1 systemd System and Service Manager
5 Announcements of Future Feature Removals and Incompatible Changes:
7 * Support for automatic flushing of the nscd user/group database caches
8 will be dropped in a future release.
10 * Support for cgroup v1 ('legacy' and 'hybrid' hierarchies) is now
11 considered obsolete and systemd by default will refuse to boot under
12 it. To forcibly reenable cgroup v1 support,
13 SYSTEMD_CGROUP_ENABLE_LEGACY_FORCE=1 must be set on kernel command
14 line. The meson option 'default-hierarchy=' is also deprecated, i.e.
15 only cgroup v2 ('unified' hierarchy) can be selected as build-time
18 * Previously, systemd-networkd did not explicitly remove any bridge
19 VLAN IDs assigned on bridge master and ports. Since version 256, if a
20 .network file for an interface has at least one valid setting in the
21 [BridgeVLAN] section, then all assigned VLAN IDs on the interface
22 that are not configured in the .network file are removed.
24 * systemd-gpt-auto-generator will stop generating units for ESP or
25 XBOOTLDR partitions if it finds mount entries for or below the /boot/
26 or /efi/ hierarchies in /etc/fstab. This is to prevent the generator
27 from interfering with systems where the ESP is explicitly configured
28 to be mounted at some path, for example /boot/efi/ (this type of
29 setup is obsolete, but still commonly found).
31 * The behavior of systemd-sleep and systemd-homed has been updated to
32 freeze user sessions when entering the various sleep modes or when
33 locking a homed-managed home area. This is known to cause issues with
34 the proprietary NVIDIA drivers. Packagers of the NVIDIA proprietary
35 drivers may want to add drop-in configuration files that set
36 SYSTEMD_SLEEP_FREEZE_USER_SESSION=false for systemd-suspend.service
37 and related services, and SYSTEMD_HOME_LOCK_FREEZE_SESSION=false for
38 systemd-homed.service.
40 * systemd-tmpfiles and systemd-sysusers, when given a relative
41 configuration file path (with at least one directory separator '/'),
42 will open the file directly, instead of searching for the given
43 partial path in the standard locations. The old mode wasn't useful
44 because tmpfiles.d/ and sysusers.d/ configuration has a flat
45 structure with no subdirectories under the standard locations and
46 this change makes it easier to work with local files with those
49 * systemd-tmpfiles now properly applies nested configuration to 'R' and
50 'D' stanzas. For example, with the combination of 'R /foo' and 'x
51 /foo/bar', /foo/bar will now be excluded from removal.
53 General Changes and New Features:
55 * Various programs will now attempt to load the main configuration file
56 from locations below /usr/lib/, /usr/local/lib/, and /run/, not just
57 below /etc/. For example, systemd-logind will look for
58 /etc/systemd/logind.conf, /run/systemd/logind.conf,
59 /usr/local/lib/systemd/logind.conf, and /usr/lib/systemd/logind.conf,
60 and use the first file that is found. This means that the search
61 logic for the main config file and for drop-ins is now the same.
63 Similarly, kernel-install will look for the config files in
64 /usr/lib/kernel/ and the other search locations, and now also
67 systemd-udevd now supports drop-ins for udev.conf.
69 * A new 'systemd-vpick' binary has been added. It implements the new
70 vpick protocol, where a "*.v/" directory may contain multiple files
71 which have versions (following the UAPI version format specification)
72 embedded in the file name. The files are ordered by version and
73 the newest one is selected.
75 systemd-nspawn --image=/--directory=, systemd-dissect, systemd-portabled,
76 and the RootDirectory=, RootImage=, ExtensionImages=, and
77 ExtensionDirectories= settings for units now support the vpick
78 protocol and allow the latest version to be selected automatically if
79 a "*.v/" directory is specified as the source.
81 * Encrypted service credentials can now be made accessible to
82 unprivileged users. systemd-creds gained new options --user/--uid=
83 for encrypting/decrypting a credential for a specific user.
85 * New command-line tool 'importctl' to download, import, and export
86 disk images via systemd-importd is added with the following verbs:
87 pull-tar, pull-raw, import-tar, import-raw, import-fs, export-tar,
88 export-raw, list-transfers, and cancel-transfer. This functionality
89 was previously available in "machinectl", where it was used
90 exclusively for machine images. The new "importctl" generalizes this
91 for sysext, confext, and portable service images.
95 * New system manager setting ProtectSystem= has been added. It is
96 analogous to the unit setting, but applies to the whole system. It is
97 enabled by default in the initrd.
99 * New unit setting WantsMountsFor= has been added. It is analogous to
100 RequiresMountsFor=, but creates a Wants= dependency instead of
101 Requires=. This new logic is now used in various places where mounts
102 were added as dependencies for other settings (WorkingDirectory=-…,
103 PrivateTmp=yes, cryptsetup lines with 'nofail').
105 * New unit setting MemoryZSwapWriteback= can be used to control the new
106 memory.zswap.writeback cgroup knob added in kernel 6.8.
108 * The manager gained a org.freedesktop.systemd1.StartAuxiliaryScope()
109 D-Bus method to devolve some processes from a service into a new
110 scope. This new scope will remain running, even when the original
111 service unit is restarted or stopped. This allows a service unit to
112 split out some worker processes which need to continue running.
113 Control group properties of the new scope are copied from the
114 originating unit, so various limits are retained.
116 * Units now expose properties EffectiveMemoryMax=,
117 EffectiveMemoryHigh=, and EffectiveTasksMax=, which report the
118 most stringent limit systemd is aware of for the given unit.
120 * A new unit file specifier %D expands to $XDG_DATA_HOME (for user
121 services) or /usr/share/ (for system services).
123 * AllowedCPUs= now supports specifier expansion.
125 * What= setting in .mount and .swap units now accepts fstab-style
126 identifiers, for example UUID=… or LABEL=….
128 * RestrictNetworkInterfaces= now supports alternative network interface
131 * PAMName= now implies SetLoginEnvironment=yes.
133 * systemd.firstboot=no can be used on the kernel command-line to
134 disable interactive queries, but allow other first boot configuration
135 to happen based on credentials.
137 * The system's hostname can be configured via the systemd.hostname
140 * The systemd binary will no longer chainload sysvinit's "telinit"
141 binary when called under the init/telinit name on a system that isn't
142 booted with systemd. This previously has been supported to make sure
143 a distribution that has both init systems installed can reasonably
144 switch from one to the other via a simple reboot. Distributions
145 apparently have lost interest in this, and the functionality has not
146 been supported on the primary distribution this was still intended
147 for for a long time, and hence has been removed now.
149 * A new concept called "capsules" has been introduced. "Capsules" wrap
150 additional per-user service managers, whose users are transient and
151 are only defined as long as the service manager is running. (This is
152 implemented via DynamicUser=1), allowing a user manager to be used to
153 manager a group of processes without needing to create an actual user
154 account. These service managers run with home directories of
155 /var/lib/capsules/<capsule-name> and can contain regular services and
156 other units. A capsule is started via a simple "systemctl start
157 capsule@<name>.service". See the capsule@.service(5) man page for
160 Various systemd tools (including, and most importantly, systemctl and
161 systemd-run) have been updated to interact with capsules via the new
162 "--capsule="/"-C" switch.
164 * .socket units gained a new setting PassFileDescriptorsToExec=, taking
165 a boolean value. If set to true the file descriptors the socket unit
166 encapsulates are passed to the ExecStartPost=, ExecStopPre=,
167 ExecStopPost= using the usual $LISTEN_FDS interface. This may be used
168 for doing additional initializations on the sockets once they are
169 allocated. (For example, to install an additional eBPF program on
172 * The .socket setting MaxConnectionsPerSource= (which so far put a
173 limit on concurrent connections per IP in Accept=yes socket units),
174 now also has an effect on AF_UNIX sockets: it will put a limit on the
175 number of simultaneous connections from the same source UID (as
176 determined via SO_PEERCRED). This is useful for implementing IPC
177 services in a simple Accept=yes mode.
179 * The service manager will now maintain a counter of soft reboot cycles
180 the system went through. It may be queried via the D-Bus APIs.
182 * systemd's execution logic now supports the new pidfd_spawn() API
183 introduced by glibc 2.39, which allows us to invoke a subprocess in a
184 target cgroup and get a pidfd back in a single operation.
186 * systemd/PID 1 will now send an additional sd_notify() message to its
187 supervising VMM or container manager reporting the selected hostname
188 ("X_SYSTEMD_HOSTNAME=") and machine ID ("X_SYSTEMD_MACHINE_ID=") at
189 boot. Moreover, the service manager will send additional sd_notify()
190 messages ("X_SYSTEMD_UNIT_ACTIVE=") whenever a target unit is
191 reached. This can be used by VMMs/container managers to schedule
192 access to the system precisely. For example, the moment a system
193 reports "ssh-access.target" being reached a VMM/container manager
194 knows it can now connect to the system via SSH. Finally, a new
195 sd_notify() message ("X_SYSTEMD_SIGNALS_LEVEL=2") is sent the moment
196 PID 1 has successfully completed installation of its various UNIX
197 process signal handlers (i.e. the moment where SIGRTMIN+4 sent to
198 PID 1 will start to have the effect of shutting down the system
203 * systemd-journald can now forward journal entries to a socket
204 (AF_INET, AF_INET6, AF_UNIX, or AF_VSOCK). The socket can be
205 specified in journald.conf via a new option ForwardAddress= or via
206 the 'journald.forward_address' credential. Log records are sent in
207 the Journal Export Format. A related setting MaxLevelSocket= has been
208 added to control the maximum log levels for the messages sent to this
211 * systemd-vmspawn gained a new --forward-journal= option to forward the
212 virtual machine's journal entries to the host. This is done over a
213 AF_VSOCK socket, i.e. it does not require networking in the guest.
215 * journalctl gained option '-i' as a shortcut for --file=.
217 * journalctl gained a new -T/--exclude-identifier= option to filter
218 out certain syslog identifiers.
220 * journalctl gained a new --list-namespaces option.
222 * systemd-journal-remote now also accepts AF_VSOCK and AF_UNIX sockets
223 (so it can be used to receive entries forwarded by systemd-journald).
225 * systemd-journal-gatewayd allows restricting the time range of
226 retrieved entries with a new "realtime=[<since>]:[<until>]" URL
229 * systemd-cat gained a new option --namespace= to specify the target
230 journal namespace to which the output shall be connected.
232 * systemd-bsod gained a new option --tty= to specify the output TTY
236 * /dev/ now contains symlinks that combine by-path and by-{label,uuid}
239 /dev/disk/by-path/<path>/by-<label|uuid|…>/<label|uuid|…>
241 This allows distinguishing partitions with identical contents on
242 multiple storage devices. This is useful, for example, when copying
243 raw disk contents between devices.
245 * systemd-udevd now creates persistent /dev/media/by-path/ symlinks for
246 media controllers. For example, the uvcvideo driver may create
247 /dev/media0 which will be linked as
248 /dev/media/by-path/pci-0000:04:00.3-usb-0:1:1.0-media-controller.
250 * A new unit systemd-udev-load-credentials.service has been added
251 to pick up udev.conf drop-ins and udev rules from credentials.
253 * An allowlist/denylist may be specified to filter which sysfs
254 attributes are used when crafting network interface names. Those
255 lists are stored as hwdb entries
256 ID_NET_NAME_ALLOW_<sysfsattr>=0|1
258 ID_NET_NAME_ALLOW=0|1.
260 The goal is to avoid unexpected changes to interface names when the
261 kernel is updated and new sysfs attributes become visible.
263 * A new unit tpm2.target has been added to provide a synchronization
264 point for units which expect the TPM hardware to be available. A new
265 generator "systemd-tpm2-generator" has been added that will insert
266 this target whenever it detects that the firmware has initialized a
267 TPM, but Linux hasn't loaded a driver for it yet.
269 * systemd-backlight now properly supports numbered devices which the
270 kernel creates to avoid collisions in the leds subsystem.
272 * systemd-hwdb update operation can be disabled with a new environment
273 variable SYSTEMD_HWDB_UPDATE_BYPASS=1.
277 * systemd-hostnamed now exposes the machine ID and boot ID via
278 D-Bus. It also exposes the hosts AF_VSOCK CID, if available.
280 * systemd-hostnamed now provides a basic Varlink interface.
282 * systemd-hostnamed exports the full data in os-release(5) and
283 machine-info(5) via D-Bus and Varlink.
285 * hostnamectl now shows the system's product UUID and hardware serial
290 * systemd-networkd now provides a basic Varlink interface.
292 * systemd-networkd's ARP proxy support gained a new option to configure
293 a private VLAN variant of the proxy ARP supported by the kernel under
294 the name IPv4ProxyARPPrivateVLAN=.
296 * systemd-networkd now exports the NamespaceId and NamespaceNSID
297 properties via D-Bus and Varlink. (which expose the inode and NSID of
298 the network namespace the networkd instance manages)
300 * systemd-networkd now supports IPv6RetransmissionTimeSec= and
301 UseRetransmissionTime= settings in .network files to configure
302 retransmission time for IPv6 neighbor solicitation messages.
304 * networkctl gained new verbs 'mask' and 'unmask' for masking networkd
305 configuration files such as .network files.
307 * 'networkctl edit --runtime' allows editing volatile configuration
308 under /run/systemd/network/.
310 * The implementation behind TTLPropagate= network setting has been
311 removed and the setting is now ignored.
313 * systemd-network-generator will now pick up .netdev/.link/.network/
314 networkd.conf configuration from system credentials.
316 * systemd-networkd will now pick up wireguard secrets from
319 * systemd-networkd's Varlink API now supports enumerating LLDP peers.
321 * .link files now support new Property=, ImportProperty=,
322 UnsetProperty= fields for setting udev properties on a link.
324 * The various .link files that systemd ships for interfaces that are
325 supposed to be managed by systemd-networkd only now carry a
326 ID_NET_MANAGED_BY=io.systemd.Network udev property ensuring that
327 other network management solutions honouring this udev property do
328 not come into conflict with networkd, trying to manage these
331 * .link files now support a new ReceivePacketSteeringCPUMask= setting
332 for configuring which CPUs to steer incoming packets to.
336 * systemd-nspawn now provides a /run/systemd/nspawn/unix-export/
337 directory where the container payload can expose AF_UNIX sockets to
338 allow them them to be accessed from outside.
340 * systemd-nspawn will tint the terminal background for containers in a
341 blueish color. This can be controller with the new --background=
344 * systemd-nspawn gained support for the 'owneridmap' option for --bind=
345 mounts to map the target directory owner from inside the container to
346 the owner of the directory bound from the host filesystem.
348 * systemd-nspawn now supports moving Wi-Fi network devices into a
349 container, just like other network interfaces.
353 * systemd-resolved now reads RFC 8914 EDE error codes provided by
354 upstream DNS services.
356 * systemd-resolved and resolvectl now support RFC 9460 SVCB and HTTPS
357 records, as well as RFC 2915 NAPTR records.
359 * resolvectl gained a new option --relax-single-label= to allow
360 querying single-label hostnames via unicast DNS on a per-query basis.
362 * systemd-resolved's Varlink IPC interface now supports resolving
363 DNS-SD services as well as an API for resolving raw DNS RRs.
365 * systemd-resolved's .dnssd DNS_SD service description files now
366 support DNS-SD "subtypes" via the new SubType= setting.
368 * systemd-resolved's configuration may now be reloaded without
369 restarting the service. (i.e. "systemctl reload systemd-resolved" is
374 * An sshd config drop-in to allow ssh keys acquired via userdbctl (for
375 example expose by homed accounts) to be used for authorization of
376 incoming SSH connections.
378 * A small new unit generator "systemd-ssh-generator" has been added. It
379 checks if the sshd binary is installed. If so, it binds it via
380 per-connection socket activation to various sockets depending on the
383 • If the system is run in a VM providing AF_VSOCK support, it
384 automatically binds sshd to AF_VSOCK port 22.
386 • If the system is invoked as a full-OS container and the container
387 manager pre-mounts a directory /run/host/unix-export/, it will
388 bind sshd to an AF_UNIX socket /run/host/unix-export/ssh. The
389 idea is the container manager bind mounts the directory to an
390 appropriate place on the host as well, so that the AF_UNIX socket
391 may be used to easily connect from the host to the container.
393 • sshd is also bound to an AF_UNIX socket
394 /run/ssh-unix-local/socket, which may be to use ssh/sftp in a
395 "sudo"-like fashion to access resources of other local users.
397 • Via the kernel command line option "systemd.ssh_listen=" and the
398 system credential "ssh.listen" sshd may be bound to additional,
399 explicitly configured options, including AF_INET/AF_INET6 ports.
401 In particular the first two mechanisms should make dealing with local
402 VMs and full OS containers a lot easier, as SSH connections will
403 *just* *work* from the host – even if no networking is available
406 systemd-ssh-generator optionally generates a per-connection
407 socket activation service file wrapping sshd. This is only done if
408 the distribution does not provide one on its own under the name
409 "sshd@.service". The generated unit only works correctly if the SSH
410 privilege separation ("privsep") directory exists. Unfortunately
411 distributions vary wildly where they place this directory. An
412 incomprehensive list:
414 • /usr/share/empty.sshd/ (new fedora)
417 • /run/sshd/ (debian/ubuntu?)
419 If the SSH privsep directory is placed below /var/ or /run/ care
420 needs to be taken that the directory is created automatically at boot
421 if needed, since these directories possibly or always come up
422 empty. This can be done via a tmpfiles.d/ drop-in. You may use the
423 "sshdprivsepdir" meson option provided by systemd to configure the
424 directory, in case you want systemd to create the directory as needed
425 automatically, if your distribution does not cover this natively.
427 Recommendations to distributions, in order to make things just work:
429 • Please provide a per-connection SSH service file under the name
432 • Please move the SSH privsep dir into /usr/ (so that it is truly
433 immutable on image-based operating systems, is strictly under
434 package manager control, and never requires recreation if the
435 system boots up with an empty /run/ or /var/).
437 • As an extension of this: please consider following Fedora's lead
438 here, and use /usr/share/empty.sshd/ to minimize needless
439 differences between distributions.
441 • If your distribution insists on placing the directory in /var/ or
442 /run/ then please at least provide a tmpfiles.d/ drop-in to
443 recreate it automatically at boot, so that the sshd binary just
444 works, regardless in which context it is called.
446 * A small tool "systemd-ssh-proxy" has been added, which is supposed to
447 act as counterpart to "systemd-ssh-generator". It's a small plug-in
448 for the SSH client (via ProxyCommand/ProxyUseFdpass) to allow it to
449 connect to AF_VSOCK or AF_UNIX sockets. Example: "ssh vsock/4711"
450 connects to a local VM with cid 4711, or "ssh
451 unix/run/ssh-unix-local/socket" to connect to the local host via the
452 AF_UNIX socket /run/ssh-unix-local/socket.
454 systemd-boot and systemd-stub and Related Tools:
456 * TPM 1.2 PCR measurement support has been removed from systemd-stub.
457 TPM 1.2 is obsolete and – due to the (by today's standards) weak
458 cryptographic algorithms it only supports – does not actually provide
459 the security benefits it's supposed to provide. Given that the rest
460 of systemd's codebase never supported TPM 1.2, the support has now
461 been removed from systemd-stub as well.
463 * systemd-stub will now measure its payload via the new EFI
464 Confidential Computing APIs (CC), in addition to the pre-existing
467 * confexts are loaded by systemd-stub from the ESP as well.
469 * kernel-install gained support for --root= for the 'list' verb.
471 * bootctl now provides a basic Varlink interface and can be run as a
472 daemon via a template unit.
474 * systemd-measure gained new options --certificate=, --private-key=,
475 and --private-key-source= to allow using OpenSSL's "engines" or
476 "providers" as the signing mechanism to use when creating signed
477 TPM2 PCR measurement values.
479 * ukify gained support for signing of PCR signatures via OpenSSL's
480 engines and providers.
482 * ukify now supports zboot kernels.
484 * systemd-boot now supports passing additional kernel command line
485 switches to invoked kernels via an SMBIOS Type #11 string
486 "io.systemd.boot.kernel-cmdline-extra". This is similar to the
487 pre-existing support for this in systemd-stub, but also applies to
488 Type #1 Boot Loader Specification Entries.
490 * systemd-boot's automatic SecureBoot enrollment support gained support
491 for enrolling "dbx" too (Previously, only db/KEK/PK enrollment was
492 supported). It also now supports UEFI "Custom" mode.
494 * The pcrlock policy is saved in an unencrypted credential file
495 "pcrlock.<entry-token>.cred" under XBOOTLDR/ESP in the
496 /loader/credentials/ directory. It will be picked up at boot by
497 systemd-stub and passed to the initrd, where it can be used to unlock
498 the root file system.
500 * systemd-pcrlock gained an --entry-token= option to configure the
503 * systemd-pcrlock now provides a basic Varlink interface and can be run
504 as a daemon via a template unit.
506 * systemd-pcrlock's TPM nvindex access policy has been modified, this
507 means that previous pcrlock policies stored in nvindexes are
508 invalidated. They must be removed (systemd-pcrlock remove-policy) and
509 recreated (systemd-pcrlock make-policy). For the time being
510 systemd-pcrlock remains an experimental feature, but it is expected
511 to become stable in the next release, i.e. v257.
513 * systemd-pcrlock's --recovery-pin= switch now takes three values:
514 "hide", "show", "query". If "show" is selected the automatically
515 generated recovery PIN is shown to the user. If "query" is selected
516 then the PIN is queried from the user.
520 * systemd-run is now a multi-call binary. When invoked as 'run0', it
521 provides as interface similar to 'sudo', with all arguments starting
522 at the first non-option parameter being treated the command to invoke
523 as root. Unlike 'sudo' and similar tools, it does not make use of
524 setuid binaries or other privilege escalation methods, but instead
525 runs the specified command as a transient unit, which is started by
526 the system service manager, so privileges are dropped, rather than
527 gained, thus implementing a much more robust and safe security
528 model. As usual, authorization is managed via Polkit.
530 * systemd-run/run0 will now tint the terminal background on supported
531 terminals: in a reddish tone when invoking a root service, in a
532 yellowish tone otherwise. This may be controlled and turned off via
533 the new --background= switch.
535 * systemd-run gained a new option '--ignore-failure' to suppress
540 * 'systemctl edit --stdin' allows creation of unit files and drop-ins
541 with contents supplied via standard input. This is useful when creating
542 configuration programmatically; the tool takes care of figuring out
543 the file name, creating any directories, and reloading the manager
546 * 'systemctl disable --now' and 'systemctl mask --now' now work
547 correctly with template units.
549 * 'systemd-analyze architectures' lists known CPU architectures.
551 * 'systemd-analyze --json=…' is supported for 'architectures',
552 'capability', 'exit-status'.
554 * 'systemd-tmpfiles --purge' will purge (remove) all files and
555 directories created via tmpfiles.d configuration.
557 * systemd-id128 gained new options --no-pager, --no-legend, and
560 * hostnamectl gained '-j' as shortcut for '--json=pretty' or
563 * loginctl now supports -j/--json=.
565 * resolvectl now supports -j/--json= for --type=.
567 * systemd-tmpfiles gained a new option --dry-run to print what would be
568 done without actually taking action.
570 * varlinkctl gained a new --collect switch to collect all responses of
571 a method call that supports multiple replies and turns it into a
574 * systemd-dissect gained a new --make-archive option to generate an
575 archive file (tar.gz and similar) from a disk image.
579 * systemd-vmspawn gained a new --firmware= option to configure or list
580 firmware definitions for Qemu, a new --tpm= option to enable or
581 disable the use of a software TPM, a new --linux= option to specify a
582 kernel binary for direct kernel boot, a new --initrd= option to
583 specify an initrd for direct kernel boot, a new -D/--directory option
584 to use a plain directory as the root file system, a new
585 --private-users option similar to the one in systemd-nspawn, new
586 options --bind= and --bind-ro= to bind part of the host's file system
587 hierarchy into the guest, a new --extra-drive= option to attach
588 additional storage, and -n/--network-tap/--network-user-mode to
589 configure networking.
591 * A new systemd-vmspawn@.service can be used to launch systemd-vmspawn
594 * systemd-vmspawn gained the new --console= and --background= switches
595 that control how to interact with the VM. As before, by default an
596 interactive terminal interface is provided, but now with a background
597 tinted with a greenish hue.
599 * systemd-vmspawn can now register its VMs with systemd-machined,
600 controlled via the --register= switch.
602 * machinectl's start command (and related) can now invoke images either
603 as containers via `systemd-nspawn` (switch is --runner=nspawn, the
604 default) or as VMs via `systemd-vmspawn` (switch is --runner=vmspawn,
607 * systemd-vmspawn now supports two switches --pass-ssh-key= and
608 --ssh-key-type= to optionally set up transient SSH keys to pass to the
609 invoked VMs in order to be able to SSH into them once booted.
611 * systemd-vmspawn will now enable various "HyperV enlightenments" and
612 the "VM Generation ID" on the VMs.
614 * A new environment variable $SYSTEMD_VMSPAWN_QEMU_EXTRA may carry
615 additional qemu command line options to pass to qemu.
619 * systemd-repart gained new options --generate-fstab= and
620 --generate-crypttab= to write out fstab and crypttab files matching the
621 generated partitions.
623 * systemd-repart gained a new option --private-key-source= to allow
624 using OpenSSL's "engines" or "providers" as the signing mechanism to
625 use when creating verity signature partitions.
627 * systemd-repart gained a new DefaultSubvolume= setting in repart.d/
628 drop-ins that allow configuring the default btrfs subvolume for newly
629 formatted btrfs file systems.
633 * libsystemd gained new call sd_bus_creds_new_from_pidfd() to get a
634 credentials object for a pidfd and sd_bus_creds_get_pidfd_dup() to
635 retrieve the pidfd from a credentials object.
637 * sd-bus' credentials logic will now also acquire peer's UNIX group
638 lists and peer's pidfd if supported and requested.
640 * RPM macro %_kernel_install_dir has been added with the path
641 to the directory for kernel-install plugins.
643 * The liblz4, libzstd, liblzma, libkmod, libgcrypt dependencies have
644 been changed from regular shared library dependencies into dlopen()
647 * The sd-journal API gained a new call
648 sd_journal_stream_fd_with_namespace() which is just like
649 sd_journal_stream_fd() but creates a log stream targeted at a
650 specific log namespace.
652 * The sd-id128 API gained a new API call
653 sd_id128_get_invocation_app_specific() for acquiring an app-specific
654 ID that is derived from the service invocation ID.
656 systemd-cryptsetup/systemd-cryptenroll:
658 * systemd-cryptenroll can now enroll directly with a PKCS11 public key
659 (instead of a certificate).
661 * systemd-cryptsetup/systemd-cryptenroll now may lock a disk against a
662 PKCS#11 provided EC key (before it only supported RSA).
664 * systemd-cryptsetup gained support for crypttab option
665 link-volume-key= to link the volume key into the kernel keyring when
666 the volume is opened.
668 * systemd-cryptenroll will no longer enable Dictionary Attack
669 Protection (i.e. turn on NO_DA) for TPM enrollments that do not
670 involve a PIN. DA should not be necessary in that case (since key
671 entropy is high enough to make this unnecessary), but risks
672 accidental lock-out in case of unexpected PCR changes.
674 * systemd-cryptenroll now supports enrolling a new slot while unlocking
675 the old slot via TPM2 (previously unlocking only worked via password
680 * The remaining documentation that was on
681 https://freedesktop.org/wiki/Software/systemd/ has been moved to
684 * A new text describing the VM integration interfaces of systemd has
687 https://systemd.io/VM_INTERFACE
689 * The sd_notify() man page has gained an example with C code that shows
690 how to implement the interface in C without involving libsystemd.
692 systemd-homed, systemd-logind, systemd-userdbd:
694 * systemd-homed now supports unlocking of home directories when logging
695 in via SSH. Previously home directories needed to be unlocked before
696 an SSH login is attempted.
698 * JSON User Records have been extended with a separate public storage
699 area called "User Record Blob Directories". This is intended to store
700 the user's background image, avatar picture, and other similar items
701 which are too large to fit into the User Record itself.
703 systemd-homed, userdbctl, and homectl gained support for blob
704 directories. homectl gained --avatar= and --login-background= to
705 control two specific items of the blob directories.
707 * A new "additionalLanguages" field has been added to JSON user records
708 (as supported by systemd-homed and systemd-userdbd), which is closely
709 related to the pre-existing "preferredLanguage", and allows
710 specifying multiple additional languages for the user account. It is
711 used to initialize the $LANGUAGES environment variable when used.
713 * A new pair of "preferredSessionType" and "preferredSessionLauncher"
714 fields have been added to JSON user records, that may be used to
715 control which kind of desktop session to preferable activate on
718 * homectl gained a new verb 'firstboot', and a new
719 systemd-homed-firstboot.service unit uses this verb to create users
720 in a first boot environment, either from system credentials or by
721 querying interactively.
723 * systemd-logind now supports a new "background-light" session class
724 which does not pull in the user@.service unit. This is intended in
725 particular for lighter weight per-user cron jobs which do require any
726 per-user service manager to be around.
728 * The per-user service manager will now be tracked as a distinct "manager"
729 session type among logind sessions of each user.
731 * homectl now supports an --offline mode, by which certain account
732 properties can be changed without unlocking the home directory.
734 * systemd-logind gained a new
735 org.freedesktop.login1.Manager.ListSessionsEx() method that provides
736 additional metadata compared to ListSessions(). loginctl makes use of
737 this to list additional fields in list-sessions.
739 * systemd-logind gained a new org.freedesktop.login1.Manager.Sleep()
740 method that automatically redirects to SuspendThenHibernate(),
741 Suspend(), HybridSleep(), or Hibernate(), depending on what is
742 supported and configured, a new configuration setting SleepOperation=,
743 and an accompanying helper method
744 org.freedesktop.login1.Manager.CanSleep() and property
745 org.freedesktop.login1.Manager.SleepOperation.
747 'systemctl sleep' calls the new method to automatically put the
748 machine to sleep in the most appropriate way.
750 Credential Management:
752 * systemd-creds now provides a Varlink IPC API for encrypting and
753 decrypting credentials.
755 * systemd-creds' "tpm2-absent" key selection has been renamed to
756 "null", since that's what it actually does: "encrypt" and "sign"
757 with a fixed null key. --with-key=null should only be used in very
758 specific cases, as it provides zero integrity or confidentiality
759 protections. (i.e. it's only safe to use as fallback in environments
760 lacking both a TPM and access to the root fs to use the host
761 encryption key, or when integrity is provided some other way.)
763 * systemd-creds gained a new switch --allow-null. If specified, the
764 "decrypt" verb will decode encrypted credentials that use the "null"
765 key (by default this is refused, since using the "null" key defeats
766 the authenticated encryption normally done).
770 * The sleep.conf configuration file gained a new MemorySleepMode=
771 setting for configuring the sleep mode in more detail.
773 * A tiny new service systemd-hibernate-clear.service has been added
774 which clears hibernation information from the HibernateLocation EFI
775 variable, in case the resume device is gone. Normally, this variable
776 is supposed to be cleaned up by the code that initiates the resume
777 from hibernation image. But when the device is missing and that code
778 doesn't run, this service will now do the necessary work, ensuring
779 that no outdated hibernation image information remains on subsequent
782 Unprivileged User Namespaces & Mounts:
784 * A small new service systemd-nsresourced.service has been added. It
785 provides a Varlink IPC API that assigns a free, transiently allocated
786 64K UID/GID range to an uninitialized user namespace a client
787 provides. It may be used to implement unprivileged container managers
788 and other programs that need dynamic user ID ranges. It also provides
789 interfaces to then delegate mount file descriptors, control groups
790 and network interfaces to user namespaces set up this way.
792 * A small new service systemd-mountfsd.service has been added. It
793 provides a Varlink IPC API for mounting DDI images, and returning a set
794 of mount file descriptors for it. If a user namespace fd is provided
795 as input, then the mounts are registered with the user namespace. To
796 ensure trust in the image it must provide Verity information (or
797 alternatively interactive polkit authentication is required).
799 * The systemd-dissect tool now can access DDIs fully unprivileged by
800 using systemd-nsresourced/systemd-mountfsd.
802 * If the service manager runs unprivileged (i.e. systemd --user) it now
803 supports RootImage= for accessing DDI images, also implemented via
804 the systemd-nsresourced/systemd-mountfsd.
806 * systemd-nspawn may now operate without privileges, if a suitable DDI
807 is provided via --image=, again implemented via
808 systemd-nsresourced/systemd-mountfsd.
812 * timedatectl and machinectl gained option '-P', an alias for
813 '--value --property=…'.
815 * Various tools that pretty-print config files will now highlight
816 configuration directives.
818 * varlinkctl gained support for the "ssh:" transport. This requires
819 OpenSSH 9.4 or newer.
821 * systemd-sysext gained support for enabling system extensions in
822 mutable fashion, where a writeable upperdir is stored under
823 /var/lib/extensions.mutable/, and a new --mutable= option to
824 configure this behaviour. An "ephemeral" mode is not also supported
825 where the mutable layer is configured to be a tmpfs that is
826 automatically released when the system extensions are reattached.
828 * Coredumps are now retained for two weeks by default (instead of three
831 * portablectl --copy= parameter gained a new 'mixed' argument, that will
832 result in resources owned by the OS (e.g.: portable profiles) to be linked
833 but resources owned by the portable image (e.g.: the unit files and the
834 images themselves) to be copied.
836 * systemd will now register MIME types for various of its file types
837 (e.g. journal files, DDIs, encrypted credentials …) via the XDG
838 shared-mime-info infrastructure. (Files of these types will thus be
839 recognized as their own thing in desktop file managers such as GNOME
842 * systemd-dissect will now show the detected sector size of a given DDI
843 in its default output.
845 * systemd-portabled now generates recognizable structured log messages
846 whenever a portable service is attached or detached.
848 * Verity signature checking in userspace (i.e. checking against
849 /etc/verity.d/ keys) when activating DDIs can now be turned on/off
850 via a kernel command line option systemd.allow_userspace_verity= and
851 an environment variable SYSTEMD_ALLOW_USERSPACE_VERITY=.
853 * ext4/xfs file system quota handling has been reworked, so that
854 quotacheck and quotaon are now invoked as per-file-system templated
855 services (as opposed to single system-wide singletons), similar in
856 style to the fsck, growfs, pcrfs logic. This means file systems with
857 quota enabled can now be reasonably enabled at runtime of the system,
860 * "systemd-analyze dot" will now also show BindsTo= dependencies.
862 * systemd-debug-generator gained the ability add in arbitrary units
863 based on them being passed in via system credentials.
865 * A new kernel command-line option systemd.default_debug_tty= can be
866 used to specify the TTY for the debug shell, independently of
867 enabling or disabling it.
869 Contributions from: A S Alam, AKHIL KUMAR, Abraham Samuel Adekunle,
870 Adrian Vovk, Adrian Wannenmacher, Alan Liang, Alberto Planas,
871 Alexander Zavyalov, Anders Jonsson, Andika Triwidada, Andres Beltran,
872 Andrew Sayers, Antonio Alvarez Feijoo, Arthur Zamarin, Artur Pak,
873 AtariDreams, Benjamin Franzke, Bernhard M. Wiedemann, Black-Hole1,
874 Bryan Jacobs, Burak Gerz, Carlos Garnacho, Chandra Pratap,
875 Chris Simons, Christian Wesselhoeft, Clayton Craft, Colin Geniet,
876 Colin Walters, Costa Tsaousis, Cristian Rodríguez, Daan De Meyer,
877 Damien Challet, Dan Streetman, David Tardon, David Venhoek,
878 Diego Viola, Dionna Amalie Glaze, Dmitry Konishchev,
879 Edson Juliano Drosdeck, Eisuke Kawashima, Eli Schwartz,
880 Emanuele Giuseppe Esposito, Eric Daigle, Evgeny Vereshchagin,
881 Felix Riemann, Fernando Fernandez Mancera, Florian Schmaus, Franck Bui,
882 Frantisek Sumsal, Friedrich Altheide, Gabríel Arthúr Pétursson,
883 Gaël Donval, Georges Basile Stavracas Neto, Gerd Hoffmann, GNOME
884 Foundation, Guilhem Lettron, Göran Uddeborg, Hans de Goede, Harald
885 Brinkmann, Heinrich Schuchardt, Henry Li, Holger Assmann, Ivan Kruglov,
886 Ivan Shapovalov, Jakub Sitnicki, James Muir, Jan Engelhardt, Jan Macku,
887 Jeff King, JmbFountain, Joakim Nohlgård, Julius Alexandre,
888 Jörg Behrmann, Keian, Kirk, Kristian Klausen, Krzesimir Nowak,
889 Lars Ellenberg, Lennart Poettering, Luca Boccassi, Ludwig Nussel,
890 Lukáš Nykrýn, Luxiter, Maanya Goenka, Mariano Giménez,
891 Markus Merklinger, Martin Ivicic, Martin Trigaux, Martin Wilck,
892 Matt Layher, Matt Muggeridge, Matteo Croce, Matthias Lisin,
893 Max Gautier, Max Staudt, Michael Biebl, Michal Koutný, Michal Sekletár,
894 Mike Gilbert, Mike Yuan, Mikko Ylinen, MkfsSion, MrSmör,
895 Nandakumar Raghavan, Nick Cao, Nick Rosbrook, Ole Peder Brandtzæg,
896 Ondrej Kozina, Oğuz Ersen, Pablo Méndez Hernández, Pierre GRASSER,
897 Piotr Drąg, QuonXF, Rafaël Kooi, Raito Bezarius, Reid Wahl,
898 Renjaya Raga Zenta, Richard Maw, Roland Hieber, Ronan Pigott, Rose,
899 Ross Burton, Sam Leonard, Samuel BF, Sergei Zhmylev, Sergey A, Shulhan,
900 SidhuRupinder, Sludge, Stuart Hayhurst, Susant Sahani,
901 Takashi Sakamoto, Temuri Doghonadze, Thilo Fromm, Thomas Blume,
902 TobiPeterG, Tomáš Pecka, Topi Miettinen, Tycho Andersen, Unique-Usman,
903 Usman Akinyemi, Vasiliy Kovalev, Vasiliy Stelmachenok,
904 Vishal Chillara Srinivas, Vitaly Kuznetsov, Vito Caputo,
905 Vladimir Stoiakin, Werner Sembach, Will Springer, Winterhuman,
906 Xiaotian Wu, Yu Watanabe, Yuri Chornoivan, Zbigniew Jędrzejewski-Szmek,
907 Zmyeir, aslepykh, chenjiayi, cpackham-atlnz, cunshunxia, djantti,
908 hfavisado, hulkoba, ksaleem, medusalix, mille-feuille, mkubiak, mooo,
909 msizanoen, networkException, nl6720, r-vdp, runiq, samuelvw01,
910 sharad3001, sushmbha, wangyuhang, zzywysm, İ. Ensar Gülşen,
911 Štěpán Němec, 我超厉害, 김인수
913 — Happy Place, 2024-04-XX
917 Announcements of Future Feature Removals and Incompatible Changes:
919 * Support for split-usr (/usr/ mounted separately during late boot,
920 instead of being mounted by the initrd before switching to the rootfs)
921 and unmerged-usr (parallel directories /bin/ and /usr/bin/, /lib/ and
922 /usr/lib/, …) has been removed. For more details, see:
923 https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html
925 * We intend to remove cgroup v1 support from a systemd release after
926 the end of 2023. If you run services that make explicit use of
927 cgroup v1 features (i.e. the "legacy hierarchy" with separate
928 hierarchies for each controller), please implement compatibility with
929 cgroup v2 (i.e. the "unified hierarchy") sooner rather than later.
930 Most of Linux userspace has been ported over already.
932 * Support for System V service scripts is now deprecated and will be
933 removed in a future release. Please make sure to update your software
934 *now* to include a native systemd unit file instead of a legacy
935 System V script to retain compatibility with future systemd releases.
937 * Support for the SystemdOptions EFI variable is deprecated.
938 'bootctl systemd-efi-options' will emit a warning when used. It seems
939 that this feature is little-used and it is better to use alternative
940 approaches like credentials and confexts. The plan is to drop support
941 altogether at a later point, but this might be revisited based on
944 * systemd-run's switch --expand-environment= which currently is disabled
945 by default when combined with --scope, will be changed in a future
946 release to be enabled by default.
948 * "systemctl switch-root" is now restricted to initrd transitions only.
950 Transitions between real systems should be done with
951 "systemctl soft-reboot" instead.
953 * The "ip=off" and "ip=none" kernel command line options interpreted by
954 systemd-network-generator will now result in IPv6RA + link-local
955 addressing being disabled, too. Previously DHCP was turned off, but
956 IPv6RA and IPv6 link-local addressing was left enabled.
958 * The NAMING_BRIDGE_MULTIFUNCTION_SLOT naming scheme has been deprecated
961 * SuspendMode=, HibernateState= and HybridSleepState= in the [Sleep]
962 section of systemd-sleep.conf are now deprecated and have no effect.
963 They did not (and could not) take any value other than the respective
964 default. HybridSleepMode= is also deprecated, and will now always use
965 the 'suspend' disk mode.
969 * The way services are spawned has been overhauled. Previously, a
970 process was forked that shared all of the manager's memory (via
971 copy-on-write) while doing all the required setup (e.g.: mount
972 namespaces, CGroup configuration, etc.) before exec'ing the target
973 executable. This was problematic for various reasons: several glibc
974 APIs were called that are not supposed to be used after a fork but
975 before an exec, copy-on-write meant that if either process (the
976 manager or the child) touched a memory page a copy was triggered, and
977 also the memory footprint of the child process was that of the
978 manager, but with the memory limits of the service. From this version
979 onward, the new process is spawned using CLONE_VM and CLONE_VFORK
980 semantics via posix_spawn(3), and it immediately execs a new internal
981 binary, systemd-executor, that receives the configuration to apply
982 via memfd, and sets up the process before exec'ing the target
983 executable. The systemd-executor binary is pinned by file descriptor
984 by each manager instance (system and users), and the reference is
985 updated on daemon-reexec - it is thus important to reexec all running
986 manager instances when the systemd-executor and/or libsystemd*
987 libraries are updated on the filesystem.
989 * Most of the internal process tracking is being changed to use PIDFDs
990 instead of PIDs when the kernel supports it, to improve robustness
993 * A new option SurviveFinalKillSignal= can be used to configure the
994 unit to be skipped in the final SIGTERM/SIGKILL spree on shutdown.
995 This is part of the required configuration to let a unit's processes
996 survive a soft-reboot operation.
998 * System extension images (sysext) can now set
999 EXTENSION_RELOAD_MANAGER=1 in their extension-release files to
1000 automatically reload the service manager (PID 1) when
1001 merging/refreshing/unmerging on boot. Generally, while this can be
1002 used to ship services in system extension images it's recommended to
1003 do that via portable services instead.
1005 * The ExtensionImages= and ExtensionDirectories= options now support
1006 confexts images/directories.
1008 * A new option NFTSet= provides a method for integrating dynamic cgroup
1009 IDs into firewall rules with NFT sets. The benefit of using this
1010 setting is to be able to use control group as a selector in firewall
1011 rules easily and this in turn allows more fine grained filtering.
1012 Also, NFT rules for cgroup matching use numeric cgroup IDs, which
1013 change every time a service is restarted, making them hard to use in
1014 systemd environment.
1016 * A new option CoredumpReceive= can be set for service and scope units,
1017 together with Delegate=yes, to make systemd-coredump on the host
1018 forward core files from processes crashing inside the delegated
1019 CGroup subtree to systemd-coredump running in the container. This new
1020 option is by default used by systemd-nspawn containers that use the
1023 * A new ConditionSecurity=measured-uki option is now available, to ensure
1024 a unit can only run when the system has been booted from a measured UKI.
1026 * MemoryAvailable= now considers physical memory if there are no CGroup
1027 memory limits set anywhere in the tree.
1029 * The $USER environment variable is now always set for services, while
1030 previously it was only set if User= was specified. A new option
1031 SetLoginEnvironment= is now supported to determine whether to also set
1032 $HOME, $LOGNAME, and $SHELL.
1034 * Socket units now support a new pair of
1035 PollLimitBurst=/PollLimitInterval= options to configure a limit on
1036 how often polling events on the file descriptors backing this unit
1037 will be considered within a time window.
1039 * Scope units can now be created using PIDFDs instead of PIDs to select
1040 the processes they should include.
1042 * Sending SIGRTMIN+18 with 0x500 as sigqueue() value will now cause the
1043 manager to dump the list of currently pending jobs.
1045 * If the kernel supports MOVE_MOUNT_BENEATH, the systemctl and
1046 machinectl bind and mount-image verbs will now cause the new mount to
1047 replace the old mount (if any), instead of overmounting it.
1049 * Units now have MemoryPeak, MemorySwapPeak, MemorySwapCurrent and
1050 MemoryZSwapCurrent properties, which respectively contain the values
1051 of the cgroup v2's memory.peak, memory.swap.peak, memory.swap.current
1052 and memory.zswap.current properties. This information is also shown in
1053 "systemctl status" output, if available.
1055 TPM2 Support + Disk Encryption & Authentication:
1057 * systemd-cryptenroll now allows specifying a PCR bank and explicit hash
1058 value in the --tpm2-pcrs= option.
1060 * systemd-cryptenroll now allows specifying a TPM2 key handle (nv
1061 index) to be used instead of the default SRK via the new
1062 --tpm2-seal-key-handle= option.
1064 * systemd-cryptenroll now allows TPM2 enrollment using only a TPM2
1065 public key (in TPM2B_PUBLIC format) – without access to the TPM2
1066 device itself – which enables offline sealing of LUKS images for a
1067 specific TPM2 chip, as long as the SRK public key is known. Pass the
1068 public to the tool via the new --tpm2-device-key= switch.
1070 * systemd-cryptsetup is now installed in /usr/bin/ and is no longer an
1071 internal-only executable.
1073 * The TPM2 Storage Root Key will now be set up, if not already present,
1074 by a new systemd-tpm2-setup.service early boot service. The SRK will
1075 be stored in PEM format and TPM2_PUBLIC format (the latter is useful
1076 for systemd-cryptenroll --tpm2-device-key=, as mentioned above) for
1077 easier access. A new "srk" verb has been added to systemd-analyze to
1078 allow extracting it on demand if it is already set up.
1080 * The internal systemd-pcrphase executable has been renamed to
1083 * The systemd-pcrextend tool gained a new --pcr= switch to override
1084 which PCR to measure into.
1086 * systemd-pcrextend now exposes a Varlink interface at
1087 io.systemd.PCRExtend that can be used to do measurements and event
1090 * TPM measurements are now also written to an event log at
1091 /run/log/systemd/tpm2-measure.log, using a derivative of the TCG
1092 Canonical Event Log format. Previously we'd only log them to the
1093 journal, where they however were subject to rotation and similar.
1095 * A new component "systemd-pcrlock" has been added that allows managing
1096 local TPM2 PCR policies for PCRs 0-7 and similar, which are hard to
1097 predict by the OS vendor because of the inherently local nature of
1098 what measurements they contain, such as firmware versions of the
1099 system and extension cards and suchlike. pcrlock can predict PCR
1100 measurements ahead of time based on various inputs, such as the local
1101 TPM2 event log, GPT partition tables, PE binaries, UKI kernels, and
1102 various other things. It can then pre-calculate a TPM2 policy from
1103 this, which it stores in an TPM2 NV index. TPM2 objects (such as disk
1104 encryption keys) can be locked against this NV index, so that they
1105 are locked against a specific combination of system firmware and
1106 state. Alternatives for each component are supported to allowlist
1107 multiple kernel versions or boot loader version simultaneously
1108 without losing access to the disk encryption keys. The tool can also
1109 be used to analyze and validate the local TPM2 event log.
1110 systemd-cryptsetup, systemd-cryptenroll, systemd-repart have all been
1111 updated to support such policies. There's currently no support for
1112 locking the system's root disk against a pcrlock policy, this will be
1113 added soon. Moreover, it is currently not possible to combine a
1114 pcrlock policy with a signed PCR policy. This component is
1115 experimental and its public interface is subject to change.
1117 systemd-boot, systemd-stub, ukify, bootctl, kernel-install:
1119 * bootctl will now show whether the system was booted from a UKI in its
1122 * systemd-boot and systemd-stub now use different project keys in their
1123 respective SBAT sections, so that they can be revoked individually if
1126 * systemd-boot will no longer load unverified Devicetree blobs when UEFI
1127 SecureBoot is enabled. For more details see:
1128 https://github.com/systemd/systemd/security/advisories/GHSA-6m6p-rjcq-334c
1130 * systemd-boot gained new hotkeys to reboot and power off the system
1131 from the boot menu ("B" and "O"). If the "auto-poweroff" and
1132 "auto-reboot" options in loader.conf are set these entries are also
1133 shown as menu items (which is useful on devices lacking a regular
1136 * systemd-boot gained a new configuration value "menu-disabled" for the
1137 set-timeout option, to allow completely disabling the boot menu,
1138 including the hotkey.
1140 * systemd-boot will now measure the content of loader.conf in TPM2
1143 * systemd-stub will now concatenate the content of all kernel
1144 command-line addons before measuring them in TPM2 PCR 12, in a single
1145 measurement, instead of measuring them individually.
1147 * systemd-stub will now measure and load Devicetree Blob addons, which
1148 are searched and loaded following the same model as the existing
1149 kernel command-line addons.
1151 * systemd-stub will now ignore unauthenticated kernel command line options
1152 passed from systemd-boot when running inside Confidential VMs with UEFI
1155 * systemd-stub will now load a Devicetree blob even if the firmware did
1156 not load any beforehand (e.g.: for ACPI systems).
1158 * ukify is no longer considered experimental, and now ships in /usr/bin/.
1160 * ukify gained a new verb inspect to describe the sections of a UKI and
1161 print the contents of the well-known sections.
1163 * ukify gained a new verb genkey to generate a set of key pairs for
1164 signing UKIs and their PCR data.
1166 * The 90-loaderentry kernel-install hook now supports installing device
1169 * kernel-install now supports the --json=, --root=, --image=, and
1170 --image-policy= options for the inspect verb.
1172 * kernel-install now supports new list and add-all verbs. The former
1173 lists all installed kernel images (if those are available in
1174 /usr/lib/modules/). The latter will install all the kernels it can
1179 * A new option --copy-from= has been added that synthesizes partition
1180 definitions from the given image, which are then applied by the
1181 systemd-repart algorithm.
1183 * A new option --copy-source= has been added, which can be used to specify
1184 a directory to which CopyFiles= is considered relative to.
1186 * New --make-ddi=confext, --make-ddi=sysext, and --make-ddi=portable
1187 options have been added to make it easier to generate these types of
1188 DDIs, without having to provide repart.d definitions for them.
1190 * The dm-verity salt and UUID will now be derived from the specified
1193 * New VerityDataBlockSizeBytes= and VerityHashBlockSizeBytes= can now be
1194 configured in repart.d/ configuration files.
1196 * A new Subvolumes= setting is now supported in repart.d/ configuration
1197 files, to indicate which directories in the target partition should be
1200 * A new --tpm2-device-key= option can be used to lock a disk against a
1201 specific TPM2 public key. This matches the same switch the
1202 systemd-cryptenroll tool now supports (see above).
1206 * The journalctl --lines= parameter now accepts +N to show the oldest N
1207 entries instead of the newest.
1209 * journald now ensures that sealing happens once per epoch, and sets a
1210 new compatibility flag to distinguish old journal files that were
1211 created before this change, for backward compatibility.
1215 * udev will now create symlinks to loopback block devices in the
1216 /dev/disk/by-loop-ref/ directory that are based on the .lo_file_name
1217 string field selected during allocation. The systemd-dissect tool and
1218 the util-linux losetup command now supports a complementing new switch
1219 --loop-ref= for selecting the string. This means a loopback block
1220 device may now be allocated under a caller-chosen reference and can
1221 subsequently be referenced without first having to look up the block
1222 device name the caller ended up with.
1224 * udev also creates symlinks to loopback block devices in the
1225 /dev/disk/by-loop-inode/ directory based on the .st_dev/st_ino fields
1226 of the inode attached to the loopback block device. This means that
1227 attaching a file to a loopback device will implicitly make a handle
1228 available to be found via that file's inode information.
1230 * udevadm info gained support for JSON output via a new --json= flag, and
1231 for filtering output using the same mechanism that udevadm trigger
1234 * The predictable network interface naming logic is extended to include
1235 the SR-IOV-R "representor" information in network interface names.
1236 This feature was intended for v254, but even though the code was
1237 merged, the part that actually enabled the feature was forgotten.
1238 It is now enabled by default and is part of the new "v255" naming
1241 * A new hwdb/rules file has been added that sets the
1242 ID_NET_AUTO_LINK_LOCAL_ONLY=1 udev property on all network interfaces
1243 that should usually only be configured with link-local addressing
1244 (IPv4LL + IPv6LL), i.e. for PC-to-PC cables ("laplink") or
1245 Thunderbolt networking. systemd-networkd and NetworkManager (soon)
1246 will make use of this information to apply an appropriate network
1247 configuration by default.
1249 * The ID_NET_DRIVER property on network interfaces is now set
1250 relatively early in the udev rule set so that other rules may rely on
1251 its use. This is implemented in a new "net-driver" udev built-in.
1255 * The "duid-only" option for DHCPv4 client's ClientIdentifier= setting
1256 is now dropped, as it never worked, hence it should not be used by
1259 * The 'prefixstable' ipv6 address generation mode now considers the SSID
1260 when generating stable addresses, so that a different stable address
1261 is used when roaming between wireless networks. If you already use
1262 'prefixstable' addresses with wireless networks, the stable address
1263 will be changed by the update.
1265 * The DHCPv4 client gained a RapidCommit option, true by default, which
1266 enables RFC4039 Rapid Commit behavior to obtain a lease in a
1267 simplified 2-message exchange instead of the typical 4-message
1268 exchange, if also supported by the DHCP server.
1270 * The DHCPv4 client gained new InitialCongestionWindow= and
1271 InitialAdvertisedReceiveWindow= options for route configurations.
1273 * The DHCPv4 client gained a new RequestAddress= option that allows
1274 to send a preferred IP address in the initial DHCPDISCOVER message.
1276 * The DHCPv4 server and client gained support for IPv6-only mode
1279 * The SendHostname= and Hostname= options are now available for the
1280 DHCPv6 client, independently of the DHCPv4= option, so that these
1281 configuration values can be set independently for each client.
1283 * The DHCPv4 and DHCPv6 client state can now be queried via D-Bus,
1284 including lease information.
1286 * The DHCPv6 client can now be configured to use a custom DUID type.
1288 * .network files gained a new IPv4ReversePathFilter= setting in the
1289 [Network] section, to control sysctl's rp_filter setting.
1291 * .network files gaiend a new HopLimit= setting in the [Route] section,
1292 to configure a per-route hop limit.
1294 * .network files gained a new TCPRetransmissionTimeoutSec= setting in
1295 the [Route] section, to configure a per-route TCP retransmission
1298 * A new directive NFTSet= provides a method for integrating network
1299 configuration into firewall rules with NFT sets. The benefit of using
1300 this setting is that static network configuration or dynamically
1301 obtained network addresses can be used in firewall rules with the
1302 indirection of NFT set types.
1304 * The [IPv6AcceptRA] section supports the following new options:
1305 UsePREF64=, UseHopLimit=, UseICMP6RateLimit=, and NFTSet=.
1307 * The [IPv6SendRA] section supports the following new options:
1308 RetransmitSec=, HopLimit=, HomeAgent=, HomeAgentLifetimeSec=, and
1309 HomeAgentPreference=.
1311 * A new [IPv6PREF64Prefix] set of options, containing Prefix= and
1312 LifetimeSec=, has been introduced to append pref64 options in router
1313 advertisements (RFC8781).
1315 * The network generator now configures the interfaces with only
1316 link-local addressing if "ip=link-local" is specified on the kernel
1319 * The prefix of the configuration files generated by the network
1320 generator from the kernel command line is now prefixed with '70-',
1321 to make them have higher precedence over the default configuration
1324 * Added a new -Ddefault-network=BOOL meson option, that causes more
1325 .network files to be installed as enabled by default. These configuration
1326 files will which match generic setups, e.g. 89-ethernet.network matches
1327 all Ethernet interfaces and enables both DHCPv4 and DHCPv6 clients.
1329 * If a ID_NET_MANAGED_BY= udev property is set on a network device and
1330 it is any other string than "io.systemd.Network" then networkd will
1331 not manage this device. This may be used to allow multiple network
1332 management services to run in parallel and assign ownership of
1333 specific devices explicitly. NetworkManager will soon implement a
1338 * systemctl is-failed now checks the system state if no unit is
1341 * systemctl will now automatically soft-reboot if a new root file system
1342 is found under /run/nextroot/ when a reboot operation is invoked.
1346 * Wall messages now work even when utmp support is disabled, using
1347 systemd-logind to query the necessary information.
1349 * systemd-logind now sends a new PrepareForShutdownWithMetadata D-Bus
1350 signal before shutdown/reboot/soft-reboot that includes additional
1351 information compared to the PrepareForShutdown signal. Currently the
1352 additional information is the type of operation that is about to be
1355 Hibernation & Suspend:
1357 * The kernel and OS versions will no longer be checked on resume from
1360 * Hibernation into swap files backed by btrfs are now
1361 supported. (Previously this was supported only for other file
1366 * A new systemd-vmspawn tool has been added, that aims to provide for VMs
1367 the same interfaces and functionality that systemd-nspawn provides for
1368 containers. For now it supports QEMU as a backend, and exposes some of
1369 its options to the user. This component is experimental and its public
1370 interface is subject to change.
1372 * "systemd-analyze plot" has gained tooltips on each unit name with
1373 related-unit information in its svg output, such as Before=,
1374 Requires=, and similar properties.
1376 * A new varlinkctl tool has been added to allow interfacing with
1377 Varlink services, and introspection has been added to all such
1378 services. This component is experimental and its public interface is
1381 * systemd-sysext and systemd-confext now expose a Varlink service
1382 at io.systemd.sysext.
1384 * portable services now accept confexts as extensions.
1386 * systemd-sysupdate now accepts directories in the MatchPattern= option.
1388 * systemd-run will now output the invocation ID of the launched
1389 transient unit and its peak memory usage.
1391 * systemd-analyze, systemd-tmpfiles, systemd-sysusers, systemd-sysctl,
1392 and systemd-binfmt gained a new --tldr option that can be used instead
1393 of --cat-config to suppress uninteresting configuration lines, such as
1394 comments and whitespace.
1396 * resolvectl gained a new "show-server-state" command that shows
1397 current statistics of the resolver. This is backed by a new
1398 DumpStatistics() Varlink method provided by systemd-resolved.
1400 * systemd-timesyncd will now emit a D-Bus signal when the LinkNTPServers
1403 * vconsole now supports KEYMAP=@kernel for preserving the kernel keymap
1406 * seccomp now supports the LoongArch64 architecture.
1408 * seccomp may now be enabled for services running as a non-root User=
1409 without NoNewPrivileges=yes.
1411 * systemd-id128 now supports a new -P option to show only values. The
1412 combination of -P and --app options is also supported.
1414 * A new pam_systemd_loadkey.so PAM module is now available, which will
1415 automatically fetch the passphrase used by cryptsetup to unlock the
1416 root file system and set it as the PAM authtok. This enables, among
1417 other things, configuring auto-unlock of the GNOME Keyring / KDE
1418 Wallet when autologin is configured.
1420 * Many meson options now use the 'feature' type, which means they
1421 take enabled/disabled/auto as values.
1423 * A new meson option -Dconfigfiledir= can be used to change where
1424 configuration files with default values are installed to.
1426 * Options and verbs in man pages are now tagged with the version they
1427 were first introduced in.
1429 * A new component "systemd-storagetm" has been added, which exposes all
1430 local block devices as NVMe-TCP devices, fully automatically. It's
1431 hooked into a new target unit storage-target-mode.target that is
1432 suppsoed to be booted into via
1433 rd.systemd.unit=storage-target-mode.target on the kernel command
1434 line. This is intended to be used for installers and debugging to
1435 quickly get access to the local disk. It's inspired by MacOS "target
1436 disk mode". This component is experimental and its public interface is
1439 * A new component "systemd-bsod" has been added, which can show logged
1440 error messages full screen, if they have a log level of LOG_EMERG log
1441 level. This component is experimental and its public interface is
1444 * The systemd-dissect tool's --with command will now set the
1445 $SYSTEMD_DISSECT_DEVICE environment variable to the block device it
1446 operates on for the invoked process.
1448 * The systemd-mount tool gained a new --tmpfs switch for mounting a new
1449 'tmpfs' instance. This is useful since it does so via .mount units
1450 and thus can be executed remotely or in containers.
1452 * The various tools in systemd that take "verbs" (such as systemctl,
1453 loginctl, machinectl, …) now will suggest a close verb name in case
1454 the user specified an unrecognized one.
1456 * libsystemd now exports a new function sd_id128_get_app_specific()
1457 that generates "app-specific" 128bit IDs from any ID. It's similar to
1458 sd_id128_get_machine_app_specific() and
1459 sd_id128_get_boot_app_specific() but takes the ID to base calculation
1460 on as input. This new functionality is also exposed in the
1461 "systemd-id128" tool where you can now combine --app= with `show`.
1463 * All tools that parse timestamps now can also parse RFC3339 style
1464 timestamps that include the "T" and Z" characters.
1466 * New documentation has been added:
1468 https://systemd.io/FILE_DESCRIPTOR_STORE
1469 https://systemd.io/TPM2_PCR_MEASUREMENTS
1470 https://systemd.io/MOUNT_REQUIREMENTS
1472 * The codebase now recognizes the suffix .confext.raw and .sysext.raw
1473 as alternative to the .raw suffix generally accepted for DDIs. It is
1474 recommended to name configuration extensions and system extensions
1475 with such suffixes, to indicate their purpose in the name.
1477 * The sd-device API gained a new function
1478 sd_device_enumerator_add_match_property_required() which allows
1479 configuring matches on properties that are strictly required. This is
1480 different from the existing sd_device_enumerator_add_match_property()
1481 matches of which one one needs to apply.
1483 * The MAC address the veth side of an nspawn container shall get
1484 assigned may now be controlled via the $SYSTEMD_NSPAWN_NETWORK_MAC
1485 environment variable.
1487 * The libiptc dependency is now implemented via dlopen(), so that tools
1488 such as networkd and nspawn no longer have a hard dependency on the
1489 shared library when compiled with support for libiptc.
1491 * New rpm macros have been added: %systemd_user_daemon_reexec does
1492 daemon-reexec for all user managers, and %systemd_postun_with_reload
1493 and %systemd_user_postun_with_reload do a reload for system and user
1496 * coredumpctl now propagates SIGTERM to the debugger process.
1498 Contributions from: 김인수, Abderrahim Kitouni, Adam Goldman,
1499 Adam Williamson, Alexandre Peixoto Ferreira, Alex Hudspith,
1500 Alvin Alvarado, André Paiusco, Antonio Alvarez Feijoo,
1501 Anton Lundin, Arian van Putten, Arseny Maslennikov, Arthur Shau,
1502 Balázs Úr, beh_10257, Benjamin Peterson, Bertrand Jacquin,
1503 Brian Norris, Charles Lee, Cheng-Chia Tseng, Chris Patterson,
1504 Christian Hergert, Christian Hesse, Christian Kirbach,
1505 Clayton Craft, commondservice, cunshunxia, Curtis Klein, cvlc12,
1506 Daan De Meyer, Daniele Medri, Daniel P. Berrangé, Daniel Rusek,
1507 Daniel Thompson, Dan Nicholson, Dan Streetman, David Rheinsberg,
1508 David Santamaría Rogado, David Tardon, dependabot[bot],
1509 Diego Viola, Dmitry V. Levin, Emanuele Giuseppe Esposito,
1510 Emil Renner Berthing, Emil Velikov, Etienne Dechamps, Fabian Vogt,
1511 felixdoerre, Felix Dörre, Florian Schmaus, Franck Bui,
1512 Frantisek Sumsal, G2-Games, Gioele Barabucci, Hugo Carvalho,
1513 huyubiao, Iago López Galeiras, IllusionMan1212, Jade Lovelace,
1514 janana, Jan Janssen, Jan Kuparinen, Jan Macku, Jeremy Fleischman,
1515 Jin Liu, jjimbo137, Joerg Behrmann, Johannes Segitz, Jordan Rome,
1516 Jordan Williams, Julien Malka, Juno Computers, Khem Raj, khm,
1517 Kingbom Dou, Kiran Vemula, Krzesimir Nowak, Laszlo Gombos,
1518 Lennart Poettering, linuxlion, Luca Boccassi, Lucas Adriano Salles,
1519 Lukas, Lukáš Nykrýn, Maanya Goenka, Maarten, Malte Poll,
1520 Marc Pervaz Boocha, Martin Beneš, Martin Joerg, Martin Wilck,
1521 Mathieu Tortuyaux, Matthias Schiffer, Maxim Mikityanskiy,
1522 Max Kellermann, Michael A Cassaniti, Michael Biebl, Michael Kuhn,
1523 Michael Vasseur, Michal Koutný, Michal Sekletár, Mike Yuan,
1524 Milton D. Miller II, mordner, msizanoen, NAHO, Nandakumar Raghavan,
1525 Neil Wilson, Nick Rosbrook, Nils K, NRK, Oğuz Ersen,
1526 Omojola Joshua, onenowy, Paul Meyer, Paymon MARANDI, pelaufer,
1527 Peter Hutterer, PhylLu, Pierre GRASSER, Piotr Drąg, Priit Laes,
1528 Rahil Bhimjiani, Raito Bezarius, Raul Cheleguini, Reto Schneider,
1529 Richard Maw, Robby Red, RoepLuke, Roland Hieber, Roland Singer,
1530 Ronan Pigott, Sam James, Sam Leonard, Sergey A, Susant Sahani,
1531 Sven Joachim, Tad Fisher, Takashi Sakamoto, Thorsten Kukuk, Tj,
1532 Tomasz Świątek, Topi Miettinen, Valentin David,
1533 Valentin Lefebvre, Victor Westerhuis, Vincent Haupert,
1534 Vishal Chillara Srinivas, Vito Caputo, Warren, Weblate,
1535 Xiaotian Wu, xinpeng wang, Yaron Shahrabani, Yo-Jung Lin,
1536 Yu Watanabe, Zbigniew Jędrzejewski-Szmek, zeroskyx,
1537 Дамјан Георгиевски, наб
1539 — Edinburgh, 2023-12-06
1543 Announcements of Future Feature Removals and Incompatible Changes:
1545 * The next release (v255) will remove support for split-usr (/usr/
1546 mounted separately during late boot, instead of being mounted by the
1547 initrd before switching to the rootfs) and unmerged-usr (parallel
1548 directories /bin/ and /usr/bin/, /lib/ and /usr/lib/, …). For more
1550 https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html
1552 * We intend to remove cgroup v1 support from a systemd release after
1553 the end of 2023. If you run services that make explicit use of
1554 cgroup v1 features (i.e. the "legacy hierarchy" with separate
1555 hierarchies for each controller), please implement compatibility with
1556 cgroup v2 (i.e. the "unified hierarchy") sooner rather than later.
1557 Most of Linux userspace has been ported over already.
1559 * Support for System V service scripts is now deprecated and will be
1560 removed in a future release. Please make sure to update your software
1561 *now* to include a native systemd unit file instead of a legacy
1562 System V script to retain compatibility with future systemd releases.
1564 * Support for the SystemdOptions EFI variable is deprecated.
1565 'bootctl systemd-efi-options' will emit a warning when used. It seems
1566 that this feature is little-used and it is better to use alternative
1567 approaches like credentials and confexts. The plan is to drop support
1568 altogether at a later point, but this might be revisited based on
1571 * EnvironmentFile= now treats the line following a comment line
1572 trailing with escape as a non comment line. For details, see:
1573 https://github.com/systemd/systemd/issues/27975
1575 * PrivateNetwork=yes and NetworkNamespacePath= now imply
1576 PrivateMounts=yes unless PrivateMounts=no is explicitly specified.
1578 * Behaviour of sandboxing options for the per-user service manager
1579 units has changed. They now imply PrivateUsers=yes, which means user
1580 namespaces will be implicitly enabled when a sandboxing option is
1581 enabled in a user unit. Enabling user namespaces has the drawback
1582 that system users will no longer be visible (and processes/files will
1583 appear as owned by 'nobody') in the user unit.
1585 By definition a sandboxed user unit should run with reduced
1586 privileges, so impact should be small. This will remove a great
1587 source of confusion that has been reported by users over the years,
1588 due to how these options require an extra setting to be manually
1589 enabled when used in the per-user service manager, which is not
1590 needed in the system service manager. For more details, see:
1591 https://lists.freedesktop.org/archives/systemd-devel/2022-December/048682.html
1593 * systemd-run's switch --expand-environment= which currently is disabled
1594 by default when combined with --scope, will be changed in a future
1595 release to be enabled by default.
1597 Security Relevant Changes:
1599 * pam_systemd will now by default pass the CAP_WAKE_ALARM ambient
1600 process capability to invoked session processes of regular users on
1601 local seats (as well as to systemd --user), unless configured
1602 otherwise via data from JSON user records, or via the PAM module's
1603 parameter list. This is useful in order allow desktop tools such as
1604 GNOME's Alarm Clock application to set a timer for
1605 CLOCK_REALTIME_ALARM that wakes up the system when it elapses. A
1606 per-user service unit file may thus use AmbientCapability= to pass
1607 the capability to invoked processes. Note that this capability is
1608 relatively narrow in focus (in particular compared to other process
1609 capabilities such as CAP_SYS_ADMIN) and we already — by default —
1610 permit more impactful operations such as system suspend to local
1615 * Memory limits that apply while the unit is activating are now
1616 supported. Previously IO and CPU settings were already supported via
1617 StartupCPUWeight= and similar. The same logic has been added for the
1618 various manager and unit memory settings (DefaultStartupMemoryLow=,
1619 StartupMemoryLow=, StartupMemoryHigh=, StartupMemoryMax=,
1620 StartupMemorySwapMax=, StartupMemoryZSwapMax=).
1622 * The service manager gained support for enqueuing POSIX signals to
1623 services that carry an additional integer value, exposing the
1624 sigqueue() system call. This is accessible via new D-Bus calls
1625 org.freedesktop.systemd1.Manager.QueueSignalUnit() and
1626 org.freedesktop.systemd1.Unit.QueueSignal(), as well as in systemctl
1627 via the new --kill-value= option.
1629 * systemctl gained a new "list-paths" verb, which shows all currently
1630 active .path units, similarly to how "systemctl list-timers" shows
1631 active timers, and "systemctl list-sockets" shows active sockets.
1633 * systemctl gained a new --when= switch which is honoured by the various
1634 forms of shutdown (i.e. reboot, kexec, poweroff, halt) and allows
1635 scheduling these operations by time, similar in fashion to how this
1636 has been supported by SysV shutdown.
1638 * If MemoryDenyWriteExecute= is enabled for a service and the kernel
1639 supports the new PR_SET_MDWE prctl() call, it is used instead of the
1640 seccomp()-based system call filter to achieve the same effect.
1642 * A new set of kernel command line options is now understood:
1643 systemd.tty.term.<name>=, systemd.tty.rows.<name>=,
1644 systemd.tty.columns.<name>= allow configuring the TTY type and
1645 dimensions for the tty specified via <name>. When systemd invokes a
1646 service on a tty (via TTYName=) it will look for these and configure
1647 the TTY accordingly. This is particularly useful in VM environments
1648 to propagate host terminal settings into the appropriate TTYs of the
1651 * A new RootEphemeral= setting is now understood in service units. It
1652 takes a boolean argument. If enabled for services that use RootImage=
1653 or RootDirectory= an ephemeral copy of the disk image or directory
1654 tree is made when the service is started. It is removed automatically
1655 when the service is stopped. That ephemeral copy is made using
1656 btrfs/xfs reflinks or btrfs snapshots, if available.
1658 * The service activation logic gained new settings RestartSteps= and
1659 RestartMaxDelaySec= which allow exponentially-growing restart
1660 intervals for Restart=.
1662 * The service activation logic gained a new setting RestartMode= which
1663 can be set to 'direct' to skip the inactive/failed states when
1664 restarting, so that dependent units are not notified until the service
1665 converges to a final (successful or failed) state. For example, this
1666 means that OnSuccess=/OnFailure= units will not be triggered until the
1667 service state has converged.
1669 * PID 1 will now automatically load the virtio_console kernel module
1670 during early initialization if running in a suitable VM. This is done
1671 so that early-boot logging can be written to the console if available.
1673 * Similarly, virtio-vsock support is loaded early in suitable VM
1674 environments. PID 1 will send sd_notify() notifications via AF_VSOCK
1675 to the VMM if configured, thus loading this early is beneficial.
1677 * A new verb "fdstore" has been added to systemd-analyze to show the
1678 current contents of the file descriptor store of a unit. This is
1679 backed by a new D-Bus call DumpUnitFileDescriptorStore() provided by
1680 the service manager.
1682 * The service manager will now set a new $FDSTORE environment variable
1683 when invoking processes for services that have the file descriptor
1686 * A new service option FileDescriptorStorePreserve= has been added that
1687 allows tuning the lifecycle of the per-service file descriptor store.
1688 If set to "yes", the entries in the fd store are retained even after
1689 the service has been fully stopped.
1691 * The "systemctl clean" command may now be used to clear the fdstore of
1694 * Unit *.preset files gained a new directive "ignore", in addition to
1695 the existing "enable" and "disable". As the name suggests, matching
1696 units are left unchanged, i.e. neither enabled nor disabled.
1698 * Service units gained a new setting DelegateSubgroup=. It takes the
1699 name of a sub-cgroup to place any processes the service manager forks
1700 off in. Previously, the service manager would place all service
1701 processes directly in the top-level cgroup it created for the
1702 service. This usually meant that main process in a service with
1703 delegation enabled would first have to create a subgroup and move
1704 itself down into it, in order to not conflict with the "no processes
1705 in inner cgroups" rule of cgroup v2. With this option, this step is
1706 now handled by PID 1.
1708 * The service manager will now look for .upholds/ directories,
1709 similarly to the existing support for .wants/ and .requires/
1710 directories. Symlinks in this directory result in Upholds=
1713 The [Install] section of unit files gained support for a new
1714 UpheldBy= directive to generate .upholds/ symlinks automatically when
1717 * The service manager now supports a new kernel command line option
1718 systemd.default_device_timeout_sec=, which may be used to override
1719 the default timeout for .device units.
1721 * A new "soft-reboot" mechanism has been added to the service manager.
1722 A "soft reboot" is similar to a regular reboot, except that it
1723 affects userspace only: the service manager shuts down any running
1724 services and other units, then optionally switches into a new root
1725 file system (mounted to /run/nextroot/), and then passes control to a
1726 systemd instance in the new file system which then starts the system
1727 up again. The kernel is not rebooted and neither is the hardware,
1728 firmware or boot loader. This provides a fast, lightweight mechanism
1729 to quickly reset or update userspace, without the latency that a full
1730 system reset involves. Moreover, open file descriptors may be passed
1731 across the soft reboot into the new system where they will be passed
1732 back to the originating services. This allows pinning resources
1733 across the reboot, thus minimizing grey-out time further. This new
1734 reboot mechanism is accessible via the new "systemctl soft-reboot"
1737 * Services using RootDirectory= or RootImage= will now have read-only
1738 access to a copy of the host's os-release file under
1739 /run/host/os-release, which will be kept up-to-date on 'soft-reboot'.
1740 This was already the case for Portable Services, and the feature has
1741 now been extended to all services that do not run off the host's
1744 * A new service setting MemoryKSM= has been added to enable kernel
1745 same-page merging individually for services.
1747 * A new service setting ImportCredentials= has been added that augments
1748 LoadCredential= and LoadCredentialEncrypted= and searches for
1749 credentials to import from the system, and supports globbing.
1751 * A new job mode "restart-dependencies" has been added to the service
1752 manager (exposed via systemctl --job-mode=). It is only valid when
1753 used with "start" jobs, and has the effect that the "start" job will
1754 be propagated as "restart" jobs to currently running units that have
1755 a BindsTo= or Requires= dependency on the started unit.
1757 * A new verb "whoami" has been added to "systemctl" which determines as
1758 part of which unit the command is being invoked. It writes the unit
1759 name to standard output. If one or more PIDs are specified reports
1760 the unit names the processes referenced by the PIDs belong to.
1762 * The system and service credential logic has been improved: there's
1763 now a clearly defined place where system provisioning tools running
1764 in the initrd can place credentials that will be imported into the
1765 system's set of credentials during the initrd → host transition: the
1766 /run/credentials/@initrd/ directory. Once the credentials placed
1767 there are imported into the system credential set they are deleted
1768 from this directory, and the directory itself is deleted afterwards
1771 * A new kernel command line option systemd.set_credential_binary= has
1772 been added, that is similar to the pre-existing
1773 systemd.set_credential= but accepts arbitrary binary credential data,
1774 encoded in Base64. Note that the kernel command line is not a
1775 recommend way to transfer credentials into a system, since it is
1776 world-readable from userspace.
1778 * The default machine ID to use may now be configured via the
1779 system.machine_id system credential. It will only be used if no
1780 machine ID was set yet on the host.
1782 * On Linux kernel 6.4 and newer system and service credentials will now
1783 be placed in a tmpfs instance that has the "noswap" mount option
1784 set. Previously, a "ramfs" instance was used. By switching to tmpfs
1785 ACL support and overall size limits can now be enforced, without
1786 compromising on security, as the memory is never paged out either
1789 * The service manager now can detect when it is running in a
1790 'Confidential Virtual Machine', and a corresponding 'cvm' value is now
1791 accepted by ConditionSecurity= for units that want to conditionalize
1792 themselves on this. systemd-detect-virt gained new 'cvm' and
1793 '--list-cvm' switches to respectively perform the detection or list
1794 all known flavours of confidential VM, depending on the vendor. The
1795 manager will publish a 'ConfidentialVirtualization' D-Bus property,
1796 and will also set a SYSTEMD_CONFIDENTIAL_VIRTUALIZATION= environment
1797 variable for unit generators. Finally, udev rules can match on a new
1798 'cvm' key that will be set when in a confidential VM.
1799 Additionally, when running in a 'Confidential Virtual Machine', SMBIOS
1800 strings and QEMU's fw_cfg protocol will not be used to import
1801 credentials and kernel command line parameters by the system manager,
1802 systemd-boot and systemd-stub, because the hypervisor is considered
1803 untrusted in this particular setting.
1807 * The sd-journal API gained a new call sd_journal_get_seqnum() to
1808 retrieve the current log record's sequence number and sequence number
1809 ID, which allows applications to order records the same way as
1810 journal does internally. The sequence number is now also exported in
1811 the JSON and "export" output of the journal.
1813 * journalctl gained a new switch --truncate-newline. If specified
1814 multi-line log records will be truncated at the first newline,
1815 i.e. only the first line of each log message will be shown.
1817 * systemd-journal-upload gained support for --namespace=, similar to
1818 the switch of the same name of journalctl.
1822 * systemd-repart's drop-in files gained a new ExcludeFiles= option which
1823 may be used to exclude certain files from the effect of CopyFiles=.
1825 * systemd-repart's Verity support now implements the Minimize= setting
1826 to minimize the size of the resulting partition.
1828 * systemd-repart gained a new --offline= switch, which may be used to
1829 control whether images shall be built "online" or "offline",
1830 i.e. whether to make use of kernel facilities such as loopback block
1831 devices and device mapper or not.
1833 * If systemd-repart is told to populate a newly created ESP or XBOOTLDR
1834 partition with some files, it will now default to VFAT rather than
1837 * systemd-repart gained a new --architecture= switch. If specified, the
1838 per-architecture GPT partition types (i.e. the root and /usr/
1839 partitions) configured in the partition drop-in files are
1840 automatically adjusted to match the specified CPU architecture, in
1841 order to simplify cross-architecture DDI building.
1843 * systemd-repart will now default to a minimum size of 300MB for XFS
1844 filesystems if no size parameter is specified. This matches what the
1845 XFS tools (xfsprogs) can support.
1847 systemd-boot, systemd-stub, ukify, bootctl, kernel-install:
1849 * gnu-efi is no longer required to build systemd-boot and systemd-stub.
1850 Instead, pyelftools is now needed, and it will be used to perform the
1851 ELF -> PE relocations at build time.
1853 * bootctl gained a new switch --print-root-device/-R that prints the
1854 block device the root file system is backed by. If specified twice,
1855 it returns the whole disk block device (as opposed to partition block
1856 device) the root file system is on. It's useful for invocations such
1857 as "cfdisk $(bootctl -RR)" to quickly show the partition table of the
1860 * systemd-stub will now look for the SMBIOS Type 1 field
1861 "io.systemd.stub.kernel-cmdline-extra" and append its value to the
1862 kernel command line it invokes. This is useful for VMMs such as qemu
1863 to pass additional kernel command lines into the system even when
1864 booting via full UEFI. The contents of the field are measured into
1867 * The KERNEL_INSTALL_LAYOUT= setting for kernel-install gained a new
1868 value "auto". With this value, a kernel will be automatically
1869 analyzed, and if it qualifies as UKI, it will be installed as if the
1870 setting was to set to "uki", otherwise as "bls".
1872 * systemd-stub can now optionally load UEFI PE "add-on" images that may
1873 contain additional kernel command line information. These "add-ons"
1874 superficially look like a regular UEFI executable, and are expected
1875 to be signed via SecureBoot/shim. However, they do not actually
1876 contain code, but instead a subset of the PE sections that UKIs
1877 support. They are supposed to provide a way to extend UKIs with
1878 additional resources in a secure and authenticated way. Currently,
1879 only the .cmdline PE section may be used in add-ons, in which case
1880 any specified string is appended to the command line embedded into
1881 the UKI itself. A new 'addon<EFI-ARCH>.efi.stub' is now provided that
1882 can be used to trivially create addons, via 'ukify' or 'objcopy'. In
1883 the future we expect other sections to be made extensible like this as
1886 * ukify has been updated to allow building these UEFI PE "add-on"
1887 images, using the new 'addon<EFI-ARCH>.efi.stub'.
1889 * ukify now accepts SBAT information to place in the .sbat PE section
1890 of UKIs and addons. If a UKI is built the SBAT information from the
1891 inner kernel is merged with any SBAT information associated with
1892 systemd-stub and the SBAT data specified on the ukify command line.
1894 * The kernel-install script has been rewritten in C, and reuses much of
1895 the infrastructure of existing tools such as bootctl. It also gained
1896 --esp-path= and --boot-path= options to override the path to the ESP,
1897 and the $BOOT partition. Options --make-entry-directory= and
1898 --entry-token= have been added as well, similar to bootctl's options
1901 * A new kernel-install plugin 60-ukify has been added which will
1902 combine kernel/initrd locally into a UKI and optionally sign them
1903 with a local key. This may be used to switch to UKI mode even on
1904 systems where a local kernel or initrd is used. (Typically UKIs are
1905 built and signed by the vendor.)
1907 * The ukify tool now supports "pesign" in addition to the pre-existing
1908 "sbsign" for signing UKIs.
1910 * systemd-measure and systemd-stub now look for the .uname PE section
1911 that should contain the kernel's "uname -r" string.
1913 * systemd-measure and ukify now calculate expected PCR hashes for a UKI
1914 "offline", i.e. without access to a TPM (physical or
1917 Memory Pressure & Control:
1919 * The sd-event API gained new calls sd_event_add_memory_pressure(),
1920 sd_event_source_set_memory_pressure_type(),
1921 sd_event_source_set_memory_pressure_period() to create and configure
1922 an event source that is called whenever the OS signals memory
1923 pressure. Another call sd_event_trim_memory() is provided that
1924 compacts the process' memory use by releasing allocated but unused
1925 malloc() memory back to the kernel. Services can also provide their
1926 own custom callback to do memory trimming. This should improve system
1927 behaviour under memory pressure, as on Linux traditionally provided
1928 no mechanism to return process memory back to the kernel if the
1929 kernel was under memory pressure. This makes use of the kernel's PSI
1930 interface. Most long-running services in systemd have been hooked up
1931 with this, and in particular systems with low memory should benefit
1934 * Service units gained new settings MemoryPressureWatch= and
1935 MemoryPressureThresholdSec= to configure the PSI memory pressure
1936 logic individually. If these options are used, the
1937 $MEMORY_PRESSURE_WATCH and $MEMORY_PRESSURE_WRITE environment
1938 variables will be set for the invoked processes to inform them about
1939 the requested memory pressure behaviour. (This is used by the
1940 aforementioned sd-events API additions, if set.)
1942 * systemd-analyze gained a new "malloc" verb that shows the output
1943 generated by glibc's malloc_info() on services that support it. Right
1944 now, only the service manager has been updated accordingly. This
1945 call requires privileges.
1947 User & Session Management:
1949 * The sd-login API gained a new call sd_session_get_username() to
1950 return the user name of the owner of a login session. It also gained
1951 a new call sd_session_get_start_time() to retrieve the time the login
1952 session started. A new call sd_session_get_leader() has been added to
1953 return the PID of the "leader" process of a session. A new call
1954 sd_uid_get_login_time() returns the time since the specified user has
1955 most recently been continuously logged in with at least one session.
1957 * JSON user records gained a new set of fields capabilityAmbientSet and
1958 capabilityBoundingSet which contain a list of POSIX capabilities to
1959 set for the logged in users in the ambient and bounding sets,
1960 respectively. homectl gained the ability to configure these two sets
1961 for users via --capability-bounding-set=/--capability-ambient-set=.
1963 * pam_systemd learnt two new module options
1964 default-capability-bounding-set= and default-capability-ambient-set=,
1965 which configure the default bounding sets for users as they are
1966 logging in, if the JSON user record doesn't specify this explicitly
1967 (see above). The built-in default for the ambient set now contains
1968 the CAP_WAKE_ALARM, thus allowing regular users who may log in
1969 locally to resume from a system suspend via a timer.
1971 * The Session D-Bus objects systemd-logind gained a new SetTTY() method
1972 call to update the TTY of a session after it has been allocated. This
1973 is useful for SSH sessions which are typically allocated first, and
1974 for which a TTY is added later.
1976 * The sd-login API gained a new call sd_pid_notifyf_with_fds() which
1977 combines the various other sd_pid_notify() flavours into one: takes a
1978 format string, an overriding PID, and a set of file descriptors to
1979 send. It also gained a new call sd_pid_notify_barrier() call which is
1980 equivalent to sd_notify_barrier() but allows the originating PID to
1983 * "loginctl list-users" and "loginctl list-sessions" will now show the
1984 state of each logged in user/session in their tabular output. It will
1985 also show the current idle state of sessions.
1989 * systemd-dissect will now show the intended CPU architecture of an
1992 * systemd-dissect will now install itself as mount helper for the "ddi"
1993 pseudo-file system type. This means you may now mount DDIs directly
1994 via /bin/mount or /etc/fstab, making full use of embedded Verity
1995 information and all other DDI features.
1997 Example: mount -t ddi myimage.raw /some/where
1999 * The systemd-dissect tool gained the new switches --attach/--detach to
2000 attach/detach a DDI to a loopback block device without mounting it.
2001 It will automatically derive the right sector size from the image
2002 and set up Verity and similar, but not mount the file systems in it.
2004 * When systemd-gpt-auto-generator or the DDI mounting logic mount an
2005 ESP or XBOOTLDR partition the MS_NOSYMFOLLOW mount option is now
2006 implied. Given that these file systems are typically untrusted, this
2007 should make mounting them automatically have less of a security
2010 * All tools that parse DDIs (such as systemd-nspawn, systemd-dissect,
2011 systemd-tmpfiles, …) now understand a new switch --image-policy= which
2012 takes a string encoding image dissection policy. With this mechanism
2013 automatic discovery and use of specific partition types and the
2014 cryptographic requirements on the partitions (Verity, LUKS, …) can be
2015 restricted, permitting better control of the exposed attack surfaces
2016 when mounting disk images. systemd-gpt-auto-generator will honour such
2017 an image policy too, configurable via the systemd.image_policy= kernel
2018 command line option. Unit files gained the RootImagePolicy=,
2019 MountImagePolicy= and ExtensionImagePolicy= to configure the same for
2020 disk images a service runs off.
2022 * systemd-analyze gained a new verb "image-policy" to validate and
2023 parse image policy strings.
2025 * systemd-dissect gained support for a new --validate switch to
2026 superficially validate DDI structure, and check whether a specific
2027 image policy allows the DDI.
2029 * systemd-dissect gained support for a new --mtree-hash switch to
2030 optionally disable calculating mtree hashes, which can be slow on
2033 * systemd-dissect --copy-to, --copy-from, --list and --mtree switches
2034 are now able to operate on directories too, other than images.
2038 * networkd's GENEVE support as gained a new .network option
2039 InheritInnerProtocol=.
2041 * The [Tunnel] section in .netdev files has gained a new setting
2042 IgnoreDontFragment for controlling the IPv4 "DF" flag of datagrams.
2044 * A new global IPv6PrivacyExtensions= setting has been added that
2045 selects the default value of the per-network setting of the same
2048 * The predictable network interface naming logic was extended to
2049 include SR-IOV-R "representor" information in network interface
2050 names. Unfortunately, this feature was not enabled by default and can
2051 only be enabled at compilation time by setting
2052 -Ddefault-net-naming-scheme=v254.
2054 * The DHCPv4 + DHCPv6 + IPv6 RA logic in networkd gained support for
2055 the RFC8910 captive portal option.
2059 * udevadm gained the new "verify" verb for validating udev rules files
2062 * udev gained a new tool "iocost" that can be used to configure QoS IO
2063 cost data based on hwdb information onto suitable block devices. Also
2064 see https://github.com/iocost-benchmark/iocost-benchmarks.
2066 TPM2 Support + Disk Encryption & Authentication:
2068 * systemd-cryptenroll/systemd-cryptsetup will now install a TPM2 SRK
2069 ("Storage Root Key") as first step in the TPM2, and then use that
2070 for binding FDE to, if TPM2 support is used. This matches
2071 recommendations of TCG (see
2072 https://trustedcomputinggroup.org/wp-content/uploads/TCG-TPM-v2.0-Provisioning-Guidance-Published-v1r1.pdf)
2074 * systemd-cryptenroll and other tools that take TPM2 PCR parameters now
2075 understand textual identifiers for these PCRs.
2077 * systemd-veritysetup + /etc/veritytab gained support for a series of
2078 new options: hash-offset=, superblock=, format=, data-block-size=,
2079 hash-block-size=, data-blocks=, salt=, uuid=, hash=, fec-device=,
2080 fec-offset=, fec-roots= to configure various aspects of a Verity
2083 * systemd-cryptsetup + /etc/crypttab gained support for a new
2084 veracrypt-pim= option for setting the Personal Iteration Multiplier
2085 of veracrypt volumes.
2087 * systemd-integritysetup + /etc/integritytab gained support for a new
2088 mode= setting for controlling the dm-integrity mode (journal, bitmap,
2089 direct) for the volume.
2091 * systemd-analyze gained a new verb "pcrs" that shows the known TPM PCR
2092 registers, their symbolic names and current values.
2096 * The ACL support in tmpfiles.d/ has been updated: if an uppercase "X"
2097 access right is specified this is equivalent to "x" but only if the
2098 inode in question already has the executable bit set for at least
2099 some user/group. Otherwise the "x" bit will be turned off.
2101 * tmpfiles.d/'s C line type now understands a new modifier "+": a line
2102 with C+ will result in a "merge" copy, i.e. all files of the source
2103 tree are copied into the target tree, even if that tree already
2104 exists, resulting in a combined tree of files already present in the
2105 target tree and those copied in.
2107 * systemd-tmpfiles gained a new --graceful switch. If specified lines
2108 with unknown users/groups will silently be skipped.
2112 * systemd-notify gained two new options --fd= and --fdname= for sending
2113 arbitrary file descriptors to the service manager (while specifying an
2114 explicit name for it).
2116 * systemd-notify gained a new --exec switch, which makes it execute the
2117 specified command line after sending the requested messages. This is
2118 useful for sending out READY=1 first, and then continuing invocation
2119 without changing process ID, so that the tool can be nicely used
2120 within an ExecStart= line of a unit file that uses Type=notify.
2122 sd-event + sd-bus APIs:
2124 * The sd-event API gained a new call sd_event_source_leave_ratelimit()
2125 which may be used to explicitly end a rate-limit state an event
2126 source might be in, resetting all rate limiting counters.
2128 * When the sd-bus library is used to make connections to AF_UNIX D-Bus
2129 sockets, it will now encode the "description" set via
2130 sd_bus_set_description() into the source socket address. It will also
2131 look for this information when accepting a connection. This is useful
2132 to track individual D-Bus connections on a D-Bus broker for debug
2137 * systemd-resolved gained a new resolved.conf setting
2138 StateRetentionSec= which may be used to retain cached DNS records
2139 even after their nominal TTL, and use them in case upstream DNS
2140 servers cannot be reached. This can be used to make name resolution
2141 more resilient in case of network problems.
2143 * resolvectl gained a new verb "show-cache" to show the current cache
2144 contents of systemd-resolved. This verb communicates with the
2145 systemd-resolved daemon and requires privileges.
2149 * Meson >= 0.60.0 is now required to build systemd.
