, in
9465 order to support daemons which chroot() at start-up.
9467 * There is a new Storage= configuration option for journald
9468 which allows configuration of where log data should go. This
9469 also provides a way to disable journal logging entirely, so
9470 that data collected is only forwarded to the console, the
9471 kernel log buffer or another syslog implementation.
9473 * Many bugfixes and optimizations
9475 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
9476 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
9477 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
9478 Shawn Landden, Tom Gundersen
9482 * "systemctl help <unit>" now shows the man page if one is
9485 * Several new man pages have been added.
9487 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
9488 MaxLevelConsole= can now be specified in
9489 journald.conf. These options allow reducing the amount of
9490 data stored on disk or forwarded by the log level.
9492 * TimerSlackNSec= can now be specified in system.conf for
9493 PID1. This allows system-wide power savings.
9495 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
9496 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
9501 * logind is now capable of (optionally) handling power and
9502 sleep keys as well as the lid switch.
9504 * journalctl now understands the syntax "journalctl
9505 /usr/bin/avahi-daemon" to get all log output of a specific
9508 * CapabilityBoundingSet= in system.conf now also influences
9509 the capability bound set of usermode helpers of the kernel.
9511 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
9512 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
9513 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
9514 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
9518 * Note that we skipped 139 releases here in order to set the
9519 new version to something that is greater than both udev's
9520 and systemd's most recent version number.
9522 * udev: all udev sources are merged into the systemd source tree now.
9523 All future udev development will happen in the systemd tree. It
9524 is still fully supported to use the udev daemon and tools without
9525 systemd running, like in initramfs or other init systems. Building
9526 udev though, will require the *build* of the systemd tree, but
9527 udev can be properly *run* without systemd.
9529 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
9530 should be used to create dead device nodes as workarounds for broken
9533 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
9534 no longer supported. udev_monitor_new_from_netlink() needs to be
9535 used to subscribe to events.
9537 * udev: when udevd is started by systemd, processes which are left
9538 behind by forking them off of udev rules, are unconditionally cleaned
9539 up and killed now after the event handling has finished. Services or
9540 daemons must be started as systemd services. Services can be
9541 pulled-in by udev to get started, but they can no longer be directly
9542 forked by udev rules.
9544 * udev: the daemon binary is called systemd-udevd now and installed
9545 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
9546 to adapt to that, create symlink, or rename the binary after building
9549 * libudev no longer provides these symbols:
9550 udev_monitor_from_socket()
9551 udev_queue_get_failed_list_entry()
9552 udev_get_{dev,sys,run}_path()
9553 The versions number was bumped and symbol versioning introduced.
9555 * systemd-loginctl and systemd-journalctl have been renamed
9556 to loginctl and journalctl to match systemctl.
9558 * The config files: /etc/systemd/systemd-logind.conf and
9559 /etc/systemd/systemd-journald.conf have been renamed to
9560 logind.conf and journald.conf. Package updates should rename
9561 the files to the new names on upgrade.
9563 * For almost all files the license is now LGPL2.1+, changed
9564 from the previous GPL2.0+. Exceptions are some minor stuff
9565 of udev (which will be changed to LGPL2.1 eventually, too),
9566 and the MIT licensed sd-daemon.[ch] library that is suitable
9567 to be used as drop-in files.
9569 * systemd and logind now handle system sleep states, in
9570 particular suspending and hibernating.
9572 * logind now implements a sleep/shutdown/idle inhibiting logic
9573 suitable for a variety of uses. Soonishly Lennart will blog
9574 about this in more detail.
9576 * var-run.mount and var-lock.mount are no longer provided
9577 (which previously bind mounted these directories to their new
9578 places). Distributions which have not converted these
9579 directories to symlinks should consider stealing these files
9580 from git history and add them downstream.
9582 * We introduced the Documentation= field for units and added
9583 this to all our shipped units. This is useful to make it
9584 easier to explore the boot and the purpose of the various
9587 * All smaller setup units (such as
9588 systemd-vconsole-setup.service) now detect properly if they
9589 are run in a container and are skipped when
9590 appropriate. This guarantees an entirely noise-free boot in
9591 Linux container environments such as systemd-nspawn.
9593 * A framework for implementing offline system updates is now
9594 integrated, for details see:
9595 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
9597 * A new service type Type=idle is available now which helps us
9598 avoiding ugly interleaving of getty output and boot status
9601 * There's now a system-wide CapabilityBoundingSet= option to
9602 globally reduce the set of capabilities for the
9603 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
9604 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
9605 even CAP_NET_ADMIN system-wide for secure systems.
9607 * There are now system-wide DefaultLimitXXX= options to
9608 globally change the defaults of the various resource limits
9609 for all units started by PID 1.
9611 * Harald Hoyer's systemd test suite has been integrated into
9612 systemd which allows easy testing of systemd builds in qemu
9613 and nspawn. (This is really awesome! Ask us for details!)
9615 * The fstab parser is now implemented as generator, not inside
9618 * systemctl will now warn you if .mount units generated from
9619 /etc/fstab are out of date due to changes in fstab that
9620 have not been read by systemd yet.