2151 * The default keymap to apply may now be chosen at build-time via the
2152 new -Ddefault-keymap= meson option.
2154 * Most of systemd's long-running services now have a generic handler of
2155 the SIGRTMIN+18 signal handler which executes various operations
2156 depending on the sigqueue() parameter sent along. For example, values
2157 0x100…0x107 allow changing the maximum log level of such
2158 services. 0x200…0x203 allow changing the log target of such
2159 services. 0x300 make the services trim their memory similarly to the
2160 automatic PSI-triggered action, see above. 0x301 make the services
2161 output their malloc_info() data to the logs.
2163 * machinectl gained new "edit" and "cat" verbs for editing .nspawn
2164 files, inspired by systemctl's verbs of the same name which edit unit
2165 files. Similarly, networkctl gained the same verbs for editing
2166 .network, .netdev, .link files.
2168 * A new syscall filter group "@sandbox" has been added that contains
2169 syscalls for sandboxing system calls such as those for seccomp and
2172 * New documentation has been added:
2174 https://systemd.io/COREDUMP
2175 https://systemd.io/MEMORY_PRESSURE
2178 * systemd-firstboot gained a new --reset option. If specified, the
2179 settings in /etc/ it knows how to initialize are reset.
2181 * systemd-sysext is now a multi-call binary and is also installed under
2182 the systemd-confext alias name (via a symlink). When invoked that way
2183 it will operate on /etc/ instead of /usr/ + /opt/. It thus becomes a
2184 powerful, atomic, secure configuration management of sorts, that
2185 locally can merge configuration from multiple confext configuration
2186 images into a single immutable tree.
2188 * The --network-macvlan=, --network-ipvlan=, --network-interface=
2189 switches of systemd-nspawn may now optionally take the intended
2190 network interface inside the container.
2192 * All our programs will now send an sd_notify() message with their exit
2193 status in the EXIT_STATUS= field when exiting, using the usual
2194 protocol, including PID 1. This is useful for VMMs and container
2195 managers to collect an exit status from a system as it shuts down, as
2196 set via "systemctl exit …". This is particularly useful in test cases
2197 and similar, as invocations via a VM can now nicely propagate an exit
2198 status to the host, similar to local processes.
2200 * systemd-run gained a new switch --expand-environment=no to disable
2201 server-side environment variable expansion in specified command
2202 lines. Expansion defaults to enabled for all execution types except
2203 --scope, where it defaults to off (and prints a warning) for backward
2204 compatibility reasons. --scope will be flipped to enabled by default
2205 too in a future release. If you are using --scope and passing a '$'
2206 character in the payload you should start explicitly using
2207 --expand-environment=yes/no according to the use case.
2209 * The systemd-system-update-generator has been updated to also look for
2210 the special flag file /etc/system-update in addition to the existing
2211 support for /system-update to decide whether to enter system update
2214 * The /dev/hugepages/ file system is now mounted with nosuid + nodev
2215 mount options by default.
2217 * systemd-fstab-generator now understands two new kernel command line
2218 options systemd.mount-extra= and systemd.swap-extra=, which configure
2219 additional mounts or swaps in a format similar to /etc/fstab. 'fsck'
2220 will be ran on these block devices, like it already happens for
2221 'root='. It also now supports the new fstab.extra and
2222 fstab.extra.initrd credentials that may contain additional /etc/fstab
2223 lines to apply at boot.
2225 * systemd-getty-generator now understands two new credentials
2226 getty.ttys.container and getty.ttys.serial. These credentials may
2227 contain a list of TTY devices – one per line – to instantiate
2228 container-getty@.service and serial-getty@.service on.
2230 * The getty/serial-getty/container-getty units now import the 'agetty.*'
2231 and 'login.*' credentials, which are consumed by the 'login' and
2232 'agetty' programs starting from util-linux v2.40.
2234 * systemd-sysupdate's sysupdate.d/ drop-ins gained a new setting
2235 PathRelativeTo=, which can be set to "esp", "xbootldr", "boot", in
2236 which case the Path= setting is taken relative to the ESP or XBOOTLDR
2237 partitions, rather than the system's root directory /. The relevant
2238 directories are automatically discovered.
2240 * The systemd-ac-power tool gained a new switch --low, which reports
2241 whether the battery charge is considered "low", similar to how the
2242 s2h suspend logic checks this state to decide whether to enter system
2243 suspend or hibernation.
2245 * The /etc/os-release file can now have two new optional fields
2246 VENDOR_NAME= and VENDOR_URL= to carry information about the vendor of
2249 * When the system hibernates, information about the device and offset
2250 used is now written to a non-volatile EFI variable. On next boot the
2251 system will attempt to resume from the location indicated in this EFI
2252 variable. This should make hibernation a lot more robust, while
2253 requiring no manual configuration of the resume location.
2255 * The $XDG_STATE_HOME environment variable (added in more recent
2256 versions of the XDG basedir specification) is now honoured to
2257 implement the StateDirectory= setting in user services.
2259 * A new component "systemd-battery-check" has been added. It may run
2260 during early boot (usually in the initrd), and checks the battery
2261 charge level of the system. In case the charge level is very low the
2262 user is notified (graphically via Plymouth – if available – as well
2263 as in text form on the console), and the system is turned off after a
2264 10s delay. The feature can be disabled by passing
2265 systemd.battery_check=0 through the kernel command line.
2267 * The 'passwdqc' library is now supported as an alternative to the
2268 'pwquality' library and can be selected at build time.
2270 Contributions from: 김인수, 07416, Addison Snelling, Adrian Vovk,
2271 Aidan Dang, Alexander Krabler, Alfred Klomp, Anatoli Babenia,
2272 Andrei Stepanov, Andrew Baxter, Antonio Alvarez Feijoo,
2273 Arian van Putten, Arthur Shau, A S Alam,
2274 Asier Sarasua Garmendia, Balló György, Bastien Nocera,
2275 Benjamin Herrenschmidt, Benjamin Raison, Bill Peterson,
2276 Brad Fitzpatrick, Brett Holman, bri, Chen Qi, Chitoku,
2277 Christian Hesse, Christoph Anton Mitterer, Christopher Gurnee,
2278 Colin Walters, Cornelius Hoffmann, Cristian Rodríguez, cunshunxia,
2279 cvlc12, Cyril Roelandt, Daan De Meyer, Daniele Medri,
2280 Daniel P. Berrangé, Daniel Rusek, Dan Streetman, David Edmundson,
2281 David Schroeder, David Tardon, dependabot[bot],
2282 Dimitri John Ledkov, Dmitrii Fomchenkov, Dmitry V. Levin, dmkUK,
2283 Dominique Martinet, don bright, drosdeck, Edson Juliano Drosdeck,
2284 Egor Ignatov, EinBaum, Emanuele Giuseppe Esposito, Eric Curtin,
2285 Erik Sjölund, Evgeny Vereshchagin, Florian Klink, Franck Bui,
2286 François Rigault, Fran Diéguez, Franklin Yu, Frantisek Sumsal,
2287 Fuminobu TAKEYAMA, Gaël PORTAY, Gerd Hoffmann, Gertalitec,
2288 Gibeom Gwon, Gustavo Noronha Silva, Hannu Lounento,
2289 Hans de Goede, Haochen Tong, HATAYAMA Daisuke, Henrik Holst,
2290 Hoe Hao Cheng, Igor Tsiglyar, Ivan Vecera, James Hilliard,
2291 Jan Engelhardt, Jan Janssen, Jan Luebbe, Jan Macku, Janne Sirén,
2292 jcg, Jeidnx, Joan Bruguera, Joerg Behrmann, jonathanmetzman,
2293 Jordan Rome, Josef Miegl, Joshua Goins, Joyce, Joyce Brum,
2294 Juno Computers, Kai Lueke, Kevin P. Fleming, Kiran Vemula, Klaus,
2295 Klaus Zipfel, Lawrence Thorpe, Lennart Poettering, licunlong,
2296 Lily Foster, Luca Boccassi, Ludwig Nussel, Luna Jernberg,
2297 maanyagoenka, Maanya Goenka, Maksim Kliazovich, Malte Poll,
2298 Marko Korhonen, Masatake YAMATO, Mateusz Poliwczak, Matt Johnston,
2299 Miao Wang, Micah Abbott, Michael A Cassaniti, Michal Koutný,
2300 Michal Sekletár, Mike Yuan, mooo, Morten Linderud, msizanoen,
2301 Nick Rosbrook, nikstur, Olivier Gayot, Omojola Joshua,
2302 Paolo Velati, Paul Barker, Pavel Borecki, Petr Menšík,
2303 Philipp Kern, Philip Withnall, Piotr Drąg, Quintin Hill,
2304 Rene Hollander, Richard Phibel, Robert Meijers, Robert Scheck,
2305 Roger Gammans, Romain Geissler, Ronan Pigott, Russell Harmon,
2306 saikat0511, Samanta Navarro, Sam James, Sam Morris,
2307 Simon Braunschmidt, Sjoerd Simons, Sorah Fukumori,
2308 Stanislaw Gruszka, Stefan Roesch, Steven Luo, Steve Ramage,
2309 Susant Sahani, taniishkaaa, Tanishka, Temuri Doghonadze,
2310 Thierry Martin, Thomas Blume, Thomas Genty, Thomas Weißschuh,
2311 Thorsten Kukuk, Times-Z, Tobias Powalowski, tofylion,
2312 Topi Miettinen, Uwe Kleine-König, Velislav Ivanov,
2313 Vitaly Kuznetsov, Vít Zikmund, Weblate, Will Fancher,
2314 William Roberts, Winterhuman, Wolfgang Müller, Xeonacid,
2315 Xiaotian Wu, Xi Ruoyao, Yuri Chornoivan, Yu Watanabe, Yuxiang Zhu,
2316 Zbigniew Jędrzejewski-Szmek, zhmylove, ZjYwMj,
2317 Дамјан Георгиевски, наб
2319 — Edinburgh, 2023-07-28
2323 Announcements of Future Feature Removals and Incompatible Changes:
2325 * We intend to remove cgroup v1 support from systemd release after the
2326 end of 2023. If you run services that make explicit use of cgroup v1
2327 features (i.e. the "legacy hierarchy" with separate hierarchies for
2328 each controller), please implement compatibility with cgroup v2 (i.e.
2329 the "unified hierarchy") sooner rather than later. Most of Linux
2330 userspace has been ported over already.
2332 * We intend to remove support for split-usr (/usr mounted separately
2333 during boot) and unmerged-usr (parallel directories /bin and
2334 /usr/bin, /lib and /usr/lib, etc). This will happen in the second
2335 half of 2023, in the first release that falls into that time window.
2336 For more details, see:
2337 https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html
2339 * We intend to change behaviour w.r.t. units of the per-user service
2340 manager and sandboxing options, so that they work without having to
2341 manually enable PrivateUsers= as well, which is not required for
2342 system units. To make this work, we will implicitly enable user
2343 namespaces (PrivateUsers=yes) when a sandboxing option is enabled in a
2344 user unit. The drawback is that system users will no longer be visible
2345 (and appear as 'nobody') to the user unit when a sandboxing option is
2346 enabled. By definition a sandboxed user unit should run with reduced
2347 privileges, so impact should be small. This will remove a great source
2348 of confusion that has been reported by users over the years, due to
2349 how these options require an extra setting to be manually enabled when
2350 used in the per-user service manager, as opposed as to the system
2351 service manager. We plan to enable this change in the next release
2352 later this year. For more details, see:
2353 https://lists.freedesktop.org/archives/systemd-devel/2022-December/048682.html
2355 Deprecations and incompatible changes:
2357 * systemctl will now warn when invoked without /proc/ mounted
2358 (e.g. when invoked after chroot() into an directory tree without the
2359 API mount points like /proc/ being set up.) Operation in such an
2360 environment is not fully supported.
2362 * The return value of 'systemctl is-active|is-enabled|is-failed' for
2363 unknown units is changed: previously 1 or 3 were returned, but now 4
2364 (EXIT_PROGRAM_OR_SERVICES_STATUS_UNKNOWN) is used as documented.
2366 * 'udevadm hwdb' subcommand is deprecated and will emit a warning.
2367 systemd-hwdb (added in 2014) should be used instead.
2369 * 'bootctl --json' now outputs a single JSON array, instead of a stream
2370 of newline-separated JSON objects.
2372 * Udev rules in 60-evdev.rules have been changed to load hwdb
2373 properties for all modalias patterns. Previously only the first
2374 matching pattern was used. This could change what properties are
2375 assigned if the user has more and less specific patterns that could
2376 match the same device, but it is expected that the change will have
2377 no effect for most users.
2379 * systemd-networkd-wait-online exits successfully when all interfaces
2380 are ready or unmanaged. Previously, if neither '--any' nor
2381 '--interface=' options were used, at least one interface had to be in
2382 configured state. This change allows the case where systemd-networkd
2383 is enabled, but no interfaces are configured, to be handled
2384 gracefully. It may occur in particular when a different network
2385 manager is also enabled and used.
2387 * Some compatibility helpers were dropped: EmergencyAction= in the user
2388 manager, as well as measuring kernel command line into PCR 8 in
2389 systemd-stub, along with the -Defi-tpm-pcr-compat compile-time
2392 * The '-Dupdate-helper-user-timeout=' build-time option has been
2393 renamed to '-Dupdate-helper-user-timeout-sec=', and now takes an
2394 integer as parameter instead of a string.
2396 * The DDI image dissection logic (which backs RootImage= in service
2397 unit files, the --image= switch in various tools such as
2398 systemd-nspawn, as well as systemd-dissect) will now only mount file
2399 systems of types btrfs, ext4, xfs, erofs, squashfs, vfat. This list
2400 can be overridden via the $SYSTEMD_DISSECT_FILE_SYSTEMS environment
2401 variable. These file systems are fairly well supported and maintained
2402 in current kernels, while others are usually more niche, exotic or
2403 legacy and thus typically do not receive the same level of security
2406 * The default per-link multicast DNS mode is changed to "yes"
2407 (that was previously "no"). As the default global multicast DNS mode
2408 has been "yes" (but can be changed by the build option), now the
2409 multicast DNS is enabled on all links by default. You can disable the
2410 multicast DNS on all links by setting MulticastDNS= in resolved.conf,
2411 or on an interface by calling "resolvectl mdns INTERFACE no".
2415 * A tool 'ukify' tool to build, measure, and sign Unified Kernel Images
2416 (UKIs) has been added. This replaces functionality provided by
2417 'dracut --uefi' and extends it with automatic calculation of PE file
2418 offsets, insertion of signed PCR policies generated by
2419 systemd-measure, support for initrd concatenation, signing of the
2420 embedded Linux image and the combined image with sbsign, and
2421 heuristics to autodetect the kernel uname and verify the splash
2424 Changes in systemd and units:
2426 * A new service type Type=notify-reload is defined. When such a unit is
2427 reloaded a UNIX process signal (typically SIGHUP) is sent to the main
2428 service process. The manager will then wait until it receives a
2429 "RELOADING=1" followed by a "READY=1" notification from the unit as
2430 response (via sd_notify()). Otherwise, this type is the same as
2431 Type=notify. A new setting ReloadSignal= may be used to change the
2432 signal to send from the default of SIGHUP.
2434 user@.service, systemd-networkd.service, systemd-udevd.service, and
2435 systemd-logind have been updated to this type.
2437 * Initrd environments which are not on a pure memory file system (e.g.
2438 overlayfs combination as opposed to tmpfs) are now supported. With
2439 this change, during the initrd → host transition ("switch root")
2440 systemd will erase all files of the initrd only when the initrd is
2441 backed by a memory file system such as tmpfs.
2443 * New per-unit MemoryZSwapMax= option has been added to configure
2444 memory.zswap.max cgroup properties (the maximum amount of zswap
2447 * A new LogFilterPatterns= option has been added for units. It may be
2448 used to specify accept/deny regular expressions for log messages
2449 generated by the unit, that shall be enforced by systemd-journald.
2450 Rejected messages are neither stored in the journal nor forwarded.
2451 This option may be used to suppress noisy or uninteresting messages
2454 * The manager has a new
2455 org.freedesktop.systemd1.Manager.GetUnitByPIDFD() D-Bus method to
2456 query process ownership via a PIDFD, which is more resilient against
2457 PID recycling issues.
2459 * Scope units now support OOMPolicy=. Login session scopes default to
2460 OOMPolicy=continue, allowing login scopes to survive the OOM killer
2461 terminating some processes in the scope.
2463 * systemd-fstab-generator now supports x-systemd.makefs option for
2464 /sysroot/ (in the initrd).
2466 * The maximum rate at which daemon reloads are executed can now be
2467 limited with the new ReloadLimitIntervalSec=/ReloadLimitBurst=
2468 options. (Or the equivalent on the kernel command line:
2469 systemd.reload_limit_interval_sec=/systemd.reload_limit_burst=). In
2470 addition, systemd now logs the originating unit and PID when a reload
2471 request is received over D-Bus.
2473 * When enabling a swap device systemd will now reinitialize the device
2474 when the page size of the swap space does not match the page size of
2475 the running kernel. Note that this requires the 'swapon' utility to
2476 provide the '--fixpgsz' option, as implemented by util-linux, and it
2477 is not supported by busybox at the time of writing.
2479 * systemd now executes generator programs in a mount namespace
2480 "sandbox" with most of the file system read-only and write access
2481 restricted to the output directories, and with a temporary /tmp/
2482 mount provided. This provides a safeguard against programming errors
2483 in the generators, but also fixes here-docs in shells, which
2484 previously didn't work in early boot when /tmp/ wasn't available
2485 yet. (This feature has no security implications, because the code is
2486 still privileged and can trivially exit the sandbox.)
2488 * The system manager will now parse a new "vmm.notify_socket"
2489 system credential, which may be supplied to a VM via SMBIOS. If
2490 found, the manager will send a "READY=1" notification on the
2491 specified socket after boot is complete. This allows readiness
2492 notification to be sent from a VM guest to the VM host over a VSOCK
2495 * The sample PAM configuration file for systemd-user@.service now
2496 includes a call to pam_namespace. This puts children of user@.service
2497 in the expected namespace. (Many distributions replace their file
2498 with something custom, so this change has limited effect.)
2500 * A new environment variable $SYSTEMD_DEFAULT_MOUNT_RATE_LIMIT_BURST
2501 can be used to override the mount units burst late limit for
2502 parsing '/proc/self/mountinfo', which was introduced in v249.
2505 * Drop-ins for init.scope changing control group resource limits are
2506 now applied, while they were previously ignored.
2508 * New build-time configuration options '-Ddefault-timeout-sec=' and
2509 '-Ddefault-user-timeout-sec=' have been added, to let distributions
2510 choose the default timeout for starting/stopping/aborting system and
2511 user units respectively.
2513 * Service units gained a new setting OpenFile= which may be used to
2514 open arbitrary files in the file system (or connect to arbitrary
2515 AF_UNIX sockets in the file system), and pass the open file
2516 descriptor to the invoked process via the usual file descriptor
2517 passing protocol. This is useful to give unprivileged services access
2518 to select files which have restrictive access modes that would
2519 normally not allow this. It's also useful in case RootDirectory= or
2520 RootImage= is used to allow access to files from the host environment
2521 (which is after all not visible from the service if these two options
2526 * The new net naming scheme "v253" has been introduced. In the new
2527 scheme, ID_NET_NAME_PATH is also set for USB devices not connected via
2528 a PCI bus. This extends the coverage of predictable interface names
2529 in some embedded systems.
2531 The "amba" bus path is now included in ID_NET_NAME_PATH, resulting in
2532 a more informative path on some embedded systems.
2534 * Partition block devices will now also get symlinks in
2535 /dev/disk/by-diskseq/<seq>-part<n>, which may be used to reference
2536 block device nodes via the kernel's "diskseq" value. Previously those
2537 symlinks were only created for the main block device.
2539 * A new operator '-=' is supported for SYMLINK variables. This allows
2540 symlinks to be unconfigured even if an earlier rule added them.
2542 * 'udevadm --trigger --settle' now also works for network devices
2543 that are being renamed.
2545 Changes in sd-boot, bootctl, and the Boot Loader Specification:
2547 * systemd-boot now passes its random seed directly to the kernel's RNG
2548 via the LINUX_EFI_RANDOM_SEED_TABLE_GUID configuration table, which
2549 means the RNG gets seeded very early in boot before userspace has
2552 * systemd-boot will pass a disk-backed random seed – even when secure
2553 boot is enabled – if it can additionally get a random seed from EFI
2554 itself (via EFI's RNG protocol), or a prior seed in
2555 LINUX_EFI_RANDOM_SEED_TABLE_GUID from a preceding bootloader.
2557 * systemd-boot-system-token.service was renamed to
2558 systemd-boot-random-seed.service and extended to always save a random
2559 seed to ESP on every boot when a compatible boot loader is used. This
2560 allows a refreshed random seed to be used in the boot loader.
2562 * systemd-boot handles various seed inputs using a domain- and
2563 field-separated hashing scheme.
2565 * systemd-boot's 'random-seed-mode' option has been removed. A system
2566 token is now always required to be present for random seeds to be
2569 * systemd-boot now supports being loaded from other locations than the
2570 ESP, for example for direct kernel boot under QEMU or when embedded
2573 * systemd-boot now parses SMBIOS information to detect
2574 virtualization. This information is used to skip some warnings which
2575 are not useful in a VM and to conditionalize other aspects of
2578 * systemd-boot now supports a new 'if-safe' mode that will perform UEFI
2579 Secure Boot automated certificate enrollment from the ESP only if it
2580 is considered 'safe' to do so. At the moment 'safe' means running in
2583 * systemd-stub now processes random seeds in the same way as
2584 systemd-boot already does, in case a unified kernel image is being
2585 used from a different bootloader than systemd-boot, or without any
2588 * bootctl will now generate a system token on all EFI systems, even
2589 virtualized ones, and is activated in the case that the system token
2590 is missing from either sd-boot and sd-stub booted systems.
2592 * bootctl now implements two new verbs: 'kernel-identify' prints the
2593 type of a kernel image file, and 'kernel-inspect' provides
2594 information about the embedded command line and kernel version of
2597 * bootctl now honours $KERNEL_INSTALL_CONF_ROOT with the same meaning
2598 as for kernel-install.
2600 * The JSON output of "bootctl list" will now contain two more fields:
2601 isDefault and isSelected are boolean fields set to true on the
2602 default and currently booted boot menu entries.
2604 * bootctl gained a new verb "unlink" for removing a boot loader entry
2605 type #1 file from disk in a safe and robust way.
2607 * bootctl also gained a new verb "cleanup" that automatically removes
2608 all files from the ESP's and XBOOTLDR's "entry-token" directory, that
2609 is not referenced anymore by any installed Type #1 boot loader
2610 specification entry. This is particularly useful in environments where
2611 a large number of entries reference the same or partly the same
2612 resources (for example, for snapshot-based setups).
2614 Changes in kernel-install:
2616 * A new "installation layout" can be configured as layout=uki. With
2617 this setting, a Boot Loader Specification Type#1 entry will not be
2618 created. Instead, a new kernel-install plugin 90-uki-copy.install
2619 will copy any .efi files from the staging area into the boot
2620 partition. A plugin to generate the UKI .efi file must be provided
2623 Changes in systemctl:
2625 * 'systemctl reboot' has dropped support for accepting a positional
2626 argument as the argument to the reboot(2) syscall. Please use the
2627 --reboot-argument= option instead.
2629 * 'systemctl disable' will now warn when called on units without
2630 install information. A new --no-warn option has been added that
2631 silences this warning.
2633 * New option '--drop-in=' can be used to tell 'systemctl edit' the name
2634 of the drop-in to edit. (Previously, 'override.conf' was always
2637 * 'systemctl list-dependencies' now respects --type= and --state=.
2639 * 'systemctl kexec' now supports XEN VMM environments.
2641 * 'systemctl edit' will now tell the invoked editor to jump into the
2642 first line with actual unit file data, skipping over synthesized
2645 Changes in systemd-networkd and related tools:
2647 * The [DHCPv4] section in .network file gained new SocketPriority=
2648 setting that assigns the Linux socket priority used by the DHCPv4 raw
2649 socket. This may be used in conjunction with the
2650 EgressQOSMaps=setting in [VLAN] section of .netdev file to send the
2651 desired ethernet 802.1Q frame priority for DHCPv4 initial
2652 packets. This cannot be achieved with netfilter mangle tables because
2653 of the raw socket bypass.
2655 * The [DHCPv4] and [IPv6AcceptRA] sections in .network file gained a
2656 new QuickAck= boolean setting that enables the TCP quick ACK mode for
2657 the routes configured by the acquired DHCPv4 lease or received router
2658 advertisements (RAs).
2660 * The RouteMetric= option (for DHCPv4, DHCPv6, and IPv6 advertised
2661 routes) now accepts three values, for high, medium, and low preference
2662 of the router (which can be set with the RouterPreference=) setting.
2664 * systemd-networkd-wait-online now supports matching via alternative
2667 * The [DHCPv6] section in .network file gained new SendRelease=
2668 setting which enables the DHCPv6 client to send release when
2669 it stops. This is the analog of the [DHCPv4] SendRelease= setting.
2670 It is enabled by default.
2672 * If the Address= setting in [Network] or [Address] sections in .network
2673 specified without its prefix length, then now systemd-networkd assumes
2674 /32 for IPv4 or /128 for IPv6 addresses.
2676 * networkctl shows network and link file dropins in status output.
2678 Changes in systemd-dissect:
2680 * systemd-dissect gained a new option --list, to print the paths of
2681 all files and directories in a DDI.
2683 * systemd-dissect gained a new option --mtree, to generate a file
2684 manifest compatible with BSD mtree(5) of a DDI
2686 * systemd-dissect gained a new option --with, to execute a command with
2687 the specified DDI temporarily mounted and used as working
2688 directory. This is for example useful to convert a DDI to "tar"
2689 simply by running it within a "systemd-dissect --with" invocation.
2691 * systemd-dissect gained a new option --discover, to search for
2692 Discoverable Disk Images (DDIs) in well-known directories of the
2693 system. This will list machine, portable service and system extension
2696 * systemd-dissect now understands 2nd stage initrd images stored as a
2697 Discoverable Disk Image (DDI).
2699 * systemd-dissect will now display the main UUID of GPT DDIs (i.e. the
2700 disk UUID stored in the GPT header) among the other data it can show.
2702 * systemd-dissect gained a new --in-memory switch to operate on an
2703 in-memory copy of the specified DDI file. This is useful to access a
2704 DDI with write access without persisting any changes. It's also
2705 useful for accessing a DDI without keeping the originating file
2708 * The DDI dissection logic will now automatically detect the intended
2709 sector size of disk images stored in files, based on the GPT
2710 partition table arrangement. Loopback block devices for such DDIs
2711 will then be configured automatically for the right sector size. This
2712 is useful to make dealing with modern 4K sector size DDIs fully
2713 automatic. The systemd-dissect tool will now show the detected sector
2714 size among the other DDI information in its output.
2716 Changes in systemd-repart:
2718 * systemd-repart gained new options --include-partitions= and
2719 --exclude-partitions= to filter operation on partitions by type UUID.
2720 This allows systemd-repart to be used to build images in which the
2721 type of one partition is set based on the contents of another
2722 partition (for example when the boot partition shall include a verity
2723 hash of the root partition).
2725 * systemd-repart also gained a --defer-partitions= option that is
2726 similar to --exclude-partitions=, but the size of the partition is
2727 still taken into account when sizing partitions, but without
2730 * systemd-repart gained a new --sector-size= option to specify what
2731 sector size should be used when an image is created.
2733 * systemd-repart now supports generating erofs file systems via
2734 CopyFiles= (a read-only file system similar to squashfs).
2736 * The Minimize= option was extended to accept "best" (which means the
2737 most minimal image possible, but may require multiple attempts) and
2738 "guess" (which means a reasonably small image).
2740 * The systemd-growfs binary now comes with a regular unit file template
2741 systemd-growfs@.service which can be instantiated directly for any
2742 desired file system. (Previously, the unit was generated dynamically
2743 by various generators, but no regular unit file template was
2746 Changes in journal tools:
2748 * Various systemd tools will append extra fields to log messages when
2749 in debug mode, or when SYSTEMD_ENABLE_LOG_CONTEXT=1 is set. Currently
2750 this includes information about D-Bus messages when sd-bus is used,
2751 e.g. DBUS_SENDER=, DBUS_DESTINATION=, and DBUS_PATH=, and information
2752 about devices when sd-device is used, e.g. DEVNAME= and DRIVER=.
2753 Details of what is logged and when are subject to change.
2755 * The systemd-journald-audit.socket can now be disabled via the usual
2756 "systemctl disable" mechanism to stop collection of audit
2757 messages. Please note that it is not enabled statically anymore and
2758 must be handled by the preset/enablement logic in package
2759 installation scripts.
2761 * New options MaxUse=, KeepFree=, MaxFileSize=, and MaxFiles= can
2762 be used to curtail disk use by systemd-journal-remote. This is
2763 similar to the options supported by systemd-journald.
2765 Changes in systemd-cryptenroll, systemd-cryptsetup, and related
2768 * When enrolling new keys systemd-cryptenroll now supports unlocking
2769 via FIDO2 tokens (option --unlock-fido2-device=). Previously, a
2770 password was strictly required to be specified.
2772 * systemd-cryptsetup now supports pre-flight requests for FIDO2 tokens
2773 (except for tokens with user verification, UV) to identify tokens
2774 before authentication. Multiple FIDO2 tokens can now be enrolled at
2775 the same time, and systemd-cryptsetup will automatically select one
2776 that corresponds to one of the available LUKS key slots.
2778 * systemd-cryptsetup now supports new options tpm2-measure-bank= and
2779 tpm2-measure-pcr= in crypttab(5). These allow specifying the TPM2 PCR
2780 bank and number into which the volume key should be measured. This is
2781 automatically enabled for the encrypted root volume discovered and
2782 activated by systemd-gpt-auto-generator.
2784 * systemd-gpt-auto-generator mounts the ESP and XBOOTLDR partitions with
2785 "noexec,nosuid,nodev".
2787 * systemd-gpt-auto-generator will now honour the rootfstype= and
2788 rootflags= kernel command line switches for root file systems it
2789 discovers, to match behaviour in case an explicit root fs is
2790 specified via root=.
2792 * systemd-pcrphase gained new options --machine-id and --file-system=
2793 to measure the machine-id and mount point information into PCR 15.
2794 New service unit files systemd-pcrmachine.service and
2795 systemd-pcrfs@.service have been added that invoke the tool with
2796 these switches during early boot.
2798 * systemd-pcrphase gained a --graceful switch will make it exit cleanly
2799 with a success exit code even if no TPM device is detected.
2801 * systemd-cryptenroll now stores the user-supplied PIN with a salt,
2802 making it harder to brute-force.
2804 Changes in other tools:
2806 * systemd-homed gained support for luksPbkdfForceIterations (the
2807 intended number of iterations for the PBKDF operation on LUKS).
2809 * Environment variables $SYSTEMD_HOME_MKFS_OPTIONS_BTRFS,
2810 $SYSTEMD_HOME_MKFS_OPTIONS_EXT4, and $SYSTEMD_HOME_MKFS_OPTIONS_XFS
2811 may now be used to specify additional arguments for mkfs when
2812 systemd-homed formats a file system.
2814 * systemd-hostnamed now exports the contents of
2815 /sys/class/dmi/id/bios_vendor and /sys/class/dmi/id/bios_date via two
2816 new D-Bus properties: FirmwareVendor and FirmwareDate. This allows
2817 unprivileged code to access those values.
2819 systemd-hostnamed also exports the SUPPORT_END= field from
2820 os-release(5) as OperatingSystemSupportEnd. hostnamectl make uses of
2821 this to show the status of the installed system.
2823 * systemd-measure gained an --append= option to sign multiple phase
2824 paths with different signing keys. This allows secrets to be
2825 accessible only in certain parts of the boot sequence. Note that
2826 'ukify' provides similar functionality in a more accessible form.
2828 * systemd-timesyncd will now write a structured log message with
2829 MESSAGE_ID set to SD_MESSAGE_TIME_BUMP when it bumps the clock based
2830 on a on-disk timestamp, similarly to what it did when reaching
2831 synchronization via NTP.
2833 * systemd-timesyncd will now update the on-disk timestamp file on each
2834 boot at least once, making it more likely that the system time
2835 increases in subsequent boots.
2837 * systemd-vconsole-setup gained support for system/service credentials:
2838 vconsole.keymap/vconsole.keymap_toggle and
2839 vconsole.font/vconsole.font_map/vconsole.font_unimap are analogous
2840 the similarly-named options in vconsole.conf.
2842 * systemd-localed will now save the XKB keyboard configuration to
2843 /etc/vconsole.conf, and also read it from there with a higher
2844 preference than the /etc/X11/xorg.conf.d/00-keyboard.conf config
2845 file. Previously, this information was stored in the former file in
2846 converted form, and only in latter file in the original form. Tools
2847 which want to access keyboard configuration can now do so from a
2850 * systemd-resolved gained support for configuring the nameservers and
2851 search domains via kernel command line (nameserver=, domain=) and
2852 credentials (network.dns, network.search_domains).
2854 * systemd-resolved will now synthesize host names for the DNS stub
2855 addresses it supports. Specifically when "_localdnsstub" is resolved,
2856 127.0.0.53 is returned, and if "_localdnsproxy" is resolved
2857 127.0.0.54 is returned.
2859 * systemd-notify will now send a "RELOADING=1" notification when called
2860 with --reloading, and "STOPPING=1" when called with --stopping. This
2861 can be used to implement notifications from units where it's easier
2862 to call a program than to use the sd-daemon library.
2864 * systemd-analyze's 'plot' command can now output its information in
2865 JSON, controlled via the --json= switch. Also, new --table, and
2866 --no-legend options have been added.
2868 * 'machinectl enable' will now automatically enable machines.target
2869 unit in addition to adding the machine unit to the target.
2871 Similarly, 'machinectl start|stop' gained a --now option to enable or
2872 disable the machine unit when starting or stopping it.
2874 * systemd-sysusers will now create /etc/ if it is missing.
2876 * systemd-sleep 'HibernateDelaySec=' setting is changed back to
2877 pre-v252's behaviour, and a new 'SuspendEstimationSec=' setting is
2878 added to provide the new initial value for the new automated battery
2879 estimation functionality. If 'HibernateDelaySec=' is set to any value,
2880 the automated estimate (and thus the automated hibernation on low
2881 battery to avoid data loss) functionality will be disabled.
2883 * Default tmpfiles.d/ configuration will now automatically create
2884 credentials storage directory '/etc/credstore/' with the appropriate,
2885 secure permissions. If '/run/credstore/' exists, its permissions will
2886 be fixed too in case they are not correct.
2888 Changes in libsystemd and shared code:
2890 * sd-bus gained new convenience functions sd_bus_emit_signal_to(),
2891 sd_bus_emit_signal_tov(), and sd_bus_message_new_signal_to().
2893 * sd-id128 functions now return -EUCLEAN (instead of -EIO) when the
2894 128-bit ID in files such as /etc/machine-id has an invalid
2895 format. They also accept NULL as output parameter in more places,
2896 which is useful when the caller only wants to validate the inputs and
2897 does not need the output value.
2899 * sd-login gained new functions sd_pidfd_get_session(),
2900 sd_pidfd_get_owner_uid(), sd_pidfd_get_unit(),
2901 sd_pidfd_get_user_unit(), sd_pidfd_get_slice(),
2902 sd_pidfd_get_user_slice(), sd_pidfd_get_machine_name(), and
2903 sd_pidfd_get_cgroup(), that are analogous to sd_pid_get_*(),
2904 but accept a PIDFD instead of a PID.
2906 * sd-path (and systemd-path) now export four new paths:
2907 SD_PATH_SYSTEMD_SYSTEM_ENVIRONMENT_GENERATOR,
2908 SD_PATH_SYSTEMD_USER_ENVIRONMENT_GENERATOR,
2909 SD_PATH_SYSTEMD_SEARCH_SYSTEM_ENVIRONMENT_GENERATOR, and
2910 SD_PATH_SYSTEMD_SEARCH_USER_ENVIRONMENT_GENERATOR,
2912 * sd_notify() now supports AF_VSOCK as transport for notification
2913 messages (in addition to the existing AF_UNIX support). This is
2914 enabled if $NOTIFY_SOCKET is set in a "vsock:CID:port" format.
2916 * Detection of chroot() environments now works if /proc/ is not
2917 mounted. This affects systemd-detect-virt --chroot, but also means
2918 that systemd tools will silently skip various operations in such an
2921 * "Lockheed Martin Hardened Security for Intel Processors" (HS SRE)
2922 virtualization is now detected.
2924 Changes in the build system:
2926 * Standalone variants of systemd-repart and systemd-shutdown may now be
2927 built (if -Dstandalone=true).
2929 * systemd-ac-power has been moved from /usr/lib/ to /usr/bin/, to, for
2930 example, allow scripts to conditionalize execution on AC power
2933 * The libp11kit library is now loaded through dlopen(3).
2935 Changes in the documentation:
2937 * Specifications that are not closely tied to systemd have moved to
2938 https://uapi-group.org/specifications/: the Boot Loader Specification
2939 and the Discoverable Partitions Specification.
2941 Contributions from: 김인수, 13r0ck, Aidan Dang, Alberto Planas,
2942 Alvin Šipraga, Andika Triwidada, AndyChi, angus-p, Anita Zhang,
2943 Antonio Alvarez Feijoo, Arsen Arsenović, asavah, Benjamin Fogle,
2944 Benjamin Tissoires, berenddeschouwer, BerndAdameit,
2945 Bernd Steinhauser, blutch112, cake03, Callum Farmer, Carlo Teubner,
2946 Charles Hardin, chris, Christian Brauner, Christian Göttsche,
2947 Cristian Rodríguez, Daan De Meyer, Dan Streetman, DaPigGuy,
2948 Darrell Kavanagh, David Tardon, dependabot[bot], Dirk Su,
2949 Dmitry V. Levin, drosdeck, Edson Juliano Drosdeck, edupont,
2950 Eric DeVolder, Erik Moqvist, Evgeny Vereshchagin, Fabian Gurtner,
2951 Felix Riemann, Franck Bui, Frantisek Sumsal, Geert Lorang,
2952 Gerd Hoffmann, Gio, Hannoskaj, Hans de Goede, Hugo Carvalho,
2953 igo95862, Ilya Leoshkevich, Ivan Shapovalov, Jacek Migacz,
2954 Jade Lovelace, Jan Engelhardt, Jan Janssen, Jan Macku, January,
2955 Jason A. Donenfeld, jcg, Jean-Tiare Le Bigot, Jelle van der Waa,
2956 Jeremy Linton, Jian Zhang, Jiayi Chen, Jia Zhang, Joerg Behrmann,
2957 Jörg Thalheim, Joshua Goins, joshuazivkovic, Joshua Zivkovic,
2958 Kai-Chuan Hsieh, Khem Raj, Koba Ko, Lennart Poettering, lichao,
2959 Li kunyu, Luca Boccassi, Luca BRUNO, Ludwig Nussel,
2960 Łukasz Stelmach, Lycowolf, marcel151, Marcus Schäfer, Marek Vasut,
2961 Mark Laws, Michael Biebl, Michał Kotyla, Michal Koutný,
2962 Michal Sekletár, Mike Gilbert, Mike Yuan, MkfsSion, ml,
2963 msizanoen1, mvzlb, MVZ Ludwigsburg, Neil Moore, Nick Rosbrook,
2964 noodlejetski, Pasha Vorobyev, Peter Cai, p-fpv, Phaedrus Leeds,
2965 Philipp Jungkamp, Quentin Deslandes, Raul Tambre, Ray Strode,
2966 reuben olinsky, Richard E. van der Luit, Richard Phibel,
2967 Ricky Tigg, Robin Humble, rogg, Rudi Heitbaum, Sam James,
2968 Samuel Cabrero, Samuel Thibault, Siddhesh Poyarekar, Simon Brand,
2969 Space Meyer, Spindle Security, Steve Ramage, Takashi Sakamoto,
2970 Thomas Haller, Tonći Galić, Topi Miettinen, Torsten Hilbrich,
2971 Tuetuopay, uerdogan, Ulrich Ölmann, Valentin David,
2972 Vitaly Kuznetsov, Vito Caputo, Waltibaba, Will Fancher,
2973 William Roberts, wouter bolsterlee, Youfu Zhang, Yu Watanabe,
2974 Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски,
2977 — Warsaw, 2023-02-15
2981 Announcements of Future Feature Removals:
2983 * We intend to remove cgroup v1 support from systemd release after the
2984 end of 2023. If you run services that make explicit use of cgroup v1
2985 features (i.e. the "legacy hierarchy" with separate hierarchies for
2986 each controller), please implement compatibility with cgroup v2 (i.e.
2987 the "unified hierarchy") sooner rather than later. Most of Linux
2988 userspace has been ported over already.
2990 * We intend to remove support for split-usr (/usr mounted separately
2991 during boot) and unmerged-usr (parallel directories /bin and
2992 /usr/bin, /lib and /usr/lib, etc). This will happen in the second
2993 half of 2023, in the first release that falls into that time window.
2994 For more details, see:
2995 https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html
2997 Compatibility Breaks:
2999 * ConditionKernelVersion= checks that use the '=' or '!=' operators
3000 will now do simple string comparisons (instead of version comparisons
3001 à la stverscmp()). Version comparisons are still done for the
3002 ordering operators '<', '>', '<=', '>='. Moreover, if no operator is
3003 specified, a shell-style glob match is now done. This creates a minor
3004 incompatibility compared to older systemd versions when the '*', '?',
3005 '[', ']' characters are used, as these will now match as shell globs
3006 instead of literally. Given that kernel version strings typically do
3007 not include these characters we expect little breakage through this
3010 * The service manager will now read the SELinux label used for SELinux
3011 access checks from the unit file at the time it loads the file.
3012 Previously, the label would be read at the moment of the access
3013 check, which was problematic since at that time the unit file might
3014 already have been updated or removed.
3018 * systemd-measure is a new tool for calculating and signing expected
3019 TPM2 PCR values for a given unified kernel image (UKI) booted via
3020 sd-stub. The public key used for the signature and the signed
3021 expected PCR information can be embedded inside the UKI. This
3022 information can be extracted from the UKI by external tools and code
3023 in the image itself and is made available to userspace in the booted
3026 systemd-cryptsetup, systemd-cryptenroll, and systemd-creds have been
3027 updated to make use of this information if available in the booted
3028 kernel: when locking an encrypted volume/credential to the TPM
3029 systemd-cryptenroll/systemd-creds will use the public key to bind the
3030 volume/credential to any kernel that carries PCR information signed
3031 by the same key pair. When unlocking such volumes/credentials
3032 systemd-cryptsetup/systemd-creds will use the signature embedded in
3033 the booted UKI to gain access.
3035 Binding TPM-based disk encryption to public keys/signatures of PCR
3036 values — instead of literal PCR values — addresses the inherent
3037 "brittleness" of traditional PCR-bound TPM disk encryption schemes:
3038 disks remain accessible even if the UKI is updated, without any TPM
3039 specific preparation during the OS update — as long as each UKI
3040 carries the necessary PCR signature information.
3042 Net effect: if you boot a properly prepared kernel, TPM-bound disk
3043 encryption now defaults to be locked to kernels which carry PCR
3044 signatures from the same key pair. Example: if a hypothetical distro
3045 FooOS prepares its UKIs like this, TPM-based disk encryption is now –
3046 by default – bound to only FooOS kernels, and encrypted volumes bound
3047 to the TPM cannot be unlocked on kernels from other sources. (But do
3048 note this behaviour requires preparation/enabling in the UKI, and of
3049 course users can always enroll non-TPM ways to unlock the volume.)
3051 * systemd-pcrphase is a new tool that is invoked at six places during
3052 system runtime, and measures additional words into TPM2 PCR 11, to
3053 mark milestones of the boot process. This allows binding access to
3054 specific TPM2-encrypted secrets to specific phases of the boot
3055 process. (Example: LUKS2 disk encryption key only accessible in the
3056 initrd, but not later.)
3058 Changes in systemd itself, i.e. the manager and units
3060 * The cpu controller is delegated to user manager units by default, and
3061 CPUWeight= settings are applied to the top-level user slice units
3062 (app.slice, background.slice, session.slice). This provides a degree
3063 of resource isolation between different user services competing for
3066 * Systemd can optionally do a full preset in the "first boot" condition
3067 (instead of just enable-only). This behaviour is controlled by the
3068 compile-time option -Dfirst-boot-full-preset. Right now it defaults
3069 to 'false', but the plan is to switch it to 'true' for the subsequent
3072 * Drop-ins are now allowed for transient units too.
3074 * Systemd will set the taint flag 'support-ended' if it detects that
3075 the OS image is past its end-of-support date. This date is declared
3076 in a new /etc/os-release field SUPPORT_END= described below.
3078 * Two new settings ConditionCredential= and AssertCredential= can be
3079 used to skip or fail units if a certain system credential is not
3082 * ConditionMemory= accepts size suffixes (K, M, G, T, …).
3084 * DefaultSmackProcessLabel= can be used in system.conf and user.conf to
3085 specify the SMACK security label to use when not specified in a unit
3088 * DefaultDeviceTimeoutSec= can be used in system.conf and user.conf to
3089 specify the default timeout when waiting for device units to
3092 * C.UTF-8 is used as the default locale if nothing else has been
3095 * [Condition|Assert]Firmware= have been extended to support certain
3096 SMBIOS fields. For example
3098 ConditionFirmware=smbios-field(board_name = "Custom Board")
3100 conditionalizes the unit to run only when
3101 /sys/class/dmi/id/board_name contains "Custom Board" (without the
3104 * ConditionFirstBoot= now correctly evaluates as true only during the
3105 boot phase of the first boot. A unit executed later, after booting
3106 has completed, will no longer evaluate this condition as true.
3108 * Socket units will now create sockets in the SELinuxContext= of the
3109 associated service unit, if any.
3111 * Boot phase transitions (start initrd → exit initrd → boot complete →
3112 shutdown) will be measured into TPM2 PCR 11, so that secrets can be
3113 bound to a specific runtime phase. E.g.: a LUKS encryption key can be
3114 unsealed only in the initrd.
3116 * Service credentials (i.e. SetCredential=/LoadCredential=/…) will now
3117 also be provided to ExecStartPre= processes.
3119 * Various units are now correctly ordered against
3120 initrd-switch-root.target where previously a conflict without
3121 ordering was configured. A stop job for those units would be queued,
3122 but without the ordering it could be executed only after
3123 initrd-switch-root.service, leading to units not being restarted in
3124 the host system as expected.
3126 * In order to fully support the IPMI watchdog driver, which has not yet
3127 been ported to the new common watchdog device interface,
3128 /dev/watchdog0 will be tried first and systemd will silently fallback
3129 to /dev/watchdog if it is not found.
3131 * New watchdog-related D-Bus properties are now published by systemd:
3132 WatchdogDevice, WatchdogLastPingTimestamp,
3133 WatchdogLastPingTimestampMonotonic.
3135 * At shutdown, API virtual files systems (proc, sys, etc.) will be
3138 * At shutdown, systemd will now log about processes blocking unmounting
3141 * A new meson build option 'clock-valid-range-usec-max' was added to
3142 allow disabling system time correction if RTC returns a timestamp far
3145 * Propagated restart jobs will no longer be discarded while a unit is
3148 * PID 1 will now import system credentials from SMBIOS Type 11 fields
3149 ("OEM vendor strings"), in addition to qemu_fwcfg. This provides a
3150 simple, fast and generic path for supplying credentials to a VM,
3151 without involving external tools such as cloud-init/ignition.
3153 * The CPUWeight= setting of unit files now accepts a new special value
3154 "idle", which configures "idle" level scheduling for the unit.
3156 * Service processes that are activated due to a .timer or .path unit
3157 triggering will now receive information about this via environment
3158 variables. Note that this is information is lossy, as activation
3159 might be coalesced and only one of the activating triggers will be
3160 reported. This is hence more suited for debugging or tracing rather
3161 than for behaviour decisions.
3163 * The riscv_flush_icache(2) system call has been added to the list of
3164 system calls allowed by default when SystemCallFilter= is used.
3166 * The selinux context derived from the target executable, instead of
3167 'init_t' used for the manager itself, is now used when creating
3168 listening sockets for units that specify SELinuxContextFromNet=yes.
3170 Changes in sd-boot, bootctl, and the Boot Loader Specification:
3172 * The Boot Loader Specification has been cleaned up and clarified.
3173 Various corner cases in version string comparisons have been fixed
3174 (e.g. comparisons for empty strings). Boot counting is now part of
3175 the main specification.
3177 * New PCRs measurements are performed during boot: PCR 11 for the
3178 kernel+initrd combo, PCR 13 for any sysext images. If a measurement
3179 took place this is now reported to userspace via the new
3180 StubPcrKernelImage and StubPcrInitRDSysExts EFI variables.
3182 * As before, systemd-stub will measure kernel parameters and system
3183 credentials into PCR 12. It will now report this fact via the
3184 StubPcrKernelParameters EFI variable to userspace.
3186 * The UEFI monotonic boot counter is now included in the updated random
3187 seed file maintained by sd-boot, providing some additional entropy.
3189 * sd-stub will use LoadImage/StartImage to execute the kernel, instead
3190 of arranging the image manually and jumping to the kernel entry
3191 point. sd-stub also installs a temporary UEFI SecurityOverride to
3192 allow the (unsigned) nested image to be booted. This is safe because
3193 the outer (signed) stub+kernel binary must have been verified before
3194 the stub was executed.
3196 * Booting in EFI mixed mode (a 64-bit kernel over 32-bit UEFI firmware)
3197 is now supported by sd-boot.
3199 * bootctl gained a bunch of new options: --all-architectures to install
3200 binaries for all supported EFI architectures, --root= and --image=
3201 options to operate on a directory or disk image, and
3202 --install-source= to specify the source for binaries to install,
3203 --efi-boot-option-description= to control the name of the boot entry.
3205 * The sd-boot stub exports a StubFeatures flag, which is used by
3206 bootctl to show features supported by the stub that was used to boot.
3208 * The PE section offsets that are used by tools that assemble unified
3209 kernel images have historically been hard-coded. This may lead to
3210 overlapping PE sections which may break on boot. The UKI will now try
3211 to detect and warn about this.
3213 Any tools that assemble UKIs must update to calculate these offsets
3214 dynamically. Future sd-stub versions may use offsets that will not
3215 work with the currently used set of hard-coded offsets!
3217 * sd-stub now accepts (and passes to the initrd and then to the full
3218 OS) new PE sections '.pcrsig' and '.pcrkey' that can be used to embed
3219 signatures of expected PCR values, to allow sealing secrets via the
3220 TPM2 against pre-calculated PCR measurements.
3222 Changes in the hardware database:
3224 * 'systemd-hwdb query' now supports the --root= option.
3226 Changes in systemctl:
3228 * systemctl now supports --state= and --type= options for the 'show'
3231 * systemctl gained a new verb 'list-automounts' to list automount
3234 * systemctl gained support for a new --image= switch to be able to
3235 operate on the specified disk image (similar to the existing --root=
3236 which operates relative to some directory).
3238 Changes in systemd-networkd:
3240 * networkd can set Linux NetLabel labels for integration with the
3241 network control in security modules via a new NetLabel= option.
3243 * The RapidCommit= is (re-)introduced to enable faster configuration
3244 via DHCPv6 (RFC 3315).
3246 * networkd gained a new option TCPCongestionControlAlgorithm= that
3247 allows setting a per-route TCP algorithm.
3249 * networkd gained a new option KeepFileDescriptor= to allow keeping a
3250 reference (file descriptor) open on TUN/TAP interfaces, which is
3251 useful to avoid link flaps while the underlying service providing the
3252 interface is being serviced.
3254 * RouteTable= now also accepts route table names.
3256 Changes in systemd-nspawn:
3258 * The --bind= and --overlay= options now support relative paths.
3260 * The --bind= option now supports a 'rootidmap' value, which will
3261 use id-mapped mounts to map the root user inside the container to the
3262 owner of the mounted directory on the host.
3264 Changes in systemd-resolved:
3266 * systemd-resolved now persists DNSOverTLS in its state file too. This
3267 fixes a problem when used in combination with NetworkManager, which
3268 sends the setting only once, causing it to be lost if resolved was
3269 restarted at any point.
3271 * systemd-resolved now exposes a Varlink socket at
3272 /run/systemd/resolve/io.systemd.Resolve.Monitor, accessible only for
3273 root. Processed DNS requests in a JSON format will be published to
3274 any clients connected to this socket.
3276 resolvectl gained a 'monitor' verb to make use of this.
3278 * systemd-resolved now treats unsupported DNSSEC algorithms as INSECURE
3279 instead of returning SERVFAIL, as per RFC:
3280 https://datatracker.ietf.org/doc/html/rfc6840#section-5.2
3282 * OpenSSL is the default crypto backend for systemd-resolved. (gnutls
3283 is still supported.)
3285 Changes in libsystemd and other libraries:
3287 * libsystemd now exports sd_bus_error_setfv() (a convenience function
3288 for setting bus errors), sd_id128_string_equal (a convenience
3289 function for 128-bit ID string comparisons), and
3290 sd_bus_message_read_strv_extend() (a function to incrementally read
3293 * libsystemd now exports sd_device_get_child_first()/_next() as a
3294 high-level interface for enumerating child devices. It also supports
3295 sd_device_new_child() for opening a child device given a device
3298 * libsystemd now exports sd_device_monitor_set()/get_description()
3299 which allow setting a custom description that will be used in log
3300 messages by sd_device_monitor*.
3302 * Private shared libraries (libsystemd-shared-nnn.so,
3303 libsystemd-core-nnn.so) are now installed into arch-specific
3304 directories to allow multi-arch installs.
3306 * A new sd-gpt.h header is now published, listing GUIDs from the
3307 Discoverable Partitions specification. For more details see:
3308 https://systemd.io/DISCOVERABLE_PARTITIONS/
3310 * A new function sd_hwdb_new_from_path() has been added to open a hwdb
3311 database given an explicit path to the file.
3313 * The signal number argument to sd_event_add_signal() now can now be
3314 ORed with the SD_EVENT_SIGNAL_PROCMASK flag, causing sigprocmask() to
3315 be automatically invoked to block the specified signal. This is
3316 useful to simplify invocations as the caller doesn't have to do this
3319 * A new convenience call sd_event_set_signal_exit() has been added to
3320 sd-event to set up signal handling so that the event loop
3321 automatically terminates cleanly on SIGTERM/SIGINT.
3323 Changes in other components:
3325 * systemd-sysusers, systemd-tmpfiles, and systemd-sysctl configuration
3326 can now be provided via the credential mechanism.
3328 * systemd-analyze gained a new verb 'compare-versions' that implements
3329 comparisons for versions strings (similarly to 'rpmdev-vercmp' and
3330 'dpkg --compare-versions').
3332 * 'systemd-analyze dump' is extended to accept glob patterns for unit
3333 names to limit the output to matching units.
3335 * tmpfiles.d/ lines can read file contents to write from a credential.
3336 The new modifier char '^' is used to specify that the argument is a
3337 credential name. This mechanism is used to automatically populate
3338 /etc/motd, /etc/issue, and /etc/hosts from credentials.
3340 * tmpfiles.d/ may now be configured to avoid changing uid/gid/mode of
3341 an inode if the specification is prefixed with ':' and the inode
3344 * Default tmpfiles.d/ configuration now carries a line to automatically
3345 use an 'ssh.authorized_keys.root' credential if provided to set up
3346 the SSH authorized_keys file for the root user.
3348 * systemd-tmpfiles will now gracefully handle absent source of "C" copy
3351 * tmpfiles.d/ F/w lines now optionally permit encoding of the payload
3352 in base64. This is useful to write arbitrary binary data into files.
3354 * The pkgconfig and rpm macros files now export the directory for user
3355 units as 'user_tmpfiles_dir' and '%_user_tmpfilesdir'.
3357 * Detection of Apple Virtualization and detection of Parallels and
3358 KubeVirt virtualization on non-x86 archs have been added.
3360 * os-release gained a new field SUPPORT_END=YYYY-MM-DD to inform the
3361 user when their system will become unsupported.
3363 * When performing suspend-then-hibernate, the system will estimate the
3364 discharge rate and use that to set the delay until hibernation and
3365 hibernate immediately instead of suspending when running from a
3366 battery and the capacity is below 5%.
3368 * systemd-sysctl gained a --strict option to fail when a sysctl
3369 setting is unknown to the kernel.
3371 * machinectl supports --force for the 'copy-to' and 'copy-from'
3374 * coredumpctl gained the --root and --image options to look for journal
3375 files under the specified root directory, image, or block device.
3377 * 'journalctl -o' and similar commands now implement a new output mode
3378 "short-delta". It is similar to "short-monotonic", but also shows the
3379 time delta between subsequent messages.
3381 * journalctl now respects the --quiet flag when verifying consistency
3384 * Journal log messages gained a new implicit field _RUNTIME_SCOPE= that
3385 will indicate whether a message was logged in the 'initrd' phase or
3386 in the 'system' phase of the boot process.
3388 * Journal files gained a new compatibility flag
3389 'HEADER_INCOMPATIBLE_COMPACT'. Files with this flag implement changes
3390 to the storage format that allow reducing size on disk. As with other
3391 compatibility flags, older journalctl versions will not be able to
3392 read journal files using this new format. The environment variable
3393 'SYSTEMD_JOURNAL_COMPACT=0' can be passed to systemd-journald to
3394 disable this functionality. It is enabled by default.
3396 * systemd-run's --working-directory= switch now works when used in
3397 combination with --scope.
3399 * portablectl gained a --force flag to skip certain sanity checks. This
3400 is implemented using new flags accepted by systemd-portabled for the
3401 *WithExtensions() D-Bus methods: SD_SYSTEMD_PORTABLE_FORCE_ATTACH
3402 flag now means that the attach/detach checks whether the units are
3403 already present and running will be skipped. Similarly,
3404 SD_SYSTEMD_PORTABLE_FORCE_SYSEXT flag means that the check whether
3405 image name matches the name declared inside of the image will be
3406 skipped. Callers must be sure to do those checks themselves if
3409 * systemd-portabled will now use the original filename to check
3410 extension-release.NAME for correctness, in case it is passed a
3413 * systemd-portabled now uses PrivateTmp=yes in the 'trusted' profile
3416 * sysext's extension-release files now support '_any' as a special
3417 value for the ID= field, to allow distribution-independent extensions
3418 (e.g.: fully statically compiled binaries, scripts). It also gained
3419 support for a new ARCHITECTURE= field that may be used to explicitly
3420 restrict an image to hosts of a specific architecture.
3422 * systemd-repart now supports creating squashfs partitions. This
3423 requires mksquashfs from squashfs-tools.
3425 * systemd-repart gained a --split flag to also generate split
3426 artifacts, i.e. a separate file for each partition. This is useful in
3427 conjunction with systemd-sysupdate or other tools, or to generate
3428 split dm-verity artifacts.
3430 * systemd-repart is now able to generate dm-verity partitions, including
3433 * systemd-repart can now set a partition UUID to zero, allowing it to
3434 be filled in later, such as when using verity partitions.
3436 * systemd-repart now supports drop-ins for its configuration files.
3438 * Package metadata logged by systemd-coredump in the system journal is
3441 * xdg-autostart-service now expands 'tilde' characters in Exec lines.
3443 * systemd-oomd now automatically links against libatomic, if available.
3445 * systemd-oomd now sends out a 'Killed' D-Bus signal when a cgroup is
3448 * scope units now also provide oom-kill status.
3450 * systemd-pstore will now try to load only the efi_pstore kernel module
3451 before running, ensuring that pstore can be used.
3453 * systemd-logind gained a new StopIdleSessionSec= option to stop an idle
3454 session after a preconfigure timeout.
3456 * systemd-homed will now wait up to 30 seconds for workers to terminate,
3457 rather than indefinitely.
3459 * homectl gained a new '--luks-sector-size=' flag that allows users to
3460 select the preferred LUKS sector size. Must be a power of 2 between 512
3461 and 4096. systemd-userdbd records gained a corresponding field.
3463 * systemd-sysusers will now respect the 'SOURCE_DATE_EPOCH' environment
3464 variable when generating the 'sp_lstchg' field, to ensure an image
3465 build can be reproducible.
3467 * 'udevadm wait' will now listen to kernel uevents too when called with
3470 * When naming network devices udev will now consult the Devicetree
3471 "alias" fields for the device.
3473 * systemd-udev will now create infiniband/by-path and
3474 infiniband/by-ibdev links for Infiniband verbs devices.
3476 * systemd-udev-trigger.service will now also prioritize input devices.
3478 * ConditionACPower= and systemd-ac-power will now assume the system is
3479 running on AC power if no battery can be found.
3481 * All features and tools using the TPM2 will now communicate with it
3482 using a bind key. Beforehand, the tpm2 support used encrypted sessions
3483 by creating a primary key that was used to encrypt traffic. This
3484 creates a problem as the key created for encrypting the traffic could
3485 be faked by an active interposer on the bus. In cases when a pin is
3486 used, a bind key will be used. The pin is used as the auth value for
3487 the seal key, aka the disk encryption key, and that auth value will be
3488 used in the session establishment. An attacker would need the pin
3489 value to create the secure session and thus an active interposer
3490 without the pin cannot interpose on TPM2 traffic.
3492 * systemd-growfs no longer requires udev to run.
3494 * systemd-backlight now will better support systems with multiple
3497 * systemd-cryptsetup's keyfile-timeout= option now also works when a
3498 device is used as a keyfile.
3500 * systemd-cryptenroll gained a new --unlock-key-file= option to get the
3501 unlocking key from a key file (instead of prompting the user). Note
3502 that this is the key for unlocking the volume in order to be able to
3503 enroll a new key, but it is not the key that is enrolled.
3505 * systemd-dissect gained a new --umount switch that will safely and
3506 synchronously unmount all partitions of an image previously mounted
3507 with 'systemd-dissect --mount'.
3509 * When using gcrypt, all systemd tools and services will now configure
3510 it to prefer the OS random number generator if present.
3512 * All example code shipped with documentation has been relicensed from CC0
3515 * Unit tests will no longer fail when running on a system without
3518 Experimental features:
3520 * BPF programs can now be compiled with bpf-gcc (requires libbpf >= 1.0
3521 and bpftool >= 7.0).
3523 * sd-boot can automatically enroll SecureBoot keys from files found on
3524 the ESP. This enrollment can be either automatic ('force' mode) or
3525 controlled by the user ('manual' mode). It is sufficient to place the
3526 SecureBoot keys in the right place in the ESP and they will be picked
3527 up by sd-boot and shown in the boot menu.
3529 * The mkosi config in systemd gained support for automatically
3530 compiling a kernel with the configuration appropriate for testing
3531 systemd. This may be useful when developing or testing systemd in
3532 tandem with the kernel.
3534 Contributions from: 김인수, Adam Williamson, adrian5, Aidan Dang,
3535 Akihiko Odaki, Alban Bedel, Albert Mikaelyan, Aleksey Vasenev,
3536 Alexander Graf, Alexander Shopov, Alexander Wilson,
3537 Alper Nebi Yasak, anarcat, Anders Jonsson, Andre Kalb,
3538 Andrew Stone, Andrey Albershteyn, Anita Zhang, Ansgar Burchardt,
3539 Antonio Alvarez Feijoo, Arnaud Ferraris, Aryan singh, asavah,
3540 Avamander, Avram Lubkin, Balázs Meskó, Bastien Nocera,
3541 Benjamin Franzke, BerndAdameit, bin456789, Celeste Liu,
3542 Chih-Hsuan Yen, Christian Brauner, Christian Göttsche,
3543 Christian Hesse, Clyde Byrd III, codefiles, Colin Walters,
3544 Cristian Rodríguez, Daan De Meyer, Daniel Braunwarth,
3545 Daniel Rusek, Dan Streetman, Darsey Litzenberger, David Edmundson,
3546 David Jaša, David Rheinsberg, David Seifert, David Tardon,
3547 dependabot[bot], Devendra Tewari, Dominique Martinet, drosdeck,
3548 Edson Juliano Drosdeck, Eduard Tolosa, eggfly, Einsler Lee,
3549 Elias Probst, Eli Schwartz, Evgeny Vereshchagin, exploide, Fei Li,
3550 Foster Snowhill, Franck Bui, Frank Dana, Frantisek Sumsal,
3551 Gerd Hoffmann, Gio, Goffredo Baroncelli, gtwang01,
3552 Guillaume W. Bres, H A, Hans de Goede, Heinrich Schuchardt,
3553 Hugo Carvalho, i-do-cpp, igo95862, j00512545, Jacek Migacz,
3554 Jade Bilkey, James Hilliard, Jan B, Janis Goldschmidt,
3555 Jan Janssen, Jan Kuparinen, Jan Luebbe, Jan Macku,
3556 Jason A. Donenfeld, Javkhlanbayar Khongorzul, Jeremy Soller,
3557 JeroenHD, jiangchuangang, João Loureiro,
3558 Joaquín Ignacio Aramendía, Jochen Sprickerhof,
3559 Johannes Schauer Marin Rodrigues, Jonas Kümmerlin,
3560 Jonas Witschel, Jonathan Kang, Jonathan Lebon, Joost Heitbrink,
3561 Jörg Thalheim, josh-gordon-fb, Joyce, Kai Lueke, lastkrick,
3562 Lennart Poettering, Leon M. George, licunlong, Li kunyu,
3563 LockBlock-dev, Loïc Collignon, Lubomir Rintel, Luca Boccassi,
3564 Luca BRUNO, Ludwig Nussel, Łukasz Stelmach, Maccraft123,
3565 Marc Kleine-Budde, Marius Vollmer, Martin Wilck, matoro,
3566 Matthias Lisin, Max Gautier, Maxim Mikityanskiy, Michael Biebl,
3567 Michal Koutný, Michal Sekletár, Michal Stanke, Mike Gilbert,
3568 Mitchell Freiderich, msizanoen1, Nick Rosbrook, nl6720, Oğuz Ersen,
3569 Oleg Solovyov, Olga Smirnova, Pablo Ceballos, Pavel Zhukov,
3570 Phaedrus Leeds, Philipp Gortan, Piotr Drąg, Pyfisch,
3571 Quentin Deslandes, Rahil Bhimjiani, Rene Hollander, Richard Huang,
3572 Richard Phibel, Rudi Heitbaum, Sam James, Sarah Brofeldt,
3573 Sean Anderson, Sebastian Scheibner, Shreenidhi Shedi,
3574 Sonali Srivastava, Steve Ramage, Suraj Krishnan, Swapnil Devesh,
3575 Takashi Sakamoto, Ted X. Toth, Temuri Doghonadze, Thomas Blume,
3576 Thomas Haller, Thomas Hebb, Tomáš Hnyk, Tomasz Paweł Gajc,
3577 Topi Miettinen, Ulrich Ölmann, undef, Uriel Corfa,
3578 Victor Westerhuis, Vincent Dagonneau, Vishal Chillara Srinivas,
3579 Vito Caputo, Weblate, Wenchao Hao, William Roberts, williamsumendap,
3580 wineway, xiaoyang, Yuri Chornoivan, Yu Watanabe,
3581 Zbigniew Jędrzejewski-Szmek, Zhaofeng Li, наб
3583 – The Great Beyond, 2022-10-31 👻
3587 Backwards-incompatible changes:
3589 * The minimum kernel version required has been bumped from 3.13 to 4.15,
3590 and CLOCK_BOOTTIME is now assumed to always exist.
3592 * C11 with GNU extensions (aka "gnu11") is now used to build our
3593 components. Public API headers are still restricted to ISO C89.
3595 * In v250, a systemd-networkd feature that automatically configures
3596 routes to addresses specified in AllowedIPs= was added and enabled by
3597 default. However, this causes network connectivity issues in many
3598 existing setups. Hence, it has been disabled by default since
3599 systemd-stable 250.3. The feature can still be used by explicitly
3600 configuring RouteTable= setting in .netdev files.
3602 * Jobs started via StartUnitWithFlags() will no longer return 'skipped'
3603 when a Condition*= check does not succeed, restoring the JobRemoved
3604 signal to the behaviour it had before v250.
3606 * The org.freedesktop.portable1 methods GetMetadataWithExtensions() and
3607 GetImageMetadataWithExtensions() have been fixed to provide an extra
3608 return parameter, containing the actual extension release metadata.
3609 The current implementation was judged to be broken and unusable, and
3610 thus the usual procedure of adding a new set of methods was skipped,
3611 and backward compatibility broken instead on the assumption that
3612 nobody can be affected given the current state of this interface.
3614 * All kernels supported by systemd mix bytes returned by RDRAND (or
3615 similar) into the entropy pool at early boot. This means that on
3616 those systems, even if /dev/urandom is not yet initialized, it still
3617 returns bytes that are of at least RDRAND quality. For that reason,
3618 we no longer have reason to invoke RDRAND from systemd itself, which
3619 has historically been a source of bugs. Furthermore, kernels ≥5.6
3620 provide the getrandom(GRND_INSECURE) interface for returning random
3621 bytes before the entropy pool is initialized without warning into
3622 kmsg, which is what we attempt to use if available. systemd's direct
3623 usage of RDRAND has been removed. x86 systems ≥Broadwell that are
3624 running an older kernel may experience kmsg warnings that were not
3625 seen with 250. For newer kernels, non-x86 systems, or older x86
3626 systems, there should be no visible changes.
3628 * sd-boot will now measure the kernel command line into TPM PCR 12
3629 rather than PCR 8. This improves usefulness of the measurements on
3630 systems where sd-boot is chainloaded from Grub. Grub measures all
3631 commands its executes into PCR 8, which makes it very hard to use
3632 reasonably, hence separate ourselves from that and use PCR 12
3633 instead, which is what certain Ubuntu editions already do. To retain
3634 compatibility with systems running older systemd systems a new meson
3635 option 'efi-tpm-pcr-compat' has been added (which defaults to false).
3636 If enabled, the measurement is done twice: into the new-style PCR 12
3637 *and* the old-style PCR 8. It's strongly advised to migrate all users
3638 to PCR 12 for this purpose in the long run, as we intend to remove
3639 this compatibility feature in two years' time.
3641 * busctl capture now writes output in the newer pcapng format instead
3644 * A udev rule that imported hwdb matches for USB devices with lowercase
3645 hexadecimal vendor/product ID digits was added in systemd 250. This
3646 has been reverted, since uppercase hexadecimal digits are supposed to
3647 be used, and we already had a rule with the appropriate match.
3649 Users might need to adjust their local hwdb entries.
3651 * arch_prctl(2) has been moved to the @default set in the syscall filters
3652 (as exposed via the SystemCallFilter= setting in service unit files).
3653 It is apparently used by the linker now.
3655 * The tmpfiles entries that create the /run/systemd/netif directory and
3656 its subdirectories were moved from tmpfiles.d/systemd.conf to
3657 tmpfiles.d/systemd-network.conf.
3659 Users might need to adjust their files that override tmpfiles.d/systemd.conf
3660 to account for this change.
3662 * The requirement for Portable Services images to contain a well-formed
3663 os-release file (i.e.: contain at least an ID field) is now enforced.
3664 This applies to base images and extensions, and also to systemd-sysext.
3666 Changes in the Boot Loader Specification, kernel-install and sd-boot:
3668 * kernel-install's and bootctl's Boot Loader Specification Type #1
3669 entry generation logic has been reworked. The user may now pick
3670 explicitly by which "token" string to name the installation's boot
3671 entries, via the new /etc/kernel/entry-token file or the new
3672 --entry-token= switch to bootctl. By default — as before — the
3673 entries are named after the local machine ID. However, in "golden
3674 image" environments, where the machine ID shall be initialized on
3675 first boot (as opposed to at installation time before first boot) the
3676 machine ID will not be available at build time. In this case the
3677 --entry-token= switch to bootctl (or the /etc/kernel/entry-token
3678 file) may be used to override the "token" for the entries, for
3679 example the IMAGE_ID= or ID= fields from /etc/os-release. This will
3680 make the OS images independent of any machine ID, and ensure that the
3681 images will not carry any identifiable information before first boot,
3682 but on the other hand means that multiple parallel installations of
3683 the very same image on the same disk cannot be supported.
3685 Summary: if you are building golden images that shall acquire
3686 identity information exclusively on first boot, make sure to both
3687 remove /etc/machine-id *and* to write /etc/kernel/entry-token to the
3688 value of the IMAGE_ID= or ID= field of /etc/os-release or another
3689 suitable identifier before deploying the image.
3691 * The Boot Loader Specification has been extended with
3692 /loader/entries.srel file located in the EFI System Partition (ESP)
3693 that disambiguates the format of the entries in the /loader/entries/
3694 directory (in order to discern them from incompatible uses of this
3695 directory by other projects). For entries that follow the
3696 Specification, the string "type1" is stored in this file.
3698 bootctl will now write this file automatically when installing the
3699 systemd-boot boot loader.
3701 * kernel-install supports a new initrd_generator= setting in
3702 /etc/kernel/install.conf, that is exported as
3703 $KERNEL_INSTALL_INITRD_GENERATOR to kernel-install plugins. This
3704 allows choosing different initrd generators.
3706 * kernel-install will now create a "staging area" (an initially-empty
3707 directory to gather files for a Boot Loader Specification Type #1
3708 entry). The path to this directory is exported as
3709 $KERNEL_INSTALL_STAGING_AREA to kernel-install plugins, which should
3710 drop files there instead of writing them directly to the final
3711 location. kernel-install will move them when all files have been
3712 prepared successfully.
3714 * New option sort-key= has been added to the Boot Loader Specification
3715 to override the sorting order of the entries in the boot menu. It is
3716 read by sd-boot and bootctl, and will be written by kernel-install,
3717 with the default value of IMAGE_ID= or ID= fields from
3718 os-release. Together, this means that on multiboot installations,
3719 entries should be grouped and sorted in a predictable way.
3721 * The sort order of boot entries has been updated: entries which have
3722 the new field sort-key= are sorted by it first, and all entries
3723 without it are ordered later. After that, entries are sorted by
3724 version so that newest entries are towards the beginning of the list.
3726 * The kernel-install tool gained a new 'inspect' verb which shows the
3727 paths and other settings used.
3729 * sd-boot can now optionally beep when the menu is shown and menu
3730 entries are selected, which can be useful on machines without a
3731 working display. (Controllable via a loader.conf setting.)
3733 * The --make-machine-id-directory= switch to bootctl has been replaced
3734 by --make-entry-directory=, given that the entry directory is not
3735 necessarily named after the machine ID, but after some other suitable
3736 ID as selected via --entry-token= described above. The old name of
3737 the option is still understood to maximize compatibility.
3739 * 'bootctl list' gained support for a new --json= switch to output boot
3740 menu entries in JSON format.
3742 * 'bootctl is-installed' now supports the --graceful, and various verbs
3743 omit output with the new option --quiet.
3745 Changes in systemd-homed:
3747 * Starting with v250 systemd-homed uses UID/GID mapping on the mounts
3748 of activated home directories it manages (if the kernel and selected
3749 file systems support it). So far it mapped three UID ranges: the
3750 range from 0…60000, the user's own UID, and the range 60514…65534,
3751 leaving everything else unmapped (in other words, the 16-bit UID range
3752 is mapped almost fully, with the exception of the UID subrange used
3753 for systemd-homed users, with one exception: the user's own UID).
3754 Unmapped UIDs may not be used for file ownership in the home
3755 directory — any chown() attempts with them will fail. With this
3756 release a fourth range is added to these mappings:
3757 524288…1879048191. This range is the UID range intended for container
3760 https://systemd.io/UIDS-GIDS
3762 This range may be used for container managers that place container OS
3763 trees in the home directory (which is a questionable approach, for
3764 quota, permission, SUID handling and network file system
3765 compatibility reasons, but nonetheless apparently commonplace). Note
3766 that this mapping is mapped 1:1 in a pass-through fashion, i.e. the
3767 UID assignments from the range are not managed or mapped by
3768 `systemd-homed`, and must be managed with other mechanisms, in the
3769 context of the local system.
3771 Typically, a better approach to user namespacing in relevant
3772 container managers would be to leave container OS trees on disk at
3773 UID offset 0, but then map them to a dynamically allocated runtime
3774 UID range via another UID mount map at container invocation
3775 time. That way user namespace UID ranges become strictly a runtime
3776 concept, and do not leak into persistent file systems, persistent
3777 user databases or persistent configuration, thus greatly simplifying
3778 handling, and improving compatibility with home directories intended
3779 to be portable like the ones managed by systemd-homed.
3781 Changes in shared libraries:
3783 * A new libsystemd-core-<version>.so private shared library is
3784 installed under /usr/lib/systemd/system, mirroring the existing
3785 libsystemd-shared-<version>.so library. This allows the total
3786 installation size to be reduced by binary code reuse.
3788 * The <version> tag used in the name of libsystemd-shared.so and
3789 libsystemd-core.so can be configured via the meson option
3790 'shared-lib-tag'. Distributions may build subsequent versions of the
3791 systemd package with unique tags (e.g. the full package version),
3792 thus allowing multiple installations of those shared libraries to be
3793 available at the same time. This is intended to fix an issue where
3794 programs that link to those libraries would fail to execute because
3795 they were installed earlier or later than the appropriate version of
3798 * The sd-id128 API gained a new call sd_id128_to_uuid_string() that is
3799 similar to sd_id128_to_string() but formats the ID in RFC 4122 UUID
3800 format instead of as a simple series of hex characters.
3802 * The sd-device API gained two new calls sd_device_new_from_devname()
3803 and sd_device_new_from_path() which permit allocating an sd_device
3804 object from a device node name or file system path.
3806 * sd-device also gained a new call sd_device_open() which will open the
3807 device node associated with a device for which an sd_device object
3808 has been allocated. The call is supposed to address races around
3809 device nodes being removed/recycled due to hotplug events, or media
3810 change events: the call checks internally whether the major/minor of
3811 the device node and the "diskseq" (in case of block devices) match
3812 with the metadata loaded in the sd_device object, thus ensuring that
3813 the device once opened really matches the provided sd_device object.
3815 Changes in PID1, systemctl, and systemd-oomd:
3817 * A new set of service monitor environment variables will be passed to
3818 OnFailure=/OnSuccess= handlers, but only if exactly one unit lists the
3819 handler unit as OnFailure=/OnSuccess=. The variables are:
3820 $MONITOR_SERVICE_RESULT, $MONITOR_EXIT_CODE, $MONITOR_EXIT_STATUS,
3821 $MONITOR_INVOCATION_ID and $MONITOR_UNIT. For cases when a single
3822 handler needs to watch multiple units, use a templated handler.
3824 * A new ExtensionDirectories= setting in service unit files allows
3825 system extensions to be loaded from a directory. (It is similar to
3826 ExtensionImages=, but takes paths to directories, instead of
3829 'portablectl attach --extension=' now also accepts directory paths.
3831 * The user.delegate and user.invocation_id extended attributes on
3832 cgroups are used in addition to trusted.delegate and
3833 trusted.invocation_id. The latter pair requires privileges to set,
3834 but the former doesn't and can be also set by the unprivileged user
3837 (Only supported on kernels ≥5.6.)
3839 * Units that were killed by systemd-oomd will now have a service result
3840 of 'oom-kill'. The number of times a service was killed is tallied
3841 in the 'user.oomd_ooms' extended attribute.
3843 The OOMPolicy= unit file setting is now also honoured by
3846 * In unit files the new %y/%Y specifiers can be used to refer to
3847 normalized unit file path, which is particularly useful for symlinked
3850 The new %q specifier resolves to the pretty hostname
3851 (i.e. PRETTY_HOSTNAME= from /etc/machine-info).
3853 The new %d specifier resolves to the credentials directory of a
3854 service (same as $CREDENTIALS_DIRECTORY).
3856 * The RootDirectory=, MountAPIVFS=, ExtensionDirectories=,
3857 *Capabilities*=, ProtectHome=, *Directory=, TemporaryFileSystem=,
3858 PrivateTmp=, PrivateDevices=, PrivateNetwork=, NetworkNamespacePath=,
3859 PrivateIPC=, IPCNamespacePath=, PrivateUsers=, ProtectClock=,
3860 ProtectKernelTunables=, ProtectKernelModules=, ProtectKernelLogs=,
3861 MountFlags= service settings now also work in unprivileged user
3862 services, i.e. those run by the user's --user service manager, as long
3863 as user namespaces are enabled on the system.
3865 * Services with Restart=always and a failing ExecCondition= will no
3866 longer be restarted, to bring ExecCondition= behaviour in line with
3867 Condition*= settings.
3869 * LoadCredential= now accepts a directory as the argument; all files
3870 from the directory will be loaded as credentials.
3872 * A new D-Bus property ControlGroupId is now exposed on service units,
3873 that encapsulates the service's numeric cgroup ID that newer kernels
3874 assign to each cgroup.
3876 * PID 1 gained support for configuring the "pre-timeout" of watchdog
3877 devices and the associated governor, via the new
3878 RuntimeWatchdogPreSec= and RuntimeWatchdogPreGovernor= configuration
3879 options in /etc/systemd/system.conf.
3881 * systemctl's --timestamp= option gained a new choice "unix", to show
3882 timestamp as unix times, i.e. seconds since 1970, Jan 1st.
3884 * A new "taint" flag named "old-kernel" is introduced which is set when
3885 the kernel systemd runs on is older then the current baseline version
3886 (see above). The flag is shown in "systemctl status" output.
3888 * Two additional taint flags "short-uid-range" and "short-gid-range"
3889 have been added as well, which are set when systemd notices it is run
3890 within a userns namespace that does not define the full 0…65535 UID
3893 * A new "unmerged-usr" taint flag has been added that is set whenever
3894 running on systems where /bin/ + /sbin/ are *not* symlinks to their
3895 counterparts in /usr/, i.e. on systems where the /usr/-merge has not
3898 * Generators invoked by PID 1 will now have a couple of useful
3899 environment variables set describing the execution context a
3900 bit. $SYSTEMD_SCOPE encodes whether the generator is called from the
3901 system service manager, or from the per-user service
3902 manager. $SYSTEMD_IN_INITRD encodes whether the generator is invoked
3903 in initrd context or on the host. $SYSTEMD_FIRST_BOOT encodes whether
3904 systemd considers the current boot to be a "first"
3905 boot. $SYSTEMD_VIRTUALIZATION encode whether virtualization is
3906 detected and which type of hypervisor/container
3907 manager. $SYSTEMD_ARCHITECTURE indicates which architecture the
3908 kernel is built for.
3910 * PID 1 will now automatically pick up system credentials from qemu's
3911 fw_cfg interface, thus allowing passing arbitrary data into VM
3912 systems similar to how this is already supported for passing them
3913 into `systemd-nspawn` containers. Credentials may now also be passed
3914 in via the new kernel command line option `systemd.set_credential=`
3915 (note that kernel command line options are world-readable during
3916 runtime, and only useful for credentials that require no
3917 confidentiality). The credentials that can be passed to unified
3918 kernels that use the `systemd-stub` UEFI stub are now similarly
3919 picked up automatically. Automatic importing of system credentials
3920 this way can be turned off via the new
3921 `systemd.import_credentials=no` kernel command line option.
3923 * LoadCredential= will now automatically look for credentials in the
3924 /etc/credstore/, /run/credstore/, /usr/lib/credstore/ directories if
3925 the argument is not an absolute path. Similarly,
3926 LoadCredentialEncrypted= will check the same directories plus
3927 /etc/credstore.encrypted/, /run/credstore.encrypted/ and
3928 /usr/lib/credstore.encrypted/. The idea is to use those directories
3929 as the system-wide location for credentials that services should pick
3932 * System and service credentials are described in great detail in a new
3935 https://systemd.io/CREDENTIALS
3937 Changes in systemd-journald:
3939 * The journal JSON export format has been added to listed of stable
3940 interfaces (https://systemd.io/PORTABILITY_AND_STABILITY/).
3942 * journalctl --list-boots now supports JSON output and the --reverse option.
3944 * Under docs/: JOURNAL_EXPORT_FORMATS was imported from the wiki and
3945 updated, BUILDING_IMAGES is new:
3947 https://systemd.io/JOURNAL_EXPORT_FORMATS
3948 https://systemd.io/BUILDING_IMAGES
3952 * Two new hwdb files have been added. One lists "handhelds" (PDAs,
3953 calculators, etc.), the other AV production devices (DJ tables,
3954 keypads, etc.) that should accessible to the seat owner user by
3957 * udevadm trigger gained a new --prioritized-subsystem= option to
3958 process certain subsystems (and all their parent devices) earlier.
3960 systemd-udev-trigger.service now uses this new option to trigger
3961 block and TPM devices first, hopefully making the boot a bit faster.
3963 * udevadm trigger now implements --type=all, --initialized-match,
3964 --initialized-nomatch to trigger both subsystems and devices, only
3965 already-initialized devices, and only devices which haven't been
3966 initialized yet, respectively.
3968 * udevadm gained a new "wait" command for safely waiting for a specific
3969 device to show up in the udev device database. This is useful in
3970 scripts that asynchronously allocate a block device (e.g. through
3971 repartitioning, or allocating a loopback device or similar) and need
3972 to synchronize on the creation to complete.
3974 * udevadm gained a new "lock" command for locking one or more block
3975 devices while formatting it or writing a partition table to it. It is
3976 an implementation of https://systemd.io/BLOCK_DEVICE_LOCKING and
3977 usable in scripts dealing with block devices.
3979 * udevadm info will show a couple of additional device fields in its
3980 output, and will not apply a limited set of coloring to line types.
3982 * udevadm info --tree will now show a tree of objects (i.e. devices and
3983 suchlike) in the /sys/ hierarchy.
3985 * Block devices will now get a new set of device symlinks in
3986 /dev/disk/by-diskseq/<nr>, which may be used to reference block
3987 device nodes via the kernel's "diskseq" value. Note that this does
3988 not guarantee that opening a device by a symlink like this will
3989 guarantee that the opened device actually matches the specified
3990 diskseq value. To be safe against races, the actual diskseq value of
3991 the opened device (BLKGETDISKSEQ ioctl()) must still be compred with
3992 the one in the symlink path.
3994 * .link files gained support for setting MDI/MID-X on a link.
3996 * .link files gained support for [Match] Firmware= setting to match on
3997 the device firmware description string. By mistake, it was previously
3998 only supported in .network files.
4000 * .link files gained support for [Link] SR-IOVVirtualFunctions= setting
4001 and [SR-IOV] section to configure SR-IOV virtual functions.
4003 Changes in systemd-networkd:
4005 * The default scope for unicast routes configured through [Route]
4006 section is changed to "link", to make the behavior consistent with
4007 "ip route" command. The manual configuration of [Route] Scope= is
4010 * A new unit systemd-networkd-wait-online@<interface>.service has been
4011 added that can be used to wait for a specific network interface to be
4014 * systemd-networkd gained a new [Bridge] Isolated=true|false setting
4015 that configures the eponymous kernel attribute on the bridge.
4017 * .netdev files now can be used to create virtual WLAN devices, and
4018 configure various settings on them, via the [WLAN] section.
4020 * .link/.network files gained support for [Match] Kind= setting to match
4021 on device kind ("bond", "bridge", "gre", "tun", "veth", etc.)
4023 This value is also shown by 'networkctl status'.
4025 * The Local= setting in .netdev files for various virtual network
4026 devices gained support for specifying, in addition to the network
4027 address, the name of a local interface which must have the specified
4030 * systemd-networkd gained a new [Tunnel] External= setting in .netdev
4031 files, to configure tunnels in external mode (a.k.a. collect metadata
4034 * [Network] L2TP= setting was removed. Please use interface specifier in
4035 Local= setting in .netdev files of corresponding L2TP interface.
4037 * New [DHCPServer] BootServerName=, BootServerAddress=, and
4038 BootFilename= settings can be used to configure the server address,
4039 server name, and file name sent in the DHCP packet (e.g. to configure
4042 Changes in systemd-resolved:
4044 * systemd-resolved is started earlier (in sysinit.target), so it
4045 available earlier and will also be started in the initrd if installed
4048 Changes in disk encryption:
4050 * systemd-cryptenroll can now control whether to require the user to
4051 enter a PIN when using TPM-based unlocking of a volume via the new
4052 --tpm2-with-pin= option.
4054 Option tpm2-pin= can be used in /etc/crypttab.
4056 * When unlocking devices via TPM, TPM2 parameter encryption is now
4057 used, to ensure that communication between CPU and discrete TPM chips
4058 cannot be eavesdropped to acquire disk encryption keys.
4060 * A new switch --fido2-credential-algorithm= has been added to
4061 systemd-cryptenroll allowing selection of the credential algorithm to
4062 use when binding encryption to FIDO2 tokens.
4064 Changes in systemd-hostnamed:
4066 * HARDWARE_VENDOR= and HARDWARE_MODEL= can be set in /etc/machine-info
4067 to override the values gleaned from the hwdb.
4069 * A ID_CHASSIS property can be set in the hwdb (for the DMI device
4070 /sys/class/dmi/id) to override the chassis that is reported by
4073 * hostnamed's D-Bus interface gained a new method GetHardwareSerial()
4074 for reading the hardware serial number, as reportd by DMI. It also
4075 exposes a new method D-Bus property FirmwareVersion that encode the
4076 firmware version of the system.
4078 Changes in other components:
4080 * /etc/locale.conf is now populated through tmpfiles.d factory /etc/
4081 handling with the values that were configured during systemd build
4082 (if /etc/locale.conf has not been created through some other
4083 mechanism). This means that /etc/locale.conf should always have
4084 reasonable contents and we avoid a potential mismatch in defaults.
4086 * The userdbctl tool will now show UID range information as part of the
4087 list of known users.
4089 * A new build-time configuration setting default-user-shell= can be
4090 used to set the default shell for user records and nspawn shell
4091 invocations (instead of the default /bin/bash).
4093 * systemd-timesyncd now provides a D-Bus API for receiving NTP server
4094 information dynamically at runtime via IPC.
4096 * The systemd-creds tool gained a new "has-tpm2" verb, which reports
4097 whether a functioning TPM2 infrastructure is available, i.e. if
4098 firmware, kernel driver and systemd all have TPM2 support enabled and
4101 * The systemd-creds tool gained support for generating encrypted
4102 credentials that are using an empty encryption key. While this
4103 provides no integrity nor confidentiality it's useful to implement
4104 codeflows that work the same on TPM-ful and TPM2-less systems. The
4105 service manager will only accept credentials "encrypted" that way if
4106 a TPM2 device cannot be detected, to ensure that credentials
4107 "encrypted" like that cannot be used to trick TPM2 systems.
4109 * When deciding whether to colorize output, all systemd programs now
4110 also check $COLORTERM (in addition to $NO_COLOR, $SYSTEMD_COLORS, and
4113 * Meson's new install_tag feature is now in use for several components,
4114 allowing to build and install select binaries only: pam, nss, devel
4115 (pkg-config files), systemd-boot, libsystemd, libudev. Example:
4116 $ meson build systemd-boot
4117 $ meson install --tags systemd-boot --no-rebuild
4118 https://mesonbuild.com/Installing.html#installation-tags
4120 * A new build configuration option has been added, to allow selecting the
4121 default compression algorithm used by systemd-journald and systemd-coredump.
4122 This allows to build-in support for decompressing all supported formats,
4123 but choose a specific one for compression. E.g.:
4124 $ meson -Ddefault-compression=xz
4126 Experimental features:
4128 * sd-boot gained a new *experimental* setting "reboot-for-bitlocker" in
4129 loader.conf that implements booting Microsoft Windows from the
4130 sd-boot in a way that first reboots the system, to reset the TPM
4131 PCRs. This improves compatibility with BitLocker's TPM use, as the
4132 PCRs will only record the Windows boot process, and not sd-boot
4133 itself, thus retaining the PCR measurements not involving sd-boot.
4134 Note that this feature is experimental for now, and is likely going
4135 to be generalized and renamed in a future release, without retaining
4136 compatibility with the current implementation.
4138 * A new systemd-sysupdate component has been added that automatically
4139 discovers, downloads, and installs A/B-style updates for the host
4140 installation itself, or container images, portable service images,
4141 and other assets. See the new systemd-sysupdate man page for updates.
4143 Contributions from: 4piu, Adam Williamson, adrian5, Albert Brox,
4144 AlexCatze, Alex Henrie, Alfonso Sánchez-Beato, Alice S,
4145 Alvin Šipraga, amarjargal, Amarjargal, Andrea Pappacoda,
4146 Andreas Rammhold, Andy Chi, Anita Zhang, Antonio Alvarez Feijoo,
4147 Arfrever Frehtes Taifersar Arahesis, ash, Bastien Nocera, Be,
4148 bearhoney, Ben Efros, Benjamin Berg, Benjamin Franzke,
4149 Brett Holman, Christian Brauner, Clyde Byrd III, Curtis Klein,
4150 Daan De Meyer, Daniele Medri, Daniel Mack, Danilo Krummrich,
4151 David, David Bond, Davide Cavalca, David Tardon, davijosw,
4152 dependabot[bot], Donald Chan, Dorian Clay, Eduard Tolosa,
4153 Elias Probst, Eli Schwartz, Erik Sjölund, Evgeny Vereshchagin,
4154 Federico Ceratto, Franck Bui, Frantisek Sumsal, Gaël PORTAY,
4155 Georges Basile Stavracas Neto, Gibeom Gwon, Goffredo Baroncelli,
4156 Grigori Goronzy, Hans de Goede, Heiko Becker, Hugo Carvalho,
4157 Jakob Lell, James Hilliard, Jan Janssen, Jason A. Donenfeld,
4158 Joan Bruguera, Joerie de Gram, Josh Triplett, Julia Kartseva,
4159 Kazuo Moriwaka, Khem Raj, ksa678491784, Lance, Lan Tian,
4160 Laura Barcziova, Lennart Poettering, Leviticoh, licunlong,
4161 Lidong Zhong, lincoln auster, Lubomir Rintel, Luca Boccassi,
4162 Luca BRUNO, lucagoc, Ludwig Nussel, Marcel Hellwig, march1993,
4163 Marco Scardovi, Mario Limonciello, Mariusz Tkaczyk,
4164 Markus Weippert, Martin, Martin Liska, Martin Wilck, Matija Skala,
4165 Matthew Blythe, Matthias Lisin, Matthijs van Duin, Matt Walton,
4166 Max Gautier, Michael Biebl, Michael Olbrich, Michal Koutný,
4167 Michal Sekletár, Mike Gilbert, MkfsSion, Morten Linderud,
4168 Nick Rosbrook, Nikolai Grigoriev, Nikolai Kostrigin,
4169 Nishal Kulkarni, Noel Kuntze, Pablo Ceballos, Peter Hutterer,
4170 Peter Morrow, Pigmy-penguin, Piotr Drąg, prumian, Richard Neill,
4171 Rike-Benjamin Schuppner, rodin-ia, Romain Naour, Ruben Kerkhof,
4172 Ryan Hendrickson, Santa Wiryaman, Sebastian Pucilowski, Seth Falco,
4173 Simon Ellmann, Sonali Srivastava, Stefan Seering,
4174 Stephen Hemminger, tawefogo, techtino, Temuri Doghonadze,
4175 Thomas Batten, Thomas Haller, Thomas Weißschuh, Tobias Stoeckmann,
4176 Tomasz Pala, Tyson Whitehead, Vishal Chillara Srinivas,
4177 Vivien Didelot, w30023233, wangyuhang, Weblate, Xiaotian Wu,
4178 yangmingtai, YmrDtnJu, Yonathan Randolph, Yutsuten, Yu Watanabe,
4179 Zbigniew Jędrzejewski-Szmek, наб
4181 — Edinburgh, 2022-05-21
4185 * Support for encrypted and authenticated credentials has been added.
4186 This extends the credential logic introduced with v247 to support
4187 non-interactive symmetric encryption and authentication, based on a
4188 key that is stored on the /var/ file system or in the TPM2 chip (if
4189 available), or the combination of both (by default if a TPM2 chip
4190 exists the combination is used, otherwise the /var/ key only). The
4191 credentials are automatically decrypted at the moment a service is
4192 started, and are made accessible to the service itself in unencrypted
4193 form. A new tool 'systemd-creds' encrypts credentials for this
4194 purpose, and two new service file settings LoadCredentialEncrypted=
4195 and SetCredentialEncrypted= configure such credentials.
4197 This feature is useful to store sensitive material such as SSL
4198 certificates, passwords and similar securely at rest and only decrypt
4199 them when needed, and in a way that is tied to the local OS
4200 installation or hardware.
4202 * systemd-gpt-auto-generator can now automatically set up discoverable
4203 LUKS2 encrypted swap partitions.
4205 * The GPT Discoverable Partitions Specification has been substantially
4206 extended with support for root and /usr/ partitions for the majority
4207 of architectures systemd supports. This includes platforms that do
4208 not natively support UEFI, because even though GPT is specified under
4209 UEFI umbrella, it is useful on other systems too. Specifically,
4210 systemd-nspawn, systemd-sysext, systemd-gpt-auto-generator and
4211 Portable Services use the concept without requiring UEFI.
4213 * The GPT Discoverable Partitions Specifications has been extended with
4214 a new set of partitions that may carry PKCS#7 signatures for Verity
4215 partitions, encoded in a simple JSON format. This implements a simple
4216 mechanism for building disk images that are fully authenticated and
4217 can be tested against a set of cryptographic certificates. This is
4218 now implemented for the various systemd tools that can operate with
4219 disk images, such as systemd-nspawn, systemd-sysext, systemd-dissect,
4220 Portable services/RootImage=, systemd-tmpfiles, and systemd-sysusers.
4221 The PKCS#7 signatures are passed to the kernel (where they are
4222 checked against certificates from the kernel keyring), or can be
4223 verified against certificates provided in userspace (via a simple
4224 drop-in file mechanism).
4226 * systemd-dissect's inspection logic will now report for which uses a
4227 disk image is intended. Specifically, it will display whether an
4228 image is suitable for booting on UEFI or in a container (using
4229 systemd-nspawn's --image= switch), whether it can be used as portable
4230 service, or attached as system extension.
4232 * The system-extension.d/ drop-in files now support a new field
4233 SYSEXT_SCOPE= that may encode which purpose a system extension image
4234 is for: one of "initrd", "system" or "portable". This is useful to
4235 make images more self-descriptive, and to ensure system extensions
4236 cannot be attached in the wrong contexts.
4238 * The os-release file learnt a new PORTABLE_PREFIXES= field which may
4239 be used in portable service images to indicate which unit prefixes
4242 * The GPT image dissection logic in systemd-nspawn/systemd-dissect/…
4243 now is able to decode images for non-native architectures as well.
4244 This allows systemd-nspawn to boot images of non-native architectures
4245 if the corresponding user mode emulator is installed and
4246 systemd-binfmtd is running.
4248 * systemd-logind gained new settings HandlePowerKeyLongPress=,
4249 HandleRebootKeyLongPress=, HandleSuspendKeyLongPress= and
4250 HandleHibernateKeyLongPress= which may be used to configure actions
4251 when the relevant keys are pressed for more than 5s. This is useful
4252 on devices that only have hardware for a subset of these keys. By
4253 default, if the reboot key is pressed long the poweroff operation is
4254 now triggered, and when the suspend key is pressed long the hibernate
4255 operation is triggered. Long pressing the other two keys currently
4256 does not trigger any operation by default.
4258 * When showing unit status updates on the console during boot and
4259 shutdown, and a service is slow to start so that the cylon animation
4260 is shown, the most recent sd_notify() STATUS= text is now shown as
4261 well. Services may use this to make the boot/shutdown output easier
4262 to understand, and to indicate what precisely a service that is slow
4263 to start or stop is waiting for. In particular, the per-user service
4264 manager instance now reports what it is doing and which service it is
4265 waiting for this way to the system service manager.
4267 * The service manager will now re-execute on reception of the
4268 SIGRTMIN+25 signal. It previously already did that on SIGTERM — but
4269 only when running as PID 1. There was no signal to request this when
4270 running as per-user service manager, i.e. as any other PID than 1.
4271 SIGRTMIN+25 works for both system and user managers.
4273 * The hardware watchdog logic in PID 1 gained support for operating
4274 with the default timeout configured in the hardware, instead of
4275 insisting on re-configuring it. Set RuntimeWatchdogSec=default to
4276 request this behavior.
4278 * A new kernel command line option systemd.watchdog_sec= is now
4279 understood which may be used to override the hardware watchdog
4280 time-out for the boot.
4282 * A new setting DefaultOOMScoreAdjust= is now supported in
4283 /etc/systemd/system.conf and /etc/systemd/user.conf. It may be used
4284 to set the default process OOM score adjustment value for processes
4285 started by the service manager. For per-user service managers this
4286 now defaults to 100, but for per-system service managers is left as
4287 is. This means that by default now services forked off the user
4288 service manager are more likely to be killed by the OOM killer than
4289 system services or the managers themselves.
4291 * A new per-service setting RestrictFileSystems= as been added that
4292 restricts the file systems a service has access to by their type.
4293 This is based on the new BPF LSM of the Linux kernel. It provides an
4294 effective way to make certain API file systems unavailable to
4295 services (and thus minimizing attack surface). A new command
4296 "systemd-analyze filesystems" has been added that lists all known
4297 file system types (and how they are grouped together under useful
4300 * Services now support a new setting RestrictNetworkInterfaces= for
4301 restricting access to specific network interfaces.
4303 * Service unit files gained new settings StartupAllowedCPUs= and
4304 StartupAllowedMemoryNodes=. These are similar to their counterparts
4305 without the "Startup" prefix and apply during the boot process
4306 only. This is useful to improve boot-time behavior of the system and
4307 assign resources differently during boot than during regular
4308 runtime. This is similar to the preexisting StartupCPUWeight=
4311 * Related to this: the various StartupXYZ= settings
4312 (i.e. StartupCPUWeight=, StartupAllowedCPUs=, …) are now also applied
4313 during shutdown. The settings not prefixed with "Startup" hence apply
4314 during regular runtime, and those that are prefixed like that apply
4315 during boot and shutdown.
4317 * A new per-unit set of conditions/asserts
4318 [Condition|Assert][Memory|CPU|IO]Pressure= have been added to make a
4319 unit skip/fail activation if the system's (or a slice's) memory/cpu/io
4320 pressure is above the configured threshold, using the kernel PSI
4321 feature. For more details see systemd.unit(5) and
4322 https://docs.kernel.org/accounting/psi.html
4324 * The combination of ProcSubset=pid and ProtectKernelTunables=yes and/or
4325 ProtectKernelLogs=yes can now be used.
4327 * The default maximum numbers of inodes have been raised from 64k to 1M
4328 for /dev/, and from 400k to 1M for /tmp/.
4330 * The per-user service manager learnt support for communicating with
4331 systemd-oomd to acquire OOM kill information.
4333 * A new service setting ExecSearchPath= has been added that allows
4334 changing the search path for executables for services. It affects
4335 where we look for the binaries specified in ExecStart= and similar,
4336 and the specified directories are also added the $PATH environment
4337 variable passed to invoked processes.
4339 * A new setting RuntimeRandomizedExtraSec= has been added for service
4340 and scope units that allows extending the runtime time-out as
4341 configured by RuntimeMaxSec= with a randomized amount.
4343 * The syntax of the service unit settings RuntimeDirectory=,
4344 StateDirectory=, CacheDirectory=, LogsDirectory= has been extended:
4345 if the specified value is now suffixed with a colon, followed by
4346 another filename, the latter will be created as symbolic link to the
4347 specified directory. This allows creating these service directories
4348 together with alias symlinks to make them available under multiple
4351 * Service unit files gained two new settings TTYRows=/TTYColumns= for
4352 configuring rows/columns of the TTY device passed to
4353 stdin/stdout/stderr of the service. This is useful to propagate TTY
4354 dimensions to a virtual machine.
4356 * A new service unit file setting ExitType= has been added that
4357 specifies when to assume a service has exited. By default systemd
4358 only watches the main process of a service. By setting
4359 ExitType=cgroup it can be told to wait for the last process in a
4362 * Automount unit files gained a new setting ExtraOptions= that can be
4363 used to configure additional mount options to pass to the kernel when
4364 mounting the autofs instance.
4366 * "Urlification" (generation of ESC sequences that generate clickable
4367 hyperlinks in modern terminals) may now be turned off altogether
4370 * Path units gained new TriggerLimitBurst= and TriggerLimitIntervalSec=
4371 settings that default to 200 and 2 s respectively. The ratelimit
4372 ensures that a path unit cannot cause PID1 to busy-loop when it is
4373 trying to trigger a service that is skipped because of a Condition*=
4374 not being satisfied. This matches the configuration and behaviour of
4377 * The TPM2/FIDO2/PKCS11 support in systemd-cryptsetup is now also built
4378 as a plug-in for cryptsetup. This means the plain cryptsetup command
4379 may now be used to unlock volumes set up this way.
4381 * The TPM2 logic in cryptsetup will now automatically detect systems
4382 where the TPM2 chip advertises SHA256 PCR banks but the firmware only
4383 updates the SHA1 banks. In such a case PCR policies will be
4384 automatically bound to the latter, not the former. This makes the PCR
4385 policies reliable, but of course do not provide the same level of
4386 trust as SHA256 banks.
4388 * The TPM2 logic in systemd-cryptsetup/systemd-cryptsetup now supports
4389 RSA primary keys in addition to ECC, improving compatibility with
4390 TPM2 chips that do not support ECC. RSA keys are much slower to use
4391 than ECC, and hence are only used if ECC is not available.
4393 * /etc/crypttab gained support for a new token-timeout= setting for
4394 encrypted volumes that allows configuration of the maximum time to
4395 wait for PKCS#11/FIDO2 tokens to be plugged in. If the time elapses
4396 the logic will query the user for a regular passphrase/recovery key
4399 * Support for activating dm-integrity volumes at boot via a new file
4400 /etc/integritytab and the tool systemd-integritysetup have been
4401 added. This is similar to /etc/crypttab and /etc/veritytab, but deals
4402 with dm-integrity instead of dm-crypt/dm-verity.
4404 * The systemd-veritysetup-generator now understands a new usrhash=
4405 kernel command line option for specifying the Verity root hash for
4406 the partition backing the /usr/ file system. A matching set of
4407 systemd.verity_usr_* kernel command line options has been added as
4408 well. These all work similar to the corresponding options for the
4411 * The sd-device API gained a new API call sd_device_get_diskseq() to
4412 return the DISKSEQ property of a device structure. The "disk
4413 sequence" concept is a new feature recently introduced to the Linux
4414 kernel that allows detecting reuse cycles of block devices, i.e. can
4415 be used to recognize when loopback block devices are reused for a
4416 different purpose or CD-ROM drives get their media changed.
4418 * A new unit systemd-boot-update.service has been added. If enabled
4419 (the default) and the sd-boot loader is detected to be installed, it
4420 is automatically updated to the newest version when out of date. This
4421 is useful to ensure the boot loader remains up-to-date, and updates
4422 automatically propagate from the OS tree in /usr/.
4424 * sd-boot will now build with SBAT by default in order to facilitate
4425 working with recent versions of Shim that require it to be present.
4427 * sd-boot can now parse Microsoft Windows' Boot Configuration Data.
4428 This is used to robustly generate boot entry titles for Windows.
4430 * A new generic target unit factory-reset.target has been added. It is
4431 hooked into systemd-logind similar in fashion to
4432 reboot/poweroff/suspend/hibernate, and is supposed to be used to
4433 initiate a factory reset operation. What precisely this operation
4434 entails is up for the implementer to decide, the primary goal of the
4435 new unit is provide a framework where to plug in the implementation
4436 and how to trigger it.
4438 * A new meson build-time option 'clock-valid-range-usec-max' has been
4439 added which takes a time in µs and defaults to 15 years. If the RTC
4440 time is noticed to be more than the specified time ahead of the
4441 built-in epoch of systemd (which by default is the release timestamp
4442 of systemd) it is assumed that the RTC is not working correctly, and
4443 the RTC is reset to the epoch. (It already is reset to the epoch when
4444 noticed to be before it.) This should increase the chance that time
4445 doesn't accidentally jump too far ahead due to faulty hardware or
4448 * A new setting SaveIntervalSec= has been added to systemd-timesyncd,
4449 which may be used to automatically save the current system time to
4450 disk in regular intervals. This is useful to maintain a roughly
4451 monotonic clock even without RTC hardware and with some robustness
4452 against abnormal system shutdown.
4454 * systemd-analyze verify gained support for a pair of new --image= +
4455 --root= switches for verifying units below a specific root
4456 directory/image instead of on the host.
4458 * systemd-analyze verify gained support for verifying unit files under
4459 an explicitly specified unit name, independently of what the filename
4462 * systemd-analyze verify gained a new switch --recursive-errors= which
4463 controls whether to only fail on errors found in the specified units
4464 or recursively any dependent units.
4466 * systemd-analyze security now supports a new --offline mode for
4467 analyzing unit files stored on disk instead of loaded units. It may
4468 be combined with --root=/--image to analyze unit files under a root
4469 directory or disk image. It also learnt a new --threshold= parameter
4470 for specifying an exposure level threshold: if the exposure level
4471 exceeds the specified value the call will fail. It also gained a new
4472 --security-policy= switch for configuring security policies to
4473 enforce on the units. A policy is a JSON file that lists which tests
4474 shall be weighted how much to determine the overall exposure
4475 level. Altogether these new features are useful for fully automatic
4476 analysis and enforcement of security policies on unit files.
4478 * systemd-analyze security gain a new --json= switch for JSON output.
4480 * systemd-analyze learnt a new --quiet switch for reducing
4481 non-essential output. It's honored by the "dot", "syscall-filter",
4482 "filesystems" commands.
4484 * systemd-analyze security gained a --profile= option that can be used
4485 to take into account a portable profile when analyzing portable
4486 services, since a lot of the security-related settings are enabled
4489 * systemd-analyze learnt a new inspect-elf verb that parses ELF core
4490 files, binaries and executables and prints metadata information,
4491 including the build-id and other info described on:
4492 https://systemd.io/COREDUMP_PACKAGE_METADATA/
4494 * .network files gained a new UplinkInterface= in the [IPv6SendRA]
4495 section, for automatically propagating DNS settings from other
4498 * The static lease DHCP server logic in systemd-networkd may now serve
4499 IP addresses outside of the configured IP pool range for the server.
4501 * CAN support in systemd-networkd gained four new settings Loopback=,
4502 OneShot=, PresumeAck=, ClassicDataLengthCode= for tweaking CAN
4503 control modes. It gained a number of further settings for tweaking
4506 * The [CAN] section in .network file gained new TimeQuantaNSec=,
4507 PropagationSegment=, PhaseBufferSegment1=, PhaseBufferSegment2=,
4508 SyncJumpWidth=, DataTimeQuantaNSec=, DataPropagationSegment=,
4509 DataPhaseBufferSegment1=, DataPhaseBufferSegment2=, and
4510 DataSyncJumpWidth= settings to control bit-timing processed by the
4513 * DHCPv4 client support in systemd-networkd learnt a new Label= option
4514 for configuring the address label to apply to configure IPv4
4517 * The [IPv6AcceptRA] section of .network files gained support for a new
4518 UseMTU= setting that may be used to control whether to apply the
4519 announced MTU settings to the local interface.
4521 * The [DHCPv4] section in .network file gained a new Use6RD= boolean
4522 setting to control whether the DHCPv4 client request and process the
4525 * The [DHCPv6PrefixDelegation] section in .network file is renamed to
4526 [DHCPPrefixDelegation], as now the prefix delegation is also supported
4527 with DHCPv4 protocol by enabling the Use6RD= setting.
4529 * The [DHCPPrefixDelegation] section in .network file gained a new
4530 setting UplinkInterface= to specify the upstream interface.
4532 * The [DHCPv6] section in .network file gained a new setting
4533 UseDelegatedPrefix= to control whether the delegated prefixes will be
4534 propagated to the downstream interfaces.
4536 * The [IPv6AcceptRA] section of .network files now understands two new
4537 settings UseGateway=/UseRoutePrefix= for explicitly configuring
4538 whether to use the relevant fields from the IPv6 Router Advertisement
4541 * The ForceDHCPv6PDOtherInformation= setting in the [DHCPv6] section
4542 has been removed. Please use the WithoutRA= and UseDelegatedPrefix=
4543 settings in the [DHCPv6] section and the DHCPv6Client= setting in the
4544 [IPv6AcceptRA] section to control when the DHCPv6 client is started
4545 and how the delegated prefixes are handled by the DHCPv6 client.
4547 * The IPv6Token= section in the [Network] section is deprecated, and
4548 the [IPv6AcceptRA] section gained the Token= setting for its
4549 replacement. The [IPv6Prefix] section also gained the Token= setting.
4550 The Token= setting gained 'eui64' mode to explicitly configure an
4551 address with the EUI64 algorithm based on the interface MAC address.
4552 The 'prefixstable' mode can now optionally take a secret key. The
4553 Token= setting in the [DHCPPrefixDelegation] section now supports all
4554 algorithms supported by the same settings in the other sections.
4556 * The [RoutingPolicyRule] section of .network file gained a new
4557 SuppressInterfaceGroup= setting.
4559 * The IgnoreCarrierLoss= setting in the [Network] section of .network
4560 files now allows a duration to be specified, controlling how long to
4561 wait before reacting to carrier loss.
4563 * The [DHCPServer] section of .network file gained a new Router=
4564 setting to specify the router address.
4566 * The [CAKE] section of .network files gained various new settings
4567 AutoRateIngress=, CompensationMode=, FlowIsolationMode=, NAT=,
4568 MPUBytes=, PriorityQueueingPreset=, FirewallMark=, Wash=, SplitGSO=,
4569 and UseRawPacketSize= for configuring CAKE.
4571 * systemd-networkd now ships with new default .network files:
4572 80-container-vb.network which matches host-side network bridge device
4573 created by systemd-nspawn's --network-bridge or --network-zone
4574 switch, and 80-6rd-tunnel.network which matches automatically created
4575 sit tunnel with 6rd prefix when the DHCP 6RD option is received.
4577 * systemd-networkd's handling of Endpoint= resolution for WireGuard
4578 interfaces has been improved.
4580 * systemd-networkd will now automatically configure routes to addresses
4581 specified in AllowedIPs=. This feature can be controlled via
4582 RouteTable= and RouteMetric= settings in [WireGuard] or
4583 [WireGuardPeer] sections.
4585 * systemd-networkd will now once again automatically generate persistent
4586 MAC addresses for batadv and bridge interfaces. Users can disable this
4587 by using MACAddress=none in .netdev files.
4589 * systemd-networkd and systemd-udevd now support IP over InfiniBand
4590 interfaces. The Kind= setting in .netdev file accepts "ipoib". And
4591 systemd.netdev files gained the [IPoIB] section.
4593 * systemd-networkd and systemd-udevd now support net.ifname_policy=
4594 option on the kernel command-line. This is implemented through the
4595 systemd-network-generator service that automatically generates
4596 appropriate .link, .network, and .netdev files.
4598 * The various systemd-udevd "ethtool" buffer settings now understand
4599 the special value "max" to configure the buffers to the maximum the
4602 * systemd-udevd's .link files may now configure a large variety of
4603 NIC coalescing settings, plus more hardware offload settings.
4605 * .link files gained a new WakeOnLanPassword= setting in the [Link]
4606 section that allows to specify a WoL "SecureOn" password on hardware
4609 * systemd-nspawn's --setenv= switch now supports an additional syntax:
4610 if only a variable name is specified (i.e. without being suffixed by
4611 a '=' character and a value) the current value of the environment
4612 variable is propagated to the container. e.g. --setenv=FOO will
4613 lookup the current value of $FOO in the environment, and pass it down
4614 to the container. Similar behavior has been added to homectl's,
4615 machinectl's and systemd-run's --setenv= switch.
4617 * systemd-nspawn gained a new switch --suppress-sync= which may be used
4618 to optionally suppress the effect of the sync()/fsync()/fdatasync()
4619 system calls for the container payload. This is useful for build
4620 system environments where safety against abnormal system shutdown is
4621 not essential as all build artifacts can be regenerated any time, but
4622 the performance win is beneficial.
4624 * systemd-nspawn will now raise the RLIMIT_NOFILE hard limit to the
4625 same value that PID 1 uses for most forked off processes.
4627 * systemd-nspawn's --bind=/--bind-ro= switches now optionally take
4628 uidmap/nouidmap options as last parameter. If "uidmap" is used the
4629 bind mounts are created with UID mapping taking place that ensures
4630 the host's file ownerships are mapped 1:1 to container file
4631 ownerships, even if user namespacing is used. This way
4632 files/directories bound into containers will no longer show up as
4633 owned by the nobody user as they typically did if no special care was
4634 taken to shift them manually.
4636 * When discovering Windows installations sd-boot will now attempt to
4637 show the Windows version.
4639 * The color scheme to use in sd-boot may now be configured at
4642 * sd-boot gained the ability to change screen resolution during
4643 boot-time, by hitting the "r" key. This will cycle through available
4644 resolutions and save the last selection.
4646 * sd-boot learnt a new hotkey "f". When pressed the system will enter
4647 firmware setup. This is useful in environments where it is difficult
4648 to hit the right keys early enough to enter the firmware, and works
4649 on any firmware regardless which key it natively uses.
4651 * sd-boot gained support for automatically booting into the menu item
4652 selected on the last boot (using the "@saved" identifier for menu
4655 * sd-boot gained support for automatically loading all EFI drivers
4656 placed in the /EFI/systemd/drivers/ subdirectory of the EFI System
4657 Partition (ESP). These drivers are loaded before the menu entries are
4658 loaded. This is useful e.g. to load additional file system drivers
4659 for the XBOOTLDR partition.
4661 * systemd-boot will now paint the input cursor on its own instead of
4662 relying on the firmware to do so, increasing compatibility with broken
4663 firmware that doesn't make the cursor reasonably visible.
4665 * sd-boot now embeds a .osrel PE section like we expect from Boot
4666 Loader Specification Type #2 Unified Kernels. This means sd-boot
4667 itself may be used in place of a Type #2 Unified Kernel. This is
4668 useful for debugging purposes as it allows chain-loading one a
4669 (development) sd-boot instance from another.
4671 * sd-boot now supports a new "devicetree" field in Boot Loader
4672 Specification Type #1 entries: if configured the specified device
4673 tree file is installed before the kernel is invoked. This is useful
4674 for installing/applying new devicetree files without updating the
4677 * Similarly, sd-stub now can read devicetree data from a PE section
4678 ".dtb" and apply it before invoking the kernel.
4680 * sd-stub (the EFI stub that can be glued in front of a Linux kernel)
4681 gained the ability to pick up credentials and sysext files, wrap them
4682 in a cpio archive, and pass as an additional initrd to the invoked
4683 Linux kernel, in effect placing those files in the /.extra/ directory
4684 of the initrd environment. This is useful to implement trusted initrd
4685 environments which are fully authenticated but still can be extended
4686 (via sysexts) and parameterized (via encrypted/authenticated
4687 credentials, see above).
4689 Credentials can be located next to the kernel image file (credentials
4690 specific to a single boot entry), or in one of the shared directories
4691 (credentials applicable to multiple boot entries).
4693 * sd-stub now comes with a full man page, that explains its feature set
4694 and how to combine a kernel image, an initrd and the stub to build a
4695 complete EFI unified kernel image, implementing Boot Loader
4696 Specification Type #2.
4698 * sd-stub may now provide the initrd to the executed kernel via the
4699 LINUX_EFI_INITRD_MEDIA_GUID EFI protocol, adding compatibility for
4700 non-x86 architectures.
4702 * bootctl learnt new set-timeout and set-timeout-oneshot commands that
4703 may be used to set the boot menu time-out of the boot loader (for all
4704 or just the subsequent boot).
4706 * bootctl and kernel-install will now read variables
4707 KERNEL_INSTALL_LAYOUT= from /etc/machine-info and layout= from
4708 /etc/kernel/install.conf. When set, it specifies the layout to use
4709 for installation directories on the boot partition, so that tools
4710 don't need to guess it based on the already-existing directories. The
4711 only value that is defined natively is "bls", corresponding to the
4713 https://systemd.io/BOOT_LOADER_SPECIFICATION/. Plugins for
4714 kernel-install that implement a different layout can declare other
4715 values for this variable.
4717 'bootctl install' will now write KERNEL_INSTALL_LAYOUT=bls, on the
4718 assumption that if the user installed sd-boot to the ESP, they intend
4719 to use the entry layout understood by sd-boot. It'll also write
4720 KERNEL_INSTALL_MACHINE_ID= if it creates any directories using the ID
4721 (and it wasn't specified in the config file yet). Similarly,
4722 kernel-install will now write KERNEL_INSTALL_MACHINE_ID= (if it
4723 wasn't specified in the config file yet). Effectively, those changes
4724 mean that the machine-id used for boot loader entry installation is
4725 "frozen" upon first use and becomes independent of the actual
4728 Configuring KERNEL_INSTALL_MACHINE_ID fixes the following problem:
4729 images created for distribution ("golden images") are built with no
4730 machine-id, so that a unique machine-id can be created on the first
4731 boot. But those images may contain boot loader entries with the
4732 machine-id used during build included in paths. Using a "frozen"
4733 value allows unambiguously identifying entries that match the
4734 specific installation, while still permitting parallel installations
4737 Configuring KERNEL_INSTALL_LAYOUT obviates the need for
4738 kernel-install to guess the installation layout. This fixes the
4739 problem where a (possibly empty) directory in the boot partition is
4740 created from a different layout causing kernel-install plugins to
4741 assume the wrong layout. A particular example of how this may happen
4742 is the grub2 package in Fedora which includes directories under /boot
4743 directly in its file list. Various other packages pull in grub2 as a
4744 dependency, so it may be installed even if unused, breaking
4745 installations that use the bls layout.
4747 * bootctl and systemd-bless-boot can now be linked statically.
4749 * systemd-sysext now optionally doesn't insist on extension-release.d/
4750 files being placed in the image under the image's file name. If the
4751 file system xattr user.extension-release.strict is set on the
4752 extension release file, it is accepted regardless of its name. This
4753 relaxes security restrictions a bit, as system extension may be
4754 attached under a wrong name this way.
4756 * udevadm's test-builtin command learnt a new --action= switch for
4757 testing the built-in with the specified action (in place of the
4760 * udevadm info gained new switches --property=/--value for showing only
4761 specific udev properties/values instead of all.
4763 * A new hwdb database has been added that contains matches for various
4764 types of signal analyzers (protocol analyzers, logic analyzers,
4765 oscilloscopes, multimeters, bench power supplies, etc.) that should
4766 be accessible to regular users.
4768 * A new hwdb database entry has been added that carries information
4769 about types of cameras (regular or infrared), and in which direction
4770 they point (front or back).
4772 * A new rule to allow console users access to rfkill by default has been
4775 * Device nodes for the Software Guard eXtension enclaves (sgx_vepc) are
4776 now also owned by the system group "sgx".
4778 * A new build-time meson option "extra-net-naming-schemes=" has been
4779 added to define additional naming schemes for udev's network
4780 interface naming logic. This is useful for enterprise distributions
4781 and similar which want to pin the schemes of certain distribution
4782 releases under a specific name and previously had to patch the
4783 sources to introduce new named schemes.
4785 * The predictable naming logic for network interfaces has been extended
4786 to generate stable names from Xen netfront device information.
4788 * hostnamed's chassis property can now be sourced from chassis-type
4789 field encoded in devicetree (in addition to the existing DMI
4792 * systemd-cgls now optionally displays cgroup IDs and extended
4793 attributes for each cgroup. (Controllable via the new --xattr= +
4794 --cgroup-id= switches.)
4796 * coredumpctl gained a new --all switch for operating on all
4797 Journal files instead of just the local ones.
4799 * systemd-coredump will now use libdw/libelf via dlopen() rather than
4800 directly linking, allowing users to easily opt-out of backtrace/metadata
4801 analysis of core files, and reduce image sizes when this is not needed.
4803 * systemd-coredump will now analyze core files with libdw/libelf in a
4804 forked, sandboxed process.
4806 * systemd-homed will now try to unmount an activate home area in
4807 regular intervals once the user logged out fully. Previously this was
4808 attempted exactly once but if the home directory was busy for some
4809 reason it was not tried again.
4811 * systemd-homed's LUKS2 home area backend will now create a BSD file
4812 system lock on the image file while the home area is active
4813 (i.e. mounted). If a home area is found to be locked, logins are
4814 politely refused. This should improve behavior when using home areas
4815 images that are accessible via the network from multiple clients, and
4816 reduce the chance of accidental file system corruption in that case.
4818 * Optionally, systemd-homed will now drop the kernel buffer cache once
4819 a user has fully logged out, configurable via the new --drop-caches=
4822 * systemd-homed now makes use of UID mapped mounts for the home areas.
4823 If the kernel and used file system support it, files are now
4824 internally owned by the "nobody" user (i.e. the user typically used
4825 for indicating "this ownership is not mapped"), and dynamically
4826 mapped to the UID used locally on the system via the UID mapping
4827 mount logic of recent kernels. This makes migrating home areas
4828 between different systems cheaper because recursively chown()ing file
4829 system trees is no longer necessary.
4831 * systemd-homed's CIFS backend now optionally supports CIFS service
4832 names with a directory suffix, in order to place home directories in
4833 a subdirectory of a CIFS share, instead of the top-level directory.
4835 * systemd-homed's CIFS backend gained support for specifying additional
4836 mount options in the JSON user record (cifsExtraMountOptions field,
4837 and --cifs-extra-mount-options= homectl switch). This is for example
4838 useful for configuring mount options such as "noserverino" that some
4839 SMB3 services require (use that to run a homed home directory from a
4840 FritzBox SMB3 share this way).
4842 * systemd-homed will now default to btrfs' zstd compression for home
4843 areas. This is inspired by Fedora's recent decision to switch to zstd
4846 * Additional mount options to use when mounting the file system of
4847 LUKS2 volumes in systemd-homed has been added. Via the
4848 $SYSTEMD_HOME_MOUNT_OPTIONS_BTRFS, $SYSTEMD_HOME_MOUNT_OPTIONS_EXT4,
4849 $SYSTEMD_HOME_MOUNT_OPTIONS_XFS environment variables to
4850 systemd-homed or via the luksExtraMountOptions user record JSON
4851 property. (Exposed via homectl --luks-extra-mount-options)
4853 * homectl's resize command now takes the special size specifications
4854 "min" and "max" to shrink/grow the home area to the minimum/maximum
4855 size possible, taking disk usage/space constraints and file system
4856 limitations into account. Resizing is now generally graceful: the
4857 logic will try to get as close to the specified size as possible, but
4858 not consider it a failure if the request couldn't be fulfilled
4861 * systemd-homed gained the ability to automatically shrink home areas
4862 on logout to their minimal size and grow them again on next
4863 login. This ensures that while inactive, a home area only takes up
4864 the minimal space necessary, but once activated, it provides
4865 sufficient space for the user's needs. This behavior is only
4866 supported if btrfs is used as file system inside the home area
4867 (because only for btrfs online growing/shrinking is implemented in
4868 the kernel). This behavior is now enabled by default, but may be
4869 controlled via the new --auto-resize-mode= setting of homectl.
4871 * systemd-homed gained support for automatically re-balancing free disk
4872 space among active home areas, in case the LUKS2 backends are used,
4873 and no explicit disk size was requested. This way disk space is
4874 automatically managed and home areas resized in regular intervals and
4875 manual resizing when disk space becomes scarce should not be
4876 necessary anymore. This behavior is only supported if btrfs is used
4877 within the home areas (as only then online shrinking and growing is
4878 supported), and may be configured via the new rebalanceWeight JSON
4879 user record field (as exposed via the new --rebalance-weight= homectl
4880 setting). Re-balancing is mostly automatic, but can also be requested
4881 explicitly via "homectl rebalance", which is synchronous, and thus
4882 may be used to wait until the rebalance run is complete.
4884 * userdbctl gained a --json= switch for configured the JSON formatting
4885 to use when outputting user or group records.
4887 * userdbctl gained a new --multiplexer= switch for explicitly
4888 configuring whether to use the systemd-userdbd server side user
4889 record resolution logic.
4891 * userdbctl's ssh-authorized-keys command learnt a new --chain switch,
4892 for chaining up another command to execute after completing the
4893 look-up. Since the OpenSSH's AuthorizedKeysCommand only allows
4894 configuration of a single command to invoke, this maybe used to
4895 invoke multiple: first userdbctl's own implementation, and then any
4896 other also configured in the command line.
4898 * The sd-event API gained a new function sd_event_add_inotify_fd() that
4899 is similar to sd_event_add_inotify() but accepts a file descriptor
4900 instead of a path in the file system for referencing the inode to
4903 * The sd-event API gained a new function
4904 sd_event_source_set_ratelimit_expire_callback() that may be used to
4905 define a callback function that is called whenever an event source
4906 leaves the rate limiting phase.
4908 * New documentation has been added explaining which steps are necessary
4909 to port systemd to a new architecture:
4911 https://systemd.io/PORTING_TO_NEW_ARCHITECTURES
4913 * The x-systemd.makefs option in /etc/fstab now explicitly supports
4914 ext2, ext3, and f2fs file systems.
4916 * Mount units and units generated from /etc/fstab entries with 'noauto'
4917 are now ordered the same as other units. Effectively, they will be
4918 started earlier (if something actually pulled them in) and stopped
4919 later, similarly to normal mount units that are part of
4920 fs-local.target. This change should be invisible to users, but
4921 should prevent those units from being stopped too early during
4924 * The systemd-getty-generator now honors a new kernel command line
4925 argument systemd.getty_auto= and a new environment variable
4926 $SYSTEMD_GETTY_AUTO that allows turning it off at boot. This is for
4927 example useful to turn off gettys inside of containers or similar
4930 * systemd-resolved now listens on a second DNS stub address: 127.0.0.54
4931 (in addition to 127.0.0.53, as before). If DNS requests are sent to
4932 this address they are propagated in "bypass" mode only, i.e. are
4933 almost not processed locally, but mostly forwarded as-is to the
4934 current upstream DNS servers. This provides a stable DNS server
4935 address that proxies all requests dynamically to the right upstream
4936 DNS servers even if these dynamically change. This stub does not do
4937 mDNS/LLMNR resolution. However, it will translate look-ups to
4938 DNS-over-TLS if necessary. This new stub is particularly useful in
4939 container/VM environments, or for tethering setups: use DNAT to
4940 redirect traffic to any IP address to this stub.
4942 * systemd-importd now honors new environment variables
4943 $SYSTEMD_IMPORT_BTRFS_SUBVOL, $SYSTEMD_IMPORT_BTRFS_QUOTA,
4944 $SYSTEMD_IMPORT_SYNC, which may be used disable btrfs subvolume
4945 generation, btrfs quota setup and disk synchronization.
4947 * systemd-importd and systemd-resolved can now be optionally built with
4948 OpenSSL instead of libgcrypt.
4950 * systemd-repart no longer requires OpenSSL.
4952 * systemd-sysusers will no longer create the redundant 'nobody' group
4953 by default, as the 'nobody' user is already created with an
4954 appropriate primary group.
4956 * If a unit uses RuntimeMaxSec, systemctl show will now display it.
4958 * systemctl show-environment gained support for --output=json.
4960 * pam_systemd will now first try to use the X11 abstract socket, and
4961 fallback to the socket file in /tmp/.X11-unix/ only if that does not
4964 * systemd-journald will no longer go back to volatile storage
4965 regardless of configuration when its unit is restarted.
4967 * Initial support for the LoongArch architecture has been added (system
4968 call lists, GPT partition table UUIDs, etc).
4970 * systemd-journald's own logging messages are now also logged to the
4971 journal itself when systemd-journald logs to /dev/kmsg.
4973 * systemd-journald now re-enables COW for archived journal files on
4974 filesystems that support COW. One benefit of this change is that
4975 archived journal files will now get compressed on btrfs filesystems
4976 that have compression enabled.
4978 * systemd-journald now deduplicates fields in a single log message
4979 before adding it to the journal. In archived journal files, it will
4980 also punch holes for unused parts and truncate the file as
4981 appropriate, leading to reductions in disk usage.
4983 * journalctl --verify was extended with more informative error
4986 * More of sd-journal's functions are now resistant against journal file
4989 * The shutdown command learnt a new option --show, to display the
4992 * A LICENSES/ directory is now included in the git tree. It contains a
4993 README.md file that explains the licenses used by source files in
4994 this repository. It also contains the text of all applicable
4995 licenses as they appear on spdx.org.
4997 Contributions from: Aakash Singh, acsfer, Adolfo Jayme Barrientos,
4998 Adrian Vovk, Albert Brox, Alberto Mardegan, Alexander Kanavin,
4999 alexlzhu, Alfonso Sánchez-Beato, Alvin Šipraga, Alyssa Ross,
5000 Amir Omidi, Anatol Pomozov, Andika Triwidada, Andreas Rammhold,
5001 Andreas Valder, Andrej Lajovic, Andrew Soutar, Andrew Stone, Andy Chi,
5002 Anita Zhang, Anssi Hannula, Antonio Alvarez Feijoo,
5003 Antony Deepak Thomas, Arnaud Ferraris, Arvid E. Picciani,
5004 Bastien Nocera, Benjamin Berg, Benjamin Herrenschmidt, Ben Stockett,
5005 Bogdan Seniuc, Boqun Feng, Carl Lei, chlorophyll-zz, Chris Packham,
5006 Christian Brauner, Christian Göttsche, Christian Wehrli,
5007 Christoph Anton Mitterer, Cristian Rodríguez, Daan De Meyer,
5008 Daniel Maixner, Dann Frazier, Dan Streetman, Davide Cavalca,
5009 David Seifert, David Tardon, dependabot[bot], Dimitri John Ledkov,
5010 Dimitri Papadopoulos, Dimitry Ishenko, Dmitry Khlebnikov,
5011 Dominique Martinet, duament, Egor, Egor Ignatov, Emil Renner Berthing,
5012 Emily Gonyer, Ettore Atalan, Evgeny Vereshchagin, Florian Klink,
5013 Franck Bui, Frantisek Sumsal, Geass-LL, Gibeom Gwon, GnunuX,
5014 Gogo Gogsi, gregzuro, Greg Zuro, Gustavo Costa, Hans de Goede,
5015 Hela Basa, Henri Chain, hikigaya58, Hugo Carvalho,
5016 Hugo Osvaldo Barrera, Iago Lopez Galeiras, Iago López Galeiras,
5017 I-dont-need-name, igo95862, Jack Dähn, James Hilliard, Jan Janssen,
5018 Jan Kuparinen, Jan Macku, Jan Palus, Jarkko Sakkinen, Jayce Fayne,
5019 jiangchuangang, jlempen, John Lindgren, Jonas Dreßler, Jonas Jelten,
5020 Jonas Witschel, Joris Hartog, José Expósito, Julia Kartseva,
5021 Kai-Heng Feng, Kai Wohlfahrt, Kay Siver Bø, KennthStailey,
5022 Kevin Kuehler, Kevin Orr, Khem Raj, Kristian Klausen, Kyle Laker,
5023 lainahai, LaserEyess, Lennart Poettering, Lia Lenckowski, longpanda,
5024 Luca Boccassi, Luca BRUNO, Ludwig Nussel, Lukas Senionis,
5025 Maanya Goenka, Maciek Borzecki, Marcel Menzel, Marco Scardovi,
5026 Marcus Harrison, Mark Boudreau, Matthijs van Duin, Mauricio Vásquez,
5027 Maxime de Roucy, Max Resch, MertsA, Michael Biebl, Michael Catanzaro,
5028 Michal Koutný, Michal Sekletár, Miika Karanki, Mike Gilbert,
5029 Milo Turner, ml, monosans, Nacho Barrientos, nassir90, Nishal Kulkarni,
5030 nl6720, Ondrej Kozina, Paulo Neves, Pavel Březina, pedro martelletto,
5031 Peter Hutterer, Peter Morrow, Piotr Drąg, Rasmus Villemoes, ratijas,
5032 Raul Tambre, rene, Riccardo Schirone, Robert-L-Turner, Robert Scheck,
5033 Ross Jennings, saikat0511, Scott Lamb, Scott Worley,
5034 Sergei Trofimovich, Sho Iizuka, Slava Bacherikov, Slimane Selyan Amiri,
5035 StefanBruens, Steven Siloti, svonohr, Taiki Sugawara, Takashi Sakamoto,
5036 Takuro Onoue, Thomas Blume, Thomas Haller, Thomas Mühlbacher,
5037 Tianlu Shao, Toke Høiland-Jørgensen, Tom Yan, Tony Asleson,
5038 Topi Miettinen, Ulrich Ölmann, Urs Ritzmann, Vincent Bernat,
5039 Vito Caputo, Vladimir Panteleev, WANG Xuerui, Wind/owZ, Wu Xiaotian,
5040 xdavidwu, Xiaotian Wu, xujing, yangmingtai, Yao Wei, Yao Wei (魏銘廷),
5041 Yegor Alexeyev, Yu Watanabe, Zbigniew Jędrzejewski-Szmek,
5042 Дамјан Георгиевски, наб
5044 — Warsaw, 2021-12-23
5048 * When operating on disk images via the --image= switch of various
5049 tools (such as systemd-nspawn or systemd-dissect), or when udev finds
5050 no 'root=' parameter on the kernel command line, and multiple
5051 suitable root or /usr/ partitions exist in the image, then a simple
5052 comparison inspired by strverscmp() is done on the GPT partition
5053 label, and the newest partition is picked. This permits a simple and
5054 generic whole-file-system A/B update logic where new operating system
5055 versions are dropped into partitions whose label is then updated with
5056 a matching version identifier.
5058 * systemd-sysusers now supports querying the passwords to set for the
5059 users it creates via the "credentials" logic introduced in v247: the
5060 passwd.hashed-password.<user> and passwd.plaintext-password.<user>
5061 credentials are consulted for the password to use (either in UNIX
5062 hashed form, or literally). By default these credentials are inherited
5063 down from PID1 (which in turn imports it from a container manager if
5064 there is one). This permits easy configuration of user passwords
5065 during first boot. Example:
5067 # systemd-nspawn -i foo.raw --volatile=yes --set-credential=passwd.plaintext-password.root:foo
5069 Note that systemd-sysusers operates in purely additive mode: it
5070 executes no operation if the declared users already exist, and hence
5071 doesn't set any passwords as effect of the command line above if the
5072 specified root user exists already in the image. (Note that
5073 --volatile=yes ensures it doesn't, though.)
5075 * systemd-firstboot now also supports querying various system
5076 parameters via the credential subsystems. Thus, as above this may be
5077 used to initialize important system parameters on first boot of
5078 previously unprovisioned images (i.e. images with a mostly empty
5081 * PID 1 may now show both the unit name and the unit description
5082 strings in its status output during boot. This may be configured with
5083 StatusUnitFormat=combined in system.conf or
5084 systemd.status-unit-format=combined on the kernel command line.
5086 * The systemd-machine-id-setup tool now supports a --image= switch for
5087 provisioning a machine ID file into an OS disk image, similar to how
5088 --root= operates on an OS file tree. This matches the existing switch
5089 of the same name for systemd-tmpfiles, systemd-firstboot, and
5090 systemd-sysusers tools.
5092 * Similarly, systemd-repart gained support for the --image= switch too.
5093 In combination with the existing --size= option, this makes the tool
5094 particularly useful for easily growing disk images in a single
5095 invocation, following the declarative rules included in the image
5098 * systemd-repart's partition configuration files gained support for a
5099 new switch MakeDirectories= which may be used to create arbitrary
5100 directories inside file systems that are created, before registering
5101 them in the partition table. This is useful in particular for root
5102 partitions to create mount point directories for other partitions
5103 included in the image. For example, a disk image that contains a
5104 root, /home/, and /var/ partitions, may set MakeDirectories=yes to
5105 create /home/ and /var/ as empty directories in the root file system
5106 on its creation, so that the resulting image can be mounted
5107 immediately, even in read-only mode.
5109 * systemd-repart's CopyBlocks= setting gained support for the special
5110 value "auto". If used, a suitable matching partition on the booted OS
5111 is found as source to copy blocks from. This is useful when
5112 implementing replicating installers, that are booted from one medium
5113 and then stream their own root partition onto the target medium.
5115 * systemd-repart's partition configuration files gained support for a
5116 Flags=, a ReadOnly= and a NoAuto= setting, allowing control of these
5117 GPT partition flags for the created partitions: this is useful for
5118 marking newly created partitions as read-only, or as not being
5119 subject for automatic mounting from creation on.
5121 * The /etc/os-release file has been extended with two new (optional)
5122 variables IMAGE_VERSION= and IMAGE_ID=, carrying identity and version
5123 information for OS images that are updated comprehensively and
5124 atomically as one image. Two new specifiers %M, %A now resolve to
5125 these two fields in the various configuration options that resolve
5128 * portablectl gained a new switch --extension= for enabling portable
5129 service images with extensions that follow the extension image
5130 concept introduced with v248, and thus allows layering multiple
5131 images when setting up the root filesystem of the service.
5133 * systemd-coredump will now extract ELF build-id information from
5134 processes dumping core and include it in the coredump report.
5135 Moreover, it will look for ELF .note.package sections with
5136 distribution packaging meta-information about the crashing process.
5137 This is useful to directly embed the rpm or deb (or any other)
5138 package name and version in ELF files, making it easy to match
5139 coredump reports with the specific package for which the software was
5140 compiled. This is particularly useful on environments with ELF files
5141 from multiple vendors, different distributions and versions, as is
5142 common today in our containerized and sand-boxed world. For further
5145 https://systemd.io/COREDUMP_PACKAGE_METADATA
5147 * A new udev hardware database has been added for FireWire devices
5150 * The "net_id" built-in of udev has been updated with three
5151 backwards-incompatible changes:
5153 - PCI hotplug slot names on s390 systems are now parsed as
5154 hexadecimal numbers. They were incorrectly parsed as decimal
5155 previously, or ignored if the name was not a valid decimal
5158 - PCI onboard indices up to 65535 are allowed. Previously, numbers
5159 above 16383 were rejected. This primarily impacts s390 systems,
5160 where values up to 65535 are used.
5162 - Invalid characters in interface names are replaced with "_".
5164 The new version of the net naming scheme is "v249". The previous
5165 scheme can be selected via the "net.naming_scheme=v247" kernel
5166 command line parameter.
5168 * sd-bus' sd_bus_is_ready() and sd_bus_is_open() calls now accept a
5169 NULL bus object, for which they will return false. Or in other words,
5170 an unallocated bus connection is neither ready nor open.
5172 * The sd-device API acquired a new API function
5173 sd_device_get_usec_initialized() that returns the monotonic time when
5174 the udev device first appeared in the database.
5176 * sd-device gained a new APIs sd_device_trigger_with_uuid() and
5177 sd_device_get_trigger_uuid(). The former is similar to
5178 sd_device_trigger() but returns a randomly generated UUID that is
5179 associated with the synthetic uevent generated by the call. This UUID
5180 may be read from the sd_device object a monitor eventually receives,
5181 via the sd_device_get_trigger_uuid(). This interface requires kernel
5182 4.13 or above to work, and allows tracking a synthetic uevent through
5183 the entire device management stack. The "udevadm trigger --settle"
5184 logic has been updated to make use of this concept if available to
5185 wait precisely for the uevents it generates. "udevadm trigger" also
5186 gained a new parameter --uuid that prints the UUID for each generated
5189 * sd-device also gained new APIs sd_device_new_from_ifname() and
5190 sd_device_new_from_ifindex() for allocating an sd-device object for
5191 the specified network interface. The former accepts an interface name
5192 (either a primary or an alternative name), the latter an interface
5195 * The native Journal protocol has been documented. Clients may talk
5196 this as alternative to the classic BSD syslog protocol for locally
5197 delivering log records to the Journal. The protocol has been stable
5198 for a long time and in fact been implemented already in a variety
5199 of alternative client libraries. This documentation makes the support
5202 https://systemd.io/JOURNAL_NATIVE_PROTOCOL
5204 * A new BPFProgram= setting has been added to service files. It may be
5205 set to a path to a loaded kernel BPF program, i.e. a path to a bpffs
5206 file, or a bind mount or symlink to one. This may be used to upload
5207 and manage BPF programs externally and then hook arbitrary systemd
5210 * The "home.arpa" domain that has been officially declared as the
5211 choice for domain for local home networks per RFC 8375 has been added
5212 to the default NTA list of resolved, since DNSSEC is generally not
5213 available on private domains.
5215 * The CPUAffinity= setting of unit files now resolves "%" specifiers.
5217 * A new ManageForeignRoutingPolicyRules= setting has been added to
5218 .network files which may be used to exclude foreign-created routing
5219 policy rules from systemd-networkd management.
5221 * systemd-network-wait-online gained two new switches -4 and -6 that
5222 may be used to tweak whether to wait for only IPv4 or only IPv6
5225 * .network files gained a new RequiredFamilyForOnline= setting to
5226 fine-tune whether to require an IPv4 or IPv6 address in order to
5227 consider an interface "online".
5229 * networkctl will now show an over-all "online" state in the per-link
5232 * In .network files a new OutgoingInterface= setting has been added to
5233 specify the output interface in bridge FDB setups.
5235 * In .network files the Multipath group ID may now be configured for
5236 [NextHop] entries, via the new Group= setting.
5238 * The DHCP server logic configured in .network files gained a new
5239 setting RelayTarget= that turns the server into a DHCP server relay.
5240 The RelayAgentCircuitId= and RelayAgentRemoteId= settings may be used
5241 to further tweak the DHCP relay behaviour.
5243 * The DHCP server logic also gained a new ServerAddress= setting in
5244 .network files that explicitly specifies the server IP address to
5245 use. If not specified, the address is determined automatically, as
5248 * The DHCP server logic in systemd-networkd gained support for static
5249 DHCP leases, configurable via the [DHCPServerStaticLease]
5250 section. This allows explicitly mapping specific MAC addresses to
5251 fixed IP addresses and vice versa.
5253 * The RestrictAddressFamilies= setting in service files now supports a
5254 new special value "none". If specified sockets of all address
5255 families will be made unavailable to services configured that way.
5257 * systemd-fstab-generator and systemd-repart have been updated to
5258 support booting from disks that carry only a /usr/ partition but no
5259 root partition yet, and where systemd-repart can add it in on the
5260 first boot. This is useful for implementing systems that ship with a
5261 single /usr/ file system, and whose root file system shall be set up
5262 and formatted on a LUKS-encrypted volume whose key is generated
5263 locally (and possibly enrolled in the TPM) during the first boot.
5265 * The [Address] section of .network files now accepts a new
5266 RouteMetric= setting that configures the routing metric to use for
5267 the prefix route created as effect of the address configuration.
5268 Similarly, the [DHCPv6PrefixDelegation] and [IPv6Prefix] sections
5269 gained matching settings for their prefix routes. (The option of the
5270 same name in the [DHCPv6] section is moved to [IPv6AcceptRA], since
5271 it conceptually belongs there; the old option is still understood for
5274 * The DHCPv6 IAID and DUID are now explicitly configurable in .network
5277 * A new udev property ID_NET_DHCP_BROADCAST on network interface
5278 devices is now honoured by systemd-networkd, controlling whether to
5279 issue DHCP offers via broadcasting. This is used to ensure that s390
5280 layer 3 network interfaces work out-of-the-box with systemd-networkd.
5282 * nss-myhostname and systemd-resolved will now synthesize address
5283 records for a new special hostname "_outbound". The name will always
5284 resolve to the local IP addresses most likely used for outbound
5285 connections towards the default routes. On multi-homed hosts this is
5286 useful to have a stable handle referring to "the" local IP address
5287 that matters most, to the point where this is defined.
5289 * The Discoverable Partition Specification has been updated with a new
5290 GPT partition flag "grow-file-system" defined for its partition
5291 types. Whenever partitions with this flag set are automatically
5292 mounted (i.e. via systemd-gpt-auto-generator or the --image= switch
5293 of systemd-nspawn or other tools; and as opposed to explicit mounting
5294 via /etc/fstab), the file system within the partition is
5295 automatically grown to the full size of the partition. If the file
5296 system size already matches the partition size this flag has no
5297 effect. Previously, this functionality has been available via the
5298 explicit x-systemd.growfs mount option, and this new flag extends
5299 this to automatically discovered mounts. A new GrowFileSystem=
5300 setting has been added to systemd-repart drop-in files that allows
5301 configuring this partition flag. This new flag defaults to on for
5302 partitions automatically created by systemd-repart, except if they
5303 are marked read-only. See the specification for further details:
5305 https://systemd.io/DISCOVERABLE_PARTITIONS
5307 * .network files gained a new setting RoutesToNTP= in the [DHCPv4]
5308 section. If enabled (which is the default), and an NTP server address
5309 is acquired through a DHCP lease on this interface an explicit route
5310 to this address is created on this interface to ensure that NTP
5311 traffic to the NTP server acquired on an interface is also routed
5312 through that interface. The pre-existing RoutesToDNS= setting that
5313 implements the same for DNS servers is now enabled by default.
5315 * A pair of service settings SocketBindAllow= + SocketBindDeny= have
5316 been added that may be used to restrict the network interfaces
5317 sockets created by the service may be bound to. This is implemented
5320 * A new ConditionFirmware= setting has been added to unit files to
5321 conditionalize on certain firmware features. At the moment it may
5322 check whether running on a UEFI system, a device.tree system, or if
5323 the system is compatible with some specified device-tree feature.
5325 * A new ConditionOSRelease= setting has been added to unit files to
5326 check os-release(5) fields. The "=", "!=", "<", "<=", ">=", ">"
5327 operators may be used to check if some field has some specific value
5328 or do an alphanumerical comparison. Equality comparisons are useful
5329 for fields like ID, but relative comparisons for fields like
5330 VERSION_ID or IMAGE_VERSION.
5332 * hostnamed gained a new Describe() D-Bus method that returns a JSON
5333 serialization of the host data it exposes. This is exposed via
5334 "hostnamectl --json=" to acquire a host identity description in JSON.
5335 It's our intention to add a similar features to most services and
5336 objects systemd manages, in order to simplify integration with
5337 program code that can consume JSON.
5339 * Similarly, networkd gained a Describe() method on its Manager and
5340 Link bus objects. This is exposed via "networkctl --json=".
5342 * hostnamectl's various "get-xyz"/"set-xyz" verb pairs
5343 (e.g. "hostnamectl get-hostname", "hostnamectl "set-hostname") have
5344 been replaced by a single "xyz" verb (e.g. "hostnamectl hostname")
5345 that is used both to get the value (when no argument is given), and
5346 to set the value (when an argument is specified). The old names
5347 continue to be supported for compatibility.
5349 * systemd-detect-virt and ConditionVirtualization= are now able to
5350 correctly identify Amazon EC2 environments.
5352 * The LogLevelMax= setting of unit files now applies not only to log
5353 messages generated *by* the service, but also to log messages
5354 generated *about* the service by PID 1. To suppress logs concerning a
5355 specific service comprehensively, set this option to a high log
5358 * bootctl gained support for a new --make-machine-id-directory= switch
5359 that allows precise control on whether to create the top-level
5360 per-machine directory in the boot partition that typically contains
5361 Type 1 boot loader entries.
5363 * During build SBAT data to include in the systemd-boot EFI PE binaries
5364 may be specified now.
5366 * /etc/crypttab learnt a new option "headless". If specified any
5367 requests to query the user interactively for passwords or PINs will
5368 be skipped. This is useful on systems that are headless, i.e. where
5369 an interactive user is generally not present.
5371 * /etc/crypttab also learnt a new option "password-echo=" that allows
5372 configuring whether the encryption password prompt shall echo the
5373 typed password and if so, do so literally or via asterisks. (The
5374 default is the same behaviour as before: provide echo feedback via
5377 * FIDO2 support in systemd-cryptenroll/systemd-cryptsetup and
5378 systemd-homed has been updated to allow explicit configuration of the
5379 "user presence" and "user verification" checks, as well as whether a
5380 PIN is required for authentication, via the new switches
5381 --fido2-with-user-presence=, --fido2-with-user-verification=,
5382 --fido2-with-client-pin= to systemd-cryptenroll and homectl. Which
5383 features are available, and may be enabled or disabled depends on the
5386 * systemd-nspawn's --private-user= switch now accepts the special value
5387 "identity" which configures a user namespacing environment with an
5388 identity mapping of 65535 UIDs. This means the container UID 0 is
5389 mapped to the host UID 0, and the UID 1 to host UID 1. On first look
5390 this doesn't appear to be useful, however it does reduce the attack
5391 surface a bit, since the resulting container will possess process
5392 capabilities only within its namespace and not on the host.
5394 * systemd-nspawn's --private-user-chown switch has been replaced by a
5395 more generic --private-user-ownership= switch that accepts one of
5396 three values: "chown" is equivalent to the old --private-user-chown,
5397 and "off" is equivalent to the absence of the old switch. The value
5398 "map" uses the new UID mapping mounts of Linux 5.12 to map ownership
5399 of files and directories of the underlying image to the chosen UID
5400 range for the container. "auto" is equivalent to "map" if UID mapping
5401 mount are supported, otherwise it is equivalent to "chown". The short
5402 -U switch systemd-nspawn now implies --private-user-ownership=auto
5403 instead of the old --private-user-chown. Effectively this means: if
5404 the backing file system supports UID mapping mounts the feature is
5405 now used by default if -U is used. Generally, it's a good idea to use
5406 UID mapping mounts instead of recursive chown()ing, since it allows
5407 running containers off immutable images (since no modifications of
5408 the images need to take place), and share images between multiple
5409 instances. Moreover, the recursive chown()ing operation is slow and
5410 can be avoided. Conceptually it's also a good thing if transient UID
5411 range uses do not leak into persistent file ownership anymore. TLDR:
5412 finally, the last major drawback of user namespacing has been
5413 removed, and -U should always be used (unless you use btrfs, where
5414 UID mapped mounts do not exist; or your container actually needs
5415 privileges on the host).
5417 * nss-systemd now synthesizes user and group shadow records in addition
5418 to the main user and group records. Thus, hashed passwords managed by
5419 systemd-homed are now accessible via the shadow database.
5421 * The userdb logic (and thus nss-systemd, and so on) now read
5422 additional user/group definitions in JSON format from the drop-in
5423 directories /etc/userdb/, /run/userdb/, /run/host/userdb/ and
5424 /usr/lib/userdb/. This is a simple and powerful mechanism for making
5425 additional users available to the system, with full integration into
5426 NSS including the shadow databases. Since the full JSON user/group
5427 record format is supported this may also be used to define users with
5428 resource management settings and other runtime settings that
5429 pam_systemd and systemd-logind enforce at login.
5431 * The userdbctl tool gained two new switches --with-dropin= and
5432 --with-varlink= which can be used to fine-tune the sources used for
5433 user database lookups.
5435 * systemd-nspawn gained a new switch --bind-user= for binding a host
5436 user account into the container. This does three things: the user's
5437 home directory is bind mounted from the host into the container,
5438 below the /run/userdb/home/ hierarchy. A free UID is picked in the
5439 container, and a user namespacing UID mapping to the host user's UID
5440 installed. And finally, a minimal JSON user and group record (along
5441 with its hashed password) is dropped into /run/host/userdb/. These
5442 records are picked up automatically by the userdb drop-in logic
5443 describe above, and allow the user to login with the same password as
5444 on the host. Effectively this means: if host and container run new
5445 enough systemd versions making a host user available to the container
5446 is trivially simple.
5448 * systemd-journal-gatewayd now supports the switches --user, --system,
5449 --merge, --file= that are equivalent to the same switches of
5450 journalctl, and permit exposing only the specified subset of the
5453 * The OnFailure= dependency between units is now augmented with a
5454 implicit reverse dependency OnFailureOf= (this new dependency cannot
5455 be configured directly it's only created as effect of an OnFailure=
5456 dependency in the reverse order — it's visible in "systemctl show"
5457 however). Similar, Slice= now has an reverse dependency SliceOf=,
5458 that is also not configurable directly, but useful to determine all
5459 units that are members of a slice.
5461 * A pair of new dependency types between units PropagatesStopTo= +
5462 StopPropagatedFrom= has been added, that allows propagation of unit
5463 stop events between two units. It operates similar to the existing
5464 PropagatesReloadTo= + ReloadPropagatedFrom= dependencies.
5466 * A new dependency type OnSuccess= has been added (plus the reverse
5467 dependency OnSuccessOf=, which cannot be configured directly, but
5468 exists only as effect of the reverse OnSuccess=). It is similar to
5469 OnFailure=, but triggers in the opposite case: when a service exits
5470 cleanly. This allows "chaining up" of services where one or more
5471 services are started once another service has successfully completed.
5473 * A new dependency type Upholds= has been added (plus the reverse
5474 dependency UpheldBy=, which cannot be configured directly, but exists
5475 only as effect of Upholds=). This dependency type is a stronger form
5476 of Wants=: if a unit has an UpHolds= dependency on some other unit
5477 and the former is active then the latter is started whenever it is
5478 found inactive (and no job is queued for it). This is an alternative
5479 to Restart= inside service units, but less configurable, and the
5480 request to uphold a unit is not encoded in the unit itself but in
5481 another unit that intends to uphold it.
5483 * The systemd-ask-password tool now also supports reading passwords
5484 from the credentials subsystem, via the new --credential= switch.
5486 * The systemd-ask-password tool learnt a new switch --emoji= which may
5487 be used to explicit control whether the lock and key emoji (🔐) is
5488 shown in the password prompt on suitable TTYs.
5490 * The --echo switch of systemd-ask-password now optionally takes a
5491 parameter that controls character echo. It may either show asterisks
5492 (default, as before), turn echo off entirely, or echo the typed
5493 characters literally.
5495 * The systemd-ask-password tool also gained a new -n switch for
5496 suppressing output of a trailing newline character when writing the
5497 acquired password to standard output, similar to /bin/echo's -n
5500 * New documentation has been added that describes the organization of
5501 the systemd source code tree:
5503 https://systemd.io/ARCHITECTURE
5505 * Units using ConditionNeedsUpdate= will no longer be activated in
5508 * It is now possible to list a template unit in the WantedBy= or
5509 RequiredBy= settings of the [Install] section of another template
5510 unit, which will be instantiated using the same instance name.
5512 * A new MemoryAvailable property is available for units. If the unit,
5513 or the slices it is part of, have a memory limit set via MemoryMax=/
5514 MemoryHigh=, MemoryAvailable will indicate how much more memory the
5515 unit can claim before hitting the limits.
5517 * systemd-coredump will now try to stay below the cgroup memory limit
5518 placed on itself or one of the slices it runs under, if the storage
5519 area for core files (/var/lib/systemd/coredump/) is placed on a tmpfs,
5520 since files written on such filesystems count toward the cgroup memory
5521 limit. If there is not enough available memory in such cases to store
5522 the core file uncompressed, systemd-coredump will skip to compressed
5523 storage directly (if enabled) and it will avoid analyzing the core file
5524 to print backtrace and metadata in the journal.
5526 * tmpfiles.d/ drop-ins gained a new '=' modifier to check if the type
5527 of a path matches the configured expectations, and remove it if not.
5529 * tmpfiles.d/'s 'Age' now accepts an 'age-by' argument, which allows to
5530 specify which of the several available filesystem timestamps (access
5531 time, birth time, change time, modification time) to look at when
5532 deciding whether a path has aged enough to be cleaned.
5534 * A new IPv6StableSecretAddress= setting has been added to .network
5535 files, which takes an IPv6 address to use as secret for IPv6 address
5538 * The [DHCPServer] logic in .network files gained support for a new
5539 UplinkInterface= setting that permits configuration of the uplink
5540 interface name to propagate DHCP lease information from.
5542 * The WakeOnLan= setting in .link files now accepts a list of flags
5543 instead of a single one, to configure multiple wake-on-LAN policies.
5545 * User-space defined tracepoints (USDT) have been added to udev at
5546 strategic locations. This is useful for tracing udev behaviour and
5547 performance with bpftrace and similar tools.
5549 * systemd-journald-upload gained a new NetworkTimeoutSec= option for
5550 setting a network timeout time.
5552 * If a system service is running in a new mount namespace (RootDirectory=
5553 and friends), all file systems will be mounted with MS_NOSUID by
5554 default, unless the system is running with SELinux enabled.
5556 * When enumerating time zones the timedatectl tool will now consult the
5557 'tzdata.zi' file shipped by the IANA time zone database package, in
5558 addition to 'zone1970.tab', as before. This makes sure time zone
5559 aliases are now correctly supported. Some distributions so far did
5560 not install this additional file, most do however. If you
5561 distribution does not install it yet, it might make sense to change
5564 * Intel HID rfkill event is no longer masked, since it's the only
5565 source of rfkill event on newer HP laptops. To have both backward and
5566 forward compatibility, userspace daemon needs to debounce duplicated
5567 events in a short time window.
5569 Contributions from: Aakash Singh, adrian5, Albert Brox,
5570 Alexander Sverdlin, Alexander Tsoy, Alexey Rubtsov, alexlzhu,
5571 Allen Webb, Alvin Šipraga, Alyssa Ross, Anders Wenhaug,
5572 Andrea Pappacoda, Anita Zhang, asavah, Balint Reczey, Bertrand Jacquin,
5573 borna-blazevic, caoxia2008cxx, Carlo Teubner, Christian Göttsche,
5574 Christian Hesse, Daniel Schaefer, Dan Streetman,
5575 David Santamaría Rogado, David Tardon, Deepak Rawat, dgcampea,
5576 Dimitri John Ledkov, ei-ke, Emilio Herrera, Emil Renner Berthing,
5577 Eric Cook, Flos Lonicerae, Franck Bui, Francois Gervais,
5578 Frantisek Sumsal, Gibeom Gwon, gitm0, Hamish Moffatt, Hans de Goede,
5579 Harsh Barsaiyan, Henri Chain, Hristo Venev, Icenowy Zheng, Igor Zhbanov,
5580 imayoda, Jakub Warczarek, James Buren, Jan Janssen, Jan Macku,
5581 Jan Synacek, Jason Francis, Jayanth Ananthapadmanaban, Jeremy Szu,
5582 Jérôme Carretero, Jesse Stricker, jiangchuangang, Joerg Behrmann,
5583 Jóhann B. Guðmundsson, Jörg Deckert, Jörg Thalheim, Juergen Hoetzel,
5584 Julia Kartseva, Kai-Heng Feng, Khem Raj, KoyamaSohei, laineantti,
5585 Lennart Poettering, LetzteInstanz, Luca Adrian L, Luca Boccassi,
5586 Lucas Magasweran, Mantas Mikulėnas, Marco Antonio Mauro, Mark Wielaard,
5587 Masahiro Matsuya, Matt Johnston, Michael Catanzaro, Michal Koutný,
5588 Michal Sekletár, Mike Crowe, Mike Kazantsev, Milan, milaq,
5589 Miroslav Suchý, Morten Linderud, nerdopolis, nl6720, Noah Meyerhans,
5590 Oleg Popov, Olle Lundberg, Ondrej Kozina, Paweł Marciniak, Perry.Yuan,
5591 Peter Hutterer, Peter Kjellerstedt, Peter Morrow, Phaedrus Leeds,
5592 plattrap, qhill, Raul Tambre, Roman Beranek, Roshan Shariff,
5593 Ryan Hendrickson, Samuel BF, scootergrisen, Sebastian Blunt,
5594 Seong-ho Cho, Sergey Bugaev, Sevan Janiyan, Sibo Dong, simmon,
5595 Simon Watts, Srinidhi Kaushik, Štěpán Němec, Steve Bonds, Susant Sahani,
5596 sverdlin, syyhao1994, Takashi Sakamoto, Topi Miettinen, tramsay,
5597 Trent Piepho, Uwe Kleine-König, Viktor Mihajlovski, Vincent Dechenaux,
5598 Vito Caputo, William A. Kennington III, Yangyang Shen, Yegor Alexeyev,
5599 Yi Gao, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, zsien, наб
5601 — Edinburgh, 2021-07-07
5605 * A concept of system extension images is introduced. Such images may
5606 be used to extend the /usr/ and /opt/ directory hierarchies at
5607 runtime with additional files (even if the file system is read-only).
5608 When a system extension image is activated, its /usr/ and /opt/
5609 hierarchies and os-release information are combined via overlayfs
5610 with the file system hierarchy of the host OS.
5612 A new systemd-sysext tool can be used to merge, unmerge, list, and
5613 refresh system extension hierarchies. See
5614 https://www.freedesktop.org/software/systemd/man/systemd-sysext.html.
5616 The systemd-sysext.service automatically merges installed system
5617 extensions during boot (before basic.target, but not in very early
5618 boot, since various file systems have to be mounted first).
5620 The SYSEXT_LEVEL= field in os-release(5) may be used to specify the
5621 supported system extension level.
5623 * A new ExtensionImages= unit setting can be used to apply the same
5624 system extension image concept from systemd-sysext to the namespaced
5625 file hierarchy of specific services, following the same rules and
5628 * Support for a new special "root=tmpfs" kernel command-line option has
5629 been added. When specified, a tmpfs is mounted on /, and mount.usr=
5630 should be used to point to the operating system implementation.
5632 * A new configuration file /etc/veritytab may be used to configure
5633 dm-verity integrity protection for block devices. Each line is in the
5634 format "volume-name data-device hash-device roothash options",
5635 similar to /etc/crypttab.
5637 * A new kernel command-line option systemd.verity.root_options= may be
5638 used to configure dm-verity behaviour for the root device.
5640 * The key file specified in /etc/crypttab (the third field) may now
5641 refer to an AF_UNIX/SOCK_STREAM socket in the file system. The key is
5642 acquired by connecting to that socket and reading from it. This
5643 allows the implementation of a service to provide key information
5644 dynamically, at the moment when it is needed.
5646 * When the hostname is set explicitly to "localhost", systemd-hostnamed
5647 will respect this. Previously such a setting would be mostly silently
5648 ignored. The goal is to honour configuration as specified by the
5651 * The fallback hostname that will be used by the system manager and
5652 systemd-hostnamed can now be configured in two new ways: by setting
5653 DEFAULT_HOSTNAME= in os-release(5), or by setting
5654 $SYSTEMD_DEFAULT_HOSTNAME in the environment block. As before, it can
5655 also be configured during compilation. The environment variable is
5656 intended for testing and local overrides, the os-release(5) field is
5657 intended to allow customization by different variants of a
5658 distribution that share the same compiled packages.
5660 * The environment block of the manager itself may be configured through
5661 a new ManagerEnvironment= setting in system.conf or user.conf. This
5662 complements existing ways to set the environment block (the kernel
5663 command line for the system manager, the inherited environment and
5664 user@.service unit file settings for the user manager).
5666 * systemd-hostnamed now exports the default hostname and the source of
5667 the configured hostname ("static", "transient", or "default") as
5670 * systemd-hostnamed now exports the "HardwareVendor" and
5671 "HardwareModel" D-Bus properties, which are supposed to contain a
5672 pair of cleaned up, human readable strings describing the system's
5673 vendor and model. It's typically sourced from the firmware's DMI
5674 tables, but may be augmented from a new hwdb database. hostnamectl
5675 shows this in the status output.
5677 * Support has been added to systemd-cryptsetup for extracting the
5678 PKCS#11 token URI and encrypted key from the LUKS2 JSON embedded
5679 metadata header. This allows the information how to open the
5680 encrypted device to be embedded directly in the device and obviates
5681 the need for configuration in an external file.
5683 * systemd-cryptsetup gained support for unlocking LUKS2 volumes using
5684 TPM2 hardware, as well as FIDO2 security tokens (in addition to the
5685 pre-existing support for PKCS#11 security tokens).
5687 * systemd-repart may enroll encrypted partitions using TPM2
5688 hardware. This may be useful for example to create an encrypted /var
5689 partition bound to the machine on first boot.
5691 * A new systemd-cryptenroll tool has been added to enroll TPM2, FIDO2
5692 and PKCS#11 security tokens to LUKS volumes, list and destroy
5695 https://0pointer.net/blog/unlocking-luks2-volumes-with-tpm2-fido2-pkcs11-security-hardware-on-systemd-248.html
5697 It also supports enrolling "recovery keys" and regular passphrases.
5699 * The libfido2 dependency is now based on dlopen(), so that the library
5700 is used at runtime when installed, but is not a hard runtime
5703 * systemd-cryptsetup gained support for two new options in
5704 /etc/crypttab: "no-write-workqueue" and "no-read-workqueue" which
5705 request synchronous processing of encryption/decryption IO.
5707 * The manager may be configured at compile time to use the fexecve()
5708 instead of the execve() system call when spawning processes. Using
5709 fexecve() closes a window between checking the security context of an
5710 executable and spawning it, but unfortunately the kernel displays
5711 stale information in the process' "comm" field, which impacts ps
5714 * The configuration option -Dcompat-gateway-hostname has been dropped.
5715 "_gateway" is now the only supported name.
5717 * The ConditionSecurity=tpm2 unit file setting may be used to check if
5718 the system has at least one TPM2 (tpmrm class) device.
5720 * A new ConditionCPUFeature= has been added that may be used to
5721 conditionalize units based on CPU features. For example,
5722 ConditionCPUFeature=rdrand will condition a unit so that it is only
5723 run when the system CPU supports the RDRAND opcode.
5725 * The existing ConditionControlGroupController= setting has been
5726 extended with two new values "v1" and "v2". "v2" means that the
5727 unified v2 cgroup hierarchy is used, and "v1" means that legacy v1
5728 hierarchy or the hybrid hierarchy are used.
5730 * A new PrivateIPC= setting on a unit file allows executed processes to
5731 be moved into a private IPC namespace, with separate System V IPC
5732 identifiers and POSIX message queues.
5734 A new IPCNamespacePath= allows the unit to be joined to an existing
5737 * The tables of system calls in seccomp filters are now automatically
5738 generated from kernel lists exported on
5739 https://fedora.juszkiewicz.com.pl/syscalls.html.
5741 The following architectures should now have complete lists:
5742 alpha, arc, arm64, arm, i386, ia64, m68k, mips64n32, mips64, mipso32,
5743 powerpc, powerpc64, s390, s390x, tilegx, sparc, x86_64, x32.
5745 * The MountAPIVFS= service file setting now additionally mounts a tmpfs
5746 on /run/ if it is not already a mount point. A writable /run/ has
5747 always been a requirement for a functioning system, but this was not
5748 guaranteed when using a read-only image.
5750 Users can always specify BindPaths= or InaccessiblePaths= as
5751 overrides, and they will take precedence. If the host's root mount
5752 point is used, there is no change in behaviour.
5754 * New bind mounts and file system image mounts may be injected into the
5755 mount namespace of a service (without restarting it). This is exposed
5756 respectively as 'systemctl bind <unit> <path>…' and
5757 'systemctl mount-image <unit> <image>…'.
5759 * The StandardOutput= and StandardError= settings can now specify files
5760 to be truncated for output (as "truncate:<path>").
5762 * The ExecPaths= and NoExecPaths= settings may be used to specify
5763 noexec for parts of the file system.
5765 * sd-bus has a new function sd_bus_open_user_machine() to open a
5766 connection to the session bus of a specific user in a local container
5767 or on the local host. This is exposed in the existing -M switch to
5768 systemctl and similar tools:
5770 systemctl --user -M lennart@foobar start foo
5772 This will connect to the user bus of a user "lennart" in container
5773 "foobar". If no container name is specified, the specified user on
5774 the host itself is connected to
5776 systemctl --user -M lennart@ start quux
5778 * sd-bus also gained a convenience function sd_bus_message_send() to
5779 simplify invocations of sd_bus_send(), taking only a single
5780 parameter: the message to send.
5782 * sd-event allows rate limits to be set on event sources, for dealing
5783 with high-priority event sources that might starve out others. See
5784 the new man page sd_event_source_set_ratelimit(3) for details.
5786 * systemd.link files gained a [Link] Promiscuous= switch, which allows
5787 the device to be raised in promiscuous mode.
5789 New [Link] TransmitQueues= and ReceiveQueues= settings allow the
5790 number of TX and RX queues to be configured.
5792 New [Link] TransmitQueueLength= setting allows the size of the TX
5793 queue to be configured.
5795 New [Link] GenericSegmentOffloadMaxBytes= and
5796 GenericSegmentOffloadMaxSegments= allow capping the packet size and
5797 the number of segments accepted in Generic Segment Offload.
5799 * systemd-networkd gained support for the "B.A.T.M.A.N. advanced"
5800 wireless routing protocol that operates on ISO/OSI Layer 2 only and
5801 uses ethernet frames to route/bridge packets. This encompasses a new
5802 "batadv" netdev Type=, a new [BatmanAdvanced] section with a bunch of
5803 new settings in .netdev files, and a new BatmanAdvanced= setting in
5806 * systemd.network files gained a [Network] RouteTable= configuration
5807 switch to select the routing policy table.
5809 systemd.network files gained a [RoutingPolicyRule] Type=
5810 configuration switch (one of "blackhole, "unreachable", "prohibit").
5812 systemd.network files gained a [IPv6AcceptRA] RouteDenyList= and
5813 RouteAllowList= settings to ignore/accept route advertisements from
5814 routers matching specified prefixes. The DenyList= setting has been
5815 renamed to PrefixDenyList= and a new PrefixAllowList= option has been
5818 systemd.network files gained a [DHCPv6] UseAddress= setting to
5819 optionally ignore the address provided in the lease.
5821 systemd.network files gained a [DHCPv6PrefixDelegation]
5822 ManageTemporaryAddress= switch.
5824 systemd.network files gained a new ActivationPolicy= setting which
5825 allows configuring how the UP state of an interface shall be managed,
5826 i.e. whether the interface is always upped, always downed, or may be
5827 upped/downed by the user using "ip link set dev".
5829 * The default for the Broadcast= setting in .network files has slightly
5830 changed: the broadcast address will not be configured for wireguard
5833 * systemd.netdev files gained a [VLAN] Protocol=, IngressQOSMaps=,
5834 EgressQOSMaps=, and [MACVLAN] BroadcastMulticastQueueLength=
5835 configuration options for VLAN packet handling.
5837 * udev rules may now set log_level= option. This allows debug logs to
5838 be enabled for select events, e.g. just for a specific subsystem or
5839 even a single device.
5841 * udev now exports the VOLUME_ID, LOGICAL_VOLUME_ID, VOLUME_SET_ID, and
5842 DATA_PREPARED_ID properties for block devices with ISO9660 file
5845 * udev now exports decoded DMI information about installed memory slots
5846 as device properties under the /sys/class/dmi/id/ pseudo device.
5848 * /dev/ is not mounted noexec anymore. This didn't provide any
5849 significant security benefits and would conflict with the executable
5850 mappings used with /dev/sgx device nodes. The previous behaviour can
5851 be restored for individual services with NoExecPaths=/dev (or by allow-
5852 listing and excluding /dev from ExecPaths=).
5854 * Permissions for /dev/vsock are now set to 0o666, and /dev/vhost-vsock
5855 and /dev/vhost-net are owned by the kvm group.
5857 * The hardware database has been extended with a list of fingerprint
5858 readers that correctly support USB auto-suspend using data from
5861 * systemd-resolved can now answer DNSSEC questions through the stub
5862 resolver interface in a way that allows local clients to do DNSSEC
5863 validation themselves. For a question with DO+CD set, it'll proxy the
5864 DNS query and respond with a mostly unmodified packet received from
5865 the upstream server.
5867 * systemd-resolved learnt a new boolean option CacheFromLocalhost= in
5868 resolved.conf. If true the service will provide caching even for DNS
5869 lookups made to an upstream DNS server on the 127.0.0.1/::1
5870 addresses. By default (and when the option is false) systemd-resolved
5871 will not cache such lookups, in order to avoid duplicate local
5872 caching, under the assumption the local upstream server caches
5875 * systemd-resolved now implements RFC5001 NSID in its local DNS
5876 stub. This may be used by local clients to determine whether they are
5877 talking to the DNS resolver stub or a different DNS server.
5879 * When resolving host names and other records resolvectl will now
5880 report where the data was acquired from (i.e. the local cache, the
5881 network, locally synthesized, …) and whether the network traffic it
5882 effected was encrypted or not. Moreover the tool acquired a number of
5883 new options --cache=, --synthesize=, --network=, --zone=,
5884 --trust-anchor=, --validate= that take booleans and may be used to
5885 tweak a lookup, i.e. whether it may be answered from cached
5886 information, locally synthesized information, information acquired
5887 through the network, the local mDNS/LLMNR zone, the DNSSEC trust
5888 anchor, and whether DNSSEC validation shall be executed for the
5891 * systemd-nspawn gained a new --ambient-capability= setting
5892 (AmbientCapability= in .nspawn files) to configure ambient
5893 capabilities passed to the container payload.
5895 * systemd-nspawn gained the ability to configure the firewall using the
5896 nftables subsystem (in addition to the existing iptables
5897 support). Similarly, systemd-networkd's IPMasquerade= option now
5898 supports nftables as back-end, too. In both cases NAT on IPv6 is now
5899 supported too, in addition to IPv4 (the iptables back-end still is
5902 "IPMasquerade=yes", which was the same as "IPMasquerade=ipv4" before,
5903 retains its meaning, but has been deprecated. Please switch to either
5904 "ivp4" or "both" (if covering IPv6 is desired).
5906 * systemd-importd will now download .verity and .roothash.p7s files
5907 along with the machine image (as exposed via machinectl pull-raw).
5909 * systemd-oomd now gained a new DefaultMemoryPressureDurationSec=
5910 setting to configure the time a unit's cgroup needs to exceed memory
5911 pressure limits before action will be taken, and a new
5912 ManagedOOMPreference=none|avoid|omit setting to avoid killing certain
5915 systemd-oomd is now considered fully supported (the usual
5916 backwards-compatibility promises apply). Swap is not required for
5917 operation, but it is still recommended.
5919 * systemd-timesyncd gained a new ConnectionRetrySec= setting which
5920 configures the retry delay when trying to contact servers.
5922 * systemd-stdio-bridge gained --system/--user options to connect to the
5923 system bus (previous default) or the user session bus.
5925 * systemd-localed may now call locale-gen to generate missing locales
5926 on-demand (UTF-8-only). This improves integration with Debian-based
5927 distributions (Debian/Ubuntu/PureOS/Tanglu/...) and Arch Linux.
5929 * systemctl --check-inhibitors=true may now be used to obey inhibitors
5930 even when invoked non-interactively. The old --ignore-inhibitors
5931 switch is now deprecated and replaced by --check-inhibitors=false.
5933 * systemctl import-environment will now emit a warning when called
5934 without any arguments (i.e. to import the full environment block of
5935 the called program). This command will usually be invoked from a
5936 shell, which means that it'll inherit a bunch of variables which are
5937 specific to that shell, and usually to the TTY the shell is connected
5938 to, and don't have any meaning in the global context of the system or
5939 user service manager. Instead, only specific variables should be
5940 imported into the manager environment block.
5942 Similarly, programs which update the manager environment block by
5943 directly calling the D-Bus API of the manager, should also push
5944 specific variables, and not the full inherited environment.
5946 * systemctl's status output now shows unit state with a more careful
5947 choice of Unicode characters: units in maintenance show a "○" symbol
5948 instead of the usual "●", failed units show "×", and services being
5951 * coredumpctl gained a --debugger-arguments= switch to pass arguments
5952 to the debugger. It also gained support for showing coredump info in
5953 a simple JSON format.
5955 * systemctl/loginctl/machinectl's --signal= option now accept a special
5956 value "list", which may be used to show a brief table with known
5957 process signals and their numbers.
5959 * networkctl now shows the link activation policy in status.
5961 * Various tools gained --pager/--no-pager/--json= switches to
5962 enable/disable the pager and provide JSON output.
5964 * Various tools now accept two new values for the SYSTEMD_COLORS
5965 environment variable: "16" and "256", to configure how many terminal
5966 colors are used in output.
5968 * less 568 or newer is now required for the auto-paging logic of the
5969 various tools. Hyperlink ANSI sequences in terminal output are now
5970 used even if a pager is used, and older versions of less are not able
5971 to display these sequences correctly. SYSTEMD_URLIFY=0 may be used to
5972 disable this output again.
5974 * Builds with support for separate / and /usr/ hierarchies ("split-usr"
5975 builds, non-merged-usr builds) are now officially deprecated. A
5976 warning is emitted during build. Support is slated to be removed in
5977 about a year (when the Debian Bookworm release development starts).
5979 * Systems with the legacy cgroup v1 hierarchy are now marked as
5980 "tainted", to make it clearer that using the legacy hierarchy is not
5983 * systemd-localed will now refuse to configure a keymap which is not
5984 installed in the file system. This is intended as a bug fix, but
5985 could break cases where systemd-localed was used to configure the
5986 keymap in advanced of it being installed. It is necessary to install
5987 the keymap file first.
5989 * The main git development branch has been renamed to 'main'.
5991 * mmcblk[0-9]boot[0-9] devices will no longer be probed automatically
5992 for partitions, as in the vast majority of cases they contain none
5993 and are used internally by the bootloader (eg: uboot).
5995 * systemd will now set the $SYSTEMD_EXEC_PID environment variable for
5996 spawned processes to the PID of the process itself. This may be used
5997 by programs for detecting whether they were forked off by the service
5998 manager itself or are a process forked off further down the tree.
6000 * The sd-device API gained four new calls: sd_device_get_action() to
6001 determine the uevent add/remove/change/… action the device object has
6002 been seen for, sd_device_get_seqno() to determine the uevent sequence
6003 number, sd_device_new_from_stat_rdev() to allocate a new sd_device
6004 object from stat(2) data of a device node, and sd_device_trigger() to
6005 write to the 'uevent' attribute of a device.
6007 * For most tools the --no-legend= switch has been replaced by
6008 --legend=no and --legend=yes, to force whether tables are shown with
6011 * Units acquired a new property "Markers" that takes a list of zero,
6012 one or two of the following strings: "needs-reload" and
6013 "needs-restart". These markers may be set via "systemctl
6014 set-property". Once a marker is set, "systemctl reload-or-restart
6015 --marked" may be invoked to execute the operation the units are
6016 marked for. This is useful for package managers that want to mark
6017 units for restart/reload while updating, but effect the actual
6018 operations at a later step at once.
6020 * The sd_bus_message_read_strv() API call of sd-bus may now also be
6021 used to parse arrays of D-Bus signatures and D-Bus paths, in addition
6024 * bootctl will now report whether the UEFI firmware used a TPM2 device
6025 and measured the boot process into it.
6027 * systemd-tmpfiles learnt support for a new environment variable
6028 $SYSTEMD_TMPFILES_FORCE_SUBVOL which takes a boolean value. If true
6029 the v/q/Q lines in tmpfiles.d/ snippets will create btrfs subvolumes
6030 even if the root fs of the system is not itself a btrfs volume.
6032 * systemd-detect-virt/ConditionVirtualization= will now explicitly
6033 detect Docker/Podman environments where possible. Moreover, they
6034 should be able to generically detect any container manager as long as
6035 it assigns the container a cgroup.
6037 * portablectl gained a new "reattach" verb for detaching/reattaching a
6038 portable service image, useful for updating images on-the-fly.
6040 * Intel SGX enclave device nodes (which expose a security feature of
6041 newer Intel CPUs) will now be owned by a new system group "sgx".
6043 Contributions from: Adam Nielsen, Adrian Vovk, AJ Jordan, Alan Perry,
6044 Alastair Pharo, Alexander Batischev, Ali Abdallah, Andrew Balmos,
6045 Anita Zhang, Annika Wickert, Ansgar Burchardt, Antonio Terceiro,
6046 Antonius Frie, Ardy, Arian van Putten, Ariel Fermani, Arnaud T,
6047 A S Alam, Bastien Nocera, Benjamin Berg, Benjamin Robin, Björn Daase,
6048 caoxia, Carlo Wood, Charles Lee, ChopperRob, chri2, Christian Ehrhardt,
6049 Christian Hesse, Christopher Obbard, clayton craft, corvusnix, cprn,
6050 Daan De Meyer, Daniele Medri, Daniel Rusek, Dan Sanders, Dan Streetman,
6051 Darren Ng, David Edmundson, David Tardon, Deepak Rawat, Devon Pringle,
6052 Dmitry Borodaenko, dropsignal, Einsler Lee, Endre Szabo,
6053 Evgeny Vereshchagin, Fabian Affolter, Fangrui Song, Felipe Borges,
6054 feliperodriguesfr, Felix Stupp, Florian Hülsmann, Florian Klink,
6055 Florian Westphal, Franck Bui, Frantisek Sumsal, Gablegritule,
6056 Gaël PORTAY, Gaurav, Giedrius Statkevičius, Greg Depoire-Ferrer,
6057 Gustavo Costa, Hans de Goede, Hela Basa, heretoenhance, hide,
6058 Iago López Galeiras, igo95862, Ilya Dmitrichenko, Jameer Pathan,
6059 Jan Tojnar, Jiehong, Jinyuan Si, Joerg Behrmann, John Slade,
6060 Jonathan G. Underwood, Jonathan McDowell, Josh Triplett, Joshua Watt,
6061 Julia Cartwright, Julien Humbert, Kairui Song, Karel Zak,
6062 Kevin Backhouse, Kevin P. Fleming, Khem Raj, Konomi, krissgjeng,
6063 l4gfcm, Lajos Veres, Lennart Poettering, Lincoln Ramsay, Luca Boccassi,
6064 Luca BRUNO, Lucas Werkmeister, Luka Kudra, Luna Jernberg,
6065 Marc-André Lureau, Martin Wilck, Matthias Klumpp, Matt Turner,
6066 Michael Gisbers, Michael Marley, Michael Trapp, Michal Fabik,
6067 Michał Kopeć, Michal Koutný, Michal Sekletár, Michele Guerini Rocco,
6068 Mike Gilbert, milovlad, moson-mo, Nick, nihilix-melix, Oğuz Ersen,
6069 Ondrej Mosnacek, pali, Pavel Hrdina, Pavel Sapezhko, Perry Yuan,
6070 Peter Hutterer, Pierre Dubouilh, Piotr Drąg, Pjotr Vertaalt,
6071 Richard Laager, RussianNeuroMancer, Sam Lunt, Sebastiaan van Stijn,
6072 Sergey Bugaev, shenyangyang4, simmon, Simonas Kazlauskas,
6073 Slimane Selyan Amiri, Stefan Agner, Steve Ramage, Susant Sahani,
6074 Sven Mueller, Tad Fisher, Takashi Iwai, Thomas Haller, Tom Shield,
6075 Topi Miettinen, Torsten Hilbrich, tpgxyz, Tyler Hicks, ulf-f,
6076 Ulrich Ölmann, Vincent Pelletier, Vinnie Magro, Vito Caputo, Vlad,
6077 walbit-de, Whired Planck, wouter bolsterlee, Xℹ Ruoyao, Yangyang Shen,
6078 Yuri Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek,
6079 Zmicer Turok, Дамјан Георгиевски
6081 — Berlin, 2021-03-30
6085 * KERNEL API INCOMPATIBILITY: Linux 4.14 introduced two new uevents
6086 "bind" and "unbind" to the Linux device model. When this kernel
6087 change was made, systemd-udevd was only minimally updated to handle
6088 and propagate these new event types. The introduction of these new
6089 uevents (which are typically generated for USB devices and devices
6090 needing a firmware upload before being functional) resulted in a
6091 number of issues which we so far didn't address. We hoped the kernel
6092 maintainers would themselves address these issues in some form, but
6093 that did not happen. To handle them properly, many (if not most) udev
6094 rules files shipped in various packages need updating, and so do many
6095 programs that monitor or enumerate devices with libudev or sd-device,
6096 or otherwise process uevents. Please note that this incompatibility
6097 is not fault of systemd or udev, but caused by an incompatible kernel
6098 change that happened back in Linux 4.14, but is becoming more and
6099 more visible as the new uevents are generated by more kernel drivers.