9622 * systemd is now suitable for usage in initrds. Dracut has
9623 already been updated to make use of this. With this in place
9624 initrds get a slight bit faster but primarily are much
9625 easier to introspect and debug since "systemctl status" in
9626 the host system can be used to introspect initrd services,
9627 and the journal from the initrd is kept around too.
9629 * systemd-delta has been added, a tool to explore differences
9630 between user/admin configuration and vendor defaults.
9632 * PrivateTmp= now affects both /tmp and /var/tmp.
9634 * Boot time status messages are now much prettier and feature
9635 proper english language. Booting up systemd has never been
9638 * Read-ahead pack files now include the inode number of all
9639 files to pre-cache. When the inode changes the pre-caching
9640 is not attempted. This should be nicer to deal with updated
9641 packages which might result in changes of read-ahead
9644 * We now temporaritly lower the kernel's read_ahead_kb variable
9645 when collecting read-ahead data to ensure the kernel's
9646 built-in read-ahead does not add noise to our measurements
9647 of necessary blocks to pre-cache.
9649 * There's now RequiresMountsFor= to add automatic dependencies
9650 for all mounts necessary for a specific file system path.
9652 * MountAuto= and SwapAuto= have been removed from
9653 system.conf. Mounting file systems at boot has to take place
9656 * nspawn now learned a new switch --uuid= to set the machine
9657 ID on the command line.
9659 * nspawn now learned the -b switch to automatically search
9662 * vt102 is now the default TERM for serial TTYs, upgraded from
9665 * systemd-logind now works on VT-less systems.
9667 * The build tree has been reorganized. The individual
9668 components now have directories of their own.
9670 * A new condition type ConditionPathIsReadWrite= is now available.
9672 * nspawn learned the new -C switch to create cgroups for the
9673 container in other hierarchies.
9675 * We now have support for hardware watchdogs, configurable in
9678 * The scheduled shutdown logic now has a public API.
9680 * We now mount /tmp as tmpfs by default, but this can be
9681 masked and /etc/fstab can override it.
9683 * Since udisks does not make use of /media anymore we are not
9684 mounting a tmpfs on it anymore.
9686 * journalctl gained a new --local switch to only interleave
9687 locally generated journal files.
9689 * We can now load the IMA policy at boot automatically.
9691 * The GTK tools have been split off into a systemd-ui.
9693 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
9694 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
9695 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
9696 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
9697 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
9698 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
9699 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
9700 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
9701 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
9706 * This is mostly a bugfix release
9708 * Support optional initialization of the machine ID from the
9709 KVM or container configured UUID.
9711 * Support immediate reboots with "systemctl reboot -ff"
9713 * Show /etc/os-release data in systemd-analyze output
9715 * Many bugfixes for the journal, including endianness fixes and
9716 ensuring that disk space enforcement works
9718 * sd-login.h is C++ compatible again
9720 * Extend the /etc/os-release format on request of the Debian
9723 * We now refuse non-UTF8 strings used in various configuration
9724 and unit files. This is done to ensure we do not pass invalid
9725 data over D-Bus or expose it elsewhere.
9727 * Register Mimo USB Screens as suitable for automatic seat
9730 * Read SELinux client context from journal clients in a race
9733 * Reorder configuration file lookup order. /etc now always
9734 overrides /run in order to allow the administrator to always
9735 and unconditionally override vendor-supplied or
9736 automatically generated data.
9738 * The various user visible bits of the journal now have man
9739 pages. We still lack man pages for the journal API calls
9742 * We now ship all man pages in HTML format again in the
9745 Contributions from: Dave Reisner, Dirk Eibach, Frederic
9746 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
9747 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
9752 * This is mostly a bugfix release
9754 * systems lacking /etc/os-release are no longer supported.
9756 * Various functionality updates to libsystemd-login.so
9758 * Track class of PAM logins to distinguish greeters from
9761 Contributions from: Kay Sievers, Lennart Poettering, Michael
9766 * This is an important bugfix release for v41.
9768 * Building man pages is now optional which should be useful
9769 for those building systemd from git but unwilling to install
9772 * Watchdog support for supervising services is now usable. In
9773 a future release support for hardware watchdogs
9774 (i.e. /dev/watchdog) will be added building on this.
9776 * Service start rate limiting is now configurable and can be
9777 turned off per service. When a start rate limit is hit a
9778 reboot can automatically be triggered.
9780 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
9782 Contributions from: Benjamin Franzke, Bill Nottingham,
9783 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
9784 Schmidt, Michał Górny, Piotr Drąg
9788 * The systemd binary is installed /usr/lib/systemd/systemd now;
9789 An existing /sbin/init symlink needs to be adapted with the
9792 * The code that loads kernel modules has been ported to invoke
9793 libkmod directly, instead of modprobe. This means we do not
9794 support systems with module-init-tools anymore.
9796 * Watchdog support is now already useful, but still not
9799 * A new kernel command line option systemd.setenv= is
9800 understood to set system wide environment variables
9801 dynamically at boot.
9803 * We now limit the set of capabilities of systemd-journald.