6101 To minimize issues resulting from this kernel change (but not avoid
6102 them entirely) starting with systemd-udevd 247 the udev "tags"
6103 concept (which is a concept for marking and filtering devices during
6104 enumeration and monitoring) has been reworked: udev tags are now
6105 "sticky", meaning that once a tag is assigned to a device it will not
6106 be removed from the device again until the device itself is removed
6107 (i.e. unplugged). This makes sure that any application monitoring
6108 devices that match a specific tag is guaranteed to both see uevents
6109 where the device starts being relevant, and those where it stops
6110 being relevant (the latter now regularly happening due to the new
6111 "unbind" uevent type). The udev tags concept is hence now a concept
6112 tied to a *device* instead of a device *event* — unlike for example
6113 udev properties whose lifecycle (as before) is generally tied to a
6114 device event, meaning that the previously determined properties are
6115 forgotten whenever a new uevent is processed.
6117 With the newly redefined udev tags concept, sometimes it's necessary
6118 to determine which tags are the ones applied by the most recent
6119 uevent/database update, in order to discern them from those
6120 originating from earlier uevents/database updates of the same
6121 device. To accommodate for this a new automatic property CURRENT_TAGS
6122 has been added that works similar to the existing TAGS property but
6123 only lists tags set by the most recent uevent/database
6124 update. Similarly, the libudev/sd-device API has been updated with
6125 new functions to enumerate these 'current' tags, in addition to the
6126 existing APIs that now enumerate the 'sticky' ones.
6128 To properly handle "bind"/"unbind" on Linux 4.14 and newer it is
6129 essential that all udev rules files and applications are updated to
6130 handle the new events. Specifically:
6132 • All rule files that currently use a header guard similar to
6133 ACTION!="add|change",GOTO="xyz_end" should be updated to use
6134 ACTION=="remove",GOTO="xyz_end" instead, so that the
6135 properties/tags they add are also applied whenever "bind" (or
6136 "unbind") is seen. (This is most important for all physical device
6137 types — those for which "bind" and "unbind" are currently
6138 generated, for all other device types this change is still
6139 recommended but not as important — but certainly prepares for
6140 future kernel uevent type additions).
6142 • Similarly, all code monitoring devices that contains an 'if' branch
6143 discerning the "add" + "change" uevent actions from all other
6144 uevents actions (i.e. considering devices only relevant after "add"
6145 or "change", and irrelevant on all other events) should be reworked
6146 to instead negatively check for "remove" only (i.e. considering
6147 devices relevant after all event types, except for "remove", which
6148 invalidates the device). Note that this also means that devices
6149 should be considered relevant on "unbind", even though conceptually
6150 this — in some form — invalidates the device. Since the precise
6151 effect of "unbind" is not generically defined, devices should be
6152 considered relevant even after "unbind", however I/O errors
6153 accessing the device should then be handled gracefully.
6155 • Any code that uses device tags for deciding whether a device is
6156 relevant or not most likely needs to be updated to use the new
6157 udev_device_has_current_tag() API (or sd_device_has_current_tag()
6158 in case sd-device is used), to check whether the tag is set at the
6159 moment an uevent is seen (as opposed to the existing
6160 udev_device_has_tag() API which checks if the tag ever existed on
6161 the device, following the API concept redefinition explained
6164 We are very sorry for this breakage and the requirement to update
6165 packages using these interfaces. We'd again like to underline that
6166 this is not caused by systemd/udev changes, but result of a kernel
6169 * UPCOMING INCOMPATIBILITY: So far most downstream distribution
6170 packages have not retriggered devices once the udev package (or any
6171 auxiliary package installing additional udev rules) is updated. We
6172 intend to work with major distributions to change this, so that
6173 "udevadm trigger -c change" is issued on such upgrades, ensuring that
6174 the updated ruleset is applied to the devices already discovered, so
6175 that (asynchronously) after the upgrade completed the udev database
6176 is consistent with the updated rule set. This means udev rules must
6177 be ready to be retriggered with a "change" action any time, and
6178 result in correct and complete udev database entries. While the
6179 majority of udev rule files known to us currently get this right,
6180 some don't. Specifically, there are udev rules files included in
6181 various packages that only set udev properties on the "add" action,
6182 but do not handle the "change" action. If a device matching those
6183 rules is retriggered with the "change" action (as is intended here)
6184 it would suddenly lose the relevant properties. This always has been
6185 problematic, but as soon as all udev devices are triggered on relevant
6186 package upgrades this will become particularly so. It is strongly
6187 recommended to fix offending rules so that they can handle a "change"
6188 action at any time, and acquire all necessary udev properties even
6189 then. Or in other words: the header guard mentioned above
6190 (ACTION=="remove",GOTO="xyz_end") is the correct approach to handle
6191 this, as it makes sure rules are rerun on "change" correctly, and
6192 accumulate the correct and complete set of udev properties. udev rule
6193 definitions that cannot handle "change" events being triggered at
6194 arbitrary times should be considered buggy.
6196 * The MountAPIVFS= service file setting now defaults to on if
6197 RootImage= and RootDirectory= are used, which means that with those
6198 two settings /proc/, /sys/ and /dev/ are automatically properly set
6199 up for services. Previous behaviour may be restored by explicitly
6200 setting MountAPIVFS=off.
6202 * Since PAM 1.2.0 (2015) configuration snippets may be placed in
6203 /usr/lib/pam.d/ in addition to /etc/pam.d/. If a file exists in the
6204 latter it takes precedence over the former, similar to how most of
6205 systemd's own configuration is handled. Given that PAM stack
6206 definitions are primarily put together by OS vendors/distributions
6207 (though possibly overridden by users), this systemd release moves its
6208 own PAM stack configuration for the "systemd-user" PAM service (i.e.
6209 for the PAM session invoked by the per-user user@.service instance)
6210 from /etc/pam.d/ to /usr/lib/pam.d/. We recommend moving all
6211 packages' vendor versions of their PAM stack definitions from
6212 /etc/pam.d/ to /usr/lib/pam.d/, but if such OS-wide migration is not
6213 desired the location to which systemd installs its PAM stack
6214 configuration may be changed via the -Dpamconfdir Meson option.
6216 * The runtime dependencies on libqrencode, libpcre2, libidn/libidn2,
6217 libpwquality and libcryptsetup have been changed to be based on
6218 dlopen(): instead of regular dynamic library dependencies declared in
6219 the binary ELF headers, these libraries are now loaded on demand
6220 only, if they are available. If the libraries cannot be found the
6221 relevant operations will fail gracefully, or a suitable fallback
6222 logic is chosen. This is supposed to be useful for general purpose
6223 distributions, as it allows minimizing the list of dependencies the
6224 systemd packages pull in, permitting building of more minimal OS
6225 images, while still making use of these "weak" dependencies should
6226 they be installed. Since many package managers automatically
6227 synthesize package dependencies from ELF shared library dependencies,
6228 some additional manual packaging work has to be done now to replace
6229 those (slightly downgraded from "required" to "recommended" or
6230 whatever is conceptually suitable for the package manager). Note that
6231 this change does not alter build-time behaviour: as before the
6232 build-time dependencies have to be installed during build, even if
6233 they now are optional during runtime.
6235 * sd-event.h gained a new call sd_event_add_time_relative() for
6236 installing timers relative to the current time. This is mostly a
6237 convenience wrapper around the pre-existing sd_event_add_time() call
6238 which installs absolute timers.
6240 * sd-event event sources may now be placed in a new "exit-on-failure"
6241 mode, which may be controlled via the new
6242 sd_event_source_get_exit_on_failure() and
6243 sd_event_source_set_exit_on_failure() functions. If enabled, any
6244 failure returned by the event source handler functions will result in
6245 exiting the event loop (unlike the default behaviour of just
6246 disabling the event source but continuing with the event loop). This
6247 feature is useful to set for all event sources that define "primary"
6248 program behaviour (where failure should be fatal) in contrast to
6249 "auxiliary" behaviour (where failure should remain local).
6251 * Most event source types sd-event supports now accept a NULL handler
6252 function, in which case the event loop is exited once the event
6253 source is to be dispatched, using the userdata pointer — converted to
6254 a signed integer — as exit code of the event loop. Previously this
6255 was supported for IO and signal event sources already. Exit event
6256 sources still do not support this (simply because it makes little
6257 sense there, as the event loop is already exiting when they are
6260 * A new per-unit setting RootImageOptions= has been added which allows
6261 tweaking the mount options for any file system mounted as effect of
6262 the RootImage= setting.
6264 * Another new per-unit setting MountImages= has been added, that allows
6265 mounting additional disk images into the file system tree accessible
6268 * Timer units gained a new FixedRandomDelay= boolean setting. If
6269 enabled, the random delay configured with RandomizedDelaySec= is
6270 selected in a way that is stable on a given system (though still
6271 different for different units).
6273 * Socket units gained a new setting Timestamping= that takes "us", "ns"
6274 or "off". This controls the SO_TIMESTAMP/SO_TIMESTAMPNS socket
6277 * systemd-repart now generates JSON output when requested with the new
6280 * systemd-machined's OpenMachineShell() bus call will now pass
6281 additional policy metadata data fields to the PolicyKit
6282 authentication request.
6284 * systemd-tmpfiles gained a new -E switch, which is equivalent to
6285 --exclude-prefix=/dev --exclude-prefix=/proc --exclude=/run
6286 --exclude=/sys. It's particularly useful in combination with --root=,
6287 when operating on OS trees that do not have any of these four runtime
6288 directories mounted, as this means no files below these subtrees are
6289 created or modified, since those mount points should probably remain
6292 * systemd-tmpfiles gained a new --image= switch which is like --root=,
6293 but takes a disk image instead of a directory as argument. The
6294 specified disk image is mounted inside a temporary mount namespace
6295 and the tmpfiles.d/ drop-ins stored in the image are executed and
6296 applied to the image. systemd-sysusers similarly gained a new
6297 --image= switch, that allows the sysusers.d/ drop-ins stored in the
6298 image to be applied onto the image.
6300 * Similarly, the journalctl command also gained an --image= switch,
6301 which is a quick one-step solution to look at the log data included
6304 * journalctl's --output=cat option (which outputs the log content
6305 without any metadata, just the pure text messages) will now make use
6306 of terminal colors when run on a suitable terminal, similarly to the
6309 * JSON group records now support a "description" string that may be
6310 used to add a human-readable textual description to such groups. This
6311 is supposed to match the user's GECOS field which traditionally
6312 didn't have a counterpart for group records.
6314 * The "systemd-dissect" tool that may be used to inspect OS disk images
6315 and that was previously installed to /usr/lib/systemd/ has now been
6316 moved to /usr/bin/, reflecting its updated status of an officially
6317 supported tool with a stable interface. It gained support for a new
6318 --mkdir switch which when combined with --mount has the effect of
6319 creating the directory to mount the image to if it is missing
6320 first. It also gained two new commands --copy-from and --copy-to for
6321 copying files and directories in and out of an OS image without the
6322 need to manually mount it. It also acquired support for a new option
6323 --json= to generate JSON output when inspecting an OS image.
6325 * The cgroup2 file system is now mounted with the
6326 "memory_recursiveprot" mount option, supported since kernel 5.7. This
6327 means that the MemoryLow= and MemoryMin= unit file settings now apply
6328 recursively to whole subtrees.
6330 * systemd-homed now defaults to using the btrfs file system — if
6331 available — when creating home directories in LUKS volumes. This may
6332 be changed with the DefaultFileSystemType= setting in homed.conf.
6333 It's now the default file system in various major distributions and
6334 has the major benefit for homed that it can be grown and shrunk while
6335 mounted, unlike the other contenders ext4 and xfs, which can both be
6336 grown online, but not shrunk (in fact xfs is the technically most
6337 limited option here, as it cannot be shrunk at all).
6339 * JSON user records managed by systemd-homed gained support for
6340 "recovery keys". These are basically secondary passphrases that can
6341 unlock user accounts/home directories. They are computer-generated
6342 rather than user-chosen, and typically have greater entropy.
6343 homectl's --recovery-key= option may be used to add a recovery key to
6344 a user account. The generated recovery key is displayed as a QR code,
6345 so that it can be scanned to be kept in a safe place. This feature is
6346 particularly useful in combination with systemd-homed's support for
6347 FIDO2 or PKCS#11 authentication, as a secure fallback in case the
6348 security tokens are lost. Recovery keys may be entered wherever the
6349 system asks for a password.
6351 * systemd-homed now maintains a "dirty" flag for each LUKS encrypted
6352 home directory which indicates that a home directory has not been
6353 deactivated cleanly when offline. This flag is useful to identify
6354 home directories for which the offline discard logic did not run when
6355 offlining, and where it would be a good idea to log in again to catch
6358 * systemctl gained a new parameter --timestamp= which may be used to
6359 change the style in which timestamps are output, i.e. whether to show
6360 them in local timezone or UTC, or whether to show µs granularity.
6362 * Alibaba's "pouch" container manager is now detected by
6363 systemd-detect-virt, ConditionVirtualization= and similar
6364 constructs. Similar, they now also recognize IBM PowerVM machine
6367 * systemd-nspawn has been reworked to use the /run/host/incoming/ as
6368 place to use for propagating external mounts into the
6369 container. Similarly /run/host/notify is now used as the socket path
6370 for container payloads to communicate with the container manager
6371 using sd_notify(). The container manager now uses the
6372 /run/host/inaccessible/ directory to place "inaccessible" file nodes
6373 of all relevant types which may be used by the container payload as
6374 bind mount source to over-mount inodes to make them inaccessible.
6375 /run/host/container-manager will now be initialized with the same
6376 string as the $container environment variable passed to the
6377 container's PID 1. /run/host/container-uuid will be initialized with
6378 the same string as $container_uuid. This means the /run/host/
6379 hierarchy is now the primary way to make host resources available to
6380 the container. The Container Interface documents these new files and
6383 https://systemd.io/CONTAINER_INTERFACE
6385 * Support for the "ConditionNull=" unit file condition has been
6386 deprecated and undocumented for 6 years. systemd started to warn
6387 about its use 1.5 years ago. It has now been removed entirely.
6389 * sd-bus.h gained a new API call sd_bus_error_has_names(), which takes
6390 a sd_bus_error struct and a list of error names, and checks if the
6391 error matches one of these names. It's a convenience wrapper that is
6392 useful in cases where multiple errors shall be handled the same way.
6394 * A new system call filter list "@known" has been added, that contains
6395 all system calls known at the time systemd was built.
6397 * Behaviour of system call filter allow lists has changed slightly:
6398 system calls that are contained in @known will result in EPERM by
6399 default, while those not contained in it result in ENOSYS. This
6400 should improve compatibility because known system calls will thus be
6401 communicated as prohibited, while unknown (and thus newer ones) will
6402 be communicated as not implemented, which hopefully has the greatest
6403 chance of triggering the right fallback code paths in client
6406 * "systemd-analyze syscall-filter" will now show two separate sections
6407 at the bottom of the output: system calls known during systemd build
6408 time but not included in any of the filter groups shown above, and
6409 system calls defined on the local kernel but known during systemd
6412 * If the $SYSTEMD_LOG_SECCOMP=1 environment variable is set for
6413 systemd-nspawn all system call filter violations will be logged by
6414 the kernel (audit). This is useful for tracking down system calls
6415 invoked by container payloads that are prohibited by the container's
6416 system call filter policy.
6418 * If the $SYSTEMD_SECCOMP=0 environment variable is set for
6419 systemd-nspawn (and other programs that use seccomp) all seccomp
6420 filtering is turned off.
6422 * Two new unit file settings ProtectProc= and ProcSubset= have been
6423 added that expose the hidepid= and subset= mount options of procfs.
6424 All processes of the unit will only see processes in /proc that are
6425 are owned by the unit's user. This is an important new sandboxing
6426 option that is recommended to be set on all system services. All
6427 long-running system services that are included in systemd itself set
6428 this option now. This option is only supported on kernel 5.8 and
6429 above, since the hidepid= option supported on older kernels was not a
6430 per-mount option but actually applied to the whole PID namespace.
6432 * Socket units gained a new boolean setting FlushPending=. If enabled
6433 all pending socket data/connections are flushed whenever the socket
6434 unit enters the "listening" state, i.e. after the associated service
6437 * The unit file setting NUMAMask= gained a new "all" value: when used,
6438 all existing NUMA nodes are added to the NUMA mask.
6440 * A new "credentials" logic has been added to system services. This is
6441 a simple mechanism to pass privileged data to services in a safe and
6442 secure way. It's supposed to be used to pass per-service secret data
6443 such as passwords or cryptographic keys but also associated less
6444 private information such as user names, certificates, and similar to
6445 system services. Each credential is identified by a short user-chosen
6446 name and may contain arbitrary binary data. Two new unit file
6447 settings have been added: SetCredential= and LoadCredential=. The
6448 former allows setting a credential to a literal string, the latter
6449 sets a credential to the contents of a file (or data read from a
6450 user-chosen AF_UNIX stream socket). Credentials are passed to the
6451 service via a special credentials directory, one file for each
6452 credential. The path to the credentials directory is passed in a new
6453 $CREDENTIALS_DIRECTORY environment variable. Since the credentials
6454 are passed in the file system they may be easily referenced in
6455 ExecStart= command lines too, thus no explicit support for the
6456 credentials logic in daemons is required (though ideally daemons
6457 would look for the bits they need in $CREDENTIALS_DIRECTORY
6458 themselves automatically, if set). The $CREDENTIALS_DIRECTORY is
6459 backed by unswappable memory if privileges allow it, immutable if
6460 privileges allow it, is accessible only to the service's UID, and is
6461 automatically destroyed when the service stops.
6463 * systemd-nspawn supports the same credentials logic. It can both
6464 consume credentials passed to it via the aforementioned
6465 $CREDENTIALS_DIRECTORY protocol as well as pass these credentials on
6466 to its payload. The service manager/PID 1 has been updated to match
6467 this: it can also accept credentials from the container manager that
6468 invokes it (in fact: any process that invokes it), and passes them on
6469 to its services. Thus, credentials can be propagated recursively down
6470 the tree: from a system's service manager to a systemd-nspawn
6471 service, to the service manager that runs as container payload and to
6472 the service it runs below. Credentials may also be added on the
6473 systemd-nspawn command line, using new --set-credential= and
6474 --load-credential= command line switches that match the
6475 aforementioned service settings.
6477 * systemd-repart gained new settings Format=, Encrypt=, CopyFiles= in
6478 the partition drop-ins which may be used to format/LUKS
6479 encrypt/populate any created partitions. The partitions are
6480 encrypted/formatted/populated before they are registered in the
6481 partition table, so that they appear atomically: either the
6482 partitions do not exist yet or they exist fully encrypted, formatted,
6483 and populated — there is no time window where they are
6484 "half-initialized". Thus the system is robust to abrupt shutdown: if
6485 the tool is terminated half-way during its operations on next boot it
6486 will start from the beginning.
6488 * systemd-repart's --size= operation gained a new "auto" value. If
6489 specified, and operating on a loopback file it is automatically sized
6490 to the minimal size the size constraints permit. This is useful to
6491 use "systemd-repart" as an image builder for minimally sized images.
6493 * systemd-resolved now gained a third IPC interface for requesting name
6494 resolution: besides D-Bus and local DNS to 127.0.0.53 a Varlink
6495 interface is now supported. The nss-resolve NSS module has been
6496 modified to use this new interface instead of D-Bus. Using Varlink
6497 has a major benefit over D-Bus: it works without a broker service,
6498 and thus already during earliest boot, before the dbus daemon has
6499 been started. This means name resolution via systemd-resolved now
6500 works at the same time systemd-networkd operates: from earliest boot
6501 on, including in the initrd.
6503 * systemd-resolved gained support for a new DNSStubListenerExtra=
6504 configuration file setting which may be used to specify additional IP
6505 addresses the built-in DNS stub shall listen on, in addition to the
6506 main one on 127.0.0.53:53.
6508 * Name lookups issued via systemd-resolved's D-Bus and Varlink
6509 interfaces (and thus also via glibc NSS if nss-resolve is used) will
6510 now honour a trailing dot in the hostname: if specified the search
6511 path logic is turned off. Thus "resolvectl query foo." is now
6512 equivalent to "resolvectl query --search=off foo.".
6514 * systemd-resolved gained a new D-Bus property "ResolvConfMode" that
6515 exposes how /etc/resolv.conf is currently managed: by resolved (and
6516 in which mode if so) or another subsystem. "resolvctl" will display
6517 this property in its status output.
6519 * The resolv.conf snippets systemd-resolved provides will now set "."
6520 as the search domain if no other search domain is known. This turns
6521 off the derivation of an implicit search domain by nss-dns for the
6522 hostname, when the hostname is set to an FQDN. This change is done to
6523 make nss-dns using resolv.conf provided by systemd-resolved behave
6524 more similarly to nss-resolve.
6526 * systemd-tmpfiles' file "aging" logic (i.e. the automatic clean-up of
6527 /tmp/ and /var/tmp/ based on file timestamps) now looks at the
6528 "birth" time (btime) of a file in addition to the atime, mtime, and
6531 * systemd-analyze gained a new verb "capability" that lists all known
6532 capabilities by the systemd build and by the kernel.
6534 * If a file /usr/lib/clock-epoch exists, PID 1 will read its mtime and
6535 advance the system clock to it at boot if it is noticed to be before
6536 that time. Previously, PID 1 would only advance the time to an epoch
6537 time that is set during build-time. With this new file OS builders
6538 can change this epoch timestamp on individual OS images without
6539 having to rebuild systemd.
6541 * systemd-logind will now listen to the KEY_RESTART key from the Linux
6542 input layer and reboot the system if it is pressed, similarly to how
6543 it already handles KEY_POWER, KEY_SUSPEND or KEY_SLEEP. KEY_RESTART
6544 was originally defined in the Multimedia context (to restart playback
6545 of a song or film), but is now primarily used in various embedded
6546 devices for "Reboot" buttons. Accordingly, systemd-logind will now
6547 honour it as such. This may configured in more detail via the new
6548 HandleRebootKey= and RebootKeyIgnoreInhibited=.
6550 * systemd-nspawn/systemd-machined will now reconstruct hardlinks when
6551 copying OS trees, for example in "systemd-nspawn --ephemeral",
6552 "systemd-nspawn --template=", "machinectl clone" and similar. This is
6553 useful when operating with OSTree images, which use hardlinks heavily
6554 throughout, and where such copies previously resulting in "exploding"
6557 * systemd-nspawn's --console= setting gained support for a new
6558 "autopipe" value, which is identical to "interactive" when invoked on
6559 a TTY, and "pipe" otherwise.
6561 * systemd-networkd's .network files gained support for explicitly
6562 configuring the multicast membership entries of bridge devices in the
6563 [BridgeMDB] section. It also gained support for the PIE queuing
6564 discipline in the [FlowQueuePIE] sections.
6566 * systemd-networkd's .netdev files may now be used to create "BareUDP"
6567 tunnels, configured in the new [BareUDP] setting.
6569 * systemd-networkd's Gateway= setting in .network files now accepts the
6570 special values "_dhcp4" and "_ipv6ra" to configure additional,
6571 locally defined, explicit routes to the gateway acquired via DHCP or
6572 IPv6 Router Advertisements. The old setting "_dhcp" is deprecated,
6573 but still accepted for backwards compatibility.
6575 * systemd-networkd's [IPv6PrefixDelegation] section and
6576 IPv6PrefixDelegation= options have been renamed as [IPv6SendRA] and
6577 IPv6SendRA= (the old names are still accepted for backwards
6580 * systemd-networkd's .network files gained the DHCPv6PrefixDelegation=
6581 boolean setting in [Network] section. If enabled, the delegated prefix
6582 gained by another link will be configured, and an address within the
6583 prefix will be assigned.
6585 * systemd-networkd's .network files gained the Announce= boolean setting
6586 in [DHCPv6PrefixDelegation] section. When enabled, the delegated
6587 prefix will be announced through IPv6 router advertisement (IPv6 RA).
6588 The setting is enabled by default.
6590 * VXLAN tunnels may now be marked as independent of any underlying
6591 network interface via the new Independent= boolean setting.
6593 * systemctl gained support for two new verbs: "service-log-level" and
6594 "service-log-target" may be used on services that implement the
6595 generic org.freedesktop.LogControl1 D-Bus interface to dynamically
6596 adjust the log level and target. All of systemd's long-running
6597 services support this now, but ideally all system services would
6598 implement this interface to make the system more uniformly
6601 * The SystemCallErrorNumber= unit file setting now accepts the new
6602 "kill" and "log" actions, in addition to arbitrary error number
6603 specifications as before. If "kill" the processes are killed on the
6604 event, if "log" the offending system call is audit logged.
6606 * A new SystemCallLog= unit file setting has been added that accepts a
6607 list of system calls that shall be logged about (audit).
6609 * The OS image dissection logic (as used by RootImage= in unit files or
6610 systemd-nspawn's --image= switch) has gained support for identifying
6611 and mounting explicit /usr/ partitions, which are now defined in the
6612 discoverable partition specification. This should be useful for
6613 environments where the root file system is
6614 generated/formatted/populated dynamically on first boot and combined
6615 with an immutable /usr/ tree that is supplied by the vendor.
6617 * In the final phase of shutdown, within the systemd-shutdown binary
6618 we'll now try to detach MD devices (i.e software RAID) in addition to
6619 loopback block devices and DM devices as before. This is supposed to
6620 be a safety net only, in order to increase robustness if things go
6621 wrong. Storage subsystems are expected to properly detach their
6622 storage volumes during regular shutdown already (or in case of
6623 storage backing the root file system: in the initrd hook we return to
6626 * If the SYSTEMD_LOG_TID environment variable is set all systemd tools
6627 will now log the thread ID in their log output. This is useful when
6628 working with heavily threaded programs.
6630 * If the SYSTEMD_RDRAND environment variable is set to "0", systemd will
6631 not use the RDRAND CPU instruction. This is useful in environments
6632 such as replay debuggers where non-deterministic behaviour is not
6635 * The autopaging logic in systemd's various tools (such as systemctl)
6636 has been updated to turn on "secure" mode in "less"
6637 (i.e. $LESSECURE=1) if execution in a "sudo" environment is
6638 detected. This disables invoking external programs from the pager,
6639 via the pipe logic. This behaviour may be overridden via the new
6640 $SYSTEMD_PAGERSECURE environment variable.
6642 * Units which have resource limits (.service, .mount, .swap, .slice,
6643 .socket, and .slice) gained new configuration settings
6644 ManagedOOMSwap=, ManagedOOMMemoryPressure=, and
6645 ManagedOOMMemoryPressureLimitPercent= that specify resource pressure
6646 limits and optional action taken by systemd-oomd.
6648 * A new service systemd-oomd has been added. It monitors resource
6649 contention for selected parts of the unit hierarchy using the PSI
6650 information reported by the kernel, and kills processes when memory
6651 or swap pressure is above configured limits. This service is only
6652 enabled by default in developer mode (see below) and should be
6653 considered a preview in this release. Behaviour details and option
6654 names are subject to change without the usual backwards-compatibility
6657 * A new helper oomctl has been added to introspect systemd-oomd state.
6658 It is only enabled by default in developer mode and should be
6659 considered a preview without the usual backwards-compatibility
6662 * New meson option -Dcompat-mutable-uid-boundaries= has been added. If
6663 enabled, systemd reads the system UID boundaries from /etc/login.defs
6664 at runtime, instead of using the built-in values selected during
6665 build. This is an option to improve compatibility for upgrades from
6666 old systems. It's strongly recommended not to make use of this
6667 functionality on new systems (or even enable it during build), as it
6668 makes something runtime-configurable that is mostly an implementation
6669 detail of the OS, and permits avoidable differences in deployments
6670 that create all kinds of problems in the long run.
6672 * New meson option '-Dmode=developer|release' has been added. When
6673 'developer', additional checks and features are enabled that are
6674 relevant during upstream development, e.g. verification that
6675 semi-automatically-generated documentation has been properly updated
6676 following API changes. Those checks are considered hints for
6677 developers and are not actionable in downstream builds. In addition,
6678 extra features that are not ready for general consumption may be
6679 enabled in developer mode. It is thus recommended to set
6680 '-Dmode=release' in end-user and distro builds.
6682 * systemd-cryptsetup gained support for processing detached LUKS
6683 headers specified on the kernel command line via the header=
6684 parameter of the luks.options= kernel command line option. The same
6685 device/path syntax as for key files is supported for header files
6688 * The "net_id" built-in of udev has been updated to ignore ACPI _SUN
6689 slot index data for devices that are connected through a PCI bridge
6690 where the _SUN index is associated with the bridge instead of the
6691 network device itself. Previously this would create ambiguous device
6692 naming if multiple network interfaces were connected to the same PCI
6693 bridge. Since this is a naming scheme incompatibility on systems that
6694 possess hardware like this it has been introduced as new naming
6695 scheme "v247". The previous scheme can be selected via the
6696 "net.naming_scheme=v245" kernel command line parameter.
6698 * ConditionFirstBoot= semantics have been modified to be safe towards
6699 abnormal system power-off during first boot. Specifically, the
6700 "systemd-machine-id-commit.service" service now acts as boot
6701 milestone indicating when the first boot process is sufficiently
6702 complete in order to not consider the next following boot also a
6703 first boot. If the system is reset before this unit is reached the
6704 first time, the next boot will still be considered a first boot; once
6705 it has been reached, no further boots will be considered a first
6706 boot. The "first-boot-complete.target" unit now acts as official hook
6707 point to order against this. If a service shall be run on every boot
6708 until the first boot fully succeeds it may thus be ordered before
6709 this target unit (and pull it in) and carry ConditionFirstBoot=
6712 * bootctl's set-default and set-oneshot commands now accept the three
6713 special strings "@default", "@oneshot", "@current" in place of a boot
6714 entry id. These strings are resolved to the current default and
6715 oneshot boot loader entry, as well as the currently booted one. Thus
6716 a command "bootctl set-default @current" may be used to make the
6717 currently boot menu item the new default for all subsequent boots.
6719 * "systemctl edit" has been updated to show the original effective unit
6720 contents in commented form in the text editor.
6722 * Units in user mode are now segregated into three new slices:
6723 session.slice (units that form the core of graphical session),
6724 app.slice ("normal" user applications), and background.slice
6725 (low-priority tasks). Unless otherwise configured, user units are
6726 placed in app.slice. The plan is to add resource limits and
6727 protections for the different slices in the future.
6729 * New GPT partition types for RISCV32/64 for the root and /usr
6730 partitions, and their associated Verity partitions have been defined,
6731 and are now understood by systemd-gpt-auto-generator, and the OS
6732 image dissection logic.
6734 Contributions from: Adolfo Jayme Barrientos, afg, Alec Moskvin, Alyssa
6735 Ross, Amitanand Chikorde, Andrew Hangsleben, Anita Zhang, Ansgar
6736 Burchardt, Arian van Putten, Aurelien Jarno, Axel Rasmussen, bauen1,
6737 Beniamino Galvani, Benjamin Berg, Bjørn Mork, brainrom, Chandradeep
6738 Dey, Charles Lee, Chris Down, Christian Göttsche, Christof Efkemann,
6739 Christoph Ruegge, Clemens Gruber, Daan De Meyer, Daniele Medri, Daniel
6740 Mack, Daniel Rusek, Dan Streetman, David Tardon, Dimitri John Ledkov,
6741 Dmitry Borodaenko, Elias Probst, Elisei Roca, ErrantSpore, Etienne
6742 Doms, Fabrice Fontaine, fangxiuning, Felix Riemann, Florian Klink,
6743 Franck Bui, Frantisek Sumsal, fwSmit, George Rawlinson, germanztz,
6744 Gibeom Gwon, Glen Whitney, Gogo Gogsi, Göran Uddeborg, Grant Mathews,
6745 Hans de Goede, Hans Ulrich Niedermann, Haochen Tong, Harald Seiler,
6746 huangyong, Hubert Kario, igo95862, Ikey Doherty, Insun Pyo, Jan Chren,
6747 Jan Schlüter, Jérémy Nouhaud, Jian-Hong Pan, Joerg Behrmann, Jonathan
6748 Lebon, Jörg Thalheim, Josh Brobst, Juergen Hoetzel, Julien Humbert,
6749 Kai-Chuan Hsieh, Kairui Song, Kamil Dudka, Kir Kolyshkin, Kristijan
6750 Gjoshev, Kyle Huey, Kyle Russell, Lee Whalen, Lennart Poettering,
6751 lichangze, Luca Boccassi, Lucas Werkmeister, Luca Weiss, Marc
6752 Kleine-Budde, Marco Wang, Martin Wilck, Marti Raudsepp, masmullin2000,
6753 Máté Pozsgay, Matt Fenwick, Michael Biebl, Michael Scherer, Michal
6754 Koutný, Michal Sekletár, Michal Suchanek, Mikael Szreder, Milo
6755 Casagrande, mirabilos, Mitsuha_QuQ, mog422, Muhammet Kara, Nazar
6756 Vinnichuk, Nicholas Narsing, Nicolas Fella, Njibhu, nl6720, Oğuz Ersen,
6757 Olivier Le Moal, Ondrej Kozina, onlybugreports, Pass Automated Testing
6758 Suite, Pat Coulthard, Pavel Sapezhko, Pedro Ruiz, perry_yuan, Peter
6759 Hutterer, Phaedrus Leeds, PhoenixDiscord, Piotr Drąg, Plan C,
6760 Purushottam choudhary, Rasmus Villemoes, Renaud Métrich, Robert Marko,
6761 Roman Beranek, Ronan Pigott, Roy Chen (陳彥廷), RussianNeuroMancer,
6762 Samanta Navarro, Samuel BF, scootergrisen, Sorin Ionescu, Steve Dodd,
6763 Susant Sahani, Timo Rothenpieler, Tobias Hunger, Tobias Kaufmann, Topi
6764 Miettinen, vanou, Vito Caputo, Weblate, Wen Yang, Whired Planck,
6765 williamvds, Yu, Li-Yu, Yuri Chornoivan, Yu Watanabe, Zbigniew
6766 Jędrzejewski-Szmek, Zmicer Turok, Дамјан Георгиевски
6768 – Warsaw, 2020-11-26
6772 * The service manager gained basic support for cgroup v2 freezer. Units
6773 can now be suspended or resumed either using new systemctl verbs,
6774 freeze and thaw respectively, or via D-Bus.
6776 * PID 1 may now automatically load pre-compiled AppArmor policies from
6777 /etc/apparmor/earlypolicy during early boot.
6779 * The CPUAffinity= setting in service unit files now supports a new
6780 special value "numa" that causes the CPU affinity masked to be set
6781 based on the NUMA mask.
6783 * systemd will now log about all left-over processes remaining in a
6784 unit when the unit is stopped. It will now warn about services using
6785 KillMode=none, as this is generally an unsafe thing to make use of.
6787 * Two new unit file settings
6788 ConditionPathIsEncrypted=/AssertPathIsEncrypted= have been
6789 added. They may be used to check whether a specific file system path
6790 resides on a block device that is encrypted on the block level
6791 (i.e. using dm-crypt/LUKS).
6793 * Another pair of new settings ConditionEnvironment=/AssertEnvironment=
6794 has been added that may be used for simple environment checks. This
6795 is particularly useful when passing in environment variables from a
6796 container manager (or from PAM in case of the systemd --user
6799 * .service unit files now accept a new setting CoredumpFilter= which
6800 allows configuration of the memory sections coredumps of the
6801 service's processes shall include.
6803 * .mount units gained a new ReadWriteOnly= boolean option. If set
6804 it will not be attempted to mount a file system read-only if mounting
6805 in read-write mode doesn't succeed. An option x-systemd.rw-only is
6806 available in /etc/fstab to control the same.
6808 * .socket units gained a new boolean setting PassPacketInfo=. If
6809 enabled, the kernel will attach additional per-packet metadata to all
6810 packets read from the socket, as an ancillary message. This controls
6811 the IP_PKTINFO, IPV6_RECVPKTINFO, NETLINK_PKTINFO socket options,
6812 depending on socket type.
6814 * .service units gained a new setting RootHash= which may be used to
6815 specify the root hash for verity enabled disk images which are
6816 specified in RootImage=. RootVerity= may be used to specify a path to
6817 the Verity data matching a RootImage= file system. (The latter is
6818 only useful for images that do not contain the Verity data embedded
6819 into the same image that carries a GPT partition table following the
6820 Discoverable Partition Specification). Similarly, systemd-nspawn
6821 gained a new switch --verity-data= that takes a path to a file with
6822 the verity data of the disk image supplied in --image=, if the image
6823 doesn't contain the verity data itself.
6825 * .service units gained a new setting RootHashSignature= which takes
6826 either a base64 encoded PKCS#7 signature of the root hash specified
6827 with RootHash=, or a path to a file to read the signature from. This
6828 allows validation of the root hash against public keys available in
6829 the kernel keyring, and is only supported on recent kernels
6830 (>= 5.4)/libcryptsetup (>= 2.30). A similar switch has been added to
6831 systemd-nspawn and systemd-dissect (--root-hash-sig=). Support for
6832 this mechanism has also been added to systemd-veritysetup.
6834 * .service unit files gained two new options
6835 TimeoutStartFailureMode=/TimeoutStopFailureMode= that may be used to
6836 tune behaviour if a start or stop timeout is hit, i.e. whether to
6837 terminate the service with SIGTERM, SIGABRT or SIGKILL.
6839 * Most options in systemd that accept hexadecimal values prefixed with
6840 0x in additional to the usual decimal notation now also support octal
6841 notation when the 0o prefix is used and binary notation if the 0b
6844 * Various command line parameters and configuration file settings that
6845 configure key or certificate files now optionally take paths to
6846 AF_UNIX sockets in the file system. If configured that way a stream
6847 connection is made to the socket and the required data read from
6848 it. This is a simple and natural extension to the existing regular
6849 file logic, and permits other software to provide keys or
6850 certificates via simple IPC services, for example when unencrypted
6851 storage on disk is not desired. Specifically, systemd-networkd's
6852 Wireguard and MACSEC key file settings as well as
6853 systemd-journal-gatewayd's and systemd-journal-remote's PEM
6854 key/certificate parameters support this now.
6856 * Unit files, tmpfiles.d/ snippets, sysusers.d/ snippets and other
6857 configuration files that support specifier expansion learnt six new
6858 specifiers: %a resolves to the current architecture, %o/%w/%B/%W
6859 resolve to the various ID fields from /etc/os-release, %l resolves to
6860 the "short" hostname of the system, i.e. the hostname configured in
6861 the kernel truncated at the first dot.
6863 * Support for the .include syntax in unit files has been removed. The
6864 concept has been obsolete for 6 years and we started warning about
6865 its pending removal 2 years ago (also see NEWS file below). It's
6868 * StandardError= and StandardOutput= in unit files no longer support
6869 the "syslog" and "syslog-console" switches. They were long removed
6870 from the documentation, but will now result in warnings when used,
6871 and be converted to "journal" and "journal+console" automatically.
6873 * If the service setting User= is set to the "nobody" user, a warning
6874 message is now written to the logs (but the value is nonetheless
6875 accepted). Setting User=nobody is unsafe, since the primary purpose
6876 of the "nobody" user is to own all files whose owner cannot be mapped
6877 locally. It's in particular used by the NFS subsystem and in user
6878 namespacing. By running a service under this user's UID it might get
6879 read and even write access to all these otherwise unmappable files,
6880 which is quite likely a major security problem.
6882 * tmpfs mounts automatically created by systemd (/tmp, /run, /dev/shm,
6883 and others) now have a size and inode limits applied (50% of RAM for
6884 /tmp and /dev/shm, 10% of RAM for other mounts, etc.). Please note
6885 that the implicit kernel default is 50% too, so there is no change
6886 in the size limit for /tmp and /dev/shm.
6888 * nss-mymachines lost support for resolution of users and groups, and
6889 now only does resolution of hostnames. This functionality is now
6890 provided by nss-systemd. Thus, the 'mymachines' entry should be
6891 removed from the 'passwd:' and 'group:' lines in /etc/nsswitch.conf
6892 (and 'systemd' added if it is not already there).
6894 * A new kernel command line option systemd.hostname= has been added
6895 that allows controlling the hostname that is initialized early during
6898 * A kernel command line option "udev.blockdev_read_only" has been
6899 added. If specified all hardware block devices that show up are
6900 immediately marked as read-only by udev. This option is useful for
6901 making sure that a specific boot under no circumstances modifies data
6902 on disk. Use "blockdev --setrw" to undo the effect of this, per
6905 * A new boolean kernel command line option systemd.swap= has been
6906 added, which may be used to turn off automatic activation of swap
6907 devices listed in /etc/fstab.
6909 * New kernel command line options systemd.condition_needs_update= and
6910 systemd.condition_first_boot= have been added, which override the
6911 result of the ConditionNeedsUpdate= and ConditionFirstBoot=
6914 * A new kernel command line option systemd.clock_usec= has been added
6915 that allows setting the system clock to the specified time in µs
6916 since Jan 1st, 1970 early during boot. This is in particular useful
6917 in order to make test cases more reliable.
6919 * The fs.suid_dumpable sysctl is set to 2 / "suidsafe". This allows
6920 systemd-coredump to save core files for suid processes. When saving
6921 the core file, systemd-coredump will use the effective uid and gid of
6922 the process that faulted.
6924 * The /sys/module/kernel/parameters/crash_kexec_post_notifiers file is
6925 now automatically set to "Y" at boot, in order to enable pstore
6926 generation for collection with systemd-pstore.
6928 * We provide a set of udev rules to enable auto-suspend on PCI and USB
6929 devices that were tested to correctly support it. Previously, this
6930 was distributed as a set of udev rules, but has now been replaced by
6931 by a set of hwdb entries (and a much shorter udev rule to take action
6932 if the device modalias matches one of the new hwdb entries).
6934 As before, entries are periodically imported from the database
6935 maintained by the ChromiumOS project. If you have a device that
6936 supports auto-suspend correctly and where it should be enabled by
6937 default, please submit a patch that adds it to the database (see
6938 /usr/lib/udev/hwdb.d/60-autosuspend.hwdb).
6940 * systemd-udevd gained the new configuration option timeout_signal= as well
6941 as a corresponding kernel command line option udev.timeout_signal=.
6942 The option can be used to configure the UNIX signal that the main
6943 daemon sends to the worker processes on timeout. Setting the signal
6944 to SIGABRT is useful for debugging.
6946 * .link files managed by systemd-udevd gained options RxFlowControl=,
6947 TxFlowControl=, AutoNegotiationFlowControl= in the [Link] section, in
6948 order to configure various flow control parameters. They also gained
6949 RxMiniBufferSize= and RxJumboBufferSize= in order to configure jumbo
6950 frame ring buffer sizes.
6952 * networkd.conf gained a new boolean setting ManageForeignRoutes=. If
6953 enabled systemd-networkd manages all routes configured by other tools.
6955 * .network files managed by systemd-networkd gained a new section
6956 [SR-IOV], in order to configure SR-IOV capable network devices.
6958 * systemd-networkd's [IPv6Prefix] section in .network files gained a
6959 new boolean setting Assign=. If enabled an address from the prefix is
6960 automatically assigned to the interface.
6962 * systemd-networkd gained a new section [DHCPv6PrefixDelegation] which
6963 controls delegated prefixes assigned by DHCPv6 client. The section
6964 has three settings: SubnetID=, Assign=, and Token=. The setting
6965 SubnetID= allows explicit configuration of the preferred subnet that
6966 systemd-networkd's Prefix Delegation logic assigns to interfaces. If
6967 Assign= is enabled (which is the default) an address from any acquired
6968 delegated prefix is automatically chosen and assigned to the
6969 interface. The setting Token= specifies an optional address generation
6972 * systemd-networkd's [Network] section gained a new setting
6973 IPv4AcceptLocal=. If enabled the interface accepts packets with local
6976 * systemd-networkd gained support for configuring the HTB queuing
6977 discipline in the [HierarchyTokenBucket] and
6978 [HierarchyTokenBucketClass] sections. Similar the "pfifo" qdisc may
6979 be configured in the [PFIFO] section, "GRED" in
6980 [GenericRandomEarlyDetection], "SFB" in [StochasticFairBlue], "cake"
6981 in [CAKE], "PIE" in [PIE], "DRR" in [DeficitRoundRobinScheduler] and
6982 [DeficitRoundRobinSchedulerClass], "BFIFO" in [BFIFO],
6983 "PFIFOHeadDrop" in [PFIFOHeadDrop], "PFIFOFast" in [PFIFOFast], "HHF"
6984 in [HeavyHitterFilter], "ETS" in [EnhancedTransmissionSelection] and
6985 "QFQ" in [QuickFairQueueing] and [QuickFairQueueingClass].
6987 * systemd-networkd gained support for a new Termination= setting in the
6988 [CAN] section for configuring the termination resistor. It also
6989 gained a new ListenOnly= setting for controlling whether to only
6990 listen on CAN interfaces, without interfering with traffic otherwise
6991 (which is useful for debugging/monitoring CAN network
6992 traffic). DataBitRate=, DataSamplePoint=, FDMode=, FDNonISO= have
6993 been added to configure various CAN-FD aspects.
6995 * systemd-networkd's [DHCPv6] section gained a new option WithoutRA=.
6996 When enabled, DHCPv6 will be attempted right-away without requiring an
6997 Router Advertisement packet suggesting it first (i.e. without the 'M'
6998 or 'O' flags set). The [IPv6AcceptRA] section gained a boolean option
6999 DHCPv6Client= that may be used to turn off the DHCPv6 client even if
7000 the RA packets suggest it.
7002 * systemd-networkd's [DHCPv4] section gained a new setting UseGateway=
7003 which may be used to turn off use of the gateway information provided
7004 by the DHCP lease. A new FallbackLeaseLifetimeSec= setting may be
7005 used to configure how to process leases that lack a lifetime option.
7007 * systemd-networkd's [DHCPv4] and [DHCPServer] sections gained a new
7008 setting SendVendorOption= allowing configuration of additional vendor
7009 options to send in the DHCP requests/responses. The [DHCPv6] section
7010 gained a new SendOption= setting for sending arbitrary DHCP
7011 options. RequestOptions= has been added to request arbitrary options
7012 from the server. UserClass= has been added to set the DHCP user class
7015 * systemd-networkd's [DHCPServer] section gained a new set of options
7016 EmitPOP3=/POP3=, EmitSMTP=/SMTP=, EmitLPR=/LPR= for including server
7017 information about these three protocols in the DHCP lease. It also
7018 gained support for including "MUD" URLs ("Manufacturer Usage
7019 Description"). Support for "MUD" URLs was also added to the LLDP
7020 stack, configurable in the [LLDP] section in .network files.
7022 * The Mode= settings in [MACVLAN] and [MACVTAP] now support 'source'
7023 mode. Also, the sections now support a new setting SourceMACAddress=.
7025 * systemd-networkd's .netdev files now support a new setting
7026 VLANProtocol= in the [Bridge] section that allows configuration of
7027 the VLAN protocol to use.
7029 * systemd-networkd supports a new Group= setting in the [Link] section
7030 of the .network files, to control the link group.
7032 * systemd-networkd's [Network] section gained a new
7033 IPv6LinkLocalAddressGenerationMode= setting, which specifies how IPv6
7034 link local address is generated.
7036 * A new default .network file is now shipped that matches TUN/TAP
7037 devices that begin with "vt-" in their name. Such interfaces will
7038 have IP routing onto the host links set up automatically. This is
7039 supposed to be used by VM managers to trivially acquire a network
7040 interface which is fully set up for host communication, simply by
7041 carefully picking an interface name to use.
7043 * systemd-networkd's [DHCPv6] section gained a new setting RouteMetric=
7044 which sets the route priority for routes specified by the DHCP server.
7046 * systemd-networkd's [DHCPv6] section gained a new setting VendorClass=
7047 which configures the vendor class information sent to DHCP server.
7049 * The BlackList= settings in .network files' [DHCPv4] and
7050 [IPv6AcceptRA] sections have been renamed DenyList=. The old names
7051 are still understood to provide compatibility.
7053 * networkctl gained the new "forcerenew" command for forcing all DHCP
7054 server clients to renew their lease. The interface "status" output
7055 will now show numerous additional fields of information about an
7056 interface. There are new "up" and "down" commands to bring specific
7057 interfaces up or down.
7059 * systemd-resolved's DNS= configuration option now optionally accepts a
7060 port number (after ":") and a host name (after "#"). When the host
7061 name is specified, the DNS-over-TLS certificate is validated to match
7062 the specified hostname. Additionally, in case of IPv6 addresses, an
7063 interface may be specified (after "%").
7065 * systemd-resolved may be configured to forward single-label DNS names.
7066 This is not standard-conformant, but may make sense in setups where
7067 public DNS servers are not used.
7069 * systemd-resolved's DNS-over-TLS support gained SNI validation.
7071 * systemd-nspawn's --resolv-conf= switch gained a number of new
7072 supported values. Specifically, options starting with "replace-" are
7073 like those prefixed "copy-" but replace any existing resolv.conf
7074 file. And options ending in "-uplink" and "-stub" can now be used to
7075 propagate other flavours of resolv.conf into the container (as
7076 defined by systemd-resolved).
7078 * The various programs included in systemd can now optionally output
7079 their log messages on stderr prefixed with a timestamp, controlled by
7080 the $SYSTEMD_LOG_TIME environment variable.
7082 * systemctl gained a new "-P" switch that is a shortcut for "--value
7085 * "systemctl list-units" and "systemctl list-machines" no longer hide
7086 their first output column with --no-legend. To hide the first column,
7089 * "systemctl reboot" takes the option "--reboot-argument=".
7090 The optional positional argument to "systemctl reboot" is now
7091 being deprecated in favor of this option.
7093 * systemd-run gained a new switch --slice-inherit. If specified the
7094 unit it generates is placed in the same slice as the systemd-run
7097 * systemd-journald gained support for zstd compression of large fields
7098 in journal files. The hash tables in journal files have been hardened
7099 against hash collisions. This is an incompatible change and means
7100 that journal files created with new systemd versions are not readable
7101 with old versions. If the $SYSTEMD_JOURNAL_KEYED_HASH boolean
7102 environment variable for systemd-journald.service is set to 0 this
7103 new hardening functionality may be turned off, so that generated
7104 journal files remain compatible with older journalctl
7107 * journalctl will now include a clickable link in the default output for
7108 each log message for which a URL with further documentation is
7109 known. This is only supported on terminal emulators that support
7110 clickable hyperlinks, and is turned off if a pager is used (since
7111 "less" still doesn't support hyperlinks,
7112 unfortunately). Documentation URLs may be included in log messages
7113 either by including a DOCUMENTATION= journal field in it, or by
7114 associating a journal message catalog entry with the log message's
7115 MESSAGE_ID, which then carries a "Documentation:" tag.
7117 * journald.conf gained a new boolean setting Audit= that may be used to
7118 control whether systemd-journald will enable audit during
7121 * when systemd-journald's log stream is broken up into multiple lines
7122 because the PID of the sender changed this is indicated in the
7123 generated log records via the _LINE_BREAK=pid-change field.
7125 * journalctl's "-o cat" output mode will now show one or more journal
7126 fields specified with --output-fields= instead of unconditionally
7127 MESSAGE=. This is useful to retrieve a very specific set of fields
7128 without any decoration.
7130 * The sd-journal.h API gained two new functions:
7131 sd_journal_enumerate_available_unique() and
7132 sd_journal_enumerate_available_data() that operate like their
7133 counterparts that lack the _available_ in the name, but skip items
7134 that cannot be read and processed by the local implementation
7135 (i.e. are compressed in an unsupported format or such),
7137 * coredumpctl gained a new --file= switch, matching the same one in
7138 journalctl: a specific journal file may be specified to read the
7141 * coredumps collected by systemd-coredump may now be compressed using
7144 * systemd-binfmt gained a new switch --unregister for unregistering all
7145 registered entries at once. This is now invoked automatically at
7146 shutdown, so that binary formats registered with the "F" flag will
7147 not block clean file system unmounting.
7149 * systemd-notify's --pid= switch gained new values: "parent", "self",
7150 "auto" for controlling which PID to send to the service manager: the
7151 systemd-notify process' PID, or the one of the process invoking it.
7153 * systemd-logind's Session bus object learnt a new method call
7154 SetType() for temporarily updating the session type of an already
7155 allocated session. This is useful for upgrading tty sessions to
7156 graphical ones once a compositor is invoked.
7158 * systemd-socket-proxy gained a new switch --exit-idle-time= for
7159 configuring an exit-on-idle time.
7161 * systemd-repart's --empty= setting gained a new value "create". If
7162 specified a new empty regular disk image file is created under the
7163 specified name. Its size may be specified with the new --size=
7164 option. The latter is also supported without the "create" mode, in
7165 order to grow existing disk image files to the specified size. These
7166 two new options are useful when creating or manipulating disk images
7167 instead of operating on actual block devices.
7169 * systemd-repart drop-ins now support a new UUID= setting to control
7170 the UUID to assign to a newly created partition.
7172 * systemd-repart's SizeMin= per-partition parameter now defaults to 10M
7175 * systemd-repart's Label= setting now support the usual, simple
7176 specifier expansion.
7178 * systemd-homed's LUKS backend gained the ability to discard empty file
7179 system blocks automatically when the user logs out. This is enabled
7180 by default to ensure that home directories take minimal space when
7181 logged out but get full size guarantees when logged in. This may be
7182 controlled with the new --luks-offline-discard= switch to homectl.
7184 * If systemd-homed detects that /home/ is encrypted as a whole it will
7185 now default to the directory or subvolume backends instead of the
7186 LUKS backend, in order to avoid double encryption. The default
7187 storage and file system may now be configured explicitly, too, via
7188 the new /etc/systemd/homed.conf configuration file.
7190 * systemd-homed now supports unlocking home directories with FIDO2
7191 security tokens that support the 'hmac-secret' extension, in addition
7192 to the existing support for PKCS#11 security token unlocking
7193 support. Note that many recent hardware security tokens support both
7194 interfaces. The FIDO2 support is accessible via homectl's
7195 --fido2-device= option.
7197 * homectl's --pkcs11-uri= setting now accepts two special parameters:
7198 if "auto" is specified and only one suitable PKCS#11 security token
7199 is plugged in, its URL is automatically determined and enrolled for
7200 unlocking the home directory. If "list" is specified a brief table of
7201 suitable PKCS#11 security tokens is shown. Similar, the new
7202 --fido2-device= option also supports these two special values, for
7203 automatically selecting and listing suitable FIDO2 devices.
7205 * The /etc/crypttab tmp option now optionally takes an argument
7206 selecting the file system to use. Moreover, the default is now
7207 changed from ext2 to ext4.
7209 * There's a new /etc/crypttab option "keyfile-erase". If specified the
7210 key file listed in the same line is removed after use, regardless if
7211 volume activation was successful or not. This is useful if the key
7212 file is only acquired transiently at runtime and shall be erased
7213 before the system continues to boot.
7215 * There's also a new /etc/crypttab option "try-empty-password". If
7216 specified, before asking the user for a password it is attempted to
7217 unlock the volume with an empty password. This is useful for
7218 installing encrypted images whose password shall be set on first boot
7219 instead of at installation time.
7221 * systemd-cryptsetup will now attempt to load the keys to unlock
7222 volumes with automatically from files in
7223 /etc/cryptsetup-keys.d/<volume>.key and
7224 /run/cryptsetup-keys.d/<volume>.key, if any of these files exist.
7226 * systemd-cryptsetup may now activate Microsoft BitLocker volumes via
7227 /etc/crypttab, during boot.
7229 * logind.conf gained a new RuntimeDirectoryInodesMax= setting to
7230 control the inode limit for the per-user $XDG_RUNTIME_DIR tmpfs
7233 * A new generator systemd-xdg-autostart-generator has been added. It
7234 generates systemd unit files from XDG autostart .desktop files, and
7235 may be used to let the systemd user instance manage services that are
7236 started automatically as part of the desktop session.
7238 * "bootctl" gained a new verb "reboot-to-firmware" that may be used
7239 to query and change the firmware's 'Reboot Into Firmware Interface'
7242 * systemd-firstboot gained a new switch --kernel-command-line= that may
7243 be used to initialize the /etc/kernel/cmdline file of the image. It
7244 also gained a new switch --root-password-hashed= which is like
7245 --root-password= but accepts a pre-hashed UNIX password as
7246 argument. The new option --delete-root-password may be used to unset
7247 any password for the root user (dangerous!). The --root-shell= switch
7248 may be used to control the shell to use for the root account. A new
7249 --force option may be used to override any already set settings with
7250 the parameters specified on the command line (by default, the tool
7251 will not override what has already been set before, i.e. is purely
7254 * systemd-firstboot gained support for a new --image= switch, which is
7255 similar to --root= but accepts the path to a disk image file, on
7256 which it then operates.
7258 * A new sd-path.h API has been added to libsystemd. It provides a
7259 simple API for retrieving various search paths and primary
7260 directories for various resources.
7262 * A new call sd_notify_barrier() has been added to the sd-daemon.h
7263 API. The call will block until all previously sent sd_notify()
7264 messages have been processed by the service manager. This is useful
7265 to remove races caused by a process already having disappeared at the
7266 time a notification message is processed by the service manager,
7267 making correct attribution impossible. The systemd-notify tool will
7268 now make use of this call implicitly, but this can be turned off again
7269 via the new --no-block switch.
7271 * When sending a file descriptor (fd) to the service manager to keep
7272 track of, using the sd_notify() mechanism, a new parameter FDPOLL=0
7273 may be specified. If passed the service manager will refrain from
7274 poll()ing on the file descriptor. Traditionally (and when the
7275 parameter is not specified), the service manager will poll it for
7276 POLLHUP or POLLERR events, and immediately close the fds in that
7279 * The service manager (PID1) gained a new D-Bus method call
7280 SetShowStatus() which may be used to control whether it shall show
7281 boot-time status output on the console. This method has a similar
7282 effect to sending SIGRTMIN+20/SIGRTMIN+21 to PID 1.
7284 * The sd-bus API gained a number of convenience functions that take
7285 va_list arguments rather than "...". For example, there's now
7286 sd_bus_call_methodv() to match sd_bus_call_method(). Those calls make
7287 it easier to build wrappers that accept variadic arguments and want
7288 to pass a ready va_list structure to sd-bus.
7290 * sd-bus vtable entries can have a new SD_BUS_VTABLE_ABSOLUTE_OFFSET
7291 flag which alters how the userdata pointer to pass to the callbacks
7292 is determined. When the flag is set, the offset field is converted
7293 as-is into a pointer, without adding it to the object pointer the
7294 vtable is associated with.
7296 * sd-bus now exposes four new functions:
7297 sd_bus_interface_name_is_valid() + sd_bus_service_name_is_valid() +
7298 sd_bus_member_name_is_valid() + sd_bus_object_path_is_valid() will
7299 validate strings to check if they qualify as various D-Bus concepts.
7301 * The sd-bus API gained the SD_BUS_METHOD_WITH_ARGS(),
7302 SD_BUS_METHOD_WITH_ARGS_OFFSET() and SD_BUS_SIGNAL_WITH_ARGS() macros
7303 that simplify adding argument names to D-Bus methods and signals.
7305 * The man pages for the sd-bus and sd-hwdb APIs have been completed.
7307 * Various D-Bus APIs of systemd daemons now have man pages that
7308 document the methods, signals and properties.
7310 * The expectations on user/group name syntax are now documented in
7311 detail; documentation on how classic home directories may be
7312 converted into home directories managed by homed has been added;
7313 documentation regarding integration of homed/userdb functionality in
7314 desktops has been added:
7316 https://systemd.io/USER_NAMES
7317 https://systemd.io/CONVERTING_TO_HOMED
7318 https://systemd.io/USERDB_AND_DESKTOPS
7320 * Documentation for the on-disk Journal file format has been updated
7321 and has now moved to:
7323 https://systemd.io/JOURNAL_FILE_FORMAT
7325 * The interface for containers (https://systemd.io/CONTAINER_INTERFACE)
7326 has been extended by a set of environment variables that expose
7327 select fields from the host's os-release file to the container
7328 payload. Similarly, host's os-release files can be mounted into the
7329 container underneath /run/host. Together, those mechanisms provide a
7330 standardized way to expose information about the host to the
7331 container payload. Both interfaces are implemented in systemd-nspawn.
7333 * All D-Bus services shipped in systemd now implement the generic
7334 LogControl1 D-Bus API which allows clients to change log level +
7335 target of the service during runtime.
7337 * Only relevant for developers: the mkosi.default symlink has been
7338 dropped from version control. Please create a symlink to one of the
7339 distribution-specific defaults in .mkosi/ based on your preference.
7341 Contributions from: 24bisquitz, Adam Nielsen, Alan Perry, Alexander
7342 Malafeev, Amitanand.Chikorde, Alin Popa, Alvin Šipraga, Amos Bird,
7343 Andreas Rammhold, AndreRH, Andrew Doran, Anita Zhang, Ankit Jain,
7344 antznin, Arnaud Ferraris, Arthur Moraes do Lago, Arusekk, Balaji
7345 Punnuru, Balint Reczey, Bastien Nocera, bemarek, Benjamin Berg,
7346 Benjamin Dahlhoff, Benjamin Robin, Chris Down, Chris Kerr, Christian
7347 Göttsche, Christian Hesse, Christian Oder, Ciprian Hacman, Clinton Roy,
7348 codicodi, Corey Hinshaw, Daan De Meyer, Dana Olson, Dan Callaghan,
7349 Daniel Fullmer, Daniel Rusek, Dan Streetman, Dave Reisner, David
7350 Edmundson, David Wood, Denis Pronin, Diego Escalante Urrelo, Dimitri
7351 John Ledkov, dolphrundgren, duguxy, Einsler Lee, Elisei Roca, Emmanuel
7352 Garette, Eric Anderson, Eric DeVolder, Evgeny Vereshchagin,
7353 ExtinctFire, fangxiuning, Ferran Pallarès Roca, Filipe Brandenburger,
7354 Filippo Falezza, Finn, Florian Klink, Florian Mayer, Franck Bui,
7355 Frantisek Sumsal, gaurav, Georg Müller, Gergely Polonkai, Giedrius
7356 Statkevičius, Gigadoc2, gogogogi, Gaurav Singh, gzjsgdsb, Hans de
7357 Goede, Haochen Tong, ianhi, ignapk, Jakov Smolic, James T. Lee, Jan
7358 Janssen, Jan Klötzke, Jan Palus, Jay Burger, Jeremy Cline, Jérémy
7359 Rosen, Jian-Hong Pan, Jiri Slaby, Joel Shapiro, Joerg Behrmann, Jörg
7360 Thalheim, Jouke Witteveen, Kai-Heng Feng, Kenny Levinsen, Kevin
7361 Kuehler, Kumar Kartikeya Dwivedi, layderv, laydervus, Lénaïc Huard,
7362 Lennart Poettering, Lidong Zhong, Luca Boccassi, Luca BRUNO, Lucas
7363 Werkmeister, Lukas Klingsbo, Lukáš Nykrýn, Łukasz Stelmach, Maciej
7364 S. Szmigiero, MadMcCrow, Marc-André Lureau, Marcel Holtmann, Marc
7365 Kleine-Budde, Martin Hundebøll, Matthew Leeds, Matt Ranostay, Maxim
7366 Fomin, MaxVerevkin, Michael Biebl, Michael Chapman, Michael Gubbels,
7367 Michael Marley, Michał Bartoszkiewicz, Michal Koutný, Michal Sekletár,
7368 Mike Gilbert, Mike Kazantsev, Mikhail Novosyolov, ml, Motiejus Jakštys,
7369 nabijaczleweli, nerdopolis, Niccolò Maggioni, Niklas Hambüchen, Norbert
7370 Lange, Paul Cercueil, pelzvieh, Peter Hutterer, Piero La Terza, Pieter
7371 Lexis, Piotr Drąg, Rafael Fontenelle, Richard Petri, Ronan Pigott, Ross
7372 Lagerwall, Rubens Figueiredo, satmandu, Sean-StarLabs, Sebastian
7373 Jennen, sterlinghughes, Surhud More, Susant Sahani, szb512, Thomas
7374 Haller, Tobias Hunger, Tom, Tomáš Pospíšek, Tomer Shechner, Tom Hughes,
7375 Topi Miettinen, Tudor Roman, Uwe Kleine-König, Valery0xff, Vito Caputo,
7376 Vladimir Panteleev, Vladyslav Tronko, Wen Yang, Yegor Vialov, Yigal
7377 Korman, Yi Gao, YmrDtnJu, Yuri Chornoivan, Yu Watanabe, Zbigniew
7378 Jędrzejewski-Szmek, Zhu Li, Дамјан Георгиевски, наб
7380 – Warsaw, 2020-07-30
7384 * A new tool "systemd-repart" has been added, that operates as an
7385 idempotent declarative repartitioner for GPT partition tables.
7386 Specifically, a set of partitions that must or may exist can be
7387 configured via drop-in files, and during every boot the partition
7388 table on disk is compared with these files, creating missing
7389 partitions or growing existing ones based on configurable relative
7390 and absolute size constraints. The tool is strictly incremental,
7391 i.e. does not delete, shrink or move partitions, but only adds and
7392 grows them. The primary use-case is OS images that ship in minimized
7393 form, that on first boot are grown to the size of the underlying
7394 block device or augmented with additional partitions. For example,
7395 the root partition could be extended to cover the whole disk, or a
7396 swap or /home partitions could be added on first boot. It can also be
7397 used for systems that use an A/B update scheme but ship images with
7398 just the A partition, with B added on first boot. The tool is
7399 primarily intended to be run in the initrd, shortly before
7400 transitioning into the host OS, but can also be run after the
7401 transition took place. It automatically discovers the disk backing
7402 the root file system, and should hence not require any additional
7403 configuration besides the partition definition drop-ins. If no
7404 configuration drop-ins are present, no action is taken.
7406 * A new component "userdb" has been added, along with a small daemon
7407 "systemd-userdbd.service" and a client tool "userdbctl". The framework
7408 allows defining rich user and group records in a JSON format,
7409 extending on the classic "struct passwd" and "struct group"
7410 structures. Various components in systemd have been updated to
7411 process records in this format, including systemd-logind and
7412 pam-systemd. The user records are intended to be extensible, and
7413 allow setting various resource management, security and runtime
7414 parameters that shall be applied to processes and sessions of the
7415 user as they log in. This facility is intended to allow associating
7416 such metadata directly with user/group records so that they can be
7417 produced, extended and consumed in unified form. We hope that
7418 eventually frameworks such as sssd will generate records this way, so
7419 that for the first time resource management and various other
7420 per-user settings can be configured in LDAP directories and then
7421 provided to systemd (specifically to systemd-logind and pam-system)
7422 to apply on login. For further details see:
7424 https://systemd.io/USER_RECORD
7425 https://systemd.io/GROUP_RECORD
7426 https://systemd.io/USER_GROUP_API
7428 * A small new service systemd-homed.service has been added, that may be
7429 used to securely manage home directories with built-in encryption.
7430 The complete user record data is unified with the home directory,
7431 thus making home directories naturally migratable. Its primary
7432 back-end is based on LUKS volumes, but fscrypt, plain directories,
7433 and other storage schemes are also supported. This solves a couple of
7434 problems we saw with traditional ways to manage home directories, in
7435 particular when it comes to encryption. For further discussion of
7436 this, see the video of Lennart's talk at AllSystemsGo! 2019:
7438 https://media.ccc.de/v/ASG2019-164-reinventing-home-directories
7440 For further details about the format and expectations on home
7441 directories this new daemon makes, see:
7443 https://systemd.io/HOME_DIRECTORY
7445 * systemd-journald is now multi-instantiable. In addition to the main
7446 instance systemd-journald.service there's now a template unit
7447 systemd-journald@.service, with each instance defining a new named
7448 log 'namespace' (whose name is specified via the instance part of the
7449 unit name). A new unit file setting LogNamespace= has been added,
7450 taking such a namespace name, that assigns services to the specified
7451 log namespaces. As each log namespace is serviced by its own
7452 independent journal daemon, this functionality may be used to improve
7453 performance and increase isolation of applications, at the price of
7454 losing global message ordering. Each instance of journald has a
7455 separate set of configuration files, with possibly different disk
7456 usage limitations and other settings.
7458 journalctl now takes a new option --namespace= to show logs from a
7459 specific log namespace. The sd-journal.h API gained
7460 sd_journal_open_namespace() for opening the log stream of a specific
7461 log namespace. systemd-journald also gained the ability to exit on
7462 idle, which is useful in the context of log namespaces, as this means
7463 log daemons for log namespaces can be activated automatically on
7464 demand and will stop automatically when no longer used, minimizing
7467 * When systemd-tmpfiles copies a file tree using the 'C' line type it
7468 will now label every copied file according to the SELinux database.
7470 * When systemd/PID 1 detects it is used in the initrd it will now boot
7471 into initrd.target rather than default.target by default. This should
7472 make it simpler to build initrds with systemd as for many cases the
7473 only difference between a host OS image and an initrd image now is
7474 the presence of the /etc/initrd-release file.
7476 * A new kernel command line option systemd.cpu_affinity= is now
7477 understood. It's equivalent to the CPUAffinity= option in
7478 /etc/systemd/system.conf and allows setting the CPU mask for PID 1
7479 itself and the default for all other processes.
7481 * When systemd/PID 1 is reloaded (with systemctl daemon-reload or
7482 equivalent), the SELinux database is now reloaded, ensuring that
7483 sockets and other file system objects are generated taking the new
7484 database into account.
7486 * systemd/PID 1 accepts a new "systemd.show-status=error" setting, and
7487 "quiet" has been changed to imply that instead of
7488 "systemd.show-status=auto". In this mode, only messages about errors
7489 and significant delays in boot are shown on the console.
7491 * The sd-event.h API gained native support for the new Linux "pidfd"
7492 concept. This permits watching processes using file descriptors
7493 instead of PID numbers, which fixes a number of races and makes
7494 process supervision more robust and efficient. All of systemd's
7495 components will now use pidfds if the kernel supports it for process
7496 watching, with the exception of PID 1 itself, unfortunately. We hope
7497 to move PID 1 to exclusively using pidfds too eventually, but this
7498 requires some more kernel work first. (Background: PID 1 watches
7499 processes using waitid() with the P_ALL flag, and that does not play
7500 together nicely with pidfds yet.)
7502 * Closely related to this, the sd-event.h API gained two new calls
7503 sd_event_source_send_child_signal() (for sending a signal to a
7504 watched process) and sd_event_source_get_child_process_own() (for
7505 marking a process so that it is killed automatically whenever the
7506 event source watching it is freed).
7508 * systemd-networkd gained support for configuring Token Bucket Filter
7509 (TBF) parameters in its qdisc configuration support. Similarly,
7510 support for Stochastic Fairness Queuing (SFQ), Controlled-Delay
7511 Active Queue Management (CoDel), and Fair Queue (FQ) has been added.
7513 * systemd-networkd gained support for Intermediate Functional Block
7514 (IFB) network devices.
7516 * systemd-networkd gained support for configuring multi-path IP routes,
7517 using the new MultiPathRoute= setting in the [Route] section.
7519 * systemd-networkd's DHCPv4 client has been updated to support a new
7520 SendDecline= option. If enabled, duplicate address detection is done
7521 after a DHCP offer is received from the server. If a conflict is
7522 detected, the address is declined. The DHCPv4 client also gained
7523 support for a new RouteMTUBytes= setting that allows to configure the
7524 MTU size to be used for routes generated from DHCPv4 leases.
7526 * The PrefixRoute= setting in systemd-networkd's [Address] section of
7527 .network files has been deprecated, and replaced by AddPrefixRoute=,
7528 with its sense inverted.
7530 * The Gateway= setting of [Route] sections of .network files gained
7531 support for a special new value "_dhcp". If set, the configured
7532 static route uses the gateway host configured via DHCP.
7534 * New User= and SuppressPrefixLength= settings have been implemented
7535 for the [RoutingPolicyRule] section of .network files to configure
7536 source routing based on UID ranges and prefix length, respectively.
7538 * The Type= match property of .link files has been generalized to
7539 always match the device type shown by 'networkctl status', even for
7540 devices where udev does not set DEVTYPE=. This allows e.g. Type=ether
7543 * sd-bus gained a new API call sd_bus_message_sensitive() that marks a
7544 D-Bus message object as "sensitive". Those objects are erased from
7545 memory when they are freed. This concept is intended to be used for
7546 messages that contain security sensitive data. A new flag
7547 SD_BUS_VTABLE_SENSITIVE has been introduced as well to mark methods
7548 in sd-bus vtables, causing any incoming and outgoing messages of
7549 those methods to be implicitly marked as "sensitive".
7551 * sd-bus gained a new API call sd_bus_message_dump() for dumping the
7552 contents of a message (or parts thereof) to standard output for
7555 * systemd-sysusers gained support for creating users with the primary
7556 group named differently than the user.
7558 * systemd-growfs (i.e. the x-systemd.growfs mount option in /etc/fstab)
7559 gained support for growing XFS partitions. Previously it supported
7560 only ext4 and btrfs partitions.
7562 * The support for /etc/crypttab gained a new x-initrd.attach option. If
7563 set, the specified encrypted volume is unlocked already in the
7564 initrd. This concept corresponds to the x-initrd.mount option in
7567 * systemd-cryptsetup gained native support for unlocking encrypted
7568 volumes utilizing PKCS#11 smartcards, i.e. for example to bind
7569 encryption of volumes to YubiKeys. This is exposed in the new
7570 pkcs11-uri= option in /etc/crypttab.
7572 * The /etc/fstab support in systemd now supports two new mount options
7573 x-systemd.{required,wanted}-by=, for explicitly configuring the units
7574 that the specified mount shall be pulled in by, in place of
7575 the usual local-fs.target/remote-fs.target.
7577 * The https://systemd.io/ web site has been relaunched, directly
7578 populated with most of the documentation included in the systemd
7579 repository. systemd also acquired a new logo, thanks to Tobias
7582 * systemd-udevd gained support for managing "alternative" network
7583 interface names, as supported by new Linux kernels. For the first
7584 time this permits assigning multiple (and longer!) names to a network
7585 interface. systemd-udevd will now by default assign the names
7586 generated via all supported naming schemes to each interface. This
7587 may be further tweaked with .link files and the AlternativeName= and
7588 AlternativeNamesPolicy= settings. Other components of systemd have
7589 been updated to support the new alternative names wherever
7590 appropriate. For example, systemd-nspawn will now generate
7591 alternative interface names for the host-facing side of container
7592 veth links based on the full container name without truncation.
7594 * systemd-nspawn interface naming logic has been updated in another way
7595 too: if the main interface name (i.e. as opposed to new-style
7596 "alternative" names) based on the container name is truncated, a
7597 simple hashing scheme is used to give different interface names to
7598 multiple containers whose names all begin with the same prefix. Since
7599 this changes the primary interface names pointing to containers if
7600 truncation happens, the old scheme may still be requested by
7601 selecting an older naming scheme, via the net.naming_scheme= kernel
7602 command line option.
7604 * PrivateUsers= in service files now works in services run by the
7605 systemd --user per-user instance of the service manager.
7607 * A new per-service sandboxing option ProtectClock= has been added that
7608 locks down write access to the system clock. It takes away device
7609 node access to /dev/rtc as well as the system calls that set the
7610 system clock and the CAP_SYS_TIME and CAP_WAKE_ALARM capabilities.
7611 Note that this option does not affect access to auxiliary services
7612 that allow changing the clock, for example access to
7615 * The systemd-id128 tool gained a new "show" verb for listing or
7616 resolving a number of well-known UUIDs/128-bit IDs, currently mostly
7617 GPT partition table types.
7619 * The Discoverable Partitions Specification has been updated to support
7620 /var and /var/tmp partition discovery. Support for this has been
7621 added to systemd-gpt-auto-generator. For details see:
7623 https://systemd.io/DISCOVERABLE_PARTITIONS
7625 * "systemctl list-unit-files" has been updated to show a new column
7626 with the suggested enablement state based on the vendor preset files
7627 for the respective units.
7629 * "systemctl" gained a new option "--with-dependencies". If specified
7630 commands such as "systemctl status" or "systemctl cat" will now show
7631 all specified units along with all units they depend on.
7633 * networkctl gained support for showing per-interface logs in its
7636 * systemd-networkd-wait-online gained support for specifying the maximum
7637 operational state to wait for, and to wait for interfaces to
7640 * The [Match] section of .link and .network files now supports a new
7641 option PermanentMACAddress= which may be used to check against the
7642 permanent MAC address of a network device even if a randomized MAC
7645 * The [TrafficControlQueueingDiscipline] section in .network files has
7646 been renamed to [NetworkEmulator] with the "NetworkEmulator" prefix
7647 dropped from the individual setting names.
7649 * Any .link and .network files that have an empty [Match] section (this
7650 also includes empty and commented-out files) will now be
7651 rejected. systemd-udev and systemd-networkd started warning about
7652 such files in version 243.
7654 * systemd-logind will now validate access to the operation of changing
7655 the virtual terminal via a polkit action. By default, only users
7656 with at least one session on a local VT are granted permission.
7658 * When systemd sets up PAM sessions that invoked service processes
7659 shall run in, the pam_setcred() API is now invoked, thus permitting
7660 PAM modules to set additional credentials for the processes.
7662 * portablectl attach/detach verbs now accept --now and --enable options
7663 to combine attachment with enablement and invocation, or detachment
7664 with stopping and disablement.
7666 * UPGRADE ISSUE: a bug where some jobs were trimmed as redundant was
7667 fixed, which in turn exposed bugs in unit configuration of services
7668 which have Type=oneshot and should only run once, but do not have
7669 RemainAfterExit=yes set. Without RemainAfterExit=yes, a one-shot
7670 service may be started again after exiting successfully, for example
7671 as a dependency in another transaction. Affected services included
7672 some internal systemd services (most notably
7673 systemd-vconsole-setup.service, which was updated to have
7674 RemainAfterExit=yes), and plymouth-start.service. Please ensure that
7675 plymouth has been suitably updated or patched before upgrading to
7676 this systemd release. See
7677 https://bugzilla.redhat.com/show_bug.cgi?id=1807771 for some
7678 additional discussion.
7680 Contributions from: AJ Bagwell, Alin Popa, Andreas Rammhold, Anita
7681 Zhang, Ansgar Burchardt, Antonio Russo, Arian van Putten, Ashley Davis,
7682 Balint Reczey, Bart Willems, Bastien Nocera, Benjamin Dahlhoff, Charles
7683 (Chas) Williams, cheese1, Chris Down, Chris Murphy, Christian Ehrhardt,
7684 Christian Göttsche, cvoinf, Daan De Meyer, Daniele Medri, Daniel Rusek,
7685 Daniel Shahaf, Dann Frazier, Dan Streetman, Dariusz Gadomski, David
7686 Michael, Dimitri John Ledkov, Emmanuel Bourg, Evgeny Vereshchagin,
7687 ezst036, Felipe Sateler, Filipe Brandenburger, Florian Klink, Franck
7688 Bui, Fran Dieguez, Frantisek Sumsal, Greg "GothAck" Miell, Guilhem
7689 Lettron, Guillaume Douézan-Grard, Hans de Goede, HATAYAMA Daisuke, Iain
7690 Lane, James Buren, Jan Alexander Steffens (heftig), Jérémy Rosen, Jin
7691 Park, Jun'ichi Nomura, Kai Krakow, Kevin Kuehler, Kevin P. Fleming,
7692 Lennart Poettering, Leonid Bloch, Leonid Evdokimov, lothrond, Luca
7693 Boccassi, Lukas K, Lynn Kirby, Mario Limonciello, Mark Deneen, Matthew
7694 Leeds, Michael Biebl, Michal Koutný, Michal Sekletár, Mike Auty, Mike
7695 Gilbert, mtron, nabijaczleweli, Naïm Favier, Nate Jones, Norbert Lange,
7696 Oliver Giles, Paul Davey, Paul Menzel, Peter Hutterer, Piotr Drąg, Rafa
7697 Couto, Raphael, rhn, Robert Scheck, Rocka, Romain Naour, Ryan Attard,
7698 Sascha Dewald, Shengjing Zhu, Slava Kardakov, Spencer Michaels, Sylvain
7699 Plantefeve, Stanislav Angelovič, Susant Sahani, Thomas Haller, Thomas
7700 Schmitt, Timo Schlüßler, Timo Wilken, Tobias Bernard, Tobias Klauser,
7701 Tobias Stoeckmann, Topi Miettinen, tsia, WataruMatsuoka, Wieland
7702 Hoffmann, Wilhelm Schuster, Will Fleming, xduugu, Yong Cong Sin, Yuri
7703 Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zeyu
7706 – Warsaw, 2020-03-06
7710 * Support for the cpuset cgroups v2 controller has been added.
7711 Processes may be restricted to specific CPUs using the new
7712 AllowedCPUs= setting, and to specific memory NUMA nodes using the new
7713 AllowedMemoryNodes= setting.
7715 * The signal used in restart jobs (as opposed to e.g. stop jobs) may
7716 now be configured using a new RestartKillSignal= setting. This
7717 allows units which signals to request termination to implement
7718 different behaviour when stopping in preparation for a restart.
7720 * "systemctl clean" may now be used also for socket, mount, and swap
7723 * systemd will also read configuration options from the EFI variable
7724 SystemdOptions. This may be used to configure systemd behaviour when
7725 modifying the kernel command line is inconvenient, but configuration
7726 on disk is read too late, for example for the options related to
7727 cgroup hierarchy setup. 'bootctl systemd-efi-options' may be used to
7728 set the EFI variable.
7730 * systemd will now disable printk ratelimits in early boot. This should
7731 allow us to capture more logs from the early boot phase where normal
7732 storage is not available and the kernel ring buffer is used for
7733 logging. Configuration on the kernel command line has higher priority
7734 and overrides the systemd setting.
7736 systemd programs which log to /dev/kmsg directly use internal
7737 ratelimits to prevent runaway logging. (Normally this is only used
7738 during early boot, so in practice this change has very little
7741 * Unit files now support top level dropin directories of the form
7742 <unit_type>.d/ (e.g. service.d/) that may be used to add configuration
7743 that affects all corresponding unit files.
7745 * systemctl gained support for 'stop --job-mode=triggering' which will
7746 stop the specified unit and any units which could trigger it.
7748 * Unit status display now includes units triggering and triggered by
7749 the unit being shown.
7751 * The RuntimeMaxSec= setting is now supported by scopes, not just
7752 .service units. This is particularly useful for PAM sessions which
7753 create a scope unit for the user login. systemd.runtime_max_sec=
7754 setting may used with the pam_systemd module to limit the duration
7755 of the PAM session, for example for time-limited logins.
7757 * A new @pkey system call group is now defined to make it easier to
7758 allow-list memory protection syscalls for containers and services
7759 which need to use them.
7761 * systemd-udevd: removed the 30s timeout for killing stale workers on
7762 exit. systemd-udevd now waits for workers to finish. The hard-coded
7763 exit timeout of 30s was too short for some large installations, where
7764 driver initialization could be prematurely interrupted during initrd
7765 processing if the root file system had been mounted and init was
7766 preparing to switch root. If udevd is run without systemd and workers
7767 are hanging while udevd receives an exit signal, udevd will now exit
7768 when udev.event_timeout is reached for the last hanging worker. With
7769 systemd, the exit timeout can additionally be configured using
7770 TimeoutStopSec= in systemd-udevd.service.
7772 * udev now provides a program (fido_id) that identifies FIDO CTAP1
7773 ("U2F")/CTAP2 security tokens based on the usage declared in their
7774 report and descriptor and outputs suitable environment variables.
7775 This replaces the externally maintained allow lists of all known
7776 security tokens that were used previously.
7778 * Automatically generated autosuspend udev rules for allow-listed
7779 devices have been imported from the Chromium OS project. This should
7780 improve power saving with many more devices.
7782 * udev gained a new "CONST{key}=value" setting that allows matching
7783 against system-wide constants without forking a helper binary.
7784 Currently "arch" and "virt" keys are supported.
7786 * udev now opens CDROMs in non-exclusive mode when querying their
7787 capabilities. This should fix issues where other programs trying to
7788 use the CDROM cannot gain access to it, but carries a risk of
7789 interfering with programs writing to the disk, if they did not open
7790 the device in exclusive mode as they should.
7792 * systemd-networkd does not create a default route for IPv4 link local
7793 addressing anymore. The creation of the route was unexpected and was
7794 breaking routing in various cases, but people who rely on it being
7795 created implicitly will need to adjust. Such a route may be requested
7796 with DefaultRouteOnDevice=yes.
7798 Similarly, systemd-networkd will not assign a link-local IPv6 address
7799 when IPv6 link-local routing is not enabled.
7801 * Receive and transmit buffers may now be configured on links with
7802 the new RxBufferSize= and TxBufferSize= settings.
7804 * systemd-networkd may now advertise additional IPv6 routes. A new
7805 [IPv6RoutePrefix] section with Route= and LifetimeSec= options is
7808 * systemd-networkd may now configure "next hop" routes using the
7809 [NextHop] section and Gateway= and Id= settings.
7811 * systemd-networkd will now retain DHCP config on restarts by default
7812 (but this may be overridden using the KeepConfiguration= setting).
7813 The default for SendRelease= has been changed to true.
7815 * The DHCPv4 client now uses the OPTION_INFORMATION_REFRESH_TIME option
7816 received from the server.
7818 The client will use the received SIP server list if UseSIP=yes is
7821 The client may be configured to request specific options from the
7822 server using a new RequestOptions= setting.
7824 The client may be configured to send arbitrary options to the server
7825 using a new SendOption= setting.
7827 A new IPServiceType= setting has been added to configure the "IP
7828 service type" value used by the client.
7830 * The DHCPv6 client learnt a new PrefixDelegationHint= option to
7831 request prefix hints in the DHCPv6 solicitation.
7833 * The DHCPv4 server may be configured to send arbitrary options using
7834 a new SendOption= setting.
7836 * The DHCPv4 server may now be configured to emit SIP server list using
7837 the new EmitSIP= and SIP= settings.
7839 * systemd-networkd and networkctl may now renew DHCP leases on demand.
7840 networkctl has a new 'networkctl renew' verb.
7842 * systemd-networkd may now reconfigure links on demand. networkctl
7843 gained two new verbs: "reload" will reload the configuration, and
7844 "reconfigure DEVICE…" will reconfigure one or more devices.
7846 * .network files may now match on SSID and BSSID of a wireless network,
7847 i.e. the access point name and hardware address using the new SSID=
7848 and BSSID= options. networkctl will display the current SSID and
7849 BSSID for wireless links.
7851 .network files may also match on the wireless network type using the
7852 new WLANInterfaceType= option.
7854 * systemd-networkd now includes default configuration that enables
7855 link-local addressing when connected to an ad-hoc wireless network.
7857 * systemd-networkd may configure the Traffic Control queueing
7858 disciplines in the kernel using the new
7859 [TrafficControlQueueingDiscipline] section and Parent=,
7860 NetworkEmulatorDelaySec=, NetworkEmulatorDelayJitterSec=,
7861 NetworkEmulatorPacketLimit=, NetworkEmulatorLossRate=,
7862 NetworkEmulatorDuplicateRate= settings.
7864 * systemd-tmpfiles gained a new w+ setting to append to files.
7866 * systemd-analyze dump will now report when the memory configuration in
7867 the kernel does not match what systemd has configured (usually,
7868 because some external program has modified the kernel configuration
7871 * systemd-analyze gained a new --base-time= switch instructs the
7872 'calendar' verb to resolve times relative to that timestamp instead
7873 of the present time.
7875 * journalctl --update-catalog now produces deterministic output (making
7876 reproducible image builds easier).
7878 * A new devicetree-overlay setting is now documented in the Boot Loader
7881 * The default value of the WatchdogSec= setting used in systemd
7882 services (the ones bundled with the project itself) may be set at
7883 configuration time using the -Dservice-watchdog= setting. If set to
7884 empty, the watchdogs will be disabled.
7886 * systemd-resolved validates IP addresses in certificates now when GnuTLS
7889 * libcryptsetup >= 2.0.1 is now required.
7891 * A configuration option -Duser-path= may be used to override the $PATH
7892 used by the user service manager. The default is again to use the same
7893 path as the system manager.
7895 * The systemd-id128 tool gained a new switch "-u" (or "--uuid") for
7896 outputting the 128-bit IDs in UUID format (i.e. in the "canonical
7899 * Service units gained a new sandboxing option ProtectKernelLogs= which
7900 makes sure the program cannot get direct access to the kernel log
7901 buffer anymore, i.e. the syslog() system call (not to be confused
7902 with the API of the same name in libc, which is not affected), the
7903 /proc/kmsg and /dev/kmsg nodes and the CAP_SYSLOG capability are made
7904 inaccessible to the service. It's recommended to enable this setting
7905 for all services that should not be able to read from or write to the
7906 kernel log buffer, which are probably almost all.
7908 Contributions from: Aaron Plattner, Alcaro, Anita Zhang, Balint Reczey,
7909 Bastien Nocera, Baybal Ni, Benjamin Bouvier, Benjamin Gilbert, Carlo
7910 Teubner, cbzxt, Chen Qi, Chris Down, Christian Rebischke, Claudio
7911 Zumbo, ClydeByrdIII, crashfistfight, Cyprien Laplace, Daniel Edgecumbe,
7912 Daniel Gorbea, Daniel Rusek, Daniel Stuart, Dan Streetman, David
7913 Pedersen, David Tardon, Dimitri John Ledkov, Dominique Martinet, Donald
7914 A. Cupp Jr, Evgeny Vereshchagin, Fabian Henneke, Filipe Brandenburger,
7915 Franck Bui, Frantisek Sumsal, Georg Müller, Hans de Goede, Haochen
7916 Tong, HATAYAMA Daisuke, Iwan Timmer, Jan Janssen, Jan Kundrát, Jan
7917 Synacek, Jan Tojnar, Jay Strict, Jérémy Rosen, Jóhann B. Guðmundsson,
7918 Jonas Jelten, Jonas Thelemann, Justin Trudell, J. Xing, Kai-Heng Feng,
7919 Kenneth D'souza, Kevin Becker, Kevin Kuehler, Lennart Poettering,
7920 Léonard Gérard, Lorenz Bauer, Luca Boccassi, Maciej Stanczew, Mario
7921 Limonciello, Marko Myllynen, Mark Stosberg, Martin Wilck, matthiasroos,
7922 Michael Biebl, Michael Olbrich, Michael Tretter, Michal Sekletar,
7923 Michal Sekletár, Michal Suchanek, Mike Gilbert, Mike Kazantsev, Nicolas
7924 Douma, nikolas, Norbert Lange, pan93412, Pascal de Bruijn, Paul Menzel,
7925 Pavel Hrdina, Peter Wu, Philip Withnall, Piotr Drąg, Rafael Fontenelle,
7926 Renaud Métrich, Riccardo Schirone, RoadrunnerWMC, Ronan Pigott, Ryan
7927 Attard, Sebastian Wick, Serge, Siddharth Chandrasekara, Steve Ramage,
7928 Steve Traylen, Susant Sahani, Thibault Nélis, Tim Teichmann, Tom
7929 Fitzhenry, Tommy J, Torsten Hilbrich, Vito Caputo, ypf791, Yu Watanabe,
7930 Zach Smith, Zbigniew Jędrzejewski-Szmek
7932 – Warsaw, 2019-11-29
7936 * This release enables unprivileged programs (i.e. requiring neither
7937 setuid nor file capabilities) to send ICMP Echo (i.e. ping) requests
7938 by turning on the "net.ipv4.ping_group_range" sysctl of the Linux
7939 kernel for the whole UNIX group range, i.e. all processes. This
7940 change should be reasonably safe, as the kernel support for it was
7941 specifically implemented to allow safe access to ICMP Echo for
7942 processes lacking any privileges. If this is not desirable, it can be
7943 disabled again by setting the parameter to "1 0".
7945 * Previously, filters defined with SystemCallFilter= would have the
7946 effect that any calling of an offending system call would terminate
7947 the calling thread. This behaviour never made much sense, since
7948 killing individual threads of unsuspecting processes is likely to
7949 create more problems than it solves. With this release the default
7950 action changed from killing the thread to killing the whole
7951 process. For this to work correctly both a kernel version (>= 4.14)
7952 and a libseccomp version (>= 2.4.0) supporting this new seccomp
7953 action is required. If an older kernel or libseccomp is used the old
7954 behaviour continues to be used. This change does not affect any
7955 services that have no system call filters defined, or that use
7956 SystemCallErrorNumber= (and thus see EPERM or another error instead
7957 of being killed when calling an offending system call). Note that
7958 systemd documentation always claimed that the whole process is
7959 killed. With this change behaviour is thus adjusted to match the
7962 * On 64 bit systems, the "kernel.pid_max" sysctl is now bumped to
7963 4194304 by default, i.e. the full 22bit range the kernel allows, up
7964 from the old 16-bit range. This should improve security and
7965 robustness, as PID collisions are made less likely (though certainly
7966 still possible). There are rumours this might create compatibility
7967 problems, though at this moment no practical ones are known to
7968 us. Downstream distributions are hence advised to undo this change in
7969 their builds if they are concerned about maximum compatibility, but
7970 for everybody else we recommend leaving the value bumped. Besides
7971 improving security and robustness this should also simplify things as
7972 the maximum number of allowed concurrent tasks was previously bounded
7973 by both "kernel.pid_max" and "kernel.threads-max" and now effectively
7974 only a single knob is left ("kernel.threads-max"). There have been
7975 concerns that usability is affected by this change because larger PID
7976 numbers are harder to type, but we believe the change from 5 digits
7977 to 7 digits doesn't hamper usability.
7979 * MemoryLow= and MemoryMin= gained hierarchy-aware counterparts,
7980 DefaultMemoryLow= and DefaultMemoryMin=, which can be used to
7981 hierarchically set default memory protection values for a particular
7982 subtree of the unit hierarchy.
7984 * Memory protection directives can now take a value of zero, allowing
7985 explicit opting out of a default value propagated by an ancestor.
7987 * systemd now defaults to the "unified" cgroup hierarchy setup during
7988 build-time, i.e. -Ddefault-hierarchy=unified is now the build-time
7989 default. Previously, -Ddefault-hierarchy=hybrid was the default. This
7990 change reflects the fact that cgroupsv2 support has matured
7991 substantially in both systemd and in the kernel, and is clearly the
7992 way forward. Downstream production distributions might want to
7993 continue to use -Ddefault-hierarchy=hybrid (or even =legacy) for
7994 their builds as unfortunately the popular container managers have not
7995 caught up with the kernel API changes.
7997 * Man pages are not built by default anymore (html pages were already
7998 disabled by default), to make development builds quicker. When
7999 building systemd for a full installation with documentation, meson
8000 should be called with -Dman=true and/or -Dhtml=true as appropriate.
8001 The default was changed based on the assumption that quick one-off or
8002 repeated development builds are much more common than full optimized
8003 builds for installation, and people need to pass various other
8004 options to when doing "proper" builds anyway, so the gain from making
8005 development builds quicker is bigger than the one time disruption for
8008 Two scripts are created in the *build* directory to generate and
8009 preview man and html pages on demand, e.g.:
8011 build/man/man systemctl
8012 build/man/html systemd.index
8014 * libidn2 is used by default if both libidn2 and libidn are installed.
8015 Please use -Dlibidn=true if libidn is preferred.
8017 * The D-Bus "wire format" of the CPUAffinity= attribute is changed on
8018 big-endian machines. Before, bytes were written and read in native
8019 machine order as exposed by the native libc __cpu_mask interface.
8020 Now, little-endian order is always used (CPUs 0–7 are described by
8021 bits 0–7 in byte 0, CPUs 8–15 are described by byte 1, and so on).
8022 This change fixes D-Bus calls that cross endianness boundary.
8024 The presentation format used for CPUAffinity= by "systemctl show" and
8025 "systemd-analyze dump" is changed to present CPU indices instead of
8026 the raw __cpu_mask bitmask. For example, CPUAffinity=0-1 would be
8027 shown as CPUAffinity=03000000000000000000000000000… (on
8028 little-endian) or CPUAffinity=00000000000000300000000000000… (on
8029 64-bit big-endian), and is now shown as CPUAffinity=0-1, matching the
8030 input format. The maximum integer that will be printed in the new
8031 format is 8191 (four digits), while the old format always used a very
8032 long number (with the length varying by architecture), so they can be
8033 unambiguously distinguished.
8035 * /usr/sbin/halt.local is no longer supported. Implementation in
8036 distributions was inconsistent and it seems this functionality was
8039 To replace this functionality, users should:
8040 - either define a new unit and make it a dependency of final.target
8041 (systemctl add-wants final.target my-halt-local.service)
8042 - or move the shutdown script to /usr/lib/systemd/system-shutdown/
8043 and ensure that it accepts "halt", "poweroff", "reboot", and
8044 "kexec" as an argument, see the description in systemd-shutdown(8).
8046 * When a [Match] section in .link or .network file is empty (contains
8047 no match patterns), a warning will be emitted. Please add any "match
8048 all" pattern instead, e.g. OriginalName=* or Name=* in case all
8049 interfaces should really be matched.
8051 * A new setting NUMAPolicy= may be used to set process memory
8052 allocation policy. This setting can be specified in
8053 /etc/systemd/system.conf and hence will set the default policy for
8054 PID1. The default policy can be overridden on a per-service
8055 basis. The related setting NUMAMask= is used to specify NUMA node
8056 mask that should be associated with the selected policy.
8058 * PID 1 will now listen to Out-Of-Memory (OOM) events the kernel
8059 generates when processes it manages are reaching their memory limits,
8060 and will place their units in a special state, and optionally kill or
8061 stop the whole unit.
8063 * The service manager will now expose bus properties for the IO
8064 resources used by units. This information is also shown in "systemctl
8065 status" now (for services that have IOAccounting=yes set). Moreover,
8066 the IO accounting data is included in the resource log message
8067 generated whenever a unit stops.
8069 * Units may now configure an explicit timeout to wait for when killed
8070 with SIGABRT, for example when a service watchdog is hit. Previously,
8071 the regular TimeoutStopSec= timeout was applied in this case too —
8072 now a separate timeout may be set using TimeoutAbortSec=.
8074 * Services may now send a special WATCHDOG=trigger message with
8075 sd_notify() to trigger an immediate "watchdog missed" event, and thus
8076 trigger service termination. This is useful both for testing watchdog
8077 handling, but also for defining error paths in services, that shall
8078 be handled the same way as watchdog events.
8080 * There are two new per-unit settings IPIngressFilterPath= and
8081 IPEgressFilterPath= which allow configuration of a BPF program
8082 (usually by specifying a path to a program uploaded to /sys/fs/bpf/)
8083 to apply to the IP packet ingress/egress path of all processes of a
8084 unit. This is useful to allow running systemd services with BPF
8085 programs set up externally.
8087 * systemctl gained a new "clean" verb for removing the state, cache,
8088 runtime or logs directories of a service while it is terminated. The
8089 new verb may also be used to remove the state maintained on disk for
8090 timer units that have Persistent= configured.
8092 * During the last phase of shutdown systemd will now automatically
8093 increase the log level configured in the "kernel.printk" sysctl so
8094 that any relevant loggable events happening during late shutdown are
8095 made visible. Previously, loggable events happening so late during
8096 shutdown were generally lost if the "kernel.printk" sysctl was set to
8097 high thresholds, as regular logging daemons are terminated at that
8098 time and thus nothing is written to disk.
8100 * If processes terminated during the last phase of shutdown do not exit
8101 quickly systemd will now show their names after a short time, to make
8102 debugging easier. After a longer timeout they are forcibly killed,
8105 * journalctl (and the other tools that display logs) will now highlight
8106 warnings in yellow (previously, both LOG_NOTICE and LOG_WARNING where
8107 shown in bright bold, now only LOG_NOTICE is). Moreover, audit logs
8108 are now shown in blue color, to separate them visually from regular
8109 logs. References to configuration files are now turned into clickable
8110 links on terminals that support that.
8112 * systemd-journald will now stop logging to /var/log/journal during
8113 shutdown when /var/ is on a separate mount, so that it can be
8114 unmounted safely during shutdown.
8116 * systemd-resolved gained support for a new 'strict' DNS-over-TLS mode.
8118 * systemd-resolved "Cache=" configuration option in resolved.conf has
8119 been extended to also accept the 'no-negative' value. Previously,
8120 only a boolean option was allowed (yes/no), having yes as the
8121 default. If this option is set to 'no-negative', negative answers are
8122 not cached while the old cache heuristics are used positive answers.
8123 The default remains unchanged.
8125 * The predictable naming scheme for network devices now supports
8126 generating predictable names for "netdevsim" devices.
8128 Moreover, the "en" prefix was dropped from the ID_NET_NAME_ONBOARD
8131 Those two changes form a new net.naming_scheme= entry. Distributions
8132 which want to preserve naming stability may want to set the
8133 -Ddefault-net-naming-scheme= configuration option.
8135 * systemd-networkd now supports MACsec, nlmon, IPVTAP and Xfrm
8136 interfaces natively.
8138 * systemd-networkd's bridge FDB support now allows configuration of a
8139 destination address for each entry (Destination=), as well as the
8140 VXLAN VNI (VNI=), as well as an option to declare what an entry is
8141 associated with (AssociatedWith=).
8143 * systemd-networkd's DHCPv4 support now understands a new MaxAttempts=
8144 option for configuring the maximum number of DHCP lease requests. It
8145 also learnt a new BlackList= option for deny-listing DHCP servers (a
8146 similar setting has also been added to the IPv6 RA client), as well
8147 as a SendRelease= option for configuring whether to send a DHCP
8148 RELEASE message when terminating.
8150 * systemd-networkd's DHCPv4 and DHCPv6 stacks can now be configured
8151 separately in the [DHCPv4] and [DHCPv6] sections.
8153 * systemd-networkd's DHCP support will now optionally create an
8154 implicit host route to the DNS server specified in the DHCP lease, in
8155 addition to the routes listed explicitly in the lease. This should
8156 ensure that in multi-homed systems DNS traffic leaves the systems on
8157 the interface that acquired the DNS server information even if other
8158 routes such as default routes exist. This behaviour may be turned on
8159 with the new RoutesToDNS= option.
8161 * systemd-networkd's VXLAN support gained a new option
8162 GenericProtocolExtension= for enabling VXLAN Generic Protocol
8163 Extension support, as well as IPDoNotFragment= for setting the IP
8164 "Don't fragment" bit on outgoing packets. A similar option has been
8165 added to the GENEVE support.
8167 * In systemd-networkd's [Route] section you may now configure
8168 FastOpenNoCookie= for configuring per-route TCP fast-open support, as
8169 well as TTLPropagate= for configuring Label Switched Path (LSP) TTL
8170 propagation. The Type= setting now supports local, broadcast,
8171 anycast, multicast, any, xresolve routes, too.
8173 * systemd-networkd's [Network] section learnt a new option
8174 DefaultRouteOnDevice= for automatically configuring a default route
8175 onto the network device.
8177 * systemd-networkd's bridging support gained two new options ProxyARP=
8178 and ProxyARPWifi= for configuring proxy ARP behaviour as well as
8179 MulticastRouter= for configuring multicast routing behaviour. A new
8180 option MulticastIGMPVersion= may be used to change bridge's multicast
8181 Internet Group Management Protocol (IGMP) version.
8183 * systemd-networkd's FooOverUDP support gained the ability to configure
8184 local and peer IP addresses via Local= and Peer=. A new option
8185 PeerPort= may be used to configure the peer's IP port.
8187 * systemd-networkd's TUN support gained a new setting VnetHeader= for
8188 tweaking Generic Segment Offload support.
8190 * The address family for policy rules may be specified using the new
8191 Family= option in the [RoutingPolicyRule] section.
8193 * networkctl gained a new "delete" command for removing virtual network
8194 devices, as well as a new "--stats" switch for showing device
8197 * networkd.conf gained a new setting SpeedMeter= and
8198 SpeedMeterIntervalSec=, to measure bitrate of network interfaces. The
8199 measured speed may be shown by 'networkctl status'.
8201 * "networkctl status" now displays MTU and queue lengths, and more
8202 detailed information about VXLAN and bridge devices.
8204 * systemd-networkd's .network and .link files gained a new Property=
8205 setting in the [Match] section, to match against devices with
8206 specific udev properties.
8208 * systemd-networkd's tunnel support gained a new option
8209 AssignToLoopback= for selecting whether to use the loopback device
8210 "lo" as underlying device.
8212 * systemd-networkd's MACAddress= setting in the [Neighbor] section has
8213 been renamed to LinkLayerAddress=, and it now allows configuration of
8216 * systemd-networkd's handling of the kernel's disable_ipv6 sysctl is
8217 simplified: systemd-networkd will disable the sysctl (enable IPv6) if
8218 IPv6 configuration (static or DHCPv6) was found for a given
8219 interface. It will not touch the sysctl otherwise.
8221 * The order of entries is $PATH used by the user manager instance was
8222 changed to put bin/ entries before the corresponding sbin/ entries.
8223 It is recommended to not rely on this order, and only ever have one
8224 binary with a given name in the system paths under /usr.
8226 * A new tool systemd-network-generator has been added that may generate
8227 .network, .netdev and .link files from IP configuration specified on
8228 the kernel command line in the format used by Dracut.
8230 * The CriticalConnection= setting in .network files is now deprecated,
8231 and replaced by a new KeepConfiguration= setting which allows more
8232 detailed configuration of the IP configuration to keep in place.
8234 * systemd-analyze gained a few new verbs:
8236 - "systemd-analyze timestamp" parses and converts timestamps. This is
8237 similar to the existing "systemd-analyze calendar" command which
8238 does the same for recurring calendar events.
8240 - "systemd-analyze timespan" parses and converts timespans (i.e.
8241 durations as opposed to points in time).
8243 - "systemd-analyze condition" will parse and test ConditionXYZ=
8246 - "systemd-analyze exit-status" will parse and convert exit status
8247 codes to their names and back.
8249 - "systemd-analyze unit-files" will print a list of all unit
8250 file paths and unit aliases.
8252 * SuccessExitStatus=, RestartPreventExitStatus=, and
8253 RestartForceExitStatus= now accept exit status names (e.g. "DATAERR"
8254 is equivalent to "65"). Those exit status name mappings may be
8255 displayed with the systemd-analyze exit-status verb describe above.
8257 * systemd-logind now exposes a per-session SetBrightness() bus call,
8258 which may be used to securely change the brightness of a kernel
8259 brightness device, if it belongs to the session's seat. By using this
8260 call unprivileged clients can make changes to "backlight" and "leds"
8261 devices securely with strict requirements on session membership.
8262 Desktop environments may use this to generically make brightness
8263 changes to such devices without shipping private SUID binaries or
8264 udev rules for that purpose.
8266 * "udevadm info" gained a --wait-for-initialization switch to wait for
8267 a device to be initialized.
8269 * systemd-hibernate-resume-generator will now look for resumeflags= on
8270 the kernel command line, which is similar to rootflags= and may be
8271 used to configure device timeout for the hibernation device.
8273 * sd-event learnt a new API call sd_event_source_disable_unref() for
8274 disabling and unref'ing an event source in a single function. A
8275 related call sd_event_source_disable_unrefp() has been added for use
8276 with gcc's cleanup extension.
8278 * The sd-id128.h public API gained a new definition
8279 SD_ID128_UUID_FORMAT_STR for formatting a 128-bit ID in UUID format
8282 * "busctl introspect" gained a new switch --xml-interface for dumping
8283 XML introspection data unmodified.
8285 * PID 1 may now show the unit name instead of the unit description
8286 string in its status output during boot. This may be configured in
8287 the StatusUnitFormat= setting in /etc/systemd/system.conf or the
8288 kernel command line option systemd.status_unit_format=.
8290 * PID 1 now understands a new option KExecWatchdogSec= in
8291 /etc/systemd/system.conf to set a watchdog timeout for kexec reboots.
8292 Previously watchdog functionality was only available for regular
8293 reboots. The new setting defaults to off, because we don't know in
8294 the general case if the watchdog will be reset after kexec (some
8295 drivers do reset it, but not all), and the new userspace might not be
8296 configured to handle the watchdog.
8298 Moreover, the old ShutdownWatchdogSec= setting has been renamed to
8299 RebootWatchdogSec= to more clearly communicate what it is about. The
8300 old name is still accepted for compatibility.
8302 * The systemd.debug_shell kernel command line option now optionally
8303 takes a tty name to spawn the debug shell on, which allows a
8304 different tty to be selected than the built-in default.
8306 * Service units gained a new ExecCondition= setting which will run
8307 before ExecStartPre= and either continue execution of the unit (for
8308 clean exit codes), stop execution without marking the unit failed
8309 (for exit codes 1 through 254), or stop execution and fail the unit
8310 (for exit code 255 or abnormal termination).
8312 * A new service systemd-pstore.service has been added that pulls data
8313 from /sys/fs/pstore/ and saves it to /var/lib/pstore for later
8316 * timedatectl gained new verbs for configuring per-interface NTP
8317 service configuration for systemd-timesyncd.
8319 * "localectl list-locales" won't list non-UTF-8 locales anymore. It's
8320 2019. (You can set non-UTF-8 locales though, if you know their name.)
8322 * If variable assignments in sysctl.d/ files are prefixed with "-" any
8323 failures to apply them are now ignored.
8325 * systemd-random-seed.service now optionally credits entropy when
8326 applying the seed to the system. Set $SYSTEMD_RANDOM_SEED_CREDIT to
8327 true for the service to enable this behaviour, but please consult the
8328 documentation first, since this comes with a couple of caveats.
8330 * systemd-random-seed.service is now a synchronization point for full
8331 initialization of the kernel's entropy pool. Services that require
8332 /dev/urandom to be correctly initialized should be ordered after this
8335 * The systemd-boot boot loader has been updated to optionally maintain
8336 a random seed file in the EFI System Partition (ESP). During the boot
8337 phase, this random seed is read and updated with a new seed
8338 cryptographically derived from it. Another derived seed is passed to
8339 the OS. The latter seed is then credited to the kernel's entropy pool
8340 very early during userspace initialization (from PID 1). This allows
8341 systems to boot up with a fully initialized kernel entropy pool from
8342 earliest boot on, and thus entirely removes all entropy pool
8343 initialization delays from systems using systemd-boot. Special care
8344 is taken to ensure different seeds are derived on system images
8345 replicated to multiple systems. "bootctl status" will show whether
8346 a seed was received from the boot loader.
8348 * bootctl gained two new verbs:
8350 - "bootctl random-seed" will generate the file in ESP and an EFI
8351 variable to allow a random seed to be passed to the OS as described
8354 - "bootctl is-installed" checks whether systemd-boot is currently
8357 * bootctl will warn if it detects that boot entries are misconfigured
8358 (for example if the kernel image was removed without purging the
8361 * A new document has been added describing systemd's use and support
8362 for the kernel's entropy pool subsystem:
8364 https://systemd.io/RANDOM_SEEDS
8366 * When the system is hibernated the swap device to write the
8367 hibernation image to is now automatically picked from all available
8368 swap devices, preferring the swap device with the highest configured
8369 priority over all others, and picking the device with the most free
8370 space if there are multiple devices with the highest priority.
8372 * /etc/crypttab support has learnt a new keyfile-timeout= per-device
8373 option that permits selecting the timeout how long to wait for a
8374 device with an encryption key before asking for the password.
8376 * IOWeight= has learnt to properly set the IO weight when using the
8377 BFQ scheduler officially found in kernels 5.0+.
8379 * A new mailing list has been created for reporting of security issues:
8380 systemd-security@redhat.com. For mode details, see
8381 https://systemd.io/CONTRIBUTING#security-vulnerability-reports.
8383 Contributions from: Aaron Barany, Adrian Bunk, Alan Jenkins, Albrecht
8384 Lohofener, Andrej Valek, Anita Zhang, Arian van Putten, Balint Reczey,
8385 Bastien Nocera, Ben Boeckel, Benjamin Robin, camoz, Chen Qi, Chris
8386 Chiu, Chris Down, Christian Göttsche, Christian Kellner, Clinton Roy,
8387 Connor Reeder, Daniel Black, Daniel Lublin, Daniele Medri, Dan
8388 Streetman, Dave Reisner, Dave Ross, David Art, David Tardon, Debarshi
8389 Ray, Dimitri John Ledkov, Dominick Grift, Donald Buczek, Douglas
8390 Christman, Eric DeVolder, EtherGraf, Evgeny Vereshchagin, Feldwor,
8391 Felix Riemann, Florian Dollinger, Francesco Pennica, Franck Bui,
8392 Frantisek Sumsal, Franz Pletz, frederik, Hans de Goede, Iago López
8393 Galeiras, Insun Pyo, Ivan Shapovalov, Iwan Timmer, Jack, Jakob
8394 Unterwurzacher, Jan Chren, Jan Klötzke, Jan Losinski, Jan Pokorný, Jan
8395 Synacek, Jan-Michael Brummer, Jeka Pats, Jeremy Soller, Jérémy Rosen,
8396 Jiri Pirko, Joe Lin, Joerg Behrmann, Joe Richey, Jóhann B. Guðmundsson,
8397 Johannes Christ, Johannes Schmitz, Jonathan Rouleau, Jorge Niedbalski,
8398 Jörg Thalheim, Kai Krakow, Kai Lüke, Karel Zak, Kashyap Chamarthy,
8399 Krayushkin Konstantin, Lennart Poettering, Lubomir Rintel, Luca
8400 Boccassi, Luís Ferreira, Marc-André Lureau, Markus Felten, Martin Pitt,
8401 Matthew Leeds, Mattias Jernberg, Michael Biebl, Michael Olbrich,
8402 Michael Prokop, Michael Stapelberg, Michael Zhivich, Michal Koutný,
8403 Michal Sekletar, Mike Gilbert, Milan Broz, Miroslav Lichvar, mpe85,
8404 Mr-Foo, Network Silence, Oliver Harley, pan93412, Paul Menzel, pEJipE,
8405 Peter A. Bigot, Philip Withnall, Piotr Drąg, Rafael Fontenelle, Robert
8406 Scheck, Roberto Santalla, Ronan Pigott, root, RussianNeuroMancer,
8407 Sebastian Jennen, shinygold, Shreyas Behera, Simon Schricker, Susant
8408 Sahani, Thadeu Lima de Souza Cascardo, Theo Ouzhinski, Thiebaud
8409 Weksteen, Thomas Haller, Thomas Weißschuh, Tomas Mraz, Tommi Rantala,
8410 Topi Miettinen, VD-Lycos, ven, Vladimir Yerilov, Wieland Hoffmann,
8411 William A. Kennington III, William Wold, Xi Ruoyao, Yuri Chornoivan,
8412 Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zhang Xianwei
8414 – Camerino, 2019-09-03
8418 * In .link files, MACAddressPolicy=persistent (the default) is changed
8419 to cover more devices. For devices like bridges, tun, tap, bond, and
8420 similar interfaces that do not have other identifying information,
8421 the interface name is used as the basis for persistent seed for MAC
8422 and IPv4LL addresses. The way that devices that were handled
8423 previously is not changed, and this change is about covering more
8424 devices then previously by the "persistent" policy.
8426 MACAddressPolicy=random may be used to force randomized MACs and
8427 IPv4LL addresses for a device if desired.
8429 Hint: the log output from udev (at debug level) was enhanced to
8430 clarify what policy is followed and which attributes are used.
8431 `SYSTEMD_LOG_LEVEL=debug udevadm test-builtin net_setup_link /sys/class/net/<name>`
8432 may be used to view this.
8434 Hint: if a bridge interface is created without any slaves, and gains
8435 a slave later, then now the bridge does not inherit slave's MAC.
8436 To inherit slave's MAC, for example, create the following file:
8438 # /etc/systemd/network/98-bridge-inherit-mac.link
8443 MACAddressPolicy=none
8446 * The .device units generated by systemd-fstab-generator and other
8447 generators do not automatically pull in the corresponding .mount unit
8448 as a Wants= dependency. This means that simply plugging in the device
8449 will not cause the mount unit to be started automatically. But please
8450 note that the mount unit may be started for other reasons, in
8451 particular if it is part of local-fs.target, and any unit which
8452 (transitively) depends on local-fs.target is started.
8454 * networkctl list/status/lldp now accept globbing wildcards for network
8455 interface names to match against all existing interfaces.
8457 * The $PIDFILE environment variable is set to point the absolute path
8458 configured with PIDFile= for processes of that service.
8460 * The fallback DNS server list was augmented with Cloudflare public DNS
8461 servers. Use `-Ddns-servers=` to set a different fallback.
8463 * A new special target usb-gadget.target will be started automatically
8464 when a USB Device Controller is detected (which means that the system
8465 is a USB peripheral).
8467 * A new unit setting CPUQuotaPeriodSec= assigns the time period
8468 relatively to which the CPU time quota specified by CPUQuota= is
8471 * A new unit setting ProtectHostname= may be used to prevent services
8472 from modifying hostname information (even if they otherwise would
8473 have privileges to do so).
8475 * A new unit setting NetworkNamespacePath= may be used to specify a
8476 namespace for service or socket units through a path referring to a
8477 Linux network namespace pseudo-file.
8479 * The PrivateNetwork= setting and JoinsNamespaceOf= dependencies now
8480 have an effect on .socket units: when used the listening socket is
8481 created within the configured network namespace instead of the host
8484 * ExecStart= command lines in unit files may now be prefixed with ':'
8485 in which case environment variable substitution is
8486 disabled. (Supported for the other ExecXYZ= settings, too.)
8488 * .timer units gained two new boolean settings OnClockChange= and
8489 OnTimezoneChange= which may be used to also trigger a unit when the
8490 system clock is changed or the local timezone is
8491 modified. systemd-run has been updated to make these options easily
8492 accessible from the command line for transient timers.
8494 * Two new conditions for units have been added: ConditionMemory= may be
8495 used to conditionalize a unit based on installed system
8496 RAM. ConditionCPUs= may be used to conditionalize a unit based on
8497 installed CPU cores.
8499 * The @default system call filter group understood by SystemCallFilter=
8500 has been updated to include the new rseq() system call introduced in
8503 * A new time-set.target has been added that indicates that the system
8504 time has been set from a local source (possibly imprecise). The
8505 existing time-sync.target is stronger and indicates that the time has
8506 been synchronized with a precise external source. Services where
8507 approximate time is sufficient should use the new target.
8509 * "systemctl start" (and related commands) learnt a new
8510 --show-transaction option. If specified brief information about all
8511 jobs queued because of the requested operation is shown.
8513 * systemd-networkd recognizes a new operation state 'enslaved', used
8514 (instead of 'degraded' or 'carrier') for interfaces which form a
8515 bridge, bond, or similar, and an new 'degraded-carrier' operational
8516 state used for the bond or bridge master interface when one of the
8517 enslaved devices is not operational.
8519 * .network files learnt the new IgnoreCarrierLoss= option for leaving
8520 networks configured even if the carrier is lost.
8522 * The RequiredForOnline= setting in .network files may now specify a
8523 minimum operational state required for the interface to be considered
8524 "online" by systemd-networkd-wait-online. Related to this
8525 systemd-networkd-wait-online gained a new option --operational-state=
8526 to configure the same, and its --interface= option was updated to
8527 optionally also take an operational state specific for an interface.
8529 * systemd-networkd-wait-online gained a new setting --any for waiting
8530 for only one of the requested interfaces instead of all of them.
8532 * systemd-networkd now implements L2TP tunnels.
8534 * Two new .network settings UseAutonomousPrefix= and UseOnLinkPrefix=
8535 may be used to cause autonomous and onlink prefixes received in IPv6
8536 Router Advertisements to be ignored.
8538 * New MulticastFlood=, NeighborSuppression=, and Learning= .network
8539 file settings may be used to tweak bridge behaviour.
8541 * The new TripleSampling= option in .network files may be used to
8542 configure CAN triple sampling.
8544 * A new .netdev settings PrivateKeyFile= and PresharedKeyFile= may be
8545 used to point to private or preshared key for a WireGuard interface.
8547 * /etc/crypttab now supports the same-cpu-crypt and
8548 submit-from-crypt-cpus options to tweak encryption work scheduling
8551 * systemd-tmpfiles will now take a BSD file lock before operating on a
8552 contents of directory. This may be used to temporarily exclude
8553 directories from aging by taking the same lock (useful for example
8554 when extracting a tarball into /tmp or /var/tmp as a privileged user,
8555 which might create files with really old timestamps, which
8556 nevertheless should not be deleted). For further details, see:
8558 https://systemd.io/TEMPORARY_DIRECTORIES
8560 * systemd-tmpfiles' h line type gained support for the
8561 FS_PROJINHERIT_FL ('P') file attribute (introduced in kernel 4.5),
8562 controlling project quota inheritance.
8564 * sd-boot and bootctl now implement support for an Extended Boot Loader
8565 (XBOOTLDR) partition, that is intended to be mounted to /boot, in
8566 addition to the ESP partition mounted to /efi or /boot/efi.
8567 Configuration file fragments, kernels, initrds and other EFI images
8568 to boot will be loaded from both the ESP and XBOOTLDR partitions.
8569 The XBOOTLDR partition was previously described by the Boot Loader
8570 Specification, but implementation was missing in sd-boot. Support for
8571 this concept allows using the sd-boot boot loader in more
8572 conservative scenarios where the boot loader itself is placed in the
8573 ESP but the kernels to boot (and their metadata) in a separate
8576 * A system may now be booted with systemd.volatile=overlay on the
8577 kernel command line, which causes the root file system to be set up
8578 an overlayfs mount combining the root-only root directory with a
8579 writable tmpfs. In this setup, the underlying root device is not
8580 modified, and any changes are lost at reboot.
8582 * Similar, systemd-nspawn can now boot containers with a volatile
8583 overlayfs root with the new --volatile=overlay switch.
8585 * systemd-nspawn can now consume OCI runtime bundles using a new
8586 --oci-bundle= option. This implementation is fully usable, with most
8587 features in the specification implemented, but since this a lot of
8588 new code and functionality, this feature should most likely not
8589 be used in production yet.
8591 * systemd-nspawn now supports various options described by the OCI
8592 runtime specification on the command-line and in .nspawn files:
8593 --inaccessible=/Inaccessible= may be used to mask parts of the file
8594 system tree, --console=/--pipe may be used to configure how standard
8595 input, output, and error are set up.
8597 * busctl learned the `emit` verb to generate D-Bus signals.
8599 * systemd-analyze cat-config may be used to gather and display
8600 configuration spread over multiple files, for example system and user
8601 presets, tmpfiles.d, sysusers.d, udev rules, etc.
8603 * systemd-analyze calendar now takes an optional new parameter
8604 --iterations= which may be used to show a maximum number of iterations
8605 the specified expression will elapse next.
8607 * The sd-bus C API gained support for naming method parameters in the
8610 * systemd-logind gained D-Bus APIs to specify the "reboot parameter"
8611 the reboot() system call expects.
8613 * journalctl learnt a new --cursor-file= option that points to a file
8614 from which a cursor should be loaded in the beginning and to which
8615 the updated cursor should be stored at the end.
8617 * ACRN hypervisor and Windows Subsystem for Linux (WSL) are now
8618 detected by systemd-detect-virt (and may also be used in
8619 ConditionVirtualization=).
8621 * The behaviour of systemd-logind may now be modified with environment
8622 variables $SYSTEMD_REBOOT_TO_FIRMWARE_SETUP,
8623 $SYSTEMD_REBOOT_TO_BOOT_LOADER_MENU, and
8624 $SYSTEMD_REBOOT_TO_BOOT_LOADER_ENTRY. They cause logind to either
8625 skip the relevant operation completely (when set to false), or to
8626 create a flag file in /run/systemd (when set to true), instead of
8627 actually commencing the real operation when requested. The presence
8628 of /run/systemd/reboot-to-firmware-setup,
8629 /run/systemd/reboot-to-boot-loader-menu, and
8630 /run/systemd/reboot-to-boot-loader-entry, may be used by alternative
8631 boot loader implementations to replace some steps logind performs
8632 during reboot with their own operations.
8634 * systemctl can be used to request a reboot into the boot loader menu
8635 or a specific boot loader entry with the new --boot-load-menu= and
8636 --boot-loader-entry= options to a reboot command. (This requires a
8637 boot loader that supports this, for example sd-boot.)
8639 * kernel-install will no longer unconditionally create the output
8640 directory (e.g. /efi/<machine-id>/<kernel-version>) for boot loader
8641 snippets, but will do only if the machine-specific parent directory
8642 (i.e. /efi/<machine-id>/) already exists. bootctl has been modified
8643 to create this parent directory during sd-boot installation.
8645 This makes it easier to use kernel-install with plugins which support
8646 a different layout of the bootloader partitions (for example grub2).
8648 * During package installation (with `ninja install`), we would create
8649 symlinks for getty@tty1.service, systemd-networkd.service,
8650 systemd-networkd.socket, systemd-resolved.service,
8651 remote-cryptsetup.target, remote-fs.target,
8652 systemd-networkd-wait-online.service, and systemd-timesyncd.service
8653 in /etc, as if `systemctl enable` was called for those units, to make
8654 the system usable immediately after installation. Now this is not
8655 done anymore, and instead calling `systemctl preset-all` is
8656 recommended after the first installation of systemd.
8658 * A new boolean sandboxing option RestrictSUIDSGID= has been added that
8659 is built on seccomp. When turned on creation of SUID/SGID files is
8662 * The NoNewPrivileges= and the new RestrictSUIDSGID= options are now
8663 implied if DynamicUser= is turned on for a service. This hardens
8664 these services, so that they neither can benefit from nor create
8665 SUID/SGID executables. This is a minor compatibility breakage, given
8666 that when DynamicUser= was first introduced SUID/SGID behaviour was
8667 unaffected. However, the security benefit of these two options is
8668 substantial, and the setting is still relatively new, hence we opted
8669 to make it mandatory for services with dynamic users.
8671 Contributions from: Adam Jackson, Alexander Tsoy, Andrey Yashkin,
8672 Andrzej Pietrasiewicz, Anita Zhang, Balint Reczey, Beniamino Galvani,
8673 Ben Iofel, Benjamin Berg, Benjamin Dahlhoff, Chris, Chris Morin,
8674 Christopher Wong, Claudius Ellsel, Clemens Gruber, dana, Daniel Black,
8675 Davide Cavalca, David Michael, David Rheinsberg, emersion, Evgeny
8676 Vereshchagin, Filipe Brandenburger, Franck Bui, Frantisek Sumsal,
8677 Giacinto Cifelli, Hans de Goede, Hugo Kindel, Ignat Korchagin, Insun
8678 Pyo, Jan Engelhardt, Jonas Dorel, Jonathan Lebon, Jonathon Kowalski,
8679 Jörg Sommer, Jörg Thalheim, Jussi Pakkanen, Kai-Heng Feng, Lennart
8680 Poettering, Lubomir Rintel, Luís Ferreira, Martin Pitt, Matthias
8681 Klumpp, Michael Biebl, Michael Niewöhner, Michael Olbrich, Michal
8682 Sekletar, Mike Lothian, Paul Menzel, Piotr Drąg, Riccardo Schirone,
8683 Robin Elvedi, Roman Kulikov, Ronald Tschalär, Ross Burton, Ryan
8684 Gonzalez, Sebastian Krzyszkowiak, Stephane Chazelas, StKob, Susant
8685 Sahani, Sylvain Plantefève, Szabolcs Fruhwald, Taro Yamada, Theo
8686 Ouzhinski, Thomas Haller, Tobias Jungel, Tom Yan, Tony Asleson, Topi
8687 Miettinen, unixsysadmin, Van Laser, Vesa Jääskeläinen, Yu, Li-Yu,
8688 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
8690 — Warsaw, 2019-04-11
8694 * The default locale can now be configured at compile time. Otherwise,
8695 a suitable default will be selected automatically (one of C.UTF-8,
8696 en_US.UTF-8, and C).
8698 * The version string shown by systemd and other tools now includes the
8699 git commit hash when built from git. An override may be specified
8700 during compilation, which is intended to be used by distributions to
8701 include the package release information.
8703 * systemd-cat can now filter standard input and standard error streams
8704 for different syslog priorities using the new --stderr-priority=
8707 * systemd-journald and systemd-journal-remote reject entries which
8708 contain too many fields (CVE-2018-16865) and set limits on the
8709 process' command line length (CVE-2018-16864).
8711 * $DBUS_SESSION_BUS_ADDRESS environment variable is set by pam_systemd
8714 * A new network device NamePolicy "keep" is implemented for link files,
8715 and used by default in 99-default.link (the fallback configuration
8716 provided by systemd). With this policy, if the network device name
8717 was already set by userspace, the device will not be renamed again.
8718 This matches the naming scheme that was implemented before
8719 systemd-240. If naming-scheme < 240 is specified, the "keep" policy
8720 is also enabled by default, even if not specified. Effectively, this
8721 means that if naming-scheme >= 240 is specified, network devices will
8722 be renamed according to the configuration, even if they have been
8723 renamed already, if "keep" is not specified as the naming policy in
8724 the .link file. The 99-default.link file provided by systemd includes
8725 "keep" for backwards compatibility, but it is recommended for user
8726 installed .link files to *not* include it.
8728 The "kernel" policy, which keeps kernel names declared to be
8729 "persistent", now works again as documented.
8731 * kernel-install script now optionally takes the paths to one or more
8732 initrd files, and passes them to all plugins.
8734 * The mincore() system call has been dropped from the @system-service
8735 system call filter group, as it is pretty exotic and may potentially
8736 used for side-channel attacks.
8738 * -fPIE is dropped from compiler and linker options. Please specify
8739 -Db_pie=true option to meson to build position-independent
8740 executables. Note that the meson option is supported since meson-0.49.
8742 * The fs.protected_regular and fs.protected_fifos sysctls, which were
8743 added in Linux 4.19 to make some data spoofing attacks harder, are
8744 now enabled by default. While this will hopefully improve the
8745 security of most installations, it is technically a backwards
8746 incompatible change; to disable these sysctls again, place the
8747 following lines in /etc/sysctl.d/60-protected.conf or a similar file:
8749 fs.protected_regular = 0
8750 fs.protected_fifos = 0
8752 Note that the similar hardlink and symlink protection has been
8753 enabled since v199, and may be disabled likewise.
8755 * The files read from the EnvironmentFile= setting in unit files now
8756 parse backslashes inside quotes literally, matching the behaviour of
8759 * udevadm trigger, udevadm control, udevadm settle and udevadm monitor
8760 now automatically become NOPs when run in a chroot() environment.
8762 * The tmpfiles.d/ "C" line type will now copy directory trees not only
8763 when the destination is so far missing, but also if it already exists
8764 as a directory and is empty. This is useful to cater for systems
8765 where directory trees are put together from multiple separate mount
8766 points but otherwise empty.
8768 * A new function sd_bus_close_unref() (and the associated
8769 sd_bus_close_unrefp()) has been added to libsystemd, that combines
8770 sd_bus_close() and sd_bus_unref() in one.
8772 * udevadm control learnt a new option for --ping for testing whether a
8773 systemd-udevd instance is running and reacting.
8775 * udevadm trigger learnt a new option for --wait-daemon for waiting
8776 systemd-udevd daemon to be initialized.
8778 Contributions from: Aaron Plattner, Alberts Muktupāvels, Alex Mayer,
8779 Ayman Bagabas, Beniamino Galvani, Burt P, Chris Down, Chris Lamb, Chris
8780 Morin, Christian Hesse, Claudius Ellsel, dana, Daniel Axtens, Daniele
8781 Medri, Dave Reisner, David Santamaría Rogado, Diego Canuhe, Dimitri
8782 John Ledkov, Evgeny Vereshchagin, Fabrice Fontaine, Filipe
8783 Brandenburger, Franck Bui, Frantisek Sumsal, govwin, Hans de Goede,
8784 James Hilliard, Jan Engelhardt, Jani Uusitalo, Jan Janssen, Jan
8785 Synacek, Jonathan McDowell, Jonathan Roemer, Jonathon Kowalski, Joost
8786 Heitbrink, Jörg Thalheim, Lance, Lennart Poettering, Louis Taylor,
8787 Lucas Werkmeister, Mantas Mikulėnas, Marc-Antoine Perennou,
8788 marvelousblack, Michael Biebl, Michael Sloan, Michal Sekletar, Mike
8789 Auty, Mike Gilbert, Mikhail Kasimov, Neil Brown, Niklas Hambüchen,
8790 Patrick Williams, Paul Seyfert, Peter Hutterer, Philip Withnall, Roger
8791 James, Ronnie P. Thomas, Ryan Gonzalez, Sam Morris, Stephan Edel,
8792 Stephan Gerhold, Susant Sahani, Taro Yamada, Thomas Haller, Topi
8793 Miettinen, YiFei Zhu, YmrDtnJu, YunQiang Su, Yu Watanabe, Zbigniew
8794 Jędrzejewski-Szmek, zsergeant77, Дамјан Георгиевски
8796 — Berlin, 2019-02-14
8800 * NoNewPrivileges=yes has been set for all long-running services
8801 implemented by systemd. Previously, this was problematic due to
8802 SELinux (as this would also prohibit the transition from PID1's label
8803 to the service's label). This restriction has since been lifted, but
8804 an SELinux policy update is required.
8805 (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.)
8807 * DynamicUser=yes is dropped from systemd-networkd.service,
8808 systemd-resolved.service and systemd-timesyncd.service, which was
8809 enabled in v239 for systemd-networkd.service and systemd-resolved.service,
8810 and since v236 for systemd-timesyncd.service. The users and groups
8811 systemd-network, systemd-resolve and systemd-timesync are created
8812 by systemd-sysusers again. Distributors or system administrators
8813 may need to create these users and groups if they not exist (or need
8814 to re-enable DynamicUser= for those units) while upgrading systemd.
8815 Also, the clock file for systemd-timesyncd may need to move from
8816 /var/lib/private/systemd/timesync/clock to /var/lib/systemd/timesync/clock.
8818 * When unit files are loaded from disk, previously systemd would
8819 sometimes (depending on the unit loading order) load units from the
8820 target path of symlinks in .wants/ or .requires/ directories of other
8821 units. This meant that unit could be loaded from different paths
8822 depending on whether the unit was requested explicitly or as a
8823 dependency of another unit, not honouring the priority of directories
8824 in search path. It also meant that it was possible to successfully
8825 load and start units which are not found in the unit search path, as
8826 long as they were requested as a dependency and linked to from
8827 .wants/ or .requires/. The target paths of those symlinks are not
8828 used for loading units anymore and the unit file must be found in
8831 * A new service type has been added: Type=exec. It's very similar to
8832 Type=simple but ensures the service manager will wait for both fork()
8833 and execve() of the main service binary to complete before proceeding
8834 with follow-up units. This is primarily useful so that the manager
8835 propagates any errors in the preparation phase of service execution
8836 back to the job that requested the unit to be started. For example,
8837 consider a service that has ExecStart= set to a file system binary
8838 that doesn't exist. With Type=simple starting the unit would be
8839 considered instantly successful, as only fork() has to complete
8840 successfully and the manager does not wait for execve(), and hence
8841 its failure is seen "too late". With the new Type=exec service type
8842 starting the unit will fail, as the manager will wait for the
8843 execve() and notice its failure, which is then propagated back to the
8846 NOTE: with the next release 241 of systemd we intend to change the
8847 systemd-run tool to default to Type=exec for transient services
8848 started by it. This should be mostly safe, but in specific corner
8849 cases might result in problems, as the systemd-run tool will then
8850 block on NSS calls (such as user name look-ups due to User=) done
8851 between the fork() and execve(), which under specific circumstances
8852 might cause problems. It is recommended to specify "-p Type=simple"
8853 explicitly in the few cases where this applies. For regular,
8854 non-transient services (i.e. those defined with unit files on disk)
8855 we will continue to default to Type=simple.
8857 * The Linux kernel's current default RLIMIT_NOFILE resource limit for
8858 userspace processes is set to 1024 (soft) and 4096
8859 (hard). Previously, systemd passed this on unmodified to all
8860 processes it forked off. With this systemd release the hard limit
8861 systemd passes on is increased to 512K, overriding the kernel's
8862 defaults and substantially increasing the number of simultaneous file
8863 descriptors unprivileged userspace processes can allocate. Note that
8864 the soft limit remains at 1024 for compatibility reasons: the
8865 traditional UNIX select() call cannot deal with file descriptors >=
8866 1024 and increasing the soft limit globally might thus result in
8867 programs unexpectedly allocating a high file descriptor and thus
8868 failing abnormally when attempting to use it with select() (of
8869 course, programs shouldn't use select() anymore, and prefer
8870 poll()/epoll, but the call unfortunately remains undeservedly popular
8871 at this time). This change reflects the fact that file descriptor
8872 handling in the Linux kernel has been optimized in more recent
8873 kernels and allocating large numbers of them should be much cheaper
8874 both in memory and in performance than it used to be. Programs that
8875 want to take benefit of the increased limit have to "opt-in" into
8876 high file descriptors explicitly by raising their soft limit. Of
8877 course, when they do that they must acknowledge that they cannot use
8878 select() anymore (and neither can any shared library they use — or
8879 any shared library used by any shared library they use and so on).
8880 Which default hard limit is most appropriate is of course hard to
8881 decide. However, given reports that ~300K file descriptors are used
8882 in real-life applications we believe 512K is sufficiently high as new
8883 default for now. Note that there are also reports that using very
8884 high hard limits (e.g. 1G) is problematic: some software allocates
8885 large arrays with one element for each potential file descriptor
8886 (Java, …) — a high hard limit thus triggers excessively large memory
8887 allocations in these applications. Hopefully, the new default of 512K
8888 is a good middle ground: higher than what real-life applications
8889 currently need, and low enough for avoid triggering excessively large
8890 allocations in problematic software. (And yes, somebody should fix
8893 * The fs.nr_open and fs.file-max sysctls are now automatically bumped
8894 to the highest possible values, as separate accounting of file
8895 descriptors is no longer necessary, as memcg tracks them correctly as
8896 part of the memory accounting anyway. Thus, from the four limits on
8897 file descriptors currently enforced (fs.file-max, fs.nr_open,
8898 RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two,
8899 and keep only the latter two. A set of build-time options
8900 (-Dbump-proc-sys-fs-file-max=false and -Dbump-proc-sys-fs-nr-open=false)
8901 has been added to revert this change in behaviour, which might be
8902 an option for systems that turn off memcg in the kernel.
8904 * When no /etc/locale.conf file exists (and hence no locale settings
8905 are in place), systemd will now use the "C.UTF-8" locale by default,
8906 and set LANG= to it. This locale is supported by various
8907 distributions including Fedora, with clear indications that upstream
8908 glibc is going to make it available too. This locale enables UTF-8
8909 mode by default, which appears appropriate for 2018.
8911 * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by
8912 default. This effectively switches the RFC3704 Reverse Path filtering
8913 from Strict mode to Loose mode. This is more appropriate for hosts
8914 that have multiple links with routes to the same networks (e.g.
8915 a client with a Wi-Fi and Ethernet both connected to the internet).
8917 Consult the kernel documentation for details on this sysctl:
8918 https://docs.kernel.org/networking/ip-sysctl.html
8920 * The v239 change to turn on "net.ipv4.tcp_ecn" by default has been
8923 * CPUAccounting=yes no longer enables the CPU controller when using
8924 kernel 4.15+ and the unified cgroup hierarchy, as required accounting
8925 statistics are now provided independently from the CPU controller.
8927 * Support for disabling a particular cgroup controller within a sub-tree
8928 has been added through the DisableControllers= directive.
8930 * cgroup_no_v1=all on the kernel command line now also implies
8931 using the unified cgroup hierarchy, unless one explicitly passes
8932 systemd.unified_cgroup_hierarchy=0 on the kernel command line.
8934 * The new "MemoryMin=" unit file property may now be used to set the
8935 memory usage protection limit of processes invoked by the unit. This
8936 controls the cgroup v2 memory.min attribute. Similarly, the new
8937 "IODeviceLatencyTargetSec=" property has been added, wrapping the new
8938 cgroup v2 io.latency cgroup property for configuring per-service I/O
8941 * systemd now supports the cgroup v2 devices BPF logic, as counterpart
8942 to the cgroup v1 "devices" cgroup controller.
8944 * systemd-escape now is able to combine --unescape with --template. It
8945 also learnt a new option --instance for extracting and unescaping the
8946 instance part of a unit name.
8948 * sd-bus now provides the sd_bus_message_readv() which is similar to
8949 sd_bus_message_read() but takes a va_list object. The pair
8950 sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout()
8951 has been added for configuring the default method call timeout to
8952 use. sd_bus_error_move() may be used to efficiently move the contents
8953 from one sd_bus_error structure to another, invalidating the
8954 source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may
8955 be used to control whether a bus connection object is automatically
8956 flushed when an sd-event loop is exited.
8958 * When processing classic BSD syslog log messages, journald will now
8959 save the original time-stamp string supplied in the new
8960 SYSLOG_TIMESTAMP= journal field. This permits consumers to
8961 reconstruct the original BSD syslog message more correctly.
8963 * StandardOutput=/StandardError= in service files gained support for
8964 new "append:…" parameters, for connecting STDOUT/STDERR of a service
8965 to a file, and appending to it.
8967 * The signal to use as last step of killing of unit processes is now
8968 configurable. Previously it was hard-coded to SIGKILL, which may now
8969 be overridden with the new KillSignal= setting. Note that this is the
8970 signal used when regular termination (i.e. SIGTERM) does not suffice.
8971 Similarly, the signal used when aborting a program in case of a
8972 watchdog timeout may now be configured too (WatchdogSignal=).
8974 * The XDG_SESSION_DESKTOP environment variable may now be configured in
8975 the pam_systemd argument line, using the new desktop= switch. This is
8976 useful to initialize it properly from a display manager without
8977 having to touch C code.
8979 * Most configuration options that previously accepted percentage values
8980 now also accept permille values with the '‰' suffix (instead of '%').
8982 * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for
8985 * systemd-resolved's configuration file resolved.conf gained a new
8986 option ReadEtcHosts= which may be used to turn off processing and
8987 honoring /etc/hosts entries.
8989 * The "--wait" switch may now be passed to "systemctl
8990 is-system-running", in which case the tool will synchronously wait
8991 until the system finished start-up.
8993 * hostnamed gained a new bus call to determine the DMI product UUID.
8995 * On x86-64 systemd will now prefer using the RDRAND processor
8996 instruction over /dev/urandom whenever it requires randomness that
8997 neither has to be crypto-grade nor should be reproducible. This
8998 should substantially reduce the amount of entropy systemd requests
8999 from the kernel during initialization on such systems, though not
9000 reduce it to zero. (Why not zero? systemd still needs to allocate
9001 UUIDs and such uniquely, which require high-quality randomness.)
9003 * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP
9004 tunnels. It also gained a new option ForceDHCPv6PDOtherInformation=
9005 for forcing the "Other Information" bit in IPv6 RA messages. The
9006 bonding logic gained four new options AdActorSystemPriority=,
9007 AdUserPortKey=, AdActorSystem= for configuring various 802.3ad
9008 aspects, and DynamicTransmitLoadBalancing= for enabling dynamic
9009 shuffling of flows. The tunnel logic gained a new
9010 IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid
9011 Deployment. The policy rule logic gained four new options IPProtocol=,
9012 SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained
9013 support for the MulticastToUnicast= option. networkd also gained
9014 support for configuring static IPv4 ARP or IPv6 neighbor entries.
9016 * .preset files (as read by 'systemctl preset') may now be used to
9017 instantiate services.
9019 * /etc/crypttab now understands the sector-size= option to configure
9020 the sector size for an encrypted partition.
9022 * Key material for encrypted disks may now be placed on a formatted
9023 medium, and referenced from /etc/crypttab by the UUID of the file
9024 system, followed by "=" suffixed by the path to the key file.
9026 * The "collect" udev component has been removed without replacement, as
9027 it is neither used nor maintained.
9029 * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=,
9030 LogsDirectory=, ConfigurationDirectory= settings are used in a
9031 service the executed processes will now receive a set of environment
9032 variables containing the full paths of these directories.
9033 Specifically, RUNTIME_DIRECTORY=, STATE_DIRECTORY, CACHE_DIRECTORY,
9034 LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options
9035 are used. Note that these options may be used multiple times per
9036 service in which case the resulting paths will be concatenated and
9037 separated by colons.
9039 * Predictable interface naming has been extended to cover InfiniBand
9040 NICs. They will be exposed with an "ib" prefix.
9042 * tmpfiles.d/ line types may now be suffixed with a '-' character, in
9043 which case the respective line failing is ignored.
9045 * .link files may now be used to configure the equivalent to the
9046 "ethtool advertise" commands.
9048 * The sd-device.h and sd-hwdb.h APIs are now exported, as an
9049 alternative to libudev.h. Previously, the latter was just an internal
9050 wrapper around the former, but now these two APIs are exposed
9053 * sd-id128.h gained a new function sd_id128_get_boot_app_specific()
9054 which calculates an app-specific boot ID similar to how
9055 sd_id128_get_machine_app_specific() generates an app-specific machine
9058 * A new tool systemd-id128 has been added that can be used to determine
9059 and generate various 128-bit IDs.
9061 * /etc/os-release gained two new standardized fields DOCUMENTATION_URL=
9064 * systemd-hibernate-resume-generator will now honor the "noresume"
9065 kernel command line option, in which case it will bypass resuming
9066 from any hibernated image.
9068 * The systemd-sleep.conf configuration file gained new options
9069 AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=,
9070 AllowHybridSleep= for prohibiting specific sleep modes even if the
9071 kernel exports them.
9073 * portablectl is now officially supported and has thus moved to
9076 * bootctl learnt the two new commands "set-default" and "set-oneshot"
9077 for setting the default boot loader item to boot to (either
9078 persistently or only for the next boot). This is currently only
9079 compatible with sd-boot, but may be implemented on other boot loaders
9080 too, that follow the boot loader interface. The updated interface is
9081 now documented here:
9083 https://systemd.io/BOOT_LOADER_INTERFACE
9085 * A new kernel command line option systemd.early_core_pattern= is now
9086 understood which may be used to influence the core_pattern PID 1
9087 installs during early boot.
9089 * busctl learnt two new options -j and --json= for outputting method
9090 call replies, properties and monitoring output in JSON.
9092 * journalctl's JSON output now supports simple ANSI coloring as well as
9093 a new "json-seq" mode for generating RFC7464 output.
9095 * Unit files now support the %g/%G specifiers that resolve to the UNIX
9096 group/GID of the service manager runs as, similar to the existing
9097 %u/%U specifiers that resolve to the UNIX user/UID.
9099 * systemd-logind learnt a new global configuration option
9100 UserStopDelaySec= that may be set in logind.conf. It specifies how
9101 long the systemd --user instance shall remain started after a user
9102 logs out. This is useful to speed up repetitive re-connections of the
9103 same user, as it means the user's service manager doesn't have to be
9104 stopped/restarted on each iteration, but can be reused between
9105 subsequent options. This setting defaults to 10s. systemd-logind also
9106 exports two new properties on its Manager D-Bus objects indicating
9107 whether the system's lid is currently closed, and whether the system
9110 * systemd gained support for a generic boot counting logic, which
9111 generically permits automatic reverting to older boot loader entries
9112 if newer updated ones don't work. The boot loader side is implemented
9113 in sd-boot, but is kept open for other boot loaders too. For details
9116 https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT
9118 * The SuccessAction=/FailureAction= unit file settings now learnt two
9119 new parameters: "exit" and "exit-force", which result in immediate
9120 exiting of the service manager, and are only useful in systemd --user
9121 and container environments.
9123 * Unit files gained support for a pair of options
9124 FailureActionExitStatus=/SuccessActionExitStatus= for configuring the
9125 exit status to use as service manager exit status when
9126 SuccessAction=/FailureAction= is set to exit or exit-force.
9128 * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service
9129 options may now be used to configure the log rate limiting applied by
9130 journald per-service.
9132 * systemd-analyze gained a new verb "timespan" for parsing and
9133 normalizing time span values (i.e. strings like "5min 7s 8us").
9135 * systemd-analyze also gained a new verb "security" for analyzing the
9136 security and sand-boxing settings of services in order to determine an
9137 "exposure level" for them, indicating whether a service would benefit
9138 from more sand-boxing options turned on for them.
9140 * "systemd-analyze syscall-filter" will now also show system calls
9141 supported by the local kernel but not included in any of the defined
9144 * .nspawn files now understand the Ephemeral= setting, matching the
9145 --ephemeral command line switch.
9147 * sd-event gained the new APIs sd_event_source_get_floating() and
9148 sd_event_source_set_floating() for controlling whether a specific
9149 event source is "floating", i.e. destroyed along with the even loop
9152 * Unit objects on D-Bus gained a new "Refs" property that lists all
9153 clients that currently have a reference on the unit (to ensure it is
9156 * The JoinControllers= option in system.conf is no longer supported, as
9157 it didn't work correctly, is hard to support properly, is legacy (as
9158 the concept only exists on cgroup v1) and apparently wasn't used.
9160 * Journal messages that are generated whenever a unit enters the failed
9161 state are now tagged with a unique MESSAGE_ID. Similarly, messages
9162 generated whenever a service process exits are now made recognizable,
9163 too. A tagged message is also emitted whenever a unit enters the
9164 "dead" state on success.
9166 * systemd-run gained a new switch --working-directory= for configuring
9167 the working directory of the service to start. A shortcut -d is
9168 equivalent, setting the working directory of the service to the
9169 current working directory of the invoking program. The new --shell
9170 (or just -S) option has been added for invoking the $SHELL of the
9171 caller as a service, and implies --pty --same-dir --wait --collect
9172 --service-type=exec. Or in other words, "systemd-run -S" is now the
9173 quickest way to quickly get an interactive in a fully clean and
9174 well-defined system service context.
9176 * machinectl gained a new verb "import-fs" for importing an OS tree
9177 from a directory. Moreover, when a directory or tarball is imported
9178 and single top-level directory found with the OS itself below the OS
9179 tree is automatically mangled and moved one level up.
9181 * systemd-importd will no longer set up an implicit btrfs loop-back
9182 file system on /var/lib/machines. If one is already set up, it will
9183 continue to be used.
9185 * A new generator "systemd-run-generator" has been added. It will
9186 synthesize a unit from one or more program command lines included in
9187 the kernel command line. This is very useful in container managers
9190 # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"'
9192 This will run "systemd-nspawn" on an image, invoke the specified
9193 command line and immediately shut down the container again, returning
9194 the command line's exit code.
9196 * The block device locking logic is now documented:
9198 https://systemd.io/BLOCK_DEVICE_LOCKING
9200 * loginctl and machinectl now optionally output the various tables in
9201 JSON using the --output= switch. It is our intention to add similar
9202 support to systemctl and all other commands.
9204 * udevadm's query and trigger verb now optionally take a .device unit
9207 * systemd-udevd's network naming logic now understands a new
9208 net.naming_scheme= kernel command line switch, which may be used to
9209 pick a specific version of the naming scheme. This helps stabilizing
9210 interface names even as systemd/udev are updated and the naming logic
9213 * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and
9214 SD_ID128_ALLF to test if a 128-bit ID is set to all 0xFF bytes, and to
9215 initialize one to all 0xFF.
9217 * After loading the SELinux policy systemd will now recursively relabel
9218 all files and directories listed in
9219 /run/systemd/relabel-extra.d/*.relabel (which should be simple
9220 newline separated lists of paths) in addition to the ones it already
9221 implicitly relabels in /run, /dev and /sys. After the relabelling is
9222 completed the *.relabel files (and /run/systemd/relabel-extra.d/) are
9223 removed. This is useful to permit initrds (i.e. code running before
9224 the SELinux policy is in effect) to generate files in the host
9225 filesystem safely and ensure that the correct label is applied during
9226 the transition to the host OS.
9228 * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding
9229 mknod() handling in user namespaces. Previously mknod() would always
9230 fail with EPERM in user namespaces. Since 4.18 mknod() will succeed
9231 but device nodes generated that way cannot be opened, and attempts to
9232 open them result in EPERM. This breaks the "graceful fallback" logic
9233 in systemd's PrivateDevices= sand-boxing option. This option is
9234 implemented defensively, so that when systemd detects it runs in a
9235 restricted environment (such as a user namespace, or an environment
9236 where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD)
9237 where device nodes cannot be created the effect of PrivateDevices= is
9238 bypassed (following the logic that 2nd-level sand-boxing is not
9239 essential if the system systemd runs in is itself already sand-boxed
9240 as a whole). This logic breaks with 4.18 in container managers where
9241 user namespacing is used: suddenly PrivateDevices= succeeds setting
9242 up a private /dev/ file system containing devices nodes — but when
9243 these are opened they don't work.
9245 At this point it is recommended that container managers utilizing
9246 user namespaces that intend to run systemd in the payload explicitly
9247 block mknod() with seccomp or similar, so that the graceful fallback
9250 We are very sorry for the breakage and the requirement to change
9251 container configurations for newer kernels. It's purely caused by an
9252 incompatible kernel change. The relevant kernel developers have been
9253 notified about this userspace breakage quickly, but they chose to
9256 * PermissionsStartOnly= setting is deprecated (but is still supported
9257 for backwards compatibility). The same functionality is provided by
9258 the more flexible "+", "!", and "!!" prefixes to ExecStart= and other
9261 * $DBUS_SESSION_BUS_ADDRESS environment variable is not set by
9262 pam_systemd anymore.
9264 * The naming scheme for network devices was changed to always rename
9265 devices, even if they were already renamed by userspace. The "kernel"
9266 policy was changed to only apply as a fallback, if no other naming
9269 * The requirements to build systemd is bumped to meson-0.46 and
9272 Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander
9273 Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson,
9274 Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov,
9275 asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt
9276 Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen
9277 Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius
9278 Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn
9279 Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner,
9280 David Anderson, Davide Cavalca, David Leeds, David Malcolm, David
9281 Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus,
9282 Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters,
9283 Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad,
9284 Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank
9285 Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe
9286 Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit
9287 Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan
9288 Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld,
9289 javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi
9290 Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens,
9291 Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi,
9292 Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry
9293 Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz
9294 Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier,
9295 Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin
9296 Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko
9297 Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck,
9298 Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich,
9299 Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal
9300 Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal
9301 Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby,
9302 Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł
9303 Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller,
9304 Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez,
9305 Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam
9306 Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher,
9307 Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee
9308 (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen
9309 Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim,
9310 Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas
9311 Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias
9312 Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore
9313 Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech
9314 Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward,
9315 Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe,
9316 Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein
9318 — Warsaw, 2018-12-21
9322 * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
9323 builtin will name network interfaces differently than in previous
9324 versions for virtual network interfaces created with SR-IOV and NPAR
9325 and for devices where the PCI network controller device does not have
9326 a slot number associated.
9328 SR-IOV virtual devices are now named based on the name of the parent
9329 interface, with a suffix of "v<N>", where <N> is the virtual device
9330 number. Previously those virtual devices were named as if completely
9333 The ninth and later NPAR virtual devices will be named following the
9334 scheme used for the first eight NPAR partitions. Previously those
9335 devices were not renamed and the kernel default (eth<n>) was used.
9337 "net_id" will also generate names for PCI devices where the PCI
9338 network controller device does not have an associated slot number
9339 itself, but one of its parents does. Previously those devices were
9340 not renamed and the kernel default (eth<n>) was used.
9342 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
9343 systemd-logind.service. Since v235, IPAddressDeny=any has been set to
9344 the unit. So, it is expected that the default behavior of
9345 systemd-logind is not changed. However, if distribution packagers or
9346 administrators disabled or modified IPAddressDeny= setting by a
9347 drop-in config file, then it may be necessary to update the file to
9348 re-enable AF_INET and AF_INET6 to support network user name services,
9351 * When the RestrictNamespaces= unit property is specified multiple
9352 times, then the specified types are merged now. Previously, only the
9353 last assignment was used. So, if distribution packagers or
9354 administrators modified the setting by a drop-in config file, then it
9355 may be necessary to update the file.
9357 * When OnFailure= is used in combination with Restart= on a service
9358 unit, then the specified units will no longer be triggered on
9359 failures that result in restarting. Previously, the specified units
9360 would be activated each time the unit failed, even when the unit was
9361 going to be restarted automatically. This behaviour contradicted the
9362 documentation. With this release the code is adjusted to match the
9365 * systemd-tmpfiles will now print a notice whenever it encounters
9366 tmpfiles.d/ lines referencing the /var/run/ directory. It will
9367 recommend reworking them to use the /run/ directory instead (for
9368 which /var/run/ is simply a symlinked compatibility alias). This way
9369 systemd-tmpfiles can properly detect line conflicts and merge lines
9370 referencing the same file by two paths, without having to access
9373 * systemctl disable/unmask/preset/preset-all cannot be used with
9374 --runtime. Previously this was allowed, but resulted in unintuitive
9375 behaviour that wasn't useful. systemctl disable/unmask will now undo
9376 both runtime and persistent enablement/masking, i.e. it will remove
9377 any relevant symlinks both in /run and /etc.
9379 * Note that all long-running system services shipped with systemd will
9380 now default to a system call allow list (rather than a deny list, as
9381 before). In particular, systemd-udevd will now enforce one too. For
9382 most cases this should be safe, however downstream distributions
9383 which disabled sandboxing of systemd-udevd (specifically the
9384 MountFlags= setting), might want to disable this security feature
9385 too, as the default allow-listing will prohibit all mount, swap,
9386 reboot and clock changing operations from udev rules.
9388 * sd-boot acquired new loader configuration settings to optionally turn
9389 off Windows and MacOS boot partition discovery as well as
9390 reboot-into-firmware menu items. It is also able to pick a better
9391 screen resolution for HiDPI systems, and now provides loader
9392 configuration settings to change the resolution explicitly.
9394 * systemd-resolved now supports DNS-over-TLS. It's still
9395 turned off by default, use DNSOverTLS=opportunistic to turn it on in
9396 resolved.conf. We intend to make this the default as soon as couple
9397 of additional techniques for optimizing the initial latency caused by
9398 establishing a TLS/TCP connection are implemented.
9400 * systemd-resolved.service and systemd-networkd.service now set
9401 DynamicUser=yes. The users systemd-resolve and systemd-network are
9402 not created by systemd-sysusers anymore.
9404 NOTE: This has a chance of breaking nss-ldap and similar NSS modules
9405 that embed a network facing module into any process using getpwuid()
9406 or related call: the dynamic allocation of the user ID for
9407 systemd-resolved.service means the service manager has to check NSS
9408 if the user name is already taken when forking off the service. Since
9409 the user in the common case won't be defined in /etc/passwd the
9410 lookup is likely to trigger nss-ldap which in turn might use NSS to
9411 ask systemd-resolved for hostname lookups. This will hence result in
9412 a deadlock: a user name lookup in order to start
9413 systemd-resolved.service will result in a hostname lookup for which
9414 systemd-resolved.service needs to be started already. There are
9415 multiple ways to work around this problem: pre-allocate the
9416 "systemd-resolve" user on such systems, so that nss-ldap won't be
9417 triggered; or use a different NSS package that doesn't do networking
9418 in-process but provides a local asynchronous name cache; or configure
9419 the NSS package to avoid lookups for UIDs in the range `pkg-config
9420 systemd --variable=dynamicuidmin` … `pkg-config systemd
9421 --variable=dynamicuidmax`, so that it does not consider itself
9422 authoritative for the same UID range systemd allocates dynamic users
9425 * The systemd-resolve tool has been renamed to resolvectl (it also
9426 remains available under the old name, for compatibility), and its
9427 interface is now verb-based, similar in style to the other <xyz>ctl
9428 tools, such as systemctl or loginctl.
9430 * The resolvectl/systemd-resolve tool also provides 'resolvconf'
9431 compatibility. It may be symlinked under the 'resolvconf' name, in
9432 which case it will take arguments and input compatible with the
9433 Debian and FreeBSD resolvconf tool.
9435 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
9436 where the system initially suspends, and after a timeout resumes and
9439 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
9440 set the client will only send a DUID as client identifier. (EDIT: the
9441 option was broken, and was dropped in v255.)
9443 * The nss-systemd glibc NSS module will now enumerate dynamic users and
9444 groups in effect. Previously, it could resolve UIDs/GIDs to user
9445 names/groups and vice versa, but did not support enumeration.
9447 * journald's Compress= configuration setting now optionally accepts a
9448 byte threshold value. All journal objects larger than this threshold
9449 will be compressed, smaller ones will not. Previously this threshold
9450 was not configurable and set to 512.
9452 * A new system.conf setting NoNewPrivileges= is now available which may
9453 be used to turn off acquisition of new privileges system-wide
9454 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
9455 for all its children). Note that turning this option on means setuid
9456 binaries and file system capabilities lose their special powers.
9457 While turning on this option is a big step towards a more secure
9458 system, doing so is likely to break numerous pre-existing UNIX tools,
9459 in particular su and sudo.
9461 * A new service systemd-time-sync-wait.service has been added. If
9462 enabled it will delay the time-sync.target unit at boot until time
9463 synchronization has been received from the network. This
9464 functionality is useful on systems lacking a local RTC or where it is
9465 acceptable that the boot process shall be delayed by external network
9468 * When hibernating, systemd will now inform the kernel of the image
9469 write offset, on kernels new enough to support this. This means swap
9470 files should work for hibernation now.
9472 * When loading unit files, systemd will now look for drop-in unit files
9473 extensions in additional places. Previously, for a unit file name
9474 "foo-bar-baz.service" it would look for dropin files in
9475 "foo-bar-baz.service.d/*.conf". Now, it will also look in
9476 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
9477 service name truncated after all inner dashes. This scheme allows
9478 writing drop-ins easily that apply to a whole set of unit files at
9479 once. It's particularly useful for mount and slice units (as their
9480 naming is prefix based), but is also useful for service and other
9481 units, for packages that install multiple unit files at once,
9482 following a strict naming regime of beginning the unit file name with
9483 the package's name. Two new specifiers are now supported in unit
9484 files to match this: %j and %J are replaced by the part of the unit
9485 name following the last dash.
9487 * Unit files and other configuration files that support specifier
9488 expansion now understand another three new specifiers: %T and %V will
9489 resolve to /tmp and /var/tmp respectively, or whatever temporary
9490 directory has been set for the calling user. %E will expand to either
9491 /etc (for system units) or $XDG_CONFIG_HOME (for user units).
9493 * The ExecStart= lines of unit files are no longer required to
9494 reference absolute paths. If non-absolute paths are specified the
9495 specified binary name is searched within the service manager's
9496 built-in $PATH, which may be queried with 'systemd-path
9497 search-binaries-default'. It's generally recommended to continue to
9498 use absolute paths for all binaries specified in unit files.
9500 * Units gained a new load state "bad-setting", which is used when a
9501 unit file was loaded, but contained fatal errors which prevent it
9502 from being started (for example, a service unit has been defined
9503 lacking both ExecStart= and ExecStop= lines).
9505 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
9506 support alternative debuggers, for example lldb. The old name
9507 continues to be available however, for compatibility reasons. Use the
9508 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
9509 to pick an alternative debugger instead of the default gdb.
9511 * systemctl and the other tools will now output escape sequences that
9512 generate proper clickable hyperlinks in various terminal emulators
9513 where useful (for example, in the "systemctl status" output you can
9514 now click on the unit file name to quickly open it in the
9515 editor/viewer of your choice). Note that not all terminal emulators
9516 support this functionality yet, but many do. Unfortunately, the
9517 "less" pager doesn't support this yet, hence this functionality is
9518 currently automatically turned off when a pager is started (which
9519 happens quite often due to auto-paging). We hope to remove this
9520 limitation as soon as "less" learns these escape sequences. This new
9521 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
9522 environment variable. For details on these escape sequences see:
9523 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
9525 * networkd's .network files now support a new IPv6MTUBytes= option for
9526 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
9527 option in the [Route] section to configure the MTU to use for
9528 specific routes. It also gained support for configuration of the DHCP
9529 "UserClass" option through the new UserClass= setting. It gained
9530 three new options in the new [CAN] section for configuring CAN
9531 networks. The MULTICAST and ALLMULTI interface flags may now be
9532 controlled explicitly with the new Multicast= and AllMulticast=
9535 * networkd will now automatically make use of the kernel's route
9536 expiration feature, if it is available.
9538 * udevd's .link files now support setting the number of receive and
9539 transmit channels, using the RxChannels=, TxChannels=,
9540 OtherChannels=, CombinedChannels= settings.
9542 * Support for UDPSegmentationOffload= has been removed, given its
9543 limited support in hardware, and waning software support.
9545 * networkd's .netdev files now support creating "netdevsim" interfaces.
9547 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
9548 to query the unit belonging to a specific kernel control group.
9550 * systemd-analyze gained a new verb "cat-config", which may be used to
9551 dump the contents of any configuration file, with all its matching
9552 drop-in files added in, and honouring the usual search and masking
9553 logic applied to systemd configuration files. For example use
9554 "systemd-analyze cat-config systemd/system.conf" to get the complete
9555 system configuration file of systemd how it would be loaded by PID 1
9556 itself. Similar to this, various tools such as systemd-tmpfiles or
9557 systemd-sysusers, gained a new option "--cat-config", which does the
9558 corresponding operation for their own configuration settings. For
9559 example, "systemd-tmpfiles --cat-config" will now output the full
9560 list of tmpfiles.d/ lines in place.
9562 * timedatectl gained three new verbs: "show" shows bus properties of
9563 systemd-timedated, "timesync-status" shows the current NTP
9564 synchronization state of systemd-timesyncd, and "show-timesync"
9565 shows bus properties of systemd-timesyncd.
9567 * systemd-timesyncd gained a bus interface on which it exposes details
9570 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
9571 understood by systemd-timedated. It takes a colon-separated list of
9572 unit names of NTP client services. The list is used by
9573 "timedatectl set-ntp".
9575 * systemd-nspawn gained a new --rlimit= switch for setting initial
9576 resource limits for the container payload. There's a new switch
9577 --hostname= to explicitly override the container's hostname. A new
9578 --no-new-privileges= switch may be used to control the
9579 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
9580 --oom-score-adjust= switch controls the OOM scoring adjustment value
9581 for the payload. The new --cpu-affinity= switch controls the CPU
9582 affinity of the container payload. The new --resolv-conf= switch
9583 allows more detailed control of /etc/resolv.conf handling of the
9584 container. Similarly, the new --timezone= switch allows more detailed
9585 control of /etc/localtime handling of the container.
9587 * systemd-detect-virt gained a new --list switch, which will print a
9588 list of all currently known VM and container environments.
9590 * Support for "Portable Services" has been added, see
9591 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
9592 experimental, but this is expected to change soon. Reflecting this
9593 experimental state, the "portablectl" binary is not installed into
9594 /usr/bin yet. The binary has to be called with the full path
9595 /usr/lib/systemd/portablectl instead.
9597 * journalctl's and systemctl's -o switch now knows a new log output
9598 mode "with-unit". The output it generates is very similar to the
9599 regular "short" mode, but displays the unit name instead of the
9600 syslog tag for each log line. Also, the date is shown with timezone
9601 information. This mode is probably more useful than the classic
9602 "short" output mode for most purposes, except where pixel-perfect
9603 compatibility with classic /var/log/messages formatting is required.
9605 * A new --dump-bus-properties switch has been added to the systemd
9606 binary, which may be used to dump all supported D-Bus properties.
9607 (Options which are still supported, but are deprecated, are *not*
9610 * sd-bus gained a set of new calls:
9611 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
9612 enable/disable the "floating" state of a bus slot object,
9613 i.e. whether the slot object pins the bus it is allocated for into
9614 memory or if the bus slot object gets disconnected when the bus goes
9615 away. sd_bus_open_with_description(),
9616 sd_bus_open_user_with_description(),
9617 sd_bus_open_system_with_description() may be used to allocate bus
9618 objects and set their description string already during allocation.
9620 * sd-event gained support for watching inotify events from the event
9621 loop, in an efficient way, sharing inotify handles between multiple
9622 users. For this a new function sd_event_add_inotify() has been added.
9624 * sd-event and sd-bus gained support for calling special user-supplied
9625 destructor functions for userdata pointers associated with
9626 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
9627 functions sd_bus_slot_set_destroy_callback,
9628 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
9629 sd_bus_track_get_destroy_callback,
9630 sd_event_source_set_destroy_callback,
9631 sd_event_source_get_destroy_callback have been added.
9633 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
9635 * PID 1 will now automatically reschedule .timer units whenever the
9636 local timezone changes. (They previously got rescheduled
9637 automatically when the system clock changed.)
9639 * New documentation has been added to document cgroups delegation,
9640 portable services and the various code quality tools we have set up:
9642 https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md
9643 https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md
9644 https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md
9646 * The Boot Loader Specification has been added to the source tree.
9648 https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md
9650 While moving it into our source tree we have updated it and further
9651 changes are now accepted through the usual github PR workflow.
9653 * pam_systemd will now look for PAM userdata fields systemd.memory_max,
9654 systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
9655 earlier PAM modules. The data in these fields is used to initialize
9656 the session scope's resource properties. Thus external PAM modules
9657 may now configure per-session limits, for example sourced from
9658 external user databases.
9660 * socket units with Accept=yes will now maintain a "refused" counter in
9661 addition to the existing "accepted" counter, counting connections
9662 refused due to the enforced limits.
9664 * The "systemd-path search-binaries-default" command may now be use to
9665 query the default, built-in $PATH PID 1 will pass to the services it
9668 * A new unit file setting PrivateMounts= has been added. It's a boolean
9669 option. If enabled the unit's processes are invoked in their own file
9670 system namespace. Note that this behaviour is also implied if any
9671 other file system namespacing options (such as PrivateTmp=,
9672 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
9673 primarily useful for services that do not use any of the other file
9674 system namespacing options. One such service is systemd-udevd.service
9675 where this is now used by default.
9677 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
9678 when the system is booted in UEFI "secure mode".
9680 * A new unit "system-update-pre.target" is added, which defines an
9681 optional synchronization point for offline system updates, as
9682 implemented by the pre-existing "system-update.target" unit. It
9683 allows ordering services before the service that executes the actual
9684 update process in a generic way.
9686 * Systemd now emits warnings whenever .include syntax is used.
9688 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
9689 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
9690 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
9691 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
9692 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
9693 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
9694 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
9695 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
9696 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
9697 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
9698 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
9699 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
9700 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
9701 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
9702 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
9703 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
9704 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
9705 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
9706 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
9707 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
9708 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
9709 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
9710 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
9711 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
9712 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
9713 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
9714 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
9715 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
9716 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
9718 — Berlin, 2018-06-22
9722 * The MemoryAccounting= unit property now defaults to on. After
9723 discussions with the upstream control group maintainers we learnt
9724 that the negative impact of cgroup memory accounting on current
9725 kernels is finally relatively minimal, so that it should be safe to
9726 enable this by default without affecting system performance. Besides
9727 memory accounting only task accounting is turned on by default, all
9728 other forms of resource accounting (CPU, IO, IP) remain off for now,
9729 because it's not clear yet that their impact is small enough to move
9730 from opt-in to opt-out. We recommend downstreams to leave memory
9731 accounting on by default if kernel 4.14 or higher is primarily
9732 used. On very resource constrained systems or when support for old
9733 kernels is a necessity, -Dmemory-accounting-default=false can be used
9734 to revert this change.
9736 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
9737 %udev_rules_update) and the journal catalog (%journal_catalog_update)
9738 from the upgrade scriptlets of individual packages now do nothing.
9739 Transfiletriggers have been added which will perform those updates
9740 once at the end of the transaction.
9742 Similar transfiletriggers have been added to execute any sysctl.d
9743 and binfmt.d rules. Thus, it should be unnecessary to provide any
9744 scriptlets to execute this configuration from package installation
9747 * systemd-sysusers gained a mode where the configuration to execute is
9748 specified on the command line, but this configuration is not executed
9749 directly, but instead it is merged with the configuration on disk,
9750 and the result is executed. This is useful for package installation
9751 scripts which want to create the user before installing any files on
9752 disk (in case some of those files are owned by that user), while
9753 still allowing local admin overrides.
9755 This functionality is exposed to rpm scriptlets through a new
9756 %sysusers_create_package macro. Old %sysusers_create and
9757 %sysusers_create_inline macros are deprecated.
9759 A transfiletrigger for sysusers.d configuration is now installed,
9760 which means that it should be unnecessary to call systemd-sysusers from
9761 package installation scripts, unless the package installs any files
9762 owned by those newly-created users, in which case
9763 %sysusers_create_package should be used.
9765 * Analogous change has been done for systemd-tmpfiles: it gained a mode
9766 where the command-line configuration is merged with the configuration
9767 on disk. This is exposed as the new %tmpfiles_create_package macro,
9768 and %tmpfiles_create is deprecated. A transfiletrigger is installed
9769 for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
9770 from package installation scripts.
9772 * sysusers.d configuration for a user may now also specify the group
9773 number, in addition to the user number ("u username 123:456"), or
9774 without the user number ("u username -:456").
9776 * Configution items for systemd-sysusers can now be specified as
9777 positional arguments when the new --inline switch is used.
9779 * The login shell of users created through sysusers.d may now be
9780 specified (previously, it was always /bin/sh for root and
9781 /sbin/nologin for other users).
9783 * systemd-analyze gained a new --global switch to look at global user
9784 configuration. It also gained a unit-paths verb to list the unit load
9785 paths that are compiled into systemd (which can be used with
9786 --systemd, --user, or --global).
9788 * udevadm trigger gained a new --settle/-w option to wait for any
9789 triggered events to finish (but just those, and not any other events
9790 which are triggered meanwhile).
9792 * The action that systemd-logind takes when the lid is closed and the
9793 machine is connected to external power can now be configured using
9794 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
9795 was determined by HandleLidSwitch=, and, for backwards compatibility,
9796 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
9798 * journalctl will periodically call sd_journal_process() to make it
9799 resilient against inotify queue overruns when journal files are
9800 rotated very quickly.
9802 * Two new functions in libsystemd — sd_bus_get_n_queued_read and
9803 sd_bus_get_n_queued_write — may be used to check the number of
9804 pending bus messages.
9806 * systemd gained a new
9807 org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
9808 which can be used to migrate foreign processes to scope and service
9809 units. The primary user for this new API is systemd itself: the
9810 systemd --user instance uses this call of the systemd --system
9811 instance to migrate processes if it itself gets the request to
9812 migrate processes and the kernel refuses this due to access
9813 restrictions. Thanks to this "systemd-run --scope --user …" works
9814 again in pure cgroup v2 environments when invoked from the user
9817 * A new TemporaryFileSystem= setting can be used to mask out part of
9818 the real file system tree with tmpfs mounts. This may be combined
9819 with BindPaths= and BindReadOnlyPaths= to hide files or directories
9820 not relevant to the unit, while still allowing some paths lower in
9821 the tree to be accessed.
9823 ProtectHome=tmpfs may now be used to hide user home and runtime
9824 directories from units, in a way that is mostly equivalent to
9825 "TemporaryFileSystem=/home /run/user /root".
9827 * Non-service units are now started with KeyringMode=shared by default.
9828 This means that mount and swapon and other mount tools have access
9829 to keys in the main keyring.
9831 * /sys/fs/bpf is now mounted automatically.
9833 * QNX virtualization is now detected by systemd-detect-virt and may
9834 be used in ConditionVirtualization=.
9836 * IPAccounting= may now be enabled also for slice units.
9838 * A new -Dsplit-bin= build configuration switch may be used to specify
9839 whether bin and sbin directories are merged, or if they should be
9840 included separately in $PATH and various listings of executable
9841 directories. The build configuration scripts will try to autodetect
9842 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
9843 system, but distributions are encouraged to configure this
9846 * A new -Dok-color= build configuration switch may be used to change
9847 the colour of "OK" status messages.
9849 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
9850 PrivateNetwork=yes was buggy in previous versions of systemd. This
9851 means that after the upgrade and daemon-reexec, any such units must
9854 * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
9855 will not exclude read-only files owned by root from cleanup.
9857 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
9858 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
9859 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
9860 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
9861 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
9862 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
9863 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
9864 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
9865 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
9866 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
9867 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
9868 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
9869 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
9870 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
9871 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
9872 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
9874 — Warsaw, 2018-03-05
9878 * Some keyboards come with a zoom see-saw or rocker which until now got
9879 mapped to the Linux "zoomin/out" keys in hwdb. However, these
9880 keycodes are not recognized by any major desktop. They now produce
9881 Up/Down key events so that they can be used for scrolling.
9883 * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
9884 slightly: previously, if an argument was specified for lines of this
9885 type (i.e. the right-most column was set) this string was appended to
9886 existing files each time systemd-tmpfiles was run. This behaviour was
9887 different from what the documentation said, and not particularly
9888 useful, as repeated systemd-tmpfiles invocations would not be
9889 idempotent and grow such files without bounds. With this release
9890 behaviour has been altered to match what the documentation says:
9891 lines of this type only have an effect if the indicated files don't
9892 exist yet, and only then the argument string is written to the file.
9894 * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
9895 systemd-tmpfiles behaviour: previously, read-only files owned by root
9896 were always excluded from the file "aging" algorithm (i.e. the
9897 automatic clean-up of directories like /tmp based on
9898 atime/mtime/ctime). We intend to drop this restriction, and age files
9899 by default even when owned by root and read-only. This behaviour was
9900 inherited from older tools, but there have been requests to remove
9901 it, and it's not obvious why this restriction was made in the first
9902 place. Please speak up now, if you are aware of software that requires
9903 this behaviour, otherwise we'll remove the restriction in v238.
9905 * A new environment variable $SYSTEMD_OFFLINE is now understood by
9906 systemctl. It takes a boolean argument. If on, systemctl assumes it
9907 operates on an "offline" OS tree, and will not attempt to talk to the
9908 service manager. Previously, this mode was implicitly enabled if a
9909 chroot() environment was detected, and this new environment variable
9910 now provides explicit control.
9912 * .path and .socket units may now be created transiently, too.
9913 Previously only service, mount, automount and timer units were
9914 supported as transient units. The systemd-run tool has been updated
9915 to expose this new functionality, you may hence use it now to bind
9916 arbitrary commands to path or socket activation on-the-fly from the
9917 command line. Moreover, almost all properties are now exposed for the
9918 unit types that already supported transient operation.
9920 * The systemd-mount command gained support for a new --owner= parameter
9921 which takes a user name, which is then resolved and included in uid=
9922 and gid= mount options string of the file system to mount.
9924 * A new unit condition ConditionControlGroupController= has been added
9925 that checks whether a specific cgroup controller is available.
9927 * Unit files, udev's .link files, and systemd-networkd's .netdev and
9928 .network files all gained support for a new condition
9929 ConditionKernelVersion= for checking against specific kernel
9932 * In systemd-networkd, the [IPVLAN] section in .netdev files gained
9933 support for configuring device flags in the Flags= setting. In the
9934 same files, the [Tunnel] section gained support for configuring
9935 AllowLocalRemote=. The [Route] section in .network files gained
9936 support for configuring InitialCongestionWindow=,
9937 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
9938 understands RapidCommit=.
9940 * systemd-networkd's DHCPv6 support gained support for Prefix
9943 * sd-bus gained support for a new "watch-bind" feature. When this
9944 feature is enabled, an sd_bus connection may be set up to connect to
9945 an AF_UNIX socket in the file system as soon as it is created. This
9946 functionality is useful for writing early-boot services that
9947 automatically connect to the system bus as soon as it is started,
9948 without ugly time-based polling. systemd-networkd and
9949 systemd-resolved have been updated to make use of this
9950 functionality. busctl exposes this functionality in a new
9951 --watch-bind= command line switch.
9953 * sd-bus will now optionally synthesize a local "Connected" signal as
9954 soon as a D-Bus connection is set up fully. This message mirrors the
9955 already existing "Disconnected" signal which is synthesized when the
9956 connection is terminated. This signal is generally useful but
9957 particularly handy in combination with the "watch-bind" feature
9958 described above. Synthesizing of this message has to be requested
9959 explicitly through the new API call sd_bus_set_connected_signal(). In
9960 addition a new call sd_bus_is_ready() has been added that checks
9961 whether a connection is fully set up (i.e. between the "Connected" and
9962 "Disconnected" signals).
9964 * sd-bus gained two new calls sd_bus_request_name_async() and
9965 sd_bus_release_name_async() for asynchronously registering bus
9966 names. Similar, there is now sd_bus_add_match_async() for installing
9967 a signal match asynchronously. All of systemd's own services have
9968 been updated to make use of these calls. Doing these operations
9969 asynchronously has two benefits: it reduces the risk of deadlocks in
9970 case of cyclic dependencies between bus services, and it speeds up
9971 service initialization since synchronization points for bus
9972 round-trips are removed.
9974 * sd-bus gained two new calls sd_bus_match_signal() and
9975 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
9976 and sd_bus_add_match_async() but instead of taking a D-Bus match
9977 string take match fields as normal function parameters.
9979 * sd-bus gained two new calls sd_bus_set_sender() and
9980 sd_bus_message_set_sender() for setting the sender name of outgoing
9981 messages (either for all outgoing messages or for just one specific
9982 one). These calls are only useful in direct connections as on
9983 brokered connections the broker fills in the sender anyway,
9984 overwriting whatever the client filled in.
9986 * sd-event gained a new pseudo-handle that may be specified on all API
9987 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
9988 used this refers to the default event loop object of the calling
9989 thread. Note however that this does not implicitly allocate one —
9990 which has to be done prior by using sd_event_default(). Similarly
9991 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
9992 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
9993 to the default bus of the specified type of the calling thread. Here
9994 too this does not implicitly allocate bus connection objects, this
9995 has to be done prior with sd_bus_default() and friends.
9997 * sd-event gained a new call pair
9998 sd_event_source_{get|set}_io_fd_own(). This may be used to request
9999 automatic closure of the file descriptor an IO event source watches
10000 when the event source is destroyed.
10002 * systemd-networkd gained support for natively configuring WireGuard
10005 * In previous versions systemd synthesized user records both for the
10006 "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
10007 internally. In order to simplify distribution-wide renames of the
10008 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
10009 new transitional flag file has been added: if
10010 /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
10011 user and group record within the systemd codebase is disabled.
10013 * systemd-notify gained a new --uid= option for selecting the source
10014 user/UID to use for notification messages sent to the service
10017 * journalctl gained a new --grep= option to list only entries in which
10018 the message matches a certain pattern. By default matching is case
10019 insensitive if the pattern is lowercase, and case sensitive
10020 otherwise. Option --case-sensitive=yes|no can be used to override
10021 this an specify case sensitivity or case insensitivity.
10023 * There's now a "systemd-analyze service-watchdogs" command for printing
10024 the current state of the service runtime watchdog, and optionally
10025 enabling or disabling the per-service watchdogs system-wide if given a
10026 boolean argument (i.e. the concept you configure in WatchdogSec=), for
10027 debugging purposes. There's also a kernel command line option
10028 systemd.service_watchdogs= for controlling the same.
10030 * Two new "log-level" and "log-target" options for systemd-analyze were
10031 added that merge the now deprecated get-log-level, set-log-level and
10032 get-log-target, set-log-target pairs. The deprecated options are still
10033 understood for backwards compatibility. The two new options print the
10034 current value when no arguments are given, and set them when a
10035 level/target is given as an argument.
10037 * sysusers.d's "u" lines now optionally accept both a UID and a GID
10038 specification, separated by a ":" character, in order to create users
10039 where UID and GID do not match.
10041 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
10042 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
10043 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
10044 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
10045 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
10046 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
10047 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
10048 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
10049 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
10050 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
10051 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
10052 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
10053 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
10054 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
10055 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
10056 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
10057 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
10058 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
10059 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
10060 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
10067 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
10068 in v235 has been extended to also set the dummy.ko module option
10069 numdummies=0, preventing the kernel from automatically creating
10070 dummy0. All dummy interfaces must now be explicitly created.
10072 * Unknown '%' specifiers in configuration files are now rejected. This
10073 applies to units and tmpfiles.d configuration. Any percent characters
10074 that are followed by a letter or digit that are not supposed to be
10075 interpreted as the beginning of a specifier should be escaped by
10076 doubling ("%%"). (So "size=5%" is still accepted, as well as
10077 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
10078 valid specifiers today.)
10080 * systemd-resolved now maintains a new dynamic
10081 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
10082 recommended to make /etc/resolv.conf a symlink to it. This file
10083 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
10084 includes dynamically acquired search domains, achieving more correct
10085 DNS resolution by software that bypasses local DNS APIs such as NSS.
10087 * The "uaccess" udev tag has been dropped from /dev/kvm and
10088 /dev/dri/renderD*. These devices now have the 0666 permissions by
10089 default (but this may be changed at build-time). /dev/dri/renderD*
10090 will now be owned by the "render" group along with /dev/kfd.
10092 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
10093 systemd-journal-gatewayd.service and
10094 systemd-journal-upload.service. This means "nss-systemd" must be
10095 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
10096 services are resolved properly.
10098 * In /etc/fstab two new mount options are now understood:
10099 x-systemd.makefs and x-systemd.growfs. The former has the effect that
10100 the configured file system is formatted before it is mounted, the
10101 latter that the file system is resized to the full block device size
10102 after it is mounted (i.e. if the file system is smaller than the
10103 partition it resides on, it's grown). This is similar to the fsck
10104 logic in /etc/fstab, and pulls in systemd-makefs@.service and
10105 systemd-growfs@.service as necessary, similar to
10106 systemd-fsck@.service. Resizing is currently only supported on ext4
10109 * In systemd-networkd, the IPv6 RA logic now optionally may announce
10110 DNS server and domain information.
10112 * Support for the LUKS2 on-disk format for encrypted partitions has
10113 been added. This requires libcryptsetup2 during compilation and
10116 * The systemd --user instance will now signal "readiness" when its
10117 basic.target unit has been reached, instead of when the run queue ran
10118 empty for the first time.