9805 * We now set SIGPIPE to ignore by default, since it only is
9806 useful in shell pipelines, and has little use in general
9807 code. This can be disabled with IgnoreSIPIPE=no in unit
9810 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
9811 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
9816 * This is mostly a bugfix release
9818 * We now expose the reason why a service failed in the
9819 "Result" D-Bus property.
9821 * Rudimentary service watchdog support (will be completed over
9822 the next few releases.)
9824 * When systemd forks off in order execute some service we will
9825 now immediately changes its argv[0] to reflect which process
9826 it will execute. This is useful to minimize the time window
9827 with a generic argv[0], which makes bootcharts more useful
9829 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
9830 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
9831 Mike Kazantsev, Ray Strode
9835 * This is mostly a test release, but incorporates many
9838 * New systemd-cgtop tool to show control groups by their
9841 * Linking against libacl for ACLs is optional again. If
9842 disabled, support tracking device access for active logins
9843 goes becomes unavailable, and so does access to the user
9844 journals by the respective users.
9846 * If a group "adm" exists, journal files are automatically
9847 owned by them, thus allow members of this group full access
9848 to the system journal as well as all user journals.
9850 * The journal now stores the SELinux context of the logging
9851 client for all entries.
9853 * Add C++ inclusion guards to all public headers
9855 * New output mode "cat" in the journal to print only text
9856 messages, without any meta data like date or time.
9858 * Include tiny X server wrapper as a temporary stop-gap to
9859 teach XOrg udev display enumeration. This is used by display
9860 managers such as gdm, and will go away as soon as XOrg
9861 learned native udev hotplugging for display devices.
9863 * Add new systemd-cat tool for executing arbitrary programs
9864 with STDERR/STDOUT connected to the journal. Can also act as
9865 BSD logger replacement, and does so by default.
9867 * Optionally store all locally generated coredumps in the
9868 journal along with meta data.
9870 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
9871 writing short strings to files (for usage for /sys), and for
9872 creating symlinks, character and block device nodes.
9874 * New unit file option ControlGroupPersistent= to make cgroups
9875 persistent, following the mechanisms outlined in
9876 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
9878 * Support multiple local RTCs in a sane way
9880 * No longer monopolize IO when replaying readahead data on
9881 rotating disks, since we might starve non-file-system IO to
9882 death, since fanotify() will not see accesses done by blkid,
9885 * Do not show kernel threads in systemd-cgls anymore, unless
9886 requested with new -k switch.
9888 Contributions from: Dan Horák, Kay Sievers, Lennart
9889 Poettering, Michal Schmidt
9893 * This is mostly a test release, but incorporates many
9896 * The git repository moved to:
9897 git://anongit.freedesktop.org/systemd/systemd
9898 ssh://git.freedesktop.org/git/systemd/systemd
9900 * First release with the journal
9901 http://0pointer.de/blog/projects/the-journal.html
9903 * The journal replaces both systemd-kmsg-syslogd and
9904 systemd-stdout-bridge.
9906 * New sd_pid_get_unit() API call in libsystemd-logind
9908 * Many systemadm clean-ups
9910 * Introduce remote-fs-pre.target which is ordered before all
9911 remote mounts and may be used to start services before all
9914 * Added Mageia support
9916 * Add bash completion for systemd-loginctl
9918 * Actively monitor PID file creation for daemons which exit in
9919 the parent process before having finished writing the PID
9920 file in the daemon process. Daemons which do this need to be
9921 fixed (i.e. PID file creation must have finished before the
9922 parent exits), but we now react a bit more gracefully to them.
9924 * Add colourful boot output, mimicking the well-known output
9925 of existing distributions.
9927 * New option PassCredentials= for socket units, for
9928 compatibility with a recent kernel ABI breakage.
9930 * /etc/rc.local is now hooked in via a generator binary, and
9931 thus will no longer act as synchronization point during
9934 * systemctl list-unit-files now supports --root=.
9936 * systemd-tmpfiles now understands two new commands: z, Z for
9937 relabelling files according to the SELinux database. This is
9938 useful to apply SELinux labels to specific files in /sys,
9941 * Output of SysV services is now forwarded to both the console
9942 and the journal by default, not only just the console.
9944 * New man pages for all APIs from libsystemd-login.
9946 * The build tree got reorganized and the build system is a
9947 lot more modular allowing embedded setups to specifically
9948 select the components of systemd they are interested in.
9950 * Support for Linux systems lacking the kernel VT subsystem is
9953 * configure's --with-rootdir= got renamed to
9954 --with-rootprefix= to follow the naming used by udev and
9957 * Unless specified otherwise we will now install to /usr instead
9958 of /usr/local by default.
9960 * Processes with '@' in argv[0][0] are now excluded from the
9961 final shut-down killing spree, following the logic explained
9963 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
9965 * All processes remaining in a service cgroup when we enter
9966 the START or START_PRE states are now killed with
9967 SIGKILL. That means it is no longer possible to spawn
9968 background processes from ExecStart= lines (which was never
9969 supported anyway, and bad style).
9971 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
9972 reloading of units together.
9974 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
9975 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
9976 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
9977 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
9978 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
projects / platform / upstream / systemd.git / blob