10120 * Tmpfiles.d with user configuration are now also supported.
10121 systemd-tmpfiles gained a new --user switch, and snippets placed in
10122 ~/.config/user-tmpfiles.d/ and corresponding directories will be
10123 executed by systemd-tmpfiles --user running in the new
10124 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
10125 running in the user session.
10127 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
10128 %S resolves to the top-level state directory (/var/lib for the system
10129 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
10130 top-level cache directory (/var/cache for the system instance,
10131 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
10132 logs directory (/var/log for the system instance,
10133 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
10134 existing %t specifier, that resolves to the top-level runtime
10135 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
10138 * journalctl learnt a new parameter --output-fields= for limiting the
10139 set of journal fields to output in verbose and JSON output modes.
10141 * systemd-timesyncd's configuration file gained a new option
10142 RootDistanceMaxSec= for setting the maximum root distance of servers
10143 it'll use, as well as the new options PollIntervalMinSec= and
10144 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
10146 * bootctl gained a new command "list" for listing all available boot
10147 menu items on systems that follow the boot loader specification.
10149 * systemctl gained a new --dry-run switch that shows what would be done
10150 instead of doing it, and is currently supported by the shutdown and
10153 * ConditionSecurity= can now detect the TOMOYO security module.
10155 * Unit file [Install] sections are now also respected in unit drop-in
10156 files. This is intended to be used by drop-ins under /usr/lib/.
10158 * systemd-firstboot may now also set the initial keyboard mapping.
10160 * Udev "changed" events for devices which are exposed as systemd
10161 .device units are now propagated to units specified in
10162 ReloadPropagatedFrom= as reload requests.
10164 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
10165 unit template name (i.e. a name in the form of 'foobar@.service',
10166 without the instance component between the '@' and - the '.'), then
10167 the escaped sysfs path of the device is automatically used as the
10170 * SystemCallFilter= in unit files has been extended so that an "errno"
10171 can be specified individually for each system call. Example:
10172 SystemCallFilter=~uname:EILSEQ.
10174 * The cgroup delegation logic has been substantially updated. Delegate=
10175 now optionally takes a list of controllers (instead of a boolean, as
10176 before), which lists the controllers to delegate at least.
10178 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
10180 * A new LogLevelMax= setting configures the maximum log level any
10181 process of the service may log at (i.e. anything with a lesser
10182 priority than what is specified is automatically dropped). A new
10183 LogExtraFields= setting allows configuration of additional journal
10184 fields to attach to all log records generated by any of the unit's
10187 * New StandardInputData= and StandardInputText= settings along with the
10188 new option StandardInput=data may be used to configure textual or
10189 binary data that shall be passed to the executed service process via
10190 standard input, encoded in-line in the unit file.
10192 * StandardInput=, StandardOutput= and StandardError= may now be used to
10193 connect stdin/stdout/stderr of executed processes directly with a
10194 file or AF_UNIX socket in the file system, using the new "file:" option.
10196 * A new unit file option CollectMode= has been added, that allows
10197 tweaking the garbage collection logic for units. It may be used to
10198 tell systemd to garbage collect units that have failed automatically
10199 (normally it only GCs units that exited successfully). systemd-run
10200 and systemd-mount expose this new functionality with a new -G option.
10202 * "machinectl bind" may now be used to bind mount non-directories
10203 (i.e. regularfiles, devices, fifos, sockets).
10205 * systemd-analyze gained a new verb "calendar" for validating and
10206 testing calendar time specifications to use for OnCalendar= in timer
10207 units. Besides validating the expression it will calculate the next
10208 time the specified expression would elapse.
10210 * In addition to the pre-existing FailureAction= unit file setting
10211 there's now SuccessAction=, for configuring a shutdown action to
10212 execute when a unit completes successfully. This is useful in
10213 particular inside containers that shall terminate after some workload
10214 has been completed. Also, both options are now supported for all unit
10215 types, not just services.
10217 * networkds's IP rule support gained two new options
10218 IncomingInterface= and OutgoingInterface= for configuring the incoming
10219 and outgoing interfaces of configured rules. systemd-networkd also
10220 gained support for "vxcan" network devices.
10222 * networkd gained a new setting RequiredForOnline=, taking a
10223 boolean. If set, systemd-wait-online will take it into consideration
10224 when determining that the system is up, otherwise it will ignore the
10225 interface for this purpose.
10227 * The sd_notify() protocol gained support for a new operation: with
10228 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
10229 store again, ahead of POLLHUP or POLLERR when they are removed
10232 * A new document doc/UIDS-GIDS.md has been added to the source tree,
10233 that documents the UID/GID range and assignment assumptions and
10234 requirements of systemd.
10236 * The watchdog device PID 1 will ping may now be configured through the
10237 WatchdogDevice= configuration file setting, or by setting the
10238 systemd.watchdog_service= kernel command line option.
10240 * systemd-resolved's gained support for registering DNS-SD services on
10241 the local network using MulticastDNS. Services may either be
10242 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
10243 the same dir below /run, /usr/lib), or through its D-Bus API.
10245 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
10246 extend the effective start, runtime, and stop time. The service must
10247 continue to send EXTEND_TIMEOUT_USEC within the period specified to
10248 prevent the service manager from making the service as timedout.
10250 * systemd-resolved's DNSSEC support gained support for RFC 8080
10251 (Ed25519 keys and signatures).
10253 * The systemd-resolve command line tool gained a new set of options
10254 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
10255 --set-nta= and --revert to configure per-interface DNS configuration
10256 dynamically during runtime. It's useful for pushing DNS information
10257 into systemd-resolved from DNS hook scripts that various interface
10258 managing software supports (such as pppd).
10260 * systemd-nspawn gained a new --network-namespace-path= command line
10261 option, which may be used to make a container join an existing
10262 network namespace, by specifying a path to a "netns" file.
10264 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
10265 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
10266 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
10267 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
10268 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
10269 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
10270 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
10271 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
10272 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
10273 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
10274 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
10275 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
10276 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
10277 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
10278 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
10279 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
10280 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
10281 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
10282 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
10283 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
10284 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
10285 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
10286 Jędrzejewski-Szmek, Zeal Jagannatha
10288 — Berlin, 2017-12-14
10292 * INCOMPATIBILITY: systemd-logind.service and other long-running
10293 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
10294 communication with the outside. This generally improves security of
10295 the system, and is in almost all cases a safe and good choice, as
10296 these services do not and should not provide any network-facing
10297 functionality. However, systemd-logind uses the glibc NSS API to
10298 query the user database. This creates problems on systems where NSS
10299 is set up to directly consult network services for user database
10300 lookups. In particular, this creates incompatibilities with the
10301 "nss-nis" module, which attempts to directly contact the NIS/YP
10302 network servers it is configured for, and will now consistently
10303 fail. In such cases, it is possible to turn off IP sandboxing for
10304 systemd-logind.service (set IPAddressDeny= in its [Service] section
10305 to the empty string, via a .d/ unit file drop-in). Downstream
10306 distributions might want to update their nss-nis packaging to include
10307 such a drop-in snippet, accordingly, to hide this incompatibility
10308 from the user. Another option is to make use of glibc's nscd service
10309 to proxy such network requests through a privilege-separated, minimal
10310 local caching daemon, or to switch to more modern technologies such
10311 sssd, whose NSS hook-ups generally do not involve direct network
10312 access. In general, we think it's definitely time to question the
10313 implementation choices of nss-nis, i.e. whether it's a good idea
10314 today to embed a network-facing loadable module into all local
10315 processes that need to query the user database, including the most
10316 trivial and benign ones, such as "ls". For more details about
10317 IPAddressDeny= see below.
10319 * A new modprobe.d drop-in is now shipped by default that sets the
10320 bonding module option max_bonds=0. This overrides the kernel default,
10321 to avoid conflicts and ambiguity as to whether or not bond0 should be
10322 managed by systemd-networkd or not. This resolves multiple issues
10323 with bond0 properties not being applied, when bond0 is configured
10324 with systemd-networkd. Distributors may choose to not package this,
10325 however in that case users will be prevented from correctly managing
10326 bond0 interface using systemd-networkd.
10328 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
10329 which print the logging level and target of the system manager. They
10330 complement the existing "set-log-level" and "set-log-target" verbs
10331 used to change those values.
10333 * journald.conf gained a new boolean setting ReadKMsg= which defaults
10334 to on. If turned off kernel log messages will not be read by
10335 systemd-journald or included in the logs. It also gained a new
10336 setting LineMax= for configuring the maximum line length in
10337 STDOUT/STDERR log streams. The new default for this value is 48K, up
10338 from the previous hardcoded 2048.
10340 * A new unit setting RuntimeDirectoryPreserve= has been added, which
10341 allows more detailed control of what to do with a runtime directory
10342 configured with RuntimeDirectory= (i.e. a directory below /run or
10343 $XDG_RUNTIME_DIR) after a unit is stopped.
10345 * The RuntimeDirectory= setting for units gained support for creating
10346 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
10347 one top-level directory.
10349 * Units gained new options StateDirectory=, CacheDirectory=,
10350 LogsDirectory= and ConfigurationDirectory= which are closely related
10351 to RuntimeDirectory= but manage per-service directories below
10352 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
10353 possible to write unit files which when activated automatically gain
10354 properly owned service specific directories in these locations, thus
10355 making unit files self-contained and increasing compatibility with
10356 stateless systems and factory reset where /etc or /var are
10357 unpopulated at boot. Matching these new settings there's also
10358 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
10359 ConfigurationDirectoryMode= for configuring the access mode of these
10360 directories. These settings are particularly useful in combination
10361 with DynamicUser=yes as they provide secure, properly-owned,
10362 writable, and stateful locations for storage, excluded from the
10363 sandbox that such services live in otherwise.
10365 * Automake support has been removed from this release. systemd is now
10368 * systemd-journald will now aggressively cache client metadata during
10369 runtime, speeding up log write performance under pressure. This comes
10370 at a small price though: as much of the metadata is read
10371 asynchronously from /proc/ (and isn't implicitly attached to log
10372 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
10373 metadata stored alongside a log entry might be slightly
10374 out-of-date. Previously it could only be slightly newer than the log
10375 message. The time window is small however, and given that the kernel
10376 is unlikely to be improved anytime soon in this regard, this appears
10379 * nss-myhostname/systemd-resolved will now by default synthesize an
10380 A/AAAA resource record for the "_gateway" hostname, pointing to the
10381 current default IP gateway. Previously it did that for the "gateway"
10382 name, hampering adoption, as some distributions wanted to leave that
10383 hostname open for local use. The old behaviour may still be
10384 requested at build time.
10386 * systemd-networkd's [Address] section in .network files gained a new
10387 Scope= setting for configuring the IP address scope. The [Network]
10388 section gained a new boolean setting ConfigureWithoutCarrier= that
10389 tells systemd-networkd to ignore link sensing when configuring the
10390 device. The [DHCP] section gained a new Anonymize= boolean option for
10391 turning on a number of options suggested in RFC 7844. A new
10392 [RoutingPolicyRule] section has been added for configuring the IP
10393 routing policy. The [Route] section has gained support for a new
10394 Type= setting which permits configuring
10395 blackhole/unreachable/prohibit routes.
10397 * The [VRF] section in .netdev files gained a new Table= setting for
10398 configuring the routing table to use. The [Tunnel] section gained a
10399 new Independent= boolean field for configuring tunnels independent of
10400 an underlying network interface. The [Bridge] section gained a new
10401 GroupForwardMask= option for configuration of propagation of link
10402 local frames between bridge ports.
10404 * The WakeOnLan= setting in .link files gained support for a number of
10405 new modes. A new TCP6SegmentationOffload= setting has been added for
10406 configuring TCP/IPv6 hardware segmentation offload.
10408 * The IPv6 RA sender implementation may now optionally send out RDNSS
10409 and RDNSSL records to supply DNS configuration to peers.
10411 * systemd-nspawn gained support for a new --system-call-filter= command
10412 line option for adding and removing entries in the default system
10413 call filter it applies. Moreover systemd-nspawn has been changed to
10414 implement a system call allow list instead of a deny list.
10416 * systemd-run gained support for a new --pipe command line option. If
10417 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
10418 are directly passed on to the activated transient service
10419 executable. This allows invoking arbitrary processes as systemd
10420 services (for example to take benefit of dependency management,
10421 accounting management, resource management or log management that is
10422 done automatically for services) — while still allowing them to be
10423 integrated in a classic UNIX shell pipeline.
10425 * When a service sends RELOAD=1 via sd_notify() and reload propagation
10426 using ReloadPropagationTo= is configured, a reload is now propagated
10427 to configured units. (Previously this was only done on explicitly
10428 requested reloads, using "systemctl reload" or an equivalent
10431 * For each service unit a restart counter is now kept: it is increased
10432 each time the service is restarted due to Restart=, and may be
10433 queried using "systemctl show -p NRestarts …".
10435 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
10436 @signal and @timer have been added, for usage with SystemCallFilter=
10437 in unit files and the new --system-call-filter= command line option
10438 of systemd-nspawn (see above).
10440 * ExecStart= lines in unit files gained two new modifiers: when a
10441 command line is prefixed with "!" the command will be executed as
10442 configured, except for the credentials applied by
10443 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
10444 "+", but does still apply namespacing options unlike "+". There's
10445 also "!!" now, which is mostly identical, but becomes a NOP on
10446 systems that support ambient capabilities. This is useful to write
10447 unit files that work with ambient capabilities where possible but
10448 automatically fall back to traditional privilege dropping mechanisms
10449 on systems where this is not supported.
10451 * ListenNetlink= settings in socket units now support RDMA netlink
10454 * A new unit file setting LockPersonality= has been added which permits
10455 locking down the chosen execution domain ("personality") of a service
10458 * A new special target "getty-pre.target" has been added, which is
10459 ordered before all text logins, and may be used to order services
10460 before textual logins acquire access to the console.
10462 * systemd will now attempt to load the virtio-rng.ko kernel module very
10463 early on if a VM environment supporting this is detected. This should
10464 improve entropy during early boot in virtualized environments.
10466 * A _netdev option is now supported in /etc/crypttab that operates in a
10467 similar way as the same option in /etc/fstab: it permits configuring
10468 encrypted devices that need to be ordered after the network is up.
10469 Following this logic, two new special targets
10470 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
10471 added that are to cryptsetup.target what remote-fs.target and
10472 remote-fs-pre.target are to local-fs.target.
10474 * Service units gained a new UnsetEnvironment= setting which permits
10475 unsetting specific environment variables for services that are
10476 normally passed to it (for example in order to mask out locale
10477 settings for specific services that can't deal with it).
10479 * Units acquired a new boolean option IPAccounting=. When turned on, IP
10480 traffic accounting (packet count as well as byte count) is done for
10481 the service, and shown as part of "systemctl status" or "systemd-run
10484 * Service units acquired two new options IPAddressAllow= and
10485 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
10486 for configuring a simple IP access control list for all sockets of
10487 the unit. These options are available also on .slice and .socket
10488 units, permitting flexible access list configuration for individual
10489 services as well as groups of services (as defined by a slice unit),
10490 including system-wide. Note that IP ACLs configured this way are
10491 enforced on every single IPv4 and IPv6 socket created by any process
10492 of the service unit, and apply to ingress as well as egress traffic.
10494 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
10495 structured log message is generated each time the unit is stopped,
10496 containing information about the consumed resources of this
10499 * A new setting KeyringMode= has been added to unit files, which may be
10500 used to control how the kernel keyring is set up for executed
10503 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
10504 "systemctl kexec" and "systemctl exit" are now always asynchronous in
10505 behaviour (that is: these commands return immediately after the
10506 operation was enqueued instead of waiting for the operation to
10507 complete). Previously, "systemctl poweroff" and "systemctl reboot"
10508 were asynchronous on systems using systemd-logind (i.e. almost
10509 always, and like they were on sysvinit), and the other three commands
10510 were unconditionally synchronous. With this release this is cleaned
10511 up, and callers will see the same asynchronous behaviour on all
10512 systems for all five operations.
10514 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
10517 * .timer units now accept calendar specifications in other timezones
10518 than UTC or the local timezone.
10520 * The tmpfiles snippet var.conf has been changed to create
10521 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
10522 the "utmp" group already, and it appears to be generally understood
10523 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
10524 databases. Previously this was implemented correctly for all these
10525 databases excepts btmp, which has been opened up like this now
10526 too. Note that while the other databases are world-readable
10527 (i.e. 0644), btmp is not and remains more restrictive.
10529 * The systemd-resolve tool gained a new --reset-server-features
10530 switch. When invoked like this systemd-resolved will forget
10531 everything it learnt about the features supported by the configured
10532 upstream DNS servers, and restarts the feature probing logic on the
10533 next resolver look-up for them at the highest feature level
10536 * The status dump systemd-resolved sends to the logs upon receiving
10537 SIGUSR1 now also includes information about all DNS servers it is
10538 configured to use, and the features levels it probed for them.
10540 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
10541 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
10542 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
10543 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
10544 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
10545 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
10546 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
10547 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
10548 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
10549 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
10550 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
10551 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
10552 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
10553 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
10554 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
10555 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
10556 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
10557 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
10558 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
10559 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
10561 — Berlin, 2017-10-06
10565 * Meson is now supported as build system in addition to Automake. It is
10566 our plan to remove Automake in one of our next releases, so that
10567 Meson becomes our exclusive build system. Hence, please start using
10568 the Meson build system in your downstream packaging. There's plenty
10569 of documentation around how to use Meson, the extremely brief
10572 ./autogen.sh && ./configure && make && sudo make install
10576 meson build && ninja -C build && sudo ninja -C build install
10578 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
10579 which permits configuring a timeout on the time a job is
10580 running. This is particularly useful for setting timeouts on jobs for
10583 * Unit files gained two new options ConditionUser= and ConditionGroup=
10584 for conditionalizing units based on the identity of the user/group
10585 running a systemd user instance.
10587 * systemd-networkd now understands a new FlowLabel= setting in the
10588 [VXLAN] section of .network files, as well as a Priority= in
10589 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
10590 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
10591 gained support for configuration of GENEVE links, and IPv6 address
10592 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
10594 * .link files now understand a new Port= setting.
10596 * systemd-networkd's DHCP support gained support for DHCP option 119
10597 (domain search list).
10599 * systemd-networkd gained support for serving IPv6 address ranges using
10600 the Router Advertisement protocol. The new .network configuration
10601 section [IPv6Prefix] may be used to configure the ranges to
10602 serve. This is implemented based on a new, minimal, native server
10603 implementation of RA.
10605 * journalctl's --output= switch gained support for a new parameter
10606 "short-iso-precise" for a mode where timestamps are shown as precise
10609 * systemd-udevd's "net_id" builtin may now generate stable network
10610 interface names from IBM PowerVM VIO devices as well as ACPI platform
10613 * MulticastDNS support in systemd-resolved may now be explicitly
10614 enabled/disabled using the new MulticastDNS= configuration file
10617 * systemd-resolved may now optionally use libidn2 instead of the libidn
10618 for processing internationalized domain names. Support for libidn2
10619 should be considered experimental and should not be enabled by
10622 * "machinectl pull-tar" and related call may now do verification of
10623 downloaded images using SUSE-style .sha256 checksum files in addition
10624 to the already existing support for validating using Ubuntu-style
10627 * sd-bus gained support for a new sd_bus_message_appendv() call which
10628 is va_list equivalent of sd_bus_message_append().
10630 * sd-boot gained support for validating images using SHIM/MOK.
10632 * The SMACK code learnt support for "onlycap".
10634 * systemd-mount --umount is now much smarter in figuring out how to
10635 properly unmount a device given its mount or device path.
10637 * The code to call libnss_dns as a fallback from libnss_resolve when
10638 the communication with systemd-resolved fails was removed. This
10639 fallback was redundant and interfered with the [!UNAVAIL=return]
10640 suffix. See nss-resolve(8) for the recommended configuration.
10642 * systemd-logind may now be restarted without losing state. It stores
10643 the file descriptors for devices it manages in the system manager
10644 using the FDSTORE= mechanism. Please note that further changes in
10645 other components may be required to make use of this (for example
10646 Xorg has code to listen for stops of systemd-logind and terminate
10647 itself when logind is stopped or restarted, in order to avoid using
10648 stale file descriptors for graphical devices, which is now
10649 counterproductive and must be reverted in order for restarts of
10650 systemd-logind to be safe. See
10651 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
10653 * All kernel-install plugins are called with the environment variable
10654 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
10655 /etc/machine-id. If the machine ID could not be determined,
10656 $KERNEL_INSTALL_MACHINE_ID will be empty. Plugins should not put
10657 anything in the entry directory (passed as the second argument) if
10658 $KERNEL_INSTALL_MACHINE_ID is empty. For backwards compatibility, a
10659 temporary directory is passed as the entry directory and removed
10660 after all the plugins exit.
10662 * If KERNEL_INSTALL_MACHINE_ID is set in /etc/machine-info, kernel-install
10663 will now use its value as the machine ID instead of the machine ID
10664 from /etc/machine-id. If KERNEL_INSTALL_MACHINE_ID isn't set in
10665 /etc/machine-info and no machine ID is set in /etc/machine-id,
10666 kernel-install will try to store the current machine ID there as
10667 KERNEL_INSTALL_MACHINE_ID. If there is no machine ID, kernel-install
10668 will generate a new UUID, store it in /etc/machine-info as
10669 KERNEL_INSTALL_MACHINE_ID and use it as the machine ID.
10671 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
10672 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
10673 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
10674 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
10675 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
10676 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
10677 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
10678 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
10679 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
10680 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
10681 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
10682 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
10683 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
10684 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
10685 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
10686 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
10687 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
10688 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
10689 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
10690 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
10691 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
10692 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
10693 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
10694 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
10695 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
10696 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
10697 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
10698 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
10699 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
10702 — Berlin, 2017-07-12
10706 * The "hybrid" control group mode has been modified to improve
10707 compatibility with "legacy" cgroups-v1 setups. Specifically, the
10708 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
10709 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
10710 cgroups-v1 hierarchy), the only externally visible change being that
10711 the cgroups-v2 hierarchy is also mounted, to
10712 /sys/fs/cgroup/unified. This should provide a large degree of
10713 compatibility with "legacy" cgroups-v1, while taking benefit of the
10714 better management capabilities of cgroups-v2.
10716 * The default control group setup mode may be selected both a boot-time
10717 via a set of kernel command line parameters (specifically:
10718 systemd.unified_cgroup_hierarchy= and
10719 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
10720 default selected on the configure command line
10721 (--with-default-hierarchy=). The upstream default is "hybrid"
10722 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
10723 this will change in a future systemd version to be "unified" (pure
10724 cgroups-v2 mode). The third option for the compile time option is
10725 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
10726 distributions to default to "hybrid" mode for release distributions,
10727 starting with v233. We recommend "unified" for development
10728 distributions (specifically: distributions such as Fedora's rawhide)
10729 as that's where things are headed in the long run. Use "legacy" for
10730 greatest stability and compatibility only.
10732 * Note one current limitation of "unified" and "hybrid" control group
10733 setup modes: the kernel currently does not permit the systemd --user
10734 instance (i.e. unprivileged code) to migrate processes between two
10735 disconnected cgroup subtrees, even if both are managed and owned by
10736 the user. This effectively means "systemd-run --user --scope" doesn't
10737 work when invoked from outside of any "systemd --user" service or
10738 scope. Specifically, it is not supported from session scopes. We are
10739 working on fixing this in a future systemd version. (See #3388 for
10740 further details about this.)
10742 * DBus policy files are now installed into /usr rather than /etc. Make
10743 sure your system has dbus >= 1.9.18 running before upgrading to this
10744 version, or override the install path with --with-dbuspolicydir= .
10746 * All python scripts shipped with systemd (specifically: the various
10747 tests written in Python) now require Python 3.
10749 * systemd unit tests can now run standalone (without the source or
10750 build directories), and can be installed into /usr/lib/systemd/tests/
10751 with 'make install-tests'.
10753 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
10754 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
10757 * Support for the %c, %r, %R specifiers in unit files has been
10758 removed. Specifiers are not supposed to be dependent on configuration
10759 in the unit file itself (so that they resolve the same regardless
10760 where used in the unit files), but these specifiers were influenced
10761 by the Slice= option.
10763 * The shell invoked by debug-shell.service now defaults to /bin/sh in
10764 all cases. If distributions want to use a different shell for this
10765 purpose (for example Fedora's /sbin/sushell) they need to specify
10766 this explicitly at configure time using --with-debug-shell=.
10768 * The confirmation spawn prompt has been reworked to offer the
10771 (c)ontinue, proceed without asking anymore
10772 (D)ump, show the state of the unit
10773 (f)ail, don't execute the command and pretend it failed
10775 (i)nfo, show a short summary of the unit
10776 (j)obs, show jobs that are in progress
10777 (s)kip, don't execute the command and pretend it succeeded
10778 (y)es, execute the command
10780 The 'n' choice for the confirmation spawn prompt has been removed,
10781 because its meaning was confusing.
10783 The prompt may now also be redirected to an alternative console by
10784 specifying the console as parameter to systemd.confirm_spawn=.
10786 * Services of Type=notify require a READY=1 notification to be sent
10787 during startup. If no such message is sent, the service now fails,
10788 even if the main process exited with a successful exit code.
10790 * Services that fail to start up correctly now always have their
10791 ExecStopPost= commands executed. Previously, they'd enter "failed"
10792 state directly, without executing these commands.
10794 * The option MulticastDNS= of network configuration files has acquired
10795 an actual implementation. With MulticastDNS=yes a host can resolve
10796 names of remote hosts and reply to mDNS A and AAAA requests.
10798 * When units are about to be started an additional check is now done to
10799 ensure that all dependencies of type BindsTo= (when used in
10800 combination with After=) have been started.
10802 * systemd-analyze gained a new verb "syscall-filter" which shows which
10803 system call groups are defined for the SystemCallFilter= unit file
10804 setting, and which system calls they contain.
10806 * A new system call filter group "@filesystem" has been added,
10807 consisting of various file system related system calls. Group
10808 "@reboot" has been added, covering reboot, kexec and shutdown related
10809 calls. Finally, group "@swap" has been added covering swap
10810 configuration related calls.
10812 * A new unit file option RestrictNamespaces= has been added that may be
10813 used to restrict access to the various process namespace types the
10814 Linux kernel provides. Specifically, it may be used to take away the
10815 right for a service unit to create additional file system, network,
10816 user, and other namespaces. This sandboxing option is particularly
10817 relevant due to the high amount of recently discovered namespacing
10818 related vulnerabilities in the kernel.
10820 * systemd-udev's .link files gained support for a new AutoNegotiation=
10821 setting for configuring Ethernet auto-negotiation.
10823 * systemd-networkd's .network files gained support for a new
10824 ListenPort= setting in the [DHCP] section to explicitly configure the
10825 UDP client port the DHCP client shall listen on.
10827 * .network files gained a new Unmanaged= boolean setting for explicitly
10828 excluding one or more interfaces from management by systemd-networkd.
10830 * The systemd-networkd ProxyARP= option has been renamed to
10831 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
10832 renamed to ReduceARPProxy=. The old names continue to be available
10835 * systemd-networkd gained support for configuring IPv6 Proxy NDP
10836 addresses via the new IPv6ProxyNDPAddress= .network file setting.
10838 * systemd-networkd's bonding device support gained support for two new
10839 configuration options ActiveSlave= and PrimarySlave=.
10841 * The various options in the [Match] section of .network files gained
10842 support for negative matching.
10844 * New systemd-specific mount options are now understood in /etc/fstab:
10846 x-systemd.mount-timeout= may be used to configure the maximum
10847 permitted runtime of the mount command.
10849 x-systemd.device-bound may be set to bind a mount point to its
10850 backing device unit, in order to automatically remove a mount point
10851 if its backing device is unplugged. This option may also be
10852 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
10853 on the block device, which is now automatically set for all CDROM
10854 drives, so that mounted CDs are automatically unmounted when they are
10855 removed from the drive.
10857 x-systemd.after= and x-systemd.before= may be used to explicitly
10858 order a mount after or before another unit or mount point.
10860 * Enqueued start jobs for device units are now automatically garbage
10861 collected if there are no jobs waiting for them anymore.
10863 * systemctl list-jobs gained two new switches: with --after, for every
10864 queued job the jobs it's waiting for are shown; with --before the
10865 jobs which it's blocking are shown.
10867 * systemd-nspawn gained support for ephemeral boots from disk images
10868 (or in other words: --ephemeral and --image= may now be
10869 combined). Moreover, ephemeral boots are now supported for normal
10870 directories, even if the backing file system is not btrfs. Of course,
10871 if the file system does not support file system snapshots or
10872 reflinks, the initial copy operation will be relatively expensive, but
10873 this should still be suitable for many use cases.
10875 * Calendar time specifications in .timer units now support
10876 specifications relative to the end of a month by using "~" instead of
10877 "-" as separator between month and day. For example, "*-02~03" means
10878 "the third last day in February". In addition a new syntax for
10879 repeated events has been added using the "/" character. For example,
10880 "9..17/2:00" means "every two hours from 9am to 5pm".
10882 * systemd-socket-proxyd gained a new parameter --connections-max= for
10883 configuring the maximum number of concurrent connections.
10885 * sd-id128 gained a new API for generating unique IDs for the host in a
10886 way that does not leak the machine ID. Specifically,
10887 sd_id128_get_machine_app_specific() derives an ID based on the
10888 machine ID in a well-defined, non-reversible, stable way. This is
10889 useful whenever an identifier for the host is needed but where the
10890 identifier shall not be useful to identify the system beyond the
10891 scope of the application itself. (Internally this uses HMAC-SHA256 as
10892 keyed hash function using the machine ID as input.)
10894 * NotifyAccess= gained a new supported value "exec". When set
10895 notifications are accepted from all processes systemd itself invoked,
10896 including all control processes.
10898 * .nspawn files gained support for defining overlay mounts using the
10899 Overlay= and OverlayReadOnly= options. Previously this functionality
10900 was only available on the systemd-nspawn command line.
10902 * systemd-nspawn's --bind= and --overlay= options gained support for
10903 bind/overlay mounts whose source lies within the container tree by
10904 prefixing the source path with "+".
10906 * systemd-nspawn's --bind= and --overlay= options gained support for
10907 automatically allocating a temporary source directory in /var/tmp
10908 that is removed when the container dies. Specifically, if the source
10909 directory is specified as empty string this mechanism is selected. An
10910 example usage is --overlay=+/var::/var, which creates an overlay
10911 mount based on the original /var contained in the image, overlaid
10912 with a temporary directory in the host's /var/tmp. This way changes
10913 to /var are automatically flushed when the container shuts down.
10915 * systemd-nspawn --image= option does now permit raw file system block
10916 devices (in addition to images containing partition tables, as
10919 * The disk image dissection logic in systemd-nspawn gained support for
10920 automatically setting up LUKS encrypted as well as Verity protected
10921 partitions. When a container is booted from an encrypted image the
10922 passphrase is queried at start-up time. When a container with Verity
10923 data is started, the root hash is search in a ".roothash" file
10924 accompanying the disk image (alternatively, pass the root hash via
10925 the new --root-hash= command line option).
10927 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
10928 be used to dissect disk images the same way as systemd-nspawn does
10929 it, following the Bootable Partition Specification. It may even be
10930 used to mount disk images with complex partition setups (including
10931 LUKS and Verity partitions) to a local host directory, in order to
10932 inspect them. This tool is not considered public API (yet), and is
10933 thus not installed into /usr/bin. Please do not rely on its
10934 existence, since it might go away or be changed in later systemd
10937 * A new generator "systemd-verity-generator" has been added, similar in
10938 style to "systemd-cryptsetup-generator", permitting automatic setup of
10939 Verity root partitions when systemd boots up. In order to make use of
10940 this your partition setup should follow the Discoverable Partitions
10941 Specification, and the GPT partition ID of the root file system
10942 partition should be identical to the upper 128-bit of the Verity root
10943 hash. The GPT partition ID of the Verity partition protecting it
10944 should be the lower 128-bit of the Verity root hash. If the partition
10945 image follows this model it is sufficient to specify a single
10946 "roothash=" kernel command line argument to both configure which root
10947 image and verity partition to use as well as the root hash for
10948 it. Note that systemd-nspawn's Verity support follows the same
10949 semantics, meaning that disk images with proper Verity data in place
10950 may be booted in containers with systemd-nspawn as well as on
10951 physical systems via the verity generator. Also note that the "mkosi"
10952 tool available at https://github.com/systemd/mkosi has been updated
10953 to generate Verity protected disk images following this scheme. In
10954 fact, it has been updated to generate disk images that optionally
10955 implement a complete UEFI SecureBoot trust chain, involving a signed
10956 kernel and initrd image that incorporates such a root hash as well as
10957 a Verity-enabled root partition.
10959 * The hardware database (hwdb) udev supports has been updated to carry
10960 accelerometer quirks.
10962 * All system services are now run with a fresh kernel keyring set up
10963 for them. The invocation ID is stored by default in it, thus
10964 providing a safe, non-overridable way to determine the invocation
10965 ID of each service.
10967 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
10968 options for bind mounting arbitrary paths in a service-specific
10969 way. When these options are used, arbitrary host or service files and
10970 directories may be mounted to arbitrary locations in the service's
10973 * Documentation has been added that lists all of systemd's low-level
10974 environment variables:
10976 https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md
10978 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
10979 whether a specific socket file descriptor matches a specified socket
10982 * systemd-firstboot has been updated to check for the
10983 systemd.firstboot= kernel command line option. It accepts a boolean
10984 and when set to false the first boot questions are skipped.
10986 * systemd-fstab-generator has been updated to check for the
10987 systemd.volatile= kernel command line option, which either takes an
10988 optional boolean parameter or the special value "state". If used the
10989 system may be booted in a "volatile" boot mode. Specifically,
10990 "systemd.volatile" is used, the root directory will be mounted as
10991 tmpfs, and only /usr is mounted from the actual root file system. If
10992 "systemd.volatile=state" is used, the root directory will be mounted
10993 as usual, but /var is mounted as tmpfs. This concept provides similar
10994 functionality as systemd-nspawn's --volatile= option, but provides it
10995 on physical boots. Use this option for implementing stateless
10996 systems, or testing systems with all state and/or configuration reset
10997 to the defaults. (Note though that many distributions are not
10998 prepared to boot up without a populated /etc or /var, though.)
11000 * systemd-gpt-auto-generator gained support for LUKS encrypted root
11001 partitions. Previously it only supported LUKS encrypted partitions
11002 for all other uses, except for the root partition itself.
11004 * Socket units gained support for listening on AF_VSOCK sockets for
11005 communication in virtualized QEMU environments.
11007 * The "configure" script gained a new option --with-fallback-hostname=
11008 for specifying the fallback hostname to use if none is configured in
11009 /etc/hostname. For example, by specifying
11010 --with-fallback-hostname=fedora it is possible to default to a
11011 hostname of "fedora" on pristine installations.
11013 * systemd-cgls gained support for a new --unit= switch for listing only
11014 the control groups of a specific unit. Similar --user-unit= has been
11015 added for listing only the control groups of a specific user unit.
11017 * systemd-mount gained a new --umount switch for unmounting a mount or
11018 automount point (and all mount/automount points below it).
11020 * systemd will now refuse full configuration reloads (via systemctl
11021 daemon-reload and related calls) unless at least 16MiB of free space
11022 are available in /run. This is a safety precaution in order to ensure
11023 that generators can safely operate after the reload completed.
11025 * A new unit file option RootImage= has been added, which has a similar
11026 effect as RootDirectory= but mounts the service's root directory from
11027 a disk image instead of plain directory. This logic reuses the same
11028 image dissection and mount logic that systemd-nspawn already uses,
11029 and hence supports any disk images systemd-nspawn supports, including
11030 those following the Discoverable Partition Specification, as well as
11031 Verity enabled images. This option enables systemd to run system
11032 services directly off disk images acting as resource bundles,
11033 possibly even including full integrity data.
11035 * A new MountAPIVFS= unit file option has been added, taking a boolean
11036 argument. If enabled /proc, /sys and /dev (collectively called the
11037 "API VFS") will be mounted for the service. This is only relevant if
11038 RootDirectory= or RootImage= is used for the service, as these mounts
11039 are of course in place in the host mount namespace anyway.
11041 * systemd-nspawn gained support for a new --pivot-root= switch. If
11042 specified the root directory within the container image is pivoted to
11043 the specified mount point, while the original root disk is moved to a
11044 different place. This option enables booting of ostree images
11045 directly with systemd-nspawn.
11047 * The systemd build scripts will no longer complain if the NTP server
11048 addresses are not changed from the defaults. Google now supports
11049 these NTP servers officially. We still recommend downstreams to
11050 properly register an NTP pool with the NTP pool project though.
11052 * coredumpctl gained a new "--reverse" option for printing the list
11053 of coredumps in reverse order.
11055 * coredumpctl will now show additional information about truncated and
11056 inaccessible coredumps, as well as coredumps that are still being
11057 processed. It also gained a new --quiet switch for suppressing
11058 additional informational message in its output.
11060 * coredumpctl gained support for only showing coredumps newer and/or
11061 older than specific timestamps, using the new --since= and --until=
11062 options, reminiscent of journalctl's options by the same name.
11064 * The systemd-coredump logic has been improved so that it may be reused
11065 to collect backtraces in non-compiled languages, for example in
11066 scripting languages such as Python.
11068 * machinectl will now show the UID shift of local containers, if user
11069 namespacing is enabled for them.
11071 * systemd will now optionally run "environment generator" binaries at
11072 configuration load time. They may be used to add environment
11073 variables to the environment block passed to services invoked. One
11074 user environment generator is shipped by default that sets up
11075 environment variables based on files dropped into /etc/environment.d
11076 and ~/.config/environment.d/.
11078 * systemd-resolved now includes the new, recently published 2017 DNSSEC
11081 * hostnamed has been updated to report a new chassis type of
11082 "convertible" to cover "foldable" laptops that can both act as a
11083 tablet and as a laptop, such as various Lenovo Yoga devices.
11085 Contributions from: Adrián López, Alexander Galanin, Alexander
11086 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
11087 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
11088 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
11089 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
11090 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
11091 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
11092 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
11093 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
11094 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
11095 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
11096 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
11097 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
11098 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
11099 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
11100 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
11101 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
11102 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
11103 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
11104 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
11105 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
11106 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
11107 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
11108 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
11109 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
11110 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
11111 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
11112 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
11115 — Berlin, 2017-03-01
11119 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
11120 RestrictAddressFamilies= enabled. These sandboxing options should
11121 generally be compatible with the various external udev call-out
11122 binaries we are aware of, however there may be exceptions, in
11123 particular when exotic languages for these call-outs are used. In
11124 this case, consider turning off these settings locally.
11126 * The new RemoveIPC= option can be used to remove IPC objects owned by
11127 the user or group of a service when that service exits.
11129 * The new ProtectKernelModules= option can be used to disable explicit
11130 load and unload operations of kernel modules by a service. In
11131 addition access to /usr/lib/modules is removed if this option is set.
11133 * ProtectSystem= option gained a new value "strict", which causes the
11134 whole file system tree with the exception of /dev, /proc, and /sys,
11135 to be remounted read-only for a service.
11137 * The new ProtectKernelTunables= option can be used to disable
11138 modification of configuration files in /sys and /proc by a service.
11139 Various directories and files are remounted read-only, so access is
11140 restricted even if the file permissions would allow it.
11142 * The new ProtectControlGroups= option can be used to disable write
11143 access by a service to /sys/fs/cgroup.
11145 * Various systemd services have been hardened with
11146 ProtectKernelTunables=yes, ProtectControlGroups=yes,
11147 RestrictAddressFamilies=.
11149 * Support for dynamically creating users for the lifetime of a service
11150 has been added. If DynamicUser=yes is specified, user and group IDs
11151 will be allocated from the range 61184…65519 for the lifetime of the
11152 service. They can be resolved using the new nss-systemd.so NSS
11153 module. The module must be enabled in /etc/nsswitch.conf. Services
11154 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
11155 any resources allocated by the service will be cleaned up when the
11156 service exits. They also have ProtectHome=read-only and
11157 ProtectSystem=strict enabled, so they are not able to make any
11158 permanent modifications to the system.
11160 * The nss-systemd module also always resolves root and nobody, making
11161 it possible to have no /etc/passwd or /etc/group files in minimal
11162 container or chroot environments.
11164 * Services may be started with their own user namespace using the new
11165 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
11166 under which the service is running are mapped. All other users are
11169 * Support for the cgroup namespace has been added to systemd-nspawn. If
11170 supported by kernel, the container system started by systemd-nspawn
11171 will have its own view of the cgroup hierarchy. This new behaviour
11172 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
11174 * The new MemorySwapMax= option can be used to limit the maximum swap
11175 usage under the unified cgroup hierarchy.
11177 * Support for the CPU controller in the unified cgroup hierarchy has
11178 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
11179 options. This controller requires out-of-tree patches for the kernel
11180 and the support is provisional.
11182 * Mount and automount units may now be created transiently
11183 (i.e. dynamically at runtime via the bus API, instead of requiring
11184 unit files in the file system).
11186 * systemd-mount is a new tool which may mount file systems – much like
11187 mount(8), optionally pulling in additional dependencies through
11188 transient .mount and .automount units. For example, this tool
11189 automatically runs fsck on a backing block device before mounting,
11190 and allows the automount logic to be used dynamically from the
11191 command line for establishing mount points. This tool is particularly
11192 useful when dealing with removable media, as it will ensure fsck is
11193 run – if necessary – before the first access and that the file system
11194 is quickly unmounted after each access by utilizing the automount
11195 logic. This maximizes the chance that the file system on the
11196 removable media stays in a clean state, and if it isn't in a clean
11197 state is fixed automatically.
11199 * LazyUnmount=yes option for mount units has been added to expose the
11200 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
11203 * /efi will be used as the mount point of the EFI boot partition, if
11204 the directory is present, and the mount point was not configured
11205 through other means (e.g. fstab). If /efi directory does not exist,
11206 /boot will be used as before. This makes it easier to automatically
11207 mount the EFI partition on systems where /boot is used for something
11210 * When operating on GPT disk images for containers, systemd-nspawn will
11211 now mount the ESP to /boot or /efi according to the same rules as PID
11212 1 running on a host. This allows tools like "bootctl" to operate
11213 correctly within such containers, in order to make container images
11214 bootable on physical systems.
11216 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
11218 * Two new user session targets have been added to support running
11219 graphical sessions under the systemd --user instance:
11220 graphical-session.target and graphical-session-pre.target. See
11221 systemd.special(7) for a description of how those targets should be
11224 * The vconsole initialization code has been significantly reworked to
11225 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
11226 support unicode keymaps. Font and keymap configuration will now be
11227 copied to all allocated virtual consoles.
11229 * FreeBSD's bhyve virtualization is now detected.
11231 * Information recorded in the journal for core dumps now includes the
11232 contents of /proc/mountinfo and the command line of the process at
11233 the top of the process hierarchy (which is usually the init process
11236 * systemd-journal-gatewayd learned the --directory= option to serve
11237 files from the specified location.
11239 * journalctl --root=… can be used to peruse the journal in the
11240 /var/log/ directories inside of a container tree. This is similar to
11241 the existing --machine= option, but does not require the container to
11244 * The hardware database has been extended to support
11245 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
11248 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
11249 specify the click rate for mice which include a horizontal wheel with
11250 a click rate that is different than the one for the vertical wheel.
11252 * systemd-run gained a new --wait option that makes service execution
11253 synchronous. (Specifically, the command will not return until the
11254 specified service binary exited.)
11256 * systemctl gained a new --wait option that causes the start command to
11257 wait until the units being started have terminated again.
11259 * A new journal output mode "short-full" has been added which displays
11260 timestamps with abbreviated English day names and adds a timezone
11261 suffix. Those timestamps include more information than the default
11262 "short" output mode, and can be passed directly to journalctl's
11263 --since= and --until= options.
11265 * /etc/resolv.conf will be bind-mounted into containers started by
11266 systemd-nspawn, if possible, so any changes to resolv.conf contents
11267 are automatically propagated to the container.
11269 * The number of instances for socket-activated services originating
11270 from a single IP address can be limited with
11271 MaxConnectionsPerSource=, extending the existing setting of
11274 * systemd-networkd gained support for vcan ("Virtual CAN") interface
11277 * .netdev and .network configuration can now be extended through
11280 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
11281 Segmentation Offload, Generic Receive Offload, Large Receive Offload
11282 can be enabled and disabled using the new UDPSegmentationOffload=,
11283 TCPSegmentationOffload=, GenericSegmentationOffload=,
11284 GenericReceiveOffload=, LargeReceiveOffload= options in the
11285 [Link] section of .link files.
11287 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
11288 Port VLAN ID can be configured for bridge devices using the new STP=,
11289 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
11290 section of .netdev files.
11292 * The route table to which routes received over DHCP or RA should be
11293 added can be configured with the new RouteTable= option in the [DHCP]
11294 and [IPv6AcceptRA] sections of .network files.
11296 * The Address Resolution Protocol can be disabled on links managed by
11297 systemd-networkd using the ARP=no setting in the [Link] section of
11300 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
11301 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
11302 encode information about the result and exit codes of the current
11303 service runtime cycle.
11305 * systemd-sysctl will now configure kernel parameters in the order
11306 they occur in the configuration files. This matches what sysctl
11307 has been traditionally doing.
11309 * kernel-install "plugins" that are executed to perform various
11310 tasks after a new kernel is added and before an old one is removed
11311 can now return a special value to terminate the procedure and
11312 prevent any later plugins from running.
11314 * Journald's SplitMode=login setting has been deprecated. It has been
11315 removed from documentation, and its use is discouraged. In a future
11316 release it will be completely removed, and made equivalent to current
11317 default of SplitMode=uid.
11319 * Storage=both option setting in /etc/systemd/coredump.conf has been
11320 removed. With fast LZ4 compression storing the core dump twice is not
11323 * The --share-system systemd-nspawn option has been replaced with an
11324 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
11325 this functionality is discouraged. In addition the variables
11326 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
11327 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
11328 individual namespaces.
11330 * "machinectl list" now shows the IP address of running containers in
11331 the output, as well as OS release information.
11333 * "loginctl list" now shows the TTY of each session in the output.
11335 * sd-bus gained new API calls sd_bus_track_set_recursive(),
11336 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
11337 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
11338 tracking objects in a "recursive" mode, where a single client can be
11339 counted multiple times, if it takes multiple references.
11341 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
11342 sd_bus_get_exit_on_disconnect(). They may be used to make a
11343 process using sd-bus automatically exit if the bus connection is
11346 * Bus clients of the service manager may now "pin" loaded units into
11347 memory, by taking an explicit reference on them. This is useful to
11348 ensure the client can retrieve runtime data about the service even
11349 after the service completed execution. Taking such a reference is
11350 available only for privileged clients and should be helpful to watch
11351 running services in a race-free manner, and in particular collect
11352 information about exit statuses and results.
11354 * The nss-resolve module has been changed to strictly return UNAVAIL
11355 when communication via D-Bus with resolved failed, and NOTFOUND when
11356 a lookup completed but was negative. This means it is now possible to
11357 neatly configure fallbacks using nsswitch.conf result checking
11358 expressions. Taking benefit of this, the new recommended
11359 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
11361 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
11363 * A new setting CtrlAltDelBurstAction= has been added to
11364 /etc/systemd/system.conf which may be used to configure the precise
11365 behaviour if the user on the console presses Ctrl-Alt-Del more often
11366 than 7 times in 2s. Previously this would unconditionally result in
11367 an expedited, immediate reboot. With this new setting the precise
11368 operation may be configured in more detail, and also turned off
11371 * In .netdev files two new settings RemoteChecksumTx= and
11372 RemoteChecksumRx= are now understood that permit configuring the
11373 remote checksumming logic for VXLAN networks.
11375 * The service manager learnt a new "invocation ID" concept for invoked
11376 services. Each runtime cycle of a service will get a new invocation
11377 ID (a 128-bit random UUID) assigned that identifies the current
11378 run of the service uniquely and globally. A new invocation ID
11379 is generated each time a service starts up. The journal will store
11380 the invocation ID of a service along with any logged messages, thus
11381 making the invocation ID useful for matching the online runtime of a
11382 service with the offline log data it generated in a safe way without
11383 relying on synchronized timestamps. In many ways this new service
11384 invocation ID concept is similar to the kernel's boot ID concept that
11385 uniquely and globally identifies the runtime of each boot. The
11386 invocation ID of a service is passed to the service itself via an
11387 environment variable ($INVOCATION_ID). A new bus call
11388 GetUnitByInvocationID() has been added that is similar to GetUnit()
11389 but instead of retrieving the bus path for a unit by its name
11390 retrieves it by its invocation ID. The returned path is valid only as
11391 long as the passed invocation ID is current.
11393 * systemd-resolved gained a new "DNSStubListener" setting in
11394 resolved.conf. It either takes a boolean value or the special values
11395 "udp" and "tcp", and configures whether to enable the stub DNS
11396 listener on 127.0.0.53:53.
11398 * IP addresses configured via networkd may now carry additional
11399 configuration settings supported by the kernel. New options include:
11400 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
11401 PrefixRoute=, AutoJoin=.
11403 * The PAM configuration fragment file for "user@.service" shipped with
11404 systemd (i.e. the --user instance of systemd) has been stripped to
11405 the minimum necessary to make the system boot. Previously, it
11406 contained Fedora-specific stanzas that did not apply to other
11407 distributions. It is expected that downstream distributions add
11408 additional configuration lines, matching their needs to this file,
11409 using it only as rough template of what systemd itself needs. Note
11410 that this reduced fragment does not even include an invocation of
11411 pam_limits which most distributions probably want to add, even though
11412 systemd itself does not need it. (There's also the new build time
11413 option --with-pamconfdir=no to disable installation of the PAM
11414 fragment entirely.)
11416 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
11417 capability is now also dropped from its set (in addition to
11418 CAP_SYS_MKNOD as before).
11420 * In service unit files it is now possible to connect a specific named
11421 file descriptor with stdin/stdout/stdout of an executed service. The
11422 name may be specified in matching .socket units using the
11423 FileDescriptorName= setting.
11425 * A number of journal settings may now be configured on the kernel
11426 command line. Specifically, the following options are now understood:
11427 systemd.journald.max_level_console=,
11428 systemd.journald.max_level_store=,
11429 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
11430 systemd.journald.max_level_wall=.
11432 * "systemctl is-enabled --full" will now show by which symlinks a unit
11433 file is enabled in the unit dependency tree.
11435 * Support for VeraCrypt encrypted partitions has been added to the
11436 "cryptsetup" logic and /etc/crypttab.
11438 * systemd-detect-virt gained support for a new --private-users switch
11439 that checks whether the invoking processes are running inside a user
11440 namespace. Similar, a new special value "private-users" for the
11441 existing ConditionVirtualization= setting has been added, permitting
11442 skipping of specific units in user namespace environments.
11444 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
11445 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
11446 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
11447 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
11448 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
11449 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
11450 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
11451 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
11452 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
11453 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
11454 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
11455 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
11456 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
11457 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
11458 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
11459 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
11460 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
11461 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
11462 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
11463 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
11464 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
11465 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
11466 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
11467 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
11468 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
11469 Jędrzejewski-Szmek, Zeal Jagannatha
11471 — Santa Fe, 2016-11-03
11475 * In service units the various ExecXYZ= settings have been extended
11476 with an additional special character as first argument of the
11477 assigned value: if the character '+' is used the specified command
11478 line it will be run with full privileges, regardless of User=,
11479 Group=, CapabilityBoundingSet= and similar options. The effect is
11480 similar to the existing PermissionsStartOnly= option, but allows
11481 configuration of this concept for each executed command line
11484 * Services may now alter the service watchdog timeout at runtime by
11485 sending a WATCHDOG_USEC= message via sd_notify().
11487 * MemoryLimit= and related unit settings now optionally take percentage
11488 specifications. The percentage is taken relative to the amount of
11489 physical memory in the system (or in case of containers, the assigned
11490 amount of memory). This allows scaling service resources neatly with
11491 the amount of RAM available on the system. Similarly, systemd-logind's
11492 RuntimeDirectorySize= option now also optionally takes percentage
11495 * In similar fashion TasksMax= takes percentage values now, too. The
11496 value is taken relative to the configured maximum number of processes
11497 on the system. The per-service task maximum has been changed to 15%
11498 using this functionality. (Effectively this is an increase of 512 →
11499 4915 for service units, given the kernel's default pid_max setting.)
11501 * Calendar time specifications in .timer units now understand a ".."
11502 syntax for time ranges. Example: "4..7:10" may now be used for
11503 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
11506 * The InaccessableDirectories=, ReadOnlyDirectories= and
11507 ReadWriteDirectories= unit file settings have been renamed to
11508 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
11509 applied to all kinds of file nodes, and not just directories, with
11510 the exception of symlinks. Specifically these settings may now be
11511 used on block and character device nodes, UNIX sockets and FIFOS as
11512 well as regular files. The old names of these settings remain
11513 available for compatibility.
11515 * systemd will now log about all service processes it kills forcibly
11516 (using SIGKILL) because they remained after the clean shutdown phase
11517 of the service completed. This should help identifying services that
11518 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
11519 systemd-logind's configuration a similar log message is generated for
11520 processes killed at the end of each session due to this setting.
11522 * systemd will now set the $JOURNAL_STREAM environment variable for all
11523 services whose stdout/stderr are connected to the Journal (which
11524 effectively means by default: all services). The variable contains
11525 the device and inode number of the file descriptor used for
11526 stdout/stderr. This may be used by invoked programs to detect whether
11527 their stdout/stderr is connected to the Journal, in which case they
11528 can switch over to direct Journal communication, thus being able to
11529 pass extended, structured metadata along with their log messages. As
11530 one example, this is now used by glib's logging primitives.
11532 * When using systemd's default tmp.mount unit for /tmp, the mount point
11533 will now be established with the "nosuid" and "nodev" options. This
11534 avoids privilege escalation attacks that put traps and exploits into
11535 /tmp. However, this might cause problems if you e.g. put container
11536 images or overlays into /tmp; if you need this, override tmp.mount's
11537 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
11540 * systemd now supports the "memory" cgroup controller also on
11543 * The systemd-cgtop tool now optionally takes a control group path as
11544 command line argument. If specified, the control group list shown is
11545 limited to subgroups of that group.
11547 * The SystemCallFilter= unit file setting gained support for
11548 pre-defined, named system call filter sets. For example
11549 SystemCallFilter=@clock is now an effective way to make all clock
11550 changing-related system calls unavailable to a service. A number of
11551 similar pre-defined groups are defined. Writing system call filters
11552 for system services is simplified substantially with this new
11553 concept. Accordingly, all of systemd's own, long-running services now
11554 enable system call filtering based on this, by default.
11556 * A new service setting MemoryDenyWriteExecute= has been added, taking
11557 a boolean value. If turned on, a service may no longer create memory
11558 mappings that are writable and executable at the same time. This
11559 enhances security for services where this is enabled as it becomes
11560 harder to dynamically write and then execute memory in exploited
11561 service processes. This option has been enabled for all of systemd's
11562 own long-running services.
11564 * A new RestrictRealtime= service setting has been added, taking a
11565 boolean argument. If set the service's processes may no longer
11566 acquire realtime scheduling. This improves security as realtime
11567 scheduling may otherwise be used to easily freeze the system.
11569 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
11570 value. This may be used for requesting that the system manager inside
11571 of the container reports start-up completion to nspawn which then
11572 propagates this notification further to the service manager
11573 supervising nspawn itself. A related option NotifyReady= in .nspawn
11574 files has been added too. This functionality allows ordering of the
11575 start-up of multiple containers using the usual systemd ordering
11578 * machinectl gained a new command "stop" that is an alias for
11581 * systemd-resolved gained support for contacting DNS servers on
11582 link-local IPv6 addresses.
11584 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
11585 its caches. A method call for requesting the same operation has been
11586 added to the bus API too, and is made available via "systemd-resolve
11589 * systemd-resolve gained a new --status switch. If passed a brief
11590 summary of the used DNS configuration with per-interface information
11593 * resolved.conf gained a new Cache= boolean option, defaulting to
11594 on. If turned off local DNS caching is disabled. This comes with a
11595 performance penalty in particular when DNSSEC is enabled. Note that
11596 resolved disables its internal caching implicitly anyway, when the
11597 configured DNS server is on a host-local IP address such as ::1 or
11598 127.0.0.1, thus automatically avoiding double local caching.
11600 * systemd-resolved now listens on the local IP address 127.0.0.53:53
11601 for DNS requests. This improves compatibility with local programs
11602 that do not use the libc NSS or systemd-resolved's bus APIs for name
11603 resolution. This minimal DNS service is only available to local
11604 programs and does not implement the full DNS protocol, but enough to
11605 cover local DNS clients. A new, static resolv.conf file, listing just
11606 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
11607 now recommended to make /etc/resolv.conf a symlink to this file in
11608 order to route all DNS lookups to systemd-resolved, regardless if
11609 done via NSS, the bus API or raw DNS packets. Note that this local
11610 DNS service is not as fully featured as the libc NSS or
11611 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
11612 used to deliver link-local address information (as this implies
11613 sending a local interface index along), LLMNR/mDNS support via this
11614 interface is severely restricted. It is thus strongly recommended for
11615 all applications to use the libc NSS API or native systemd-resolved
11618 * systemd-networkd's bridge support learned a new setting
11619 VLANFiltering= for controlling VLAN filtering. Moreover a new section
11620 in .network files has been added for configuring VLAN bridging in
11621 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
11623 * systemd-networkd's IPv6 Router Advertisement code now makes use of
11624 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
11625 now be acquired without relying on DHCPv6. Two new options
11626 UseDomains= and UseDNS= have been added to configure this behaviour.
11628 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
11629 renamed IPv6AcceptRA=, without altering its behaviour. The old
11630 setting name remains available for compatibility reasons.
11632 * The systemd-networkd VTI/VTI6 tunneling support gained new options
11633 Key=, InputKey= and OutputKey=.
11635 * systemd-networkd gained support for VRF ("Virtual Routing Function")
11636 interface configuration.
11638 * "systemctl edit" may now be used to create new unit files by
11639 specifying the --force switch.
11641 * sd-event gained a new function sd_event_get_iteration() for
11642 requesting the current iteration counter of the event loop. It starts
11643 at zero and is increased by one with each event loop iteration.
11645 * A new rpm macro %systemd_ordering is provided by the macros.systemd
11646 file. It can be used in lieu of %systemd_requires in packages which
11647 don't use any systemd functionality and are intended to be installed
11648 in minimal containers without systemd present. This macro provides
11649 ordering dependencies to ensure that if the package is installed in
11650 the same rpm transaction as systemd, systemd will be installed before
11651 the scriptlets for the package are executed, allowing unit presets
11654 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
11655 been added to simplify packaging of generators.
11657 * The os-release file gained VERSION_CODENAME field for the
11658 distribution nickname (e.g. VERSION_CODENAME=woody).
11660 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
11661 can be set to disable parsing of metadata and the creation
11662 of persistent symlinks for that device.
11664 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
11665 to make them available to logged-in users has been reverted.
11667 * Much of the common code of the various systemd components is now
11668 built into an internal shared library libsystemd-shared-231.so
11669 (incorporating the systemd version number in the name, to be updated
11670 with future releases) that the components link to. This should
11671 decrease systemd footprint both in memory during runtime and on
11672 disk. Note that the shared library is not for public use, and is
11673 neither API nor ABI stable, but is likely to change with every new
11674 released update. Packagers need to make sure that binaries
11675 linking to libsystemd-shared.so are updated in step with the
11678 * Configuration for "mkosi" is now part of the systemd
11679 repository. mkosi is a tool to easily build legacy-free OS images,
11680 and is available on github: https://github.com/systemd/mkosi. If
11681 "mkosi" is invoked in the build tree a new raw OS image is generated
11682 incorporating the systemd sources currently being worked on and a
11683 clean, fresh distribution installation. The generated OS image may be
11684 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
11685 UEFI PC. This functionality is particularly useful to easily test
11686 local changes made to systemd in a pristine, defined environment. See
11687 doc/HACKING for details.
11689 * configure learned the --with-support-url= option to specify the
11690 distribution's bugtracker.
11692 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
11693 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
11694 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
11695 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
11696 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
11697 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
11698 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
11699 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
11700 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
11701 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
11702 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
11703 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
11704 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
11705 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
11706 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
11707 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
11708 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
11709 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
11710 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
11712 — Berlin, 2016-07-25
11716 * DNSSEC is now turned on by default in systemd-resolved (in
11717 "allow-downgrade" mode), but may be turned off during compile time by
11718 passing "--with-default-dnssec=no" to "configure" (and of course,
11719 during runtime with DNSSEC= in resolved.conf). We recommend
11720 downstreams to leave this on at least during development cycles and
11721 report any issues with the DNSSEC logic upstream. We are very
11722 interested in collecting feedback about the DNSSEC validator and its
11723 limitations in the wild. Note however, that DNSSEC support is
11724 probably nothing downstreams should turn on in stable distros just
11725 yet, as it might create incompatibilities with a few DNS servers and
11726 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
11727 automatically whenever we detect such incompatible setups, but there
11728 might be systems we do not cover yet. Hence: please help us testing
11729 the DNSSEC code, leave this on where you can, report back, but then
11730 again don't consider turning this on in your stable, LTS or
11731 production release just yet. (Note that you have to enable
11732 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
11733 and its DNSSEC mode for hostname resolution from local
11736 * systemd-resolve conveniently resolves DANE records with the --tlsa
11737 option and OPENPGPKEY records with the --openpgp option. It also
11738 supports dumping raw DNS record data via the new --raw= switch.
11740 * systemd-logind will now by default terminate user processes that are
11741 part of the user session scope unit (session-XX.scope) when the user
11742 logs out. This behavior is controlled by the KillUserProcesses=
11743 setting in logind.conf, and the previous default of "no" is now
11744 changed to "yes". This means that user sessions will be properly
11745 cleaned up after, but additional steps are necessary to allow
11746 intentionally long-running processes to survive logout.
11748 While the user is logged in at least once, user@.service is running,
11749 and any service that should survive the end of any individual login
11750 session can be started at a user service or scope using systemd-run.
11751 systemd-run(1) man page has been extended with an example which shows
11752 how to run screen in a scope unit underneath user@.service. The same
11753 command works for tmux.
11755 After the user logs out of all sessions, user@.service will be
11756 terminated too, by default, unless the user has "lingering" enabled.
11757 To effectively allow users to run long-term tasks even if they are
11758 logged out, lingering must be enabled for them. See loginctl(1) for
11759 details. The default polkit policy was modified to allow users to
11760 set lingering for themselves without authentication.
11762 Previous defaults can be restored at compile time by the
11763 --without-kill-user-processes option to "configure".
11765 * systemd-logind gained new configuration settings SessionsMax= and
11766 InhibitorsMax=, both with a default of 8192. It will not register new
11767 user sessions or inhibitors above this limit.
11769 * systemd-logind will now reload configuration on SIGHUP.
11771 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
11772 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
11773 enable. Also, support for the "io" cgroup controller in the unified
11774 hierarchy has been added, so that the "memory", "pids" and "io" are
11775 now the controllers that are supported on the unified hierarchy.
11777 WARNING: it is not possible to use previous systemd versions with
11778 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
11779 is necessary to also update systemd in the initramfs if using the
11780 unified hierarchy. An updated SELinux policy is also required.
11782 * LLDP support has been extended, and both passive (receive-only) and
11783 active (sender) modes are supported. Passive mode ("routers-only") is
11784 enabled by default in systemd-networkd. Active LLDP mode is enabled
11785 by default for containers on the internal network. The "networkctl
11786 lldp" command may be used to list information gathered. "networkctl
11787 status" will also show basic LLDP information on connected peers now.
11789 * The IAID and DUID unique identifier sent in DHCP requests may now be
11790 configured for the system and each .network file managed by
11791 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
11793 * systemd-networkd gained support for configuring proxy ARP support for
11794 each interface, via the ProxyArp= setting in .network files. It also
11795 gained support for configuring the multicast querier feature of
11796 bridge devices, via the new MulticastQuerier= setting in .netdev
11797 files. Similarly, snooping on the IGMP traffic can be controlled
11798 via the new setting MulticastSnooping=.
11800 A new setting PreferredLifetime= has been added for addresses
11801 configured in .network file to configure the lifetime intended for an
11804 The systemd-networkd DHCP server gained the option EmitRouter=, which
11805 defaults to yes, to configure whether the DHCP Option 3 (Router)
11808 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
11809 systemd-socket-activate and installed into /usr/bin. It is now fully
11812 * systemd-journald now uses separate threads to flush changes to disk
11813 when closing journal files, thus reducing impact of slow disk I/O on
11814 logging performance.
11816 * The sd-journal API gained two new calls
11817 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
11818 can be used to open journal files using file descriptors instead of
11819 file or directory paths. sd_journal_open_container() has been
11820 deprecated, sd_journal_open_directory_fd() should be used instead
11821 with the flag SD_JOURNAL_OS_ROOT.
11823 * journalctl learned a new output mode "-o short-unix" that outputs log
11824 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
11825 UTC). It also gained support for a new --no-hostname setting to
11826 suppress the hostname column in the family of "short" output modes.
11828 * systemd-ask-password now optionally skips printing of the password to
11829 stdout with --no-output which can be useful in scripts.
11831 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
11832 (devices tagged with ID_MAKER_TOOL) are now tagged with
11833 "uaccess" and are available to logged in users.
11835 * The DeviceAllow= unit setting now supports specifiers (with "%").
11837 * "systemctl show" gained a new --value switch, which allows print a
11838 only the contents of a specific unit property, without also printing
11839 the property's name. Similar support was added to "show*" verbs
11840 of loginctl and machinectl that output "key=value" lists.
11842 * A new unit type "generated" was added for files dynamically generated
11843 by generator tools. Similarly, a new unit type "transient" is used
11844 for unit files created using the runtime API. "systemctl enable" will
11845 refuse to operate on such files.
11847 * A new command "systemctl revert" has been added that may be used to
11848 revert to the vendor version of a unit file, in case local changes
11849 have been made by adding drop-ins or overriding the unit file.
11851 * "machinectl clean" gained a new verb to automatically remove all or
11852 just hidden container images.
11854 * systemd-tmpfiles gained support for a new line type "e" for emptying
11855 directories, if they exist, without creating them if they don't.
11857 * systemd-nspawn gained support for automatically patching the UID/GIDs
11858 of the owners and the ACLs of all files and directories in a
11859 container tree to match the UID/GID user namespacing range selected
11860 for the container invocation. This mode is enabled via the new
11861 --private-users-chown switch. It also gained support for
11862 automatically choosing a free, previously unused UID/GID range when
11863 starting a container, via the new --private-users=pick setting (which
11864 implies --private-users-chown). Together, these options for the first
11865 time make user namespacing for nspawn containers fully automatic and
11866 thus deployable. The systemd-nspawn@.service template unit file has
11867 been changed to use this functionality by default.
11869 * systemd-nspawn gained a new --network-zone= switch, that allows
11870 creating ad-hoc virtual Ethernet links between multiple containers,
11871 that only exist as long as at least one container referencing them is
11872 running. This allows easy connecting of multiple containers with a
11873 common link that implements an Ethernet broadcast domain. Each of
11874 these network "zones" may be named relatively freely by the user, and
11875 may be referenced by any number of containers, but each container may
11876 only reference one of these "zones". On the lower level, this is
11877 implemented by an automatically managed bridge network interface for
11878 each zone, that is created when the first container referencing its
11879 zone is created and removed when the last one referencing its zone
11882 * The default start timeout may now be configured on the kernel command
11883 line via systemd.default_timeout_start_sec=. It was already
11884 configurable via the DefaultTimeoutStartSec= option in
11885 /etc/systemd/system.conf.
11887 * Socket units gained a new TriggerLimitIntervalSec= and
11888 TriggerLimitBurst= setting to configure a limit on the activation
11889 rate of the socket unit.
11891 * The LimitNICE= setting now optionally takes normal UNIX nice values
11892 in addition to the raw integer limit value. If the specified
11893 parameter is prefixed with "+" or "-" and is in the range -20…19 the
11894 value is understood as UNIX nice value. If not prefixed like this it
11895 is understood as raw RLIMIT_NICE limit.
11897 * Note that the effect of the PrivateDevices= unit file setting changed
11898 slightly with this release: the per-device /dev file system will be
11899 mounted read-only from this version on, and will have "noexec"
11900 set. This (minor) change of behavior might cause some (exceptional)
11901 legacy software to break, when PrivateDevices=yes is set for its
11902 service. Please leave PrivateDevices= off if you run into problems
11905 * systemd-bootchart has been split out to a separate repository:
11906 https://github.com/systemd/systemd-bootchart
11908 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
11909 merged into the kernel in its current form.
11911 * The compatibility libraries libsystemd-daemon.so,
11912 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
11913 which have been deprecated since systemd-209 have been removed along
11914 with the corresponding pkg-config files. All symbols provided by
11915 those libraries are provided by libsystemd.so.
11917 * The Capabilities= unit file setting has been removed (it is ignored
11918 for backwards compatibility). AmbientCapabilities= and
11919 CapabilityBoundingSet= should be used instead.
11921 * A new special target has been added, initrd-root-device.target,
11922 which creates a synchronization point for dependencies of the root
11923 device in early userspace. Initramfs builders must ensure that this
11924 target is now included in early userspace.
11926 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
11927 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
11928 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
11929 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
11930 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
11931 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
11932 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
11933 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
11934 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
11935 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
11936 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
11937 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
11938 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
11939 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
11940 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
11941 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
11942 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
11943 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
11944 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
11945 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
11946 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
11947 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
11948 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
11949 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
11950 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
11953 — Fairfax, 2016-05-21
11957 * The systemd-resolved DNS resolver service has gained a substantial
11958 set of new features, most prominently it may now act as a DNSSEC
11959 validating stub resolver. DNSSEC mode is currently turned off by
11960 default, but is expected to be turned on by default in one of the
11961 next releases. For now, we invite everybody to test the DNSSEC logic
11962 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
11963 service also gained a full set of D-Bus interfaces, including calls
11964 to configure DNS and DNSSEC settings per link (for use by external
11965 network management software). systemd-resolved and systemd-networkd
11966 now distinguish between "search" and "routing" domains. The former
11967 are used to qualify single-label names, the latter are used purely
11968 for routing lookups within certain domains to specific links.
11969 resolved now also synthesizes RRs for all entries from /etc/hosts.
11971 * The systemd-resolve tool (which is a client utility for
11972 systemd-resolved) has been improved considerably and is now fully
11973 supported and documented. Hence it has moved from /usr/lib/systemd to
11976 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
11979 * The coredump collection logic has been reworked: when a coredump is
11980 collected it is now written to disk, compressed and processed
11981 (including stacktrace extraction) from a new instantiated service
11982 systemd-coredump@.service, instead of directly from the
11983 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
11984 processing large coredumps can take up a substantial amount of
11985 resources and time, and this previously happened entirely outside of
11986 systemd's service supervision. With the new logic the core_pattern
11987 hook only does minimal metadata collection before passing off control
11988 to the new instantiated service, which is configured with a time
11989 limit, a nice level and other settings to minimize negative impact on
11990 the rest of the system. Also note that the new logic will honour the
11991 RLIMIT_CORE setting of the crashed process, which now allows users
11992 and processes to turn off coredumping for their processes by setting
11995 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
11996 and all forked processes by default. Previously, PID 1 would leave
11997 the setting at "0" for all processes, as set by the kernel. Note that
11998 the resource limit traditionally has no effect on the generated
11999 coredumps on the system if the /proc/sys/kernel/core_pattern hook
12000 logic is used. Since the limit is now honoured (see above) its
12001 default has been changed so that the coredumping logic is enabled by
12002 default for all processes, while allowing specific opt-out.
12004 * When the stacktrace is extracted from processes of system users, this
12005 is now done as "systemd-coredump" user, in order to sandbox this
12006 potentially security sensitive parsing operation. (Note that when
12007 processing coredumps of normal users this is done under the user ID
12008 of process that crashed, as before.) Packagers should take notice
12009 that it is now necessary to create the "systemd-coredump" system user
12010 and group at package installation time.
12012 * The systemd-activate socket activation testing tool gained support
12013 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
12014 and --seqpacket switches. It also has been extended to support both
12015 new-style and inetd-style file descriptor passing. Use the new
12016 --inetd switch to request inetd-style file descriptor passing.
12018 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
12019 variable, which takes a boolean value. If set to false, ANSI color
12020 output is disabled in the tools even when run on a terminal that
12023 * The VXLAN support in networkd now supports two new settings
12024 DestinationPort= and PortRange=.
12026 * A new systemd.machine_id= kernel command line switch has been added,
12027 that may be used to set the machine ID in /etc/machine-id if it is
12028 not initialized yet. This command line option has no effect if the
12029 file is already initialized.
12031 * systemd-nspawn gained a new --as-pid2 switch that invokes any
12032 specified command line as PID 2 rather than PID 1 in the
12033 container. In this mode PID 1 is a minimal stub init process that
12034 implements the special POSIX and Linux semantics of PID 1 regarding
12035 signal and child process management. Note that this stub init process
12036 is implemented in nspawn itself and requires no support from the
12037 container image. This new logic is useful to support running
12038 arbitrary commands in the container, as normal processes are
12039 generally not prepared to run as PID 1.
12041 * systemd-nspawn gained a new --chdir= switch for setting the current
12042 working directory for the process started in the container.
12044 * "journalctl /dev/sda" will now output all kernel log messages for
12045 specified device from the current boot, in addition to all devices
12046 that are parents of it. This should make log output about devices
12047 pretty useful, as long as kernel drivers attach enough metadata to
12048 the log messages. (The usual SATA drivers do.)
12050 * The sd-journal API gained two new calls
12051 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
12052 that report whether log data from /run or /var has been found.
12054 * journalctl gained a new switch "--fields" that prints all journal
12055 record field names currently in use in the journal. This is backed
12056 by two new sd-journal API calls sd_journal_enumerate_fields() and
12057 sd_journal_restart_fields().
12059 * Most configurable timeouts in systemd now expect an argument of
12060 "infinity" to turn them off, instead of "0" as before. The semantics
12061 from now on is that a timeout of "0" means "now", and "infinity"
12062 means "never". To maintain backwards compatibility, "0" continues to
12063 turn off previously existing timeout settings.
12065 * "systemctl reload-or-try-restart" has been renamed to "systemctl
12066 try-reload-or-restart" to clarify what it actually does: the "try"
12067 logic applies to both reloading and restarting, not just restarting.
12068 The old name continues to be accepted for compatibility.
12070 * On boot-up, when PID 1 detects that the system clock is behind the
12071 release date of the systemd version in use, the clock is now set
12072 to the latter. Previously, this was already done in timesyncd, in order
12073 to avoid running with clocks set to the various clock epochs such as
12074 1902, 1938 or 1970. With this change the logic is now done in PID 1
12075 in addition to timesyncd during early boot-up, so that it is enforced
12076 before the first process is spawned by systemd. Note that the logic
12077 in timesyncd remains, as it is more comprehensive and ensures
12078 clock monotonicity by maintaining a persistent timestamp file in
12079 /var. Since /var is generally not available in earliest boot or the
12080 initrd, this part of the logic remains in timesyncd, and is not done
12083 * Support for tweaking details in net_cls.class_id through the
12084 NetClass= configuration directive has been removed, as the kernel
12085 people have decided to deprecate that controller in cgroup v2.
12086 Userspace tools such as nftables are moving over to setting rules
12087 that are specific to the full cgroup path of a task, which obsoletes
12088 these controllers anyway. The NetClass= directive is kept around for
12089 legacy compatibility reasons. For a more in-depth description of the
12090 kernel change, please refer to the respective upstream commit:
12092 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
12094 * A new service setting RuntimeMaxSec= has been added that may be used
12095 to specify a maximum runtime for a service. If the timeout is hit, the
12096 service is terminated and put into a failure state.
12098 * A new service setting AmbientCapabilities= has been added. It allows
12099 configuration of additional Linux process capabilities that are
12100 passed to the activated processes. This is only available on very
12103 * The process resource limit settings in service units may now be used
12104 to configure hard and soft limits individually.
12106 * The various libsystemd APIs such as sd-bus or sd-event now publicly
12107 expose support for gcc's __attribute__((cleanup())) C extension.
12108 Specifically, for many object destructor functions alternative
12109 versions have been added that have names suffixed with "p" and take a
12110 pointer to a pointer to the object to destroy, instead of just a
12111 pointer to the object itself. This is useful because these destructor
12112 functions may be used directly as parameters to the cleanup
12113 construct. Internally, systemd has been a heavy user of this GCC
12114 extension for a long time, and with this change similar support is
12115 now available to consumers of the library outside of systemd. Note
12116 that by using this extension in your sources compatibility with old
12117 and strictly ANSI compatible C compilers is lost. However, all gcc or
12118 LLVM versions of recent years support this extension.
12120 * Timer units gained support for a new setting RandomizedDelaySec= that
12121 allows configuring some additional randomized delay to the configured
12122 time. This is useful to spread out timer events to avoid load peaks in
12123 clusters or larger setups.
12125 * Calendar time specifications now support sub-second accuracy.
12127 * Socket units now support listening on SCTP and UDP-lite protocol
12130 * The sd-event API now comes with a full set of man pages.
12132 * Older versions of systemd contained experimental support for
12133 compressing journal files and coredumps with the LZ4 compressor that
12134 was not compatible with the lz4 binary (due to API limitations of the
12135 lz4 library). This support has been removed; only support for files
12136 compatible with the lz4 binary remains. This LZ4 logic is now
12137 officially supported and no longer considered experimental.
12139 * The dkr image import logic has been removed again from importd. dkr's
12140 micro-services focus doesn't fit into the machine image focus of
12141 importd, and quickly got out of date with the upstream dkr API.
12143 * Creation of the /run/lock/lockdev/ directory was dropped from
12144 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
12145 been available for many years. If you still need this, you need to
12146 create your own tmpfiles.d config file with:
12148 d /run/lock/lockdev 0775 root lock -
12150 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
12151 and RebootArgument= have been moved from the [Service] section of
12152 unit files to [Unit], and they are now supported on all unit types,
12153 not just service units. Of course, systemd will continue to
12154 understand these settings also at the old location, in order to
12155 maintain compatibility.
12157 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
12158 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
12159 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
12160 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
12161 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
12162 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
12163 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
12164 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
12165 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
12166 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
12167 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
12168 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
12169 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
12170 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
12171 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
12172 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
12173 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
12174 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
12175 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
12177 — Berlin, 2016-02-11
12181 * A number of properties previously only settable in unit
12182 files are now also available as properties to set when
12183 creating transient units programmatically via the bus, as it
12184 is exposed with systemd-run's --property=
12185 setting. Specifically, these are: SyslogIdentifier=,
12186 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
12187 EnvironmentFile=, ReadWriteDirectories=,
12188 ReadOnlyDirectories=, InaccessibleDirectories=,
12189 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
12191 * When creating transient services via the bus API it is now
12192 possible to pass in a set of file descriptors to use as
12193 STDIN/STDOUT/STDERR for the invoked process.
12195 * Slice units may now be created transiently via the bus APIs,
12196 similar to the way service and scope units may already be
12197 created transiently.
12199 * Wherever systemd expects a calendar timestamp specification
12200 (like in journalctl's --since= and --until= switches) UTC
12201 timestamps are now supported. Timestamps suffixed with "UTC"
12202 are now considered to be in Universal Time Coordinated
12203 instead of the local timezone. Also, timestamps may now
12204 optionally be specified with sub-second accuracy. Both of
12205 these additions also apply to recurring calendar event
12206 specification, such as OnCalendar= in timer units.
12208 * journalctl gained a new "--sync" switch that asks the
12209 journal daemon to write all so far unwritten log messages to
12210 disk and sync the files, before returning.
12212 * systemd-tmpfiles learned two new line types "q" and "Q" that
12213 operate like "v", but also set up a basic btrfs quota
12214 hierarchy when used on a btrfs file system with quota
12217 * tmpfiles' "v", "q" and "Q" will now create a plain directory
12218 instead of a subvolume (even on a btrfs file system) if the
12219 root directory is a plain directory, and not a
12220 subvolume. This should simplify things with certain chroot()
12221 environments which are not aware of the concept of btrfs
12224 * systemd-detect-virt gained a new --chroot switch to detect
12225 whether execution takes place in a chroot() environment.
12227 * CPUAffinity= now takes CPU index ranges in addition to
12228 individual indexes.
12230 * The various memory-related resource limit settings (such as
12231 LimitAS=) now understand the usual K, M, G, … suffixes to
12232 the base of 1024 (IEC). Similar, the time-related resource
12233 limit settings understand the usual min, h, day, … suffixes
12236 * There's a new system.conf setting DefaultTasksMax= to
12237 control the default TasksMax= setting for services and
12238 scopes running on the system. (TasksMax= is the primary
12239 setting that exposes the "pids" cgroup controller on systemd
12240 and was introduced in the previous systemd release.) The
12241 setting now defaults to 512, which means services that are
12242 not explicitly configured otherwise will only be able to
12243 create 512 processes or threads at maximum, from this
12244 version on. Note that this means that thread- or
12245 process-heavy services might need to be reconfigured to set
12246 TasksMax= to a higher value. It is sufficient to set
12247 TasksMax= in these specific unit files to a higher value, or
12248 even "infinity". Similar, there's now a logind.conf setting
12249 UserTasksMax= that defaults to 4096 and limits the total
12250 number of processes or tasks each user may own
12251 concurrently. nspawn containers also have the TasksMax=
12252 value set by default now, to 8192. Note that all of this
12253 only has an effect if the "pids" cgroup controller is
12254 enabled in the kernel. The general benefit of these changes
12255 should be a more robust and safer system, that provides a
12256 certain amount of per-service fork() bomb protection.
12258 * systemd-nspawn gained the new --network-veth-extra= switch
12259 to define additional and arbitrarily-named virtual Ethernet
12260 links between the host and the container.
12262 * A new service execution setting PassEnvironment= has been
12263 added that allows importing select environment variables
12264 from PID1's environment block into the environment block of
12267 * Timer units gained support for a new RemainAfterElapse=
12268 setting which takes a boolean argument. It defaults to on,
12269 exposing behaviour unchanged to previous releases. If set to
12270 off, timer units are unloaded after they elapsed if they
12271 cannot elapse again. This is particularly useful for
12272 transient timer units, which shall not stay around longer
12273 than until they first elapse.
12275 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
12276 default now (the kernel default is 16). This is beneficial
12277 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
12278 allows substantially larger numbers of queued
12279 datagrams. This should increase the capability of systemd to
12280 parallelize boot-up, as logging and sd_notify() are unlikely
12281 to stall execution anymore. If you need to change the value
12282 from the new defaults, use the usual sysctl.d/ snippets.
12284 * The compression framing format used by the journal or
12285 coredump processing has changed to be in line with what the
12286 official LZ4 tools generate. LZ4 compression support in
12287 systemd was considered unsupported previously, as the format
12288 was not compatible with the normal tools. With this release
12289 this has changed now, and it is hence safe for downstream
12290 distributions to turn it on. While not compressing as well
12291 as the XZ, LZ4 is substantially faster, which makes
12292 it a good default choice for the compression logic in the
12293 journal and in coredump handling.
12295 * Any reference to /etc/mtab has been dropped from
12296 systemd. The file has been obsolete since a while, but
12297 systemd refused to work on systems where it was incorrectly
12298 set up (it should be a symlink or non-existent). Please make
12299 sure to update to util-linux 2.27.1 or newer in conjunction
12300 with this systemd release, which also drops any reference to
12301 /etc/mtab. If you maintain a distribution make sure that no
12302 software you package still references it, as this is a
12303 likely source of bugs. There's also a glibc bug pending,
12304 asking for removal of any reference to this obsolete file:
12306 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
12308 Note that only util-linux versions built with
12309 --enable-libmount-force-mountinfo are supported.
12311 * Support for the ".snapshot" unit type has been removed. This
12312 feature turned out to be little useful and little used, and
12313 has now been removed from the core and from systemctl.
12315 * The dependency types RequiresOverridable= and
12316 RequisiteOverridable= have been removed from systemd. They
12317 have been used only very sparingly to our knowledge and
12318 other options that provide a similar effect (such as
12319 systemctl --mode=ignore-dependencies) are much more useful
12320 and commonly used. Moreover, they were only half-way
12321 implemented as the option to control behaviour regarding
12322 these dependencies was never added to systemctl. By removing
12323 these dependency types the execution engine becomes a bit
12324 simpler. Unit files that use these dependencies should be
12325 changed to use the non-Overridable dependency types
12326 instead. In fact, when parsing unit files with these
12327 options, that's what systemd will automatically convert them
12328 too, but it will also warn, asking users to fix the unit
12329 files accordingly. Removal of these dependency types should
12330 only affect a negligible number of unit files in the wild.
12332 * Behaviour of networkd's IPForward= option changed
12333 (again). It will no longer maintain a per-interface setting,
12334 but propagate one way from interfaces where this is enabled
12335 to the global kernel setting. The global setting will be
12336 enabled when requested by a network that is set up, but
12337 never be disabled again. This change was made to make sure
12338 IPv4 and IPv6 behaviour regarding packet forwarding is
12339 similar (as the Linux IPv6 stack does not support
12340 per-interface control of this setting) and to minimize
12343 * In unit files the behaviour of %u, %U, %h, %s has
12344 changed. These specifiers will now unconditionally resolve
12345 to the various user database fields of the user that the
12346 systemd instance is running as, instead of the user
12347 configured in the specific unit via User=. Note that this
12348 effectively doesn't change much, as resolving of these
12349 specifiers was already turned off in the --system instance
12350 of systemd, as we cannot do NSS lookups from PID 1. In the
12351 --user instance of systemd these specifiers where correctly
12352 resolved, but hardly made any sense, since the user instance
12353 lacks privileges to do user switches anyway, and User= is
12354 hence useless. Moreover, even in the --user instance of
12355 systemd behaviour was awkward as it would only take settings
12356 from User= assignment placed before the specifier into
12357 account. In order to unify and simplify the logic around
12358 this the specifiers will now always resolve to the
12359 credentials of the user invoking the manager (which in case
12360 of PID 1 is the root user).
12362 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
12363 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
12364 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
12365 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
12366 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
12367 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
12368 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
12369 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
12370 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
12371 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
12372 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
12373 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
12374 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
12375 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
12378 — Berlin, 2015-11-18
12382 * systemd now depends on util-linux v2.27. More specifically,
12383 the newly added mount monitor feature in libmount now
12384 replaces systemd's former own implementation.
12386 * libmount mandates /etc/mtab not to be regular file, and
12387 systemd now enforces this condition at early boot.
12388 /etc/mtab has been deprecated and warned about for a very
12389 long time, so systems running systemd should already have
12390 stopped having this file around as anything else than a
12391 symlink to /proc/self/mounts.
12393 * Support for the "pids" cgroup controller has been added. It
12394 allows accounting the number of tasks in a cgroup and
12395 enforcing limits on it. This adds two new setting
12396 TasksAccounting= and TasksMax= to each unit, as well as a
12397 global option DefaultTasksAccounting=.
12399 * Support for the "net_cls" cgroup controller has been added.
12400 It allows assigning a net class ID to each task in the
12401 cgroup, which can then be used in firewall rules and traffic
12402 shaping configurations. Note that the kernel netfilter net
12403 class code does not currently work reliably for ingress
12404 packets on unestablished sockets.
12406 This adds a new config directive called NetClass= to CGroup
12407 enabled units. Allowed values are positive numbers for fixed
12408 assignments and "auto" for picking a free value
12411 * 'systemctl is-system-running' now returns 'offline' if the
12412 system is not booted with systemd. This command can now be
12413 used as a substitute for 'systemd-notify --booted'.
12415 * Watchdog timeouts have been increased to 3 minutes for all
12416 in-tree service files. Apparently, disk IO issues are more
12417 frequent than we hoped, and user reported >1 minute waiting
12420 * 'machine-id-commit' functionality has been merged into
12421 'machine-id-setup --commit'. The separate binary has been
12424 * The WorkingDirectory= directive in unit files may now be set
12425 to the special value '~'. In this case, the working
12426 directory is set to the home directory of the user
12427 configured in User=.
12429 * "machinectl shell" will now open the shell in the home
12430 directory of the selected user by default.
12432 * The CrashChVT= configuration file setting is renamed to
12433 CrashChangeVT=, following our usual logic of not
12434 abbreviating unnecessarily. The old directive is still
12435 supported for compat reasons. Also, this directive now takes
12436 an integer value between 1 and 63, or a boolean value. The
12437 formerly supported '-1' value for disabling stays around for
12440 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
12441 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
12442 RootDirectory= properties can now be set for transient
12445 * The systemd-analyze tool gained a new "set-log-target" verb
12446 to change the logging target the system manager logs to
12447 dynamically during runtime. This is similar to how
12448 "systemd-analyze set-log-level" already changes the log
12451 * In nspawn /sys is now mounted as tmpfs, with only a selected
12452 set of subdirectories mounted in from the real sysfs. This
12453 enhances security slightly, and is useful for ensuring user
12454 namespaces work correctly.
12456 * Support for USB FunctionFS activation has been added. This
12457 allows implementation of USB gadget services that are
12458 activated as soon as they are requested, so that they don't
12459 have to run continuously, similar to classic socket
12462 * The "systemctl exit" command now optionally takes an
12463 additional parameter that sets the exit code to return from
12464 the systemd manager when exiting. This is only relevant when
12465 running the systemd user instance, or when running the
12466 system instance in a container.
12468 * sd-bus gained the new API calls sd_bus_path_encode_many()
12469 and sd_bus_path_decode_many() that allow easy encoding and
12470 decoding of multiple identifier strings inside a D-Bus
12471 object path. Another new call sd_bus_default_flush_close()
12472 has been added to flush and close per-thread default
12475 * systemd-cgtop gained support for a -M/--machine= switch to
12476 show the control groups within a certain container only.
12478 * "systemctl kill" gained support for an optional --fail
12479 switch. If specified the requested operation will fail of no
12480 processes have been killed, because the unit had no
12481 processes attached, or similar.
12483 * A new systemd.crash_reboot=1 kernel command line option has
12484 been added that triggers a reboot after crashing. This can
12485 also be set through CrashReboot= in systemd.conf.
12487 * The RuntimeDirectory= setting now understands unit
12488 specifiers like %i or %f.
12490 * A new (still internal) library API sd-ipv4acd has been added,
12491 that implements address conflict detection for IPv4. It's
12492 based on code from sd-ipv4ll, and will be useful for
12493 detecting DHCP address conflicts.
12495 * File descriptors passed during socket activation may now be
12496 named. A new API sd_listen_fds_with_names() is added to
12497 access the names. The default names may be overridden,
12498 either in the .socket file using the FileDescriptorName=
12499 parameter, or by passing FDNAME= when storing the file
12500 descriptors using sd_notify().
12502 * systemd-networkd gained support for:
12504 - Setting the IPv6 Router Advertisement settings via
12505 IPv6AcceptRouterAdvertisements= in .network files.
12507 - Configuring the HelloTimeSec=, MaxAgeSec= and
12508 ForwardDelaySec= bridge parameters in .netdev files.
12510 - Configuring PreferredSource= for static routes in
12513 * The "ask-password" framework used to query for LUKS harddisk
12514 passwords or SSL passwords during boot gained support for
12515 caching passwords in the kernel keyring, if it is
12516 available. This makes sure that the user only has to type in
12517 a passphrase once if there are multiple objects to unlock
12518 with the same one. Previously, such password caching was
12519 available only when Plymouth was used; this moves the
12520 caching logic into the systemd codebase itself. The
12521 "systemd-ask-password" utility gained a new --keyname=
12522 switch to control which kernel keyring key to use for
12523 caching a password in. This functionality is also useful for
12524 enabling display managers such as gdm to automatically
12525 unlock the user's GNOME keyring if its passphrase, the
12526 user's password and the harddisk password are the same, if
12527 gdm-autologin is used.
12529 * When downloading tar or raw images using "machinectl
12530 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
12531 file is now also downloaded, if it is available and stored
12532 next to the image file.
12534 * Units of type ".socket" gained a new boolean setting
12535 Writable= which is only useful in conjunction with
12536 ListenSpecial=. If true, enables opening the specified
12537 special file in O_RDWR mode rather than O_RDONLY mode.
12539 * systemd-rfkill has been reworked to become a singleton
12540 service that is activated through /dev/rfkill on each rfkill
12541 state change and saves the settings to disk. This way,
12542 systemd-rfkill is now compatible with devices that exist
12543 only intermittendly, and even restores state if the previous
12544 system shutdown was abrupt rather than clean.
12546 * The journal daemon gained support for vacuuming old journal
12547 files controlled by the number of files that shall remain,
12548 in addition to the already existing control by size and by
12549 date. This is useful as journal interleaving performance
12550 degrades with too many separate journal files, and allows
12551 putting an effective limit on them. The new setting defaults
12552 to 100, but this may be changed by setting SystemMaxFiles=
12553 and RuntimeMaxFiles= in journald.conf. Also, the
12554 "journalctl" tool gained the new --vacuum-files= switch to
12555 manually vacuum journal files to leave only the specified
12556 number of files in place.
12558 * udev will now create /dev/disk/by-path links for ATA devices
12559 on kernels where that is supported.
12561 * Galician, Serbian, Turkish and Korean translations were added.
12563 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
12564 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
12565 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
12566 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
12567 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
12568 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
12569 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
12570 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
12571 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
12572 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
12573 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
12574 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
12575 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
12576 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
12577 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
12578 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
12579 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
12580 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
12582 — Berlin, 2015-10-07
12586 * The DHCP implementation of systemd-networkd gained a set of
12589 - The DHCP server now supports emitting DNS and NTP
12590 information. It may be enabled and configured via
12591 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
12592 and NTP information is enabled, but no servers are
12593 configured, the corresponding uplink information (if there
12594 is any) is propagated.
12596 - Server and client now support transmission and reception
12597 of timezone information. It can be configured via the
12598 newly introduced network options UseTimezone=,
12599 EmitTimezone=, and Timezone=. Transmission of timezone
12600 information is enabled between host and containers by
12601 default now: the container will change its local timezone
12602 to what the host has set.
12604 - Lease timeouts can now be configured via
12605 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
12607 - The DHCP server improved on the stability of
12608 leases. Clients are more likely to get the same lease
12609 information back, even if the server loses state.
12611 - The DHCP server supports two new configuration options to
12612 control the lease address pool metrics, PoolOffset= and
12615 * The encapsulation limit of tunnels in systemd-networkd may
12616 now be configured via 'EncapsulationLimit='. It allows
12617 modifying the maximum additional levels of encapsulation
12618 that are permitted to be prepended to a packet.
12620 * systemd now supports the concept of user buses replacing
12621 session buses, if used with dbus-1.10 (and enabled via dbus
12622 --enable-user-session). It previously only supported this on
12623 kdbus-enabled systems, and this release expands this to
12624 'dbus-daemon' systems.
12626 * systemd-networkd now supports predictable interface names
12627 for virtio devices.
12629 * systemd now optionally supports the new Linux kernel
12630 "unified" control group hierarchy. If enabled via the kernel
12631 command-line option 'systemd.unified_cgroup_hierarchy=1',
12632 systemd will try to mount the unified cgroup hierarchy
12633 directly on /sys/fs/cgroup. If not enabled, or not
12634 available, systemd will fall back to the legacy cgroup
12635 hierarchy setup, as before. Host system and containers can
12636 mix and match legacy and unified hierarchies as they
12637 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
12638 environment variable to individually select the hierarchy to
12639 use for executed containers. By default, nspawn will use the
12640 unified hierarchy for the containers if the host uses the
12641 unified hierarchy, and the legacy hierarchy otherwise.
12642 Please note that at this point the unified hierarchy is an
12643 experimental kernel feature and is likely to change in one
12644 of the next kernel releases. Therefore, it should not be
12645 enabled by default in downstream distributions yet. The
12646 minimum required kernel version for the unified hierarchy to
12647 work is 4.2. Note that when the unified hierarchy is used
12648 for the first time delegated access to controllers is
12649 safe. Because of this systemd-nspawn containers will get
12650 access to controllers now, as will systemd user
12651 sessions. This means containers and user sessions may now
12652 manage their own resources, partitioning up what the system
12655 * A new special scope unit "init.scope" has been introduced
12656 that encapsulates PID 1 of the system. It may be used to
12657 determine resource usage and enforce resource limits on PID
12658 1 itself. PID 1 hence moved out of the root of the control
12661 * The cgtop tool gained support for filtering out kernel
12662 threads when counting tasks in a control group. Also, the
12663 count of processes is now recursively summed up by
12664 default. Two options -k and --recursive= have been added to
12665 revert to old behaviour. The tool has also been updated to
12666 work correctly in containers now.
12668 * systemd-nspawn's --bind= and --bind-ro= options have been
12669 extended to allow creation of non-recursive bind mounts.
12671 * libsystemd gained two new calls sd_pid_get_cgroup() and
12672 sd_peer_get_cgroup() which return the control group path of
12673 a process or peer of a connected AF_UNIX socket. This
12674 function call is particularly useful when implementing
12675 delegated subtrees support in the control group hierarchy.
12677 * The "sd-event" event loop API of libsystemd now supports
12678 correct dequeuing of real-time signals, without losing
12681 * When systemd requests a polkit decision when managing units it
12682 will now add additional fields to the request, including unit
12683 name and desired operation. This enables more powerful polkit
12684 policies, that make decisions depending on these parameters.
12686 * nspawn learnt support for .nspawn settings files, that may
12687 accompany the image files or directories of containers, and
12688 may contain additional settings for the container. This is
12689 an alternative to configuring container parameters via the
12690 nspawn command line.
12692 Contributions from: Cristian Rodríguez, Daniel Mack, David
12693 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
12694 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
12695 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
12696 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
12697 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
12698 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
12699 Andersen, Tom Gundersen, Torstein Husebø
12701 — Berlin, 2015-09-08
12705 * machinectl gained a new verb 'shell' which opens a fresh
12706 shell on the target container or the host. It is similar to
12707 the existing 'login' command of machinectl, but spawns the
12708 shell directly without prompting for username or
12709 password. The pseudo machine '.host' now refers to the local
12710 host and is used by default. Hence, 'machinectl shell' can
12711 be used as replacement for 'su -' which spawns a session as
12712 a fresh systemd unit in a way that is fully isolated from
12713 the originating session.
12715 * systemd-networkd learned to cope with private-zone DHCP
12716 options and allows other programs to query the values.
12718 * SELinux access control when enabling/disabling units is no
12719 longer enforced with this release. The previous implementation
12720 was incorrect, and a new corrected implementation is not yet
12721 available. As unit file operations are still protected via
12722 polkit and D-Bus policy this is not a security problem. Yet,
12723 distributions which care about optimal SELinux support should
12724 probably not stabilize on this release.
12726 * sd-bus gained support for matches of type "arg0has=", that
12727 test for membership of strings in string arrays sent in bus
12730 * systemd-resolved now dumps the contents of its DNS and LLMNR
12731 caches to the logs on reception of the SIGUSR1 signal. This
12732 is useful to debug DNS behaviour.
12734 * The coredumpctl tool gained a new --directory= option to
12735 operate on journal files in a specific directory.
12737 * "systemctl reboot" and related commands gained a new
12738 "--message=" option which may be used to set a free-text
12739 wall message when shutting down or rebooting the
12740 system. This message is also logged, which is useful for
12741 figuring out the reason for a reboot or shutdown a
12744 * The "systemd-resolve-host" tool's -i switch now takes
12745 network interface numbers as alternative to interface names.
12747 * A new unit file setting for services has been introduced:
12748 UtmpMode= allows configuration of how precisely systemd
12749 handles utmp and wtmp entries for the service if this is
12750 enabled. This allows writing services that appear similar to
12751 user sessions in the output of the "w", "who", "last" and
12754 * systemd-resolved will now locally synthesize DNS resource
12755 records for the "localhost" and "gateway" domains as well as
12756 the local hostname. This should ensure that clients querying
12757 RRs via resolved will get similar results as those going via
12758 NSS, if nss-myhostname is enabled.
12760 Contributions from: Alastair Hughes, Alex Crawford, Daniel
12761 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
12762 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
12763 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
12764 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
12765 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
12766 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
12767 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
12768 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
12769 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
12770 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
12771 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
12773 — Berlin, 2015-08-27
12777 * The systemd-efi-boot-generator functionality was merged into
12778 systemd-gpt-auto-generator.
12780 * systemd-networkd now supports Group Policy for vxlan
12781 devices. It can be enabled via the new boolean configuration
12782 option called 'GroupPolicyExtension='.
12784 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
12785 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
12786 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
12788 — Berlin, 2015-07-31
12792 * The python-systemd code has been removed from the systemd repository.
12793 A new repository has been created which accommodates the code from
12794 now on, and we kindly ask distributions to create a separate package
12795 for this: https://github.com/systemd/python-systemd
12797 * The systemd daemon will now reload its main configuration
12798 (/etc/systemd/system.conf) on daemon-reload.
12800 * sd-dhcp now exposes vendor specific extensions via
12801 sd_dhcp_lease_get_vendor_specific().
12803 * systemd-networkd gained a number of new configuration options.
12805 - A new boolean configuration option for TAP devices called
12806 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
12807 device, thus allowing to send and receive GSO packets.
12809 - A new tunnel configuration option called 'CopyDSCP='.
12810 If enabled, the DSCP field of ip6 tunnels is copied into the
12811 decapsulated packet.
12813 - A set of boolean bridge configuration options were added.
12814 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
12815 and 'UnicastFlood=' are now parsed by networkd and applied to the
12816 respective bridge link device via the respective IFLA_BRPORT_*
12819 - A new string configuration option to override the hostname sent
12820 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
12821 is true, networkd will use the configured hostname instead of the
12822 system hostname when sending DHCP requests.
12824 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
12825 networkd will configure the IPv6 flow-label of the tunnel device
12826 according to RFC2460.
12828 - The 'macvtap' virtual network devices are now supported, similar to
12829 the already supported 'macvlan' devices.
12831 * systemd-resolved now implements RFC5452 to improve resilience against
12832 cache poisoning. Additionally, source port randomization is enabled
12833 by default to further protect against DNS spoofing attacks.
12835 * nss-mymachines now supports translating UIDs and GIDs of running
12836 containers with user-namespaces enabled. If a container 'foo'
12837 translates a host uid 'UID' to the container uid 'TUID', then
12838 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
12839 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
12840 mapped as 'vg-foo-TGID'.
12842 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
12843 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
12844 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
12845 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
12846 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
12847 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
12848 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
12849 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
12850 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
12851 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
12853 — Berlin, 2015-07-29
12857 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
12858 There are no known issues with current sysfs, and udev does not need
12859 or should be used to work around such bugs.
12861 * udev does no longer enable USB HID power management. Several reports
12862 indicate, that some devices cannot handle that setting.
12864 * The udev accelerometer helper was removed. The functionality
12865 is now fully included in iio-sensor-proxy. But this means,
12866 older iio-sensor-proxy versions will no longer provide
12867 accelerometer/orientation data with this systemd version.
12868 Please upgrade iio-sensor-proxy to version 1.0.
12870 * networkd gained a new configuration option IPv6PrivacyExtensions=
12871 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
12872 for Stateless Address") on selected networks.
12874 * For the sake of fewer build-time dependencies and less code in the
12875 main repository, the python bindings are about to be removed in the
12876 next release. A new repository has been created which accommodates
12877 the code from now on, and we kindly ask distributions to create a
12878 separate package for this. The removal will take place in v223.
12880 https://github.com/systemd/python-systemd
12882 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
12883 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
12884 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
12885 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
12886 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
12887 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
12888 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
12889 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
12890 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
12891 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
12893 — Berlin, 2015-07-07
12897 * The sd-bus.h and sd-event.h APIs have now been declared
12898 stable and have been added to the official interface of
12899 libsystemd.so. sd-bus implements an alternative D-Bus client
12900 library, that is relatively easy to use, very efficient and
12901 supports both classic D-Bus as well as kdbus as transport
12902 backend. sd-event is a generic event loop abstraction that
12903 is built around Linux epoll, but adds features such as event
12904 prioritization or efficient timer handling. Both APIs are good
12905 choices for C programs looking for a bus and/or event loop
12906 implementation that is minimal and does not have to be
12907 portable to other kernels.
12909 * kdbus support is no longer compile-time optional. It is now
12910 always built-in. However, it can still be disabled at
12911 runtime using the kdbus=0 kernel command line setting, and
12912 that setting may be changed to default to off, by specifying
12913 --disable-kdbus at build-time. Note though that the kernel
12914 command line setting has no effect if the kdbus.ko kernel
12915 module is not installed, in which case kdbus is (obviously)
12916 also disabled. We encourage all downstream distributions to
12917 begin testing kdbus by adding it to the kernel images in the
12918 development distributions, and leaving kdbus support in
12921 * The minimal required util-linux version has been bumped to
12924 * Support for chkconfig (--enable-chkconfig) was removed in
12925 favor of calling an abstraction tool
12926 /lib/systemd/systemd-sysv-install. This needs to be
12927 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
12928 in README for details.
12930 * If there's a systemd unit and a SysV init script for the
12931 same service name, and the user executes "systemctl enable"
12932 for it (or a related call), then this will now enable both
12933 (or execute the related operation on both), not just the
12936 * The libudev API documentation has been converted from gtkdoc
12939 * gudev has been removed from the systemd tree, it is now an
12942 * The systemd-cgtop tool learnt a new --raw switch to generate
12943 "raw" (machine parsable) output.
12945 * networkd's IPForwarding= .network file setting learnt the
12946 new setting "kernel", which ensures that networkd does not
12947 change the IP forwarding sysctl from the default kernel
12950 * The systemd-logind bus API now exposes a new boolean
12951 property "Docked" that reports whether logind considers the
12952 system "docked", i.e. connected to a docking station or not.
12954 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
12955 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
12956 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
12957 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
12958 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
12959 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
12960 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
12961 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
12962 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
12963 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
12964 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
12965 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
12966 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
12967 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
12968 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
12969 Fink, Zbigniew Jędrzejewski-Szmek
12971 — Berlin, 2015-06-19
12975 * The gudev library has been extracted into a separate repository
12976 available at: https://git.gnome.org/browse/libgudev/
12977 It is now managed as part of the Gnome project. Distributions
12978 are recommended to pass --disable-gudev to systemd and use
12979 gudev from the Gnome project instead. gudev is still included
12980 in systemd, for now. It will be removed soon, though. Please
12981 also see the announcement-thread on systemd-devel:
12982 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
12984 * systemd now exposes a CPUUsageNSec= property for each
12985 service unit on the bus, that contains the overall consumed
12986 CPU time of a service (the sum of what each process of the
12987 service consumed). This value is only available if
12988 CPUAccounting= is turned on for a service, and is then shown
12989 in the "systemctl status" output.
12991 * Support for configuring alternative mappings of the old SysV
12992 runlevels to systemd targets has been removed. They are now
12993 hardcoded in a way that runlevels 2, 3, 4 all map to
12994 multi-user.target and 5 to graphical.target (which
12995 previously was already the default behaviour).
12997 * The auto-mounter logic gained support for mount point
12998 expiry, using a new TimeoutIdleSec= setting in .automount
12999 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
13001 * The EFI System Partition (ESP) as mounted to /boot by
13002 systemd-efi-boot-generator will now be unmounted
13003 automatically after 2 minutes of not being used. This should
13004 minimize the risk of ESP corruptions.
13006 * New /etc/fstab options x-systemd.requires= and
13007 x-systemd.requires-mounts-for= are now supported to express
13008 additional dependencies for mounts. This is useful for
13009 journaling file systems that support external journal
13010 devices or overlay file systems that require underlying file
13011 systems to be mounted.
13013 * systemd does not support direct live-upgrades (via systemctl
13014 daemon-reexec) from versions older than v44 anymore. As no
13015 distribution we are aware of shipped such old versions in a
13016 stable release this should not be problematic.
13018 * When systemd forks off a new per-connection service instance
13019 it will now set the $REMOTE_ADDR environment variable to the
13020 remote IP address, and $REMOTE_PORT environment variable to
13021 the remote IP port. This behaviour is similar to the
13022 corresponding environment variables defined by CGI.
13024 * systemd-networkd gained support for uplink failure
13025 detection. The BindCarrier= option allows binding interface
13026 configuration dynamically to the link sense of other
13027 interfaces. This is useful to achieve behaviour like in
13030 * systemd-networkd gained support for configuring the DHCP
13031 client identifier to use when requesting leases.
13033 * systemd-networkd now has a per-network UseNTP= option to
13034 configure whether NTP server information acquired via DHCP
13035 is passed on to services like systemd-timesyncd.
13037 * systemd-networkd gained support for vti6 tunnels.
13039 * Note that systemd-networkd manages the sysctl variable
13040 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
13041 it is configured for since v219. The variable controls IP
13042 forwarding, and is a per-interface alternative to the global
13043 /proc/sys/net/ipv[46]/ip_forward. This setting is
13044 configurable in the IPForward= option, which defaults to
13045 "no". This means if networkd is used for an interface it is
13046 no longer sufficient to set the global sysctl option to turn
13047 on IP forwarding! Instead, the .network file option
13048 IPForward= needs to be turned on! Note that the
13049 implementation of this behaviour was broken in v219 and has
13050 been fixed in v220.
13052 * Many bonding and vxlan options are now configurable in
13055 * systemd-nspawn gained a new --property= setting to set unit
13056 properties for the container scope. This is useful for
13057 setting resource parameters (e.g. "CPUShares=500") on
13058 containers started from the command line.
13060 * systemd-nspawn gained a new --private-users= switch to make
13061 use of user namespacing available on recent Linux kernels.
13063 * systemd-nspawn may now be called as part of a shell pipeline
13064 in which case the pipes used for stdin and stdout are passed
13065 directly to the process invoked in the container, without
13066 indirection via a pseudo tty.
13068 * systemd-nspawn gained a new switch to control the UNIX
13069 signal to use when killing the init process of the container
13070 when shutting down.
13072 * systemd-nspawn gained a new --overlay= switch for mounting
13073 overlay file systems into the container using the new kernel
13076 * When a container image is imported via systemd-importd and
13077 the host file system is not btrfs, a loopback block device
13078 file is created in /var/lib/machines.raw with a btrfs file
13079 system inside. It is then mounted to /var/lib/machines to
13080 enable btrfs features for container management. The loopback
13081 file and btrfs file system is grown as needed when container
13082 images are imported via systemd-importd.
13084 * systemd-machined/systemd-importd gained support for btrfs
13085 quota, to enforce container disk space limits on disk. This
13086 is exposed in "machinectl set-limit".
13088 * systemd-importd now can import containers from local .tar,
13089 .raw and .qcow2 images, and export them to .tar and .raw. It
13090 can also import dkr v2 images now from the network (on top
13093 * systemd-importd gained support for verifying downloaded
13094 images with gpg2 (previously only gpg1 was supported).
13096 * systemd-machined, systemd-logind, systemd: most bus calls are
13097 now accessible to unprivileged processes via polkit. Also,
13098 systemd-logind will now allow users to kill their own sessions
13099 without further privileges or authorization.
13101 * systemd-shutdownd has been removed. This service was
13102 previously responsible for implementing scheduled shutdowns
13103 as exposed in /usr/bin/shutdown's time parameter. This
13104 functionality has now been moved into systemd-logind and is
13105 accessible via a bus interface.
13107 * "systemctl reboot" gained a new switch --firmware-setup that
13108 can be used to reboot into the EFI firmware setup, if that
13109 is available. systemd-logind now exposes an API on the bus
13110 to trigger such reboots, in case graphical desktop UIs want
13111 to cover this functionality.
13113 * "systemctl enable", "systemctl disable" and "systemctl mask"
13114 now support a new "--now" switch. If specified the units
13115 that are enabled will also be started, and the ones
13116 disabled/masked also stopped.
13118 * The Gummiboot EFI boot loader tool has been merged into
13119 systemd, and renamed to "systemd-boot". The bootctl tool has been
13120 updated to support systemd-boot.
13122 * An EFI kernel stub has been added that may be used to create
13123 kernel EFI binaries that contain not only the actual kernel,
13124 but also an initrd, boot splash, command line and OS release
13125 information. This combined binary can then be signed as a
13126 single image, so that the firmware can verify it all in one
13127 step. systemd-boot has special support for EFI binaries created
13128 like this and can extract OS release information from them
13129 and show them in the boot menu. This functionality is useful
13130 to implement cryptographically verified boot schemes.
13132 * Optional support has been added to systemd-fsck to pass
13133 fsck's progress report to an AF_UNIX socket in the file
13136 * udev will no longer create device symlinks for all block devices by
13137 default. A deny list for excluding special block devices from this
13138 logic has been turned into an allow list that requires picking block
13139 devices explicitly that require device symlinks.
13141 * A new (currently still internal) API sd-device.h has been
13142 added to libsystemd. This modernized API is supposed to
13143 replace libudev eventually. In fact, already much of libudev
13144 is now just a wrapper around sd-device.h.
13146 * A new hwdb database for storing metadata about pointing
13147 stick devices has been added.
13149 * systemd-tmpfiles gained support for setting file attributes
13150 similar to the "chattr" tool with new 'h' and 'H' lines.
13152 * systemd-journald will no longer unconditionally set the
13153 btrfs NOCOW flag on new journal files. This is instead done
13154 with tmpfiles snippet using the new 'h' line type. This
13155 allows easy disabling of this logic, by masking the
13156 journal-nocow.conf tmpfiles file.
13158 * systemd-journald will now translate audit message types to
13159 human readable identifiers when writing them to the
13160 journal. This should improve readability of audit messages.
13162 * The LUKS logic gained support for the offset= and skip=
13163 options in /etc/crypttab, as previously implemented by
13166 * /usr/lib/os-release gained a new optional field VARIANT= for
13167 distributions that support multiple variants (such as a
13168 desktop edition, a server edition, …)
13170 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
13171 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
13172 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
13173 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
13174 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
13175 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
13176 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
13177 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
13178 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
13179 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
13180 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
13181 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
13182 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
13183 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
13184 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
13185 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
13186 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
13187 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
13188 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
13189 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
13190 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
13191 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
13192 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
13193 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
13194 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
13195 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
13196 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
13198 — Berlin, 2015-05-22
13202 * Introduce a new API "sd-hwdb.h" for querying the hardware
13203 metadata database. With this minimal interface one can query
13204 and enumerate the udev hwdb, decoupled from the old libudev
13205 library. libudev's interface for this is now only a wrapper
13206 around sd-hwdb. A new tool systemd-hwdb has been added to
13207 interface with and update the database.
13209 * When any of systemd's tools copies files (for example due to
13210 tmpfiles' C lines) a btrfs reflink will attempted first,
13211 before bytewise copying is done.
13213 * systemd-nspawn gained a new --ephemeral switch. When
13214 specified a btrfs snapshot is taken of the container's root
13215 directory, and immediately removed when the container
13216 terminates again. Thus, a container can be started whose
13217 changes never alter the container's root directory, and are
13218 lost on container termination. This switch can also be used
13219 for starting a container off the root file system of the
13220 host without affecting the host OS. This switch is only
13221 available on btrfs file systems.
13223 * systemd-nspawn gained a new --template= switch. It takes the
13224 path to a container tree to use as template for the tree
13225 specified via --directory=, should that directory be
13226 missing. This allows instantiating containers dynamically,
13227 on first run. This switch is only available on btrfs file
13230 * When a .mount unit refers to a mount point on which multiple
13231 mounts are stacked, and the .mount unit is stopped all of
13232 the stacked mount points will now be unmounted until no
13233 mount point remains.
13235 * systemd now has an explicit notion of supported and
13236 unsupported unit types. Jobs enqueued for unsupported unit
13237 types will now fail with an "unsupported" error code. More
13238 specifically .swap, .automount and .device units are not
13239 supported in containers, .busname units are not supported on
13240 non-kdbus systems. .swap and .automount are also not
13241 supported if their respective kernel compile time options
13244 * machinectl gained support for two new "copy-from" and
13245 "copy-to" commands for copying files from a running
13246 container to the host or vice versa.
13248 * machinectl gained support for a new "bind" command to bind
13249 mount host directories into local containers. This is
13250 currently only supported for nspawn containers.
13252 * networkd gained support for configuring bridge forwarding
13253 database entries (fdb) from .network files.
13255 * A new tiny daemon "systemd-importd" has been added that can
13256 download container images in tar, raw, qcow2 or dkr formats,
13257 and make them available locally in /var/lib/machines, so
13258 that they can run as nspawn containers. The daemon can GPG
13259 verify the downloads (not supported for dkr, since it has no
13260 provisions for verifying downloads). It will transparently
13261 decompress bz2, xz, gzip compressed downloads if necessary,
13262 and restore sparse files on disk. The daemon uses privilege
13263 separation to ensure the actual download logic runs with
13264 fewer privileges than the daemon itself. machinectl has
13265 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
13266 make the functionality of importd available to the
13267 user. With this in place the Fedora and Ubuntu "Cloud"
13268 images can be downloaded and booted as containers unmodified
13269 (the Fedora images lack the appropriate GPG signature files
13270 currently, so they cannot be verified, but this will change
13271 soon, hopefully). Note that downloading images is currently
13272 only fully supported on btrfs.
13274 * machinectl is now able to list container images found in
13275 /var/lib/machines, along with some metadata about sizes of
13276 disk and similar. If the directory is located on btrfs and
13277 quota is enabled, this includes quota display. A new command
13278 "image-status" has been added that shows additional
13279 information about images.
13281 * machinectl is now able to clone container images
13282 efficiently, if the underlying file system (btrfs) supports
13283 it, with the new "machinectl clone" command. It also
13284 gained commands for renaming and removing images, as well as
13285 marking them read-only or read-write (supported also on
13286 legacy file systems).
13288 * networkd gained support for collecting LLDP network
13289 announcements, from hardware that supports this. This is
13290 shown in networkctl output.
13292 * systemd-run gained support for a new -t (--pty) switch for
13293 invoking a binary on a pty whose input and output is
13294 connected to the invoking terminal. This allows executing
13295 processes as system services while interactively
13296 communicating with them via the terminal. Most interestingly
13297 this is supported across container boundaries. Invoking
13298 "systemd-run -t /bin/bash" is an alternative to running a
13299 full login session, the difference being that the former
13300 will not register a session, nor go through the PAM session
13303 * tmpfiles gained support for a new "v" line type for creating
13304 btrfs subvolumes. If the underlying file system is a legacy
13305 file system, this automatically degrades to creating a
13306 normal directory. Among others /var/lib/machines is now
13307 created like this at boot, should it be missing.
13309 * The directory /var/lib/containers/ has been deprecated and
13310 been replaced by /var/lib/machines. The term "machines" has
13311 been used in the systemd context as generic term for both
13312 VMs and containers, and hence appears more appropriate for
13313 this, as the directory can also contain raw images bootable
13316 * systemd-nspawn when invoked with -M but without --directory=
13317 or --image= is now capable of searching for the container
13318 root directory, subvolume or disk image automatically, in
13319 /var/lib/machines. systemd-nspawn@.service has been updated
13320 to make use of this, thus allowing it to be used for raw
13323 * A new machines.target unit has been introduced that is
13324 supposed to group all containers/VMs invoked as services on
13325 the system. systemd-nspawn@.service has been updated to
13326 integrate with that.
13328 * machinectl gained a new "start" command, for invoking a
13329 container as a service. "machinectl start foo" is mostly
13330 equivalent to "systemctl start systemd-nspawn@foo.service",
13331 but handles escaping in a nicer way.
13333 * systemd-nspawn will now mount most of the cgroupfs tree
13334 read-only into each container, with the exception of the
13335 container's own subtree in the name=systemd hierarchy.
13337 * journald now sets the special FS_NOCOW file flag for its
13338 journal files. This should improve performance on btrfs, by
13339 avoiding heavy fragmentation when journald's write-pattern
13340 is used on COW file systems. It degrades btrfs' data
13341 integrity guarantees for the files to the same levels as for
13342 ext3/ext4 however. This should be OK though as journald does
13343 its own data integrity checks and all its objects are
13344 checksummed on disk. Also, journald should handle btrfs disk
13345 full events a lot more gracefully now, by processing SIGBUS
13346 errors, and not relying on fallocate() anymore.
13348 * When journald detects that journal files it is writing to
13349 have been deleted it will immediately start new journal
13352 * systemd now provides a way to store file descriptors
13353 per-service in PID 1. This is useful for daemons to ensure
13354 that fds they require are not lost during a daemon
13355 restart. The fds are passed to the daemon on the next
13356 invocation in the same way socket activation fds are
13357 passed. This is now used by journald to ensure that the
13358 various sockets connected to all the system's stdout/stderr
13359 are not lost when journald is restarted. File descriptors
13360 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
13361 an extension to sd_notify(). Note that a limit is enforced
13362 on the number of fds a service can store in PID 1, and it
13363 defaults to 0, so that no fds may be stored, unless this is
13364 explicitly turned on.
13366 * The default TERM variable to use for units connected to a
13367 terminal, when no other value is explicitly is set is now
13368 vt220 rather than vt102. This should be fairly safe still,
13369 but allows PgUp/PgDn work.
13371 * The /etc/crypttab option header= as known from Debian is now
13374 * "loginctl user-status" and "loginctl session-status" will
13375 now show the last 10 lines of log messages of the
13376 user/session following the status output. Similar,
13377 "machinectl status" will show the last 10 log lines
13378 associated with a virtual machine or container
13379 service. (Note that this is usually not the log messages
13380 done in the VM/container itself, but simply what the
13381 container manager logs. For nspawn this includes all console
13384 * "loginctl session-status" without further argument will now
13385 show the status of the session of the caller. Similar,
13386 "lock-session", "unlock-session", "activate",
13387 "enable-linger", "disable-linger" may now be called without
13388 session/user parameter in which case they apply to the
13389 caller's session/user.
13391 * An X11 session scriptlet is now shipped that uploads
13392 $DISPLAY and $XAUTHORITY into the environment of the systemd
13393 --user daemon if a session begins. This should improve
13394 compatibility with X11 enabled applications run as systemd
13397 * Generators are now subject to masking via /etc and /run, the
13398 same way as unit files.
13400 * networkd .network files gained support for configuring
13401 per-link IPv4/IPv6 packet forwarding as well as IPv4
13402 masquerading. This is by default turned on for veth links to
13403 containers, as registered by systemd-nspawn. This means that
13404 nspawn containers run with --network-veth will now get
13405 automatic routed access to the host's networks without any
13406 further configuration or setup, as long as networkd runs on
13409 * systemd-nspawn gained the --port= (-p) switch to expose TCP
13410 or UDP posts of a container on the host. With this in place
13411 it is possible to run containers with private veth links
13412 (--network-veth), and have their functionality exposed on
13413 the host as if their services were running directly on the
13416 * systemd-nspawn's --network-veth switch now gained a short
13417 version "-n", since with the changes above it is now truly
13418 useful out-of-the-box. The systemd-nspawn@.service has been
13419 updated to make use of it too by default.
13421 * systemd-nspawn will now maintain a per-image R/W lock, to
13422 ensure that the same image is not started more than once
13423 writable. (It's OK to run an image multiple times
13424 simultaneously in read-only mode.)
13426 * systemd-nspawn's --image= option is now capable of
13427 dissecting and booting MBR and GPT disk images that contain
13428 only a single active Linux partition. Previously it
13429 supported only GPT disk images with proper GPT type
13430 IDs. This allows running cloud images from major
13431 distributions directly with systemd-nspawn, without
13434 * In addition to collecting mouse dpi data in the udev
13435 hardware database, there's now support for collecting angle
13436 information for mouse scroll wheels. The database is
13437 supposed to guarantee similar scrolling behavior on mice
13438 that it knows about. There's also support for collecting
13439 information about Touchpad types.
13441 * udev's input_id built-in will now also collect touch screen
13442 dimension data and attach it to probed devices.
13444 * /etc/os-release gained support for a Distribution Privacy
13447 * networkd gained support for creating "ipvlan", "gretap",
13448 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
13450 * systemd-tmpfiles gained support for "a" lines for setting
13453 * systemd-nspawn will now mount /tmp in the container to
13454 tmpfs, automatically.
13456 * systemd now exposes the memory.usage_in_bytes cgroup
13457 attribute and shows it for each service in the "systemctl
13458 status" output, if available.
13460 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
13461 immediate reboot is triggered. This useful if shutdown is
13462 hung and is unable to complete, to expedite the
13463 operation. Note that this kind of reboot will still unmount
13464 all file systems, and hence should not result in fsck being
13465 run on next reboot.
13467 * A .device unit for an optical block device will now be
13468 considered active only when a medium is in the drive. Also,
13469 mount units are now bound to their backing devices thus
13470 triggering automatic unmounting when devices become
13471 unavailable. With this in place systemd will now
13472 automatically unmount left-over mounts when a CD-ROM is
13473 ejected or a USB stick is yanked from the system.
13475 * networkd-wait-online now has support for waiting for
13476 specific interfaces only (with globbing), and for giving up
13477 after a configurable timeout.
13479 * networkd now exits when idle. It will be automatically
13480 restarted as soon as interfaces show up, are removed or
13481 change state. networkd will stay around as long as there is
13482 at least one DHCP state machine or similar around, that keep
13485 * networkd may now configure IPv6 link-local addressing in
13486 addition to IPv4 link-local addressing.
13488 * The IPv6 "token" for use in SLAAC may now be configured for
13489 each .network interface in networkd.
13491 * Routes configured with networkd may now be assigned a scope
13494 * networkd's [Match] sections now support globbing and lists
13495 of multiple space-separated matches per item.
13497 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
13498 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
13499 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
13500 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
13501 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
13502 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
13503 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
13504 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
13505 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
13506 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
13507 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
13508 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
13509 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
13510 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
13511 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
13512 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
13513 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
13514 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
13515 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
13516 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
13517 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
13518 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
13519 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
13520 Hoffmann, Zbigniew Jędrzejewski-Szmek
13522 — Berlin, 2015-02-16
13526 * When querying unit file enablement status (for example via
13527 "systemctl is-enabled"), a new state "indirect" is now known
13528 which indicates that a unit might not be enabled itself, but
13529 another unit listed in its Also= setting might be.
13531 * Similar to the various existing ConditionXYZ= settings for
13532 units, there are now matching AssertXYZ= settings. While
13533 failing conditions cause a unit to be skipped, but its job
13534 to succeed, failing assertions declared like this will cause
13535 a unit start operation and its job to fail.
13537 * hostnamed now knows a new chassis type "embedded".
13539 * systemctl gained a new "edit" command. When used on a unit
13540 file, this allows extending unit files with .d/ drop-in
13541 configuration snippets or editing the full file (after
13542 copying it from /usr/lib to /etc). This will invoke the
13543 user's editor (as configured with $EDITOR), and reload the
13544 modified configuration after editing.
13546 * "systemctl status" now shows the suggested enablement state
13547 for a unit, as declared in the (usually vendor-supplied)
13548 system preset files.
13550 * nss-myhostname will now resolve the single-label hostname
13551 "gateway" to the locally configured default IP routing
13552 gateways, ordered by their metrics. This assigns a stable
13553 name to the used gateways, regardless which ones are
13554 currently configured. Note that the name will only be
13555 resolved after all other name sources (if nss-myhostname is
13556 configured properly) and should hence not negatively impact
13557 systems that use the single-label hostname "gateway" in
13560 * systemd-inhibit now allows filtering by mode when listing
13563 * Scope and service units gained a new "Delegate" boolean
13564 property, which, when set, allows processes running inside the
13565 unit to further partition resources. This is primarily
13566 useful for systemd user instances as well as container
13569 * journald will now pick up audit messages directly from
13570 the kernel, and log them like any other log message. The
13571 audit fields are split up and fully indexed. This means that
13572 journalctl in many ways is now a (nicer!) alternative to
13573 ausearch, the traditional audit client. Note that this
13574 implements only a minimal audit client. If you want the
13575 special audit modes like reboot-on-log-overflow, please use
13576 the traditional auditd instead, which can be used in
13577 parallel to journald.
13579 * The ConditionSecurity= unit file option now understands the
13580 special string "audit" to check whether auditing is
13583 * journalctl gained two new commands --vacuum-size= and
13584 --vacuum-time= to delete old journal files until the
13585 remaining ones take up no more than the specified size on disk,
13586 or are not older than the specified time.
13588 * A new, native PPPoE library has been added to sd-network,
13589 systemd's library of light-weight networking protocols. This
13590 library will be used in a future version of networkd to
13591 enable PPPoE communication without an external pppd daemon.
13593 * The busctl tool now understands a new "capture" verb that
13594 works similar to "monitor", but writes a packet capture
13595 trace to STDOUT that can be redirected to a file which is
13596 compatible with libcap's capture file format. This can then
13597 be loaded in Wireshark and similar tools to inspect bus
13600 * The busctl tool now understands a new "tree" verb that shows
13601 the object trees of a specific service on the bus, or of all
13604 * The busctl tool now understands a new "introspect" verb that
13605 shows all interfaces and members of objects on the bus,
13606 including their signature and values. This is particularly
13607 useful to get more information about bus objects shown by
13608 the new "busctl tree" command.
13610 * The busctl tool now understands new verbs "call",
13611 "set-property" and "get-property" for invoking bus method
13612 calls, setting and getting bus object properties in a
13615 * busctl gained a new --augment-creds= argument that controls
13616 whether the tool shall augment credential information it
13617 gets from the bus with data from /proc, in a possibly
13620 * nspawn's --link-journal= switch gained two new values
13621 "try-guest" and "try-host" that work like "guest" and
13622 "host", but do not fail if the host has no persistent
13623 journaling enabled. -j is now equivalent to
13624 --link-journal=try-guest.
13626 * macvlan network devices created by nspawn will now have
13627 stable MAC addresses.
13629 * A new SmackProcessLabel= unit setting has been added, which
13630 controls the SMACK security label processes forked off by
13631 the respective unit shall use.
13633 * If compiled with --enable-xkbcommon, systemd-localed will
13634 verify x11 keymap settings by compiling the given keymap. It
13635 will spew out warnings if the compilation fails. This
13636 requires libxkbcommon to be installed.
13638 * When a coredump is collected, a larger number of metadata
13639 fields is now collected and included in the journal records
13640 created for it. More specifically, control group membership,
13641 environment variables, memory maps, working directory,
13642 chroot directory, /proc/$PID/status, and a list of open file
13643 descriptors is now stored in the log entry.
13645 * The udev hwdb now contains DPI information for mice. For
13648 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
13650 * All systemd programs that read standalone configuration
13651 files in /etc now also support a corresponding series of
13652 .conf.d configuration directories in /etc/, /run/,
13653 /usr/local/lib/, /usr/lib/, and (if configured with
13654 --enable-split-usr) /lib/. In particular, the following
13655 configuration files now have corresponding configuration
13656 directories: system.conf user.conf, logind.conf,
13657 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
13658 resolved.conf, timesyncd.conf, journal-remote.conf, and
13659 journal-upload.conf. Note that distributions should use the
13660 configuration directories in /usr/lib/; the directories in
13661 /etc/ are reserved for the system administrator.
13663 * systemd-rfkill will no longer take the rfkill device name
13664 into account when storing rfkill state on disk, as the name
13665 might be dynamically assigned and not stable. Instead, the
13666 ID_PATH udev variable combined with the rfkill type (wlan,
13667 bluetooth, …) is used.
13669 * A new service systemd-machine-id-commit.service has been
13670 added. When used on systems where /etc is read-only during
13671 boot, and /etc/machine-id is not initialized (but an empty
13672 file), this service will copy the temporary machine ID
13673 created as replacement into /etc after the system is fully
13674 booted up. This is useful for systems that are freshly
13675 installed with a non-initialized machine ID, but should get
13676 a fixed machine ID for subsequent boots.
13678 * networkd's .netdev files now provide a large set of
13679 configuration parameters for VXLAN devices. Similarly, the
13680 bridge port cost parameter is now configurable in .network
13681 files. There's also new support for configuring IP source
13682 routing. networkd .link files gained support for a new
13683 OriginalName= match that is useful to match against the
13684 original interface name the kernel assigned. .network files
13685 may include MTU= and MACAddress= fields for altering the MTU
13686 and MAC address while being connected to a specific network
13689 * The LUKS logic gained supported for configuring
13690 UUID-specific key files. There's also new support for naming
13691 LUKS device from the kernel command line, using the new
13692 luks.name= argument.
13694 * Timer units may now be transiently created via the bus API
13695 (this was previously already available for scope and service
13696 units). In addition it is now possible to create multiple
13697 transient units at the same time with a single bus call. The
13698 "systemd-run" tool has been updated to make use of this for
13699 running commands on a specified time, in at(1)-style.
13701 * tmpfiles gained support for "t" lines, for assigning
13702 extended attributes to files. Among other uses this may be
13703 used to assign SMACK labels to files.
13705 Contributions from: Alin Rauta, Alison Chaiken, Andrej
13706 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
13707 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
13708 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
13709 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
13710 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
13711 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
13712 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
13713 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
13714 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
13715 Michael Biebl, Michael Chapman, Michael Marineau, Michal
13716 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
13717 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
13718 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
13719 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
13720 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
13721 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
13722 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
13724 — Berlin, 2014-12-10
13728 * journalctl gained the new options -t/--identifier= to match
13729 on the syslog identifier (aka "tag"), as well as --utc to
13730 show log timestamps in the UTC timezone. journalctl now also
13731 accepts -n/--lines=all to disable line capping in a pager.
13733 * journalctl gained a new switch, --flush, that synchronously
13734 flushes logs from /run/log/journal to /var/log/journal if
13735 persistent storage is enabled. systemd-journal-flush.service
13736 now waits until the operation is complete.
13738 * Services can notify the manager before they start a reload
13739 (by sending RELOADING=1) or shutdown (by sending
13740 STOPPING=1). This allows the manager to track and show the
13741 internal state of daemons and closes a race condition when
13742 the process is still running but has closed its D-Bus
13745 * Services with Type=oneshot do not have to have any ExecStart
13748 * User units are now loaded also from
13749 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
13750 /run/systemd/user directory that was already previously
13751 supported, but is under the control of the user.
13753 * Job timeouts (i.e. timeouts on the time a job that is
13754 queued stays in the run queue) can now optionally result in
13755 immediate reboot or power-off actions (JobTimeoutAction= and
13756 JobTimeoutRebootArgument=). This is useful on ".target"
13757 units, to limit the maximum time a target remains
13758 undispatched in the run queue, and to trigger an emergency
13759 operation in such a case. This is now used by default to
13760 turn off the system if boot-up (as defined by everything in
13761 basic.target) hangs and does not complete for at least
13762 15min. Also, if power-off or reboot hang for at least 30min
13763 an immediate power-off/reboot operation is triggered. This
13764 functionality is particularly useful to increase reliability
13765 on embedded devices, but also on laptops which might
13766 accidentally get powered on when carried in a backpack and
13767 whose boot stays stuck in a hard disk encryption passphrase
13770 * systemd-logind can be configured to also handle lid switch
13771 events even when the machine is docked or multiple displays
13772 are attached (HandleLidSwitchDocked= option).
13774 * A helper binary and a service have been added which can be
13775 used to resume from hibernation in the initramfs. A
13776 generator will parse the resume= option on the kernel
13777 command line to trigger resume.
13779 * A user console daemon systemd-consoled has been
13780 added. Currently, it is a preview, and will so far open a
13781 single terminal on each session of the user marked as
13782 Desktop=systemd-console.
13784 * Route metrics can be specified for DHCP routes added by
13787 * The SELinux context of socket-activated services can be set
13788 from the information provided by the networking stack
13789 (SELinuxContextFromNet= option).
13791 * Userspace firmware loading support has been removed and
13792 the minimum supported kernel version is thus bumped to 3.7.
13794 * Timeout for udev workers has been increased from 1 to 3
13795 minutes, but a warning will be printed after 1 minute to
13796 help diagnose kernel modules that take a long time to load.
13798 * Udev rules can now remove tags on devices with TAG-="foobar".
13800 * systemd's readahead implementation has been removed. In many
13801 circumstances it didn't give expected benefits even for
13802 rotational disk drives and was becoming less relevant in the
13803 age of SSDs. As none of the developers has been using
13804 rotating media anymore, and nobody stepped up to actively
13805 maintain this component of systemd it has now been removed.
13807 * Swap units can use Options= to specify discard options.
13808 Discard options specified for swaps in /etc/fstab are now
13811 * Docker containers are now detected as a separate type of
13814 * The Password Agent protocol gained support for queries where
13815 the user input is shown, useful e.g. for user names.
13816 systemd-ask-password gained a new --echo option to turn that
13819 * The default sysctl.d/ snippets will now set:
13821 net.core.default_qdisc = fq_codel
13823 This selects Fair Queuing Controlled Delay as the default
13824 queuing discipline for network interfaces. fq_codel helps
13825 fight the network bufferbloat problem. It is believed to be
13826 a good default with no tuning required for most workloads.
13827 Downstream distributions may override this choice. On 10Gbit
13828 servers that do not do forwarding, "fq" may perform better.
13829 Systems without a good clocksource should use "pfifo_fast".
13831 * If kdbus is enabled during build a new option BusPolicy= is
13832 available for service units, that allows locking all service
13833 processes into a stricter bus policy, in order to limit
13834 access to various bus services, or even hide most of them
13835 from the service's view entirely.
13837 * networkctl will now show the .network and .link file
13838 networkd has applied to a specific interface.
13840 * sd-login gained a new API call sd_session_get_desktop() to
13841 query which desktop environment has been selected for a
13844 * UNIX utmp support is now compile-time optional to support
13845 legacy-free systems.
13847 * systemctl gained two new commands "add-wants" and
13848 "add-requires" for pulling in units from specific targets
13851 * If the word "rescue" is specified on the kernel command line
13852 the system will now boot into rescue mode (aka
13853 rescue.target), which was previously available only by
13854 specifying "1" or "systemd.unit=rescue.target" on the kernel
13855 command line. This new kernel command line option nicely
13856 mirrors the already existing "emergency" kernel command line
13859 * New kernel command line options mount.usr=, mount.usrflags=,
13860 mount.usrfstype= have been added that match root=, rootflags=,
13861 rootfstype= but allow mounting a specific file system to
13864 * The $NOTIFY_SOCKET is now also passed to control processes of
13865 services, not only the main process.
13867 * This version reenables support for fsck's -l switch. This
13868 means at least version v2.25 of util-linux is required for
13869 operation, otherwise dead-locks on device nodes may
13870 occur. Again: you need to update util-linux to at least
13871 v2.25 when updating systemd to v217.
13873 * The "multi-seat-x" tool has been removed from systemd, as
13874 its functionality has been integrated into X servers 1.16,
13875 and the tool is hence redundant. It is recommended to update
13876 display managers invoking this tool to simply invoke X
13877 directly from now on, again.
13879 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
13880 message flag has been added for all of systemd's polkit
13881 authenticated method calls has been added. In particular this
13882 now allows optional interactive authorization via polkit for
13883 many of PID1's privileged operations such as unit file
13884 enabling and disabling.
13886 * "udevadm hwdb --update" learnt a new switch "--usr" for
13887 placing the rebuilt hardware database in /usr instead of
13888 /etc. When used only hardware database entries stored in
13889 /usr will be used, and any user database entries in /etc are
13890 ignored. This functionality is useful for vendors to ship a
13891 pre-built database on systems where local configuration is
13892 unnecessary or unlikely.
13894 * Calendar time specifications in .timer units now also
13895 understand the strings "semi-annually", "quarterly" and
13896 "minutely" as shortcuts (in addition to the preexisting
13897 "annually", "hourly", …).
13899 * systemd-tmpfiles will now correctly create files in /dev
13900 at boot which are marked for creation only at boot. It is
13901 recommended to always create static device nodes with 'c!'
13902 and 'b!', so that they are created only at boot and not
13903 overwritten at runtime.
13905 * When the watchdog logic is used for a service (WatchdogSec=)
13906 and the watchdog timeout is hit the service will now be
13907 terminated with SIGABRT (instead of just SIGTERM), in order
13908 to make sure a proper coredump and backtrace is
13909 generated. This ensures that hanging services will result in
13910 similar coredump/backtrace behaviour as services that hit a
13911 segmentation fault.
13913 Contributions from: Andreas Henriksson, Andrei Borzenkov,
13914 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
13915 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
13916 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
13917 Herrmann, David Sommerseth, David Strauss, Emil Renner
13918 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
13919 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
13920 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
13921 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
13922 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
13923 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
13924 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
13925 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
13926 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
13927 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
13928 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
13929 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
13930 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
13931 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
13932 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
13933 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
13936 — Berlin, 2014-10-28
13940 * timedated no longer reads NTP implementation unit names from
13941 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
13942 implementations should add a
13944 Conflicts=systemd-timesyncd.service
13946 to their unit files to take over and replace systemd's NTP
13947 default functionality.
13949 * systemd-sysusers gained a new line type "r" for configuring
13950 which UID/GID ranges to allocate system users/groups
13951 from. Lines of type "u" may now add an additional column
13952 that specifies the home directory for the system user to be
13953 created. Also, systemd-sysusers may now optionally read user
13954 information from STDIN instead of a file. This is useful for
13955 invoking it from RPM preinst scriptlets that need to create
13956 users before the first RPM file is installed since these
13957 files might need to be owned by them. A new
13958 %sysusers_create_inline RPM macro has been introduced to do
13959 just that. systemd-sysusers now updates the shadow files as
13960 well as the user/group databases, which should enhance
13961 compatibility with certain tools like grpck.
13963 * A number of bus APIs of PID 1 now optionally consult polkit to
13964 permit access for otherwise unprivileged clients under certain
13965 conditions. Note that this currently doesn't support
13966 interactive authentication yet, but this is expected to be
13967 added eventually, too.
13969 * /etc/machine-info now has new fields for configuring the
13970 deployment environment of the machine, as well as the
13971 location of the machine. hostnamectl has been updated with
13972 new command to update these fields.
13974 * systemd-timesyncd has been updated to automatically acquire
13975 NTP server information from systemd-networkd, which might
13976 have been discovered via DHCP.
13978 * systemd-resolved now includes a caching DNS stub resolver
13979 and a complete LLMNR name resolution implementation. A new
13980 NSS module "nss-resolve" has been added which can be used
13981 instead of glibc's own "nss-dns" to resolve hostnames via
13982 systemd-resolved. Hostnames, addresses and arbitrary RRs may
13983 be resolved via systemd-resolved D-Bus APIs. In contrast to
13984 the glibc internal resolver systemd-resolved is aware of
13985 multi-homed system, and keeps DNS server and caches separate
13986 and per-interface. Queries are sent simultaneously on all
13987 interfaces that have DNS servers configured, in order to
13988 properly handle VPNs and local LANs which might resolve
13989 separate sets of domain names. systemd-resolved may acquire
13990 DNS server information from systemd-networkd automatically,
13991 which in turn might have discovered them via DHCP. A tool
13992 "systemd-resolve-host" has been added that may be used to
13993 query the DNS logic in resolved. systemd-resolved implements
13994 IDNA and automatically uses IDNA or UTF-8 encoding depending
13995 on whether classic DNS or LLMNR is used as transport. In the
13996 next releases we intend to add a DNSSEC and mDNS/DNS-SD
13997 implementation to systemd-resolved.
13999 * A new NSS module nss-mymachines has been added, that
14000 automatically resolves the names of all local registered
14001 containers to their respective IP addresses.
14003 * A new client tool "networkctl" for systemd-networkd has been
14004 added. It currently is entirely passive and will query
14005 networking configuration from udev, rtnetlink and networkd,
14006 and present it to the user in a very friendly
14007 way. Eventually, we hope to extend it to become a full
14008 control utility for networkd.
14010 * .socket units gained a new DeferAcceptSec= setting that
14011 controls the kernels' TCP_DEFER_ACCEPT sockopt for
14012 TCP. Similarly, support for controlling TCP keep-alive
14013 settings has been added (KeepAliveTimeSec=,
14014 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
14015 turning off Nagle's algorithm on TCP has been added
14018 * logind learned a new session type "web", for use in projects
14019 like Cockpit which register web clients as PAM sessions.
14021 * timer units with at least one OnCalendar= setting will now
14022 be started only after time-sync.target has been
14023 reached. This way they will not elapse before the system
14024 clock has been corrected by a local NTP client or
14025 similar. This is particular useful on RTC-less embedded
14026 machines, that come up with an invalid system clock.
14028 * systemd-nspawn's --network-veth= switch should now result in
14029 stable MAC addresses for both the outer and the inner side
14032 * systemd-nspawn gained a new --volatile= switch for running
14033 container instances with /etc or /var unpopulated.
14035 * The kdbus client code has been updated to use the new Linux
14036 3.17 memfd subsystem instead of the old kdbus-specific one.
14038 * systemd-networkd's DHCP client and server now support
14039 FORCERENEW. There are also new configuration options to
14040 configure the vendor client identifier and broadcast mode
14043 * systemd will no longer inform the kernel about the current
14044 timezone, as this is necessarily incorrect and racy as the
14045 kernel has no understanding of DST and similar
14046 concepts. This hence means FAT timestamps will be always
14047 considered UTC, similar to what Android is already
14048 doing. Also, when the RTC is configured to the local time
14049 (rather than UTC) systemd will never synchronize back to it,
14050 as this might confuse Windows at a later boot.
14052 * systemd-analyze gained a new command "verify" for offline
14053 validation of unit files.
14055 * systemd-networkd gained support for a couple of additional
14056 settings for bonding networking setups. Also, the metric for
14057 statically configured routes may now be configured. For
14058 network interfaces where this is appropriate the peer IP
14059 address may now be configured.
14061 * systemd-networkd's DHCP client will no longer request
14062 broadcasting by default, as this tripped up some networks.
14063 For hardware where broadcast is required the feature should
14064 be switched back on using RequestBroadcast=yes.
14066 * systemd-networkd will now set up IPv4LL addresses (when
14067 enabled) even if DHCP is configured successfully.
14069 * udev will now default to respect network device names given
14070 by the kernel when the kernel indicates that these are
14071 predictable. This behavior can be tweaked by changing
14072 NamePolicy= in the relevant .link file.
14074 * A new library systemd-terminal has been added that
14075 implements full TTY stream parsing and rendering. This
14076 library is supposed to be used later on for implementing a
14077 full userspace VT subsystem, replacing the current kernel
14080 * A new tool systemd-journal-upload has been added to push
14081 journal data to a remote system running
14082 systemd-journal-remote.
14084 * journald will no longer forward all local data to another
14085 running syslog daemon. This change has been made because
14086 rsyslog (which appears to be the most commonly used syslog
14087 implementation these days) no longer makes use of this, and
14088 instead pulls the data out of the journal on its own. Since
14089 forwarding the messages to a non-existent syslog server is
14090 more expensive than we assumed we have now turned this
14091 off. If you run a syslog server that is not a recent rsyslog
14092 version, you have to turn this option on again
14093 (ForwardToSyslog= in journald.conf).
14095 * journald now optionally supports the LZ4 compressor for
14096 larger journal fields. This compressor should perform much
14097 better than XZ which was the previous default.
14099 * machinectl now shows the IP addresses of local containers,
14100 if it knows them, plus the interface name of the container.
14102 * A new tool "systemd-escape" has been added that makes it
14103 easy to escape strings to build unit names and similar.
14105 * sd_notify() messages may now include a new ERRNO= field
14106 which is parsed and collected by systemd and shown among the
14107 "systemctl status" output for a service.
14109 * A new component "systemd-firstboot" has been added that
14110 queries the most basic systemd information (timezone,
14111 hostname, root password) interactively on first
14112 boot. Alternatively it may also be used to provision these
14113 things offline on OS images installed into directories.
14115 * The default sysctl.d/ snippets will now set
14117 net.ipv4.conf.default.promote_secondaries=1
14119 This has the benefit of no flushing secondary IP addresses
14120 when primary addresses are removed.
14122 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
14123 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
14124 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
14125 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
14126 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
14127 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
14128 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
14129 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
14130 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
14131 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
14132 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
14133 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
14134 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
14135 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
14136 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
14138 — Berlin, 2014-08-19
14142 * A new tool systemd-sysusers has been added. This tool
14143 creates system users and groups in /etc/passwd and
14144 /etc/group, based on static declarative system user/group
14145 definitions in /usr/lib/sysusers.d/. This is useful to
14146 enable factory resets and volatile systems that boot up with
14147 an empty /etc directory, and thus need system users and
14148 groups created during early boot. systemd now also ships
14149 with two default sysusers.d/ files for the most basic
14150 users and groups systemd and the core operating system
14153 * A new tmpfiles snippet has been added that rebuilds the
14154 essential files in /etc on boot, should they be missing.
14156 * A directive for ensuring automatic clean-up of
14157 /var/cache/man/ has been removed from the default
14158 configuration. This line should now be shipped by the man
14159 implementation. The necessary change has been made to the
14160 man-db implementation. Note that you need to update your man
14161 implementation to one that ships this line, otherwise no
14162 automatic clean-up of /var/cache/man will take place.
14164 * A new condition ConditionNeedsUpdate= has been added that
14165 may conditionalize services to only run when /etc or /var
14166 are "older" than the vendor operating system resources in
14167 /usr. This is useful for reconstructing or updating /etc
14168 after an offline update of /usr or a factory reset, on the
14169 next reboot. Services that want to run once after such an
14170 update or reset should use this condition and order
14171 themselves before the new systemd-update-done.service, which
14172 will mark the two directories as fully updated. A number of
14173 service files have been added making use of this, to rebuild
14174 the udev hardware database, the journald message catalog and
14175 dynamic loader cache (ldconfig). The systemd-sysusers tool
14176 described above also makes use of this now. With this in
14177 place it is now possible to start up a minimal operating
14178 system with /etc empty cleanly. For more information on the
14179 concepts involved see this recent blog story:
14181 https://0pointer.de/blog/projects/stateless.html
14183 * A new system group "input" has been introduced, and all
14184 input device nodes get this group assigned. This is useful
14185 for system-level software to get access to input devices. It
14186 complements what is already done for "audio" and "video".
14188 * systemd-networkd learnt minimal DHCPv4 server support in
14189 addition to the existing DHCPv4 client support. It also
14190 learnt DHCPv6 client and IPv6 Router Solicitation client
14191 support. The DHCPv4 client gained support for static routes
14192 passed in from the server. Note that the [DHCPv4] section
14193 known in older systemd-networkd versions has been renamed to
14194 [DHCP] and is now also used by the DHCPv6 client. Existing
14195 .network files using settings of this section should be
14196 updated, though compatibility is maintained. Optionally, the
14197 client hostname may now be sent to the DHCP server.
14199 * networkd gained support for vxlan virtual networks as well
14200 as tun/tap and dummy devices.
14202 * networkd gained support for automatic allocation of address
14203 ranges for interfaces from a system-wide pool of
14204 addresses. This is useful for dynamically managing a large
14205 number of interfaces with a single network configuration
14206 file. In particular this is useful to easily assign
14207 appropriate IP addresses to the veth links of a large number
14208 of nspawn instances.
14210 * RPM macros for processing sysusers, sysctl and binfmt
14211 drop-in snippets at package installation time have been
14214 * The /etc/os-release file should now be placed in
14215 /usr/lib/os-release. The old location is automatically
14216 created as symlink. /usr/lib is the more appropriate
14217 location of this file, since it shall actually describe the
14218 vendor operating system shipped in /usr, and not the
14219 configuration stored in /etc.
14221 * .mount units gained a new boolean SloppyOptions= setting
14222 that maps to mount(8)'s -s option which enables permissive
14223 parsing of unknown mount options.
14225 * tmpfiles learnt a new "L+" directive which creates a symlink
14226 but (unlike "L") deletes a pre-existing file first, should
14227 it already exist and not already be the correct
14228 symlink. Similarly, "b+", "c+" and "p+" directives have been
14229 added as well, which create block and character devices, as
14230 well as fifos in the filesystem, possibly removing any
14231 pre-existing files of different types.
14233 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
14234 'argument' field (which so far specified the source to
14235 symlink/copy the files from) is now optional. If omitted the
14236 same file os copied from /usr/share/factory/ suffixed by the
14237 full destination path. This is useful for populating /etc
14238 with essential files, by copying them from vendor defaults
14239 shipped in /usr/share/factory/etc.
14241 * A new command "systemctl preset-all" has been added that
14242 applies the service preset settings to all installed unit
14243 files. A new switch --preset-mode= has been added that
14244 controls whether only enable or only disable operations
14247 * A new command "systemctl is-system-running" has been added
14248 that allows checking the overall state of the system, for
14249 example whether it is fully up and running.
14251 * When the system boots up with an empty /etc, the equivalent
14252 to "systemctl preset-all" is executed during early boot, to
14253 make sure all default services are enabled after a factory
14256 * systemd now contains a minimal preset file that enables the
14257 most basic services systemd ships by default.
14259 * Unit files' [Install] section gained a new DefaultInstance=
14260 field for defining the default instance to create if a
14261 template unit is enabled with no instance specified.
14263 * A new passive target cryptsetup-pre.target has been added
14264 that may be used by services that need to make they run and
14265 finish before the first LUKS cryptographic device is set up.
14267 * The /dev/loop-control and /dev/btrfs-control device nodes
14268 are now owned by the "disk" group by default, opening up
14269 access to this group.
14271 * systemd-coredump will now automatically generate a
14272 stack trace of all core dumps taking place on the system,
14273 based on elfutils' libdw library. This stack trace is logged
14276 * systemd-coredump may now optionally store coredumps directly
14277 on disk (in /var/lib/systemd/coredump, possibly compressed),
14278 instead of storing them unconditionally in the journal. This
14279 mode is the new default. A new configuration file
14280 /etc/systemd/coredump.conf has been added to configure this
14281 and other parameters of systemd-coredump.
14283 * coredumpctl gained a new "info" verb to show details about a
14284 specific coredump. A new switch "-1" has also been added
14285 that makes sure to only show information about the most
14286 recent entry instead of all entries. Also, as the tool is
14287 generally useful now the "systemd-" prefix of the binary
14288 name has been removed. Distributions that want to maintain
14289 compatibility with the old name should add a symlink from
14290 the old name to the new name.
14292 * journald's SplitMode= now defaults to "uid". This makes sure
14293 that unprivileged users can access their own coredumps with
14294 coredumpctl without restrictions.
14296 * New kernel command line options "systemd.wants=" (for
14297 pulling an additional unit during boot), "systemd.mask="
14298 (for masking a specific unit for the boot), and
14299 "systemd.debug-shell" (for enabling the debug shell on tty9)
14300 have been added. This is implemented in the new generator
14301 "systemd-debug-generator".
14303 * systemd-nspawn will now by default filter a couple of
14304 syscalls for containers, among them those required for
14305 kernel module loading, direct x86 IO port access, swap
14306 management, and kexec. Most importantly though
14307 open_by_handle_at() is now prohibited for containers,
14308 closing a hole similar to a recently discussed vulnerability
14309 in docker regarding access to files on file hierarchies the
14310 container should normally not have access to. Note that, for
14311 nspawn, we generally make no security claims anyway (and
14312 this is explicitly documented in the man page), so this is
14313 just a fix for one of the most obvious problems.
14315 * A new man page file-hierarchy(7) has been added that
14316 contains a minimized, modernized version of the file system
14317 layout systemd expects, similar in style to the FHS
14318 specification or hier(5). A new tool systemd-path(1) has
14319 been added to query many of these paths for the local
14322 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
14323 longer done. Since the directory now has a per-user size
14324 limit, and is cleaned on logout this appears unnecessary,
14325 in particular since this now brings the lifecycle of this
14326 directory closer in line with how IPC objects are handled.
14328 * systemd.pc now exports a number of additional directories,
14329 including $libdir (which is useful to identify the library
14330 path for the primary architecture of the system), and a
14331 couple of drop-in directories.
14333 * udev's predictable network interface names now use the dev_port
14334 sysfs attribute, introduced in linux 3.15 instead of dev_id to
14335 distinguish between ports of the same PCI function. dev_id should
14336 only be used for ports using the same HW address, hence the need
14339 * machined has been updated to export the OS version of a
14340 container (read from /etc/os-release and
14341 /usr/lib/os-release) on the bus. This is now shown in
14342 "machinectl status" for a machine.
14344 * A new service setting RestartForceExitStatus= has been
14345 added. If configured to a set of exit signals or process
14346 return values, the service will be restarted when the main
14347 daemon process exits with any of them, regardless of the
14350 * systemctl's -H switch for connecting to remote systemd
14351 machines has been extended so that it may be used to
14352 directly connect to a specific container on the
14353 host. "systemctl -H root@foobar:waldi" will now connect as
14354 user "root" to host "foobar", and then proceed directly to
14355 the container named "waldi". Note that currently you have to
14356 authenticate as user "root" for this to work, as entering
14357 containers is a privileged operation.
14359 Contributions from: Andreas Henriksson, Benjamin Steinwender,
14360 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
14361 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
14362 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
14363 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
14364 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
14365 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
14366 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
14367 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
14368 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
14369 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
14370 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
14372 — Berlin, 2014-07-03
14376 * As an experimental feature, udev now tries to lock the
14377 disk device node (flock(LOCK_SH|LOCK_NB)) while it
14378 executes events for the disk or any of its partitions.
14379 Applications like partitioning programs can lock the
14380 disk device node (flock(LOCK_EX)) and claim temporary
14381 device ownership that way; udev will entirely skip all event
14382 handling for this disk and its partitions. If the disk
14383 was opened for writing, the close will trigger a partition
14384 table rescan in udev's "watch" facility, and if needed
14385 synthesize "change" events for the disk and all its partitions.
14386 This is now unconditionally enabled, and if it turns out to
14387 cause major problems, we might turn it on only for specific
14388 devices, or might need to disable it entirely. Device Mapper
14389 devices are excluded from this logic.
14391 * We temporarily dropped the "-l" switch for fsck invocations,
14392 since they collide with the flock() logic above. util-linux
14393 upstream has been changed already to avoid this conflict,
14394 and we will re-add "-l" as soon as util-linux with this
14395 change has been released.
14397 * The dependency on libattr has been removed. Since a long
14398 time, the extended attribute calls have moved to glibc, and
14399 libattr is thus unnecessary.
14401 * Virtualization detection works without privileges now. This
14402 means the systemd-detect-virt binary no longer requires
14403 CAP_SYS_PTRACE file capabilities, and our daemons can run
14404 with fewer privileges.
14406 * systemd-networkd now runs under its own "systemd-network"
14407 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
14408 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
14409 loses the ability to write to files owned by root this way.
14411 * Similarly, systemd-resolved now runs under its own
14412 "systemd-resolve" user with no capabilities remaining.
14414 * Similarly, systemd-bus-proxyd now runs under its own
14415 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
14417 * systemd-networkd gained support for setting up "veth"
14418 virtual Ethernet devices for container connectivity, as well
14419 as GRE and VTI tunnels.
14421 * systemd-networkd will no longer automatically attempt to
14422 manually load kernel modules necessary for certain tunnel
14423 transports. Instead, it is assumed the kernel loads them
14424 automatically when required. This only works correctly on
14425 very new kernels. On older kernels, please consider adding
14426 the kernel modules to /etc/modules-load.d/ as a work-around.
14428 * The resolv.conf file systemd-resolved generates has been
14429 moved to /run/systemd/resolve/. If you have a symlink from
14430 /etc/resolv.conf, it might be necessary to correct it.
14432 * Two new service settings, ProtectHome= and ProtectSystem=,
14433 have been added. When enabled, they will make the user data
14434 (such as /home) inaccessible or read-only and the system
14435 (such as /usr) read-only, for specific services. This allows
14436 very light-weight per-service sandboxing to avoid
14437 modifications of user data or system files from
14438 services. These two new switches have been enabled for all
14439 of systemd's long-running services, where appropriate.
14441 * Socket units gained new SocketUser= and SocketGroup=
14442 settings to set the owner user and group of AF_UNIX sockets
14443 and FIFOs in the file system.
14445 * Socket units gained a new RemoveOnStop= setting. If enabled,
14446 all FIFOS and sockets in the file system will be removed
14447 when the specific socket unit is stopped.
14449 * Socket units gained a new Symlinks= setting. It takes a list
14450 of symlinks to create to file system sockets or FIFOs
14451 created by the specific Unix sockets. This is useful to
14452 manage symlinks to socket nodes with the same lifecycle as
14455 * The /dev/log socket and /dev/initctl FIFO have been moved to
14456 /run, and have been replaced by symlinks. This allows
14457 connecting to these facilities even if PrivateDevices=yes is
14458 used for a service (which makes /dev/log itself unavailable,
14459 but /run is left). This also has the benefit of ensuring
14460 that /dev only contains device nodes, directories and
14461 symlinks, and nothing else.
14463 * sd-daemon gained two new calls sd_pid_notify() and
14464 sd_pid_notifyf(). They are similar to sd_notify() and
14465 sd_notifyf(), but allow overriding of the source PID of
14466 notification messages if permissions permit this. This is
14467 useful to send notify messages on behalf of a different
14468 process (for example, the parent process). The
14469 systemd-notify tool has been updated to make use of this
14470 when sending messages (so that notification messages now
14471 originate from the shell script invoking systemd-notify and
14472 not the systemd-notify process itself. This should minimize
14473 a race where systemd fails to associate notification
14474 messages to services when the originating process already
14477 * A new "on-abnormal" setting for Restart= has been added. If
14478 set, it will result in automatic restarts on all "abnormal"
14479 reasons for a process to exit, which includes unclean
14480 signals, core dumps, timeouts and watchdog timeouts, but
14481 does not include clean and unclean exit codes or clean
14482 signals. Restart=on-abnormal is an alternative for
14483 Restart=on-failure for services that shall be able to
14484 terminate and avoid restarts on certain errors, by
14485 indicating so with an unclean exit code. Restart=on-failure
14486 or Restart=on-abnormal is now the recommended setting for
14487 all long-running services.
14489 * If the InaccessibleDirectories= service setting points to a
14490 mount point (or if there are any submounts contained within
14491 it), it is now attempted to completely unmount it, to make
14492 the file systems truly unavailable for the respective
14495 * The ReadOnlyDirectories= service setting and
14496 systemd-nspawn's --read-only parameter are now recursively
14497 applied to all submounts, too.
14499 * Mount units may now be created transiently via the bus APIs.
14501 * The support for SysV and LSB init scripts has been removed
14502 from the systemd daemon itself. Instead, it is now
14503 implemented as a generator that creates native systemd units
14504 from these scripts when needed. This enables us to remove a
14505 substantial amount of legacy code from PID 1, following the
14506 fact that many distributions only ship a very small number
14507 of LSB/SysV init scripts nowadays.
14509 * Privileged Xen (dom0) domains are not considered
14510 virtualization anymore by the virtualization detection
14511 logic. After all, they generally have unrestricted access to
14512 the hardware and usually are used to manage the unprivileged
14515 * systemd-tmpfiles gained a new "C" line type, for copying
14516 files or entire directories.
14518 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
14519 lines. So far, they have been non-globbing versions of the
14520 latter, and have thus been redundant. In future, it is
14521 recommended to only use "z". "m" has hence been removed
14522 from the documentation, even though it stays supported.
14524 * A tmpfiles snippet to recreate the most basic structure in
14525 /var has been added. This is enough to create the /var/run →
14526 /run symlink and create a couple of structural
14527 directories. This allows systems to boot up with an empty or
14528 volatile /var. Of course, while with this change, the core OS
14529 now is capable with dealing with a volatile /var, not all
14530 user services are ready for it. However, we hope that sooner
14531 or later, many service daemons will be changed upstream so
14532 that they are able to automatically create their necessary
14533 directories in /var at boot, should they be missing. This is
14534 the first step to allow state-less systems that only require
14535 the vendor image for /usr to boot.
14537 * systemd-nspawn has gained a new --tmpfs= switch to mount an
14538 empty tmpfs instance to a specific directory. This is
14539 particularly useful for making use of the automatic
14540 reconstruction of /var (see above), by passing --tmpfs=/var.
14542 * Access modes specified in tmpfiles snippets may now be
14543 prefixed with "~", which indicates that they shall be masked
14544 by whether the existing file or directory is currently
14545 writable, readable or executable at all. Also, if specified,
14546 the sgid/suid/sticky bits will be masked for all
14549 * A new passive target unit "network-pre.target" has been
14550 added which is useful for services that shall run before any
14551 network is configured, for example firewall scripts.
14553 * The "floppy" group that previously owned the /dev/fd*
14554 devices is no longer used. The "disk" group is now used
14555 instead. Distributions should probably deprecate usage of
14558 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
14559 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
14560 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
14561 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
14562 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
14563 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
14566 — Berlin, 2014-06-11
14570 * A new "systemd-timesyncd" daemon has been added for
14571 synchronizing the system clock across the network. It
14572 implements an SNTP client. In contrast to NTP
14573 implementations such as chrony or the NTP reference server,
14574 this only implements a client side, and does not bother with
14575 the full NTP complexity, focusing only on querying time from
14576 one remote server and synchronizing the local clock to
14577 it. Unless you intend to serve NTP to networked clients or
14578 want to connect to local hardware clocks, this simple NTP
14579 client should be more than appropriate for most
14580 installations. The daemon runs with minimal privileges, and
14581 has been hooked up with networkd to only operate when
14582 network connectivity is available. The daemon saves the
14583 current clock to disk every time a new NTP sync has been
14584 acquired, and uses this to possibly correct the system clock
14585 early at bootup, in order to accommodate for systems that
14586 lack an RTC such as the Raspberry Pi and embedded devices,
14587 and to make sure that time monotonically progresses on these
14588 systems, even if it is not always correct. To make use of
14589 this daemon, a new system user and group "systemd-timesync"
14590 needs to be created on installation of systemd.
14592 * The queue "seqnum" interface of libudev has been disabled, as
14593 it was generally incompatible with device namespacing as
14594 sequence numbers of devices go "missing" if the devices are
14595 part of a different namespace.
14597 * "systemctl list-timers" and "systemctl list-sockets" gained
14598 a --recursive switch for showing units of these types also
14599 for all local containers, similar in style to the already
14600 supported --recursive switch for "systemctl list-units".
14602 * A new RebootArgument= setting has been added for service
14603 units, which may be used to specify a kernel reboot argument
14604 to use when triggering reboots with StartLimitAction=.
14606 * A new FailureAction= setting has been added for service
14607 units which may be used to specify an operation to trigger
14608 when a service fails. This works similarly to
14609 StartLimitAction=, but unlike it, controls what is done
14610 immediately rather than only after several attempts to
14611 restart the service in question.
14613 * hostnamed got updated to also expose the kernel name,
14614 release, and version on the bus. This is useful for
14615 executing commands like hostnamectl with the -H switch.
14616 systemd-analyze makes use of this to properly display
14617 details when running non-locally.
14619 * The bootchart tool can now show cgroup information in the
14620 graphs it generates.
14622 * The CFS CPU quota cgroup attribute is now exposed for
14623 services. The new CPUQuota= switch has been added for this
14624 which takes a percentage value. Setting this will have the
14625 result that a service may never get more CPU time than the
14626 specified percentage, even if the machine is otherwise idle.
14628 * systemd-networkd learned IPIP and SIT tunnel support.
14630 * LSB init scripts exposing a dependency on $network will now
14631 get a dependency on network-online.target rather than simply
14632 network.target. This should bring LSB handling closer to
14633 what it was on SysV systems.
14635 * A new fsck.repair= kernel option has been added to control
14636 how fsck shall deal with unclean file systems at boot.
14638 * The (.ini) configuration file parser will now silently ignore
14639 sections whose names begin with "X-". This may be used to maintain
14640 application-specific extension sections in unit files.
14642 * machined gained a new API to query the IP addresses of
14643 registered containers. "machinectl status" has been updated
14644 to show these addresses in its output.
14646 * A new call sd_uid_get_display() has been added to the
14647 sd-login APIs for querying the "primary" session of a
14648 user. The "primary" session of the user is elected from the
14649 user's sessions and generally a graphical session is
14650 preferred over a text one.
14652 * A minimal systemd-resolved daemon has been added. It
14653 currently simply acts as a companion to systemd-networkd and
14654 manages resolv.conf based on per-interface DNS
14655 configuration, possibly supplied via DHCP. In the long run
14656 we hope to extend this into a local DNSSEC enabled DNS and
14659 * The systemd-networkd-wait-online tool is now enabled by
14660 default. It will delay network-online.target until a network
14661 connection has been configured. The tool primarily integrates
14662 with networkd, but will also make a best effort to make sense
14663 of network configuration performed in some other way.
14665 * Two new service options StartupCPUShares= and
14666 StartupBlockIOWeight= have been added that work similarly to
14667 CPUShares= and BlockIOWeight= however only apply during
14668 system startup. This is useful to prioritize certain services
14669 differently during bootup than during normal runtime.
14671 * hostnamed has been changed to prefer the statically
14672 configured hostname in /etc/hostname (unless set to
14673 'localhost' or empty) over any dynamic one supplied by
14674 dhcp. With this change, the rules for picking the hostname
14675 match more closely the rules of other configuration settings
14676 where the local administrator's configuration in /etc always
14677 overrides any other settings.
14679 Contributions from: Ali H. Caliskan, Alison Chaiken, Bas van
14680 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
14681 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
14682 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
14683 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
14684 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
14685 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
14686 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
14687 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
14688 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
14689 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
14690 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
14691 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
14692 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
14693 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
14694 Lindskog, WaLyong Cho, Will Woods, Zbigniew
14697 — Beijing, 2014-05-28
14701 * When restoring the screen brightness at boot, stay away from
14702 the darkest setting or from the lowest 5% of the available
14703 range, depending on which is the larger value of both. This
14704 should effectively protect the user from rebooting into a
14705 black screen, should the brightness have been set to minimum
14708 * sd-login gained a new sd_machine_get_class() call to
14709 determine the class ("vm" or "container") of a machine
14710 registered with machined.
14712 * sd-login gained new calls
14713 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
14714 to query the identity of the peer of a local AF_UNIX
14715 connection. They operate similarly to their sd_pid_get_xyz()
14718 * PID 1 will now maintain a system-wide system state engine
14719 with the states "starting", "running", "degraded",
14720 "maintenance", "stopping". These states are bound to system
14721 startup, normal runtime, runtime with at least one failed
14722 service, rescue/emergency mode and system shutdown. This
14723 state is shown in the "systemctl status" output when no unit
14724 name is passed. It is useful to determine system state, in
14725 particularly when doing so for many systems or containers at
14728 * A new command "list-machines" has been added to "systemctl"
14729 that lists all local OS containers and shows their system
14730 state (see above), if systemd runs inside of them.
14732 * systemctl gained a new "-r" switch to recursively enumerate
14733 units on all local containers, when used with the
14734 "list-unit" command (which is the default one that is
14735 executed when no parameters are specified).
14737 * The GPT automatic partition discovery logic will now honour
14738 two GPT partition flags: one may be set on a partition to
14739 cause it to be mounted read-only, and the other may be set
14740 on a partition to ignore it during automatic discovery.
14742 * Two new GPT type UUIDs have been added for automatic root
14743 partition discovery, for 32-bit and 64-bit ARM. This is not
14744 particularly useful for discovering the root directory on
14745 these architectures during bare-metal boots (since UEFI is
14746 not common there), but still very useful to allow booting of
14747 ARM disk images in nspawn with the -i option.
14749 * MAC addresses of interfaces created with nspawn's
14750 --network-interface= switch will now be generated from the
14751 machine name, and thus be stable between multiple invocations
14754 * logind will now automatically remove all IPC objects owned
14755 by a user if she or he fully logs out. This makes sure that
14756 users who are logged out cannot continue to consume IPC
14757 resources. This covers SysV memory, semaphores and message
14758 queues as well as POSIX shared memory and message
14759 queues. Traditionally, SysV and POSIX IPC had no lifecycle
14760 limits. With this functionality, that is corrected. This may
14761 be turned off by using the RemoveIPC= switch of logind.conf.
14763 * The systemd-machine-id-setup and tmpfiles tools gained a
14764 --root= switch to operate on a specific root directory,
14767 * journald can now forward logged messages to the TTYs of all
14768 logged in users ("wall"). This is the default for all
14769 emergency messages now.
14771 * A new tool systemd-journal-remote has been added to stream
14772 journal log messages across the network.
14774 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
14775 controller trees are mounted into it. Note that the
14776 directories mounted beneath it are not read-only. This is a
14777 security measure and is particularly useful because glibc
14778 actually includes a search logic to pick any tmpfs it can
14779 find to implement shm_open() if /dev/shm is not available
14780 (which it might very well be in namespaced setups).
14782 * machinectl gained a new "poweroff" command to cleanly power
14783 down a local OS container.
14785 * The PrivateDevices= unit file setting will now also drop the
14786 CAP_MKNOD capability from the capability bound set, and
14787 imply DevicePolicy=closed.
14789 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
14790 comprehensively on all long-running systemd services where
14791 this is appropriate.
14793 * systemd-udevd will now run in a disassociated mount
14794 namespace. To mount directories from udev rules, make sure to
14795 pull in mount units via SYSTEMD_WANTS properties.
14797 * The kdbus support gained support for uploading policy into
14798 the kernel. sd-bus gained support for creating "monitoring"
14799 connections that can eavesdrop into all bus communication
14800 for debugging purposes.
14802 * Timestamps may now be specified in seconds since the UNIX
14803 epoch Jan 1st, 1970 by specifying "@" followed by the value
14806 * Native tcpwrap support in systemd has been removed. tcpwrap
14807 is old code, not really maintained anymore and has serious
14808 shortcomings, and better options such as firewalls
14809 exist. For setups that require tcpwrap usage, please
14810 consider invoking your socket-activated service via tcpd,
14811 like on traditional inetd.
14813 * A new system.conf configuration option
14814 DefaultTimerAccuracySec= has been added that controls the
14815 default AccuracySec= setting of .timer units.
14817 * Timer units gained a new WakeSystem= switch. If enabled,
14818 timers configured this way will cause the system to resume
14819 from system suspend (if the system supports that, which most
14822 * Timer units gained a new Persistent= switch. If enabled,
14823 timers configured this way will save to disk when they have
14824 been last triggered. This information is then used on next
14825 reboot to possible execute overdue timer events, that
14826 could not take place because the system was powered off.
14827 This enables simple anacron-like behaviour for timer units.
14829 * systemctl's "list-timers" will now also list the time a
14830 timer unit was last triggered in addition to the next time
14831 it will be triggered.
14833 * systemd-networkd will now assign predictable IPv4LL
14834 addresses to its local interfaces.
14836 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
14837 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
14838 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
14839 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
14840 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
14841 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
14842 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
14843 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
14846 — Berlin, 2014-03-25
14850 * A new unit file setting RestrictAddressFamilies= has been
14851 added to restrict which socket address families unit
14852 processes gain access to. This takes address family names
14853 like "AF_INET" or "AF_UNIX", and is useful to minimize the
14854 attack surface of services via exotic protocol stacks. This
14855 is built on seccomp system call filters.
14857 * Two new unit file settings RuntimeDirectory= and
14858 RuntimeDirectoryMode= have been added that may be used to
14859 manage a per-daemon runtime directories below /run. This is
14860 an alternative for setting up directory permissions with
14861 tmpfiles snippets, and has the advantage that the runtime
14862 directory's lifetime is bound to the daemon runtime and that
14863 the daemon starts up with an empty directory each time. This
14864 is particularly useful when writing services that drop
14865 privileges using the User= or Group= setting.
14867 * The DeviceAllow= unit setting now supports globbing for
14868 matching against device group names.
14870 * The systemd configuration file system.conf gained new
14871 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
14872 DefaultMemoryAccounting= to globally turn on/off accounting
14873 for specific resources (cgroups) for all units. These
14874 settings may still be overridden individually in each unit
14877 * systemd-gpt-auto-generator is now able to discover /srv and
14878 root partitions in addition to /home and swap partitions. It
14879 also supports LUKS-encrypted partitions now. With this in
14880 place, automatic discovery of partitions to mount following
14881 the Discoverable Partitions Specification
14882 (https://systemd.io/DISCOVERABLE_PARTITIONS/)
14883 is now a lot more complete. This allows booting without
14884 /etc/fstab and without root= on the kernel command line on
14885 systems prepared appropriately.
14887 * systemd-nspawn gained a new --image= switch which allows
14888 booting up disk images and Linux installations on any block
14889 device that follow the Discoverable Partitions Specification
14890 (see above). This means that installations made with
14891 appropriately updated installers may now be started and
14892 deployed using container managers, completely
14893 unmodified. (We hope that libvirt-lxc will add support for
14894 this feature soon, too.)
14896 * systemd-nspawn gained a new --network-macvlan= setting to
14897 set up a private macvlan interface for the
14898 container. Similarly, systemd-networkd gained a new
14899 Kind=macvlan setting in .netdev files.
14901 * systemd-networkd now supports configuring local addresses
14904 * A new tool systemd-network-wait-online has been added to
14905 synchronously wait for network connectivity using
14908 * The sd-bus.h bus API gained a new sd_bus_track object for
14909 tracking the lifecycle of bus peers. Note that sd-bus.h is
14910 still not a public API though (unless you specify
14911 --enable-kdbus on the configure command line, which however
14912 voids your warranty and you get no API stability guarantee).
14914 * The $XDG_RUNTIME_DIR runtime directories for each user are
14915 now individual tmpfs instances, which has the benefit of
14916 introducing separate pools for each user, with individual
14917 size limits, and thus making sure that unprivileged clients
14918 can no longer negatively impact the system or other users by
14919 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
14920 RuntimeDirectorySize= has been introduced that allows
14921 controlling the default size limit for all users. It
14922 defaults to 10% of the available physical memory. This is no
14923 replacement for quotas on tmpfs though (which the kernel
14924 still does not support), as /dev/shm and /tmp are still
14925 shared resources used by both the system and unprivileged
14928 * logind will now automatically turn off automatic suspending
14929 on laptop lid close when more than one display is
14930 connected. This was previously expected to be implemented
14931 individually in desktop environments (such as GNOME),
14932 however has been added to logind now, in order to fix a
14933 boot-time race where a desktop environment might not have
14934 been started yet and thus not been able to take an inhibitor
14935 lock at the time where logind already suspends the system
14936 due to a closed lid.
14938 * logind will now wait at least 30s after each system
14939 suspend/resume cycle, and 3min after system boot before
14940 suspending the system due to a closed laptop lid. This
14941 should give USB docking stations and similar enough time to
14942 be probed and configured after system resume and boot in
14943 order to then act as suspend blocker.
14945 * systemd-run gained a new --property= setting which allows
14946 initialization of resource control properties (and others)
14947 for the created scope or service unit. Example: "systemd-run
14948 --property=BlockIOWeight=10 updatedb" may be used to run
14949 updatedb at a low block IO scheduling weight.
14951 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
14952 now also work in --scope mode.
14954 * When systemd is compiled with kdbus support, basic support
14955 for enforced policies is now in place. (Note that enabling
14956 kdbus still voids your warranty and no API compatibility
14957 promises are made.)
14959 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
14960 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
14961 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
14962 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
14963 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
14964 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
14965 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
14966 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
14967 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
14968 Zbigniew Jędrzejewski-Szmek
14970 — Berlin, 2014-03-12
14974 * systemd will now relabel /dev after loading the SMACK policy
14975 according to SMACK rules.
14977 * A new unit file option AppArmorProfile= has been added to
14978 set the AppArmor profile for the processes of a unit.
14980 * A new condition check ConditionArchitecture= has been added
14981 to conditionalize units based on the system architecture, as
14982 reported by uname()'s "machine" field.
14984 * systemd-networkd now supports matching on the system
14985 virtualization, architecture, kernel command line, hostname
14988 * logind is now a lot more aggressive when suspending the
14989 machine due to a closed laptop lid. Instead of acting only
14990 on the lid close action, it will continuously watch the lid
14991 status and act on it. This is useful for laptops where the
14992 power button is on the outside of the chassis so that it can
14993 be reached without opening the lid (such as the Lenovo
14994 Yoga). On those machines, logind will now immediately
14995 re-suspend the machine if the power button has been
14996 accidentally pressed while the laptop was suspended and in a
14997 backpack or similar.
14999 * logind will now watch SW_DOCK switches and inhibit reaction
15000 to the lid switch if it is pressed. This means that logind
15001 will not suspend the machine anymore if the lid is closed
15002 and the system is docked, if the laptop supports SW_DOCK
15003 notifications via the input layer. Note that ACPI docking
15004 stations do not generate this currently. Also note that this
15005 logic is usually not fully sufficient and Desktop
15006 Environments should take a lid switch inhibitor lock when an
15007 external display is connected, as systemd will not watch
15010 * nspawn will now make use of the devices cgroup controller by
15011 default, and only permit creation of and access to the usual
15012 API device nodes like /dev/null or /dev/random, as well as
15013 access to (but not creation of) the pty devices.
15015 * We will now ship a default .network file for
15016 systemd-networkd that automatically configures DHCP for
15017 network interfaces created by nspawn's --network-veth or
15018 --network-bridge= switches.
15020 * systemd will now understand the usual M, K, G, T suffixes
15021 according to SI conventions (i.e. to the base 1000) when
15022 referring to throughput and hardware metrics. It will stay
15023 with IEC conventions (i.e. to the base 1024) for software
15024 metrics, according to what is customary according to
15025 Wikipedia. We explicitly document which base applies for
15026 each configuration option.
15028 * The DeviceAllow= setting in unit files now supports a syntax to
15029 allow-list an entire group of devices node majors at once, based on
15030 the /proc/devices listing. For example, with the string "char-pts",
15031 it is now possible to allow-list all current and future pseudo-TTYs
15034 * sd-event learned a new "post" event source. Event sources of
15035 this type are triggered by the dispatching of any event
15036 source of a type that is not "post". This is useful for
15037 implementing clean-up and check event sources that are
15038 triggered by other work being done in the program.
15040 * systemd-networkd is no longer statically enabled, but uses
15041 the usual [Install] sections so that it can be
15042 enabled/disabled using systemctl. It still is enabled by
15045 * When creating a veth interface pair with systemd-nspawn, the
15046 host side will now be prefixed with "vb-" if
15047 --network-bridge= is used, and with "ve-" if --network-veth
15048 is used. This way, it is easy to distinguish these cases on
15049 the host, for example to apply different configuration to
15050 them with systemd-networkd.
15052 * The compatibility libraries for libsystemd-journal.so,
15053 libsystem-id128.so, libsystemd-login.so and
15054 libsystemd-daemon.so do not make use of IFUNC
15055 anymore. Instead, we now build libsystemd.so multiple times
15056 under these alternative names. This means that the footprint
15057 is drastically increased, but given that these are
15058 transitional compatibility libraries, this should not matter
15059 much. This change has been made necessary to support the ARM
15060 platform for these compatibility libraries, as the ARM
15061 toolchain is not really at the same level as the toolchain
15062 for other architectures like x86 and does not support
15063 IFUNC. Please make sure to use --enable-compat-libs only
15064 during a transitional period!
15066 * The .include syntax has been deprecated and is not documented
15067 anymore. Drop-in files in .d directories should be used instead.
15069 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
15070 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
15071 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
15072 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
15073 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
15074 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
15075 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
15076 Zbigniew Jędrzejewski-Szmek
15078 — Berlin, 2014-02-24
15082 * A new component "systemd-networkd" has been added that can
15083 be used to configure local network interfaces statically or
15084 via DHCP. It is capable of bringing up bridges, VLANs, and
15085 bonding. Currently, no hook-ups for interactive network
15086 configuration are provided. Use this for your initrd,
15087 container, embedded, or server setup if you need a simple,
15088 yet powerful, network configuration solution. This
15089 configuration subsystem is quite nifty, as it allows wildcard
15090 hotplug matching in interfaces. For example, with a single
15091 configuration snippet, you can configure that all Ethernet
15092 interfaces showing up are automatically added to a bridge,
15093 or similar. It supports link-sensing and more.
15095 * A new tool "systemd-socket-proxyd" has been added which can
15096 act as a bidirectional proxy for TCP sockets. This is
15097 useful for adding socket activation support to services that
15098 do not actually support socket activation, including virtual
15099 machines and the like.
15101 * Add a new tool to save/restore rfkill state on
15104 * Save/restore state of keyboard backlights in addition to
15105 display backlights on shutdown/boot.
15107 * udev learned a new SECLABEL{} construct to label device
15108 nodes with a specific security label when they appear. For
15109 now, only SECLABEL{selinux} is supported, but the syntax is
15110 prepared for additional security frameworks.
15112 * udev gained a new scheme to configure link-level attributes
15113 from files in /etc/systemd/network/*.link. These files can
15114 match against MAC address, device path, driver name and type,
15115 and will apply attributes like the naming policy, link speed,
15116 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
15117 address assignment policy (randomized, …).
15119 * The configuration of network interface naming rules for
15120 "permanent interface names" has changed: a new NamePolicy=
15121 setting in the [Link] section of .link files determines the
15122 priority of possible naming schemes (onboard, slot, MAC,
15123 path). The default value of this setting is determined by
15124 /usr/lib/net/links/99-default.link. Old
15125 80-net-name-slot.rules udev configuration file has been
15126 removed, so local configuration overriding this file should
15127 be adapted to override 99-default.link instead.
15129 * When the User= switch is used in a unit file, also
15130 initialize $SHELL= based on the user database entry.
15132 * systemd no longer depends on libdbus. All communication is
15133 now done with sd-bus, systemd's low-level bus library
15136 * kdbus support has been added to PID 1 itself. When kdbus is
15137 enabled, this causes PID 1 to set up the system bus and
15138 enable support for a new ".busname" unit type that
15139 encapsulates bus name activation on kdbus. It works a little
15140 bit like ".socket" units, except for bus names. A new
15141 generator has been added that converts classic dbus1 service
15142 activation files automatically into native systemd .busname
15143 and .service units.
15145 * sd-bus: add a light-weight vtable implementation that allows
15146 defining objects on the bus with a simple static const
15147 vtable array of its methods, signals and properties.
15149 * systemd will not generate or install static dbus
15150 introspection data anymore to /usr/share/dbus-1/interfaces,
15151 as the precise format of these files is unclear, and
15152 nothing makes use of it.
15154 * A proxy daemon is now provided to proxy clients connecting
15155 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
15156 compatibility with classic D-Bus.
15158 * A bus driver implementation has been added that supports the
15159 classic D-Bus bus driver calls on kdbus, also for
15160 compatibility purposes.
15162 * A new API "sd-event.h" has been added that implements a
15163 minimal event loop API built around epoll. It provides a
15164 couple of features that direct epoll usage is lacking:
15165 prioritization of events, scales to large numbers of timer
15166 events, per-event timer slack (accuracy), system-wide
15167 coalescing of timer events, exit handlers, watchdog
15168 supervision support using systemd's sd_notify() API, child
15171 * A new API "sd-rntl.h" has been added that provides an API
15172 around the route netlink interface of the kernel, similar in
15173 style to "sd-bus.h".
15175 * A new API "sd-dhcp-client.h" has been added that provides a
15176 small DHCPv4 client-side implementation. This is used by
15177 "systemd-networkd".
15179 * There is a new kernel command line option
15180 "systemd.restore_state=0|1". When set to "0", none of the
15181 systemd tools will restore saved runtime state to hardware
15182 devices. More specifically, the rfkill and backlight states
15185 * The FsckPassNo= compatibility option in mount/service units
15186 has been removed. The fstab generator will now add the
15187 necessary dependencies automatically, and does not require
15188 PID1's support for that anymore.
15190 * journalctl gained a new switch, --list-boots, that lists
15191 recent boots with their times and boot IDs.
15193 * The various tools like systemctl, loginctl, timedatectl,
15194 busctl, systemd-run, … have gained a new switch "-M" to
15195 connect to a specific, local OS container (as direct
15196 connection, without requiring SSH). This works on any
15197 container that is registered with machined, such as those
15198 created by libvirt-lxc or nspawn.
15200 * systemd-run and systemd-analyze also gained support for "-H"
15201 to connect to remote hosts via SSH. This is particularly
15202 useful for systemd-run because it enables queuing of jobs
15203 onto remote systems.
15205 * machinectl gained a new command "login" to open a getty
15206 login in any local container. This works with any container
15207 that is registered with machined (such as those created by
15208 libvirt-lxc or nspawn), and which runs systemd inside.
15210 * machinectl gained a new "reboot" command that may be used to
15211 trigger a reboot on a specific container that is registered
15212 with machined. This works on any container that runs an init
15213 system of some kind.
15215 * systemctl gained a new "list-timers" command to print a nice
15216 listing of installed timer units with the times they elapse
15219 * Alternative reboot() parameters may now be specified on the
15220 "systemctl reboot" command line and are passed to the
15221 reboot() system call.
15223 * systemctl gained a new --job-mode= switch to configure the
15224 mode to queue a job with. This is a more generic version of
15225 --fail, --irreversible, and --ignore-dependencies, which are
15226 still available but not advertised anymore.
15228 * /etc/systemd/system.conf gained new settings to configure
15229 various default timeouts of units, as well as the default
15230 start limit interval and burst. These may still be overridden
15233 * PID1 will now export on the bus profile data of the security
15234 policy upload process (such as the SELinux policy upload to
15237 * journald: when forwarding logs to the console, include
15238 timestamps (following the setting in
15239 /sys/module/printk/parameters/time).
15241 * OnCalendar= in timer units now understands the special
15242 strings "yearly" and "annually". (Both are equivalent)
15244 * The accuracy of timer units is now configurable with the new
15245 AccuracySec= setting. It defaults to 1min.
15247 * A new dependency type JoinsNamespaceOf= has been added that
15248 allows running two services within the same /tmp and network
15249 namespace, if PrivateNetwork= or PrivateTmp= are used.
15251 * A new command "cat" has been added to systemctl. It outputs
15252 the original unit file of a unit, and concatenates the
15253 contents of additional "drop-in" unit file snippets, so that
15254 the full configuration is shown.
15256 * systemctl now supports globbing on the various "list-xyz"
15257 commands, like "list-units" or "list-sockets", as well as on
15258 those commands which take multiple unit names.
15260 * journalctl's --unit= switch gained support for globbing.
15262 * All systemd daemons now make use of the watchdog logic so
15263 that systemd automatically notices when they hang.
15265 * If the $container_ttys environment variable is set,
15266 getty-generator will automatically spawn a getty for each
15267 listed tty. This is useful for container managers to request
15268 login gettys to be spawned on as many ttys as needed.
15270 * %h, %s, %U specifier support is not available anymore when
15271 used in unit files for PID 1. This is because NSS calls are
15272 not safe from PID 1. They stay available for --user
15273 instances of systemd, and as special case for the root user.
15275 * loginctl gained a new "--no-legend" switch to turn off output
15276 of the legend text.
15278 * The "sd-login.h" API gained three new calls:
15279 sd_session_is_remote(), sd_session_get_remote_user(),
15280 sd_session_get_remote_host() to query information about
15283 * The udev hardware database now also carries vendor/product
15284 information of SDIO devices.
15286 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
15287 determine whether watchdog notifications are requested by
15288 the system manager.
15290 * Socket-activated per-connection services now include a
15291 short description of the connection parameters in the
15294 * tmpfiles gained a new "--boot" option. When this is not used,
15295 only lines where the command character is not suffixed with
15296 "!" are executed. When this option is specified, those
15297 options are executed too. This partitions tmpfiles
15298 directives into those that can be safely executed at any
15299 time, and those which should be run only at boot (for
15300 example, a line that creates /run/nologin).
15302 * A new API "sd-resolve.h" has been added which provides a simple
15303 asynchronous wrapper around glibc NSS hostname resolution
15304 calls, such as getaddrinfo(). In contrast to glibc's
15305 getaddrinfo_a(), it does not use signals. In contrast to most
15306 other asynchronous name resolution libraries, this one does
15307 not reimplement DNS, but reuses NSS, so that alternate
15308 hostname resolution systems continue to work, such as mDNS,
15309 LDAP, etc. This API is based on libasyncns, but it has been
15310 cleaned up for inclusion in systemd.
15312 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
15313 "sd-daemon.h" are no longer found in individual libraries
15314 libsystemd-journal.so, libsystemd-login.so,
15315 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
15316 merged them into a single library, libsystemd.so, which
15317 provides all symbols. The reason for this is cyclic
15318 dependencies, as these libraries tend to use each other's
15319 symbols. So far, we have managed to workaround that by linking
15320 a copy of a good part of our code into each of these
15321 libraries again and again, which, however, makes certain
15322 things hard to do, like sharing static variables. Also, it
15323 substantially increases footprint. With this change, there
15324 is only one library for the basic APIs systemd
15325 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
15326 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
15327 library as well, however are subject to the --enable-kdbus
15328 switch (see below). Note that "sd-dhcp-client.h" is not part
15329 of this library (this is because it only consumes, never
15330 provides, services of/to other APIs). To make the transition
15331 easy from the separate libraries to the unified one, we
15332 provide the --enable-compat-libs compile-time switch which
15333 will generate stub libraries that are compatible with the
15334 old ones but redirect all calls to the new one.
15336 * All of the kdbus logic and the new APIs "sd-bus.h",
15337 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
15338 and "sd-utf8.h" are compile-time optional via the
15339 "--enable-kdbus" switch, and they are not compiled in by
15340 default. To make use of kdbus, you have to explicitly enable
15341 the switch. Note however, that neither the kernel nor the
15342 userspace API for all of this is considered stable yet. We
15343 want to maintain the freedom to still change the APIs for
15344 now. By specifying this build-time switch, you acknowledge
15345 that you are aware of the instability of the current
15348 * Also, note that while kdbus is pretty much complete,
15349 it lacks one thing: proper policy support. This means you
15350 can build a fully working system with all features; however,
15351 it will be highly insecure. Policy support will be added in
15352 one of the next releases, at the same time that we will
15353 declare the APIs stable.
15355 * When the kernel command line argument "kdbus" is specified,
15356 systemd will automatically load the kdbus.ko kernel module. At
15357 this stage of development, it is only useful for testing kdbus
15358 and should not be used in production. Note: if "--enable-kdbus"
15359 is specified, and the kdbus.ko kernel module is available, and
15360 "kdbus" is added to the kernel command line, the entire system
15361 runs with kdbus instead of dbus-daemon, with the above mentioned
15362 problem of missing the system policy enforcement. Also a future
15363 version of kdbus.ko or a newer systemd will not be compatible with
15364 each other, and will unlikely be able to boot the machine if only
15365 one of them is updated.
15367 * systemctl gained a new "import-environment" command which
15368 uploads the caller's environment (or parts thereof) into the
15369 service manager so that it is inherited by services started
15370 by the manager. This is useful to upload variables like
15371 $DISPLAY into the user service manager.
15373 * A new PrivateDevices= switch has been added to service units
15374 which allows running a service with a namespaced /dev
15375 directory that does not contain any device nodes for
15376 physical devices. More specifically, it only includes devices
15377 such as /dev/null, /dev/urandom, and /dev/zero which are API
15380 * logind has been extended to support behaviour like VT
15381 switching on seats that do not support a VT. This makes
15382 multi-session available on seats that are not the first seat
15383 (seat0), and on systems where kernel support for VTs has
15384 been disabled at compile-time.
15386 * If a process holds a delay lock for system sleep or shutdown
15387 and fails to release it in time, we will now log its
15388 identity. This makes it easier to identify processes that
15389 cause slow suspends or power-offs.
15391 * When parsing /etc/crypttab, support for a new key-slot=
15392 option as supported by Debian is added. It allows indicating
15393 which LUKS slot to use on disk, speeding up key loading.
15395 * The sd_journal_sendv() API call has been checked and
15396 officially declared to be async-signal-safe so that it may
15397 be invoked from signal handlers for logging purposes.
15399 * Boot-time status output is now enabled automatically after a
15400 short timeout if boot does not progress, in order to give
15401 the user an indication what she or he is waiting for.
15403 * The boot-time output has been improved to show how much time
15404 remains until jobs expire.
15406 * The KillMode= switch in service units gained a new possible
15407 value "mixed". If set, and the unit is shut down, then the
15408 initial SIGTERM signal is sent only to the main daemon
15409 process, while the following SIGKILL signal is sent to
15410 all remaining processes of the service.
15412 * When a scope unit is registered, a new property "Controller"
15413 may be set. If set to a valid bus name, systemd will send a
15414 RequestStop() signal to this name when it would like to shut
15415 down the scope. This may be used to hook manager logic into
15416 the shutdown logic of scope units. Also, scope units may now
15417 be put in a special "abandoned" state, in which case the
15418 manager process which created them takes no further
15419 responsibilities for it.
15421 * When reading unit files, systemd will now verify
15422 the access mode of these files, and warn about certain
15423 suspicious combinations. This has been added to make it
15424 easier to track down packaging bugs where unit files are
15425 marked executable or world-writable.
15427 * systemd-nspawn gained a new "--setenv=" switch to set
15428 container-wide environment variables. The similar option in
15429 systemd-activate was renamed from "--environment=" to
15430 "--setenv=" for consistency.
15432 * systemd-nspawn has been updated to create a new kdbus domain
15433 for each container that is invoked, thus allowing each
15434 container to have its own set of system and user buses,
15435 independent of the host.
15437 * systemd-nspawn gained a new --drop-capability= switch to run
15438 the container with less capabilities than the default. Both
15439 --drop-capability= and --capability= now take the special
15440 string "all" for dropping or keeping all capabilities.
15442 * systemd-nspawn gained new switches for executing containers
15443 with specific SELinux labels set.
15445 * systemd-nspawn gained a new --quiet switch to not generate
15446 any additional output but the container's own console
15449 * systemd-nspawn gained a new --share-system switch to run a
15450 container without PID namespacing enabled.
15452 * systemd-nspawn gained a new --register= switch to control
15453 whether the container is registered with systemd-machined or
15454 not. This is useful for containers that do not run full
15455 OS images, but only specific apps.
15457 * systemd-nspawn gained a new --keep-unit which may be used
15458 when invoked as the only program from a service unit, and
15459 results in registration of the unit service itself in
15460 systemd-machined, instead of a newly opened scope unit.
15462 * systemd-nspawn gained a new --network-interface= switch for
15463 moving arbitrary interfaces to the container. The new
15464 --network-veth switch creates a virtual Ethernet connection
15465 between host and container. The new --network-bridge=
15466 switch then allows assigning the host side of this virtual
15467 Ethernet connection to a bridge device.
15469 * systemd-nspawn gained a new --personality= switch for
15470 setting the kernel personality for the container. This is
15471 useful when running a 32-bit container on a 64-bit host. A
15472 similar option Personality= is now also available for service
15475 * logind will now also track a "Desktop" identifier for each
15476 session which encodes the desktop environment of it. This is
15477 useful for desktop environments that want to identify
15478 multiple running sessions of itself easily.
15480 * A new SELinuxContext= setting for service units has been
15481 added that allows setting a specific SELinux execution
15482 context for a service.
15484 * Most systemd client tools will now honour $SYSTEMD_LESS for
15485 settings of the "less" pager. By default, these tools will
15486 override $LESS to allow certain operations to work, such as
15487 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
15488 influence this logic.
15490 * systemd's "seccomp" hook-up has been changed to make use of
15491 the libseccomp library instead of using its own
15492 implementation. This has benefits for portability among
15495 * For usage together with SystemCallFilter=, a new
15496 SystemCallErrorNumber= setting has been introduced that
15497 allows configuration of a system error number to be returned
15498 on filtered system calls, instead of immediately killing the
15499 process. Also, SystemCallArchitectures= has been added to
15500 limit access to system calls of a particular architecture
15501 (in order to turn off support for unused secondary
15502 architectures). There is also a global
15503 SystemCallArchitectures= setting in system.conf now to turn
15504 off support for non-native system calls system-wide.
15506 * systemd requires a kernel with a working name_to_handle_at(),
15507 please see the kernel config requirements in the README file.
15509 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
15510 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
15511 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
15512 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
15513 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
15514 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
15515 Elia Pinto, Florian Weimer, George McCollister, Goffredo
15516 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
15517 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
15518 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
15519 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
15520 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
15521 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
15522 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
15523 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
15524 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
15525 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
15526 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
15527 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
15528 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
15529 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
15530 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
15531 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
15532 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
15534 — Berlin, 2014-02-20
15538 * logind has gained support for facilitating privileged input
15539 and drm device access for unprivileged clients. This work is
15540 useful to allow Wayland display servers (and similar
15541 programs, such as kmscon) to run under the user's ID and
15542 access input and drm devices which are normally
15543 protected. When this is used (and the kernel is new enough)
15544 logind will "mute" IO on the file descriptors passed to
15545 Wayland as long as it is in the background and "unmute" it
15546 if it returns into the foreground. This allows secure
15547 session switching without allowing background sessions to
15548 eavesdrop on input and display data. This also introduces
15549 session switching support if VT support is turned off in the
15550 kernel, and on seats that are not seat0.
15552 * A new kernel command line option luks.options= is understood
15553 now which allows specifying LUKS options for usage for LUKS
15554 encrypted partitions specified with luks.uuid=.
15556 * tmpfiles.d(5) snippets may now use specifier expansion in
15557 path names. More specifically %m, %b, %H, %v, are now
15558 replaced by the local machine id, boot id, hostname, and
15559 kernel version number.
15561 * A new tmpfiles.d(5) command "m" has been introduced which
15562 may be used to change the owner/group/access mode of a file
15563 or directory if it exists, but do nothing if it does not.
15565 * This release removes high-level support for the
15566 MemorySoftLimit= cgroup setting. The underlying kernel
15567 cgroup attribute memory.soft_limit= is currently badly
15568 designed and likely to be removed from the kernel API in its
15569 current form, hence we should not expose it for now.
15571 * The memory.use_hierarchy cgroup attribute is now enabled for
15572 all cgroups systemd creates in the memory cgroup
15573 hierarchy. This option is likely to be come the built-in
15574 default in the kernel anyway, and the non-hierarchical mode
15575 never made much sense in the intrinsically hierarchical
15578 * A new field _SYSTEMD_SLICE= is logged along with all journal
15579 messages containing the slice a message was generated
15580 from. This is useful to allow easy per-customer filtering of
15581 logs among other things.
15583 * systemd-journald will no longer adjust the group of journal
15584 files it creates to the "systemd-journal" group. Instead we
15585 rely on the journal directory to be owned by the
15586 "systemd-journal" group, and its setgid bit set, so that the
15587 kernel file system layer will automatically enforce that
15588 journal files inherit this group assignment. The reason for
15589 this change is that we cannot allow NSS look-ups from
15590 journald which would be necessary to resolve
15591 "systemd-journal" to a numeric GID, because this might
15592 create deadlocks if NSS involves synchronous queries to
15593 other daemons (such as nscd, or sssd) which in turn are
15594 logging clients of journald and might block on it, which
15595 would then dead lock. A tmpfiles.d(5) snippet included in
15596 systemd will make sure the setgid bit and group are
15597 properly set on the journal directory if it exists on every
15598 boot. However, we recommend adjusting it manually after
15599 upgrades too (or from RPM scriptlets), so that the change is
15600 not delayed until next reboot.
15602 * Backlight and random seed files in /var/lib/ have moved into
15603 the /var/lib/systemd/ directory, in order to centralize all
15604 systemd generated files in one directory.
15606 * Boot time performance measurements (as displayed by
15607 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
15608 performance information if that's available to determine how
15609 much time BIOS and boot loader initialization required. With
15610 a sufficiently new BIOS you hence no longer need to boot
15611 with Gummiboot to get access to such information.
15613 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
15614 Cristian Rodríguez, Dave Reisner, David Herrmann, David
15615 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
15616 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
15617 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
15618 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
15619 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
15621 — Berlin, 2013-10-02
15625 * The Restart= option for services now understands a new
15626 on-watchdog setting, which will restart the service
15627 automatically if the service stops sending out watchdog keep
15628 alive messages (as configured with WatchdogSec=).
15630 * The getty generator (which is responsible for bringing up a
15631 getty on configured serial consoles) will no longer only
15632 start a getty on the primary kernel console but on all
15633 others, too. This makes the order in which console= is
15634 specified on the kernel command line less important.
15636 * libsystemd-logind gained a new sd_session_get_vt() call to
15637 retrieve the VT number of a session.
15639 * If the option "tries=0" is set for an entry of /etc/crypttab
15640 its passphrase is queried indefinitely instead of any
15641 maximum number of tries.
15643 * If a service with a configure PID file terminates its PID
15644 file will now be removed automatically if it still exists
15645 afterwards. This should put an end to stale PID files.
15647 * systemd-run will now also take relative binary path names
15648 for execution and no longer insists on absolute paths.
15650 * InaccessibleDirectories= and ReadOnlyDirectories= now take
15651 paths that are optionally prefixed with "-" to indicate that
15652 it should not be considered a failure if they do not exist.
15654 * journalctl -o (and similar commands) now understands a new
15655 output mode "short-precise", it is similar to "short" but
15656 shows timestamps with usec accuracy.
15658 * The option "discard" (as known from Debian) is now
15659 synonymous to "allow-discards" in /etc/crypttab. In fact,
15660 "discard" is preferred now (since it is easier to remember
15663 * Some licensing clean-ups were made, so that more code is now
15664 LGPL-2.1 licensed than before.
15666 * A minimal tool to save/restore the display backlight
15667 brightness across reboots has been added. It will store the
15668 backlight setting as late as possible at shutdown, and
15669 restore it as early as possible during reboot.
15671 * A logic to automatically discover and enable home and swap
15672 partitions on GPT disks has been added. With this in place
15673 /etc/fstab becomes optional for many setups as systemd can
15674 discover certain partitions located on the root disk
15675 automatically. Home partitions are recognized under their
15676 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
15677 partitions are recognized under their GPT type ID
15678 0657fd6da4ab43c484e50933c84b4f4f.
15680 * systemd will no longer pass any environment from the kernel
15681 or initrd to system services. If you want to set an
15682 environment for all services, do so via the kernel command
15683 line systemd.setenv= assignment.
15685 * The systemd-sysctl tool no longer natively reads the file
15686 /etc/sysctl.conf. If desired, the file should be symlinked
15687 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
15688 legacy support by a symlink rather than built-in code, it
15689 also makes the otherwise hidden order of application of the
15690 different files visible. (Note that this partly reverts to a
15691 pre-198 application order of sysctl knobs!)
15693 * The "systemctl set-log-level" and "systemctl dump" commands
15694 have been moved to systemd-analyze.
15696 * systemd-run learned the new --remain-after-exit switch,
15697 which causes the scope unit not to be cleaned up
15698 automatically after the process terminated.
15700 * tmpfiles learned a new --exclude-prefix= switch to exclude
15701 certain paths from operation.
15703 * journald will now automatically flush all messages to disk
15704 as soon as a message at the log level CRIT, ALERT or EMERG
15707 Contributions from: Andrew Cook, Brandon Philips, Christian
15708 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
15709 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
15710 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
15711 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
15712 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
15713 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
15714 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
15715 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
15716 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
15717 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
15718 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
15719 William Giokas, Zbigniew Jędrzejewski-Szmek
15721 — Berlin, 2013-09-13
15725 * The documentation has been updated to cover the various new
15726 concepts introduced with 205.
15728 * Unit files now understand the new %v specifier which
15729 resolves to the kernel version string as returned by "uname
15732 * systemctl now supports filtering the unit list output by
15733 load state, active state and sub state, using the new
15734 --state= parameter.
15736 * "systemctl status" will now show the results of the
15737 condition checks (like ConditionPathExists= and similar) of
15738 the last start attempts of the unit. They are also logged to
15741 * "journalctl -b" may now be used to look for boot output of a
15742 specific boot. Try "journalctl -b -1" for the previous boot,
15743 but the syntax is substantially more powerful.
15745 * "journalctl --show-cursor" has been added which prints the
15746 cursor string the last shown log line. This may then be used
15747 with the new "journalctl --after-cursor=" switch to continue
15748 browsing logs from that point on.
15750 * "journalctl --force" may now be used to force regeneration
15753 * Creation of "dead" device nodes has been moved from udev
15754 into kmod and tmpfiles. Previously, udev would read the kmod
15755 databases to pre-generate dead device nodes based on meta
15756 information contained in kernel modules, so that these would
15757 be auto-loaded on access rather then at boot. As this
15758 does not really have much to do with the exposing actual
15759 kernel devices to userspace this has always been slightly
15760 alien in the udev codebase. Following the new scheme kmod
15761 will now generate a runtime snippet for tmpfiles from the
15762 module meta information and it now is tmpfiles' job to the
15763 create the nodes. This also allows overriding access and
15764 other parameters for the nodes using the usual tmpfiles
15765 facilities. As side effect this allows us to remove the
15766 CAP_SYS_MKNOD capability bit from udevd entirely.
15768 * logind's device ACLs may now be applied to these "dead"
15769 devices nodes too, thus finally allowing managed access to
15770 devices such as /dev/snd/sequencer without loading the
15771 backing module right-away.
15773 * A new RPM macro has been added that may be used to apply
15774 tmpfiles configuration during package installation.
15776 * systemd-detect-virt and ConditionVirtualization= now can
15777 detect User-Mode-Linux machines (UML).
15779 * journald will now implicitly log the effective capabilities
15780 set of processes in the message metadata.
15782 * systemd-cryptsetup has gained support for TrueCrypt volumes.
15784 * The initrd interface has been simplified (more specifically,
15785 support for passing performance data via environment
15786 variables and fsck results via files in /run has been
15787 removed). These features were non-essential, and are
15788 nowadays available in a much nicer way by having systemd in
15789 the initrd serialize its state and have the hosts systemd
15790 deserialize it again.
15792 * The udev "keymap" data files and tools to apply keyboard
15793 specific mappings of scan to key codes, and force-release
15794 scan code lists have been entirely replaced by a udev
15795 "keyboard" builtin and a hwdb data file.
15797 * systemd will now honour the kernel's "quiet" command line
15798 argument also during late shutdown, resulting in a
15799 completely silent shutdown when used.
15801 * There's now an option to control the SO_REUSEPORT socket
15802 option in .socket units.
15804 * Instance units will now automatically get a per-template
15805 subslice of system.slice unless something else is explicitly
15806 configured. For example, instances of sshd@.service will now
15807 implicitly be placed in system-sshd.slice rather than
15808 system.slice as before.
15810 * Test coverage support may now be enabled at build time.
15812 Contributions from: Dave Reisner, Frederic Crozat, Harald
15813 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
15814 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
15815 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
15816 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
15817 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
15818 Giokas, Zbigniew Jędrzejewski-Szmek
15820 — Berlin, 2013-07-23
15824 * Two new unit types have been introduced:
15826 Scope units are very similar to service units, however, are
15827 created out of pre-existing processes — instead of PID 1
15828 forking off the processes. By using scope units it is
15829 possible for system services and applications to group their
15830 own child processes (worker processes) in a powerful way
15831 which then maybe used to organize them, or kill them
15832 together, or apply resource limits on them.
15834 Slice units may be used to partition system resources in an
15835 hierarchical fashion and then assign other units to them. By
15836 default there are now three slices: system.slice (for all
15837 system services), user.slice (for all user sessions),
15838 machine.slice (for VMs and containers).
15840 Slices and scopes have been introduced primarily in
15841 context of the work to move cgroup handling to a
15842 single-writer scheme, where only PID 1
15843 creates/removes/manages cgroups.
15845 * There's a new concept of "transient" units. In contrast to
15846 normal units these units are created via an API at runtime,
15847 not from configuration from disk. More specifically this
15848 means it is now possible to run arbitrary programs as
15849 independent services, with all execution parameters passed
15850 in via bus APIs rather than read from disk. Transient units
15851 make systemd substantially more dynamic then it ever was,
15852 and useful as a general batch manager.
15854 * logind has been updated to make use of scope and slice units
15855 for managing user sessions. As a user logs in he will get
15856 his own private slice unit, to which all sessions are added
15857 as scope units. We also added support for automatically
15858 adding an instance of user@.service for the user into the
15859 slice. Effectively logind will no longer create cgroup
15860 hierarchies on its own now, it will defer entirely to PID 1
15861 for this by means of scope, service and slice units. Since
15862 user sessions this way become entities managed by PID 1
15863 the output of "systemctl" is now a lot more comprehensive.
15865 * A new mini-daemon "systemd-machined" has been added which
15866 may be used by virtualization managers to register local
15867 VMs/containers. nspawn has been updated accordingly, and
15868 libvirt will be updated shortly. machined will collect a bit
15869 of meta information about the VMs/containers, and assign
15870 them their own scope unit (see above). The collected
15871 meta-data is then made available via the "machinectl" tool,
15872 and exposed in "ps" and similar tools. machined/machinectl
15873 is compile-time optional.
15875 * As discussed earlier, the low-level cgroup configuration
15876 options ControlGroup=, ControlGroupModify=,
15877 ControlGroupPersistent=, ControlGroupAttribute= have been
15878 removed. Please use high-level attribute settings instead as
15879 well as slice units.
15881 * A new bus call SetUnitProperties() has been added to alter
15882 various runtime parameters of a unit. This is primarily
15883 useful to alter cgroup parameters dynamically in a nice way,
15884 but will be extended later on to make more properties
15885 modifiable at runtime. systemctl gained a new set-properties
15886 command that wraps this call.
15888 * A new tool "systemd-run" has been added which can be used to
15889 run arbitrary command lines as transient services or scopes,
15890 while configuring a number of settings via the command
15891 line. This tool is currently very basic, however already
15892 very useful. We plan to extend this tool to even allow
15893 queuing of execution jobs with time triggers from the
15894 command line, similar in fashion to "at".
15896 * nspawn will now inform the user explicitly that kernels with
15897 audit enabled break containers, and suggest the user to turn
15900 * Support for detecting the IMA and AppArmor security
15901 frameworks with ConditionSecurity= has been added.
15903 * journalctl gained a new "-k" switch for showing only kernel
15904 messages, mimicking dmesg output; in addition to "--user"
15905 and "--system" switches for showing only user's own logs
15908 * systemd-delta can now show information about drop-in
15909 snippets extending unit files.
15911 * libsystemd-bus has been substantially updated but is still
15912 not available as public API.
15914 * systemd will now look for the "debug" argument on the kernel
15915 command line and enable debug logging, similar to what
15916 "systemd.log_level=debug" already did before.
15918 * "systemctl set-default", "systemctl get-default" has been
15919 added to configure the default.target symlink, which
15920 controls what to boot into by default.
15922 * "systemctl set-log-level" has been added as a convenient
15923 way to raise and lower systemd logging threshold.
15925 * "systemd-analyze plot" will now show the time the various
15926 generators needed for execution, as well as information
15927 about the unit file loading.
15929 * libsystemd-journal gained a new sd_journal_open_files() call
15930 for opening specific journal files. journactl also gained a
15931 new switch to expose this new functionality. Previously we
15932 only supported opening all files from a directory, or all
15933 files from the system, as opening individual files only is
15934 racy due to journal file rotation.
15936 * systemd gained the new DefaultEnvironment= setting in
15937 /etc/systemd/system.conf to set environment variables for
15940 * If a privileged process logs a journal message with the
15941 OBJECT_PID= field set, then journald will automatically
15942 augment this with additional OBJECT_UID=, OBJECT_GID=,
15943 OBJECT_COMM=, OBJECT_EXE=, … fields. This is useful if
15944 system services want to log events about specific client
15945 processes. journactl/systemctl has been updated to make use
15946 of this information if all log messages regarding a specific
15949 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
15950 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
15951 Reisner, David Coppa, David King, David Strauss, Eelco
15952 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
15953 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
15954 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
15955 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
15956 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
15957 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
15958 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
15959 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
15960 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
15961 Łukasz Stelmach, 장동준
15965 * The Python bindings gained some minimal support for the APIs
15966 exposed by libsystemd-logind.
15968 * ConditionSecurity= gained support for detecting SMACK. Since
15969 this condition already supports SELinux and AppArmor we only
15970 miss IMA for this. Patches welcome!
15972 Contributions from: Karol Lewandowski, Lennart Poettering,
15973 Zbigniew Jędrzejewski-Szmek
15977 * systemd-nspawn will now create /etc/resolv.conf if
15978 necessary, before bind-mounting the host's file onto it.
15980 * systemd-nspawn will now store meta information about a
15981 container on the container's cgroup as extended attribute
15982 fields, including the root directory.
15984 * The cgroup hierarchy has been reworked in many ways. All
15985 objects any of the components systemd creates in the cgroup
15986 tree are now suffixed. More specifically, user sessions are
15987 now placed in cgroups suffixed with ".session", users in
15988 cgroups suffixed with ".user", and nspawn containers in
15989 cgroups suffixed with ".nspawn". Furthermore, all cgroup
15990 names are now escaped in a simple scheme to avoid collision
15991 of userspace object names with kernel filenames. This work
15992 is preparation for making these objects relocatable in the
15993 cgroup tree, in order to allow easy resource partitioning of
15994 these objects without causing naming conflicts.
15996 * systemctl list-dependencies gained the new switches
15997 --plain, --reverse, --after and --before.
15999 * systemd-inhibit now shows the process name of processes that
16000 have taken an inhibitor lock.
16002 * nss-myhostname will now also resolve "localhost"
16003 implicitly. This makes /etc/hosts an optional file and
16004 nicely handles that on IPv6 ::1 maps to both "localhost" and
16005 the local hostname.
16007 * libsystemd-logind.so gained a new call
16008 sd_get_machine_names() to enumerate running containers and
16009 VMs (currently only supported by very new libvirt and
16010 nspawn). sd_login_monitor can now be used to watch
16011 VMs/containers coming and going.
16013 * .include is not allowed recursively anymore, and only in
16014 unit files. Usually it is better to use drop-in snippets in
16015 .d/*.conf anyway, as introduced with systemd 198.
16017 * systemd-analyze gained a new "critical-chain" command that
16018 determines the slowest chain of units run during system
16019 boot-up. It is very useful for tracking down where
16020 optimizing boot time is the most beneficial.
16022 * systemd will no longer allow manipulating service paths in
16023 the name=systemd:/system cgroup tree using ControlGroup= in
16024 units. (But is still fine with it in all other dirs.)
16026 * There's a new systemd-nspawn@.service service file that may
16027 be used to easily run nspawn containers as system
16028 services. With the container's root directory in
16029 /var/lib/container/foobar it is now sufficient to run
16030 "systemctl start systemd-nspawn@foobar.service" to boot it.
16032 * systemd-cgls gained a new parameter "--machine" to list only
16033 the processes within a certain container.
16035 * ConditionSecurity= now can check for "apparmor". We still
16036 are lacking checks for SMACK and IMA for this condition
16037 check though. Patches welcome!
16039 * A new configuration file /etc/systemd/sleep.conf has been
16040 added that may be used to configure which kernel operation
16041 systemd is supposed to execute when "suspend", "hibernate"
16042 or "hybrid-sleep" is requested. This makes the new kernel
16043 "freeze" state accessible to the user.
16045 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
16046 the passed argument if applicable.
16048 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
16049 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
16050 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
16051 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
16052 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
16053 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
16054 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
16059 * The output of 'systemctl list-jobs' got some polishing. The
16060 '--type=' argument may now be passed more than once. A new
16061 command 'systemctl list-sockets' has been added which shows
16062 a list of kernel sockets systemd is listening on with the
16063 socket units they belong to, plus the units these socket
16066 * The experimental libsystemd-bus library got substantial
16067 updates to work in conjunction with the (also experimental)
16068 kdbus kernel project. It works well enough to exchange
16069 messages with some sophistication. Note that kdbus is not
16070 ready yet, and the library is mostly an elaborate test case
16071 for now, and not installable.
16073 * systemd gained a new unit 'systemd-static-nodes.service'
16074 that generates static device nodes earlier during boot, and
16075 can run in conjunction with udev.
16077 * libsystemd-login gained a new call sd_pid_get_user_unit()
16078 to retrieve the user systemd unit a process is running
16079 in. This is useful for systems where systemd is used as
16082 * systemd-nspawn now places all containers in the new /machine
16083 top-level cgroup directory in the name=systemd
16084 hierarchy. libvirt will soon do the same, so that we get a
16085 uniform separation of /system, /user and /machine for system
16086 services, user processes and containers/virtual
16087 machines. This new cgroup hierarchy is also useful to stick
16088 stable names to specific container instances, which can be
16089 recognized later this way (this name may be controlled
16090 via systemd-nspawn's new -M switch). libsystemd-login also
16091 gained a new call sd_pid_get_machine_name() to retrieve the
16092 name of the container/VM a specific process belongs to.
16094 * bootchart can now store its data in the journal.
16096 * libsystemd-journal gained a new call
16097 sd_journal_add_conjunction() for AND expressions to the
16098 matching logic. This can be used to express more complex
16099 logical expressions.
16101 * journactl can now take multiple --unit= and --user-unit=
16104 * The cryptsetup logic now understands the "luks.key=" kernel
16105 command line switch for specifying a file to read the
16106 decryption key from. Also, if a configured key file is not
16107 found the tool will now automatically fall back to prompting
16110 * Python systemd.journal module was updated to wrap recently
16111 added functions from libsystemd-journal. The interface was
16112 changed to bring the low level interface in s.j._Reader
16113 closer to the C API, and the high level interface in
16114 s.j.Reader was updated to wrap and convert all data about
16117 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
16118 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
16119 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
16120 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
16121 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
16122 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
16126 * journalctl --update-catalog now understands a new --root=
16127 option to operate on catalogs found in a different root
16130 * During shutdown after systemd has terminated all running
16131 services a final killing loop kills all remaining left-over
16132 processes. We will now print the name of these processes
16133 when we send SIGKILL to them, since this usually indicates a
16136 * If /etc/crypttab refers to password files stored on
16137 configured mount points automatic dependencies will now be
16138 generated to ensure the specific mount is established first
16139 before the key file is attempted to be read.
16141 * 'systemctl status' will now show information about the
16142 network sockets a socket unit is listening on.
16144 * 'systemctl status' will also shown information about any
16145 drop-in configuration file for units. (Drop-In configuration
16146 files in this context are files such as
16147 /etc/systemd/system/foobar.service.d/*.conf)
16149 * systemd-cgtop now optionally shows summed up CPU times of
16150 cgroups. Press '%' while running cgtop to switch between
16151 percentage and absolute mode. This is useful to determine
16152 which cgroups use up the most CPU time over the entire
16153 runtime of the system. systemd-cgtop has also been updated
16154 to be 'pipeable' for processing with further shell tools.
16156 * 'hostnamectl set-hostname' will now allow setting of FQDN
16159 * The formatting and parsing of time span values has been
16160 changed. The parser now understands fractional expressions
16161 such as "5.5h". The formatter will now output fractional
16162 expressions for all time spans under 1min, i.e. "5.123456s"
16163 rather than "5s 123ms 456us". For time spans under 1s
16164 millisecond values are shown, for those under 1ms
16165 microsecond values are shown. This should greatly improve
16166 all time-related output of systemd.
16168 * libsystemd-login and libsystemd-journal gained new
16169 functions for querying the poll() events mask and poll()
16170 timeout value for integration into arbitrary event
16173 * localectl gained the ability to list available X11 keymaps
16174 (models, layouts, variants, options).
16176 * 'systemd-analyze dot' gained the ability to filter for
16177 specific units via shell-style globs, to create smaller,
16178 more useful graphs. I.e. it is now possible to create simple
16179 graphs of all the dependencies between only target units, or
16180 of all units that Avahi has dependencies with.
16182 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
16183 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
16184 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
16185 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
16186 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
16187 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
16188 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
16192 * The boot-time readahead implementation for rotating media
16193 will now read the read-ahead data in multiple passes which
16194 consist of all read requests made in equidistant time
16195 intervals. This means instead of strictly reading read-ahead
16196 data in its physical order on disk we now try to find a
16197 middle ground between physical and access time order.
16199 * /etc/os-release files gained a new BUILD_ID= field for usage
16200 on operating systems that provide continuous builds of OS
16203 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
16204 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
16205 William Douglas, Zbigniew Jędrzejewski-Szmek
16209 * systemd-python gained an API exposing libsystemd-daemon.
16211 * The SMACK setup logic gained support for uploading CIPSO
16214 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
16215 ReadOnlyDirectories= and InaccessibleDirectories= has
16216 changed. The private /tmp and /var/tmp directories are now
16217 shared by all processes of a service (which means
16218 ExecStartPre= may now leave data in /tmp that ExecStart= of
16219 the same service can still access). When a service is
16220 stopped its temporary directories are immediately deleted
16221 (normal clean-up with tmpfiles is still done in addition to
16224 * By default, systemd will now set a couple of sysctl
16225 variables in the kernel: the safe sysrq options are turned
16226 on, IP route verification is turned on, and source routing
16227 disabled. The recently added hardlink and softlink
16228 protection of the kernel is turned on. These settings should
16229 be reasonably safe, and good defaults for all new systems.
16231 * The predictable network naming logic may now be turned off
16232 with a new kernel command line switch: net.ifnames=0.
16234 * A new libsystemd-bus module has been added that implements a
16235 pretty complete D-Bus client library. For details see:
16237 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
16239 * journald will now explicitly flush the journal files to disk
16240 at the latest 5min after each write. The file will then also
16241 be marked offline until the next write. This should increase
16242 reliability in case of a crash. The synchronization delay
16243 can be configured via SyncIntervalSec= in journald.conf.
16245 * There's a new remote-fs-setup.target unit that can be used
16246 to pull in specific services when at least one remote file
16247 system is to be mounted.
16249 * There are new targets timers.target and paths.target as
16250 canonical targets to pull user timer and path units in
16251 from. This complements sockets.target with a similar
16252 purpose for socket units.
16254 * libudev gained a new call udev_device_set_attribute_value()
16255 to set sysfs attributes of a device.
16257 * The udev daemon now sets the default number of worker
16258 processes executed in parallel based on the number of available
16259 CPUs instead of the amount of available RAM. This is supposed
16260 to provide a more reliable default and limit a too aggressive
16261 parallelism for setups with 1000s of devices connected.
16263 Contributions from: Auke Kok, Colin Walters, Cristian
16264 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
16265 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
16266 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
16267 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
16268 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
16269 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
16270 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
16271 Zbigniew Jędrzejewski-Szmek
16275 * Configuration of unit files may now be extended via drop-in
16276 files without having to edit/override the unit files
16277 themselves. More specifically, if the administrator wants to
16278 change one value for a service file foobar.service he can
16279 now do so by dropping in a configuration snippet into
16280 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
16281 will load all these snippets and apply them on top of the
16282 main unit configuration file, possibly extending or
16283 overriding its settings. Using these drop-in snippets is
16284 generally nicer than the two earlier options for changing
16285 unit files locally: copying the files from
16286 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
16287 them there; or creating a new file in /etc/systemd/system/
16288 that incorporates the original one via ".include". Drop-in
16289 snippets into these .d/ directories can be placed in any
16290 directory systemd looks for units in, and the usual
16291 overriding semantics between /usr/lib, /etc and /run apply
16294 * Most unit file settings which take lists of items can now be
16295 reset by assigning the empty string to them. For example,
16296 normally, settings such as Environment=FOO=BAR append a new
16297 environment variable assignment to the environment block,
16298 each time they are used. By assigning Environment= the empty
16299 string the environment block can be reset to empty. This is
16300 particularly useful with the .d/*.conf drop-in snippets
16301 mentioned above, since this adds the ability to reset list
16302 settings from vendor unit files via these drop-ins.
16304 * systemctl gained a new "list-dependencies" command for
16305 listing the dependencies of a unit recursively.
16307 * Inhibitors are now honored and listed by "systemctl
16308 suspend", "systemctl poweroff" (and similar) too, not only
16309 GNOME. These commands will also list active sessions by
16312 * Resource limits (as exposed by the various control group
16313 controllers) can now be controlled dynamically at runtime
16314 for all units. More specifically, you can now use a command
16315 like "systemctl set-cgroup-attr foobar.service cpu.shares
16316 2000" to alter the CPU shares a specific service gets. These
16317 settings are stored persistently on disk, and thus allow the
16318 administrator to easily adjust the resource usage of
16319 services with a few simple commands. This dynamic resource
16320 management logic is also available to other programs via the
16321 bus. Almost any kernel cgroup attribute and controller is
16324 * systemd-vconsole-setup will now copy all font settings to
16325 all allocated VTs, where it previously applied them only to
16328 * libsystemd-login gained the new sd_session_get_tty() API
16331 * This release drops support for a few legacy or
16332 distribution-specific LSB facility names when parsing init
16333 scripts: $x-display-manager, $mail-transfer-agent,
16334 $mail-transport-agent, $mail-transfer-agent, $smtp,
16335 $null. Also, the mail-transfer-agent.target unit backing
16336 this has been removed. Distributions which want to retain
16337 compatibility with this should carry the burden for
16338 supporting this themselves and patch support for these back
16339 in, if they really need to. Also, the facilities $syslog and
16340 $local_fs are now ignored, since systemd does not support
16341 early-boot LSB init scripts anymore, and these facilities
16342 are implied anyway for normal services. syslog.target has
16345 * There are new bus calls on PID1's Manager object for
16346 cancelling jobs, and removing snapshot units. Previously,
16347 both calls were only available on the Job and Snapshot
16348 objects themselves.
16350 * systemd-journal-gatewayd gained SSL support.
16352 * The various "environment" files, such as /etc/locale.conf
16353 now support continuation lines with a backslash ("\") as
16354 last character in the line, similarly in style (but different)
16355 to how this is supported in shells.
16357 * For normal user processes the _SYSTEMD_USER_UNIT= field is
16358 now implicitly appended to every log entry logged. systemctl
16359 has been updated to filter by this field when operating on a
16360 user systemd instance.
16362 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
16363 CAP_AUDIT_CONTROL capabilities to the capabilities set for
16364 the container. This makes it easier to boot unmodified
16365 Fedora systems in a container, which however still requires
16366 audit=0 to be passed on the kernel command line. Auditing in
16367 kernel and userspace is unfortunately still too broken in
16368 context of containers, hence we recommend compiling it out
16369 of the kernel or using audit=0. Hopefully this will be fixed
16370 one day for good in the kernel.
16372 * nspawn gained the new --bind= and --bind-ro= parameters to
16373 bind mount specific directories from the host into the
16376 * nspawn will now mount its own devpts file system instance
16377 into the container, in order not to leak pty devices from
16378 the host into the container.
16380 * systemd will now read the firmware boot time performance
16381 information from the EFI variables, if the used boot loader
16382 supports this, and takes it into account for boot performance
16383 analysis via "systemd-analyze". This is currently supported
16384 only in conjunction with Gummiboot, but could be supported
16385 by other boot loaders too. For details see:
16387 https://systemd.io/BOOT_LOADER_INTERFACE
16389 * A new generator has been added that automatically mounts the
16390 EFI System Partition (ESP) to /boot, if that directory
16391 exists, is empty, and no other file system has been
16392 configured to be mounted there.
16394 * logind will now send out PrepareForSleep(false) out
16395 unconditionally, after coming back from suspend. This may be
16396 used by applications as asynchronous notification for
16397 system resume events.
16399 * "systemctl unlock-sessions" has been added, that allows
16400 unlocking the screens of all user sessions at once, similar
16401 to how "systemctl lock-sessions" already locked all users
16402 sessions. This is backed by a new D-Bus call UnlockSessions().
16404 * "loginctl seat-status" will now show the master device of a
16405 seat. (i.e. the device of a seat that needs to be around for
16406 the seat to be considered available, usually the graphics
16409 * tmpfiles gained a new "X" line type, that allows
16410 configuration of files and directories (with wildcards) that
16411 shall be excluded from automatic cleanup ("aging").
16413 * udev default rules set the device node permissions now only
16414 at "add" events, and do not change them any longer with a
16415 later "change" event.
16417 * The log messages for lid events and power/sleep keypresses
16418 now carry a message ID.
16420 * We now have a substantially larger unit test suite, but this
16421 continues to be work in progress.
16423 * udevadm hwdb gained a new --root= parameter to change the
16424 root directory to operate relative to.
16426 * logind will now issue a background sync() request to the kernel
16427 early at shutdown, so that dirty buffers are flushed to disk early
16428 instead of at the last moment, in order to optimize shutdown
16431 * A new bootctl tool has been added that is an interface for
16432 certain boot loader operations. This is currently a preview
16433 and is likely to be extended into a small mechanism daemon
16434 like timedated, localed, hostnamed, and can be used by
16435 graphical UIs to enumerate available boot options, and
16436 request boot into firmware operations.
16438 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
16439 the rest of the package. It also has been updated to work
16440 correctly in initrds.
16442 * polkit previously has been runtime optional, and is now also
16443 compile time optional via a configure switch.
16445 * systemd-analyze has been reimplemented in C. Also "systemctl
16446 dot" has moved into systemd-analyze.
16448 * "systemctl status" with no further parameters will now print
16449 the status of all active or failed units.
16451 * Operations such as "systemctl start" can now be executed
16452 with a new mode "--irreversible" which may be used to queue
16453 operations that cannot accidentally be reversed by a later
16454 job queuing. This is by default used to make shutdown
16455 requests more robust.
16457 * The Python API of systemd now gained a new module for
16458 reading journal files.
16460 * A new tool kernel-install has been added that can install
16461 kernel images according to the Boot Loader Specification:
16463 https://systemd.io/BOOT_LOADER_SPECIFICATION
16465 * Boot time console output has been improved to provide
16466 animated boot time output for hanging jobs.
16468 * A new tool systemd-activate has been added which can be used
16469 to test socket activation with, directly from the command
16470 line. This should make it much easier to test and debug
16471 socket activation in daemons.
16473 * journalctl gained a new "--reverse" (or -r) option to show
16474 journal output in reverse order (i.e. newest line first).
16476 * journalctl gained a new "--pager-end" (or -e) option to jump
16477 to immediately jump to the end of the journal in the
16478 pager. This is only supported in conjunction with "less".
16480 * journalctl gained a new "--user-unit=" option, that works
16481 similarly to "--unit=" but filters for user units rather than
16484 * A number of unit files to ease adoption of systemd in
16485 initrds has been added. This moves some minimal logic from
16486 the various initrd implementations into systemd proper.
16488 * The journal files are now owned by a new group
16489 "systemd-journal", which exists specifically to allow access
16490 to the journal, and nothing else. Previously, we used the
16491 "adm" group for that, which however possibly covers more
16492 than just journal/log file access. This new group is now
16493 already used by systemd-journal-gatewayd to ensure this
16494 daemon gets access to the journal files and as little else
16495 as possible. Note that "make install" will also set FS ACLs
16496 up for /var/log/journal to give "adm" and "wheel" read
16497 access to it, in addition to "systemd-journal" which owns
16498 the journal files. We recommend that packaging scripts also
16499 add read access to "adm" + "wheel" to /var/log/journal, and
16500 all existing/future journal files. To normal users and
16501 administrators little changes, however packagers need to
16502 ensure to create the "systemd-journal" system group at
16503 package installation time.
16505 * The systemd-journal-gatewayd now runs as unprivileged user
16506 systemd-journal-gateway:systemd-journal-gateway. Packaging
16507 scripts need to create these system user/group at
16510 * timedated now exposes a new boolean property CanNTP that
16511 indicates whether a local NTP service is available or not.
16513 * systemd-detect-virt will now also detect xen PVs
16515 * The pstore file system is now mounted by default, if it is
16518 * In addition to the SELinux and IMA policies we will now also
16519 load SMACK policies at early boot.
16521 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
16522 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
16523 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
16524 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
16525 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
16526 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
16527 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
16528 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
16529 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
16530 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
16531 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
16532 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
16533 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
16534 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
16538 * Timer units now support calendar time events in addition to
16539 monotonic time events. That means you can now trigger a unit
16540 based on a calendar time specification such as "Thu,Fri
16541 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
16542 or fifth day of any month of the year 2013, given that it is
16543 a Thursday or a Friday. This brings timer event support
16544 considerably closer to cron's capabilities. For details on
16545 the supported calendar time specification language see
16548 * udev now supports a number of different naming policies for
16549 network interfaces for predictable names, and a combination
16550 of these policies is now the default. Please see this wiki
16551 document for details:
16553 https://www.freedesktop.org/software/systemd/man/systemd.net-naming-scheme.html
16555 * Auke Kok's bootchart implementation has been added to the
16556 systemd tree. It is an optional component that can graph the
16557 boot in quite some detail. It is one of the best bootchart
16558 implementations around and minimal in its code and
16561 * nss-myhostname has been integrated into the systemd source
16562 tree. nss-myhostname guarantees that the local hostname
16563 always stays resolvable via NSS. It has been a weak
16564 requirement of systemd-hostnamed since a long time, and
16565 since its code is actually trivial we decided to just
16566 include it in systemd's source tree. It can be turned off
16567 with a configure switch.
16569 * The read-ahead logic is now capable of properly detecting
16570 whether a btrfs file system is on SSD or rotating media, in
16571 order to optimize the read-ahead scheme. Previously, it was
16572 only capable of detecting this on traditional file systems
16575 * In udev, additional device properties are now read from the
16576 IAB in addition to the OUI database. Also, Bluetooth company
16577 identities are attached to the devices as well.
16579 * In service files %U may be used as specifier that is
16580 replaced by the configured user name of the service.
16582 * nspawn may now be invoked without a controlling TTY. This
16583 makes it suitable for invocation as its own service. This
16584 may be used to set up a simple containerized server system
16585 using only core OS tools.
16587 * systemd and nspawn can now accept socket file descriptors
16588 when they are started for socket activation. This enables
16589 implementation of socket activated nspawn
16590 containers. i.e. think about autospawning an entire OS image
16591 when the first SSH or HTTP connection is received. We expect
16592 that similar functionality will also be added to libvirt-lxc
16595 * journalctl will now suppress ANSI color codes when
16596 presenting log data.
16598 * systemctl will no longer show control group information for
16599 a unit if the control group is empty anyway.
16601 * logind can now automatically suspend/hibernate/shutdown the
16604 * /etc/machine-info and hostnamed now also expose the chassis
16605 type of the system. This can be used to determine whether
16606 the local system is a laptop, desktop, handset or
16607 tablet. This information may either be configured by the
16608 user/vendor or is automatically determined from ACPI and DMI
16609 information if possible.
16611 * A number of polkit actions are now bound together with "imply"
16612 rules. This should simplify creating UIs because many actions
16613 will now authenticate similar ones as well.
16615 * Unit files learnt a new condition ConditionACPower= which
16616 may be used to conditionalize a unit depending on whether an
16617 AC power source is connected or not, of whether the system
16618 is running on battery power.
16620 * systemctl gained a new "is-failed" verb that may be used in
16621 shell scripts and suchlike to check whether a specific unit
16622 is in the "failed" state.
16624 * The EnvironmentFile= setting in unit files now supports file
16625 globbing, and can hence be used to easily read a number of
16626 environment files at once.
16628 * systemd will no longer detect and recognize specific
16629 distributions. All distribution-specific #ifdeffery has been
16630 removed, systemd is now fully generic and
16631 distribution-agnostic. Effectively, not too much is lost as
16632 a lot of the code is still accessible via explicit configure
16633 switches. However, support for some distribution specific
16634 legacy configuration file formats has been dropped. We
16635 recommend distributions to simply adopt the configuration
16636 files everybody else uses now and convert the old
16637 configuration from packaging scripts. Most distributions
16638 already did that. If that's not possible or desirable,
16639 distributions are welcome to forward port the specific
16640 pieces of code locally from the git history.
16642 * When logging a message about a unit systemd will now always
16643 log the unit name in the message meta data.
16645 * localectl will now also discover system locale data that is
16646 not stored in locale archives, but directly unpacked.
16648 * logind will no longer unconditionally use framebuffer
16649 devices as seat masters, i.e. as devices that are required
16650 to be existing before a seat is considered preset. Instead,
16651 it will now look for all devices that are tagged as
16652 "seat-master" in udev. By default, framebuffer devices will
16653 be marked as such, but depending on local systems, other
16654 devices might be marked as well. This may be used to
16655 integrate graphics cards using closed source drivers (such
16656 as NVidia ones) more nicely into logind. Note however, that
16657 we recommend using the open source NVidia drivers instead,
16658 and no udev rules for the closed-source drivers will be
16659 shipped from us upstream.
16661 Contributions from: Adam Williamson, Alessandro Crismani, Auke
16662 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
16663 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
16664 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
16665 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
16666 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
16667 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
16668 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
16669 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
16670 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
16671 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
16672 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
16677 * udev gained support for loading additional device properties
16678 from an indexed database that is keyed by vendor/product IDs
16679 and similar device identifiers. For the beginning this
16680 "hwdb" is populated with data from the well-known PCI and
16681 USB database, but also includes PNP, ACPI and OID data. In
16682 the longer run this indexed database shall grow into
16683 becoming the one central database for non-essential
16684 userspace device metadata. Previously, data from the PCI/USB
16685 database was only attached to select devices, since the
16686 lookup was a relatively expensive operation due to O(n) time
16687 complexity (with n being the number of entries in the
16688 database). Since this is now O(1), we decided to add in this
16689 data for all devices where this is available, by
16690 default. Note that the indexed database needs to be rebuilt
16691 when new data files are installed. To achieve this you need
16692 to update your packaging scripts to invoke "udevadm hwdb
16693 --update" after installation of hwdb data files. For
16694 RPM-based distributions we introduced the new
16695 %udev_hwdb_update macro for this purpose.
16697 * The Journal gained support for the "Message Catalog", an
16698 indexed database to link up additional information with
16699 journal entries. For further details please check:
16701 https://www.freedesktop.org/wiki/Software/systemd/catalog
16703 The indexed message catalog database also needs to be
16704 rebuilt after installation of message catalog files. Use
16705 "journalctl --update-catalog" for this. For RPM-based
16706 distributions we introduced the %journal_catalog_update
16707 macro for this purpose.
16709 * The Python Journal bindings gained support for the standard
16710 Python logging framework.
16712 * The Journal API gained new functions for checking whether
16713 the underlying file system of a journal file is capable of
16714 properly reporting file change notifications, or whether
16715 applications that want to reflect journal changes "live"
16716 need to recheck journal files continuously in appropriate
16719 * It is now possible to set the "age" field for tmpfiles
16720 entries to 0, indicating that files matching this entry
16721 shall always be removed when the directories are cleaned up.
16723 * coredumpctl gained a new "gdb" verb which invokes gdb
16724 right-away on the selected coredump.
16726 * There's now support for "hybrid sleep" on kernels that
16727 support this, in addition to "suspend" and "hibernate". Use
16728 "systemctl hybrid-sleep" to make use of this.
16730 * logind's HandleSuspendKey= setting (and related settings)
16731 now gained support for a new "lock" setting to simply
16732 request the screen lock on all local sessions, instead of
16733 actually executing a suspend or hibernation.
16735 * systemd will now mount the EFI variables file system by
16738 * Socket units now gained support for configuration of the
16739 SMACK security label.
16741 * timedatectl will now output the time of the last and next
16742 daylight saving change.
16744 * We dropped support for various legacy and distro-specific
16745 concepts, such as insserv, early-boot SysV services
16746 (i.e. those for non-standard runlevels such as 'b' or 'S')
16747 or ArchLinux /etc/rc.conf support. We recommend the
16748 distributions who still need support this to either continue
16749 to maintain the necessary patches downstream, or find a
16750 different solution. (Talk to us if you have questions!)
16752 * Various systemd components will now bypass polkit checks for
16753 root and otherwise handle properly if polkit is not found to
16754 be around. This should fix most issues for polkit-less
16755 systems. Quite frankly this should have been this way since
16756 day one. It is absolutely our intention to make systemd work
16757 fine on polkit-less systems, and we consider it a bug if
16758 something does not work as it should if polkit is not around.
16760 * For embedded systems it is now possible to build udev and
16761 systemd without blkid and/or kmod support.
16763 * "systemctl switch-root" is now capable of switching root
16764 more than once. I.e. in addition to transitions from the
16765 initrd to the host OS it is now possible to transition to
16766 further OS images from the host. This is useful to implement
16767 offline updating tools.
16769 * Various other additions have been made to the RPM macros
16770 shipped with systemd. Use %udev_rules_update() after
16771 installing new udev rules files. %_udevhwdbdir,
16772 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
16773 %_sysctldir are now available which resolve to the right
16774 directories for packages to place various data files in.
16776 * journalctl gained the new --full switch (in addition to
16777 --all, to disable ellipsation for long messages.
16779 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
16780 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
16781 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
16782 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
16783 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
16784 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
16785 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
16786 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
16787 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
16791 * journalctl gained new --since= and --until= switches to
16792 filter by time. It also now supports nice filtering for
16793 units via --unit=/-u.
16795 * Type=oneshot services may use ExecReload= and do the
16798 * The journal daemon now supports time-based rotation and
16799 vacuuming, in addition to the usual disk-space based
16802 * The journal will now index the available field values for
16803 each field name. This enables clients to show pretty drop
16804 downs of available match values when filtering. The bash
16805 completion of journalctl has been updated
16806 accordingly. journalctl gained a new switch -F to list all
16807 values a certain field takes in the journal database.
16809 * More service events are now written as structured messages
16810 to the journal, and made recognizable via message IDs.
16812 * The timedated, localed and hostnamed mini-services which
16813 previously only provided support for changing time, locale
16814 and hostname settings from graphical DEs such as GNOME now
16815 also have a minimal (but very useful) text-based client
16816 utility each. This is probably the nicest way to changing
16817 these settings from the command line now, especially since
16818 it lists available options and is fully integrated with bash
16821 * There's now a new tool "systemd-coredumpctl" to list and
16822 extract coredumps from the journal.
16824 * We now install a README each in /var/log/ and
16825 /etc/rc.d/init.d explaining where the system logs and init
16826 scripts went. This hopefully should help folks who go to
16827 that dirs and look into the otherwise now empty void and
16828 scratch their heads.
16830 * When user-services are invoked (by systemd --user) the
16831 $MANAGERPID env var is set to the PID of systemd.
16833 * SIGRTMIN+24 when sent to a --user instance will now result
16834 in immediate termination of systemd.
16836 * gatewayd received numerous feature additions such as a
16837 "follow" mode, for live syncing and filtering.
16839 * browse.html now allows filtering and showing detailed
16840 information on specific entries. Keyboard navigation and
16841 mouse screen support has been added.
16843 * gatewayd/journalctl now supports HTML5/JSON
16844 Server-Sent-Events as output.
16846 * The SysV init script compatibility logic will now
16847 heuristically determine whether a script supports the
16848 "reload" verb, and only then make this available as
16849 "systemctl reload".
16851 * "systemctl status --follow" has been removed, use "journalctl
16854 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
16855 have been removed since they are hardly useful to be
16858 * And I'd like to take the opportunity to specifically mention
16859 Zbigniew for his great contributions. Zbigniew, you rock!
16861 Contributions from: Andrew Eikum, Christian Hesse, Colin
16862 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
16863 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
16864 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
16865 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
16866 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
16867 Jędrzejewski-Szmek, Сковорода Никита Андреевич
16871 * If /etc/vconsole.conf is non-existent or empty we will no
16872 longer load any console font or key map at boot by
16873 default. Instead the kernel defaults will be left
16874 intact. This is definitely the right thing to do, as no
16875 configuration should mean no configuration, and hard-coding
16876 font names that are different on all archs is probably a bad
16877 idea. Also, the kernel default key map and font should be
16878 good enough for most cases anyway, and mostly identical to
16879 the userspace fonts/key maps we previously overloaded them
16880 with. If distributions want to continue to default to a
16881 non-kernel font or key map they should ship a default
16882 /etc/vconsole.conf with the appropriate contents.
16884 Contributions from: Colin Walters, Daniel J Walsh, Dave
16885 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
16886 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
16890 * journalctl gained a new --cursor= switch to show entries
16891 starting from the specified location in the journal.
16893 * We now enforce a size limit on journal entry fields exported
16894 with "-o json" in journalctl. Fields larger than 4K will be
16895 assigned null. This can be turned off with --all.
16897 * An (optional) journal gateway daemon is now available as
16898 "systemd-journal-gatewayd.service". This service provides
16899 access to the journal via HTTP and JSON. This functionality
16900 will be used to implement live log synchronization in both
16901 pull and push modes, but has various other users too, such
16902 as easy log access for debugging of embedded devices. Right
16903 now it is already useful to retrieve the journal via HTTP:
16905 # systemctl start systemd-journal-gatewayd.service
16906 # wget http://localhost:19531/entries
16908 This will download the journal contents in a
16909 /var/log/messages compatible format. The same as JSON:
16911 # curl -H"Accept: application/json" http://localhost:19531/entries
16913 This service is also accessible via a web browser where a
16914 single static HTML5 app is served that uses the JSON logic
16915 to enable the user to do some basic browsing of the
16916 journal. This will be extended later on. Here's an example
16917 screenshot of this app in its current state:
16919 https://0pointer.de/public/journal-gatewayd
16921 Contributions from: Kay Sievers, Lennart Poettering, Robert
16922 Milasan, Tom Gundersen
16926 * The bash completion logic is now available for journalctl
16929 * We do not mount the "cpuset" controller anymore together with
16930 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
16931 started if no parameters are assigned to it. "cpuset" hence
16932 broke code that assumed it could create "cpu" groups and
16935 * journalctl -f will now subscribe to terminal size changes,
16936 and line break accordingly.
16938 Contributions from: Dave Reisner, Kay Sievers, Lennart
16939 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
16943 * nspawn will now create a symlink /etc/localtime in the
16944 container environment, copying the host's timezone
16945 setting. Previously this has been done via a bind mount, but
16946 since symlinks cannot be bind mounted this has now been
16947 changed to create/update the appropriate symlink.
16949 * journalctl -n's line number argument is now optional, and
16950 will default to 10 if omitted.
16952 * journald will now log the maximum size the journal files may
16953 take up on disk. This is particularly useful if the default
16954 built-in logic of determining this parameter from the file
16955 system size is used. Use "systemctl status
16956 systemd-journald.service" to see this information.
16958 * The multi-seat X wrapper tool has been stripped down. As X
16959 is now capable of enumerating graphics devices via udev in a
16960 seat-aware way the wrapper is not strictly necessary
16961 anymore. A stripped down temporary stop-gap is still shipped
16962 until the upstream display managers have been updated to
16963 fully support the new X logic. Expect this wrapper to be
16964 removed entirely in one of the next releases.
16966 * HandleSleepKey= in logind.conf has been split up into
16967 HandleSuspendKey= and HandleHibernateKey=. The old setting
16968 is not available anymore. X11 and the kernel are
16969 distinguishing between these keys and we should too. This
16970 also means the inhibition lock for these keys has been split
16973 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
16974 Poettering, Lukas Nykryn, Václav Pavlín
16978 * Whenever a unit changes state we will now log this to the
16979 journal and show along the unit's own log output in
16980 "systemctl status".
16982 * ConditionPathIsMountPoint= can now properly detect bind
16983 mount points too. (Previously, a bind mount of one file
16984 system to another place in the same file system could not be
16985 detected as mount, since they shared struct stat's st_dev
16988 * We will now mount the cgroup controllers cpu, cpuacct,
16989 cpuset and the controllers net_cls, net_prio together by
16992 * nspawn containers will now have a virtualized boot
16993 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
16994 over with a randomized ID at container initialization). This
16995 has the effect of making "journalctl -b" do the right thing
16998 * The JSON output journal serialization has been updated not
16999 to generate "endless" list objects anymore, but rather one
17000 JSON object per line. This is more in line how most JSON
17001 parsers expect JSON objects. The new output mode
17002 "json-pretty" has been added to provide similar output, but
17003 neatly aligned for readability by humans.
17005 * We dropped all explicit sync() invocations in the shutdown
17006 code. The kernel does this implicitly anyway in the kernel
17007 reboot() syscall. halt(8)'s -n option is now a compatibility
17010 * We now support virtualized reboot() in containers, as
17011 supported by newer kernels. We will fall back to exit() if
17012 CAP_SYS_REBOOT is not available to the container. Also,
17013 nspawn makes use of this now and will actually reboot the
17014 container if the containerized OS asks for that.
17016 * journalctl will only show local log output by default
17017 now. Use --merge (-m) to show remote log output, too.
17019 * libsystemd-journal gained the new sd_journal_get_usage()
17020 call to determine the current disk usage of all journal
17021 files. This is exposed in the new "journalctl --disk-usage"
17024 * journald gained a new configuration setting SplitMode= in
17025 journald.conf which may be used to control how user journals
17026 are split off. See journald.conf(5) for details.
17028 * A new condition type ConditionFileNotEmpty= has been added.
17030 * tmpfiles' "w" lines now support file globbing, to write
17031 multiple files at once.
17033 * We added Python bindings for the journal submission
17034 APIs. More Python APIs for a number of selected APIs will
17035 likely follow. Note that we intend to add native bindings
17036 only for the Python language, as we consider it common
17037 enough to deserve bindings shipped within systemd. There are
17038 various projects outside of systemd that provide bindings
17039 for languages such as PHP or Lua.
17041 * Many conditions will now resolve specifiers such as %i. In
17042 addition, PathChanged= and related directives of .path units
17043 now support specifiers as well.
17045 * There's now a new RPM macro definition for the system preset
17048 * journald will now warn if it ca not forward a message to the
17049 syslog daemon because its socket is full.
17051 * timedated will no longer write or process /etc/timezone,
17052 except on Debian. As we do not support late mounted /usr
17053 anymore /etc/localtime always being a symlink is now safe,
17054 and hence the information in /etc/timezone is not necessary
17057 * logind will now always reserve one VT for a text getty (VT6
17058 by default). Previously if more than 6 X sessions where
17059 started they took up all the VTs with auto-spawned gettys,
17060 so that no text gettys were available anymore.
17062 * udev will now automatically inform the btrfs kernel logic
17063 about btrfs RAID components showing up. This should make
17064 simple hotplug based btrfs RAID assembly work.
17066 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
17067 (but not for its children which will stay at the kernel
17068 default). This should allow setups with a lot more listening
17071 * systemd will now always pass the configured timezone to the
17072 kernel at boot. timedated will do the same when the timezone
17075 * logind's inhibition logic has been updated. By default,
17076 logind will now handle the lid switch, the power and sleep
17077 keys all the time, even in graphical sessions. If DEs want
17078 to handle these events on their own they should take the new
17079 handle-power-key, handle-sleep-key and handle-lid-switch
17080 inhibitors during their runtime. A simple way to achieve
17081 that is to invoke the DE wrapped in an invocation of:
17083 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch …
17085 * Access to unit operations is now checked via SELinux taking
17086 the unit file label and client process label into account.
17088 * systemd will now notify the administrator in the journal
17089 when he over-mounts a non-empty directory.
17091 * There are new specifiers that are resolved in unit files,
17092 for the hostname (%H), the machine ID (%m) and the boot ID
17095 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
17096 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
17097 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
17098 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
17099 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
17100 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
17101 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
17105 * Support for reading structured kernel messages from
17106 /dev/kmsg has now been added and is enabled by default.
17108 * Support for reading kernel messages from /proc/kmsg has now
17109 been removed. If you want kernel messages in the journal
17110 make sure to run a recent kernel (>= 3.5) that supports
17111 reading structured messages from /dev/kmsg (see
17112 above). /proc/kmsg is now exclusive property of classic
17113 syslog daemons again.
17115 * The libudev API gained the new
17116 udev_device_new_from_device_id() call.
17118 * The logic for file system namespace (ReadOnlyDirectory=,
17119 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
17120 require pivot_root() anymore. This means fewer temporary
17121 directories are created below /tmp for this feature.
17123 * nspawn containers will now see and receive all submounts
17124 made on the host OS below the root file system of the
17127 * Forward Secure Sealing is now supported for Journal files,
17128 which provide cryptographical sealing of journal files so
17129 that attackers cannot alter log history anymore without this
17130 being detectable. Lennart will soon post a blog story about
17131 this explaining it in more detail.
17133 * There are two new service settings RestartPreventExitStatus=
17134 and SuccessExitStatus= which allow configuration of exit
17135 status (exit code or signal) which will be excepted from the
17136 restart logic, resp. consider successful.
17138 * journalctl gained the new --verify switch that can be used
17139 to check the integrity of the structure of journal files and
17140 (if Forward Secure Sealing is enabled) the contents of
17143 * nspawn containers will now be run with /dev/stdin, /dev/fd/
17144 and similar symlinks pre-created. This makes running shells
17145 as container init process a lot more fun.
17147 * The fstab support can now handle PARTUUID= and PARTLABEL=
17150 * A new ConditionHost= condition has been added to match
17151 against the hostname (with globs) and machine ID. This is
17152 useful for clusters where a single OS image is used to
17153 provision a large number of hosts which shall run slightly
17154 different sets of services.
17156 * Services which hit the restart limit will now be placed in a
17159 Contributions from: Bertram Poettering, Dave Reisner, Huang
17160 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
17161 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
17165 * When running in --user mode systemd will now become a
17166 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
17167 tree a lot more organized.
17169 * A new PartOf= unit dependency type has been introduced that
17170 may be used to group services in a natural way.
17172 * "systemctl enable" may now be used to enable instances of
17175 * journalctl now prints error log levels in red, and
17176 warning/notice log levels in bright white. It also supports
17177 filtering by log level now.
17179 * cgtop gained a new -n switch (similar to top), to configure
17180 the maximum number of iterations to run for. It also gained
17181 -b, to run in batch mode (accepting no input).
17183 * The suffix ".service" may now be omitted on most systemctl
17184 command lines involving service unit names.
17186 * There's a new bus call in logind to lock all sessions, as
17187 well as a loginctl verb for it "lock-sessions".
17189 * libsystemd-logind.so gained a new call sd_journal_perror()
17190 that works similar to libc perror() but logs to the journal
17191 and encodes structured information about the error number.
17193 * /etc/crypttab entries now understand the new keyfile-size=
17196 * shutdown(8) now can send a (configurable) wall message when
17197 a shutdown is cancelled.
17199 * The mount propagation mode for the root file system will now
17200 default to "shared", which is useful to make containers work
17201 nicely out-of-the-box so that they receive new mounts from
17202 the host. This can be undone locally by running "mount
17203 --make-rprivate /" if needed.
17205 * The prefdm.service file has been removed. Distributions
17206 should maintain this unit downstream if they intend to keep
17207 it around. However, we recommend writing normal unit files
17208 for display managers instead.
17210 * Since systemd is a crucial part of the OS we will now
17211 default to a number of compiler switches that improve
17212 security (hardening) such as read-only relocations, stack
17213 protection, and suchlike.
17215 * The TimeoutSec= setting for services is now split into
17216 TimeoutStartSec= and TimeoutStopSec= to allow configuration
17217 of individual time outs for the start and the stop phase of
17220 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
17221 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
17222 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
17223 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
17224 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
17225 Gundersen, Zbigniew Jędrzejewski-Szmek
17229 * The journal and id128 C APIs are now fully documented as man
17232 * Extra safety checks have been added when transitioning from
17233 the initial RAM disk to the main system to avoid accidental
17236 * /etc/crypttab entries now understand the new keyfile-offset=
17239 * systemctl -t can now be used to filter by unit load state.
17241 * The journal C API gained the new sd_journal_wait() call to
17242 make writing synchronous journal clients easier.
17244 * journalctl gained the new -D switch to show journals from a
17245 specific directory.
17247 * journalctl now displays a special marker between log
17248 messages of two different boots.
17250 * The journal is now explicitly flushed to /var via a service
17251 systemd-journal-flush.service, rather than implicitly simply
17252 by seeing /var/log/journal to be writable.
17254 * journalctl (and the journal C APIs) can now match for much
17255 more complex expressions, with alternatives and
17258 * When transitioning from the initial RAM disk to the main
17259 system we will now kill all processes in a killing spree to
17260 ensure no processes stay around by accident.
17262 * Three new specifiers may be used in unit files: %u, %h, %s
17263 resolve to the user name, user home directory resp. user
17264 shell. This is useful for running systemd user instances.
17266 * We now automatically rotate journal files if their data
17267 object hash table gets a fill level > 75%. We also size the
17268 hash table based on the configured maximum file size. This
17269 together should lower hash collisions drastically and thus
17270 speed things up a bit.
17272 * journalctl gained the new "--header" switch to introspect
17273 header data of journal files.
17275 * A new setting SystemCallFilters= has been added to services which may
17276 be used to apply deny lists or allow lists to system calls. This is
17277 based on SECCOMP Mode 2 of Linux 3.5.
17279 * nspawn gained a new --link-journal= switch (and quicker: -j)
17280 to link the container journal with the host. This makes it
17281 very easy to centralize log viewing on the host for all
17282 guests while still keeping the journal files separated.
17284 * Many bugfixes and optimizations
17286 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
17287 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
17288 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
17293 * Several tools now understand kernel command line arguments,
17294 which are only read when run in an initial RAM disk. They
17295 usually follow closely their normal counterparts, but are
17298 * There's a new tool to analyze the readahead files that are
17299 automatically generated at boot. Use:
17301 /usr/lib/systemd/systemd-readahead analyze /.readahead
17303 * We now provide an early debug shell on tty9 if this enabled. Use:
17305 systemctl enable debug-shell.service
17307 * All plymouth related units have been moved into the Plymouth
17308 package. Please make sure to upgrade your Plymouth version
17311 * systemd-tmpfiles now supports getting passed the basename of
17312 a configuration file only, in which case it will look for it
17313 in all appropriate directories automatically.
17315 * udevadm info now takes a /dev or /sys path as argument, and
17316 does the right thing. Example:
17318 udevadm info /dev/sda
17319 udevadm info /sys/class/block/sda
17321 * systemctl now prints a warning if a unit is stopped but a
17322 unit that might trigger it continues to run. Example: a
17323 service is stopped but the socket that activates it is left
17326 * "systemctl status" will now mention if the log output was
17327 shortened due to rotation since a service has been started.
17329 * The journal API now exposes functions to determine the
17330 "cutoff" times due to rotation.
17332 * journald now understands SIGUSR1 and SIGUSR2 for triggering
17333 immediately flushing of runtime logs to /var if possible,
17334 resp. for triggering immediate rotation of the journal
17337 * It is now considered an error if a service is attempted to
17338 be stopped that is not loaded.
17340 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
17342 * systemd-analyze now supports Python 3
17344 * tmpfiles now supports cleaning up directories via aging
17345 where the first level dirs are always kept around but
17346 directories beneath it automatically aged. This is enabled
17347 by prefixing the age field with '~'.
17349 * Seat objects now expose CanGraphical, CanTTY properties
17350 which is required to deal with very fast bootups where the
17351 display manager might be running before the graphics drivers
17352 completed initialization.
17354 * Seat objects now expose a State property.
17356 * We now include RPM macros for service enabling/disabling
17357 based on the preset logic. We recommend RPM based
17358 distributions to make use of these macros if possible. This
17359 makes it simpler to reuse RPM spec files across
17362 * We now make sure that the collected systemd unit name is
17363 always valid when services log to the journal via
17366 * There's a new man page kernel-command-line(7) detailing all
17367 command line options we understand.
17369 * The fstab generator may now be disabled at boot by passing
17370 fstab=0 on the kernel command line.
17372 * A new kernel command line option modules-load= is now understood
17373 to load a specific kernel module statically, early at boot.
17375 * Unit names specified on the systemctl command line are now
17376 automatically escaped as needed. Also, if file system or
17377 device paths are specified they are automatically turned
17378 into the appropriate mount or device unit names. Example:
17380 systemctl status /home
17381 systemctl status /dev/sda
17383 * The SysVConsole= configuration option has been removed from
17384 system.conf parsing.
17386 * The SysV search path is no longer exported on the D-Bus
17389 * The Names= option has been removed from unit file parsing.
17391 * There's a new man page bootup(7) detailing the boot process.
17393 * Every unit and every generator we ship with systemd now
17394 comes with full documentation. The self-explanatory boot is
17397 * A couple of services gained "systemd-" prefixes in their
17398 name if they wrap systemd code, rather than only external
17399 code. Among them fsck@.service which is now
17400 systemd-fsck@.service.
17402 * The HaveWatchdog property has been removed from the D-Bus
17405 * systemd.confirm_spawn= on the kernel command line should now
17408 * There's a new man page crypttab(5) which details all options
17409 we actually understand.
17411 * systemd-nspawn gained a new --capability= switch to pass
17412 additional capabilities to the container.
17414 * timedated will now read known NTP implementation unit names
17415 from /usr/lib/systemd/ntp-units.d/*.list,
17416 systemd-timedated-ntp.target has been removed.
17418 * journalctl gained a new switch "-b" that lists log data of
17419 the current boot only.
17421 * The notify socket is in the abstract namespace again, in
17422 order to support daemons which chroot() at start-up.
17424 * There is a new Storage= configuration option for journald
17425 which allows configuration of where log data should go. This
17426 also provides a way to disable journal logging entirely, so
17427 that data collected is only forwarded to the console, the
17428 kernel log buffer or another syslog implementation.
17430 * Many bugfixes and optimizations
17432 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
17433 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
17434 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
17435 Shawn Landden, Tom Gundersen
17439 * "systemctl help <unit>" now shows the man page if one is
17442 * Several new man pages have been added.
17444 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
17445 MaxLevelConsole= can now be specified in
17446 journald.conf. These options allow reducing the amount of
17447 data stored on disk or forwarded by the log level.
17449 * TimerSlackNSec= can now be specified in system.conf for
17450 PID1. This allows system-wide power savings.
17452 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
17453 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
17458 * logind is now capable of (optionally) handling power and
17459 sleep keys as well as the lid switch.
17461 * journalctl now understands the syntax "journalctl
17462 /usr/bin/avahi-daemon" to get all log output of a specific
17465 * CapabilityBoundingSet= in system.conf now also influences
17466 the capability bound set of usermode helpers of the kernel.
17468 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
17469 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
17470 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
17471 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
17475 * Note that we skipped 139 releases here in order to set the
17476 new version to something that is greater than both udev's
17477 and systemd's most recent version number.
17479 * udev: all udev sources are merged into the systemd source tree now.
17480 All future udev development will happen in the systemd tree. It
17481 is still fully supported to use the udev daemon and tools without
17482 systemd running, like in initramfs or other init systems. Building
17483 udev though, will require the *build* of the systemd tree, but
17484 udev can be properly *run* without systemd.
17486 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
17487 should be used to create dead device nodes as workarounds for broken
17490 * udev: RUN+="socket:…" and udev_monitor_new_from_socket() is
17491 no longer supported. udev_monitor_new_from_netlink() needs to be
17492 used to subscribe to events.
17494 * udev: when udevd is started by systemd, processes which are left
17495 behind by forking them off of udev rules, are unconditionally cleaned
17496 up and killed now after the event handling has finished. Services or
17497 daemons must be started as systemd services. Services can be
17498 pulled-in by udev to get started, but they can no longer be directly
17499 forked by udev rules.
17501 * udev: the daemon binary is called systemd-udevd now and installed
17502 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
17503 to adapt to that, create symlink, or rename the binary after building
17506 * libudev no longer provides these symbols:
17507 udev_monitor_from_socket()
17508 udev_queue_get_failed_list_entry()
17509 udev_get_{dev,sys,run}_path()
17510 The versions number was bumped and symbol versioning introduced.
17512 * systemd-loginctl and systemd-journalctl have been renamed
17513 to loginctl and journalctl to match systemctl.
17515 * The config files: /etc/systemd/systemd-logind.conf and
17516 /etc/systemd/systemd-journald.conf have been renamed to
17517 logind.conf and journald.conf. Package updates should rename
17518 the files to the new names on upgrade.
17520 * For almost all files the license is now LGPL2.1+, changed
17521 from the previous GPL2.0+. Exceptions are some minor stuff
17522 of udev (which will be changed to LGPL2.1 eventually, too),
17523 and the MIT licensed sd-daemon.[ch] library that is suitable
17524 to be used as drop-in files.
17526 * systemd and logind now handle system sleep states, in
17527 particular suspending and hibernating.
17529 * logind now implements a sleep/shutdown/idle inhibiting logic
17530 suitable for a variety of uses. Soonishly Lennart will blog
17531 about this in more detail.
17533 * var-run.mount and var-lock.mount are no longer provided
17534 (which previously bind mounted these directories to their new
17535 places). Distributions which have not converted these
17536 directories to symlinks should consider stealing these files
17537 from git history and add them downstream.
17539 * We introduced the Documentation= field for units and added
17540 this to all our shipped units. This is useful to make it
17541 easier to explore the boot and the purpose of the various
17544 * All smaller setup units (such as
17545 systemd-vconsole-setup.service) now detect properly if they
17546 are run in a container and are skipped when
17547 appropriate. This guarantees an entirely noise-free boot in
17548 Linux container environments such as systemd-nspawn.
17550 * A framework for implementing offline system updates is now
17551 integrated, for details see:
17552 https://www.freedesktop.org/software/systemd/man/systemd.offline-updates.html
17554 * A new service type Type=idle is available now which helps us
17555 avoiding ugly interleaving of getty output and boot status
17558 * There's now a system-wide CapabilityBoundingSet= option to
17559 globally reduce the set of capabilities for the
17560 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
17561 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
17562 even CAP_NET_ADMIN system-wide for secure systems.
17564 * There are now system-wide DefaultLimitXXX= options to
17565 globally change the defaults of the various resource limits
17566 for all units started by PID 1.
17568 * Harald Hoyer's systemd test suite has been integrated into
17569 systemd which allows easy testing of systemd builds in qemu
17570 and nspawn. (This is really awesome! Ask us for details!)
17572 * The fstab parser is now implemented as generator, not inside
17575 * systemctl will now warn you if .mount units generated from
17576 /etc/fstab are out of date due to changes in fstab that
17577 have not been read by systemd yet.
17579 * systemd is now suitable for usage in initrds. Dracut has
17580 already been updated to make use of this. With this in place
17581 initrds get a slight bit faster but primarily are much
17582 easier to introspect and debug since "systemctl status" in
17583 the host system can be used to introspect initrd services,
17584 and the journal from the initrd is kept around too.
17586 * systemd-delta has been added, a tool to explore differences
17587 between user/admin configuration and vendor defaults.
17589 * PrivateTmp= now affects both /tmp and /var/tmp.
17591 * Boot time status messages are now much prettier and feature
17592 proper english language. Booting up systemd has never been
17595 * Read-ahead pack files now include the inode number of all
17596 files to pre-cache. When the inode changes the pre-caching
17597 is not attempted. This should be nicer to deal with updated
17598 packages which might result in changes of read-ahead
17601 * We now temporaritly lower the kernel's read_ahead_kb variable
17602 when collecting read-ahead data to ensure the kernel's
17603 built-in read-ahead does not add noise to our measurements
17604 of necessary blocks to pre-cache.
17606 * There's now RequiresMountsFor= to add automatic dependencies
17607 for all mounts necessary for a specific file system path.
17609 * MountAuto= and SwapAuto= have been removed from
17610 system.conf. Mounting file systems at boot has to take place
17613 * nspawn now learned a new switch --uuid= to set the machine
17614 ID on the command line.
17616 * nspawn now learned the -b switch to automatically search
17617 for an init system.
17619 * vt102 is now the default TERM for serial TTYs, upgraded from
17622 * systemd-logind now works on VT-less systems.
17624 * The build tree has been reorganized. The individual
17625 components now have directories of their own.
17627 * A new condition type ConditionPathIsReadWrite= is now available.
17629 * nspawn learned the new -C switch to create cgroups for the
17630 container in other hierarchies.
17632 * We now have support for hardware watchdogs, configurable in
17635 * The scheduled shutdown logic now has a public API.
17637 * We now mount /tmp as tmpfs by default, but this can be
17638 masked and /etc/fstab can override it.
17640 * Since udisks does not make use of /media anymore we are not
17641 mounting a tmpfs on it anymore.
17643 * journalctl gained a new --local switch to only interleave
17644 locally generated journal files.
17646 * We can now load the IMA policy at boot automatically.
17648 * The GTK tools have been split off into a systemd-ui.
17650 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
17651 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
17652 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
17653 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
17654 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
17655 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
17656 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
17657 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
17658 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
17663 * This is mostly a bugfix release
17665 * Support optional initialization of the machine ID from the
17666 KVM or container configured UUID.
17668 * Support immediate reboots with "systemctl reboot -ff"
17670 * Show /etc/os-release data in systemd-analyze output
17672 * Many bugfixes for the journal, including endianness fixes and
17673 ensuring that disk space enforcement works
17675 * sd-login.h is C++ compatible again
17677 * Extend the /etc/os-release format on request of the Debian
17680 * We now refuse non-UTF8 strings used in various configuration
17681 and unit files. This is done to ensure we do not pass invalid
17682 data over D-Bus or expose it elsewhere.
17684 * Register Mimo USB Screens as suitable for automatic seat
17687 * Read SELinux client context from journal clients in a race
17690 * Reorder configuration file lookup order. /etc now always
17691 overrides /run in order to allow the administrator to always
17692 and unconditionally override vendor-supplied or
17693 automatically generated data.
17695 * The various user visible bits of the journal now have man
17696 pages. We still lack man pages for the journal API calls
17699 * We now ship all man pages in HTML format again in the
17702 Contributions from: Dave Reisner, Dirk Eibach, Frederic
17703 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
17704 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
17709 * This is mostly a bugfix release
17711 * systems lacking /etc/os-release are no longer supported.
17713 * Various functionality updates to libsystemd-login.so
17715 * Track class of PAM logins to distinguish greeters from
17716 normal user logins.
17718 Contributions from: Kay Sievers, Lennart Poettering, Michael
17723 * This is an important bugfix release for v41.
17725 * Building man pages is now optional which should be useful
17726 for those building systemd from git but unwilling to install
17729 * Watchdog support for supervising services is now usable. In
17730 a future release support for hardware watchdogs
17731 (i.e. /dev/watchdog) will be added building on this.
17733 * Service start rate limiting is now configurable and can be
17734 turned off per service. When a start rate limit is hit a
17735 reboot can automatically be triggered.
17737 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
17739 Contributions from: Benjamin Franzke, Bill Nottingham,
17740 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
17741 Schmidt, Michał Górny, Piotr Drąg
17745 * The systemd binary is installed /usr/lib/systemd/systemd now;
17746 An existing /sbin/init symlink needs to be adapted with the
17749 * The code that loads kernel modules has been ported to invoke
17750 libkmod directly, instead of modprobe. This means we do not
17751 support systems with module-init-tools anymore.
17753 * Watchdog support is now already useful, but still not
17756 * A new kernel command line option systemd.setenv= is
17757 understood to set system wide environment variables
17758 dynamically at boot.
17760 * We now limit the set of capabilities of systemd-journald.
17762 * We now set SIGPIPE to ignore by default, since it only is
17763 useful in shell pipelines, and has little use in general
17764 code. This can be disabled with IgnoreSIPIPE=no in unit
17767 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
17768 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
17773 * This is mostly a bugfix release
17775 * We now expose the reason why a service failed in the
17776 "Result" D-Bus property.
17778 * Rudimentary service watchdog support (will be completed over
17779 the next few releases.)
17781 * When systemd forks off in order execute some service we will
17782 now immediately changes its argv[0] to reflect which process
17783 it will execute. This is useful to minimize the time window
17784 with a generic argv[0], which makes bootcharts more useful
17786 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
17787 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
17788 Mike Kazantsev, Ray Strode
17792 * This is mostly a test release, but incorporates many
17795 * New systemd-cgtop tool to show control groups by their
17798 * Linking against libacl for ACLs is optional again. If
17799 disabled, support tracking device access for active logins
17800 goes becomes unavailable, and so does access to the user
17801 journals by the respective users.
17803 * If a group "adm" exists, journal files are automatically
17804 owned by them, thus allow members of this group full access
17805 to the system journal as well as all user journals.
17807 * The journal now stores the SELinux context of the logging
17808 client for all entries.
17810 * Add C++ inclusion guards to all public headers
17812 * New output mode "cat" in the journal to print only text
17813 messages, without any meta data like date or time.
17815 * Include tiny X server wrapper as a temporary stop-gap to
17816 teach XOrg udev display enumeration. This is used by display
17817 managers such as gdm, and will go away as soon as XOrg
17818 learned native udev hotplugging for display devices.
17820 * Add new systemd-cat tool for executing arbitrary programs
17821 with STDERR/STDOUT connected to the journal. Can also act as
17822 BSD logger replacement, and does so by default.
17824 * Optionally store all locally generated coredumps in the
17825 journal along with meta data.
17827 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
17828 writing short strings to files (for usage for /sys), and for
17829 creating symlinks, character and block device nodes.
17831 * New unit file option ControlGroupPersistent= to make cgroups
17832 persistent, following the mechanisms outlined in
17833 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
17835 * Support multiple local RTCs in a sane way
17837 * No longer monopolize IO when replaying readahead data on
17838 rotating disks, since we might starve non-file-system IO to
17839 death, since fanotify() will not see accesses done by blkid,
17842 * Do not show kernel threads in systemd-cgls anymore, unless
17843 requested with new -k switch.
17845 Contributions from: Dan Horák, Kay Sievers, Lennart
17846 Poettering, Michal Schmidt
17850 * This is mostly a test release, but incorporates many
17853 * The git repository moved to:
17854 git://anongit.freedesktop.org/systemd/systemd
17855 ssh://git.freedesktop.org/git/systemd/systemd
17857 * First release with the journal
17858 https://0pointer.de/blog/projects/the-journal.html
17860 * The journal replaces both systemd-kmsg-syslogd and
17861 systemd-stdout-bridge.
17863 * New sd_pid_get_unit() API call in libsystemd-logind
17865 * Many systemadm clean-ups
17867 * Introduce remote-fs-pre.target which is ordered before all
17868 remote mounts and may be used to start services before all
17871 * Added Mageia support
17873 * Add bash completion for systemd-loginctl
17875 * Actively monitor PID file creation for daemons which exit in
17876 the parent process before having finished writing the PID
17877 file in the daemon process. Daemons which do this need to be
17878 fixed (i.e. PID file creation must have finished before the
17879 parent exits), but we now react a bit more gracefully to them.
17881 * Add colourful boot output, mimicking the well-known output
17882 of existing distributions.
17884 * New option PassCredentials= for socket units, for
17885 compatibility with a recent kernel ABI breakage.
17887 * /etc/rc.local is now hooked in via a generator binary, and
17888 thus will no longer act as synchronization point during
17891 * systemctl list-unit-files now supports --root=.
17893 * systemd-tmpfiles now understands two new commands: z, Z for
17894 relabelling files according to the SELinux database. This is
17895 useful to apply SELinux labels to specific files in /sys,
17896 among other things.
17898 * Output of SysV services is now forwarded to both the console
17899 and the journal by default, not only just the console.
17901 * New man pages for all APIs from libsystemd-login.
17903 * The build tree got reorganized and the build system is a
17904 lot more modular allowing embedded setups to specifically
17905 select the components of systemd they are interested in.
17907 * Support for Linux systems lacking the kernel VT subsystem is
17910 * configure's --with-rootdir= got renamed to
17911 --with-rootprefix= to follow the naming used by udev and
17914 * Unless specified otherwise we will now install to /usr instead
17915 of /usr/local by default.
17917 * Processes with '@' in argv[0][0] are now excluded from the
17918 final shut-down killing spree, following the logic explained
17920 https://systemd.io/ROOT_STORAGE_DAEMONS/
17922 * All processes remaining in a service cgroup when we enter
17923 the START or START_PRE states are now killed with
17924 SIGKILL. That means it is no longer possible to spawn
17925 background processes from ExecStart= lines (which was never
17926 supported anyway, and bad style).
17928 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
17929 reloading of units together.
17931 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
17932 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
17933 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
17934 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
17935 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek