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29 setup is obsolete but still commonly found).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, and the
76 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 may 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, cancel-transfer. This functionality was
89 previously available in "machinectl", where it was exclusively for
90 machine image. The new "importctl" generalizes this for sysext,
91 confext, portable service images, too.
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 with a Wants= dependency instead of
101 Requires=. This new logic is used in various places where mounts were
102 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 even if the original service unit
111 is restarted. Control group properties of the new scope are copied
112 from the originating unit, so various limits are retained.
114 * Units now expose properties EffectiveMemoryMax=,
115 EffectiveMemoryHigh=, and EffectiveTasksMax=, which report the
116 most stringent limit systemd is aware of for the given unit.
118 * A new unit file specifier %D expands to $XDG_DATA_HOME (for user
119 services) or /usr/share/ (for system services).
121 * AllowedCPUs= now supports specifier expansion.
123 * What= setting in .mount and .swap units now accepts fstab-style
124 identifiers, for example UUID=… or LABEL=….
126 * RestrictNetworkInterfaces= now supports alternative network interface
129 * PAMName= now implies SetLoginEnvironment=yes.
131 * systemd.firstboot=no can be used on the kernel command-line to
132 disable interactive queries, but allow other first boot configuration
133 to happen based on credentials.
135 * The system's hostname can be configured via the systemd.hostname
138 * The systemd binary will no longer chainload sysvinit's "telinit"
139 binary when called under the init/telinit name on a system that
140 isn't booted with systemd. This previously has been supported to make
141 sure a distribution that has both init systems installed can be
142 reasonably switched from one to the other via a simple
143 reboot. Distributions apparently have lost interest in this, and the
144 functionality has not been supported on the primary distribution this
145 was still intended for for a longer time, and hence has been removed
148 * A new concept called "capsules" has been introduced. "Capsules"
149 encapsulate additional per-user service managers, whose users are
150 transient and are only defined as long as the service manager
151 is running (implemented via DynamicUser=1). These service managers run
152 off home directories defined in /var/lib/capsules/<name>, where
153 <name> is a the capsule's name. These home directories can contain
154 regular per-user services and other units. A capsule is started via a
155 simple "systemctl start capsule@<name>.service". See the
156 capsule@.service(5) man page for further details. Various systemd
157 tools (including, and most importantly, systemctl and systemd-run)
158 have been updated to interact with capsules via the new
159 "--capsule="/"-C" switch.
161 * .socket units gained a new setting PassFileDescriptorsToExec=, taking
162 a boolean value. If set to true the file descriptors the socket unit
163 encapsulates are passed to the ExecStartPost=, ExecStopPre=,
164 ExecStopPost= using the usual $LISTEN_FDS interface. This may be used
165 for doing additional initializations on the sockets once they are
166 allocated (for example, install an additional eBPF program on them).
168 * The .socket setting MaxConnectionsPerSource= (which so far put a
169 limit on concurrent connections per IP in Accept=yes socket units),
170 now also has an effect on AF_UNIX sockets: it will put a limit on the
171 number of simultaneous connections from the same source UID (as
172 determined via SO_PEERCRED). This is useful for implementing IPC
173 services in a simple Accept=yes mode.
175 * The service manager will now maintain a counter of soft reboot cycles
176 the system went through so far. It may be queried via the D-Bus APIs.
178 * systemd's execution logic now supports the new pidfd_spawn() API
179 introduced by glibc 2.39, which allows us to invoke a subprocess in a
180 target cgroup and get a pidfd back in a single operation.
182 * systemd/PID 1 will now send an additional sd_notify() message to its
183 supervising VMM or container manager reporting the selected hostname
184 ("X_SYSTEMD_HOSTNAME=") and machine ID ("X_SYSTEMD_MACHINE_ID=") at
185 boot. Moreover, the service manager will send additional sd_notify()
186 messages ("X_SYSTEMD_UNIT_ACTIVE=") whenever a target unit is
187 reached. This can be used by VMMs/container managers to schedule
188 access to the system precisely. For example, the moment a system
189 reports "ssh-access.target" being reached a VMM/container manager
190 knows it can now connect to the system via SSH. Finally, a new
191 sd_notify() message ("X_SYSTEMD_SIGNALS_LEVEL=2") is sent the moment
192 PID 1 successfully completed installation of its various UNIX process
193 signal handlers (i.e. the moment where SIGRTMIN+4 sent to PID 1 will
194 start to have the effect of shutting down the system cleanly).
198 * systemd-journald can now forward journal entries to a socket
199 (AF_INET, AF_INET6, AF_UNIX, or AF_VSOCK). The socket can be
200 specified in journald.conf via a new option ForwardAddress= or via
201 the 'journald.forward_address' credential. Log records are sent in
202 the Journal Export Format. A related setting MaxLevelSocket= has been
203 added to control the maximum log levels for the messages sent to this
206 * systemd-vmspawn gained a new --forward-journal= option to forward the
207 virtual machine's journal entries to the host. This is done over a
208 AF_VSOCK socket, i.e. it does not require networking in the guest.
210 * journalctl gained option '-i' as a shortcut for --file=.
212 * journalctl gained a new -T/--exclude-identifier= option to filter
213 out certain syslog identifiers.
215 * journalctl gained a new --list-namespaces option.
217 * systemd-journal-remote now also accepts AF_VSOCK and AF_UNIX sockets
218 (so it can be used to receive entries forwarded by systemd-journald).
220 * systemd-journal-gatewayd allows restricting the time range of
221 retrieved entries with a new "realtime=[<since>]:[<until>]" URL
224 * systemd-cat gained a new option --namespace= to specify the target
225 journal namespace to which the output shall be connected.
227 * systemd-bsod gained a new option --tty= to specify the output TTY
231 * /dev/ now contains symlinks that combine by-path and by-{label,uuid}
234 /dev/disk/by-path/<path>/by-<label|uuid|…>/<label|uuid|…>
236 This allows distinguishing partitions with identical contents on
237 multiple storage devices. This is useful, for example, when copying
238 raw disk contents between devices.
240 * systemd-udevd now creates persistent /dev/media/by-path/ symlinks for
241 media controllers. For example, the uvcvideo driver may create
242 /dev/media0 which will be linked as
243 /dev/media/by-path/pci-0000:04:00.3-usb-0:1:1.0-media-controller.
245 * A new unit systemd-udev-load-credentials.service has been added
246 to pick up udev.conf drop-ins and udev rules from credentials.
248 * An allowlist/denylist may be specified to filter which sysfs
249 attributes are used when crafting network interface names. Those
250 lists are stored as hwdb entries
251 ID_NET_NAME_ALLOW_<sysfsattr>=0|1
253 ID_NET_NAME_ALLOW=0|1.
255 The goal is to avoid unexpected changes to interface names when the
256 kernel is updated and new sysfs attributes become visible.
258 * A new unit tpm2.target has been added to provide a synchronization
259 point for units which expect the TPM hardware to be available. A new
260 generator "systemd-tpm2-generator" has been added that will insert
261 this target whenever it detects that the firmware has initialized a
262 TPM, but Linux hasn't loaded a driver for it yet.
264 * systemd-backlight now properly supports numbered devices which the
265 kernel creates to avoid collisions in the leds subsystem.
267 * systemd-hwdb update operation can be disabled with a new environment
268 variable SYSTEMD_HWDB_UPDATE_BYPASS=1.
272 * systemd-hostnamed now exposes the machine ID and boot ID via
273 D-Bus. It also exposes the hosts AF_VSOCK CID, if available.
275 * systemd-hostnamed now provides a basic Varlink interface.
277 * systemd-hostnamed exports the full data in os-release(5) and
278 machine-info(5) via D-Bus and Varlink.
280 * hostnamectl now shows the system's product UUID and hardware serial
285 * systemd-networkd now provides a basic Varlink interface.
287 * systemd-networkd's ARP proxy support gained a new option to configure
288 a private VLAN variant of the proxy ARP supported by the kernel under
289 the name IPv4ProxyARPPrivateVLAN=.
291 * systemd-networkd now exports the NamespaceId and NamespaceNSID
292 properties via D-Bus and Varlink. (which expose the inode and NSID of
293 the network namespace the networkd instance manages)
295 * systemd-networkd now supports IPv6RetransmissionTimeSec= and
296 UseRetransmissionTime= settings in .network files to configure
297 retransmission time for IPv6 neighbor solicitation messages.
299 * networkctl gained new verbs 'mask' and 'unmask' for masking networkd
300 configuration files such as .network files.
302 * 'networkctl edit --runtime' allows editing volatile configuration
303 under /run/systemd/network/.
305 * The implementation behind TTLPropagate= network setting has been
306 removed and the setting is now ignored.
308 * systemd-network-generator will now pick up .netdev/.link/.network/
309 networkd.conf configuration from system credentials.
311 * systemd-networkd will now pick up wireguard secrets from
314 * systemd-networkd's Varlink API now supports enumerating LLDP peers.
316 * .link files now support new Property=, ImportProperty=,
317 UnsetProperty= fields for setting udev properties on a link.
319 * The various .link files that systemd ships for interfaces that are
320 supposed to be managed by systemd-networkd only now carry a
321 ID_NET_MANAGED_BY=io.systemd.Network udev property ensuring that
322 other network management solutions honouring this udev property do
323 not come into conflict with networkd, trying to manage these
326 * .link files now support a new ReceivePacketSteeringCPUMask= setting
327 for configuring which CPUs to steer incoming packets to.
331 * systemd-nspawn now provides a /run/systemd/nspawn/unix-export/
332 directory where the container payload can expose AF_UNIX sockets to
333 allow them them to be accessed from outside.
335 * systemd-nspawn will tint the terminal background for containers in a
336 blueish color. This can be controller with the new --background=
339 * systemd-nspawn gained support for the 'owneridmap' option for --bind=
340 mounts to map the target directory owner from inside the container to
341 the owner of the directory bound from the host filesystem.
343 * systemd-nspawn now supports moving Wi-Fi network devices into a
344 container, just like other network interfaces.
348 * systemd-resolved now reads RFC 8914 EDE error codes provided by
349 upstream DNS services.
351 * systemd-resolved and resolvectl now support RFC 9460 SVCB and HTTPS
352 records, as well as RFC 2915 NAPTR records.
354 * resolvectl gained a new option --relax-single-label= to allow
355 querying single-label hostnames via unicast DNS on a per-query basis.
357 * systemd-resolved's Varlink IPC interface now supports resolving
358 DNS-SD services as well as an API for resolving raw DNS RRs.
360 * systemd-resolved's .dnssd DNS_SD service description files now
361 support DNS-SD "subtypes" via the new SubType= setting.
363 * systemd-resolved's configuration may now be reloaded without
364 restarting the service. (i.e. "systemctl reload systemd-resolved" is
369 * An sshd config drop-in to allow ssh keys acquired via userdbctl (for
370 example expose by homed accounts) to be used for authorization of
371 incoming SSH connections.
373 * A small new unit generator "systemd-ssh-generator" has been added. It
374 checks if the sshd binary is installed. If so, it binds it via
375 per-connection socket activation to various sockets depending on the
378 • If the system is run in a VM providing AF_VSOCK support, it
379 automatically binds sshd to AF_VSOCK port 22.
381 • If the system is invoked as a full-OS container and the container
382 manager pre-mounts a directory /run/host/unix-export/, it will
383 bind sshd to an AF_UNIX socket /run/host/unix-export/ssh. The
384 idea is the container manager bind mounts the directory to an
385 appropriate place on the host as well, so that the AF_UNIX socket
386 may be used to easily connect from the host to the container.
388 • sshd is also bound to an AF_UNIX socket
389 /run/ssh-unix-local/socket, which may be to use ssh/sftp in a
390 "sudo"-like fashion to access resources of other local users.
392 • Via the kernel command line option "systemd.ssh_listen=" and the
393 system credential "ssh.listen" sshd may be bound to additional,
394 explicitly configured options, including AF_INET/AF_INET6 ports.
396 In particular the first two mechanisms should make dealing with local
397 VMs and full OS containers a lot easier, as SSH connections will
398 *just* *work* from the host – even if no networking is available
401 systemd-ssh-generator optionally generates a per-connection
402 socket activation service file wrapping sshd. This is only done if
403 the distribution does not provide one on its own under the name
404 "sshd@.service". The generated unit only works correctly if the SSH
405 privilege separation ("privsep") directory exists. Unfortunately
406 distributions vary wildly where they place this directory. An
407 incomprehensive list:
409 • /usr/share/empty.sshd/ (new fedora)
412 • /run/sshd/ (debian/ubuntu?)
414 If the SSH privsep directory is placed below /var/ or /run/ care
415 needs to be taken that the directory is created automatically at boot
416 if needed, since these directories possibly or always come up
417 empty. This can be done via a tmpfiles.d/ drop-in. You may use the
418 "sshdprivsepdir" meson option provided by systemd to configure the
419 directory, in case you want systemd to create the directory as needed
420 automatically, if your distribution does not cover this natively.
422 Recommendations to distributions, in order to make things just work:
424 • Please provide a per-connection SSH service file under the name
427 • Please move the SSH privsep dir into /usr/ (so that it is truly
428 immutable on image-based operating systems, is strictly under
429 package manager control, and never requires recreation if the
430 system boots up with an empty /run/ or /var/).
432 • As an extension of this: please consider following Fedora's lead
433 here, and use /usr/share/empty.sshd/ to minimize needless
434 differences between distributions.
436 • If your distribution insists on placing the directory in /var/ or
437 /run/ then please at least provide a tmpfiles.d/ drop-in to
438 recreate it automatically at boot, so that the sshd binary just
439 works, regardless in which context it is called.
441 * A small tool "systemd-ssh-proxy" has been added, which is supposed to
442 act as counterpart to "systemd-ssh-generator". It's a small plug-in
443 for the SSH client (via ProxyCommand/ProxyUseFdpass) to allow it to
444 connect to AF_VSOCK or AF_UNIX sockets. Example: "ssh vsock/4711"
445 connects to a local VM with cid 4711, or "ssh
446 unix/run/ssh-unix-local/socket" to connect to the local host via the
447 AF_UNIX socket /run/ssh-unix-local/socket.
449 systemd-boot and systemd-stub and Related Tools:
451 * TPM 1.2 PCR measurement support has been removed from systemd-stub.
452 TPM 1.2 is obsolete and – due to the (by today's standards) weak
453 cryptographic algorithms it only supports – does not actually provide
454 the security benefits it's supposed to provide. Given that the rest
455 of systemd's codebase never supported TPM 1.2, the support has now
456 been removed from systemd-stub as well.
458 * systemd-stub will now measure its payload via the new EFI
459 Confidential Computing APIs (CC), in addition to the pre-existing
462 * confexts are loaded by systemd-stub from the ESP as well.
464 * The pcrlock policy is saved in an unencrypted credential file
465 "pcrlock.<entry-token>.cred" under XBOOTLDR/ESP in the
466 /loader/credentials/ directory. It will be picked up at boot by
467 systemd-stub and passed to the initrd, where it can be used to unlock
468 the root file system.
470 * kernel-install gained support for --root= for the 'list' verb.
472 * systemd-pcrlock gained an --entry-token= option to configure the
475 * systemd-pcrlock now provides a basic Varlink interface and can be run
476 as a daemon via a template unit.
478 * bootctl now provides a basic Varlink interface and can be run as a
479 daemon via a template unit.
481 * systemd-measure gained new options --certificate=, --private-key=,
482 and --private-key-source= to allow using OpenSSL's "engines" or
483 "providers" as the signing mechanism to use when creating signed
484 TPM2 PCR measurement values.
486 * ukify gained support for signing of PCR signatures via OpenSSL's
487 engines and providers.
489 * ukify now supports zboot kernels.
491 * systemd-boot now supports passing additional kernel command line
492 switches to invoked kernels via an SMBIOS Type #11 string
493 "io.systemd.boot.kernel-cmdline-extra". This is similar to the
494 pre-existing support for this in systemd-stub, but also applies to
495 Type #1 Boot Loader Specification Entries.
497 * systemd-boot's automatic SecureBoot enrollment support gained support
498 for enrolling "dbx" too (Previously, only db/KEK/PK enrollment was
499 supported). It also now supports UEFI "Custom" mode.
503 * systemd-run is now a multi-call binary. When invoked as 'run0', it
504 provides as interface similar to 'sudo', with all arguments starting
505 at the first non-option parameter being treated the command to invoke
506 as root. Unlike 'sudo' and similar tools, it does not make use of
507 setuid binaries or other privilege escalation methods, but instead
508 runs the specified command as a transient unit, which is started by
509 the system service manager, so privileges are dropped, rather than
510 gained, thus implementing a much more robust and safe security
511 model. As usual, authorization is managed via Polkit.
513 * systemd-run/run0 will now tint the terminal background on supported
514 terminals: in a reddish tone when invoking a root service, in a
515 yellowish tone otherwise. This may be controlled and turned off via
516 the new --background= switch.
518 * systemd-run gained a new option '--ignore-failure' to suppress
523 * 'systemctl edit --stdin' allows creation of unit files and drop-ins
524 with contents supplied via standard input. This is useful when creating
525 configuration programmatically; the tool takes care of figuring out
526 the file name, creating any directories, and reloading the manager
529 * 'systemctl disable --now' and 'systemctl mask --now' now work
530 correctly with template units.
532 * 'systemd-analyze architectures' lists known CPU architectures.
534 * 'systemd-analyze --json=…' is supported for 'architectures',
535 'capability', 'exit-status'.
537 * 'systemd-tmpfiles --purge' will purge (remove) all files and
538 directories created via tmpfiles.d configuration.
540 * systemd-id128 gained new options --no-pager, --no-legend, and
543 * hostnamectl gained '-j' as shortcut for '--json=pretty' or
546 * loginctl now supports -j/--json=.
548 * resolvectl now supports -j/--json= for --type=.
550 * systemd-tmpfiles gained a new option --dry-run to print what would be
551 done without actually taking action.
553 * varlinkctl gained a new --collect switch to collect all responses of
554 a method call that supports multiple replies and turns it into a
557 * systemd-dissect gained a new --make-archive option to generate an
558 archive file (tar.gz and similar) from a disk image.
562 * systemd-vmspawn gained a new --firmware= option to configure or list
563 firmware definitions for Qemu, a new --tpm= option to enable or
564 disable the use of a software TPM, a new --linux= option to specify a
565 kernel binary for direct kernel boot, a new --initrd= option to
566 specify an initrd for direct kernel boot, a new -D/--directory option
567 to use a plain directory as the root file system, a new
568 --private-users option similar to the one in systemd-nspawn, new
569 options --bind= and --bind-ro= to bind part of the host's file system
570 hierarchy into the guest, a new --extra-drive= option to attach
571 additional storage, and -n/--network-tap/--network-user-mode to
572 configure networking.
574 * A new systemd-vmspawn@.service can be used to launch systemd-vmspawn
577 * systemd-vmspawn gained the new --console= and --background= switches
578 that control how to interact with the VM. As before, by default an
579 interactive terminal interface is provided, but now with a background
580 tinted with a greenish hue.
582 * systemd-vmspawn can now register its VMs with systemd-machined,
583 controlled via the --register= switch.
585 * machinectl's start command (and related) can now invoke images either
586 as containers via `systemd-nspawn` (switch is --runner=nspawn, the
587 default) or as VMs via `systemd-vmspawn` (switch is --runner=vmspawn,
590 * systemd-vmspawn now supports two switches --pass-ssh-key= and
591 --ssh-key-type= to optionally set up transient SSH keys to pass to the
592 invoked VMs in order to be able to SSH into them once booted.
596 * systemd-repart gained new options --generate-fstab= and
597 --generate-crypttab= to write out fstab and crypttab files matching the
598 generated partitions.
600 * systemd-repart gained a new option --private-key-source= to allow
601 using OpenSSL's "engines" or "providers" as the signing mechanism to
602 use when creating verity signature partitions.
604 * systemd-repart gained a new DefaultSubvolume= setting in repart.d/
605 drop-ins that allow configuring the default btrfs subvolume for newly
606 formatted btrfs file systems.
610 * libsystemd gained new call sd_bus_creds_new_from_pidfd() to get a
611 credentials object for a pidfd and sd_bus_creds_get_pidfd_dup() to
612 retrieve the pidfd from a credentials object.
614 * sd-bus' credentials logic will now also acquire peer's UNIX group
615 lists and peer's pidfd if supported and requested.
617 * RPM macro %_kernel_install_dir has been added with the path
618 to the directory for kernel-install plugins.
620 * The liblz4, libzstd, liblzma, libkmod, libgcrypt dependencies have
621 been changed from regular shared library dependencies into dlopen()
624 * The sd-journal API gained a new call
625 sd_journal_stream_fd_with_namespace() which is just like
626 sd_journal_stream_fd() but creates a log stream targeted at a
627 specific log namespace.
629 systemd-cryptsetup/systemd-cryptenroll:
631 * systemd-cryptenroll can now enroll directly with a PKCS11 public key
632 (instead of a certificate).
634 * systemd-cryptsetup/systemd-cryptenroll now may lock a disk against a
635 PKCS#11 provided EC key (before it only supported RSA).
637 * systemd-cryptsetup gained support for crypttab option
638 link-volume-key= to link the volume key into the kernel keyring when
639 the volume is opened.
641 * systemd-cryptenroll will no longer enable Dictionary Attack
642 Protection (i.e. turn on NO_DA) for TPM enrollments that do not
643 involve a PIN. DA should not be necessary in that case (since key
644 entropy is high enough to make this unnecessary), but risks
645 accidental lock-out in case of unexpected PCR changes.
647 * systemd-cryptenroll now supports enrolling a new slot while unlocking
648 the old slot via TPM2 (previously unlocking only worked via password
653 * The remaining documentation that was on
654 https://freedesktop.org/wiki/Software/systemd/ has been moved to
657 * A new text describing the VM integration interfaces of systemd has
660 https://systemd.io/VM_INTERFACE
662 * The sd_notify() man page has gained an example with C code that shows
663 how to implement the interface in C without involving libsystemd.
665 systemd-homed, systemd-logind, systemd-userdbd:
667 * systemd-homed now supports unlocking of home directories when logging
668 in via SSH. Previously home directories needed to be unlocked before
669 an SSH login is attempted.
671 * JSON User Records have been extended with a separate public storage
672 area called "User Record Blob Directories". This is intended to store
673 the user's background image, avatar picture, and other similar items
674 which are too large to fit into the User Record itself.
676 systemd-homed, userdbctl, and homectl gained support for blob
677 directories. homectl gained --avatar= and --login-background= to
678 control two specific items of the blob directories.
680 * A new "additionalLanguages" field has been added to JSON user records
681 (as supported by systemd-homed and systemd-userdbd), which is closely
682 related to the pre-existing "preferredLanguage", and allows
683 specifying multiple additional languages for the user account. It is
684 used to initialize the $LANGUAGES environment variable when used.
686 * A new pair of "preferredSessionType" and "preferredSessionLauncher"
687 fields have been added to JSON user records, that may be used to
688 control which kind of desktop session to preferable activate on
691 * homectl gained a new verb 'firstboot', and a new
692 systemd-homed-firstboot.service unit uses this verb to create users
693 in a first boot environment, either from system credentials or by
694 querying interactively.
696 * systemd-logind now supports a new "background-light" session class
697 which does not pull in the user@.service unit. This is intended in
698 particular for lighter weight per-user cron jobs which do require any
699 per-user service manager to be around.
701 * The per-user service manager will now be tracked as a distinct "manager"
702 session type among logind sessions of each user.
704 * homectl now supports an --offline mode, by which certain account
705 properties can be changed without unlocking the home directory.
707 * systemd-logind gained a new
708 org.freedesktop.login1.Manager.ListSessionsEx() method that provides
709 additional metadata compared to ListSessions(). loginctl makes use of
710 this to list additional fields in list-sessions.
712 * systemd-logind gained a new org.freedesktop.login1.Manager.Sleep()
713 method that automatically redirects to SuspendThenHibernate(),
714 Suspend(), HybridSleep(), or Hibernate(), depending on what is
715 supported and configured, a new configuration setting SleepOperation=,
716 and an accompanying helper method
717 org.freedesktop.login1.Manager.CanSleep() and property
718 org.freedesktop.login1.Manager.SleepOperation.
720 'systemctl sleep' calls the new method to automatically put the
721 machine to sleep in the most appropriate way.
723 Credential Management:
725 * systemd-creds now provides a Varlink IPC API for encrypting and
726 decrypting credentials.
728 * systemd-creds' "tpm2-absent" key selection has been renamed to
729 "null", since that's what it actually does: "encrypt" and "sign"
730 with a fixed null key. --with-key=null should only be used in very
731 specific cases, as it provides zero integrity or confidentiality
732 protections. (i.e. it's only safe to use as fallback in environments
733 lacking both a TPM and access to the root fs to use the host
734 encryption key, or when integrity is provided some other way.)
736 * systemd-creds gained a new switch --allow-null. If specified, the
737 "decrypt" verb will decode encrypted credentials that use the "null"
738 key (by default this is refused, since using the "null" key defeats
739 the authenticated encryption normally done).
743 * The sleep.conf configuration file gained a new MemorySleepMode=
744 setting for configuring the sleep mode in more detail.
746 * A tiny new service systemd-hibernate-clear.service has been added
747 which clears hibernation information from the HibernateLocation EFI
748 variable, in case the resume device is gone. Normally, this variable
749 is supposed to be cleaned up by the code that initiates the resume
750 from hibernation image. But when the device is missing and that code
751 doesn't run, this service will now do the necessary work, ensuring
752 that no outdated hibernation image information remains on subsequent
755 Unprivileged User Namespaces & Mounts:
757 * A small new service systemd-nsresourced.service has been added. It
758 provides a Varlink IPC API that assigns a free, transiently allocated
759 64K UID/GID range to an uninitialized user namespace a client
760 provides. It may be used to implement unprivileged container managers
761 and other programs that need dynamic user ID ranges. It also provides
762 interfaces to then delegate mount file descriptors, control groups
763 and network interfaces to user namespaces set up this way.
765 * A small new service systemd-mountfsd.service has been added. It
766 provides a Varlink IPC API for mounting DDI images, and returning a set
767 of mount file descriptors for it. If a user namespace fd is provided
768 as input, then the mounts are registered with the user namespace. To
769 ensure trust in the image it must provide Verity information (or
770 alternatively interactive polkit authentication is required).
772 * The systemd-dissect tool now can access DDIs fully unprivileged by
773 using systemd-nsresourced/systemd-mountfsd.
775 * If the service manager runs unprivileged (i.e. systemd --user) it now
776 supports RootImage= for accessing DDI images, also implemented via
777 the systemd-nsresourced/systemd-mountfsd.
779 * systemd-nspawn may now operate without privileges, if a suitable DDI
780 is provided via --image=, again implemented via
781 systemd-nsresourced/systemd-mountfsd.
785 * timedatectl and machinectl gained option '-P', an alias for
786 '--value --property=…'.
788 * Various tools that pretty-print config files will now highlight
789 configuration directives.
791 * varlinkctl gained support for the "ssh:" transport. This requires
792 OpenSSH 9.4 or newer.
794 * systemd-sysext gained support for enabling system extensions in
795 mutable fashion, where a writeable upperdir is stored under
796 /var/lib/extensions.mutable/, and a new --mutable= option to
797 configure this behaviour. An "ephemeral" mode is not also supported
798 where the mutable layer is configured to be a tmpfs that is
799 automatically released when the system extensions are reattached.
801 * Coredumps are now retained for two weeks by default (instead of three
804 * portablectl --copy= parameter gained a new 'mixed' argument, that will
805 result in resources owned by the OS (e.g.: portable profiles) to be linked
806 but resources owned by the portable image (e.g.: the unit files and the
807 images themselves) to be copied.
809 * systemd will now register MIME types for various of its file types
810 (e.g. journal files, DDIs, encrypted credentials …) via the XDG
811 shared-mime-info infrastructure. (Files of these types will thus be
812 recognized as their own thing in desktop file managers such as GNOME
815 * systemd-dissect will now show the detected sector size of a given DDI
816 in its default output.
818 * systemd-portabled now generates recognizable structured log messages
819 whenever a portable service is attached or detached.
821 * Verity signature checking in userspace (i.e. checking against
822 /etc/verity.d/ keys) when activating DDIs can now be turned on/off
823 via a kernel command line option systemd.allow_userspace_verity= and
824 an environment variable SYSTEMD_ALLOW_USERSPACE_VERITY=.
826 * ext4/xfs file system quota handling has been reworked, so that
827 quotacheck and quotaon are now invoked as per-file-system templated
828 services (as opposed to single system-wide singletons), similar in
829 style to the fsck, growfs, pcrfs logic. This means file systems with
830 quota enabled can now be reasonably enabled at runtime of the system,
833 * "systemd-analyze dot" will now also show BindsTo= dependencies.
835 * systemd-debug-generator gained the ability add in arbitrary units
836 based on them being passed in via system credentials.
838 * A new kernel command-line option systemd.default_debug_tty= can be
839 used to specify the TTY for the debug shell, independently of
840 enabling or disabling it.
842 Contributions from: A S Alam, AKHIL KUMAR, Abraham Samuel Adekunle,
843 Adrian Vovk, Adrian Wannenmacher, Alan Liang, Alberto Planas,
844 Alexander Zavyalov, Anders Jonsson, Andika Triwidada, Andres Beltran,
845 Andrew Sayers, Antonio Alvarez Feijoo, Arthur Zamarin, Artur Pak,
846 AtariDreams, Benjamin Franzke, Bernhard M. Wiedemann, Black-Hole1,
847 Bryan Jacobs, Burak Gerz, Carlos Garnacho, Chandra Pratap,
848 Chris Simons, Christian Wesselhoeft, Clayton Craft, Colin Geniet,
849 Colin Walters, Costa Tsaousis, Cristian Rodríguez, Daan De Meyer,
850 Damien Challet, Dan Streetman, David Tardon, David Venhoek,
851 Diego Viola, Dionna Amalie Glaze, Dmitry Konishchev,
852 Edson Juliano Drosdeck, Eisuke Kawashima, Eli Schwartz,
853 Emanuele Giuseppe Esposito, Eric Daigle, Evgeny Vereshchagin,
854 Felix Riemann, Fernando Fernandez Mancera, Florian Schmaus, Franck Bui,
855 Frantisek Sumsal, Friedrich Altheide, Gabríel Arthúr Pétursson,
856 Gaël Donval, Georges Basile Stavracas Neto, Gerd Hoffmann,
857 Guilhem Lettron, Göran Uddeborg, Hans de Goede, Harald Brinkmann,
858 Heinrich Schuchardt, Henry Li, Holger Assmann, Ivan Kruglov,
859 Ivan Shapovalov, Jakub Sitnicki, James Muir, Jan Engelhardt, Jan Macku,
860 Jeff King, JmbFountain, Joakim Nohlgård, Julius Alexandre,
861 Jörg Behrmann, Keian, Kirk, Kristian Klausen, Krzesimir Nowak,
862 Lars Ellenberg, Lennart Poettering, Luca Boccassi, Ludwig Nussel,
863 Lukáš Nykrýn, Luxiter, Maanya Goenka, Mariano Giménez,
864 Markus Merklinger, Martin Ivicic, Martin Trigaux, Martin Wilck,
865 Matt Layher, Matt Muggeridge, Matteo Croce, Matthias Lisin,
866 Max Gautier, Max Staudt, Michael Biebl, Michal Koutný, Michal Sekletár,
867 Mike Gilbert, Mike Yuan, Mikko Ylinen, MkfsSion, MrSmör,
868 Nandakumar Raghavan, Nick Cao, Nick Rosbrook, Ole Peder Brandtzæg,
869 Ondrej Kozina, Oğuz Ersen, Pablo Méndez Hernández, Pierre GRASSER,
870 Piotr Drąg, QuonXF, Rafaël Kooi, Raito Bezarius, Reid Wahl,
871 Renjaya Raga Zenta, Richard Maw, Roland Hieber, Ronan Pigott, Rose,
872 Ross Burton, Sam Leonard, Samuel BF, Sergei Zhmylev, Sergey A, Shulhan,
873 SidhuRupinder, Sludge, Stuart Hayhurst, Susant Sahani,
874 Takashi Sakamoto, Temuri Doghonadze, Thilo Fromm, Thomas Blume,
875 TobiPeterG, Tomáš Pecka, Topi Miettinen, Tycho Andersen, Unique-Usman,
876 Usman Akinyemi, Vasiliy Kovalev, Vasiliy Stelmachenok,
877 Vishal Chillara Srinivas, Vitaly Kuznetsov, Vito Caputo,
878 Vladimir Stoiakin, Werner Sembach, Will Springer, Winterhuman,
879 Xiaotian Wu, Yu Watanabe, Yuri Chornoivan, Zbigniew Jędrzejewski-Szmek,
880 Zmyeir, aslepykh, chenjiayi, cpackham-atlnz, cunshunxia, djantti,
881 hfavisado, hulkoba, ksaleem, medusalix, mille-feuille, mkubiak, mooo,
882 msizanoen, networkException, nl6720, r-vdp, runiq, samuelvw01,
883 sharad3001, sushmbha, wangyuhang, zzywysm, İ. Ensar Gülşen,
884 Štěpán Němec, 我超厉害, 김인수
886 — Happy Place, 2024-04-XX
890 Announcements of Future Feature Removals and Incompatible Changes:
892 * Support for split-usr (/usr/ mounted separately during late boot,
893 instead of being mounted by the initrd before switching to the rootfs)
894 and unmerged-usr (parallel directories /bin/ and /usr/bin/, /lib/ and
895 /usr/lib/, …) has been removed. For more details, see:
896 https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html
898 * We intend to remove cgroup v1 support from a systemd release after
899 the end of 2023. If you run services that make explicit use of
900 cgroup v1 features (i.e. the "legacy hierarchy" with separate
901 hierarchies for each controller), please implement compatibility with
902 cgroup v2 (i.e. the "unified hierarchy") sooner rather than later.
903 Most of Linux userspace has been ported over already.
905 * Support for System V service scripts is now deprecated and will be
906 removed in a future release. Please make sure to update your software
907 *now* to include a native systemd unit file instead of a legacy
908 System V script to retain compatibility with future systemd releases.
910 * Support for the SystemdOptions EFI variable is deprecated.
911 'bootctl systemd-efi-options' will emit a warning when used. It seems
912 that this feature is little-used and it is better to use alternative
913 approaches like credentials and confexts. The plan is to drop support
914 altogether at a later point, but this might be revisited based on
917 * systemd-run's switch --expand-environment= which currently is disabled
918 by default when combined with --scope, will be changed in a future
919 release to be enabled by default.
921 * "systemctl switch-root" is now restricted to initrd transitions only.
923 Transitions between real systems should be done with
924 "systemctl soft-reboot" instead.
926 * The "ip=off" and "ip=none" kernel command line options interpreted by
927 systemd-network-generator will now result in IPv6RA + link-local
928 addressing being disabled, too. Previously DHCP was turned off, but
929 IPv6RA and IPv6 link-local addressing was left enabled.
931 * The NAMING_BRIDGE_MULTIFUNCTION_SLOT naming scheme has been deprecated
934 * SuspendMode=, HibernateState= and HybridSleepState= in the [Sleep]
935 section of systemd-sleep.conf are now deprecated and have no effect.
936 They did not (and could not) take any value other than the respective
937 default. HybridSleepMode= is also deprecated, and will now always use
938 the 'suspend' disk mode.
942 * The way services are spawned has been overhauled. Previously, a
943 process was forked that shared all of the manager's memory (via
944 copy-on-write) while doing all the required setup (e.g.: mount
945 namespaces, CGroup configuration, etc.) before exec'ing the target
946 executable. This was problematic for various reasons: several glibc
947 APIs were called that are not supposed to be used after a fork but
948 before an exec, copy-on-write meant that if either process (the
949 manager or the child) touched a memory page a copy was triggered, and
950 also the memory footprint of the child process was that of the
951 manager, but with the memory limits of the service. From this version
952 onward, the new process is spawned using CLONE_VM and CLONE_VFORK
953 semantics via posix_spawn(3), and it immediately execs a new internal
954 binary, systemd-executor, that receives the configuration to apply
955 via memfd, and sets up the process before exec'ing the target
956 executable. The systemd-executor binary is pinned by file descriptor
957 by each manager instance (system and users), and the reference is
958 updated on daemon-reexec - it is thus important to reexec all running
959 manager instances when the systemd-executor and/or libsystemd*
960 libraries are updated on the filesystem.
962 * Most of the internal process tracking is being changed to use PIDFDs
963 instead of PIDs when the kernel supports it, to improve robustness
966 * A new option SurviveFinalKillSignal= can be used to configure the
967 unit to be skipped in the final SIGTERM/SIGKILL spree on shutdown.
968 This is part of the required configuration to let a unit's processes
969 survive a soft-reboot operation.
971 * System extension images (sysext) can now set
972 EXTENSION_RELOAD_MANAGER=1 in their extension-release files to
973 automatically reload the service manager (PID 1) when
974 merging/refreshing/unmerging on boot. Generally, while this can be
975 used to ship services in system extension images it's recommended to
976 do that via portable services instead.
978 * The ExtensionImages= and ExtensionDirectories= options now support
979 confexts images/directories.
981 * A new option NFTSet= provides a method for integrating dynamic cgroup
982 IDs into firewall rules with NFT sets. The benefit of using this
983 setting is to be able to use control group as a selector in firewall
984 rules easily and this in turn allows more fine grained filtering.
985 Also, NFT rules for cgroup matching use numeric cgroup IDs, which
986 change every time a service is restarted, making them hard to use in
989 * A new option CoredumpReceive= can be set for service and scope units,
990 together with Delegate=yes, to make systemd-coredump on the host
991 forward core files from processes crashing inside the delegated
992 CGroup subtree to systemd-coredump running in the container. This new
993 option is by default used by systemd-nspawn containers that use the
996 * A new ConditionSecurity=measured-uki option is now available, to ensure
997 a unit can only run when the system has been booted from a measured UKI.
999 * MemoryAvailable= now considers physical memory if there are no CGroup
1000 memory limits set anywhere in the tree.
1002 * The $USER environment variable is now always set for services, while
1003 previously it was only set if User= was specified. A new option
1004 SetLoginEnvironment= is now supported to determine whether to also set
1005 $HOME, $LOGNAME, and $SHELL.
1007 * Socket units now support a new pair of
1008 PollLimitBurst=/PollLimitInterval= options to configure a limit on
1009 how often polling events on the file descriptors backing this unit
1010 will be considered within a time window.
1012 * Scope units can now be created using PIDFDs instead of PIDs to select
1013 the processes they should include.
1015 * Sending SIGRTMIN+18 with 0x500 as sigqueue() value will now cause the
1016 manager to dump the list of currently pending jobs.
1018 * If the kernel supports MOVE_MOUNT_BENEATH, the systemctl and
1019 machinectl bind and mount-image verbs will now cause the new mount to
1020 replace the old mount (if any), instead of overmounting it.
1022 * Units now have MemoryPeak, MemorySwapPeak, MemorySwapCurrent and
1023 MemoryZSwapCurrent properties, which respectively contain the values
1024 of the cgroup v2's memory.peak, memory.swap.peak, memory.swap.current
1025 and memory.zswap.current properties. This information is also shown in
1026 "systemctl status" output, if available.
1028 TPM2 Support + Disk Encryption & Authentication:
1030 * systemd-cryptenroll now allows specifying a PCR bank and explicit hash
1031 value in the --tpm2-pcrs= option.
1033 * systemd-cryptenroll now allows specifying a TPM2 key handle (nv
1034 index) to be used instead of the default SRK via the new
1035 --tpm2-seal-key-handle= option.
1037 * systemd-cryptenroll now allows TPM2 enrollment using only a TPM2
1038 public key (in TPM2B_PUBLIC format) – without access to the TPM2
1039 device itself – which enables offline sealing of LUKS images for a
1040 specific TPM2 chip, as long as the SRK public key is known. Pass the
1041 public to the tool via the new --tpm2-device-key= switch.
1043 * systemd-cryptsetup is now installed in /usr/bin/ and is no longer an
1044 internal-only executable.
1046 * The TPM2 Storage Root Key will now be set up, if not already present,
1047 by a new systemd-tpm2-setup.service early boot service. The SRK will
1048 be stored in PEM format and TPM2_PUBLIC format (the latter is useful
1049 for systemd-cryptenroll --tpm2-device-key=, as mentioned above) for
1050 easier access. A new "srk" verb has been added to systemd-analyze to
1051 allow extracting it on demand if it is already set up.
1053 * The internal systemd-pcrphase executable has been renamed to
1056 * The systemd-pcrextend tool gained a new --pcr= switch to override
1057 which PCR to measure into.
1059 * systemd-pcrextend now exposes a Varlink interface at
1060 io.systemd.PCRExtend that can be used to do measurements and event
1063 * TPM measurements are now also written to an event log at
1064 /run/log/systemd/tpm2-measure.log, using a derivative of the TCG
1065 Canonical Event Log format. Previously we'd only log them to the
1066 journal, where they however were subject to rotation and similar.
1068 * A new component "systemd-pcrlock" has been added that allows managing
1069 local TPM2 PCR policies for PCRs 0-7 and similar, which are hard to
1070 predict by the OS vendor because of the inherently local nature of
1071 what measurements they contain, such as firmware versions of the
1072 system and extension cards and suchlike. pcrlock can predict PCR
1073 measurements ahead of time based on various inputs, such as the local
1074 TPM2 event log, GPT partition tables, PE binaries, UKI kernels, and
1075 various other things. It can then pre-calculate a TPM2 policy from
1076 this, which it stores in an TPM2 NV index. TPM2 objects (such as disk
1077 encryption keys) can be locked against this NV index, so that they
1078 are locked against a specific combination of system firmware and
1079 state. Alternatives for each component are supported to allowlist
1080 multiple kernel versions or boot loader version simultaneously
1081 without losing access to the disk encryption keys. The tool can also
1082 be used to analyze and validate the local TPM2 event log.
1083 systemd-cryptsetup, systemd-cryptenroll, systemd-repart have all been
1084 updated to support such policies. There's currently no support for
1085 locking the system's root disk against a pcrlock policy, this will be
1086 added soon. Moreover, it is currently not possible to combine a
1087 pcrlock policy with a signed PCR policy. This component is
1088 experimental and its public interface is subject to change.
1090 systemd-boot, systemd-stub, ukify, bootctl, kernel-install:
1092 * bootctl will now show whether the system was booted from a UKI in its
1095 * systemd-boot and systemd-stub now use different project keys in their
1096 respective SBAT sections, so that they can be revoked individually if
1099 * systemd-boot will no longer load unverified Devicetree blobs when UEFI
1100 SecureBoot is enabled. For more details see:
1101 https://github.com/systemd/systemd/security/advisories/GHSA-6m6p-rjcq-334c
1103 * systemd-boot gained new hotkeys to reboot and power off the system
1104 from the boot menu ("B" and "O"). If the "auto-poweroff" and
1105 "auto-reboot" options in loader.conf are set these entries are also
1106 shown as menu items (which is useful on devices lacking a regular
1109 * systemd-boot gained a new configuration value "menu-disabled" for the
1110 set-timeout option, to allow completely disabling the boot menu,
1111 including the hotkey.
1113 * systemd-boot will now measure the content of loader.conf in TPM2
1116 * systemd-stub will now concatenate the content of all kernel
1117 command-line addons before measuring them in TPM2 PCR 12, in a single
1118 measurement, instead of measuring them individually.
1120 * systemd-stub will now measure and load Devicetree Blob addons, which
1121 are searched and loaded following the same model as the existing
1122 kernel command-line addons.
1124 * systemd-stub will now ignore unauthenticated kernel command line options
1125 passed from systemd-boot when running inside Confidential VMs with UEFI
1128 * systemd-stub will now load a Devicetree blob even if the firmware did
1129 not load any beforehand (e.g.: for ACPI systems).
1131 * ukify is no longer considered experimental, and now ships in /usr/bin/.
1133 * ukify gained a new verb inspect to describe the sections of a UKI and
1134 print the contents of the well-known sections.
1136 * ukify gained a new verb genkey to generate a set of key pairs for
1137 signing UKIs and their PCR data.
1139 * The 90-loaderentry kernel-install hook now supports installing device
1142 * kernel-install now supports the --json=, --root=, --image=, and
1143 --image-policy= options for the inspect verb.
1145 * kernel-install now supports new list and add-all verbs. The former
1146 lists all installed kernel images (if those are available in
1147 /usr/lib/modules/). The latter will install all the kernels it can
1152 * A new option --copy-from= has been added that synthesizes partition
1153 definitions from the given image, which are then applied by the
1154 systemd-repart algorithm.
1156 * A new option --copy-source= has been added, which can be used to specify
1157 a directory to which CopyFiles= is considered relative to.
1159 * New --make-ddi=confext, --make-ddi=sysext, and --make-ddi=portable
1160 options have been added to make it easier to generate these types of
1161 DDIs, without having to provide repart.d definitions for them.
1163 * The dm-verity salt and UUID will now be derived from the specified
1166 * New VerityDataBlockSizeBytes= and VerityHashBlockSizeBytes= can now be
1167 configured in repart.d/ configuration files.
1169 * A new Subvolumes= setting is now supported in repart.d/ configuration
1170 files, to indicate which directories in the target partition should be
1173 * A new --tpm2-device-key= option can be used to lock a disk against a
1174 specific TPM2 public key. This matches the same switch the
1175 systemd-cryptenroll tool now supports (see above).
1179 * The journalctl --lines= parameter now accepts +N to show the oldest N
1180 entries instead of the newest.
1182 * journald now ensures that sealing happens once per epoch, and sets a
1183 new compatibility flag to distinguish old journal files that were
1184 created before this change, for backward compatibility.
1188 * udev will now create symlinks to loopback block devices in the
1189 /dev/disk/by-loop-ref/ directory that are based on the .lo_file_name
1190 string field selected during allocation. The systemd-dissect tool and
1191 the util-linux losetup command now supports a complementing new switch
1192 --loop-ref= for selecting the string. This means a loopback block
1193 device may now be allocated under a caller-chosen reference and can
1194 subsequently be referenced without first having to look up the block
1195 device name the caller ended up with.
1197 * udev also creates symlinks to loopback block devices in the
1198 /dev/disk/by-loop-inode/ directory based on the .st_dev/st_ino fields
1199 of the inode attached to the loopback block device. This means that
1200 attaching a file to a loopback device will implicitly make a handle
1201 available to be found via that file's inode information.
1203 * udevadm info gained support for JSON output via a new --json= flag, and
1204 for filtering output using the same mechanism that udevadm trigger
1207 * The predictable network interface naming logic is extended to include
1208 the SR-IOV-R "representor" information in network interface names.
1209 This feature was intended for v254, but even though the code was
1210 merged, the part that actually enabled the feature was forgotten.
1211 It is now enabled by default and is part of the new "v255" naming
1214 * A new hwdb/rules file has been added that sets the
1215 ID_NET_AUTO_LINK_LOCAL_ONLY=1 udev property on all network interfaces
1216 that should usually only be configured with link-local addressing
1217 (IPv4LL + IPv6LL), i.e. for PC-to-PC cables ("laplink") or
1218 Thunderbolt networking. systemd-networkd and NetworkManager (soon)
1219 will make use of this information to apply an appropriate network
1220 configuration by default.
1222 * The ID_NET_DRIVER property on network interfaces is now set
1223 relatively early in the udev rule set so that other rules may rely on
1224 its use. This is implemented in a new "net-driver" udev built-in.
1228 * The "duid-only" option for DHCPv4 client's ClientIdentifier= setting
1229 is now dropped, as it never worked, hence it should not be used by
1232 * The 'prefixstable' ipv6 address generation mode now considers the SSID
1233 when generating stable addresses, so that a different stable address
1234 is used when roaming between wireless networks. If you already use
1235 'prefixstable' addresses with wireless networks, the stable address
1236 will be changed by the update.
1238 * The DHCPv4 client gained a RapidCommit option, true by default, which
1239 enables RFC4039 Rapid Commit behavior to obtain a lease in a
1240 simplified 2-message exchange instead of the typical 4-message
1241 exchange, if also supported by the DHCP server.
1243 * The DHCPv4 client gained new InitialCongestionWindow= and
1244 InitialAdvertisedReceiveWindow= options for route configurations.
1246 * The DHCPv4 client gained a new RequestAddress= option that allows
1247 to send a preferred IP address in the initial DHCPDISCOVER message.
1249 * The DHCPv4 server and client gained support for IPv6-only mode
1252 * The SendHostname= and Hostname= options are now available for the
1253 DHCPv6 client, independently of the DHCPv4= option, so that these
1254 configuration values can be set independently for each client.
1256 * The DHCPv4 and DHCPv6 client state can now be queried via D-Bus,
1257 including lease information.
1259 * The DHCPv6 client can now be configured to use a custom DUID type.
1261 * .network files gained a new IPv4ReversePathFilter= setting in the
1262 [Network] section, to control sysctl's rp_filter setting.
1264 * .network files gaiend a new HopLimit= setting in the [Route] section,
1265 to configure a per-route hop limit.
1267 * .network files gained a new TCPRetransmissionTimeoutSec= setting in
1268 the [Route] section, to configure a per-route TCP retransmission
1271 * A new directive NFTSet= provides a method for integrating network
1272 configuration into firewall rules with NFT sets. The benefit of using
1273 this setting is that static network configuration or dynamically
1274 obtained network addresses can be used in firewall rules with the
1275 indirection of NFT set types.
1277 * The [IPv6AcceptRA] section supports the following new options:
1278 UsePREF64=, UseHopLimit=, UseICMP6RateLimit=, and NFTSet=.
1280 * The [IPv6SendRA] section supports the following new options:
1281 RetransmitSec=, HopLimit=, HomeAgent=, HomeAgentLifetimeSec=, and
1282 HomeAgentPreference=.
1284 * A new [IPv6PREF64Prefix] set of options, containing Prefix= and
1285 LifetimeSec=, has been introduced to append pref64 options in router
1286 advertisements (RFC8781).
1288 * The network generator now configures the interfaces with only
1289 link-local addressing if "ip=link-local" is specified on the kernel
1292 * The prefix of the configuration files generated by the network
1293 generator from the kernel command line is now prefixed with '70-',
1294 to make them have higher precedence over the default configuration
1297 * Added a new -Ddefault-network=BOOL meson option, that causes more
1298 .network files to be installed as enabled by default. These configuration
1299 files will which match generic setups, e.g. 89-ethernet.network matches
1300 all Ethernet interfaces and enables both DHCPv4 and DHCPv6 clients.
1302 * If a ID_NET_MANAGED_BY= udev property is set on a network device and
1303 it is any other string than "io.systemd.Network" then networkd will
1304 not manage this device. This may be used to allow multiple network
1305 management services to run in parallel and assign ownership of
1306 specific devices explicitly. NetworkManager will soon implement a
1311 * systemctl is-failed now checks the system state if no unit is
1314 * systemctl will now automatically soft-reboot if a new root file system
1315 is found under /run/nextroot/ when a reboot operation is invoked.
1319 * Wall messages now work even when utmp support is disabled, using
1320 systemd-logind to query the necessary information.
1322 * systemd-logind now sends a new PrepareForShutdownWithMetadata D-Bus
1323 signal before shutdown/reboot/soft-reboot that includes additional
1324 information compared to the PrepareForShutdown signal. Currently the
1325 additional information is the type of operation that is about to be
1328 Hibernation & Suspend:
1330 * The kernel and OS versions will no longer be checked on resume from
1333 * Hibernation into swap files backed by btrfs are now
1334 supported. (Previously this was supported only for other file
1339 * A new systemd-vmspawn tool has been added, that aims to provide for VMs
1340 the same interfaces and functionality that systemd-nspawn provides for
1341 containers. For now it supports QEMU as a backend, and exposes some of
1342 its options to the user. This component is experimental and its public
1343 interface is subject to change.
1345 * "systemd-analyze plot" has gained tooltips on each unit name with
1346 related-unit information in its svg output, such as Before=,
1347 Requires=, and similar properties.
1349 * A new varlinkctl tool has been added to allow interfacing with
1350 Varlink services, and introspection has been added to all such
1351 services. This component is experimental and its public interface is
1354 * systemd-sysext and systemd-confext now expose a Varlink service
1355 at io.systemd.sysext.
1357 * portable services now accept confexts as extensions.
1359 * systemd-sysupdate now accepts directories in the MatchPattern= option.
1361 * systemd-run will now output the invocation ID of the launched
1362 transient unit and its peak memory usage.
1364 * systemd-analyze, systemd-tmpfiles, systemd-sysusers, systemd-sysctl,
1365 and systemd-binfmt gained a new --tldr option that can be used instead
1366 of --cat-config to suppress uninteresting configuration lines, such as
1367 comments and whitespace.
1369 * resolvectl gained a new "show-server-state" command that shows
1370 current statistics of the resolver. This is backed by a new
1371 DumpStatistics() Varlink method provided by systemd-resolved.
1373 * systemd-timesyncd will now emit a D-Bus signal when the LinkNTPServers
1376 * vconsole now supports KEYMAP=@kernel for preserving the kernel keymap
1379 * seccomp now supports the LoongArch64 architecture.
1381 * seccomp may now be enabled for services running as a non-root User=
1382 without NoNewPrivileges=yes.
1384 * systemd-id128 now supports a new -P option to show only values. The
1385 combination of -P and --app options is also supported.
1387 * A new pam_systemd_loadkey.so PAM module is now available, which will
1388 automatically fetch the passphrase used by cryptsetup to unlock the
1389 root file system and set it as the PAM authtok. This enables, among
1390 other things, configuring auto-unlock of the GNOME Keyring / KDE
1391 Wallet when autologin is configured.
1393 * Many meson options now use the 'feature' type, which means they
1394 take enabled/disabled/auto as values.
1396 * A new meson option -Dconfigfiledir= can be used to change where
1397 configuration files with default values are installed to.
1399 * Options and verbs in man pages are now tagged with the version they
1400 were first introduced in.
1402 * A new component "systemd-storagetm" has been added, which exposes all
1403 local block devices as NVMe-TCP devices, fully automatically. It's
1404 hooked into a new target unit storage-target-mode.target that is
1405 suppsoed to be booted into via
1406 rd.systemd.unit=storage-target-mode.target on the kernel command
1407 line. This is intended to be used for installers and debugging to
1408 quickly get access to the local disk. It's inspired by MacOS "target
1409 disk mode". This component is experimental and its public interface is
1412 * A new component "systemd-bsod" has been added, which can show logged
1413 error messages full screen, if they have a log level of LOG_EMERG log
1414 level. This component is experimental and its public interface is
1417 * The systemd-dissect tool's --with command will now set the
1418 $SYSTEMD_DISSECT_DEVICE environment variable to the block device it
1419 operates on for the invoked process.
1421 * The systemd-mount tool gained a new --tmpfs switch for mounting a new
1422 'tmpfs' instance. This is useful since it does so via .mount units
1423 and thus can be executed remotely or in containers.
1425 * The various tools in systemd that take "verbs" (such as systemctl,
1426 loginctl, machinectl, …) now will suggest a close verb name in case
1427 the user specified an unrecognized one.
1429 * libsystemd now exports a new function sd_id128_get_app_specific()
1430 that generates "app-specific" 128bit IDs from any ID. It's similar to
1431 sd_id128_get_machine_app_specific() and
1432 sd_id128_get_boot_app_specific() but takes the ID to base calculation
1433 on as input. This new functionality is also exposed in the
1434 "systemd-id128" tool where you can now combine --app= with `show`.
1436 * All tools that parse timestamps now can also parse RFC3339 style
1437 timestamps that include the "T" and Z" characters.
1439 * New documentation has been added:
1441 https://systemd.io/FILE_DESCRIPTOR_STORE
1442 https://systemd.io/TPM2_PCR_MEASUREMENTS
1443 https://systemd.io/MOUNT_REQUIREMENTS
1445 * The codebase now recognizes the suffix .confext.raw and .sysext.raw
1446 as alternative to the .raw suffix generally accepted for DDIs. It is
1447 recommended to name configuration extensions and system extensions
1448 with such suffixes, to indicate their purpose in the name.
1450 * The sd-device API gained a new function
1451 sd_device_enumerator_add_match_property_required() which allows
1452 configuring matches on properties that are strictly required. This is
1453 different from the existing sd_device_enumerator_add_match_property()
1454 matches of which one one needs to apply.
1456 * The MAC address the veth side of an nspawn container shall get
1457 assigned may now be controlled via the $SYSTEMD_NSPAWN_NETWORK_MAC
1458 environment variable.
1460 * The libiptc dependency is now implemented via dlopen(), so that tools
1461 such as networkd and nspawn no longer have a hard dependency on the
1462 shared library when compiled with support for libiptc.
1464 * New rpm macros have been added: %systemd_user_daemon_reexec does
1465 daemon-reexec for all user managers, and %systemd_postun_with_reload
1466 and %systemd_user_postun_with_reload do a reload for system and user
1469 * coredumpctl now propagates SIGTERM to the debugger process.
1471 Contributions from: 김인수, Abderrahim Kitouni, Adam Goldman,
1472 Adam Williamson, Alexandre Peixoto Ferreira, Alex Hudspith,
1473 Alvin Alvarado, André Paiusco, Antonio Alvarez Feijoo,
1474 Anton Lundin, Arian van Putten, Arseny Maslennikov, Arthur Shau,
1475 Balázs Úr, beh_10257, Benjamin Peterson, Bertrand Jacquin,
1476 Brian Norris, Charles Lee, Cheng-Chia Tseng, Chris Patterson,
1477 Christian Hergert, Christian Hesse, Christian Kirbach,
1478 Clayton Craft, commondservice, cunshunxia, Curtis Klein, cvlc12,
1479 Daan De Meyer, Daniele Medri, Daniel P. Berrangé, Daniel Rusek,
1480 Daniel Thompson, Dan Nicholson, Dan Streetman, David Rheinsberg,
1481 David Santamaría Rogado, David Tardon, dependabot[bot],
1482 Diego Viola, Dmitry V. Levin, Emanuele Giuseppe Esposito,
1483 Emil Renner Berthing, Emil Velikov, Etienne Dechamps, Fabian Vogt,
1484 felixdoerre, Felix Dörre, Florian Schmaus, Franck Bui,
1485 Frantisek Sumsal, G2-Games, Gioele Barabucci, Hugo Carvalho,
1486 huyubiao, Iago López Galeiras, IllusionMan1212, Jade Lovelace,
1487 janana, Jan Janssen, Jan Kuparinen, Jan Macku, Jeremy Fleischman,
1488 Jin Liu, jjimbo137, Joerg Behrmann, Johannes Segitz, Jordan Rome,
1489 Jordan Williams, Julien Malka, Juno Computers, Khem Raj, khm,
1490 Kingbom Dou, Kiran Vemula, Krzesimir Nowak, Laszlo Gombos,
1491 Lennart Poettering, linuxlion, Luca Boccassi, Lucas Adriano Salles,
1492 Lukas, Lukáš Nykrýn, Maanya Goenka, Maarten, Malte Poll,
1493 Marc Pervaz Boocha, Martin Beneš, Martin Joerg, Martin Wilck,
1494 Mathieu Tortuyaux, Matthias Schiffer, Maxim Mikityanskiy,
1495 Max Kellermann, Michael A Cassaniti, Michael Biebl, Michael Kuhn,
1496 Michael Vasseur, Michal Koutný, Michal Sekletár, Mike Yuan,
1497 Milton D. Miller II, mordner, msizanoen, NAHO, Nandakumar Raghavan,
1498 Neil Wilson, Nick Rosbrook, Nils K, NRK, Oğuz Ersen,
1499 Omojola Joshua, onenowy, Paul Meyer, Paymon MARANDI, pelaufer,
1500 Peter Hutterer, PhylLu, Pierre GRASSER, Piotr Drąg, Priit Laes,
1501 Rahil Bhimjiani, Raito Bezarius, Raul Cheleguini, Reto Schneider,
1502 Richard Maw, Robby Red, RoepLuke, Roland Hieber, Roland Singer,
1503 Ronan Pigott, Sam James, Sam Leonard, Sergey A, Susant Sahani,
1504 Sven Joachim, Tad Fisher, Takashi Sakamoto, Thorsten Kukuk, Tj,
1505 Tomasz Świątek, Topi Miettinen, Valentin David,
1506 Valentin Lefebvre, Victor Westerhuis, Vincent Haupert,
1507 Vishal Chillara Srinivas, Vito Caputo, Warren, Weblate,
1508 Xiaotian Wu, xinpeng wang, Yaron Shahrabani, Yo-Jung Lin,
1509 Yu Watanabe, Zbigniew Jędrzejewski-Szmek, zeroskyx,
1510 Дамјан Георгиевски, наб
1512 — Edinburgh, 2023-12-06
1516 Announcements of Future Feature Removals and Incompatible Changes:
1518 * The next release (v255) will remove support for split-usr (/usr/
1519 mounted separately during late boot, instead of being mounted by the
1520 initrd before switching to the rootfs) and unmerged-usr (parallel
1521 directories /bin/ and /usr/bin/, /lib/ and /usr/lib/, …). For more
1523 https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html
1525 * We intend to remove cgroup v1 support from a systemd release after
1526 the end of 2023. If you run services that make explicit use of
1527 cgroup v1 features (i.e. the "legacy hierarchy" with separate
1528 hierarchies for each controller), please implement compatibility with
1529 cgroup v2 (i.e. the "unified hierarchy") sooner rather than later.
1530 Most of Linux userspace has been ported over already.
1532 * Support for System V service scripts is now deprecated and will be
1533 removed in a future release. Please make sure to update your software
1534 *now* to include a native systemd unit file instead of a legacy
1535 System V script to retain compatibility with future systemd releases.
1537 * Support for the SystemdOptions EFI variable is deprecated.
1538 'bootctl systemd-efi-options' will emit a warning when used. It seems
1539 that this feature is little-used and it is better to use alternative
1540 approaches like credentials and confexts. The plan is to drop support
1541 altogether at a later point, but this might be revisited based on
1544 * EnvironmentFile= now treats the line following a comment line
1545 trailing with escape as a non comment line. For details, see:
1546 https://github.com/systemd/systemd/issues/27975
1548 * PrivateNetwork=yes and NetworkNamespacePath= now imply
1549 PrivateMounts=yes unless PrivateMounts=no is explicitly specified.
1551 * Behaviour of sandboxing options for the per-user service manager
1552 units has changed. They now imply PrivateUsers=yes, which means user
1553 namespaces will be implicitly enabled when a sandboxing option is
1554 enabled in a user unit. Enabling user namespaces has the drawback
1555 that system users will no longer be visible (and processes/files will
1556 appear as owned by 'nobody') in the user unit.
1558 By definition a sandboxed user unit should run with reduced
1559 privileges, so impact should be small. This will remove a great
1560 source of confusion that has been reported by users over the years,
1561 due to how these options require an extra setting to be manually
1562 enabled when used in the per-user service manager, which is not
1563 needed in the system service manager. For more details, see:
1564 https://lists.freedesktop.org/archives/systemd-devel/2022-December/048682.html
1566 * systemd-run's switch --expand-environment= which currently is disabled
1567 by default when combined with --scope, will be changed in a future
1568 release to be enabled by default.
1570 Security Relevant Changes:
1572 * pam_systemd will now by default pass the CAP_WAKE_ALARM ambient
1573 process capability to invoked session processes of regular users on
1574 local seats (as well as to systemd --user), unless configured
1575 otherwise via data from JSON user records, or via the PAM module's
1576 parameter list. This is useful in order allow desktop tools such as
1577 GNOME's Alarm Clock application to set a timer for
1578 CLOCK_REALTIME_ALARM that wakes up the system when it elapses. A
1579 per-user service unit file may thus use AmbientCapability= to pass
1580 the capability to invoked processes. Note that this capability is
1581 relatively narrow in focus (in particular compared to other process
1582 capabilities such as CAP_SYS_ADMIN) and we already — by default —
1583 permit more impactful operations such as system suspend to local
1588 * Memory limits that apply while the unit is activating are now
1589 supported. Previously IO and CPU settings were already supported via
1590 StartupCPUWeight= and similar. The same logic has been added for the
1591 various manager and unit memory settings (DefaultStartupMemoryLow=,
1592 StartupMemoryLow=, StartupMemoryHigh=, StartupMemoryMax=,
1593 StartupMemorySwapMax=, StartupMemoryZSwapMax=).
1595 * The service manager gained support for enqueuing POSIX signals to
1596 services that carry an additional integer value, exposing the
1597 sigqueue() system call. This is accessible via new D-Bus calls
1598 org.freedesktop.systemd1.Manager.QueueSignalUnit() and
1599 org.freedesktop.systemd1.Unit.QueueSignal(), as well as in systemctl
1600 via the new --kill-value= option.
1602 * systemctl gained a new "list-paths" verb, which shows all currently
1603 active .path units, similarly to how "systemctl list-timers" shows
1604 active timers, and "systemctl list-sockets" shows active sockets.
1606 * systemctl gained a new --when= switch which is honoured by the various
1607 forms of shutdown (i.e. reboot, kexec, poweroff, halt) and allows
1608 scheduling these operations by time, similar in fashion to how this
1609 has been supported by SysV shutdown.
1611 * If MemoryDenyWriteExecute= is enabled for a service and the kernel
1612 supports the new PR_SET_MDWE prctl() call, it is used instead of the
1613 seccomp()-based system call filter to achieve the same effect.
1615 * A new set of kernel command line options is now understood:
1616 systemd.tty.term.<name>=, systemd.tty.rows.<name>=,
1617 systemd.tty.columns.<name>= allow configuring the TTY type and
1618 dimensions for the tty specified via <name>. When systemd invokes a
1619 service on a tty (via TTYName=) it will look for these and configure
1620 the TTY accordingly. This is particularly useful in VM environments
1621 to propagate host terminal settings into the appropriate TTYs of the
1624 * A new RootEphemeral= setting is now understood in service units. It
1625 takes a boolean argument. If enabled for services that use RootImage=
1626 or RootDirectory= an ephemeral copy of the disk image or directory
1627 tree is made when the service is started. It is removed automatically
1628 when the service is stopped. That ephemeral copy is made using
1629 btrfs/xfs reflinks or btrfs snapshots, if available.
1631 * The service activation logic gained new settings RestartSteps= and
1632 RestartMaxDelaySec= which allow exponentially-growing restart
1633 intervals for Restart=.
1635 * The service activation logic gained a new setting RestartMode= which
1636 can be set to 'direct' to skip the inactive/failed states when
1637 restarting, so that dependent units are not notified until the service
1638 converges to a final (successful or failed) state. For example, this
1639 means that OnSuccess=/OnFailure= units will not be triggered until the
1640 service state has converged.
1642 * PID 1 will now automatically load the virtio_console kernel module
1643 during early initialization if running in a suitable VM. This is done
1644 so that early-boot logging can be written to the console if available.
1646 * Similarly, virtio-vsock support is loaded early in suitable VM
1647 environments. PID 1 will send sd_notify() notifications via AF_VSOCK
1648 to the VMM if configured, thus loading this early is beneficial.
1650 * A new verb "fdstore" has been added to systemd-analyze to show the
1651 current contents of the file descriptor store of a unit. This is
1652 backed by a new D-Bus call DumpUnitFileDescriptorStore() provided by
1653 the service manager.
1655 * The service manager will now set a new $FDSTORE environment variable
1656 when invoking processes for services that have the file descriptor
1659 * A new service option FileDescriptorStorePreserve= has been added that
1660 allows tuning the lifecycle of the per-service file descriptor store.
1661 If set to "yes", the entries in the fd store are retained even after
1662 the service has been fully stopped.
1664 * The "systemctl clean" command may now be used to clear the fdstore of
1667 * Unit *.preset files gained a new directive "ignore", in addition to
1668 the existing "enable" and "disable". As the name suggests, matching
1669 units are left unchanged, i.e. neither enabled nor disabled.
1671 * Service units gained a new setting DelegateSubgroup=. It takes the
1672 name of a sub-cgroup to place any processes the service manager forks
1673 off in. Previously, the service manager would place all service
1674 processes directly in the top-level cgroup it created for the
1675 service. This usually meant that main process in a service with
1676 delegation enabled would first have to create a subgroup and move
1677 itself down into it, in order to not conflict with the "no processes
1678 in inner cgroups" rule of cgroup v2. With this option, this step is
1679 now handled by PID 1.
1681 * The service manager will now look for .upholds/ directories,
1682 similarly to the existing support for .wants/ and .requires/
1683 directories. Symlinks in this directory result in Upholds=
1686 The [Install] section of unit files gained support for a new
1687 UpheldBy= directive to generate .upholds/ symlinks automatically when
1690 * The service manager now supports a new kernel command line option
1691 systemd.default_device_timeout_sec=, which may be used to override
1692 the default timeout for .device units.
1694 * A new "soft-reboot" mechanism has been added to the service manager.
1695 A "soft reboot" is similar to a regular reboot, except that it
1696 affects userspace only: the service manager shuts down any running
1697 services and other units, then optionally switches into a new root
1698 file system (mounted to /run/nextroot/), and then passes control to a
1699 systemd instance in the new file system which then starts the system
1700 up again. The kernel is not rebooted and neither is the hardware,
1701 firmware or boot loader. This provides a fast, lightweight mechanism
1702 to quickly reset or update userspace, without the latency that a full
1703 system reset involves. Moreover, open file descriptors may be passed
1704 across the soft reboot into the new system where they will be passed
1705 back to the originating services. This allows pinning resources
1706 across the reboot, thus minimizing grey-out time further. This new
1707 reboot mechanism is accessible via the new "systemctl soft-reboot"
1710 * Services using RootDirectory= or RootImage= will now have read-only
1711 access to a copy of the host's os-release file under
1712 /run/host/os-release, which will be kept up-to-date on 'soft-reboot'.
1713 This was already the case for Portable Services, and the feature has
1714 now been extended to all services that do not run off the host's
1717 * A new service setting MemoryKSM= has been added to enable kernel
1718 same-page merging individually for services.
1720 * A new service setting ImportCredentials= has been added that augments
1721 LoadCredential= and LoadCredentialEncrypted= and searches for
1722 credentials to import from the system, and supports globbing.
1724 * A new job mode "restart-dependencies" has been added to the service
1725 manager (exposed via systemctl --job-mode=). It is only valid when
1726 used with "start" jobs, and has the effect that the "start" job will
1727 be propagated as "restart" jobs to currently running units that have
1728 a BindsTo= or Requires= dependency on the started unit.
1730 * A new verb "whoami" has been added to "systemctl" which determines as
1731 part of which unit the command is being invoked. It writes the unit
1732 name to standard output. If one or more PIDs are specified reports
1733 the unit names the processes referenced by the PIDs belong to.
1735 * The system and service credential logic has been improved: there's
1736 now a clearly defined place where system provisioning tools running
1737 in the initrd can place credentials that will be imported into the
1738 system's set of credentials during the initrd → host transition: the
1739 /run/credentials/@initrd/ directory. Once the credentials placed
1740 there are imported into the system credential set they are deleted
1741 from this directory, and the directory itself is deleted afterwards
1744 * A new kernel command line option systemd.set_credential_binary= has
1745 been added, that is similar to the pre-existing
1746 systemd.set_credential= but accepts arbitrary binary credential data,
1747 encoded in Base64. Note that the kernel command line is not a
1748 recommend way to transfer credentials into a system, since it is
1749 world-readable from userspace.
1751 * The default machine ID to use may now be configured via the
1752 system.machine_id system credential. It will only be used if no
1753 machine ID was set yet on the host.
1755 * On Linux kernel 6.4 and newer system and service credentials will now
1756 be placed in a tmpfs instance that has the "noswap" mount option
1757 set. Previously, a "ramfs" instance was used. By switching to tmpfs
1758 ACL support and overall size limits can now be enforced, without
1759 compromising on security, as the memory is never paged out either
1762 * The service manager now can detect when it is running in a
1763 'Confidential Virtual Machine', and a corresponding 'cvm' value is now
1764 accepted by ConditionSecurity= for units that want to conditionalize
1765 themselves on this. systemd-detect-virt gained new 'cvm' and
1766 '--list-cvm' switches to respectively perform the detection or list
1767 all known flavours of confidential VM, depending on the vendor. The
1768 manager will publish a 'ConfidentialVirtualization' D-Bus property,
1769 and will also set a SYSTEMD_CONFIDENTIAL_VIRTUALIZATION= environment
1770 variable for unit generators. Finally, udev rules can match on a new
1771 'cvm' key that will be set when in a confidential VM.
1772 Additionally, when running in a 'Confidential Virtual Machine', SMBIOS
1773 strings and QEMU's fw_cfg protocol will not be used to import
1774 credentials and kernel command line parameters by the system manager,
1775 systemd-boot and systemd-stub, because the hypervisor is considered
1776 untrusted in this particular setting.
1780 * The sd-journal API gained a new call sd_journal_get_seqnum() to
1781 retrieve the current log record's sequence number and sequence number
1782 ID, which allows applications to order records the same way as
1783 journal does internally. The sequence number is now also exported in
1784 the JSON and "export" output of the journal.
1786 * journalctl gained a new switch --truncate-newline. If specified
1787 multi-line log records will be truncated at the first newline,
1788 i.e. only the first line of each log message will be shown.
1790 * systemd-journal-upload gained support for --namespace=, similar to
1791 the switch of the same name of journalctl.
1795 * systemd-repart's drop-in files gained a new ExcludeFiles= option which
1796 may be used to exclude certain files from the effect of CopyFiles=.
1798 * systemd-repart's Verity support now implements the Minimize= setting
1799 to minimize the size of the resulting partition.
1801 * systemd-repart gained a new --offline= switch, which may be used to
1802 control whether images shall be built "online" or "offline",
1803 i.e. whether to make use of kernel facilities such as loopback block
1804 devices and device mapper or not.
1806 * If systemd-repart is told to populate a newly created ESP or XBOOTLDR
1807 partition with some files, it will now default to VFAT rather than
1810 * systemd-repart gained a new --architecture= switch. If specified, the
1811 per-architecture GPT partition types (i.e. the root and /usr/
1812 partitions) configured in the partition drop-in files are
1813 automatically adjusted to match the specified CPU architecture, in
1814 order to simplify cross-architecture DDI building.
1816 * systemd-repart will now default to a minimum size of 300MB for XFS
1817 filesystems if no size parameter is specified. This matches what the
1818 XFS tools (xfsprogs) can support.
1820 systemd-boot, systemd-stub, ukify, bootctl, kernel-install:
1822 * gnu-efi is no longer required to build systemd-boot and systemd-stub.
1823 Instead, pyelftools is now needed, and it will be used to perform the
1824 ELF -> PE relocations at build time.
1826 * bootctl gained a new switch --print-root-device/-R that prints the
1827 block device the root file system is backed by. If specified twice,
1828 it returns the whole disk block device (as opposed to partition block
1829 device) the root file system is on. It's useful for invocations such
1830 as "cfdisk $(bootctl -RR)" to quickly show the partition table of the
1833 * systemd-stub will now look for the SMBIOS Type 1 field
1834 "io.systemd.stub.kernel-cmdline-extra" and append its value to the
1835 kernel command line it invokes. This is useful for VMMs such as qemu
1836 to pass additional kernel command lines into the system even when
1837 booting via full UEFI. The contents of the field are measured into
1840 * The KERNEL_INSTALL_LAYOUT= setting for kernel-install gained a new
1841 value "auto". With this value, a kernel will be automatically
1842 analyzed, and if it qualifies as UKI, it will be installed as if the
1843 setting was to set to "uki", otherwise as "bls".
1845 * systemd-stub can now optionally load UEFI PE "add-on" images that may
1846 contain additional kernel command line information. These "add-ons"
1847 superficially look like a regular UEFI executable, and are expected
1848 to be signed via SecureBoot/shim. However, they do not actually
1849 contain code, but instead a subset of the PE sections that UKIs
1850 support. They are supposed to provide a way to extend UKIs with
1851 additional resources in a secure and authenticated way. Currently,
1852 only the .cmdline PE section may be used in add-ons, in which case
1853 any specified string is appended to the command line embedded into
1854 the UKI itself. A new 'addon<EFI-ARCH>.efi.stub' is now provided that
1855 can be used to trivially create addons, via 'ukify' or 'objcopy'. In
1856 the future we expect other sections to be made extensible like this as
1859 * ukify has been updated to allow building these UEFI PE "add-on"
1860 images, using the new 'addon<EFI-ARCH>.efi.stub'.
1862 * ukify now accepts SBAT information to place in the .sbat PE section
1863 of UKIs and addons. If a UKI is built the SBAT information from the
1864 inner kernel is merged with any SBAT information associated with
1865 systemd-stub and the SBAT data specified on the ukify command line.
1867 * The kernel-install script has been rewritten in C, and reuses much of
1868 the infrastructure of existing tools such as bootctl. It also gained
1869 --esp-path= and --boot-path= options to override the path to the ESP,
1870 and the $BOOT partition. Options --make-entry-directory= and
1871 --entry-token= have been added as well, similar to bootctl's options
1874 * A new kernel-install plugin 60-ukify has been added which will
1875 combine kernel/initrd locally into a UKI and optionally sign them
1876 with a local key. This may be used to switch to UKI mode even on
1877 systems where a local kernel or initrd is used. (Typically UKIs are
1878 built and signed by the vendor.)
1880 * The ukify tool now supports "pesign" in addition to the pre-existing
1881 "sbsign" for signing UKIs.
1883 * systemd-measure and systemd-stub now look for the .uname PE section
1884 that should contain the kernel's "uname -r" string.
1886 * systemd-measure and ukify now calculate expected PCR hashes for a UKI
1887 "offline", i.e. without access to a TPM (physical or
1890 Memory Pressure & Control:
1892 * The sd-event API gained new calls sd_event_add_memory_pressure(),
1893 sd_event_source_set_memory_pressure_type(),
1894 sd_event_source_set_memory_pressure_period() to create and configure
1895 an event source that is called whenever the OS signals memory
1896 pressure. Another call sd_event_trim_memory() is provided that
1897 compacts the process' memory use by releasing allocated but unused
1898 malloc() memory back to the kernel. Services can also provide their
1899 own custom callback to do memory trimming. This should improve system
1900 behaviour under memory pressure, as on Linux traditionally provided
1901 no mechanism to return process memory back to the kernel if the
1902 kernel was under memory pressure. This makes use of the kernel's PSI
1903 interface. Most long-running services in systemd have been hooked up
1904 with this, and in particular systems with low memory should benefit
1907 * Service units gained new settings MemoryPressureWatch= and
1908 MemoryPressureThresholdSec= to configure the PSI memory pressure
1909 logic individually. If these options are used, the
1910 $MEMORY_PRESSURE_WATCH and $MEMORY_PRESSURE_WRITE environment
1911 variables will be set for the invoked processes to inform them about
1912 the requested memory pressure behaviour. (This is used by the
1913 aforementioned sd-events API additions, if set.)
1915 * systemd-analyze gained a new "malloc" verb that shows the output
1916 generated by glibc's malloc_info() on services that support it. Right
1917 now, only the service manager has been updated accordingly. This
1918 call requires privileges.
1920 User & Session Management:
1922 * The sd-login API gained a new call sd_session_get_username() to
1923 return the user name of the owner of a login session. It also gained
1924 a new call sd_session_get_start_time() to retrieve the time the login
1925 session started. A new call sd_session_get_leader() has been added to
1926 return the PID of the "leader" process of a session. A new call
1927 sd_uid_get_login_time() returns the time since the specified user has
1928 most recently been continuously logged in with at least one session.
1930 * JSON user records gained a new set of fields capabilityAmbientSet and
1931 capabilityBoundingSet which contain a list of POSIX capabilities to
1932 set for the logged in users in the ambient and bounding sets,
1933 respectively. homectl gained the ability to configure these two sets
1934 for users via --capability-bounding-set=/--capability-ambient-set=.
1936 * pam_systemd learnt two new module options
1937 default-capability-bounding-set= and default-capability-ambient-set=,
1938 which configure the default bounding sets for users as they are
1939 logging in, if the JSON user record doesn't specify this explicitly
1940 (see above). The built-in default for the ambient set now contains
1941 the CAP_WAKE_ALARM, thus allowing regular users who may log in
1942 locally to resume from a system suspend via a timer.
1944 * The Session D-Bus objects systemd-logind gained a new SetTTY() method
1945 call to update the TTY of a session after it has been allocated. This
1946 is useful for SSH sessions which are typically allocated first, and
1947 for which a TTY is added later.
1949 * The sd-login API gained a new call sd_pid_notifyf_with_fds() which
1950 combines the various other sd_pid_notify() flavours into one: takes a
1951 format string, an overriding PID, and a set of file descriptors to
1952 send. It also gained a new call sd_pid_notify_barrier() call which is
1953 equivalent to sd_notify_barrier() but allows the originating PID to
1956 * "loginctl list-users" and "loginctl list-sessions" will now show the
1957 state of each logged in user/session in their tabular output. It will
1958 also show the current idle state of sessions.
1962 * systemd-dissect will now show the intended CPU architecture of an
1965 * systemd-dissect will now install itself as mount helper for the "ddi"
1966 pseudo-file system type. This means you may now mount DDIs directly
1967 via /bin/mount or /etc/fstab, making full use of embedded Verity
1968 information and all other DDI features.
1970 Example: mount -t ddi myimage.raw /some/where
1972 * The systemd-dissect tool gained the new switches --attach/--detach to
1973 attach/detach a DDI to a loopback block device without mounting it.
1974 It will automatically derive the right sector size from the image
1975 and set up Verity and similar, but not mount the file systems in it.
1977 * When systemd-gpt-auto-generator or the DDI mounting logic mount an
1978 ESP or XBOOTLDR partition the MS_NOSYMFOLLOW mount option is now
1979 implied. Given that these file systems are typically untrusted, this
1980 should make mounting them automatically have less of a security
1983 * All tools that parse DDIs (such as systemd-nspawn, systemd-dissect,
1984 systemd-tmpfiles, …) now understand a new switch --image-policy= which
1985 takes a string encoding image dissection policy. With this mechanism
1986 automatic discovery and use of specific partition types and the
1987 cryptographic requirements on the partitions (Verity, LUKS, …) can be
1988 restricted, permitting better control of the exposed attack surfaces
1989 when mounting disk images. systemd-gpt-auto-generator will honour such
1990 an image policy too, configurable via the systemd.image_policy= kernel
1991 command line option. Unit files gained the RootImagePolicy=,
1992 MountImagePolicy= and ExtensionImagePolicy= to configure the same for
1993 disk images a service runs off.
1995 * systemd-analyze gained a new verb "image-policy" to validate and
1996 parse image policy strings.
1998 * systemd-dissect gained support for a new --validate switch to
1999 superficially validate DDI structure, and check whether a specific
2000 image policy allows the DDI.
2002 * systemd-dissect gained support for a new --mtree-hash switch to
2003 optionally disable calculating mtree hashes, which can be slow on
2006 * systemd-dissect --copy-to, --copy-from, --list and --mtree switches
2007 are now able to operate on directories too, other than images.
2011 * networkd's GENEVE support as gained a new .network option
2012 InheritInnerProtocol=.
2014 * The [Tunnel] section in .netdev files has gained a new setting
2015 IgnoreDontFragment for controlling the IPv4 "DF" flag of datagrams.
2017 * A new global IPv6PrivacyExtensions= setting has been added that
2018 selects the default value of the per-network setting of the same
2021 * The predictable network interface naming logic was extended to
2022 include SR-IOV-R "representor" information in network interface
2023 names. Unfortunately, this feature was not enabled by default and can
2024 only be enabled at compilation time by setting
2025 -Ddefault-net-naming-scheme=v254.
2027 * The DHCPv4 + DHCPv6 + IPv6 RA logic in networkd gained support for
2028 the RFC8910 captive portal option.
2032 * udevadm gained the new "verify" verb for validating udev rules files
2035 * udev gained a new tool "iocost" that can be used to configure QoS IO
2036 cost data based on hwdb information onto suitable block devices. Also
2037 see https://github.com/iocost-benchmark/iocost-benchmarks.
2039 TPM2 Support + Disk Encryption & Authentication:
2041 * systemd-cryptenroll/systemd-cryptsetup will now install a TPM2 SRK
2042 ("Storage Root Key") as first step in the TPM2, and then use that
2043 for binding FDE to, if TPM2 support is used. This matches
2044 recommendations of TCG (see
2045 https://trustedcomputinggroup.org/wp-content/uploads/TCG-TPM-v2.0-Provisioning-Guidance-Published-v1r1.pdf)
2047 * systemd-cryptenroll and other tools that take TPM2 PCR parameters now
2048 understand textual identifiers for these PCRs.
2050 * systemd-veritysetup + /etc/veritytab gained support for a series of
2051 new options: hash-offset=, superblock=, format=, data-block-size=,
2052 hash-block-size=, data-blocks=, salt=, uuid=, hash=, fec-device=,
2053 fec-offset=, fec-roots= to configure various aspects of a Verity
2056 * systemd-cryptsetup + /etc/crypttab gained support for a new
2057 veracrypt-pim= option for setting the Personal Iteration Multiplier
2058 of veracrypt volumes.
2060 * systemd-integritysetup + /etc/integritytab gained support for a new
2061 mode= setting for controlling the dm-integrity mode (journal, bitmap,
2062 direct) for the volume.
2064 * systemd-analyze gained a new verb "pcrs" that shows the known TPM PCR
2065 registers, their symbolic names and current values.
2069 * The ACL support in tmpfiles.d/ has been updated: if an uppercase "X"
2070 access right is specified this is equivalent to "x" but only if the
2071 inode in question already has the executable bit set for at least
2072 some user/group. Otherwise the "x" bit will be turned off.
2074 * tmpfiles.d/'s C line type now understands a new modifier "+": a line
2075 with C+ will result in a "merge" copy, i.e. all files of the source
2076 tree are copied into the target tree, even if that tree already
2077 exists, resulting in a combined tree of files already present in the
2078 target tree and those copied in.
2080 * systemd-tmpfiles gained a new --graceful switch. If specified lines
2081 with unknown users/groups will silently be skipped.
2085 * systemd-notify gained two new options --fd= and --fdname= for sending
2086 arbitrary file descriptors to the service manager (while specifying an
2087 explicit name for it).
2089 * systemd-notify gained a new --exec switch, which makes it execute the
2090 specified command line after sending the requested messages. This is
2091 useful for sending out READY=1 first, and then continuing invocation
2092 without changing process ID, so that the tool can be nicely used
2093 within an ExecStart= line of a unit file that uses Type=notify.
2095 sd-event + sd-bus APIs:
2097 * The sd-event API gained a new call sd_event_source_leave_ratelimit()
2098 which may be used to explicitly end a rate-limit state an event
2099 source might be in, resetting all rate limiting counters.
2101 * When the sd-bus library is used to make connections to AF_UNIX D-Bus
2102 sockets, it will now encode the "description" set via
2103 sd_bus_set_description() into the source socket address. It will also
2104 look for this information when accepting a connection. This is useful
2105 to track individual D-Bus connections on a D-Bus broker for debug
2110 * systemd-resolved gained a new resolved.conf setting
2111 StateRetentionSec= which may be used to retain cached DNS records
2112 even after their nominal TTL, and use them in case upstream DNS
2113 servers cannot be reached. This can be used to make name resolution
2114 more resilient in case of network problems.
2116 * resolvectl gained a new verb "show-cache" to show the current cache
2117 contents of systemd-resolved. This verb communicates with the
2118 systemd-resolved daemon and requires privileges.
2122 * Meson >= 0.60.0 is now required to build systemd.
2124 * The default keymap to apply may now be chosen at build-time via the
2125 new -Ddefault-keymap= meson option.
2127 * Most of systemd's long-running services now have a generic handler of
2128 the SIGRTMIN+18 signal handler which executes various operations
2129 depending on the sigqueue() parameter sent along. For example, values
2130 0x100…0x107 allow changing the maximum log level of such
2131 services. 0x200…0x203 allow changing the log target of such
2132 services. 0x300 make the services trim their memory similarly to the
2133 automatic PSI-triggered action, see above. 0x301 make the services
2134 output their malloc_info() data to the logs.
2136 * machinectl gained new "edit" and "cat" verbs for editing .nspawn
2137 files, inspired by systemctl's verbs of the same name which edit unit
2138 files. Similarly, networkctl gained the same verbs for editing
2139 .network, .netdev, .link files.
2141 * A new syscall filter group "@sandbox" has been added that contains
2142 syscalls for sandboxing system calls such as those for seccomp and
2145 * New documentation has been added:
2147 https://systemd.io/COREDUMP
2148 https://systemd.io/MEMORY_PRESSURE
2151 * systemd-firstboot gained a new --reset option. If specified, the
2152 settings in /etc/ it knows how to initialize are reset.
2154 * systemd-sysext is now a multi-call binary and is also installed under
2155 the systemd-confext alias name (via a symlink). When invoked that way
2156 it will operate on /etc/ instead of /usr/ + /opt/. It thus becomes a
2157 powerful, atomic, secure configuration management of sorts, that
2158 locally can merge configuration from multiple confext configuration
2159 images into a single immutable tree.
2161 * The --network-macvlan=, --network-ipvlan=, --network-interface=
2162 switches of systemd-nspawn may now optionally take the intended
2163 network interface inside the container.
2165 * All our programs will now send an sd_notify() message with their exit
2166 status in the EXIT_STATUS= field when exiting, using the usual
2167 protocol, including PID 1. This is useful for VMMs and container
2168 managers to collect an exit status from a system as it shuts down, as
2169 set via "systemctl exit …". This is particularly useful in test cases
2170 and similar, as invocations via a VM can now nicely propagate an exit
2171 status to the host, similar to local processes.
2173 * systemd-run gained a new switch --expand-environment=no to disable
2174 server-side environment variable expansion in specified command
2175 lines. Expansion defaults to enabled for all execution types except
2176 --scope, where it defaults to off (and prints a warning) for backward
2177 compatibility reasons. --scope will be flipped to enabled by default
2178 too in a future release. If you are using --scope and passing a '$'
2179 character in the payload you should start explicitly using
2180 --expand-environment=yes/no according to the use case.
2182 * The systemd-system-update-generator has been updated to also look for
2183 the special flag file /etc/system-update in addition to the existing
2184 support for /system-update to decide whether to enter system update
2187 * The /dev/hugepages/ file system is now mounted with nosuid + nodev
2188 mount options by default.
2190 * systemd-fstab-generator now understands two new kernel command line
2191 options systemd.mount-extra= and systemd.swap-extra=, which configure
2192 additional mounts or swaps in a format similar to /etc/fstab. 'fsck'
2193 will be ran on these block devices, like it already happens for
2194 'root='. It also now supports the new fstab.extra and
2195 fstab.extra.initrd credentials that may contain additional /etc/fstab
2196 lines to apply at boot.
2198 * systemd-getty-generator now understands two new credentials
2199 getty.ttys.container and getty.ttys.serial. These credentials may
2200 contain a list of TTY devices – one per line – to instantiate
2201 container-getty@.service and serial-getty@.service on.
2203 * The getty/serial-getty/container-getty units now import the 'agetty.*'
2204 and 'login.*' credentials, which are consumed by the 'login' and
2205 'agetty' programs starting from util-linux v2.40.
2207 * systemd-sysupdate's sysupdate.d/ drop-ins gained a new setting
2208 PathRelativeTo=, which can be set to "esp", "xbootldr", "boot", in
2209 which case the Path= setting is taken relative to the ESP or XBOOTLDR
2210 partitions, rather than the system's root directory /. The relevant
2211 directories are automatically discovered.
2213 * The systemd-ac-power tool gained a new switch --low, which reports
2214 whether the battery charge is considered "low", similar to how the
2215 s2h suspend logic checks this state to decide whether to enter system
2216 suspend or hibernation.
2218 * The /etc/os-release file can now have two new optional fields
2219 VENDOR_NAME= and VENDOR_URL= to carry information about the vendor of
2222 * When the system hibernates, information about the device and offset
2223 used is now written to a non-volatile EFI variable. On next boot the
2224 system will attempt to resume from the location indicated in this EFI
2225 variable. This should make hibernation a lot more robust, while
2226 requiring no manual configuration of the resume location.
2228 * The $XDG_STATE_HOME environment variable (added in more recent
2229 versions of the XDG basedir specification) is now honoured to
2230 implement the StateDirectory= setting in user services.
2232 * A new component "systemd-battery-check" has been added. It may run
2233 during early boot (usually in the initrd), and checks the battery
2234 charge level of the system. In case the charge level is very low the
2235 user is notified (graphically via Plymouth – if available – as well
2236 as in text form on the console), and the system is turned off after a
2237 10s delay. The feature can be disabled by passing
2238 systemd.battery_check=0 through the kernel command line.
2240 * The 'passwdqc' library is now supported as an alternative to the
2241 'pwquality' library and can be selected at build time.
2243 Contributions from: 김인수, 07416, Addison Snelling, Adrian Vovk,
2244 Aidan Dang, Alexander Krabler, Alfred Klomp, Anatoli Babenia,
2245 Andrei Stepanov, Andrew Baxter, Antonio Alvarez Feijoo,
2246 Arian van Putten, Arthur Shau, A S Alam,
2247 Asier Sarasua Garmendia, Balló György, Bastien Nocera,
2248 Benjamin Herrenschmidt, Benjamin Raison, Bill Peterson,
2249 Brad Fitzpatrick, Brett Holman, bri, Chen Qi, Chitoku,
2250 Christian Hesse, Christoph Anton Mitterer, Christopher Gurnee,
2251 Colin Walters, Cornelius Hoffmann, Cristian Rodríguez, cunshunxia,
2252 cvlc12, Cyril Roelandt, Daan De Meyer, Daniele Medri,
2253 Daniel P. Berrangé, Daniel Rusek, Dan Streetman, David Edmundson,
2254 David Schroeder, David Tardon, dependabot[bot],
2255 Dimitri John Ledkov, Dmitrii Fomchenkov, Dmitry V. Levin, dmkUK,
2256 Dominique Martinet, don bright, drosdeck, Edson Juliano Drosdeck,
2257 Egor Ignatov, EinBaum, Emanuele Giuseppe Esposito, Eric Curtin,
2258 Erik Sjölund, Evgeny Vereshchagin, Florian Klink, Franck Bui,
2259 François Rigault, Fran Diéguez, Franklin Yu, Frantisek Sumsal,
2260 Fuminobu TAKEYAMA, Gaël PORTAY, Gerd Hoffmann, Gertalitec,
2261 Gibeom Gwon, Gustavo Noronha Silva, Hannu Lounento,
2262 Hans de Goede, Haochen Tong, HATAYAMA Daisuke, Henrik Holst,
2263 Hoe Hao Cheng, Igor Tsiglyar, Ivan Vecera, James Hilliard,
2264 Jan Engelhardt, Jan Janssen, Jan Luebbe, Jan Macku, Janne Sirén,
2265 jcg, Jeidnx, Joan Bruguera, Joerg Behrmann, jonathanmetzman,
2266 Jordan Rome, Josef Miegl, Joshua Goins, Joyce, Joyce Brum,
2267 Juno Computers, Kai Lueke, Kevin P. Fleming, Kiran Vemula, Klaus,
2268 Klaus Zipfel, Lawrence Thorpe, Lennart Poettering, licunlong,
2269 Lily Foster, Luca Boccassi, Ludwig Nussel, Luna Jernberg,
2270 maanyagoenka, Maanya Goenka, Maksim Kliazovich, Malte Poll,
2271 Marko Korhonen, Masatake YAMATO, Mateusz Poliwczak, Matt Johnston,
2272 Miao Wang, Micah Abbott, Michael A Cassaniti, Michal Koutný,
2273 Michal Sekletár, Mike Yuan, mooo, Morten Linderud, msizanoen,
2274 Nick Rosbrook, nikstur, Olivier Gayot, Omojola Joshua,
2275 Paolo Velati, Paul Barker, Pavel Borecki, Petr Menšík,
2276 Philipp Kern, Philip Withnall, Piotr Drąg, Quintin Hill,
2277 Rene Hollander, Richard Phibel, Robert Meijers, Robert Scheck,
2278 Roger Gammans, Romain Geissler, Ronan Pigott, Russell Harmon,
2279 saikat0511, Samanta Navarro, Sam James, Sam Morris,
2280 Simon Braunschmidt, Sjoerd Simons, Sorah Fukumori,
2281 Stanislaw Gruszka, Stefan Roesch, Steven Luo, Steve Ramage,
2282 Susant Sahani, taniishkaaa, Tanishka, Temuri Doghonadze,
2283 Thierry Martin, Thomas Blume, Thomas Genty, Thomas Weißschuh,
2284 Thorsten Kukuk, Times-Z, Tobias Powalowski, tofylion,
2285 Topi Miettinen, Uwe Kleine-König, Velislav Ivanov,
2286 Vitaly Kuznetsov, Vít Zikmund, Weblate, Will Fancher,
2287 William Roberts, Winterhuman, Wolfgang Müller, Xeonacid,
2288 Xiaotian Wu, Xi Ruoyao, Yuri Chornoivan, Yu Watanabe, Yuxiang Zhu,
2289 Zbigniew Jędrzejewski-Szmek, zhmylove, ZjYwMj,
2290 Дамјан Георгиевски, наб
2292 — Edinburgh, 2023-07-28
2296 Announcements of Future Feature Removals and Incompatible Changes:
2298 * We intend to remove cgroup v1 support from systemd release after the
2299 end of 2023. If you run services that make explicit use of cgroup v1
2300 features (i.e. the "legacy hierarchy" with separate hierarchies for
2301 each controller), please implement compatibility with cgroup v2 (i.e.
2302 the "unified hierarchy") sooner rather than later. Most of Linux
2303 userspace has been ported over already.
2305 * We intend to remove support for split-usr (/usr mounted separately
2306 during boot) and unmerged-usr (parallel directories /bin and
2307 /usr/bin, /lib and /usr/lib, etc). This will happen in the second
2308 half of 2023, in the first release that falls into that time window.
2309 For more details, see:
2310 https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html
2312 * We intend to change behaviour w.r.t. units of the per-user service
2313 manager and sandboxing options, so that they work without having to
2314 manually enable PrivateUsers= as well, which is not required for
2315 system units. To make this work, we will implicitly enable user
2316 namespaces (PrivateUsers=yes) when a sandboxing option is enabled in a
2317 user unit. The drawback is that system users will no longer be visible
2318 (and appear as 'nobody') to the user unit when a sandboxing option is
2319 enabled. By definition a sandboxed user unit should run with reduced
2320 privileges, so impact should be small. This will remove a great source
2321 of confusion that has been reported by users over the years, due to
2322 how these options require an extra setting to be manually enabled when
2323 used in the per-user service manager, as opposed as to the system
2324 service manager. We plan to enable this change in the next release
2325 later this year. For more details, see:
2326 https://lists.freedesktop.org/archives/systemd-devel/2022-December/048682.html
2328 Deprecations and incompatible changes:
2330 * systemctl will now warn when invoked without /proc/ mounted
2331 (e.g. when invoked after chroot() into an directory tree without the
2332 API mount points like /proc/ being set up.) Operation in such an
2333 environment is not fully supported.
2335 * The return value of 'systemctl is-active|is-enabled|is-failed' for
2336 unknown units is changed: previously 1 or 3 were returned, but now 4
2337 (EXIT_PROGRAM_OR_SERVICES_STATUS_UNKNOWN) is used as documented.
2339 * 'udevadm hwdb' subcommand is deprecated and will emit a warning.
2340 systemd-hwdb (added in 2014) should be used instead.
2342 * 'bootctl --json' now outputs a single JSON array, instead of a stream
2343 of newline-separated JSON objects.
2345 * Udev rules in 60-evdev.rules have been changed to load hwdb
2346 properties for all modalias patterns. Previously only the first
2347 matching pattern was used. This could change what properties are
2348 assigned if the user has more and less specific patterns that could
2349 match the same device, but it is expected that the change will have
2350 no effect for most users.
2352 * systemd-networkd-wait-online exits successfully when all interfaces
2353 are ready or unmanaged. Previously, if neither '--any' nor
2354 '--interface=' options were used, at least one interface had to be in
2355 configured state. This change allows the case where systemd-networkd
2356 is enabled, but no interfaces are configured, to be handled
2357 gracefully. It may occur in particular when a different network
2358 manager is also enabled and used.
2360 * Some compatibility helpers were dropped: EmergencyAction= in the user
2361 manager, as well as measuring kernel command line into PCR 8 in
2362 systemd-stub, along with the -Defi-tpm-pcr-compat compile-time
2365 * The '-Dupdate-helper-user-timeout=' build-time option has been
2366 renamed to '-Dupdate-helper-user-timeout-sec=', and now takes an
2367 integer as parameter instead of a string.
2369 * The DDI image dissection logic (which backs RootImage= in service
2370 unit files, the --image= switch in various tools such as
2371 systemd-nspawn, as well as systemd-dissect) will now only mount file
2372 systems of types btrfs, ext4, xfs, erofs, squashfs, vfat. This list
2373 can be overridden via the $SYSTEMD_DISSECT_FILE_SYSTEMS environment
2374 variable. These file systems are fairly well supported and maintained
2375 in current kernels, while others are usually more niche, exotic or
2376 legacy and thus typically do not receive the same level of security
2379 * The default per-link multicast DNS mode is changed to "yes"
2380 (that was previously "no"). As the default global multicast DNS mode
2381 has been "yes" (but can be changed by the build option), now the
2382 multicast DNS is enabled on all links by default. You can disable the
2383 multicast DNS on all links by setting MulticastDNS= in resolved.conf,
2384 or on an interface by calling "resolvectl mdns INTERFACE no".
2388 * A tool 'ukify' tool to build, measure, and sign Unified Kernel Images
2389 (UKIs) has been added. This replaces functionality provided by
2390 'dracut --uefi' and extends it with automatic calculation of PE file
2391 offsets, insertion of signed PCR policies generated by
2392 systemd-measure, support for initrd concatenation, signing of the
2393 embedded Linux image and the combined image with sbsign, and
2394 heuristics to autodetect the kernel uname and verify the splash
2397 Changes in systemd and units:
2399 * A new service type Type=notify-reload is defined. When such a unit is
2400 reloaded a UNIX process signal (typically SIGHUP) is sent to the main
2401 service process. The manager will then wait until it receives a
2402 "RELOADING=1" followed by a "READY=1" notification from the unit as
2403 response (via sd_notify()). Otherwise, this type is the same as
2404 Type=notify. A new setting ReloadSignal= may be used to change the
2405 signal to send from the default of SIGHUP.
2407 user@.service, systemd-networkd.service, systemd-udevd.service, and
2408 systemd-logind have been updated to this type.
2410 * Initrd environments which are not on a pure memory file system (e.g.
2411 overlayfs combination as opposed to tmpfs) are now supported. With
2412 this change, during the initrd → host transition ("switch root")
2413 systemd will erase all files of the initrd only when the initrd is
2414 backed by a memory file system such as tmpfs.
2416 * New per-unit MemoryZSwapMax= option has been added to configure
2417 memory.zswap.max cgroup properties (the maximum amount of zswap
2420 * A new LogFilterPatterns= option has been added for units. It may be
2421 used to specify accept/deny regular expressions for log messages
2422 generated by the unit, that shall be enforced by systemd-journald.
2423 Rejected messages are neither stored in the journal nor forwarded.
2424 This option may be used to suppress noisy or uninteresting messages
2427 * The manager has a new
2428 org.freedesktop.systemd1.Manager.GetUnitByPIDFD() D-Bus method to
2429 query process ownership via a PIDFD, which is more resilient against
2430 PID recycling issues.
2432 * Scope units now support OOMPolicy=. Login session scopes default to
2433 OOMPolicy=continue, allowing login scopes to survive the OOM killer
2434 terminating some processes in the scope.
2436 * systemd-fstab-generator now supports x-systemd.makefs option for
2437 /sysroot/ (in the initrd).
2439 * The maximum rate at which daemon reloads are executed can now be
2440 limited with the new ReloadLimitIntervalSec=/ReloadLimitBurst=
2441 options. (Or the equivalent on the kernel command line:
2442 systemd.reload_limit_interval_sec=/systemd.reload_limit_burst=). In
2443 addition, systemd now logs the originating unit and PID when a reload
2444 request is received over D-Bus.
2446 * When enabling a swap device systemd will now reinitialize the device
2447 when the page size of the swap space does not match the page size of
2448 the running kernel. Note that this requires the 'swapon' utility to
2449 provide the '--fixpgsz' option, as implemented by util-linux, and it
2450 is not supported by busybox at the time of writing.
2452 * systemd now executes generator programs in a mount namespace
2453 "sandbox" with most of the file system read-only and write access
2454 restricted to the output directories, and with a temporary /tmp/
2455 mount provided. This provides a safeguard against programming errors
2456 in the generators, but also fixes here-docs in shells, which
2457 previously didn't work in early boot when /tmp/ wasn't available
2458 yet. (This feature has no security implications, because the code is
2459 still privileged and can trivially exit the sandbox.)
2461 * The system manager will now parse a new "vmm.notify_socket"
2462 system credential, which may be supplied to a VM via SMBIOS. If
2463 found, the manager will send a "READY=1" notification on the
2464 specified socket after boot is complete. This allows readiness
2465 notification to be sent from a VM guest to the VM host over a VSOCK
2468 * The sample PAM configuration file for systemd-user@.service now
2469 includes a call to pam_namespace. This puts children of user@.service
2470 in the expected namespace. (Many distributions replace their file
2471 with something custom, so this change has limited effect.)
2473 * A new environment variable $SYSTEMD_DEFAULT_MOUNT_RATE_LIMIT_BURST
2474 can be used to override the mount units burst late limit for
2475 parsing '/proc/self/mountinfo', which was introduced in v249.
2478 * Drop-ins for init.scope changing control group resource limits are
2479 now applied, while they were previously ignored.
2481 * New build-time configuration options '-Ddefault-timeout-sec=' and
2482 '-Ddefault-user-timeout-sec=' have been added, to let distributions
2483 choose the default timeout for starting/stopping/aborting system and
2484 user units respectively.
2486 * Service units gained a new setting OpenFile= which may be used to
2487 open arbitrary files in the file system (or connect to arbitrary
2488 AF_UNIX sockets in the file system), and pass the open file
2489 descriptor to the invoked process via the usual file descriptor
2490 passing protocol. This is useful to give unprivileged services access
2491 to select files which have restrictive access modes that would
2492 normally not allow this. It's also useful in case RootDirectory= or
2493 RootImage= is used to allow access to files from the host environment
2494 (which is after all not visible from the service if these two options
2499 * The new net naming scheme "v253" has been introduced. In the new
2500 scheme, ID_NET_NAME_PATH is also set for USB devices not connected via
2501 a PCI bus. This extends the coverage of predictable interface names
2502 in some embedded systems.
2504 The "amba" bus path is now included in ID_NET_NAME_PATH, resulting in
2505 a more informative path on some embedded systems.
2507 * Partition block devices will now also get symlinks in
2508 /dev/disk/by-diskseq/<seq>-part<n>, which may be used to reference
2509 block device nodes via the kernel's "diskseq" value. Previously those
2510 symlinks were only created for the main block device.
2512 * A new operator '-=' is supported for SYMLINK variables. This allows
2513 symlinks to be unconfigured even if an earlier rule added them.
2515 * 'udevadm --trigger --settle' now also works for network devices
2516 that are being renamed.
2518 Changes in sd-boot, bootctl, and the Boot Loader Specification:
2520 * systemd-boot now passes its random seed directly to the kernel's RNG
2521 via the LINUX_EFI_RANDOM_SEED_TABLE_GUID configuration table, which
2522 means the RNG gets seeded very early in boot before userspace has
2525 * systemd-boot will pass a disk-backed random seed – even when secure
2526 boot is enabled – if it can additionally get a random seed from EFI
2527 itself (via EFI's RNG protocol), or a prior seed in
2528 LINUX_EFI_RANDOM_SEED_TABLE_GUID from a preceding bootloader.
2530 * systemd-boot-system-token.service was renamed to
2531 systemd-boot-random-seed.service and extended to always save a random
2532 seed to ESP on every boot when a compatible boot loader is used. This
2533 allows a refreshed random seed to be used in the boot loader.
2535 * systemd-boot handles various seed inputs using a domain- and
2536 field-separated hashing scheme.
2538 * systemd-boot's 'random-seed-mode' option has been removed. A system
2539 token is now always required to be present for random seeds to be
2542 * systemd-boot now supports being loaded from other locations than the
2543 ESP, for example for direct kernel boot under QEMU or when embedded
2546 * systemd-boot now parses SMBIOS information to detect
2547 virtualization. This information is used to skip some warnings which
2548 are not useful in a VM and to conditionalize other aspects of
2551 * systemd-boot now supports a new 'if-safe' mode that will perform UEFI
2552 Secure Boot automated certificate enrollment from the ESP only if it
2553 is considered 'safe' to do so. At the moment 'safe' means running in
2556 * systemd-stub now processes random seeds in the same way as
2557 systemd-boot already does, in case a unified kernel image is being
2558 used from a different bootloader than systemd-boot, or without any
2561 * bootctl will now generate a system token on all EFI systems, even
2562 virtualized ones, and is activated in the case that the system token
2563 is missing from either sd-boot and sd-stub booted systems.
2565 * bootctl now implements two new verbs: 'kernel-identify' prints the
2566 type of a kernel image file, and 'kernel-inspect' provides
2567 information about the embedded command line and kernel version of
2570 * bootctl now honours $KERNEL_INSTALL_CONF_ROOT with the same meaning
2571 as for kernel-install.
2573 * The JSON output of "bootctl list" will now contain two more fields:
2574 isDefault and isSelected are boolean fields set to true on the
2575 default and currently booted boot menu entries.
2577 * bootctl gained a new verb "unlink" for removing a boot loader entry
2578 type #1 file from disk in a safe and robust way.
2580 * bootctl also gained a new verb "cleanup" that automatically removes
2581 all files from the ESP's and XBOOTLDR's "entry-token" directory, that
2582 is not referenced anymore by any installed Type #1 boot loader
2583 specification entry. This is particularly useful in environments where
2584 a large number of entries reference the same or partly the same
2585 resources (for example, for snapshot-based setups).
2587 Changes in kernel-install:
2589 * A new "installation layout" can be configured as layout=uki. With
2590 this setting, a Boot Loader Specification Type#1 entry will not be
2591 created. Instead, a new kernel-install plugin 90-uki-copy.install
2592 will copy any .efi files from the staging area into the boot
2593 partition. A plugin to generate the UKI .efi file must be provided
2596 Changes in systemctl:
2598 * 'systemctl reboot' has dropped support for accepting a positional
2599 argument as the argument to the reboot(2) syscall. Please use the
2600 --reboot-argument= option instead.
2602 * 'systemctl disable' will now warn when called on units without
2603 install information. A new --no-warn option has been added that
2604 silences this warning.
2606 * New option '--drop-in=' can be used to tell 'systemctl edit' the name
2607 of the drop-in to edit. (Previously, 'override.conf' was always
2610 * 'systemctl list-dependencies' now respects --type= and --state=.
2612 * 'systemctl kexec' now supports XEN VMM environments.
2614 * 'systemctl edit' will now tell the invoked editor to jump into the
2615 first line with actual unit file data, skipping over synthesized
2618 Changes in systemd-networkd and related tools:
2620 * The [DHCPv4] section in .network file gained new SocketPriority=
2621 setting that assigns the Linux socket priority used by the DHCPv4 raw
2622 socket. This may be used in conjunction with the
2623 EgressQOSMaps=setting in [VLAN] section of .netdev file to send the
2624 desired ethernet 802.1Q frame priority for DHCPv4 initial
2625 packets. This cannot be achieved with netfilter mangle tables because
2626 of the raw socket bypass.
2628 * The [DHCPv4] and [IPv6AcceptRA] sections in .network file gained a
2629 new QuickAck= boolean setting that enables the TCP quick ACK mode for
2630 the routes configured by the acquired DHCPv4 lease or received router
2631 advertisements (RAs).
2633 * The RouteMetric= option (for DHCPv4, DHCPv6, and IPv6 advertised
2634 routes) now accepts three values, for high, medium, and low preference
2635 of the router (which can be set with the RouterPreference=) setting.
2637 * systemd-networkd-wait-online now supports matching via alternative
2640 * The [DHCPv6] section in .network file gained new SendRelease=
2641 setting which enables the DHCPv6 client to send release when
2642 it stops. This is the analog of the [DHCPv4] SendRelease= setting.
2643 It is enabled by default.
2645 * If the Address= setting in [Network] or [Address] sections in .network
2646 specified without its prefix length, then now systemd-networkd assumes
2647 /32 for IPv4 or /128 for IPv6 addresses.
2649 * networkctl shows network and link file dropins in status output.
2651 Changes in systemd-dissect:
2653 * systemd-dissect gained a new option --list, to print the paths of
2654 all files and directories in a DDI.
2656 * systemd-dissect gained a new option --mtree, to generate a file
2657 manifest compatible with BSD mtree(5) of a DDI
2659 * systemd-dissect gained a new option --with, to execute a command with
2660 the specified DDI temporarily mounted and used as working
2661 directory. This is for example useful to convert a DDI to "tar"
2662 simply by running it within a "systemd-dissect --with" invocation.
2664 * systemd-dissect gained a new option --discover, to search for
2665 Discoverable Disk Images (DDIs) in well-known directories of the
2666 system. This will list machine, portable service and system extension
2669 * systemd-dissect now understands 2nd stage initrd images stored as a
2670 Discoverable Disk Image (DDI).
2672 * systemd-dissect will now display the main UUID of GPT DDIs (i.e. the
2673 disk UUID stored in the GPT header) among the other data it can show.
2675 * systemd-dissect gained a new --in-memory switch to operate on an
2676 in-memory copy of the specified DDI file. This is useful to access a
2677 DDI with write access without persisting any changes. It's also
2678 useful for accessing a DDI without keeping the originating file
2681 * The DDI dissection logic will now automatically detect the intended
2682 sector size of disk images stored in files, based on the GPT
2683 partition table arrangement. Loopback block devices for such DDIs
2684 will then be configured automatically for the right sector size. This
2685 is useful to make dealing with modern 4K sector size DDIs fully
2686 automatic. The systemd-dissect tool will now show the detected sector
2687 size among the other DDI information in its output.
2689 Changes in systemd-repart:
2691 * systemd-repart gained new options --include-partitions= and
2692 --exclude-partitions= to filter operation on partitions by type UUID.
2693 This allows systemd-repart to be used to build images in which the
2694 type of one partition is set based on the contents of another
2695 partition (for example when the boot partition shall include a verity
2696 hash of the root partition).
2698 * systemd-repart also gained a --defer-partitions= option that is
2699 similar to --exclude-partitions=, but the size of the partition is
2700 still taken into account when sizing partitions, but without
2703 * systemd-repart gained a new --sector-size= option to specify what
2704 sector size should be used when an image is created.
2706 * systemd-repart now supports generating erofs file systems via
2707 CopyFiles= (a read-only file system similar to squashfs).
2709 * The Minimize= option was extended to accept "best" (which means the
2710 most minimal image possible, but may require multiple attempts) and
2711 "guess" (which means a reasonably small image).
2713 * The systemd-growfs binary now comes with a regular unit file template
2714 systemd-growfs@.service which can be instantiated directly for any
2715 desired file system. (Previously, the unit was generated dynamically
2716 by various generators, but no regular unit file template was
2719 Changes in journal tools:
2721 * Various systemd tools will append extra fields to log messages when
2722 in debug mode, or when SYSTEMD_ENABLE_LOG_CONTEXT=1 is set. Currently
2723 this includes information about D-Bus messages when sd-bus is used,
2724 e.g. DBUS_SENDER=, DBUS_DESTINATION=, and DBUS_PATH=, and information
2725 about devices when sd-device is used, e.g. DEVNAME= and DRIVER=.
2726 Details of what is logged and when are subject to change.
2728 * The systemd-journald-audit.socket can now be disabled via the usual
2729 "systemctl disable" mechanism to stop collection of audit
2730 messages. Please note that it is not enabled statically anymore and
2731 must be handled by the preset/enablement logic in package
2732 installation scripts.
2734 * New options MaxUse=, KeepFree=, MaxFileSize=, and MaxFiles= can
2735 be used to curtail disk use by systemd-journal-remote. This is
2736 similar to the options supported by systemd-journald.
2738 Changes in systemd-cryptenroll, systemd-cryptsetup, and related
2741 * When enrolling new keys systemd-cryptenroll now supports unlocking
2742 via FIDO2 tokens (option --unlock-fido2-device=). Previously, a
2743 password was strictly required to be specified.
2745 * systemd-cryptsetup now supports pre-flight requests for FIDO2 tokens
2746 (except for tokens with user verification, UV) to identify tokens
2747 before authentication. Multiple FIDO2 tokens can now be enrolled at
2748 the same time, and systemd-cryptsetup will automatically select one
2749 that corresponds to one of the available LUKS key slots.
2751 * systemd-cryptsetup now supports new options tpm2-measure-bank= and
2752 tpm2-measure-pcr= in crypttab(5). These allow specifying the TPM2 PCR
2753 bank and number into which the volume key should be measured. This is
2754 automatically enabled for the encrypted root volume discovered and
2755 activated by systemd-gpt-auto-generator.
2757 * systemd-gpt-auto-generator mounts the ESP and XBOOTLDR partitions with
2758 "noexec,nosuid,nodev".
2760 * systemd-gpt-auto-generator will now honour the rootfstype= and
2761 rootflags= kernel command line switches for root file systems it
2762 discovers, to match behaviour in case an explicit root fs is
2763 specified via root=.
2765 * systemd-pcrphase gained new options --machine-id and --file-system=
2766 to measure the machine-id and mount point information into PCR 15.
2767 New service unit files systemd-pcrmachine.service and
2768 systemd-pcrfs@.service have been added that invoke the tool with
2769 these switches during early boot.
2771 * systemd-pcrphase gained a --graceful switch will make it exit cleanly
2772 with a success exit code even if no TPM device is detected.
2774 * systemd-cryptenroll now stores the user-supplied PIN with a salt,
2775 making it harder to brute-force.
2777 Changes in other tools:
2779 * systemd-homed gained support for luksPbkdfForceIterations (the
2780 intended number of iterations for the PBKDF operation on LUKS).
2782 * Environment variables $SYSTEMD_HOME_MKFS_OPTIONS_BTRFS,
2783 $SYSTEMD_HOME_MKFS_OPTIONS_EXT4, and $SYSTEMD_HOME_MKFS_OPTIONS_XFS
2784 may now be used to specify additional arguments for mkfs when
2785 systemd-homed formats a file system.
2787 * systemd-hostnamed now exports the contents of
2788 /sys/class/dmi/id/bios_vendor and /sys/class/dmi/id/bios_date via two
2789 new D-Bus properties: FirmwareVendor and FirmwareDate. This allows
2790 unprivileged code to access those values.
2792 systemd-hostnamed also exports the SUPPORT_END= field from
2793 os-release(5) as OperatingSystemSupportEnd. hostnamectl make uses of
2794 this to show the status of the installed system.
2796 * systemd-measure gained an --append= option to sign multiple phase
2797 paths with different signing keys. This allows secrets to be
2798 accessible only in certain parts of the boot sequence. Note that
2799 'ukify' provides similar functionality in a more accessible form.
2801 * systemd-timesyncd will now write a structured log message with
2802 MESSAGE_ID set to SD_MESSAGE_TIME_BUMP when it bumps the clock based
2803 on a on-disk timestamp, similarly to what it did when reaching
2804 synchronization via NTP.
2806 * systemd-timesyncd will now update the on-disk timestamp file on each
2807 boot at least once, making it more likely that the system time
2808 increases in subsequent boots.
2810 * systemd-vconsole-setup gained support for system/service credentials:
2811 vconsole.keymap/vconsole.keymap_toggle and
2812 vconsole.font/vconsole.font_map/vconsole.font_unimap are analogous
2813 the similarly-named options in vconsole.conf.
2815 * systemd-localed will now save the XKB keyboard configuration to
2816 /etc/vconsole.conf, and also read it from there with a higher
2817 preference than the /etc/X11/xorg.conf.d/00-keyboard.conf config
2818 file. Previously, this information was stored in the former file in
2819 converted form, and only in latter file in the original form. Tools
2820 which want to access keyboard configuration can now do so from a
2823 * systemd-resolved gained support for configuring the nameservers and
2824 search domains via kernel command line (nameserver=, domain=) and
2825 credentials (network.dns, network.search_domains).
2827 * systemd-resolved will now synthesize host names for the DNS stub
2828 addresses it supports. Specifically when "_localdnsstub" is resolved,
2829 127.0.0.53 is returned, and if "_localdnsproxy" is resolved
2830 127.0.0.54 is returned.
2832 * systemd-notify will now send a "RELOADING=1" notification when called
2833 with --reloading, and "STOPPING=1" when called with --stopping. This
2834 can be used to implement notifications from units where it's easier
2835 to call a program than to use the sd-daemon library.
2837 * systemd-analyze's 'plot' command can now output its information in
2838 JSON, controlled via the --json= switch. Also, new --table, and
2839 --no-legend options have been added.
2841 * 'machinectl enable' will now automatically enable machines.target
2842 unit in addition to adding the machine unit to the target.
2844 Similarly, 'machinectl start|stop' gained a --now option to enable or
2845 disable the machine unit when starting or stopping it.
2847 * systemd-sysusers will now create /etc/ if it is missing.
2849 * systemd-sleep 'HibernateDelaySec=' setting is changed back to
2850 pre-v252's behaviour, and a new 'SuspendEstimationSec=' setting is
2851 added to provide the new initial value for the new automated battery
2852 estimation functionality. If 'HibernateDelaySec=' is set to any value,
2853 the automated estimate (and thus the automated hibernation on low
2854 battery to avoid data loss) functionality will be disabled.
2856 * Default tmpfiles.d/ configuration will now automatically create
2857 credentials storage directory '/etc/credstore/' with the appropriate,
2858 secure permissions. If '/run/credstore/' exists, its permissions will
2859 be fixed too in case they are not correct.
2861 Changes in libsystemd and shared code:
2863 * sd-bus gained new convenience functions sd_bus_emit_signal_to(),
2864 sd_bus_emit_signal_tov(), and sd_bus_message_new_signal_to().
2866 * sd-id128 functions now return -EUCLEAN (instead of -EIO) when the
2867 128-bit ID in files such as /etc/machine-id has an invalid
2868 format. They also accept NULL as output parameter in more places,
2869 which is useful when the caller only wants to validate the inputs and
2870 does not need the output value.
2872 * sd-login gained new functions sd_pidfd_get_session(),
2873 sd_pidfd_get_owner_uid(), sd_pidfd_get_unit(),
2874 sd_pidfd_get_user_unit(), sd_pidfd_get_slice(),
2875 sd_pidfd_get_user_slice(), sd_pidfd_get_machine_name(), and
2876 sd_pidfd_get_cgroup(), that are analogous to sd_pid_get_*(),
2877 but accept a PIDFD instead of a PID.
2879 * sd-path (and systemd-path) now export four new paths:
2880 SD_PATH_SYSTEMD_SYSTEM_ENVIRONMENT_GENERATOR,
2881 SD_PATH_SYSTEMD_USER_ENVIRONMENT_GENERATOR,
2882 SD_PATH_SYSTEMD_SEARCH_SYSTEM_ENVIRONMENT_GENERATOR, and
2883 SD_PATH_SYSTEMD_SEARCH_USER_ENVIRONMENT_GENERATOR,
2885 * sd_notify() now supports AF_VSOCK as transport for notification
2886 messages (in addition to the existing AF_UNIX support). This is
2887 enabled if $NOTIFY_SOCKET is set in a "vsock:CID:port" format.
2889 * Detection of chroot() environments now works if /proc/ is not
2890 mounted. This affects systemd-detect-virt --chroot, but also means
2891 that systemd tools will silently skip various operations in such an
2894 * "Lockheed Martin Hardened Security for Intel Processors" (HS SRE)
2895 virtualization is now detected.
2897 Changes in the build system:
2899 * Standalone variants of systemd-repart and systemd-shutdown may now be
2900 built (if -Dstandalone=true).
2902 * systemd-ac-power has been moved from /usr/lib/ to /usr/bin/, to, for
2903 example, allow scripts to conditionalize execution on AC power
2906 * The libp11kit library is now loaded through dlopen(3).
2908 Changes in the documentation:
2910 * Specifications that are not closely tied to systemd have moved to
2911 https://uapi-group.org/specifications/: the Boot Loader Specification
2912 and the Discoverable Partitions Specification.
2914 Contributions from: 김인수, 13r0ck, Aidan Dang, Alberto Planas,
2915 Alvin Šipraga, Andika Triwidada, AndyChi, angus-p, Anita Zhang,
2916 Antonio Alvarez Feijoo, Arsen Arsenović, asavah, Benjamin Fogle,
2917 Benjamin Tissoires, berenddeschouwer, BerndAdameit,
2918 Bernd Steinhauser, blutch112, cake03, Callum Farmer, Carlo Teubner,
2919 Charles Hardin, chris, Christian Brauner, Christian Göttsche,
2920 Cristian Rodríguez, Daan De Meyer, Dan Streetman, DaPigGuy,
2921 Darrell Kavanagh, David Tardon, dependabot[bot], Dirk Su,
2922 Dmitry V. Levin, drosdeck, Edson Juliano Drosdeck, edupont,
2923 Eric DeVolder, Erik Moqvist, Evgeny Vereshchagin, Fabian Gurtner,
2924 Felix Riemann, Franck Bui, Frantisek Sumsal, Geert Lorang,
2925 Gerd Hoffmann, Gio, Hannoskaj, Hans de Goede, Hugo Carvalho,
2926 igo95862, Ilya Leoshkevich, Ivan Shapovalov, Jacek Migacz,
2927 Jade Lovelace, Jan Engelhardt, Jan Janssen, Jan Macku, January,
2928 Jason A. Donenfeld, jcg, Jean-Tiare Le Bigot, Jelle van der Waa,
2929 Jeremy Linton, Jian Zhang, Jiayi Chen, Jia Zhang, Joerg Behrmann,
2930 Jörg Thalheim, Joshua Goins, joshuazivkovic, Joshua Zivkovic,
2931 Kai-Chuan Hsieh, Khem Raj, Koba Ko, Lennart Poettering, lichao,
2932 Li kunyu, Luca Boccassi, Luca BRUNO, Ludwig Nussel,
2933 Łukasz Stelmach, Lycowolf, marcel151, Marcus Schäfer, Marek Vasut,
2934 Mark Laws, Michael Biebl, Michał Kotyla, Michal Koutný,
2935 Michal Sekletár, Mike Gilbert, Mike Yuan, MkfsSion, ml,
2936 msizanoen1, mvzlb, MVZ Ludwigsburg, Neil Moore, Nick Rosbrook,
2937 noodlejetski, Pasha Vorobyev, Peter Cai, p-fpv, Phaedrus Leeds,
2938 Philipp Jungkamp, Quentin Deslandes, Raul Tambre, Ray Strode,
2939 reuben olinsky, Richard E. van der Luit, Richard Phibel,
2940 Ricky Tigg, Robin Humble, rogg, Rudi Heitbaum, Sam James,
2941 Samuel Cabrero, Samuel Thibault, Siddhesh Poyarekar, Simon Brand,
2942 Space Meyer, Spindle Security, Steve Ramage, Takashi Sakamoto,
2943 Thomas Haller, Tonći Galić, Topi Miettinen, Torsten Hilbrich,
2944 Tuetuopay, uerdogan, Ulrich Ölmann, Valentin David,
2945 Vitaly Kuznetsov, Vito Caputo, Waltibaba, Will Fancher,
2946 William Roberts, wouter bolsterlee, Youfu Zhang, Yu Watanabe,
2947 Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски,
2950 — Warsaw, 2023-02-15
2954 Announcements of Future Feature Removals:
2956 * We intend to remove cgroup v1 support from systemd release after the
2957 end of 2023. If you run services that make explicit use of cgroup v1
2958 features (i.e. the "legacy hierarchy" with separate hierarchies for
2959 each controller), please implement compatibility with cgroup v2 (i.e.
2960 the "unified hierarchy") sooner rather than later. Most of Linux
2961 userspace has been ported over already.
2963 * We intend to remove support for split-usr (/usr mounted separately
2964 during boot) and unmerged-usr (parallel directories /bin and
2965 /usr/bin, /lib and /usr/lib, etc). This will happen in the second
2966 half of 2023, in the first release that falls into that time window.
2967 For more details, see:
2968 https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html
2970 Compatibility Breaks:
2972 * ConditionKernelVersion= checks that use the '=' or '!=' operators
2973 will now do simple string comparisons (instead of version comparisons
2974 à la stverscmp()). Version comparisons are still done for the
2975 ordering operators '<', '>', '<=', '>='. Moreover, if no operator is
2976 specified, a shell-style glob match is now done. This creates a minor
2977 incompatibility compared to older systemd versions when the '*', '?',
2978 '[', ']' characters are used, as these will now match as shell globs
2979 instead of literally. Given that kernel version strings typically do
2980 not include these characters we expect little breakage through this
2983 * The service manager will now read the SELinux label used for SELinux
2984 access checks from the unit file at the time it loads the file.
2985 Previously, the label would be read at the moment of the access
2986 check, which was problematic since at that time the unit file might
2987 already have been updated or removed.
2991 * systemd-measure is a new tool for calculating and signing expected
2992 TPM2 PCR values for a given unified kernel image (UKI) booted via
2993 sd-stub. The public key used for the signature and the signed
2994 expected PCR information can be embedded inside the UKI. This
2995 information can be extracted from the UKI by external tools and code
2996 in the image itself and is made available to userspace in the booted
2999 systemd-cryptsetup, systemd-cryptenroll, and systemd-creds have been
3000 updated to make use of this information if available in the booted
3001 kernel: when locking an encrypted volume/credential to the TPM
3002 systemd-cryptenroll/systemd-creds will use the public key to bind the
3003 volume/credential to any kernel that carries PCR information signed
3004 by the same key pair. When unlocking such volumes/credentials
3005 systemd-cryptsetup/systemd-creds will use the signature embedded in
3006 the booted UKI to gain access.
3008 Binding TPM-based disk encryption to public keys/signatures of PCR
3009 values — instead of literal PCR values — addresses the inherent
3010 "brittleness" of traditional PCR-bound TPM disk encryption schemes:
3011 disks remain accessible even if the UKI is updated, without any TPM
3012 specific preparation during the OS update — as long as each UKI
3013 carries the necessary PCR signature information.
3015 Net effect: if you boot a properly prepared kernel, TPM-bound disk
3016 encryption now defaults to be locked to kernels which carry PCR
3017 signatures from the same key pair. Example: if a hypothetical distro
3018 FooOS prepares its UKIs like this, TPM-based disk encryption is now –
3019 by default – bound to only FooOS kernels, and encrypted volumes bound
3020 to the TPM cannot be unlocked on kernels from other sources. (But do
3021 note this behaviour requires preparation/enabling in the UKI, and of
3022 course users can always enroll non-TPM ways to unlock the volume.)
3024 * systemd-pcrphase is a new tool that is invoked at six places during
3025 system runtime, and measures additional words into TPM2 PCR 11, to
3026 mark milestones of the boot process. This allows binding access to
3027 specific TPM2-encrypted secrets to specific phases of the boot
3028 process. (Example: LUKS2 disk encryption key only accessible in the
3029 initrd, but not later.)
3031 Changes in systemd itself, i.e. the manager and units
3033 * The cpu controller is delegated to user manager units by default, and
3034 CPUWeight= settings are applied to the top-level user slice units
3035 (app.slice, background.slice, session.slice). This provides a degree
3036 of resource isolation between different user services competing for
3039 * Systemd can optionally do a full preset in the "first boot" condition
3040 (instead of just enable-only). This behaviour is controlled by the
3041 compile-time option -Dfirst-boot-full-preset. Right now it defaults
3042 to 'false', but the plan is to switch it to 'true' for the subsequent
3045 * Drop-ins are now allowed for transient units too.
3047 * Systemd will set the taint flag 'support-ended' if it detects that
3048 the OS image is past its end-of-support date. This date is declared
3049 in a new /etc/os-release field SUPPORT_END= described below.
3051 * Two new settings ConditionCredential= and AssertCredential= can be
3052 used to skip or fail units if a certain system credential is not
3055 * ConditionMemory= accepts size suffixes (K, M, G, T, …).
3057 * DefaultSmackProcessLabel= can be used in system.conf and user.conf to
3058 specify the SMACK security label to use when not specified in a unit
3061 * DefaultDeviceTimeoutSec= can be used in system.conf and user.conf to
3062 specify the default timeout when waiting for device units to
3065 * C.UTF-8 is used as the default locale if nothing else has been
3068 * [Condition|Assert]Firmware= have been extended to support certain
3069 SMBIOS fields. For example
3071 ConditionFirmware=smbios-field(board_name = "Custom Board")
3073 conditionalizes the unit to run only when
3074 /sys/class/dmi/id/board_name contains "Custom Board" (without the
3077 * ConditionFirstBoot= now correctly evaluates as true only during the
3078 boot phase of the first boot. A unit executed later, after booting
3079 has completed, will no longer evaluate this condition as true.
3081 * Socket units will now create sockets in the SELinuxContext= of the
3082 associated service unit, if any.
3084 * Boot phase transitions (start initrd → exit initrd → boot complete →
3085 shutdown) will be measured into TPM2 PCR 11, so that secrets can be
3086 bound to a specific runtime phase. E.g.: a LUKS encryption key can be
3087 unsealed only in the initrd.
3089 * Service credentials (i.e. SetCredential=/LoadCredential=/…) will now
3090 also be provided to ExecStartPre= processes.
3092 * Various units are now correctly ordered against
3093 initrd-switch-root.target where previously a conflict without
3094 ordering was configured. A stop job for those units would be queued,
3095 but without the ordering it could be executed only after
3096 initrd-switch-root.service, leading to units not being restarted in
3097 the host system as expected.
3099 * In order to fully support the IPMI watchdog driver, which has not yet
3100 been ported to the new common watchdog device interface,
3101 /dev/watchdog0 will be tried first and systemd will silently fallback
3102 to /dev/watchdog if it is not found.
3104 * New watchdog-related D-Bus properties are now published by systemd:
3105 WatchdogDevice, WatchdogLastPingTimestamp,
3106 WatchdogLastPingTimestampMonotonic.
3108 * At shutdown, API virtual files systems (proc, sys, etc.) will be
3111 * At shutdown, systemd will now log about processes blocking unmounting
3114 * A new meson build option 'clock-valid-range-usec-max' was added to
3115 allow disabling system time correction if RTC returns a timestamp far
3118 * Propagated restart jobs will no longer be discarded while a unit is
3121 * PID 1 will now import system credentials from SMBIOS Type 11 fields
3122 ("OEM vendor strings"), in addition to qemu_fwcfg. This provides a
3123 simple, fast and generic path for supplying credentials to a VM,
3124 without involving external tools such as cloud-init/ignition.
3126 * The CPUWeight= setting of unit files now accepts a new special value
3127 "idle", which configures "idle" level scheduling for the unit.
3129 * Service processes that are activated due to a .timer or .path unit
3130 triggering will now receive information about this via environment
3131 variables. Note that this is information is lossy, as activation
3132 might be coalesced and only one of the activating triggers will be
3133 reported. This is hence more suited for debugging or tracing rather
3134 than for behaviour decisions.
3136 * The riscv_flush_icache(2) system call has been added to the list of
3137 system calls allowed by default when SystemCallFilter= is used.
3139 * The selinux context derived from the target executable, instead of
3140 'init_t' used for the manager itself, is now used when creating
3141 listening sockets for units that specify SELinuxContextFromNet=yes.
3143 Changes in sd-boot, bootctl, and the Boot Loader Specification:
3145 * The Boot Loader Specification has been cleaned up and clarified.
3146 Various corner cases in version string comparisons have been fixed
3147 (e.g. comparisons for empty strings). Boot counting is now part of
3148 the main specification.
3150 * New PCRs measurements are performed during boot: PCR 11 for the
3151 kernel+initrd combo, PCR 13 for any sysext images. If a measurement
3152 took place this is now reported to userspace via the new
3153 StubPcrKernelImage and StubPcrInitRDSysExts EFI variables.
3155 * As before, systemd-stub will measure kernel parameters and system
3156 credentials into PCR 12. It will now report this fact via the
3157 StubPcrKernelParameters EFI variable to userspace.
3159 * The UEFI monotonic boot counter is now included in the updated random
3160 seed file maintained by sd-boot, providing some additional entropy.
3162 * sd-stub will use LoadImage/StartImage to execute the kernel, instead
3163 of arranging the image manually and jumping to the kernel entry
3164 point. sd-stub also installs a temporary UEFI SecurityOverride to
3165 allow the (unsigned) nested image to be booted. This is safe because
3166 the outer (signed) stub+kernel binary must have been verified before
3167 the stub was executed.
3169 * Booting in EFI mixed mode (a 64-bit kernel over 32-bit UEFI firmware)
3170 is now supported by sd-boot.
3172 * bootctl gained a bunch of new options: --all-architectures to install
3173 binaries for all supported EFI architectures, --root= and --image=
3174 options to operate on a directory or disk image, and
3175 --install-source= to specify the source for binaries to install,
3176 --efi-boot-option-description= to control the name of the boot entry.
3178 * The sd-boot stub exports a StubFeatures flag, which is used by
3179 bootctl to show features supported by the stub that was used to boot.
3181 * The PE section offsets that are used by tools that assemble unified
3182 kernel images have historically been hard-coded. This may lead to
3183 overlapping PE sections which may break on boot. The UKI will now try
3184 to detect and warn about this.
3186 Any tools that assemble UKIs must update to calculate these offsets
3187 dynamically. Future sd-stub versions may use offsets that will not
3188 work with the currently used set of hard-coded offsets!
3190 * sd-stub now accepts (and passes to the initrd and then to the full
3191 OS) new PE sections '.pcrsig' and '.pcrkey' that can be used to embed
3192 signatures of expected PCR values, to allow sealing secrets via the
3193 TPM2 against pre-calculated PCR measurements.
3195 Changes in the hardware database:
3197 * 'systemd-hwdb query' now supports the --root= option.
3199 Changes in systemctl:
3201 * systemctl now supports --state= and --type= options for the 'show'
3204 * systemctl gained a new verb 'list-automounts' to list automount
3207 * systemctl gained support for a new --image= switch to be able to
3208 operate on the specified disk image (similar to the existing --root=
3209 which operates relative to some directory).
3211 Changes in systemd-networkd:
3213 * networkd can set Linux NetLabel labels for integration with the
3214 network control in security modules via a new NetLabel= option.
3216 * The RapidCommit= is (re-)introduced to enable faster configuration
3217 via DHCPv6 (RFC 3315).
3219 * networkd gained a new option TCPCongestionControlAlgorithm= that
3220 allows setting a per-route TCP algorithm.
3222 * networkd gained a new option KeepFileDescriptor= to allow keeping a
3223 reference (file descriptor) open on TUN/TAP interfaces, which is
3224 useful to avoid link flaps while the underlying service providing the
3225 interface is being serviced.
3227 * RouteTable= now also accepts route table names.
3229 Changes in systemd-nspawn:
3231 * The --bind= and --overlay= options now support relative paths.
3233 * The --bind= option now supports a 'rootidmap' value, which will
3234 use id-mapped mounts to map the root user inside the container to the
3235 owner of the mounted directory on the host.
3237 Changes in systemd-resolved:
3239 * systemd-resolved now persists DNSOverTLS in its state file too. This
3240 fixes a problem when used in combination with NetworkManager, which
3241 sends the setting only once, causing it to be lost if resolved was
3242 restarted at any point.
3244 * systemd-resolved now exposes a Varlink socket at
3245 /run/systemd/resolve/io.systemd.Resolve.Monitor, accessible only for
3246 root. Processed DNS requests in a JSON format will be published to
3247 any clients connected to this socket.
3249 resolvectl gained a 'monitor' verb to make use of this.
3251 * systemd-resolved now treats unsupported DNSSEC algorithms as INSECURE
3252 instead of returning SERVFAIL, as per RFC:
3253 https://datatracker.ietf.org/doc/html/rfc6840#section-5.2
3255 * OpenSSL is the default crypto backend for systemd-resolved. (gnutls
3256 is still supported.)
3258 Changes in libsystemd and other libraries:
3260 * libsystemd now exports sd_bus_error_setfv() (a convenience function
3261 for setting bus errors), sd_id128_string_equal (a convenience
3262 function for 128-bit ID string comparisons), and
3263 sd_bus_message_read_strv_extend() (a function to incrementally read
3266 * libsystemd now exports sd_device_get_child_first()/_next() as a
3267 high-level interface for enumerating child devices. It also supports
3268 sd_device_new_child() for opening a child device given a device
3271 * libsystemd now exports sd_device_monitor_set()/get_description()
3272 which allow setting a custom description that will be used in log
3273 messages by sd_device_monitor*.
3275 * Private shared libraries (libsystemd-shared-nnn.so,
3276 libsystemd-core-nnn.so) are now installed into arch-specific
3277 directories to allow multi-arch installs.
3279 * A new sd-gpt.h header is now published, listing GUIDs from the
3280 Discoverable Partitions specification. For more details see:
3281 https://systemd.io/DISCOVERABLE_PARTITIONS/
3283 * A new function sd_hwdb_new_from_path() has been added to open a hwdb
3284 database given an explicit path to the file.
3286 * The signal number argument to sd_event_add_signal() now can now be
3287 ORed with the SD_EVENT_SIGNAL_PROCMASK flag, causing sigprocmask() to
3288 be automatically invoked to block the specified signal. This is
3289 useful to simplify invocations as the caller doesn't have to do this
3292 * A new convenience call sd_event_set_signal_exit() has been added to
3293 sd-event to set up signal handling so that the event loop
3294 automatically terminates cleanly on SIGTERM/SIGINT.
3296 Changes in other components:
3298 * systemd-sysusers, systemd-tmpfiles, and systemd-sysctl configuration
3299 can now be provided via the credential mechanism.
3301 * systemd-analyze gained a new verb 'compare-versions' that implements
3302 comparisons for versions strings (similarly to 'rpmdev-vercmp' and
3303 'dpkg --compare-versions').
3305 * 'systemd-analyze dump' is extended to accept glob patterns for unit
3306 names to limit the output to matching units.
3308 * tmpfiles.d/ lines can read file contents to write from a credential.
3309 The new modifier char '^' is used to specify that the argument is a
3310 credential name. This mechanism is used to automatically populate
3311 /etc/motd, /etc/issue, and /etc/hosts from credentials.
3313 * tmpfiles.d/ may now be configured to avoid changing uid/gid/mode of
3314 an inode if the specification is prefixed with ':' and the inode
3317 * Default tmpfiles.d/ configuration now carries a line to automatically
3318 use an 'ssh.authorized_keys.root' credential if provided to set up
3319 the SSH authorized_keys file for the root user.
3321 * systemd-tmpfiles will now gracefully handle absent source of "C" copy
3324 * tmpfiles.d/ F/w lines now optionally permit encoding of the payload
3325 in base64. This is useful to write arbitrary binary data into files.
3327 * The pkgconfig and rpm macros files now export the directory for user
3328 units as 'user_tmpfiles_dir' and '%_user_tmpfilesdir'.
3330 * Detection of Apple Virtualization and detection of Parallels and
3331 KubeVirt virtualization on non-x86 archs have been added.
3333 * os-release gained a new field SUPPORT_END=YYYY-MM-DD to inform the
3334 user when their system will become unsupported.
3336 * When performing suspend-then-hibernate, the system will estimate the
3337 discharge rate and use that to set the delay until hibernation and
3338 hibernate immediately instead of suspending when running from a
3339 battery and the capacity is below 5%.
3341 * systemd-sysctl gained a --strict option to fail when a sysctl
3342 setting is unknown to the kernel.
3344 * machinectl supports --force for the 'copy-to' and 'copy-from'
3347 * coredumpctl gained the --root and --image options to look for journal
3348 files under the specified root directory, image, or block device.
3350 * 'journalctl -o' and similar commands now implement a new output mode
3351 "short-delta". It is similar to "short-monotonic", but also shows the
3352 time delta between subsequent messages.
3354 * journalctl now respects the --quiet flag when verifying consistency
3357 * Journal log messages gained a new implicit field _RUNTIME_SCOPE= that
3358 will indicate whether a message was logged in the 'initrd' phase or
3359 in the 'system' phase of the boot process.
3361 * Journal files gained a new compatibility flag
3362 'HEADER_INCOMPATIBLE_COMPACT'. Files with this flag implement changes
3363 to the storage format that allow reducing size on disk. As with other
3364 compatibility flags, older journalctl versions will not be able to
3365 read journal files using this new format. The environment variable
3366 'SYSTEMD_JOURNAL_COMPACT=0' can be passed to systemd-journald to
3367 disable this functionality. It is enabled by default.
3369 * systemd-run's --working-directory= switch now works when used in
3370 combination with --scope.
3372 * portablectl gained a --force flag to skip certain sanity checks. This
3373 is implemented using new flags accepted by systemd-portabled for the
3374 *WithExtensions() D-Bus methods: SD_SYSTEMD_PORTABLE_FORCE_ATTACH
3375 flag now means that the attach/detach checks whether the units are
3376 already present and running will be skipped. Similarly,
3377 SD_SYSTEMD_PORTABLE_FORCE_SYSEXT flag means that the check whether
3378 image name matches the name declared inside of the image will be
3379 skipped. Callers must be sure to do those checks themselves if
3382 * systemd-portabled will now use the original filename to check
3383 extension-release.NAME for correctness, in case it is passed a
3386 * systemd-portabled now uses PrivateTmp=yes in the 'trusted' profile
3389 * sysext's extension-release files now support '_any' as a special
3390 value for the ID= field, to allow distribution-independent extensions
3391 (e.g.: fully statically compiled binaries, scripts). It also gained
3392 support for a new ARCHITECTURE= field that may be used to explicitly
3393 restrict an image to hosts of a specific architecture.
3395 * systemd-repart now supports creating squashfs partitions. This
3396 requires mksquashfs from squashfs-tools.
3398 * systemd-repart gained a --split flag to also generate split
3399 artifacts, i.e. a separate file for each partition. This is useful in
3400 conjunction with systemd-sysupdate or other tools, or to generate
3401 split dm-verity artifacts.
3403 * systemd-repart is now able to generate dm-verity partitions, including
3406 * systemd-repart can now set a partition UUID to zero, allowing it to
3407 be filled in later, such as when using verity partitions.
3409 * systemd-repart now supports drop-ins for its configuration files.
3411 * Package metadata logged by systemd-coredump in the system journal is
3414 * xdg-autostart-service now expands 'tilde' characters in Exec lines.
3416 * systemd-oomd now automatically links against libatomic, if available.
3418 * systemd-oomd now sends out a 'Killed' D-Bus signal when a cgroup is
3421 * scope units now also provide oom-kill status.
3423 * systemd-pstore will now try to load only the efi_pstore kernel module
3424 before running, ensuring that pstore can be used.
3426 * systemd-logind gained a new StopIdleSessionSec= option to stop an idle
3427 session after a preconfigure timeout.
3429 * systemd-homed will now wait up to 30 seconds for workers to terminate,
3430 rather than indefinitely.
3432 * homectl gained a new '--luks-sector-size=' flag that allows users to
3433 select the preferred LUKS sector size. Must be a power of 2 between 512
3434 and 4096. systemd-userdbd records gained a corresponding field.
3436 * systemd-sysusers will now respect the 'SOURCE_DATE_EPOCH' environment
3437 variable when generating the 'sp_lstchg' field, to ensure an image
3438 build can be reproducible.
3440 * 'udevadm wait' will now listen to kernel uevents too when called with
3443 * When naming network devices udev will now consult the Devicetree
3444 "alias" fields for the device.
3446 * systemd-udev will now create infiniband/by-path and
3447 infiniband/by-ibdev links for Infiniband verbs devices.
3449 * systemd-udev-trigger.service will now also prioritize input devices.
3451 * ConditionACPower= and systemd-ac-power will now assume the system is
3452 running on AC power if no battery can be found.
3454 * All features and tools using the TPM2 will now communicate with it
3455 using a bind key. Beforehand, the tpm2 support used encrypted sessions
3456 by creating a primary key that was used to encrypt traffic. This
3457 creates a problem as the key created for encrypting the traffic could
3458 be faked by an active interposer on the bus. In cases when a pin is
3459 used, a bind key will be used. The pin is used as the auth value for
3460 the seal key, aka the disk encryption key, and that auth value will be
3461 used in the session establishment. An attacker would need the pin
3462 value to create the secure session and thus an active interposer
3463 without the pin cannot interpose on TPM2 traffic.
3465 * systemd-growfs no longer requires udev to run.
3467 * systemd-backlight now will better support systems with multiple
3470 * systemd-cryptsetup's keyfile-timeout= option now also works when a
3471 device is used as a keyfile.
3473 * systemd-cryptenroll gained a new --unlock-key-file= option to get the
3474 unlocking key from a key file (instead of prompting the user). Note
3475 that this is the key for unlocking the volume in order to be able to
3476 enroll a new key, but it is not the key that is enrolled.
3478 * systemd-dissect gained a new --umount switch that will safely and
3479 synchronously unmount all partitions of an image previously mounted
3480 with 'systemd-dissect --mount'.
3482 * When using gcrypt, all systemd tools and services will now configure
3483 it to prefer the OS random number generator if present.
3485 * All example code shipped with documentation has been relicensed from CC0
3488 * Unit tests will no longer fail when running on a system without
3491 Experimental features:
3493 * BPF programs can now be compiled with bpf-gcc (requires libbpf >= 1.0
3494 and bpftool >= 7.0).
3496 * sd-boot can automatically enroll SecureBoot keys from files found on
3497 the ESP. This enrollment can be either automatic ('force' mode) or
3498 controlled by the user ('manual' mode). It is sufficient to place the
3499 SecureBoot keys in the right place in the ESP and they will be picked
3500 up by sd-boot and shown in the boot menu.
3502 * The mkosi config in systemd gained support for automatically
3503 compiling a kernel with the configuration appropriate for testing
3504 systemd. This may be useful when developing or testing systemd in
3505 tandem with the kernel.
3507 Contributions from: 김인수, Adam Williamson, adrian5, Aidan Dang,
3508 Akihiko Odaki, Alban Bedel, Albert Mikaelyan, Aleksey Vasenev,
3509 Alexander Graf, Alexander Shopov, Alexander Wilson,
3510 Alper Nebi Yasak, anarcat, Anders Jonsson, Andre Kalb,
3511 Andrew Stone, Andrey Albershteyn, Anita Zhang, Ansgar Burchardt,
3512 Antonio Alvarez Feijoo, Arnaud Ferraris, Aryan singh, asavah,
3513 Avamander, Avram Lubkin, Balázs Meskó, Bastien Nocera,
3514 Benjamin Franzke, BerndAdameit, bin456789, Celeste Liu,
3515 Chih-Hsuan Yen, Christian Brauner, Christian Göttsche,
3516 Christian Hesse, Clyde Byrd III, codefiles, Colin Walters,
3517 Cristian Rodríguez, Daan De Meyer, Daniel Braunwarth,
3518 Daniel Rusek, Dan Streetman, Darsey Litzenberger, David Edmundson,
3519 David Jaša, David Rheinsberg, David Seifert, David Tardon,
3520 dependabot[bot], Devendra Tewari, Dominique Martinet, drosdeck,
3521 Edson Juliano Drosdeck, Eduard Tolosa, eggfly, Einsler Lee,
3522 Elias Probst, Eli Schwartz, Evgeny Vereshchagin, exploide, Fei Li,
3523 Foster Snowhill, Franck Bui, Frank Dana, Frantisek Sumsal,
3524 Gerd Hoffmann, Gio, Goffredo Baroncelli, gtwang01,
3525 Guillaume W. Bres, H A, Hans de Goede, Heinrich Schuchardt,
3526 Hugo Carvalho, i-do-cpp, igo95862, j00512545, Jacek Migacz,
3527 Jade Bilkey, James Hilliard, Jan B, Janis Goldschmidt,
3528 Jan Janssen, Jan Kuparinen, Jan Luebbe, Jan Macku,
3529 Jason A. Donenfeld, Javkhlanbayar Khongorzul, Jeremy Soller,
3530 JeroenHD, jiangchuangang, João Loureiro,
3531 Joaquín Ignacio Aramendía, Jochen Sprickerhof,
3532 Johannes Schauer Marin Rodrigues, Jonas Kümmerlin,
3533 Jonas Witschel, Jonathan Kang, Jonathan Lebon, Joost Heitbrink,
3534 Jörg Thalheim, josh-gordon-fb, Joyce, Kai Lueke, lastkrick,
3535 Lennart Poettering, Leon M. George, licunlong, Li kunyu,
3536 LockBlock-dev, Loïc Collignon, Lubomir Rintel, Luca Boccassi,
3537 Luca BRUNO, Ludwig Nussel, Łukasz Stelmach, Maccraft123,
3538 Marc Kleine-Budde, Marius Vollmer, Martin Wilck, matoro,
3539 Matthias Lisin, Max Gautier, Maxim Mikityanskiy, Michael Biebl,
3540 Michal Koutný, Michal Sekletár, Michal Stanke, Mike Gilbert,
3541 Mitchell Freiderich, msizanoen1, Nick Rosbrook, nl6720, Oğuz Ersen,
3542 Oleg Solovyov, Olga Smirnova, Pablo Ceballos, Pavel Zhukov,
3543 Phaedrus Leeds, Philipp Gortan, Piotr Drąg, Pyfisch,
3544 Quentin Deslandes, Rahil Bhimjiani, Rene Hollander, Richard Huang,
3545 Richard Phibel, Rudi Heitbaum, Sam James, Sarah Brofeldt,
3546 Sean Anderson, Sebastian Scheibner, Shreenidhi Shedi,
3547 Sonali Srivastava, Steve Ramage, Suraj Krishnan, Swapnil Devesh,
3548 Takashi Sakamoto, Ted X. Toth, Temuri Doghonadze, Thomas Blume,
3549 Thomas Haller, Thomas Hebb, Tomáš Hnyk, Tomasz Paweł Gajc,
3550 Topi Miettinen, Ulrich Ölmann, undef, Uriel Corfa,
3551 Victor Westerhuis, Vincent Dagonneau, Vishal Chillara Srinivas,
3552 Vito Caputo, Weblate, Wenchao Hao, William Roberts, williamsumendap,
3553 wineway, xiaoyang, Yuri Chornoivan, Yu Watanabe,
3554 Zbigniew Jędrzejewski-Szmek, Zhaofeng Li, наб
3556 – The Great Beyond, 2022-10-31 👻
3560 Backwards-incompatible changes:
3562 * The minimum kernel version required has been bumped from 3.13 to 4.15,
3563 and CLOCK_BOOTTIME is now assumed to always exist.
3565 * C11 with GNU extensions (aka "gnu11") is now used to build our
3566 components. Public API headers are still restricted to ISO C89.
3568 * In v250, a systemd-networkd feature that automatically configures
3569 routes to addresses specified in AllowedIPs= was added and enabled by
3570 default. However, this causes network connectivity issues in many
3571 existing setups. Hence, it has been disabled by default since
3572 systemd-stable 250.3. The feature can still be used by explicitly
3573 configuring RouteTable= setting in .netdev files.
3575 * Jobs started via StartUnitWithFlags() will no longer return 'skipped'
3576 when a Condition*= check does not succeed, restoring the JobRemoved
3577 signal to the behaviour it had before v250.
3579 * The org.freedesktop.portable1 methods GetMetadataWithExtensions() and
3580 GetImageMetadataWithExtensions() have been fixed to provide an extra
3581 return parameter, containing the actual extension release metadata.
3582 The current implementation was judged to be broken and unusable, and
3583 thus the usual procedure of adding a new set of methods was skipped,
3584 and backward compatibility broken instead on the assumption that
3585 nobody can be affected given the current state of this interface.
3587 * All kernels supported by systemd mix bytes returned by RDRAND (or
3588 similar) into the entropy pool at early boot. This means that on
3589 those systems, even if /dev/urandom is not yet initialized, it still
3590 returns bytes that are of at least RDRAND quality. For that reason,
3591 we no longer have reason to invoke RDRAND from systemd itself, which
3592 has historically been a source of bugs. Furthermore, kernels ≥5.6
3593 provide the getrandom(GRND_INSECURE) interface for returning random
3594 bytes before the entropy pool is initialized without warning into
3595 kmsg, which is what we attempt to use if available. systemd's direct
3596 usage of RDRAND has been removed. x86 systems ≥Broadwell that are
3597 running an older kernel may experience kmsg warnings that were not
3598 seen with 250. For newer kernels, non-x86 systems, or older x86
3599 systems, there should be no visible changes.
3601 * sd-boot will now measure the kernel command line into TPM PCR 12
3602 rather than PCR 8. This improves usefulness of the measurements on
3603 systems where sd-boot is chainloaded from Grub. Grub measures all
3604 commands its executes into PCR 8, which makes it very hard to use
3605 reasonably, hence separate ourselves from that and use PCR 12
3606 instead, which is what certain Ubuntu editions already do. To retain
3607 compatibility with systems running older systemd systems a new meson
3608 option 'efi-tpm-pcr-compat' has been added (which defaults to false).
3609 If enabled, the measurement is done twice: into the new-style PCR 12
3610 *and* the old-style PCR 8. It's strongly advised to migrate all users
3611 to PCR 12 for this purpose in the long run, as we intend to remove
3612 this compatibility feature in two years' time.
3614 * busctl capture now writes output in the newer pcapng format instead
3617 * A udev rule that imported hwdb matches for USB devices with lowercase
3618 hexadecimal vendor/product ID digits was added in systemd 250. This
3619 has been reverted, since uppercase hexadecimal digits are supposed to
3620 be used, and we already had a rule with the appropriate match.
3622 Users might need to adjust their local hwdb entries.
3624 * arch_prctl(2) has been moved to the @default set in the syscall filters
3625 (as exposed via the SystemCallFilter= setting in service unit files).
3626 It is apparently used by the linker now.
3628 * The tmpfiles entries that create the /run/systemd/netif directory and
3629 its subdirectories were moved from tmpfiles.d/systemd.conf to
3630 tmpfiles.d/systemd-network.conf.
3632 Users might need to adjust their files that override tmpfiles.d/systemd.conf
3633 to account for this change.
3635 * The requirement for Portable Services images to contain a well-formed
3636 os-release file (i.e.: contain at least an ID field) is now enforced.
3637 This applies to base images and extensions, and also to systemd-sysext.
3639 Changes in the Boot Loader Specification, kernel-install and sd-boot:
3641 * kernel-install's and bootctl's Boot Loader Specification Type #1
3642 entry generation logic has been reworked. The user may now pick
3643 explicitly by which "token" string to name the installation's boot
3644 entries, via the new /etc/kernel/entry-token file or the new
3645 --entry-token= switch to bootctl. By default — as before — the
3646 entries are named after the local machine ID. However, in "golden
3647 image" environments, where the machine ID shall be initialized on
3648 first boot (as opposed to at installation time before first boot) the
3649 machine ID will not be available at build time. In this case the
3650 --entry-token= switch to bootctl (or the /etc/kernel/entry-token
3651 file) may be used to override the "token" for the entries, for
3652 example the IMAGE_ID= or ID= fields from /etc/os-release. This will
3653 make the OS images independent of any machine ID, and ensure that the
3654 images will not carry any identifiable information before first boot,
3655 but on the other hand means that multiple parallel installations of
3656 the very same image on the same disk cannot be supported.
3658 Summary: if you are building golden images that shall acquire
3659 identity information exclusively on first boot, make sure to both
3660 remove /etc/machine-id *and* to write /etc/kernel/entry-token to the
3661 value of the IMAGE_ID= or ID= field of /etc/os-release or another
3662 suitable identifier before deploying the image.
3664 * The Boot Loader Specification has been extended with
3665 /loader/entries.srel file located in the EFI System Partition (ESP)
3666 that disambiguates the format of the entries in the /loader/entries/
3667 directory (in order to discern them from incompatible uses of this
3668 directory by other projects). For entries that follow the
3669 Specification, the string "type1" is stored in this file.
3671 bootctl will now write this file automatically when installing the
3672 systemd-boot boot loader.
3674 * kernel-install supports a new initrd_generator= setting in
3675 /etc/kernel/install.conf, that is exported as
3676 $KERNEL_INSTALL_INITRD_GENERATOR to kernel-install plugins. This
3677 allows choosing different initrd generators.
3679 * kernel-install will now create a "staging area" (an initially-empty
3680 directory to gather files for a Boot Loader Specification Type #1
3681 entry). The path to this directory is exported as
3682 $KERNEL_INSTALL_STAGING_AREA to kernel-install plugins, which should
3683 drop files there instead of writing them directly to the final
3684 location. kernel-install will move them when all files have been
3685 prepared successfully.
3687 * New option sort-key= has been added to the Boot Loader Specification
3688 to override the sorting order of the entries in the boot menu. It is
3689 read by sd-boot and bootctl, and will be written by kernel-install,
3690 with the default value of IMAGE_ID= or ID= fields from
3691 os-release. Together, this means that on multiboot installations,
3692 entries should be grouped and sorted in a predictable way.
3694 * The sort order of boot entries has been updated: entries which have
3695 the new field sort-key= are sorted by it first, and all entries
3696 without it are ordered later. After that, entries are sorted by
3697 version so that newest entries are towards the beginning of the list.
3699 * The kernel-install tool gained a new 'inspect' verb which shows the
3700 paths and other settings used.
3702 * sd-boot can now optionally beep when the menu is shown and menu
3703 entries are selected, which can be useful on machines without a
3704 working display. (Controllable via a loader.conf setting.)
3706 * The --make-machine-id-directory= switch to bootctl has been replaced
3707 by --make-entry-directory=, given that the entry directory is not
3708 necessarily named after the machine ID, but after some other suitable
3709 ID as selected via --entry-token= described above. The old name of
3710 the option is still understood to maximize compatibility.
3712 * 'bootctl list' gained support for a new --json= switch to output boot
3713 menu entries in JSON format.
3715 * 'bootctl is-installed' now supports the --graceful, and various verbs
3716 omit output with the new option --quiet.
3718 Changes in systemd-homed:
3720 * Starting with v250 systemd-homed uses UID/GID mapping on the mounts
3721 of activated home directories it manages (if the kernel and selected
3722 file systems support it). So far it mapped three UID ranges: the
3723 range from 0…60000, the user's own UID, and the range 60514…65534,
3724 leaving everything else unmapped (in other words, the 16-bit UID range
3725 is mapped almost fully, with the exception of the UID subrange used
3726 for systemd-homed users, with one exception: the user's own UID).
3727 Unmapped UIDs may not be used for file ownership in the home
3728 directory — any chown() attempts with them will fail. With this
3729 release a fourth range is added to these mappings:
3730 524288…1879048191. This range is the UID range intended for container
3733 https://systemd.io/UIDS-GIDS
3735 This range may be used for container managers that place container OS
3736 trees in the home directory (which is a questionable approach, for
3737 quota, permission, SUID handling and network file system
3738 compatibility reasons, but nonetheless apparently commonplace). Note
3739 that this mapping is mapped 1:1 in a pass-through fashion, i.e. the
3740 UID assignments from the range are not managed or mapped by
3741 `systemd-homed`, and must be managed with other mechanisms, in the
3742 context of the local system.
3744 Typically, a better approach to user namespacing in relevant
3745 container managers would be to leave container OS trees on disk at
3746 UID offset 0, but then map them to a dynamically allocated runtime
3747 UID range via another UID mount map at container invocation
3748 time. That way user namespace UID ranges become strictly a runtime
3749 concept, and do not leak into persistent file systems, persistent
3750 user databases or persistent configuration, thus greatly simplifying
3751 handling, and improving compatibility with home directories intended
3752 to be portable like the ones managed by systemd-homed.
3754 Changes in shared libraries:
3756 * A new libsystemd-core-<version>.so private shared library is
3757 installed under /usr/lib/systemd/system, mirroring the existing
3758 libsystemd-shared-<version>.so library. This allows the total
3759 installation size to be reduced by binary code reuse.
3761 * The <version> tag used in the name of libsystemd-shared.so and
3762 libsystemd-core.so can be configured via the meson option
3763 'shared-lib-tag'. Distributions may build subsequent versions of the
3764 systemd package with unique tags (e.g. the full package version),
3765 thus allowing multiple installations of those shared libraries to be
3766 available at the same time. This is intended to fix an issue where
3767 programs that link to those libraries would fail to execute because
3768 they were installed earlier or later than the appropriate version of
3771 * The sd-id128 API gained a new call sd_id128_to_uuid_string() that is
3772 similar to sd_id128_to_string() but formats the ID in RFC 4122 UUID
3773 format instead of as a simple series of hex characters.
3775 * The sd-device API gained two new calls sd_device_new_from_devname()
3776 and sd_device_new_from_path() which permit allocating an sd_device
3777 object from a device node name or file system path.
3779 * sd-device also gained a new call sd_device_open() which will open the
3780 device node associated with a device for which an sd_device object
3781 has been allocated. The call is supposed to address races around
3782 device nodes being removed/recycled due to hotplug events, or media
3783 change events: the call checks internally whether the major/minor of
3784 the device node and the "diskseq" (in case of block devices) match
3785 with the metadata loaded in the sd_device object, thus ensuring that
3786 the device once opened really matches the provided sd_device object.
3788 Changes in PID1, systemctl, and systemd-oomd:
3790 * A new set of service monitor environment variables will be passed to
3791 OnFailure=/OnSuccess= handlers, but only if exactly one unit lists the
3792 handler unit as OnFailure=/OnSuccess=. The variables are:
3793 $MONITOR_SERVICE_RESULT, $MONITOR_EXIT_CODE, $MONITOR_EXIT_STATUS,
3794 $MONITOR_INVOCATION_ID and $MONITOR_UNIT. For cases when a single
3795 handler needs to watch multiple units, use a templated handler.
3797 * A new ExtensionDirectories= setting in service unit files allows
3798 system extensions to be loaded from a directory. (It is similar to
3799 ExtensionImages=, but takes paths to directories, instead of
3802 'portablectl attach --extension=' now also accepts directory paths.
3804 * The user.delegate and user.invocation_id extended attributes on
3805 cgroups are used in addition to trusted.delegate and
3806 trusted.invocation_id. The latter pair requires privileges to set,
3807 but the former doesn't and can be also set by the unprivileged user
3810 (Only supported on kernels ≥5.6.)
3812 * Units that were killed by systemd-oomd will now have a service result
3813 of 'oom-kill'. The number of times a service was killed is tallied
3814 in the 'user.oomd_ooms' extended attribute.
3816 The OOMPolicy= unit file setting is now also honoured by
3819 * In unit files the new %y/%Y specifiers can be used to refer to
3820 normalized unit file path, which is particularly useful for symlinked
3823 The new %q specifier resolves to the pretty hostname
3824 (i.e. PRETTY_HOSTNAME= from /etc/machine-info).
3826 The new %d specifier resolves to the credentials directory of a
3827 service (same as $CREDENTIALS_DIRECTORY).
3829 * The RootDirectory=, MountAPIVFS=, ExtensionDirectories=,
3830 *Capabilities*=, ProtectHome=, *Directory=, TemporaryFileSystem=,
3831 PrivateTmp=, PrivateDevices=, PrivateNetwork=, NetworkNamespacePath=,
3832 PrivateIPC=, IPCNamespacePath=, PrivateUsers=, ProtectClock=,
3833 ProtectKernelTunables=, ProtectKernelModules=, ProtectKernelLogs=,
3834 MountFlags= service settings now also work in unprivileged user
3835 services, i.e. those run by the user's --user service manager, as long
3836 as user namespaces are enabled on the system.
3838 * Services with Restart=always and a failing ExecCondition= will no
3839 longer be restarted, to bring ExecCondition= behaviour in line with
3840 Condition*= settings.
3842 * LoadCredential= now accepts a directory as the argument; all files
3843 from the directory will be loaded as credentials.
3845 * A new D-Bus property ControlGroupId is now exposed on service units,
3846 that encapsulates the service's numeric cgroup ID that newer kernels
3847 assign to each cgroup.
3849 * PID 1 gained support for configuring the "pre-timeout" of watchdog
3850 devices and the associated governor, via the new
3851 RuntimeWatchdogPreSec= and RuntimeWatchdogPreGovernor= configuration
3852 options in /etc/systemd/system.conf.
3854 * systemctl's --timestamp= option gained a new choice "unix", to show
3855 timestamp as unix times, i.e. seconds since 1970, Jan 1st.
3857 * A new "taint" flag named "old-kernel" is introduced which is set when
3858 the kernel systemd runs on is older then the current baseline version
3859 (see above). The flag is shown in "systemctl status" output.
3861 * Two additional taint flags "short-uid-range" and "short-gid-range"
3862 have been added as well, which are set when systemd notices it is run
3863 within a userns namespace that does not define the full 0…65535 UID
3866 * A new "unmerged-usr" taint flag has been added that is set whenever
3867 running on systems where /bin/ + /sbin/ are *not* symlinks to their
3868 counterparts in /usr/, i.e. on systems where the /usr/-merge has not
3871 * Generators invoked by PID 1 will now have a couple of useful
3872 environment variables set describing the execution context a
3873 bit. $SYSTEMD_SCOPE encodes whether the generator is called from the
3874 system service manager, or from the per-user service
3875 manager. $SYSTEMD_IN_INITRD encodes whether the generator is invoked
3876 in initrd context or on the host. $SYSTEMD_FIRST_BOOT encodes whether
3877 systemd considers the current boot to be a "first"
3878 boot. $SYSTEMD_VIRTUALIZATION encode whether virtualization is
3879 detected and which type of hypervisor/container
3880 manager. $SYSTEMD_ARCHITECTURE indicates which architecture the
3881 kernel is built for.
3883 * PID 1 will now automatically pick up system credentials from qemu's
3884 fw_cfg interface, thus allowing passing arbitrary data into VM
3885 systems similar to how this is already supported for passing them
3886 into `systemd-nspawn` containers. Credentials may now also be passed
3887 in via the new kernel command line option `systemd.set_credential=`
3888 (note that kernel command line options are world-readable during
3889 runtime, and only useful for credentials that require no
3890 confidentiality). The credentials that can be passed to unified
3891 kernels that use the `systemd-stub` UEFI stub are now similarly
3892 picked up automatically. Automatic importing of system credentials
3893 this way can be turned off via the new
3894 `systemd.import_credentials=no` kernel command line option.
3896 * LoadCredential= will now automatically look for credentials in the
3897 /etc/credstore/, /run/credstore/, /usr/lib/credstore/ directories if
3898 the argument is not an absolute path. Similarly,
3899 LoadCredentialEncrypted= will check the same directories plus
3900 /etc/credstore.encrypted/, /run/credstore.encrypted/ and
3901 /usr/lib/credstore.encrypted/. The idea is to use those directories
3902 as the system-wide location for credentials that services should pick
3905 * System and service credentials are described in great detail in a new
3908 https://systemd.io/CREDENTIALS
3910 Changes in systemd-journald:
3912 * The journal JSON export format has been added to listed of stable
3913 interfaces (https://systemd.io/PORTABILITY_AND_STABILITY/).
3915 * journalctl --list-boots now supports JSON output and the --reverse option.
3917 * Under docs/: JOURNAL_EXPORT_FORMATS was imported from the wiki and
3918 updated, BUILDING_IMAGES is new:
3920 https://systemd.io/JOURNAL_EXPORT_FORMATS
3921 https://systemd.io/BUILDING_IMAGES
3925 * Two new hwdb files have been added. One lists "handhelds" (PDAs,
3926 calculators, etc.), the other AV production devices (DJ tables,
3927 keypads, etc.) that should accessible to the seat owner user by
3930 * udevadm trigger gained a new --prioritized-subsystem= option to
3931 process certain subsystems (and all their parent devices) earlier.
3933 systemd-udev-trigger.service now uses this new option to trigger
3934 block and TPM devices first, hopefully making the boot a bit faster.
3936 * udevadm trigger now implements --type=all, --initialized-match,
3937 --initialized-nomatch to trigger both subsystems and devices, only
3938 already-initialized devices, and only devices which haven't been
3939 initialized yet, respectively.
3941 * udevadm gained a new "wait" command for safely waiting for a specific
3942 device to show up in the udev device database. This is useful in
3943 scripts that asynchronously allocate a block device (e.g. through
3944 repartitioning, or allocating a loopback device or similar) and need
3945 to synchronize on the creation to complete.
3947 * udevadm gained a new "lock" command for locking one or more block
3948 devices while formatting it or writing a partition table to it. It is
3949 an implementation of https://systemd.io/BLOCK_DEVICE_LOCKING and
3950 usable in scripts dealing with block devices.
3952 * udevadm info will show a couple of additional device fields in its
3953 output, and will not apply a limited set of coloring to line types.
3955 * udevadm info --tree will now show a tree of objects (i.e. devices and
3956 suchlike) in the /sys/ hierarchy.
3958 * Block devices will now get a new set of device symlinks in
3959 /dev/disk/by-diskseq/<nr>, which may be used to reference block
3960 device nodes via the kernel's "diskseq" value. Note that this does
3961 not guarantee that opening a device by a symlink like this will
3962 guarantee that the opened device actually matches the specified
3963 diskseq value. To be safe against races, the actual diskseq value of
3964 the opened device (BLKGETDISKSEQ ioctl()) must still be compred with
3965 the one in the symlink path.
3967 * .link files gained support for setting MDI/MID-X on a link.
3969 * .link files gained support for [Match] Firmware= setting to match on
3970 the device firmware description string. By mistake, it was previously
3971 only supported in .network files.
3973 * .link files gained support for [Link] SR-IOVVirtualFunctions= setting
3974 and [SR-IOV] section to configure SR-IOV virtual functions.
3976 Changes in systemd-networkd:
3978 * The default scope for unicast routes configured through [Route]
3979 section is changed to "link", to make the behavior consistent with
3980 "ip route" command. The manual configuration of [Route] Scope= is
3983 * A new unit systemd-networkd-wait-online@<interface>.service has been
3984 added that can be used to wait for a specific network interface to be
3987 * systemd-networkd gained a new [Bridge] Isolated=true|false setting
3988 that configures the eponymous kernel attribute on the bridge.
3990 * .netdev files now can be used to create virtual WLAN devices, and
3991 configure various settings on them, via the [WLAN] section.
3993 * .link/.network files gained support for [Match] Kind= setting to match
3994 on device kind ("bond", "bridge", "gre", "tun", "veth", etc.)
3996 This value is also shown by 'networkctl status'.
3998 * The Local= setting in .netdev files for various virtual network
3999 devices gained support for specifying, in addition to the network
4000 address, the name of a local interface which must have the specified
4003 * systemd-networkd gained a new [Tunnel] External= setting in .netdev
4004 files, to configure tunnels in external mode (a.k.a. collect metadata
4007 * [Network] L2TP= setting was removed. Please use interface specifier in
4008 Local= setting in .netdev files of corresponding L2TP interface.
4010 * New [DHCPServer] BootServerName=, BootServerAddress=, and
4011 BootFilename= settings can be used to configure the server address,
4012 server name, and file name sent in the DHCP packet (e.g. to configure
4015 Changes in systemd-resolved:
4017 * systemd-resolved is started earlier (in sysinit.target), so it
4018 available earlier and will also be started in the initrd if installed
4021 Changes in disk encryption:
4023 * systemd-cryptenroll can now control whether to require the user to
4024 enter a PIN when using TPM-based unlocking of a volume via the new
4025 --tpm2-with-pin= option.
4027 Option tpm2-pin= can be used in /etc/crypttab.
4029 * When unlocking devices via TPM, TPM2 parameter encryption is now
4030 used, to ensure that communication between CPU and discrete TPM chips
4031 cannot be eavesdropped to acquire disk encryption keys.
4033 * A new switch --fido2-credential-algorithm= has been added to
4034 systemd-cryptenroll allowing selection of the credential algorithm to
4035 use when binding encryption to FIDO2 tokens.
4037 Changes in systemd-hostnamed:
4039 * HARDWARE_VENDOR= and HARDWARE_MODEL= can be set in /etc/machine-info
4040 to override the values gleaned from the hwdb.
4042 * A ID_CHASSIS property can be set in the hwdb (for the DMI device
4043 /sys/class/dmi/id) to override the chassis that is reported by
4046 * hostnamed's D-Bus interface gained a new method GetHardwareSerial()
4047 for reading the hardware serial number, as reportd by DMI. It also
4048 exposes a new method D-Bus property FirmwareVersion that encode the
4049 firmware version of the system.
4051 Changes in other components:
4053 * /etc/locale.conf is now populated through tmpfiles.d factory /etc/
4054 handling with the values that were configured during systemd build
4055 (if /etc/locale.conf has not been created through some other
4056 mechanism). This means that /etc/locale.conf should always have
4057 reasonable contents and we avoid a potential mismatch in defaults.
4059 * The userdbctl tool will now show UID range information as part of the
4060 list of known users.
4062 * A new build-time configuration setting default-user-shell= can be
4063 used to set the default shell for user records and nspawn shell
4064 invocations (instead of the default /bin/bash).
4066 * systemd-timesyncd now provides a D-Bus API for receiving NTP server
4067 information dynamically at runtime via IPC.
4069 * The systemd-creds tool gained a new "has-tpm2" verb, which reports
4070 whether a functioning TPM2 infrastructure is available, i.e. if
4071 firmware, kernel driver and systemd all have TPM2 support enabled and
4074 * The systemd-creds tool gained support for generating encrypted
4075 credentials that are using an empty encryption key. While this
4076 provides no integrity nor confidentiality it's useful to implement
4077 codeflows that work the same on TPM-ful and TPM2-less systems. The
4078 service manager will only accept credentials "encrypted" that way if
4079 a TPM2 device cannot be detected, to ensure that credentials
4080 "encrypted" like that cannot be used to trick TPM2 systems.
4082 * When deciding whether to colorize output, all systemd programs now
4083 also check $COLORTERM (in addition to $NO_COLOR, $SYSTEMD_COLORS, and
4086 * Meson's new install_tag feature is now in use for several components,
4087 allowing to build and install select binaries only: pam, nss, devel
4088 (pkg-config files), systemd-boot, libsystemd, libudev. Example:
4089 $ meson build systemd-boot
4090 $ meson install --tags systemd-boot --no-rebuild
4091 https://mesonbuild.com/Installing.html#installation-tags
4093 * A new build configuration option has been added, to allow selecting the
4094 default compression algorithm used by systemd-journald and systemd-coredump.
4095 This allows to build-in support for decompressing all supported formats,
4096 but choose a specific one for compression. E.g.:
4097 $ meson -Ddefault-compression=xz
4099 Experimental features:
4101 * sd-boot gained a new *experimental* setting "reboot-for-bitlocker" in
4102 loader.conf that implements booting Microsoft Windows from the
4103 sd-boot in a way that first reboots the system, to reset the TPM
4104 PCRs. This improves compatibility with BitLocker's TPM use, as the
4105 PCRs will only record the Windows boot process, and not sd-boot
4106 itself, thus retaining the PCR measurements not involving sd-boot.
4107 Note that this feature is experimental for now, and is likely going
4108 to be generalized and renamed in a future release, without retaining
4109 compatibility with the current implementation.
4111 * A new systemd-sysupdate component has been added that automatically
4112 discovers, downloads, and installs A/B-style updates for the host
4113 installation itself, or container images, portable service images,
4114 and other assets. See the new systemd-sysupdate man page for updates.
4116 Contributions from: 4piu, Adam Williamson, adrian5, Albert Brox,
4117 AlexCatze, Alex Henrie, Alfonso Sánchez-Beato, Alice S,
4118 Alvin Šipraga, amarjargal, Amarjargal, Andrea Pappacoda,
4119 Andreas Rammhold, Andy Chi, Anita Zhang, Antonio Alvarez Feijoo,
4120 Arfrever Frehtes Taifersar Arahesis, ash, Bastien Nocera, Be,
4121 bearhoney, Ben Efros, Benjamin Berg, Benjamin Franzke,
4122 Brett Holman, Christian Brauner, Clyde Byrd III, Curtis Klein,
4123 Daan De Meyer, Daniele Medri, Daniel Mack, Danilo Krummrich,
4124 David, David Bond, Davide Cavalca, David Tardon, davijosw,
4125 dependabot[bot], Donald Chan, Dorian Clay, Eduard Tolosa,
4126 Elias Probst, Eli Schwartz, Erik Sjölund, Evgeny Vereshchagin,
4127 Federico Ceratto, Franck Bui, Frantisek Sumsal, Gaël PORTAY,
4128 Georges Basile Stavracas Neto, Gibeom Gwon, Goffredo Baroncelli,
4129 Grigori Goronzy, Hans de Goede, Heiko Becker, Hugo Carvalho,
4130 Jakob Lell, James Hilliard, Jan Janssen, Jason A. Donenfeld,
4131 Joan Bruguera, Joerie de Gram, Josh Triplett, Julia Kartseva,
4132 Kazuo Moriwaka, Khem Raj, ksa678491784, Lance, Lan Tian,
4133 Laura Barcziova, Lennart Poettering, Leviticoh, licunlong,
4134 Lidong Zhong, lincoln auster, Lubomir Rintel, Luca Boccassi,
4135 Luca BRUNO, lucagoc, Ludwig Nussel, Marcel Hellwig, march1993,
4136 Marco Scardovi, Mario Limonciello, Mariusz Tkaczyk,
4137 Markus Weippert, Martin, Martin Liska, Martin Wilck, Matija Skala,
4138 Matthew Blythe, Matthias Lisin, Matthijs van Duin, Matt Walton,
4139 Max Gautier, Michael Biebl, Michael Olbrich, Michal Koutný,
4140 Michal Sekletár, Mike Gilbert, MkfsSion, Morten Linderud,
4141 Nick Rosbrook, Nikolai Grigoriev, Nikolai Kostrigin,
4142 Nishal Kulkarni, Noel Kuntze, Pablo Ceballos, Peter Hutterer,
4143 Peter Morrow, Pigmy-penguin, Piotr Drąg, prumian, Richard Neill,
4144 Rike-Benjamin Schuppner, rodin-ia, Romain Naour, Ruben Kerkhof,
4145 Ryan Hendrickson, Santa Wiryaman, Sebastian Pucilowski, Seth Falco,
4146 Simon Ellmann, Sonali Srivastava, Stefan Seering,
4147 Stephen Hemminger, tawefogo, techtino, Temuri Doghonadze,
4148 Thomas Batten, Thomas Haller, Thomas Weißschuh, Tobias Stoeckmann,
4149 Tomasz Pala, Tyson Whitehead, Vishal Chillara Srinivas,
4150 Vivien Didelot, w30023233, wangyuhang, Weblate, Xiaotian Wu,
4151 yangmingtai, YmrDtnJu, Yonathan Randolph, Yutsuten, Yu Watanabe,
4152 Zbigniew Jędrzejewski-Szmek, наб
4154 — Edinburgh, 2022-05-21
4158 * Support for encrypted and authenticated credentials has been added.
4159 This extends the credential logic introduced with v247 to support
4160 non-interactive symmetric encryption and authentication, based on a
4161 key that is stored on the /var/ file system or in the TPM2 chip (if
4162 available), or the combination of both (by default if a TPM2 chip
4163 exists the combination is used, otherwise the /var/ key only). The
4164 credentials are automatically decrypted at the moment a service is
4165 started, and are made accessible to the service itself in unencrypted
4166 form. A new tool 'systemd-creds' encrypts credentials for this
4167 purpose, and two new service file settings LoadCredentialEncrypted=
4168 and SetCredentialEncrypted= configure such credentials.
4170 This feature is useful to store sensitive material such as SSL
4171 certificates, passwords and similar securely at rest and only decrypt
4172 them when needed, and in a way that is tied to the local OS
4173 installation or hardware.
4175 * systemd-gpt-auto-generator can now automatically set up discoverable
4176 LUKS2 encrypted swap partitions.
4178 * The GPT Discoverable Partitions Specification has been substantially
4179 extended with support for root and /usr/ partitions for the majority
4180 of architectures systemd supports. This includes platforms that do
4181 not natively support UEFI, because even though GPT is specified under
4182 UEFI umbrella, it is useful on other systems too. Specifically,
4183 systemd-nspawn, systemd-sysext, systemd-gpt-auto-generator and
4184 Portable Services use the concept without requiring UEFI.
4186 * The GPT Discoverable Partitions Specifications has been extended with
4187 a new set of partitions that may carry PKCS#7 signatures for Verity
4188 partitions, encoded in a simple JSON format. This implements a simple
4189 mechanism for building disk images that are fully authenticated and
4190 can be tested against a set of cryptographic certificates. This is
4191 now implemented for the various systemd tools that can operate with
4192 disk images, such as systemd-nspawn, systemd-sysext, systemd-dissect,
4193 Portable services/RootImage=, systemd-tmpfiles, and systemd-sysusers.
4194 The PKCS#7 signatures are passed to the kernel (where they are
4195 checked against certificates from the kernel keyring), or can be
4196 verified against certificates provided in userspace (via a simple
4197 drop-in file mechanism).
4199 * systemd-dissect's inspection logic will now report for which uses a
4200 disk image is intended. Specifically, it will display whether an
4201 image is suitable for booting on UEFI or in a container (using
4202 systemd-nspawn's --image= switch), whether it can be used as portable
4203 service, or attached as system extension.
4205 * The system-extension.d/ drop-in files now support a new field
4206 SYSEXT_SCOPE= that may encode which purpose a system extension image
4207 is for: one of "initrd", "system" or "portable". This is useful to
4208 make images more self-descriptive, and to ensure system extensions
4209 cannot be attached in the wrong contexts.
4211 * The os-release file learnt a new PORTABLE_PREFIXES= field which may
4212 be used in portable service images to indicate which unit prefixes
4215 * The GPT image dissection logic in systemd-nspawn/systemd-dissect/…
4216 now is able to decode images for non-native architectures as well.
4217 This allows systemd-nspawn to boot images of non-native architectures
4218 if the corresponding user mode emulator is installed and
4219 systemd-binfmtd is running.
4221 * systemd-logind gained new settings HandlePowerKeyLongPress=,
4222 HandleRebootKeyLongPress=, HandleSuspendKeyLongPress= and
4223 HandleHibernateKeyLongPress= which may be used to configure actions
4224 when the relevant keys are pressed for more than 5s. This is useful
4225 on devices that only have hardware for a subset of these keys. By
4226 default, if the reboot key is pressed long the poweroff operation is
4227 now triggered, and when the suspend key is pressed long the hibernate
4228 operation is triggered. Long pressing the other two keys currently
4229 does not trigger any operation by default.
4231 * When showing unit status updates on the console during boot and
4232 shutdown, and a service is slow to start so that the cylon animation
4233 is shown, the most recent sd_notify() STATUS= text is now shown as
4234 well. Services may use this to make the boot/shutdown output easier
4235 to understand, and to indicate what precisely a service that is slow
4236 to start or stop is waiting for. In particular, the per-user service
4237 manager instance now reports what it is doing and which service it is
4238 waiting for this way to the system service manager.
4240 * The service manager will now re-execute on reception of the
4241 SIGRTMIN+25 signal. It previously already did that on SIGTERM — but
4242 only when running as PID 1. There was no signal to request this when
4243 running as per-user service manager, i.e. as any other PID than 1.
4244 SIGRTMIN+25 works for both system and user managers.
4246 * The hardware watchdog logic in PID 1 gained support for operating
4247 with the default timeout configured in the hardware, instead of
4248 insisting on re-configuring it. Set RuntimeWatchdogSec=default to
4249 request this behavior.
4251 * A new kernel command line option systemd.watchdog_sec= is now
4252 understood which may be used to override the hardware watchdog
4253 time-out for the boot.
4255 * A new setting DefaultOOMScoreAdjust= is now supported in
4256 /etc/systemd/system.conf and /etc/systemd/user.conf. It may be used
4257 to set the default process OOM score adjustment value for processes
4258 started by the service manager. For per-user service managers this
4259 now defaults to 100, but for per-system service managers is left as
4260 is. This means that by default now services forked off the user
4261 service manager are more likely to be killed by the OOM killer than
4262 system services or the managers themselves.
4264 * A new per-service setting RestrictFileSystems= as been added that
4265 restricts the file systems a service has access to by their type.
4266 This is based on the new BPF LSM of the Linux kernel. It provides an
4267 effective way to make certain API file systems unavailable to
4268 services (and thus minimizing attack surface). A new command
4269 "systemd-analyze filesystems" has been added that lists all known
4270 file system types (and how they are grouped together under useful
4273 * Services now support a new setting RestrictNetworkInterfaces= for
4274 restricting access to specific network interfaces.
4276 * Service unit files gained new settings StartupAllowedCPUs= and
4277 StartupAllowedMemoryNodes=. These are similar to their counterparts
4278 without the "Startup" prefix and apply during the boot process
4279 only. This is useful to improve boot-time behavior of the system and
4280 assign resources differently during boot than during regular
4281 runtime. This is similar to the preexisting StartupCPUWeight=
4284 * Related to this: the various StartupXYZ= settings
4285 (i.e. StartupCPUWeight=, StartupAllowedCPUs=, …) are now also applied
4286 during shutdown. The settings not prefixed with "Startup" hence apply
4287 during regular runtime, and those that are prefixed like that apply
4288 during boot and shutdown.
4290 * A new per-unit set of conditions/asserts
4291 [Condition|Assert][Memory|CPU|IO]Pressure= have been added to make a
4292 unit skip/fail activation if the system's (or a slice's) memory/cpu/io
4293 pressure is above the configured threshold, using the kernel PSI
4294 feature. For more details see systemd.unit(5) and
4295 https://docs.kernel.org/accounting/psi.html
4297 * The combination of ProcSubset=pid and ProtectKernelTunables=yes and/or
4298 ProtectKernelLogs=yes can now be used.
4300 * The default maximum numbers of inodes have been raised from 64k to 1M
4301 for /dev/, and from 400k to 1M for /tmp/.
4303 * The per-user service manager learnt support for communicating with
4304 systemd-oomd to acquire OOM kill information.
4306 * A new service setting ExecSearchPath= has been added that allows
4307 changing the search path for executables for services. It affects
4308 where we look for the binaries specified in ExecStart= and similar,
4309 and the specified directories are also added the $PATH environment
4310 variable passed to invoked processes.
4312 * A new setting RuntimeRandomizedExtraSec= has been added for service
4313 and scope units that allows extending the runtime time-out as
4314 configured by RuntimeMaxSec= with a randomized amount.
4316 * The syntax of the service unit settings RuntimeDirectory=,
4317 StateDirectory=, CacheDirectory=, LogsDirectory= has been extended:
4318 if the specified value is now suffixed with a colon, followed by
4319 another filename, the latter will be created as symbolic link to the
4320 specified directory. This allows creating these service directories
4321 together with alias symlinks to make them available under multiple
4324 * Service unit files gained two new settings TTYRows=/TTYColumns= for
4325 configuring rows/columns of the TTY device passed to
4326 stdin/stdout/stderr of the service. This is useful to propagate TTY
4327 dimensions to a virtual machine.
4329 * A new service unit file setting ExitType= has been added that
4330 specifies when to assume a service has exited. By default systemd
4331 only watches the main process of a service. By setting
4332 ExitType=cgroup it can be told to wait for the last process in a
4335 * Automount unit files gained a new setting ExtraOptions= that can be
4336 used to configure additional mount options to pass to the kernel when
4337 mounting the autofs instance.
4339 * "Urlification" (generation of ESC sequences that generate clickable
4340 hyperlinks in modern terminals) may now be turned off altogether
4343 * Path units gained new TriggerLimitBurst= and TriggerLimitIntervalSec=
4344 settings that default to 200 and 2 s respectively. The ratelimit
4345 ensures that a path unit cannot cause PID1 to busy-loop when it is
4346 trying to trigger a service that is skipped because of a Condition*=
4347 not being satisfied. This matches the configuration and behaviour of
4350 * The TPM2/FIDO2/PKCS11 support in systemd-cryptsetup is now also built
4351 as a plug-in for cryptsetup. This means the plain cryptsetup command
4352 may now be used to unlock volumes set up this way.
4354 * The TPM2 logic in cryptsetup will now automatically detect systems
4355 where the TPM2 chip advertises SHA256 PCR banks but the firmware only
4356 updates the SHA1 banks. In such a case PCR policies will be
4357 automatically bound to the latter, not the former. This makes the PCR
4358 policies reliable, but of course do not provide the same level of
4359 trust as SHA256 banks.
4361 * The TPM2 logic in systemd-cryptsetup/systemd-cryptsetup now supports
4362 RSA primary keys in addition to ECC, improving compatibility with
4363 TPM2 chips that do not support ECC. RSA keys are much slower to use
4364 than ECC, and hence are only used if ECC is not available.
4366 * /etc/crypttab gained support for a new token-timeout= setting for
4367 encrypted volumes that allows configuration of the maximum time to
4368 wait for PKCS#11/FIDO2 tokens to be plugged in. If the time elapses
4369 the logic will query the user for a regular passphrase/recovery key
4372 * Support for activating dm-integrity volumes at boot via a new file
4373 /etc/integritytab and the tool systemd-integritysetup have been
4374 added. This is similar to /etc/crypttab and /etc/veritytab, but deals
4375 with dm-integrity instead of dm-crypt/dm-verity.
4377 * The systemd-veritysetup-generator now understands a new usrhash=
4378 kernel command line option for specifying the Verity root hash for
4379 the partition backing the /usr/ file system. A matching set of
4380 systemd.verity_usr_* kernel command line options has been added as
4381 well. These all work similar to the corresponding options for the
4384 * The sd-device API gained a new API call sd_device_get_diskseq() to
4385 return the DISKSEQ property of a device structure. The "disk
4386 sequence" concept is a new feature recently introduced to the Linux
4387 kernel that allows detecting reuse cycles of block devices, i.e. can
4388 be used to recognize when loopback block devices are reused for a
4389 different purpose or CD-ROM drives get their media changed.
4391 * A new unit systemd-boot-update.service has been added. If enabled
4392 (the default) and the sd-boot loader is detected to be installed, it
4393 is automatically updated to the newest version when out of date. This
4394 is useful to ensure the boot loader remains up-to-date, and updates
4395 automatically propagate from the OS tree in /usr/.
4397 * sd-boot will now build with SBAT by default in order to facilitate
4398 working with recent versions of Shim that require it to be present.
4400 * sd-boot can now parse Microsoft Windows' Boot Configuration Data.
4401 This is used to robustly generate boot entry titles for Windows.
4403 * A new generic target unit factory-reset.target has been added. It is
4404 hooked into systemd-logind similar in fashion to
4405 reboot/poweroff/suspend/hibernate, and is supposed to be used to
4406 initiate a factory reset operation. What precisely this operation
4407 entails is up for the implementer to decide, the primary goal of the
4408 new unit is provide a framework where to plug in the implementation
4409 and how to trigger it.
4411 * A new meson build-time option 'clock-valid-range-usec-max' has been
4412 added which takes a time in µs and defaults to 15 years. If the RTC
4413 time is noticed to be more than the specified time ahead of the
4414 built-in epoch of systemd (which by default is the release timestamp
4415 of systemd) it is assumed that the RTC is not working correctly, and
4416 the RTC is reset to the epoch. (It already is reset to the epoch when
4417 noticed to be before it.) This should increase the chance that time
4418 doesn't accidentally jump too far ahead due to faulty hardware or
4421 * A new setting SaveIntervalSec= has been added to systemd-timesyncd,
4422 which may be used to automatically save the current system time to
4423 disk in regular intervals. This is useful to maintain a roughly
4424 monotonic clock even without RTC hardware and with some robustness
4425 against abnormal system shutdown.
4427 * systemd-analyze verify gained support for a pair of new --image= +
4428 --root= switches for verifying units below a specific root
4429 directory/image instead of on the host.
4431 * systemd-analyze verify gained support for verifying unit files under
4432 an explicitly specified unit name, independently of what the filename
4435 * systemd-analyze verify gained a new switch --recursive-errors= which
4436 controls whether to only fail on errors found in the specified units
4437 or recursively any dependent units.
4439 * systemd-analyze security now supports a new --offline mode for
4440 analyzing unit files stored on disk instead of loaded units. It may
4441 be combined with --root=/--image to analyze unit files under a root
4442 directory or disk image. It also learnt a new --threshold= parameter
4443 for specifying an exposure level threshold: if the exposure level
4444 exceeds the specified value the call will fail. It also gained a new
4445 --security-policy= switch for configuring security policies to
4446 enforce on the units. A policy is a JSON file that lists which tests
4447 shall be weighted how much to determine the overall exposure
4448 level. Altogether these new features are useful for fully automatic
4449 analysis and enforcement of security policies on unit files.
4451 * systemd-analyze security gain a new --json= switch for JSON output.
4453 * systemd-analyze learnt a new --quiet switch for reducing
4454 non-essential output. It's honored by the "dot", "syscall-filter",
4455 "filesystems" commands.
4457 * systemd-analyze security gained a --profile= option that can be used
4458 to take into account a portable profile when analyzing portable
4459 services, since a lot of the security-related settings are enabled
4462 * systemd-analyze learnt a new inspect-elf verb that parses ELF core
4463 files, binaries and executables and prints metadata information,
4464 including the build-id and other info described on:
4465 https://systemd.io/COREDUMP_PACKAGE_METADATA/
4467 * .network files gained a new UplinkInterface= in the [IPv6SendRA]
4468 section, for automatically propagating DNS settings from other
4471 * The static lease DHCP server logic in systemd-networkd may now serve
4472 IP addresses outside of the configured IP pool range for the server.
4474 * CAN support in systemd-networkd gained four new settings Loopback=,
4475 OneShot=, PresumeAck=, ClassicDataLengthCode= for tweaking CAN
4476 control modes. It gained a number of further settings for tweaking
4479 * The [CAN] section in .network file gained new TimeQuantaNSec=,
4480 PropagationSegment=, PhaseBufferSegment1=, PhaseBufferSegment2=,
4481 SyncJumpWidth=, DataTimeQuantaNSec=, DataPropagationSegment=,
4482 DataPhaseBufferSegment1=, DataPhaseBufferSegment2=, and
4483 DataSyncJumpWidth= settings to control bit-timing processed by the
4486 * DHCPv4 client support in systemd-networkd learnt a new Label= option
4487 for configuring the address label to apply to configure IPv4
4490 * The [IPv6AcceptRA] section of .network files gained support for a new
4491 UseMTU= setting that may be used to control whether to apply the
4492 announced MTU settings to the local interface.
4494 * The [DHCPv4] section in .network file gained a new Use6RD= boolean
4495 setting to control whether the DHCPv4 client request and process the
4498 * The [DHCPv6PrefixDelegation] section in .network file is renamed to
4499 [DHCPPrefixDelegation], as now the prefix delegation is also supported
4500 with DHCPv4 protocol by enabling the Use6RD= setting.
4502 * The [DHCPPrefixDelegation] section in .network file gained a new
4503 setting UplinkInterface= to specify the upstream interface.
4505 * The [DHCPv6] section in .network file gained a new setting
4506 UseDelegatedPrefix= to control whether the delegated prefixes will be
4507 propagated to the downstream interfaces.
4509 * The [IPv6AcceptRA] section of .network files now understands two new
4510 settings UseGateway=/UseRoutePrefix= for explicitly configuring
4511 whether to use the relevant fields from the IPv6 Router Advertisement
4514 * The ForceDHCPv6PDOtherInformation= setting in the [DHCPv6] section
4515 has been removed. Please use the WithoutRA= and UseDelegatedPrefix=
4516 settings in the [DHCPv6] section and the DHCPv6Client= setting in the
4517 [IPv6AcceptRA] section to control when the DHCPv6 client is started
4518 and how the delegated prefixes are handled by the DHCPv6 client.
4520 * The IPv6Token= section in the [Network] section is deprecated, and
4521 the [IPv6AcceptRA] section gained the Token= setting for its
4522 replacement. The [IPv6Prefix] section also gained the Token= setting.
4523 The Token= setting gained 'eui64' mode to explicitly configure an
4524 address with the EUI64 algorithm based on the interface MAC address.
4525 The 'prefixstable' mode can now optionally take a secret key. The
4526 Token= setting in the [DHCPPrefixDelegation] section now supports all
4527 algorithms supported by the same settings in the other sections.
4529 * The [RoutingPolicyRule] section of .network file gained a new
4530 SuppressInterfaceGroup= setting.
4532 * The IgnoreCarrierLoss= setting in the [Network] section of .network
4533 files now allows a duration to be specified, controlling how long to
4534 wait before reacting to carrier loss.
4536 * The [DHCPServer] section of .network file gained a new Router=
4537 setting to specify the router address.
4539 * The [CAKE] section of .network files gained various new settings
4540 AutoRateIngress=, CompensationMode=, FlowIsolationMode=, NAT=,
4541 MPUBytes=, PriorityQueueingPreset=, FirewallMark=, Wash=, SplitGSO=,
4542 and UseRawPacketSize= for configuring CAKE.
4544 * systemd-networkd now ships with new default .network files:
4545 80-container-vb.network which matches host-side network bridge device
4546 created by systemd-nspawn's --network-bridge or --network-zone
4547 switch, and 80-6rd-tunnel.network which matches automatically created
4548 sit tunnel with 6rd prefix when the DHCP 6RD option is received.
4550 * systemd-networkd's handling of Endpoint= resolution for WireGuard
4551 interfaces has been improved.
4553 * systemd-networkd will now automatically configure routes to addresses
4554 specified in AllowedIPs=. This feature can be controlled via
4555 RouteTable= and RouteMetric= settings in [WireGuard] or
4556 [WireGuardPeer] sections.
4558 * systemd-networkd will now once again automatically generate persistent
4559 MAC addresses for batadv and bridge interfaces. Users can disable this
4560 by using MACAddress=none in .netdev files.
4562 * systemd-networkd and systemd-udevd now support IP over InfiniBand
4563 interfaces. The Kind= setting in .netdev file accepts "ipoib". And
4564 systemd.netdev files gained the [IPoIB] section.
4566 * systemd-networkd and systemd-udevd now support net.ifname_policy=
4567 option on the kernel command-line. This is implemented through the
4568 systemd-network-generator service that automatically generates
4569 appropriate .link, .network, and .netdev files.
4571 * The various systemd-udevd "ethtool" buffer settings now understand
4572 the special value "max" to configure the buffers to the maximum the
4575 * systemd-udevd's .link files may now configure a large variety of
4576 NIC coalescing settings, plus more hardware offload settings.
4578 * .link files gained a new WakeOnLanPassword= setting in the [Link]
4579 section that allows to specify a WoL "SecureOn" password on hardware
4582 * systemd-nspawn's --setenv= switch now supports an additional syntax:
4583 if only a variable name is specified (i.e. without being suffixed by
4584 a '=' character and a value) the current value of the environment
4585 variable is propagated to the container. e.g. --setenv=FOO will
4586 lookup the current value of $FOO in the environment, and pass it down
4587 to the container. Similar behavior has been added to homectl's,
4588 machinectl's and systemd-run's --setenv= switch.
4590 * systemd-nspawn gained a new switch --suppress-sync= which may be used
4591 to optionally suppress the effect of the sync()/fsync()/fdatasync()
4592 system calls for the container payload. This is useful for build
4593 system environments where safety against abnormal system shutdown is
4594 not essential as all build artifacts can be regenerated any time, but
4595 the performance win is beneficial.
4597 * systemd-nspawn will now raise the RLIMIT_NOFILE hard limit to the
4598 same value that PID 1 uses for most forked off processes.
4600 * systemd-nspawn's --bind=/--bind-ro= switches now optionally take
4601 uidmap/nouidmap options as last parameter. If "uidmap" is used the
4602 bind mounts are created with UID mapping taking place that ensures
4603 the host's file ownerships are mapped 1:1 to container file
4604 ownerships, even if user namespacing is used. This way
4605 files/directories bound into containers will no longer show up as
4606 owned by the nobody user as they typically did if no special care was
4607 taken to shift them manually.
4609 * When discovering Windows installations sd-boot will now attempt to
4610 show the Windows version.
4612 * The color scheme to use in sd-boot may now be configured at
4615 * sd-boot gained the ability to change screen resolution during
4616 boot-time, by hitting the "r" key. This will cycle through available
4617 resolutions and save the last selection.
4619 * sd-boot learnt a new hotkey "f". When pressed the system will enter
4620 firmware setup. This is useful in environments where it is difficult
4621 to hit the right keys early enough to enter the firmware, and works
4622 on any firmware regardless which key it natively uses.
4624 * sd-boot gained support for automatically booting into the menu item
4625 selected on the last boot (using the "@saved" identifier for menu
4628 * sd-boot gained support for automatically loading all EFI drivers
4629 placed in the /EFI/systemd/drivers/ subdirectory of the EFI System
4630 Partition (ESP). These drivers are loaded before the menu entries are
4631 loaded. This is useful e.g. to load additional file system drivers
4632 for the XBOOTLDR partition.
4634 * systemd-boot will now paint the input cursor on its own instead of
4635 relying on the firmware to do so, increasing compatibility with broken
4636 firmware that doesn't make the cursor reasonably visible.
4638 * sd-boot now embeds a .osrel PE section like we expect from Boot
4639 Loader Specification Type #2 Unified Kernels. This means sd-boot
4640 itself may be used in place of a Type #2 Unified Kernel. This is
4641 useful for debugging purposes as it allows chain-loading one a
4642 (development) sd-boot instance from another.
4644 * sd-boot now supports a new "devicetree" field in Boot Loader
4645 Specification Type #1 entries: if configured the specified device
4646 tree file is installed before the kernel is invoked. This is useful
4647 for installing/applying new devicetree files without updating the
4650 * Similarly, sd-stub now can read devicetree data from a PE section
4651 ".dtb" and apply it before invoking the kernel.
4653 * sd-stub (the EFI stub that can be glued in front of a Linux kernel)
4654 gained the ability to pick up credentials and sysext files, wrap them
4655 in a cpio archive, and pass as an additional initrd to the invoked
4656 Linux kernel, in effect placing those files in the /.extra/ directory
4657 of the initrd environment. This is useful to implement trusted initrd
4658 environments which are fully authenticated but still can be extended
4659 (via sysexts) and parameterized (via encrypted/authenticated
4660 credentials, see above).
4662 Credentials can be located next to the kernel image file (credentials
4663 specific to a single boot entry), or in one of the shared directories
4664 (credentials applicable to multiple boot entries).
4666 * sd-stub now comes with a full man page, that explains its feature set
4667 and how to combine a kernel image, an initrd and the stub to build a
4668 complete EFI unified kernel image, implementing Boot Loader
4669 Specification Type #2.
4671 * sd-stub may now provide the initrd to the executed kernel via the
4672 LINUX_EFI_INITRD_MEDIA_GUID EFI protocol, adding compatibility for
4673 non-x86 architectures.
4675 * bootctl learnt new set-timeout and set-timeout-oneshot commands that
4676 may be used to set the boot menu time-out of the boot loader (for all
4677 or just the subsequent boot).
4679 * bootctl and kernel-install will now read variables
4680 KERNEL_INSTALL_LAYOUT= from /etc/machine-info and layout= from
4681 /etc/kernel/install.conf. When set, it specifies the layout to use
4682 for installation directories on the boot partition, so that tools
4683 don't need to guess it based on the already-existing directories. The
4684 only value that is defined natively is "bls", corresponding to the
4686 https://systemd.io/BOOT_LOADER_SPECIFICATION/. Plugins for
4687 kernel-install that implement a different layout can declare other
4688 values for this variable.
4690 'bootctl install' will now write KERNEL_INSTALL_LAYOUT=bls, on the
4691 assumption that if the user installed sd-boot to the ESP, they intend
4692 to use the entry layout understood by sd-boot. It'll also write
4693 KERNEL_INSTALL_MACHINE_ID= if it creates any directories using the ID
4694 (and it wasn't specified in the config file yet). Similarly,
4695 kernel-install will now write KERNEL_INSTALL_MACHINE_ID= (if it
4696 wasn't specified in the config file yet). Effectively, those changes
4697 mean that the machine-id used for boot loader entry installation is
4698 "frozen" upon first use and becomes independent of the actual
4701 Configuring KERNEL_INSTALL_MACHINE_ID fixes the following problem:
4702 images created for distribution ("golden images") are built with no
4703 machine-id, so that a unique machine-id can be created on the first
4704 boot. But those images may contain boot loader entries with the
4705 machine-id used during build included in paths. Using a "frozen"
4706 value allows unambiguously identifying entries that match the
4707 specific installation, while still permitting parallel installations
4710 Configuring KERNEL_INSTALL_LAYOUT obviates the need for
4711 kernel-install to guess the installation layout. This fixes the
4712 problem where a (possibly empty) directory in the boot partition is
4713 created from a different layout causing kernel-install plugins to
4714 assume the wrong layout. A particular example of how this may happen
4715 is the grub2 package in Fedora which includes directories under /boot
4716 directly in its file list. Various other packages pull in grub2 as a
4717 dependency, so it may be installed even if unused, breaking
4718 installations that use the bls layout.
4720 * bootctl and systemd-bless-boot can now be linked statically.
4722 * systemd-sysext now optionally doesn't insist on extension-release.d/
4723 files being placed in the image under the image's file name. If the
4724 file system xattr user.extension-release.strict is set on the
4725 extension release file, it is accepted regardless of its name. This
4726 relaxes security restrictions a bit, as system extension may be
4727 attached under a wrong name this way.
4729 * udevadm's test-builtin command learnt a new --action= switch for
4730 testing the built-in with the specified action (in place of the
4733 * udevadm info gained new switches --property=/--value for showing only
4734 specific udev properties/values instead of all.
4736 * A new hwdb database has been added that contains matches for various
4737 types of signal analyzers (protocol analyzers, logic analyzers,
4738 oscilloscopes, multimeters, bench power supplies, etc.) that should
4739 be accessible to regular users.
4741 * A new hwdb database entry has been added that carries information
4742 about types of cameras (regular or infrared), and in which direction
4743 they point (front or back).
4745 * A new rule to allow console users access to rfkill by default has been
4748 * Device nodes for the Software Guard eXtension enclaves (sgx_vepc) are
4749 now also owned by the system group "sgx".
4751 * A new build-time meson option "extra-net-naming-schemes=" has been
4752 added to define additional naming schemes for udev's network
4753 interface naming logic. This is useful for enterprise distributions
4754 and similar which want to pin the schemes of certain distribution
4755 releases under a specific name and previously had to patch the
4756 sources to introduce new named schemes.
4758 * The predictable naming logic for network interfaces has been extended
4759 to generate stable names from Xen netfront device information.
4761 * hostnamed's chassis property can now be sourced from chassis-type
4762 field encoded in devicetree (in addition to the existing DMI
4765 * systemd-cgls now optionally displays cgroup IDs and extended
4766 attributes for each cgroup. (Controllable via the new --xattr= +
4767 --cgroup-id= switches.)
4769 * coredumpctl gained a new --all switch for operating on all
4770 Journal files instead of just the local ones.
4772 * systemd-coredump will now use libdw/libelf via dlopen() rather than
4773 directly linking, allowing users to easily opt-out of backtrace/metadata
4774 analysis of core files, and reduce image sizes when this is not needed.
4776 * systemd-coredump will now analyze core files with libdw/libelf in a
4777 forked, sandboxed process.
4779 * systemd-homed will now try to unmount an activate home area in
4780 regular intervals once the user logged out fully. Previously this was
4781 attempted exactly once but if the home directory was busy for some
4782 reason it was not tried again.
4784 * systemd-homed's LUKS2 home area backend will now create a BSD file
4785 system lock on the image file while the home area is active
4786 (i.e. mounted). If a home area is found to be locked, logins are
4787 politely refused. This should improve behavior when using home areas
4788 images that are accessible via the network from multiple clients, and
4789 reduce the chance of accidental file system corruption in that case.
4791 * Optionally, systemd-homed will now drop the kernel buffer cache once
4792 a user has fully logged out, configurable via the new --drop-caches=
4795 * systemd-homed now makes use of UID mapped mounts for the home areas.
4796 If the kernel and used file system support it, files are now
4797 internally owned by the "nobody" user (i.e. the user typically used
4798 for indicating "this ownership is not mapped"), and dynamically
4799 mapped to the UID used locally on the system via the UID mapping
4800 mount logic of recent kernels. This makes migrating home areas
4801 between different systems cheaper because recursively chown()ing file
4802 system trees is no longer necessary.
4804 * systemd-homed's CIFS backend now optionally supports CIFS service
4805 names with a directory suffix, in order to place home directories in
4806 a subdirectory of a CIFS share, instead of the top-level directory.
4808 * systemd-homed's CIFS backend gained support for specifying additional
4809 mount options in the JSON user record (cifsExtraMountOptions field,
4810 and --cifs-extra-mount-options= homectl switch). This is for example
4811 useful for configuring mount options such as "noserverino" that some
4812 SMB3 services require (use that to run a homed home directory from a
4813 FritzBox SMB3 share this way).
4815 * systemd-homed will now default to btrfs' zstd compression for home
4816 areas. This is inspired by Fedora's recent decision to switch to zstd
4819 * Additional mount options to use when mounting the file system of
4820 LUKS2 volumes in systemd-homed has been added. Via the
4821 $SYSTEMD_HOME_MOUNT_OPTIONS_BTRFS, $SYSTEMD_HOME_MOUNT_OPTIONS_EXT4,
4822 $SYSTEMD_HOME_MOUNT_OPTIONS_XFS environment variables to
4823 systemd-homed or via the luksExtraMountOptions user record JSON
4824 property. (Exposed via homectl --luks-extra-mount-options)
4826 * homectl's resize command now takes the special size specifications
4827 "min" and "max" to shrink/grow the home area to the minimum/maximum
4828 size possible, taking disk usage/space constraints and file system
4829 limitations into account. Resizing is now generally graceful: the
4830 logic will try to get as close to the specified size as possible, but
4831 not consider it a failure if the request couldn't be fulfilled
4834 * systemd-homed gained the ability to automatically shrink home areas
4835 on logout to their minimal size and grow them again on next
4836 login. This ensures that while inactive, a home area only takes up
4837 the minimal space necessary, but once activated, it provides
4838 sufficient space for the user's needs. This behavior is only
4839 supported if btrfs is used as file system inside the home area
4840 (because only for btrfs online growing/shrinking is implemented in
4841 the kernel). This behavior is now enabled by default, but may be
4842 controlled via the new --auto-resize-mode= setting of homectl.
4844 * systemd-homed gained support for automatically re-balancing free disk
4845 space among active home areas, in case the LUKS2 backends are used,
4846 and no explicit disk size was requested. This way disk space is
4847 automatically managed and home areas resized in regular intervals and
4848 manual resizing when disk space becomes scarce should not be
4849 necessary anymore. This behavior is only supported if btrfs is used
4850 within the home areas (as only then online shrinking and growing is
4851 supported), and may be configured via the new rebalanceWeight JSON
4852 user record field (as exposed via the new --rebalance-weight= homectl
4853 setting). Re-balancing is mostly automatic, but can also be requested
4854 explicitly via "homectl rebalance", which is synchronous, and thus
4855 may be used to wait until the rebalance run is complete.
4857 * userdbctl gained a --json= switch for configured the JSON formatting
4858 to use when outputting user or group records.
4860 * userdbctl gained a new --multiplexer= switch for explicitly
4861 configuring whether to use the systemd-userdbd server side user
4862 record resolution logic.
4864 * userdbctl's ssh-authorized-keys command learnt a new --chain switch,
4865 for chaining up another command to execute after completing the
4866 look-up. Since the OpenSSH's AuthorizedKeysCommand only allows
4867 configuration of a single command to invoke, this maybe used to
4868 invoke multiple: first userdbctl's own implementation, and then any
4869 other also configured in the command line.
4871 * The sd-event API gained a new function sd_event_add_inotify_fd() that
4872 is similar to sd_event_add_inotify() but accepts a file descriptor
4873 instead of a path in the file system for referencing the inode to
4876 * The sd-event API gained a new function
4877 sd_event_source_set_ratelimit_expire_callback() that may be used to
4878 define a callback function that is called whenever an event source
4879 leaves the rate limiting phase.
4881 * New documentation has been added explaining which steps are necessary
4882 to port systemd to a new architecture:
4884 https://systemd.io/PORTING_TO_NEW_ARCHITECTURES
4886 * The x-systemd.makefs option in /etc/fstab now explicitly supports
4887 ext2, ext3, and f2fs file systems.
4889 * Mount units and units generated from /etc/fstab entries with 'noauto'
4890 are now ordered the same as other units. Effectively, they will be
4891 started earlier (if something actually pulled them in) and stopped
4892 later, similarly to normal mount units that are part of
4893 fs-local.target. This change should be invisible to users, but
4894 should prevent those units from being stopped too early during
4897 * The systemd-getty-generator now honors a new kernel command line
4898 argument systemd.getty_auto= and a new environment variable
4899 $SYSTEMD_GETTY_AUTO that allows turning it off at boot. This is for
4900 example useful to turn off gettys inside of containers or similar
4903 * systemd-resolved now listens on a second DNS stub address: 127.0.0.54
4904 (in addition to 127.0.0.53, as before). If DNS requests are sent to
4905 this address they are propagated in "bypass" mode only, i.e. are
4906 almost not processed locally, but mostly forwarded as-is to the
4907 current upstream DNS servers. This provides a stable DNS server
4908 address that proxies all requests dynamically to the right upstream
4909 DNS servers even if these dynamically change. This stub does not do
4910 mDNS/LLMNR resolution. However, it will translate look-ups to
4911 DNS-over-TLS if necessary. This new stub is particularly useful in
4912 container/VM environments, or for tethering setups: use DNAT to
4913 redirect traffic to any IP address to this stub.
4915 * systemd-importd now honors new environment variables
4916 $SYSTEMD_IMPORT_BTRFS_SUBVOL, $SYSTEMD_IMPORT_BTRFS_QUOTA,
4917 $SYSTEMD_IMPORT_SYNC, which may be used disable btrfs subvolume
4918 generation, btrfs quota setup and disk synchronization.
4920 * systemd-importd and systemd-resolved can now be optionally built with
4921 OpenSSL instead of libgcrypt.
4923 * systemd-repart no longer requires OpenSSL.
4925 * systemd-sysusers will no longer create the redundant 'nobody' group
4926 by default, as the 'nobody' user is already created with an
4927 appropriate primary group.
4929 * If a unit uses RuntimeMaxSec, systemctl show will now display it.
4931 * systemctl show-environment gained support for --output=json.
4933 * pam_systemd will now first try to use the X11 abstract socket, and
4934 fallback to the socket file in /tmp/.X11-unix/ only if that does not
4937 * systemd-journald will no longer go back to volatile storage
4938 regardless of configuration when its unit is restarted.
4940 * Initial support for the LoongArch architecture has been added (system
4941 call lists, GPT partition table UUIDs, etc).
4943 * systemd-journald's own logging messages are now also logged to the
4944 journal itself when systemd-journald logs to /dev/kmsg.
4946 * systemd-journald now re-enables COW for archived journal files on
4947 filesystems that support COW. One benefit of this change is that
4948 archived journal files will now get compressed on btrfs filesystems
4949 that have compression enabled.
4951 * systemd-journald now deduplicates fields in a single log message
4952 before adding it to the journal. In archived journal files, it will
4953 also punch holes for unused parts and truncate the file as
4954 appropriate, leading to reductions in disk usage.
4956 * journalctl --verify was extended with more informative error
4959 * More of sd-journal's functions are now resistant against journal file
4962 * The shutdown command learnt a new option --show, to display the
4965 * A LICENSES/ directory is now included in the git tree. It contains a
4966 README.md file that explains the licenses used by source files in
4967 this repository. It also contains the text of all applicable
4968 licenses as they appear on spdx.org.
4970 Contributions from: Aakash Singh, acsfer, Adolfo Jayme Barrientos,
4971 Adrian Vovk, Albert Brox, Alberto Mardegan, Alexander Kanavin,
4972 alexlzhu, Alfonso Sánchez-Beato, Alvin Šipraga, Alyssa Ross,
4973 Amir Omidi, Anatol Pomozov, Andika Triwidada, Andreas Rammhold,
4974 Andreas Valder, Andrej Lajovic, Andrew Soutar, Andrew Stone, Andy Chi,
4975 Anita Zhang, Anssi Hannula, Antonio Alvarez Feijoo,
4976 Antony Deepak Thomas, Arnaud Ferraris, Arvid E. Picciani,
4977 Bastien Nocera, Benjamin Berg, Benjamin Herrenschmidt, Ben Stockett,
4978 Bogdan Seniuc, Boqun Feng, Carl Lei, chlorophyll-zz, Chris Packham,
4979 Christian Brauner, Christian Göttsche, Christian Wehrli,
4980 Christoph Anton Mitterer, Cristian Rodríguez, Daan De Meyer,
4981 Daniel Maixner, Dann Frazier, Dan Streetman, Davide Cavalca,
4982 David Seifert, David Tardon, dependabot[bot], Dimitri John Ledkov,
4983 Dimitri Papadopoulos, Dimitry Ishenko, Dmitry Khlebnikov,
4984 Dominique Martinet, duament, Egor, Egor Ignatov, Emil Renner Berthing,
4985 Emily Gonyer, Ettore Atalan, Evgeny Vereshchagin, Florian Klink,
4986 Franck Bui, Frantisek Sumsal, Geass-LL, Gibeom Gwon, GnunuX,
4987 Gogo Gogsi, gregzuro, Greg Zuro, Gustavo Costa, Hans de Goede,
4988 Hela Basa, Henri Chain, hikigaya58, Hugo Carvalho,
4989 Hugo Osvaldo Barrera, Iago Lopez Galeiras, Iago López Galeiras,
4990 I-dont-need-name, igo95862, Jack Dähn, James Hilliard, Jan Janssen,
4991 Jan Kuparinen, Jan Macku, Jan Palus, Jarkko Sakkinen, Jayce Fayne,
4992 jiangchuangang, jlempen, John Lindgren, Jonas Dreßler, Jonas Jelten,
4993 Jonas Witschel, Joris Hartog, José Expósito, Julia Kartseva,
4994 Kai-Heng Feng, Kai Wohlfahrt, Kay Siver Bø, KennthStailey,
4995 Kevin Kuehler, Kevin Orr, Khem Raj, Kristian Klausen, Kyle Laker,
4996 lainahai, LaserEyess, Lennart Poettering, Lia Lenckowski, longpanda,
4997 Luca Boccassi, Luca BRUNO, Ludwig Nussel, Lukas Senionis,
4998 Maanya Goenka, Maciek Borzecki, Marcel Menzel, Marco Scardovi,
4999 Marcus Harrison, Mark Boudreau, Matthijs van Duin, Mauricio Vásquez,
5000 Maxime de Roucy, Max Resch, MertsA, Michael Biebl, Michael Catanzaro,
5001 Michal Koutný, Michal Sekletár, Miika Karanki, Mike Gilbert,
5002 Milo Turner, ml, monosans, Nacho Barrientos, nassir90, Nishal Kulkarni,
5003 nl6720, Ondrej Kozina, Paulo Neves, Pavel Březina, pedro martelletto,
5004 Peter Hutterer, Peter Morrow, Piotr Drąg, Rasmus Villemoes, ratijas,
5005 Raul Tambre, rene, Riccardo Schirone, Robert-L-Turner, Robert Scheck,
5006 Ross Jennings, saikat0511, Scott Lamb, Scott Worley,
5007 Sergei Trofimovich, Sho Iizuka, Slava Bacherikov, Slimane Selyan Amiri,
5008 StefanBruens, Steven Siloti, svonohr, Taiki Sugawara, Takashi Sakamoto,
5009 Takuro Onoue, Thomas Blume, Thomas Haller, Thomas Mühlbacher,
5010 Tianlu Shao, Toke Høiland-Jørgensen, Tom Yan, Tony Asleson,
5011 Topi Miettinen, Ulrich Ölmann, Urs Ritzmann, Vincent Bernat,
5012 Vito Caputo, Vladimir Panteleev, WANG Xuerui, Wind/owZ, Wu Xiaotian,
5013 xdavidwu, Xiaotian Wu, xujing, yangmingtai, Yao Wei, Yao Wei (魏銘廷),
5014 Yegor Alexeyev, Yu Watanabe, Zbigniew Jędrzejewski-Szmek,
5015 Дамјан Георгиевски, наб
5017 — Warsaw, 2021-12-23
5021 * When operating on disk images via the --image= switch of various
5022 tools (such as systemd-nspawn or systemd-dissect), or when udev finds
5023 no 'root=' parameter on the kernel command line, and multiple
5024 suitable root or /usr/ partitions exist in the image, then a simple
5025 comparison inspired by strverscmp() is done on the GPT partition
5026 label, and the newest partition is picked. This permits a simple and
5027 generic whole-file-system A/B update logic where new operating system
5028 versions are dropped into partitions whose label is then updated with
5029 a matching version identifier.
5031 * systemd-sysusers now supports querying the passwords to set for the
5032 users it creates via the "credentials" logic introduced in v247: the
5033 passwd.hashed-password.<user> and passwd.plaintext-password.<user>
5034 credentials are consulted for the password to use (either in UNIX
5035 hashed form, or literally). By default these credentials are inherited
5036 down from PID1 (which in turn imports it from a container manager if
5037 there is one). This permits easy configuration of user passwords
5038 during first boot. Example:
5040 # systemd-nspawn -i foo.raw --volatile=yes --set-credential=passwd.plaintext-password.root:foo
5042 Note that systemd-sysusers operates in purely additive mode: it
5043 executes no operation if the declared users already exist, and hence
5044 doesn't set any passwords as effect of the command line above if the
5045 specified root user exists already in the image. (Note that
5046 --volatile=yes ensures it doesn't, though.)
5048 * systemd-firstboot now also supports querying various system
5049 parameters via the credential subsystems. Thus, as above this may be
5050 used to initialize important system parameters on first boot of
5051 previously unprovisioned images (i.e. images with a mostly empty
5054 * PID 1 may now show both the unit name and the unit description
5055 strings in its status output during boot. This may be configured with
5056 StatusUnitFormat=combined in system.conf or
5057 systemd.status-unit-format=combined on the kernel command line.
5059 * The systemd-machine-id-setup tool now supports a --image= switch for
5060 provisioning a machine ID file into an OS disk image, similar to how
5061 --root= operates on an OS file tree. This matches the existing switch
5062 of the same name for systemd-tmpfiles, systemd-firstboot, and
5063 systemd-sysusers tools.
5065 * Similarly, systemd-repart gained support for the --image= switch too.
5066 In combination with the existing --size= option, this makes the tool
5067 particularly useful for easily growing disk images in a single
5068 invocation, following the declarative rules included in the image
5071 * systemd-repart's partition configuration files gained support for a
5072 new switch MakeDirectories= which may be used to create arbitrary
5073 directories inside file systems that are created, before registering
5074 them in the partition table. This is useful in particular for root
5075 partitions to create mount point directories for other partitions
5076 included in the image. For example, a disk image that contains a
5077 root, /home/, and /var/ partitions, may set MakeDirectories=yes to
5078 create /home/ and /var/ as empty directories in the root file system
5079 on its creation, so that the resulting image can be mounted
5080 immediately, even in read-only mode.
5082 * systemd-repart's CopyBlocks= setting gained support for the special
5083 value "auto". If used, a suitable matching partition on the booted OS
5084 is found as source to copy blocks from. This is useful when
5085 implementing replicating installers, that are booted from one medium
5086 and then stream their own root partition onto the target medium.
5088 * systemd-repart's partition configuration files gained support for a
5089 Flags=, a ReadOnly= and a NoAuto= setting, allowing control of these
5090 GPT partition flags for the created partitions: this is useful for
5091 marking newly created partitions as read-only, or as not being
5092 subject for automatic mounting from creation on.
5094 * The /etc/os-release file has been extended with two new (optional)
5095 variables IMAGE_VERSION= and IMAGE_ID=, carrying identity and version
5096 information for OS images that are updated comprehensively and
5097 atomically as one image. Two new specifiers %M, %A now resolve to
5098 these two fields in the various configuration options that resolve
5101 * portablectl gained a new switch --extension= for enabling portable
5102 service images with extensions that follow the extension image
5103 concept introduced with v248, and thus allows layering multiple
5104 images when setting up the root filesystem of the service.
5106 * systemd-coredump will now extract ELF build-id information from
5107 processes dumping core and include it in the coredump report.
5108 Moreover, it will look for ELF .note.package sections with
5109 distribution packaging meta-information about the crashing process.
5110 This is useful to directly embed the rpm or deb (or any other)
5111 package name and version in ELF files, making it easy to match
5112 coredump reports with the specific package for which the software was
5113 compiled. This is particularly useful on environments with ELF files
5114 from multiple vendors, different distributions and versions, as is
5115 common today in our containerized and sand-boxed world. For further
5118 https://systemd.io/COREDUMP_PACKAGE_METADATA
5120 * A new udev hardware database has been added for FireWire devices
5123 * The "net_id" built-in of udev has been updated with three
5124 backwards-incompatible changes:
5126 - PCI hotplug slot names on s390 systems are now parsed as
5127 hexadecimal numbers. They were incorrectly parsed as decimal
5128 previously, or ignored if the name was not a valid decimal
5131 - PCI onboard indices up to 65535 are allowed. Previously, numbers
5132 above 16383 were rejected. This primarily impacts s390 systems,
5133 where values up to 65535 are used.
5135 - Invalid characters in interface names are replaced with "_".
5137 The new version of the net naming scheme is "v249". The previous
5138 scheme can be selected via the "net.naming_scheme=v247" kernel
5139 command line parameter.
5141 * sd-bus' sd_bus_is_ready() and sd_bus_is_open() calls now accept a
5142 NULL bus object, for which they will return false. Or in other words,
5143 an unallocated bus connection is neither ready nor open.
5145 * The sd-device API acquired a new API function
5146 sd_device_get_usec_initialized() that returns the monotonic time when
5147 the udev device first appeared in the database.
5149 * sd-device gained a new APIs sd_device_trigger_with_uuid() and
5150 sd_device_get_trigger_uuid(). The former is similar to
5151 sd_device_trigger() but returns a randomly generated UUID that is
5152 associated with the synthetic uevent generated by the call. This UUID
5153 may be read from the sd_device object a monitor eventually receives,
5154 via the sd_device_get_trigger_uuid(). This interface requires kernel
5155 4.13 or above to work, and allows tracking a synthetic uevent through
5156 the entire device management stack. The "udevadm trigger --settle"
5157 logic has been updated to make use of this concept if available to
5158 wait precisely for the uevents it generates. "udevadm trigger" also
5159 gained a new parameter --uuid that prints the UUID for each generated
5162 * sd-device also gained new APIs sd_device_new_from_ifname() and
5163 sd_device_new_from_ifindex() for allocating an sd-device object for
5164 the specified network interface. The former accepts an interface name
5165 (either a primary or an alternative name), the latter an interface
5168 * The native Journal protocol has been documented. Clients may talk
5169 this as alternative to the classic BSD syslog protocol for locally
5170 delivering log records to the Journal. The protocol has been stable
5171 for a long time and in fact been implemented already in a variety
5172 of alternative client libraries. This documentation makes the support
5175 https://systemd.io/JOURNAL_NATIVE_PROTOCOL
5177 * A new BPFProgram= setting has been added to service files. It may be
5178 set to a path to a loaded kernel BPF program, i.e. a path to a bpffs
5179 file, or a bind mount or symlink to one. This may be used to upload
5180 and manage BPF programs externally and then hook arbitrary systemd
5183 * The "home.arpa" domain that has been officially declared as the
5184 choice for domain for local home networks per RFC 8375 has been added
5185 to the default NTA list of resolved, since DNSSEC is generally not
5186 available on private domains.
5188 * The CPUAffinity= setting of unit files now resolves "%" specifiers.
5190 * A new ManageForeignRoutingPolicyRules= setting has been added to
5191 .network files which may be used to exclude foreign-created routing
5192 policy rules from systemd-networkd management.
5194 * systemd-network-wait-online gained two new switches -4 and -6 that
5195 may be used to tweak whether to wait for only IPv4 or only IPv6
5198 * .network files gained a new RequiredFamilyForOnline= setting to
5199 fine-tune whether to require an IPv4 or IPv6 address in order to
5200 consider an interface "online".
5202 * networkctl will now show an over-all "online" state in the per-link
5205 * In .network files a new OutgoingInterface= setting has been added to
5206 specify the output interface in bridge FDB setups.
5208 * In .network files the Multipath group ID may now be configured for
5209 [NextHop] entries, via the new Group= setting.
5211 * The DHCP server logic configured in .network files gained a new
5212 setting RelayTarget= that turns the server into a DHCP server relay.
5213 The RelayAgentCircuitId= and RelayAgentRemoteId= settings may be used
5214 to further tweak the DHCP relay behaviour.
5216 * The DHCP server logic also gained a new ServerAddress= setting in
5217 .network files that explicitly specifies the server IP address to
5218 use. If not specified, the address is determined automatically, as
5221 * The DHCP server logic in systemd-networkd gained support for static
5222 DHCP leases, configurable via the [DHCPServerStaticLease]
5223 section. This allows explicitly mapping specific MAC addresses to
5224 fixed IP addresses and vice versa.
5226 * The RestrictAddressFamilies= setting in service files now supports a
5227 new special value "none". If specified sockets of all address
5228 families will be made unavailable to services configured that way.
5230 * systemd-fstab-generator and systemd-repart have been updated to
5231 support booting from disks that carry only a /usr/ partition but no
5232 root partition yet, and where systemd-repart can add it in on the
5233 first boot. This is useful for implementing systems that ship with a
5234 single /usr/ file system, and whose root file system shall be set up
5235 and formatted on a LUKS-encrypted volume whose key is generated
5236 locally (and possibly enrolled in the TPM) during the first boot.
5238 * The [Address] section of .network files now accepts a new
5239 RouteMetric= setting that configures the routing metric to use for
5240 the prefix route created as effect of the address configuration.
5241 Similarly, the [DHCPv6PrefixDelegation] and [IPv6Prefix] sections
5242 gained matching settings for their prefix routes. (The option of the
5243 same name in the [DHCPv6] section is moved to [IPv6AcceptRA], since
5244 it conceptually belongs there; the old option is still understood for
5247 * The DHCPv6 IAID and DUID are now explicitly configurable in .network
5250 * A new udev property ID_NET_DHCP_BROADCAST on network interface
5251 devices is now honoured by systemd-networkd, controlling whether to
5252 issue DHCP offers via broadcasting. This is used to ensure that s390
5253 layer 3 network interfaces work out-of-the-box with systemd-networkd.
5255 * nss-myhostname and systemd-resolved will now synthesize address
5256 records for a new special hostname "_outbound". The name will always
5257 resolve to the local IP addresses most likely used for outbound
5258 connections towards the default routes. On multi-homed hosts this is
5259 useful to have a stable handle referring to "the" local IP address
5260 that matters most, to the point where this is defined.
5262 * The Discoverable Partition Specification has been updated with a new
5263 GPT partition flag "grow-file-system" defined for its partition
5264 types. Whenever partitions with this flag set are automatically
5265 mounted (i.e. via systemd-gpt-auto-generator or the --image= switch
5266 of systemd-nspawn or other tools; and as opposed to explicit mounting
5267 via /etc/fstab), the file system within the partition is
5268 automatically grown to the full size of the partition. If the file
5269 system size already matches the partition size this flag has no
5270 effect. Previously, this functionality has been available via the
5271 explicit x-systemd.growfs mount option, and this new flag extends
5272 this to automatically discovered mounts. A new GrowFileSystem=
5273 setting has been added to systemd-repart drop-in files that allows
5274 configuring this partition flag. This new flag defaults to on for
5275 partitions automatically created by systemd-repart, except if they
5276 are marked read-only. See the specification for further details:
5278 https://systemd.io/DISCOVERABLE_PARTITIONS
5280 * .network files gained a new setting RoutesToNTP= in the [DHCPv4]
5281 section. If enabled (which is the default), and an NTP server address
5282 is acquired through a DHCP lease on this interface an explicit route
5283 to this address is created on this interface to ensure that NTP
5284 traffic to the NTP server acquired on an interface is also routed
5285 through that interface. The pre-existing RoutesToDNS= setting that
5286 implements the same for DNS servers is now enabled by default.
5288 * A pair of service settings SocketBindAllow= + SocketBindDeny= have
5289 been added that may be used to restrict the network interfaces
5290 sockets created by the service may be bound to. This is implemented
5293 * A new ConditionFirmware= setting has been added to unit files to
5294 conditionalize on certain firmware features. At the moment it may
5295 check whether running on a UEFI system, a device.tree system, or if
5296 the system is compatible with some specified device-tree feature.
5298 * A new ConditionOSRelease= setting has been added to unit files to
5299 check os-release(5) fields. The "=", "!=", "<", "<=", ">=", ">"
5300 operators may be used to check if some field has some specific value
5301 or do an alphanumerical comparison. Equality comparisons are useful
5302 for fields like ID, but relative comparisons for fields like
5303 VERSION_ID or IMAGE_VERSION.
5305 * hostnamed gained a new Describe() D-Bus method that returns a JSON
5306 serialization of the host data it exposes. This is exposed via
5307 "hostnamectl --json=" to acquire a host identity description in JSON.
5308 It's our intention to add a similar features to most services and
5309 objects systemd manages, in order to simplify integration with
5310 program code that can consume JSON.
5312 * Similarly, networkd gained a Describe() method on its Manager and
5313 Link bus objects. This is exposed via "networkctl --json=".
5315 * hostnamectl's various "get-xyz"/"set-xyz" verb pairs
5316 (e.g. "hostnamectl get-hostname", "hostnamectl "set-hostname") have
5317 been replaced by a single "xyz" verb (e.g. "hostnamectl hostname")
5318 that is used both to get the value (when no argument is given), and
5319 to set the value (when an argument is specified). The old names
5320 continue to be supported for compatibility.
5322 * systemd-detect-virt and ConditionVirtualization= are now able to
5323 correctly identify Amazon EC2 environments.
5325 * The LogLevelMax= setting of unit files now applies not only to log
5326 messages generated *by* the service, but also to log messages
5327 generated *about* the service by PID 1. To suppress logs concerning a
5328 specific service comprehensively, set this option to a high log
5331 * bootctl gained support for a new --make-machine-id-directory= switch
5332 that allows precise control on whether to create the top-level
5333 per-machine directory in the boot partition that typically contains
5334 Type 1 boot loader entries.
5336 * During build SBAT data to include in the systemd-boot EFI PE binaries
5337 may be specified now.
5339 * /etc/crypttab learnt a new option "headless". If specified any
5340 requests to query the user interactively for passwords or PINs will
5341 be skipped. This is useful on systems that are headless, i.e. where
5342 an interactive user is generally not present.
5344 * /etc/crypttab also learnt a new option "password-echo=" that allows
5345 configuring whether the encryption password prompt shall echo the
5346 typed password and if so, do so literally or via asterisks. (The
5347 default is the same behaviour as before: provide echo feedback via
5350 * FIDO2 support in systemd-cryptenroll/systemd-cryptsetup and
5351 systemd-homed has been updated to allow explicit configuration of the
5352 "user presence" and "user verification" checks, as well as whether a
5353 PIN is required for authentication, via the new switches
5354 --fido2-with-user-presence=, --fido2-with-user-verification=,
5355 --fido2-with-client-pin= to systemd-cryptenroll and homectl. Which
5356 features are available, and may be enabled or disabled depends on the
5359 * systemd-nspawn's --private-user= switch now accepts the special value
5360 "identity" which configures a user namespacing environment with an
5361 identity mapping of 65535 UIDs. This means the container UID 0 is
5362 mapped to the host UID 0, and the UID 1 to host UID 1. On first look
5363 this doesn't appear to be useful, however it does reduce the attack
5364 surface a bit, since the resulting container will possess process
5365 capabilities only within its namespace and not on the host.
5367 * systemd-nspawn's --private-user-chown switch has been replaced by a
5368 more generic --private-user-ownership= switch that accepts one of
5369 three values: "chown" is equivalent to the old --private-user-chown,
5370 and "off" is equivalent to the absence of the old switch. The value
5371 "map" uses the new UID mapping mounts of Linux 5.12 to map ownership
5372 of files and directories of the underlying image to the chosen UID
5373 range for the container. "auto" is equivalent to "map" if UID mapping
5374 mount are supported, otherwise it is equivalent to "chown". The short
5375 -U switch systemd-nspawn now implies --private-user-ownership=auto
5376 instead of the old --private-user-chown. Effectively this means: if
5377 the backing file system supports UID mapping mounts the feature is
5378 now used by default if -U is used. Generally, it's a good idea to use
5379 UID mapping mounts instead of recursive chown()ing, since it allows
5380 running containers off immutable images (since no modifications of
5381 the images need to take place), and share images between multiple
5382 instances. Moreover, the recursive chown()ing operation is slow and
5383 can be avoided. Conceptually it's also a good thing if transient UID
5384 range uses do not leak into persistent file ownership anymore. TLDR:
5385 finally, the last major drawback of user namespacing has been
5386 removed, and -U should always be used (unless you use btrfs, where
5387 UID mapped mounts do not exist; or your container actually needs
5388 privileges on the host).
5390 * nss-systemd now synthesizes user and group shadow records in addition
5391 to the main user and group records. Thus, hashed passwords managed by
5392 systemd-homed are now accessible via the shadow database.
5394 * The userdb logic (and thus nss-systemd, and so on) now read
5395 additional user/group definitions in JSON format from the drop-in
5396 directories /etc/userdb/, /run/userdb/, /run/host/userdb/ and
5397 /usr/lib/userdb/. This is a simple and powerful mechanism for making
5398 additional users available to the system, with full integration into
5399 NSS including the shadow databases. Since the full JSON user/group
5400 record format is supported this may also be used to define users with
5401 resource management settings and other runtime settings that
5402 pam_systemd and systemd-logind enforce at login.
5404 * The userdbctl tool gained two new switches --with-dropin= and
5405 --with-varlink= which can be used to fine-tune the sources used for
5406 user database lookups.
5408 * systemd-nspawn gained a new switch --bind-user= for binding a host
5409 user account into the container. This does three things: the user's
5410 home directory is bind mounted from the host into the container,
5411 below the /run/userdb/home/ hierarchy. A free UID is picked in the
5412 container, and a user namespacing UID mapping to the host user's UID
5413 installed. And finally, a minimal JSON user and group record (along
5414 with its hashed password) is dropped into /run/host/userdb/. These
5415 records are picked up automatically by the userdb drop-in logic
5416 describe above, and allow the user to login with the same password as
5417 on the host. Effectively this means: if host and container run new
5418 enough systemd versions making a host user available to the container
5419 is trivially simple.
5421 * systemd-journal-gatewayd now supports the switches --user, --system,
5422 --merge, --file= that are equivalent to the same switches of
5423 journalctl, and permit exposing only the specified subset of the
5426 * The OnFailure= dependency between units is now augmented with a
5427 implicit reverse dependency OnFailureOf= (this new dependency cannot
5428 be configured directly it's only created as effect of an OnFailure=
5429 dependency in the reverse order — it's visible in "systemctl show"
5430 however). Similar, Slice= now has an reverse dependency SliceOf=,
5431 that is also not configurable directly, but useful to determine all
5432 units that are members of a slice.
5434 * A pair of new dependency types between units PropagatesStopTo= +
5435 StopPropagatedFrom= has been added, that allows propagation of unit
5436 stop events between two units. It operates similar to the existing
5437 PropagatesReloadTo= + ReloadPropagatedFrom= dependencies.
5439 * A new dependency type OnSuccess= has been added (plus the reverse
5440 dependency OnSuccessOf=, which cannot be configured directly, but
5441 exists only as effect of the reverse OnSuccess=). It is similar to
5442 OnFailure=, but triggers in the opposite case: when a service exits
5443 cleanly. This allows "chaining up" of services where one or more
5444 services are started once another service has successfully completed.
5446 * A new dependency type Upholds= has been added (plus the reverse
5447 dependency UpheldBy=, which cannot be configured directly, but exists
5448 only as effect of Upholds=). This dependency type is a stronger form
5449 of Wants=: if a unit has an UpHolds= dependency on some other unit
5450 and the former is active then the latter is started whenever it is
5451 found inactive (and no job is queued for it). This is an alternative
5452 to Restart= inside service units, but less configurable, and the
5453 request to uphold a unit is not encoded in the unit itself but in
5454 another unit that intends to uphold it.
5456 * The systemd-ask-password tool now also supports reading passwords
5457 from the credentials subsystem, via the new --credential= switch.
5459 * The systemd-ask-password tool learnt a new switch --emoji= which may
5460 be used to explicit control whether the lock and key emoji (🔐) is
5461 shown in the password prompt on suitable TTYs.
5463 * The --echo switch of systemd-ask-password now optionally takes a
5464 parameter that controls character echo. It may either show asterisks
5465 (default, as before), turn echo off entirely, or echo the typed
5466 characters literally.
5468 * The systemd-ask-password tool also gained a new -n switch for
5469 suppressing output of a trailing newline character when writing the
5470 acquired password to standard output, similar to /bin/echo's -n
5473 * New documentation has been added that describes the organization of
5474 the systemd source code tree:
5476 https://systemd.io/ARCHITECTURE
5478 * Units using ConditionNeedsUpdate= will no longer be activated in
5481 * It is now possible to list a template unit in the WantedBy= or
5482 RequiredBy= settings of the [Install] section of another template
5483 unit, which will be instantiated using the same instance name.
5485 * A new MemoryAvailable property is available for units. If the unit,
5486 or the slices it is part of, have a memory limit set via MemoryMax=/
5487 MemoryHigh=, MemoryAvailable will indicate how much more memory the
5488 unit can claim before hitting the limits.
5490 * systemd-coredump will now try to stay below the cgroup memory limit
5491 placed on itself or one of the slices it runs under, if the storage
5492 area for core files (/var/lib/systemd/coredump/) is placed on a tmpfs,
5493 since files written on such filesystems count toward the cgroup memory
5494 limit. If there is not enough available memory in such cases to store
5495 the core file uncompressed, systemd-coredump will skip to compressed
5496 storage directly (if enabled) and it will avoid analyzing the core file
5497 to print backtrace and metadata in the journal.
5499 * tmpfiles.d/ drop-ins gained a new '=' modifier to check if the type
5500 of a path matches the configured expectations, and remove it if not.
5502 * tmpfiles.d/'s 'Age' now accepts an 'age-by' argument, which allows to
5503 specify which of the several available filesystem timestamps (access
5504 time, birth time, change time, modification time) to look at when
5505 deciding whether a path has aged enough to be cleaned.
5507 * A new IPv6StableSecretAddress= setting has been added to .network
5508 files, which takes an IPv6 address to use as secret for IPv6 address
5511 * The [DHCPServer] logic in .network files gained support for a new
5512 UplinkInterface= setting that permits configuration of the uplink
5513 interface name to propagate DHCP lease information from.
5515 * The WakeOnLan= setting in .link files now accepts a list of flags
5516 instead of a single one, to configure multiple wake-on-LAN policies.
5518 * User-space defined tracepoints (USDT) have been added to udev at
5519 strategic locations. This is useful for tracing udev behaviour and
5520 performance with bpftrace and similar tools.
5522 * systemd-journald-upload gained a new NetworkTimeoutSec= option for
5523 setting a network timeout time.
5525 * If a system service is running in a new mount namespace (RootDirectory=
5526 and friends), all file systems will be mounted with MS_NOSUID by
5527 default, unless the system is running with SELinux enabled.
5529 * When enumerating time zones the timedatectl tool will now consult the
5530 'tzdata.zi' file shipped by the IANA time zone database package, in
5531 addition to 'zone1970.tab', as before. This makes sure time zone
5532 aliases are now correctly supported. Some distributions so far did
5533 not install this additional file, most do however. If you
5534 distribution does not install it yet, it might make sense to change
5537 * Intel HID rfkill event is no longer masked, since it's the only
5538 source of rfkill event on newer HP laptops. To have both backward and
5539 forward compatibility, userspace daemon needs to debounce duplicated
5540 events in a short time window.
5542 Contributions from: Aakash Singh, adrian5, Albert Brox,
5543 Alexander Sverdlin, Alexander Tsoy, Alexey Rubtsov, alexlzhu,
5544 Allen Webb, Alvin Šipraga, Alyssa Ross, Anders Wenhaug,
5545 Andrea Pappacoda, Anita Zhang, asavah, Balint Reczey, Bertrand Jacquin,
5546 borna-blazevic, caoxia2008cxx, Carlo Teubner, Christian Göttsche,
5547 Christian Hesse, Daniel Schaefer, Dan Streetman,
5548 David Santamaría Rogado, David Tardon, Deepak Rawat, dgcampea,
5549 Dimitri John Ledkov, ei-ke, Emilio Herrera, Emil Renner Berthing,
5550 Eric Cook, Flos Lonicerae, Franck Bui, Francois Gervais,
5551 Frantisek Sumsal, Gibeom Gwon, gitm0, Hamish Moffatt, Hans de Goede,
5552 Harsh Barsaiyan, Henri Chain, Hristo Venev, Icenowy Zheng, Igor Zhbanov,
5553 imayoda, Jakub Warczarek, James Buren, Jan Janssen, Jan Macku,
5554 Jan Synacek, Jason Francis, Jayanth Ananthapadmanaban, Jeremy Szu,
5555 Jérôme Carretero, Jesse Stricker, jiangchuangang, Joerg Behrmann,
5556 Jóhann B. Guðmundsson, Jörg Deckert, Jörg Thalheim, Juergen Hoetzel,
5557 Julia Kartseva, Kai-Heng Feng, Khem Raj, KoyamaSohei, laineantti,
5558 Lennart Poettering, LetzteInstanz, Luca Adrian L, Luca Boccassi,
5559 Lucas Magasweran, Mantas Mikulėnas, Marco Antonio Mauro, Mark Wielaard,
5560 Masahiro Matsuya, Matt Johnston, Michael Catanzaro, Michal Koutný,
5561 Michal Sekletár, Mike Crowe, Mike Kazantsev, Milan, milaq,
5562 Miroslav Suchý, Morten Linderud, nerdopolis, nl6720, Noah Meyerhans,
5563 Oleg Popov, Olle Lundberg, Ondrej Kozina, Paweł Marciniak, Perry.Yuan,
5564 Peter Hutterer, Peter Kjellerstedt, Peter Morrow, Phaedrus Leeds,
5565 plattrap, qhill, Raul Tambre, Roman Beranek, Roshan Shariff,
5566 Ryan Hendrickson, Samuel BF, scootergrisen, Sebastian Blunt,
5567 Seong-ho Cho, Sergey Bugaev, Sevan Janiyan, Sibo Dong, simmon,
5568 Simon Watts, Srinidhi Kaushik, Štěpán Němec, Steve Bonds, Susant Sahani,
5569 sverdlin, syyhao1994, Takashi Sakamoto, Topi Miettinen, tramsay,
5570 Trent Piepho, Uwe Kleine-König, Viktor Mihajlovski, Vincent Dechenaux,
5571 Vito Caputo, William A. Kennington III, Yangyang Shen, Yegor Alexeyev,
5572 Yi Gao, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, zsien, наб
5574 — Edinburgh, 2021-07-07
5578 * A concept of system extension images is introduced. Such images may
5579 be used to extend the /usr/ and /opt/ directory hierarchies at
5580 runtime with additional files (even if the file system is read-only).
5581 When a system extension image is activated, its /usr/ and /opt/
5582 hierarchies and os-release information are combined via overlayfs
5583 with the file system hierarchy of the host OS.
5585 A new systemd-sysext tool can be used to merge, unmerge, list, and
5586 refresh system extension hierarchies. See
5587 https://www.freedesktop.org/software/systemd/man/systemd-sysext.html.
5589 The systemd-sysext.service automatically merges installed system
5590 extensions during boot (before basic.target, but not in very early
5591 boot, since various file systems have to be mounted first).
5593 The SYSEXT_LEVEL= field in os-release(5) may be used to specify the
5594 supported system extension level.
5596 * A new ExtensionImages= unit setting can be used to apply the same
5597 system extension image concept from systemd-sysext to the namespaced
5598 file hierarchy of specific services, following the same rules and
5601 * Support for a new special "root=tmpfs" kernel command-line option has
5602 been added. When specified, a tmpfs is mounted on /, and mount.usr=
5603 should be used to point to the operating system implementation.
5605 * A new configuration file /etc/veritytab may be used to configure
5606 dm-verity integrity protection for block devices. Each line is in the
5607 format "volume-name data-device hash-device roothash options",
5608 similar to /etc/crypttab.
5610 * A new kernel command-line option systemd.verity.root_options= may be
5611 used to configure dm-verity behaviour for the root device.
5613 * The key file specified in /etc/crypttab (the third field) may now
5614 refer to an AF_UNIX/SOCK_STREAM socket in the file system. The key is
5615 acquired by connecting to that socket and reading from it. This
5616 allows the implementation of a service to provide key information
5617 dynamically, at the moment when it is needed.
5619 * When the hostname is set explicitly to "localhost", systemd-hostnamed
5620 will respect this. Previously such a setting would be mostly silently
5621 ignored. The goal is to honour configuration as specified by the
5624 * The fallback hostname that will be used by the system manager and
5625 systemd-hostnamed can now be configured in two new ways: by setting
5626 DEFAULT_HOSTNAME= in os-release(5), or by setting
5627 $SYSTEMD_DEFAULT_HOSTNAME in the environment block. As before, it can
5628 also be configured during compilation. The environment variable is
5629 intended for testing and local overrides, the os-release(5) field is
5630 intended to allow customization by different variants of a
5631 distribution that share the same compiled packages.
5633 * The environment block of the manager itself may be configured through
5634 a new ManagerEnvironment= setting in system.conf or user.conf. This
5635 complements existing ways to set the environment block (the kernel
5636 command line for the system manager, the inherited environment and
5637 user@.service unit file settings for the user manager).
5639 * systemd-hostnamed now exports the default hostname and the source of
5640 the configured hostname ("static", "transient", or "default") as
5643 * systemd-hostnamed now exports the "HardwareVendor" and
5644 "HardwareModel" D-Bus properties, which are supposed to contain a
5645 pair of cleaned up, human readable strings describing the system's
5646 vendor and model. It's typically sourced from the firmware's DMI
5647 tables, but may be augmented from a new hwdb database. hostnamectl
5648 shows this in the status output.
5650 * Support has been added to systemd-cryptsetup for extracting the
5651 PKCS#11 token URI and encrypted key from the LUKS2 JSON embedded
5652 metadata header. This allows the information how to open the
5653 encrypted device to be embedded directly in the device and obviates
5654 the need for configuration in an external file.
5656 * systemd-cryptsetup gained support for unlocking LUKS2 volumes using
5657 TPM2 hardware, as well as FIDO2 security tokens (in addition to the
5658 pre-existing support for PKCS#11 security tokens).
5660 * systemd-repart may enroll encrypted partitions using TPM2
5661 hardware. This may be useful for example to create an encrypted /var
5662 partition bound to the machine on first boot.
5664 * A new systemd-cryptenroll tool has been added to enroll TPM2, FIDO2
5665 and PKCS#11 security tokens to LUKS volumes, list and destroy
5668 https://0pointer.net/blog/unlocking-luks2-volumes-with-tpm2-fido2-pkcs11-security-hardware-on-systemd-248.html
5670 It also supports enrolling "recovery keys" and regular passphrases.
5672 * The libfido2 dependency is now based on dlopen(), so that the library
5673 is used at runtime when installed, but is not a hard runtime
5676 * systemd-cryptsetup gained support for two new options in
5677 /etc/crypttab: "no-write-workqueue" and "no-read-workqueue" which
5678 request synchronous processing of encryption/decryption IO.
5680 * The manager may be configured at compile time to use the fexecve()
5681 instead of the execve() system call when spawning processes. Using
5682 fexecve() closes a window between checking the security context of an
5683 executable and spawning it, but unfortunately the kernel displays
5684 stale information in the process' "comm" field, which impacts ps
5687 * The configuration option -Dcompat-gateway-hostname has been dropped.
5688 "_gateway" is now the only supported name.
5690 * The ConditionSecurity=tpm2 unit file setting may be used to check if
5691 the system has at least one TPM2 (tpmrm class) device.
5693 * A new ConditionCPUFeature= has been added that may be used to
5694 conditionalize units based on CPU features. For example,
5695 ConditionCPUFeature=rdrand will condition a unit so that it is only
5696 run when the system CPU supports the RDRAND opcode.
5698 * The existing ConditionControlGroupController= setting has been
5699 extended with two new values "v1" and "v2". "v2" means that the
5700 unified v2 cgroup hierarchy is used, and "v1" means that legacy v1
5701 hierarchy or the hybrid hierarchy are used.
5703 * A new PrivateIPC= setting on a unit file allows executed processes to
5704 be moved into a private IPC namespace, with separate System V IPC
5705 identifiers and POSIX message queues.
5707 A new IPCNamespacePath= allows the unit to be joined to an existing
5710 * The tables of system calls in seccomp filters are now automatically
5711 generated from kernel lists exported on
5712 https://fedora.juszkiewicz.com.pl/syscalls.html.
5714 The following architectures should now have complete lists:
5715 alpha, arc, arm64, arm, i386, ia64, m68k, mips64n32, mips64, mipso32,
5716 powerpc, powerpc64, s390, s390x, tilegx, sparc, x86_64, x32.
5718 * The MountAPIVFS= service file setting now additionally mounts a tmpfs
5719 on /run/ if it is not already a mount point. A writable /run/ has
5720 always been a requirement for a functioning system, but this was not
5721 guaranteed when using a read-only image.
5723 Users can always specify BindPaths= or InaccessiblePaths= as
5724 overrides, and they will take precedence. If the host's root mount
5725 point is used, there is no change in behaviour.
5727 * New bind mounts and file system image mounts may be injected into the
5728 mount namespace of a service (without restarting it). This is exposed
5729 respectively as 'systemctl bind <unit> <path>…' and
5730 'systemctl mount-image <unit> <image>…'.
5732 * The StandardOutput= and StandardError= settings can now specify files
5733 to be truncated for output (as "truncate:<path>").
5735 * The ExecPaths= and NoExecPaths= settings may be used to specify
5736 noexec for parts of the file system.
5738 * sd-bus has a new function sd_bus_open_user_machine() to open a
5739 connection to the session bus of a specific user in a local container
5740 or on the local host. This is exposed in the existing -M switch to
5741 systemctl and similar tools:
5743 systemctl --user -M lennart@foobar start foo
5745 This will connect to the user bus of a user "lennart" in container
5746 "foobar". If no container name is specified, the specified user on
5747 the host itself is connected to
5749 systemctl --user -M lennart@ start quux
5751 * sd-bus also gained a convenience function sd_bus_message_send() to
5752 simplify invocations of sd_bus_send(), taking only a single
5753 parameter: the message to send.
5755 * sd-event allows rate limits to be set on event sources, for dealing
5756 with high-priority event sources that might starve out others. See
5757 the new man page sd_event_source_set_ratelimit(3) for details.
5759 * systemd.link files gained a [Link] Promiscuous= switch, which allows
5760 the device to be raised in promiscuous mode.
5762 New [Link] TransmitQueues= and ReceiveQueues= settings allow the
5763 number of TX and RX queues to be configured.
5765 New [Link] TransmitQueueLength= setting allows the size of the TX
5766 queue to be configured.
5768 New [Link] GenericSegmentOffloadMaxBytes= and
5769 GenericSegmentOffloadMaxSegments= allow capping the packet size and
5770 the number of segments accepted in Generic Segment Offload.
5772 * systemd-networkd gained support for the "B.A.T.M.A.N. advanced"
5773 wireless routing protocol that operates on ISO/OSI Layer 2 only and
5774 uses ethernet frames to route/bridge packets. This encompasses a new
5775 "batadv" netdev Type=, a new [BatmanAdvanced] section with a bunch of
5776 new settings in .netdev files, and a new BatmanAdvanced= setting in
5779 * systemd.network files gained a [Network] RouteTable= configuration
5780 switch to select the routing policy table.
5782 systemd.network files gained a [RoutingPolicyRule] Type=
5783 configuration switch (one of "blackhole, "unreachable", "prohibit").
5785 systemd.network files gained a [IPv6AcceptRA] RouteDenyList= and
5786 RouteAllowList= settings to ignore/accept route advertisements from
5787 routers matching specified prefixes. The DenyList= setting has been
5788 renamed to PrefixDenyList= and a new PrefixAllowList= option has been
5791 systemd.network files gained a [DHCPv6] UseAddress= setting to
5792 optionally ignore the address provided in the lease.
5794 systemd.network files gained a [DHCPv6PrefixDelegation]
5795 ManageTemporaryAddress= switch.
5797 systemd.network files gained a new ActivationPolicy= setting which
5798 allows configuring how the UP state of an interface shall be managed,
5799 i.e. whether the interface is always upped, always downed, or may be
5800 upped/downed by the user using "ip link set dev".
5802 * The default for the Broadcast= setting in .network files has slightly
5803 changed: the broadcast address will not be configured for wireguard
5806 * systemd.netdev files gained a [VLAN] Protocol=, IngressQOSMaps=,
5807 EgressQOSMaps=, and [MACVLAN] BroadcastMulticastQueueLength=
5808 configuration options for VLAN packet handling.
5810 * udev rules may now set log_level= option. This allows debug logs to
5811 be enabled for select events, e.g. just for a specific subsystem or
5812 even a single device.
5814 * udev now exports the VOLUME_ID, LOGICAL_VOLUME_ID, VOLUME_SET_ID, and
5815 DATA_PREPARED_ID properties for block devices with ISO9660 file
5818 * udev now exports decoded DMI information about installed memory slots
5819 as device properties under the /sys/class/dmi/id/ pseudo device.
5821 * /dev/ is not mounted noexec anymore. This didn't provide any
5822 significant security benefits and would conflict with the executable
5823 mappings used with /dev/sgx device nodes. The previous behaviour can
5824 be restored for individual services with NoExecPaths=/dev (or by allow-
5825 listing and excluding /dev from ExecPaths=).
5827 * Permissions for /dev/vsock are now set to 0o666, and /dev/vhost-vsock
5828 and /dev/vhost-net are owned by the kvm group.
5830 * The hardware database has been extended with a list of fingerprint
5831 readers that correctly support USB auto-suspend using data from
5834 * systemd-resolved can now answer DNSSEC questions through the stub
5835 resolver interface in a way that allows local clients to do DNSSEC
5836 validation themselves. For a question with DO+CD set, it'll proxy the
5837 DNS query and respond with a mostly unmodified packet received from
5838 the upstream server.
5840 * systemd-resolved learnt a new boolean option CacheFromLocalhost= in
5841 resolved.conf. If true the service will provide caching even for DNS
5842 lookups made to an upstream DNS server on the 127.0.0.1/::1
5843 addresses. By default (and when the option is false) systemd-resolved
5844 will not cache such lookups, in order to avoid duplicate local
5845 caching, under the assumption the local upstream server caches
5848 * systemd-resolved now implements RFC5001 NSID in its local DNS
5849 stub. This may be used by local clients to determine whether they are
5850 talking to the DNS resolver stub or a different DNS server.
5852 * When resolving host names and other records resolvectl will now
5853 report where the data was acquired from (i.e. the local cache, the
5854 network, locally synthesized, …) and whether the network traffic it
5855 effected was encrypted or not. Moreover the tool acquired a number of
5856 new options --cache=, --synthesize=, --network=, --zone=,
5857 --trust-anchor=, --validate= that take booleans and may be used to
5858 tweak a lookup, i.e. whether it may be answered from cached
5859 information, locally synthesized information, information acquired
5860 through the network, the local mDNS/LLMNR zone, the DNSSEC trust
5861 anchor, and whether DNSSEC validation shall be executed for the
5864 * systemd-nspawn gained a new --ambient-capability= setting
5865 (AmbientCapability= in .nspawn files) to configure ambient
5866 capabilities passed to the container payload.
5868 * systemd-nspawn gained the ability to configure the firewall using the
5869 nftables subsystem (in addition to the existing iptables
5870 support). Similarly, systemd-networkd's IPMasquerade= option now
5871 supports nftables as back-end, too. In both cases NAT on IPv6 is now
5872 supported too, in addition to IPv4 (the iptables back-end still is
5875 "IPMasquerade=yes", which was the same as "IPMasquerade=ipv4" before,
5876 retains its meaning, but has been deprecated. Please switch to either
5877 "ivp4" or "both" (if covering IPv6 is desired).
5879 * systemd-importd will now download .verity and .roothash.p7s files
5880 along with the machine image (as exposed via machinectl pull-raw).
5882 * systemd-oomd now gained a new DefaultMemoryPressureDurationSec=
5883 setting to configure the time a unit's cgroup needs to exceed memory
5884 pressure limits before action will be taken, and a new
5885 ManagedOOMPreference=none|avoid|omit setting to avoid killing certain
5888 systemd-oomd is now considered fully supported (the usual
5889 backwards-compatibility promises apply). Swap is not required for
5890 operation, but it is still recommended.
5892 * systemd-timesyncd gained a new ConnectionRetrySec= setting which
5893 configures the retry delay when trying to contact servers.
5895 * systemd-stdio-bridge gained --system/--user options to connect to the
5896 system bus (previous default) or the user session bus.
5898 * systemd-localed may now call locale-gen to generate missing locales
5899 on-demand (UTF-8-only). This improves integration with Debian-based
5900 distributions (Debian/Ubuntu/PureOS/Tanglu/...) and Arch Linux.
5902 * systemctl --check-inhibitors=true may now be used to obey inhibitors
5903 even when invoked non-interactively. The old --ignore-inhibitors
5904 switch is now deprecated and replaced by --check-inhibitors=false.
5906 * systemctl import-environment will now emit a warning when called
5907 without any arguments (i.e. to import the full environment block of
5908 the called program). This command will usually be invoked from a
5909 shell, which means that it'll inherit a bunch of variables which are
5910 specific to that shell, and usually to the TTY the shell is connected
5911 to, and don't have any meaning in the global context of the system or
5912 user service manager. Instead, only specific variables should be
5913 imported into the manager environment block.
5915 Similarly, programs which update the manager environment block by
5916 directly calling the D-Bus API of the manager, should also push
5917 specific variables, and not the full inherited environment.
5919 * systemctl's status output now shows unit state with a more careful
5920 choice of Unicode characters: units in maintenance show a "○" symbol
5921 instead of the usual "●", failed units show "×", and services being
5924 * coredumpctl gained a --debugger-arguments= switch to pass arguments
5925 to the debugger. It also gained support for showing coredump info in
5926 a simple JSON format.
5928 * systemctl/loginctl/machinectl's --signal= option now accept a special
5929 value "list", which may be used to show a brief table with known
5930 process signals and their numbers.
5932 * networkctl now shows the link activation policy in status.
5934 * Various tools gained --pager/--no-pager/--json= switches to
5935 enable/disable the pager and provide JSON output.
5937 * Various tools now accept two new values for the SYSTEMD_COLORS
5938 environment variable: "16" and "256", to configure how many terminal
5939 colors are used in output.
5941 * less 568 or newer is now required for the auto-paging logic of the
5942 various tools. Hyperlink ANSI sequences in terminal output are now
5943 used even if a pager is used, and older versions of less are not able
5944 to display these sequences correctly. SYSTEMD_URLIFY=0 may be used to
5945 disable this output again.
5947 * Builds with support for separate / and /usr/ hierarchies ("split-usr"
5948 builds, non-merged-usr builds) are now officially deprecated. A
5949 warning is emitted during build. Support is slated to be removed in
5950 about a year (when the Debian Bookworm release development starts).
5952 * Systems with the legacy cgroup v1 hierarchy are now marked as
5953 "tainted", to make it clearer that using the legacy hierarchy is not
5956 * systemd-localed will now refuse to configure a keymap which is not
5957 installed in the file system. This is intended as a bug fix, but
5958 could break cases where systemd-localed was used to configure the
5959 keymap in advanced of it being installed. It is necessary to install
5960 the keymap file first.
5962 * The main git development branch has been renamed to 'main'.
5964 * mmcblk[0-9]boot[0-9] devices will no longer be probed automatically
5965 for partitions, as in the vast majority of cases they contain none
5966 and are used internally by the bootloader (eg: uboot).
5968 * systemd will now set the $SYSTEMD_EXEC_PID environment variable for
5969 spawned processes to the PID of the process itself. This may be used
5970 by programs for detecting whether they were forked off by the service
5971 manager itself or are a process forked off further down the tree.
5973 * The sd-device API gained four new calls: sd_device_get_action() to
5974 determine the uevent add/remove/change/… action the device object has
5975 been seen for, sd_device_get_seqno() to determine the uevent sequence
5976 number, sd_device_new_from_stat_rdev() to allocate a new sd_device
5977 object from stat(2) data of a device node, and sd_device_trigger() to
5978 write to the 'uevent' attribute of a device.
5980 * For most tools the --no-legend= switch has been replaced by
5981 --legend=no and --legend=yes, to force whether tables are shown with
5984 * Units acquired a new property "Markers" that takes a list of zero,
5985 one or two of the following strings: "needs-reload" and
5986 "needs-restart". These markers may be set via "systemctl
5987 set-property". Once a marker is set, "systemctl reload-or-restart
5988 --marked" may be invoked to execute the operation the units are
5989 marked for. This is useful for package managers that want to mark
5990 units for restart/reload while updating, but effect the actual
5991 operations at a later step at once.
5993 * The sd_bus_message_read_strv() API call of sd-bus may now also be
5994 used to parse arrays of D-Bus signatures and D-Bus paths, in addition
5997 * bootctl will now report whether the UEFI firmware used a TPM2 device
5998 and measured the boot process into it.
6000 * systemd-tmpfiles learnt support for a new environment variable
6001 $SYSTEMD_TMPFILES_FORCE_SUBVOL which takes a boolean value. If true
6002 the v/q/Q lines in tmpfiles.d/ snippets will create btrfs subvolumes
6003 even if the root fs of the system is not itself a btrfs volume.
6005 * systemd-detect-virt/ConditionVirtualization= will now explicitly
6006 detect Docker/Podman environments where possible. Moreover, they
6007 should be able to generically detect any container manager as long as
6008 it assigns the container a cgroup.
6010 * portablectl gained a new "reattach" verb for detaching/reattaching a
6011 portable service image, useful for updating images on-the-fly.
6013 * Intel SGX enclave device nodes (which expose a security feature of
6014 newer Intel CPUs) will now be owned by a new system group "sgx".
6016 Contributions from: Adam Nielsen, Adrian Vovk, AJ Jordan, Alan Perry,
6017 Alastair Pharo, Alexander Batischev, Ali Abdallah, Andrew Balmos,
6018 Anita Zhang, Annika Wickert, Ansgar Burchardt, Antonio Terceiro,
6019 Antonius Frie, Ardy, Arian van Putten, Ariel Fermani, Arnaud T,
6020 A S Alam, Bastien Nocera, Benjamin Berg, Benjamin Robin, Björn Daase,
6021 caoxia, Carlo Wood, Charles Lee, ChopperRob, chri2, Christian Ehrhardt,
6022 Christian Hesse, Christopher Obbard, clayton craft, corvusnix, cprn,
6023 Daan De Meyer, Daniele Medri, Daniel Rusek, Dan Sanders, Dan Streetman,
6024 Darren Ng, David Edmundson, David Tardon, Deepak Rawat, Devon Pringle,
6025 Dmitry Borodaenko, dropsignal, Einsler Lee, Endre Szabo,
6026 Evgeny Vereshchagin, Fabian Affolter, Fangrui Song, Felipe Borges,
6027 feliperodriguesfr, Felix Stupp, Florian Hülsmann, Florian Klink,
6028 Florian Westphal, Franck Bui, Frantisek Sumsal, Gablegritule,
6029 Gaël PORTAY, Gaurav, Giedrius Statkevičius, Greg Depoire-Ferrer,
6030 Gustavo Costa, Hans de Goede, Hela Basa, heretoenhance, hide,
6031 Iago López Galeiras, igo95862, Ilya Dmitrichenko, Jameer Pathan,
6032 Jan Tojnar, Jiehong, Jinyuan Si, Joerg Behrmann, John Slade,
6033 Jonathan G. Underwood, Jonathan McDowell, Josh Triplett, Joshua Watt,
6034 Julia Cartwright, Julien Humbert, Kairui Song, Karel Zak,
6035 Kevin Backhouse, Kevin P. Fleming, Khem Raj, Konomi, krissgjeng,
6036 l4gfcm, Lajos Veres, Lennart Poettering, Lincoln Ramsay, Luca Boccassi,
6037 Luca BRUNO, Lucas Werkmeister, Luka Kudra, Luna Jernberg,
6038 Marc-André Lureau, Martin Wilck, Matthias Klumpp, Matt Turner,
6039 Michael Gisbers, Michael Marley, Michael Trapp, Michal Fabik,
6040 Michał Kopeć, Michal Koutný, Michal Sekletár, Michele Guerini Rocco,
6041 Mike Gilbert, milovlad, moson-mo, Nick, nihilix-melix, Oğuz Ersen,
6042 Ondrej Mosnacek, pali, Pavel Hrdina, Pavel Sapezhko, Perry Yuan,
6043 Peter Hutterer, Pierre Dubouilh, Piotr Drąg, Pjotr Vertaalt,
6044 Richard Laager, RussianNeuroMancer, Sam Lunt, Sebastiaan van Stijn,
6045 Sergey Bugaev, shenyangyang4, simmon, Simonas Kazlauskas,
6046 Slimane Selyan Amiri, Stefan Agner, Steve Ramage, Susant Sahani,
6047 Sven Mueller, Tad Fisher, Takashi Iwai, Thomas Haller, Tom Shield,
6048 Topi Miettinen, Torsten Hilbrich, tpgxyz, Tyler Hicks, ulf-f,
6049 Ulrich Ölmann, Vincent Pelletier, Vinnie Magro, Vito Caputo, Vlad,
6050 walbit-de, Whired Planck, wouter bolsterlee, Xℹ Ruoyao, Yangyang Shen,
6051 Yuri Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek,
6052 Zmicer Turok, Дамјан Георгиевски
6054 — Berlin, 2021-03-30
6058 * KERNEL API INCOMPATIBILITY: Linux 4.14 introduced two new uevents
6059 "bind" and "unbind" to the Linux device model. When this kernel
6060 change was made, systemd-udevd was only minimally updated to handle
6061 and propagate these new event types. The introduction of these new
6062 uevents (which are typically generated for USB devices and devices
6063 needing a firmware upload before being functional) resulted in a
6064 number of issues which we so far didn't address. We hoped the kernel
6065 maintainers would themselves address these issues in some form, but
6066 that did not happen. To handle them properly, many (if not most) udev
6067 rules files shipped in various packages need updating, and so do many
6068 programs that monitor or enumerate devices with libudev or sd-device,
6069 or otherwise process uevents. Please note that this incompatibility
6070 is not fault of systemd or udev, but caused by an incompatible kernel
6071 change that happened back in Linux 4.14, but is becoming more and
6072 more visible as the new uevents are generated by more kernel drivers.
6074 To minimize issues resulting from this kernel change (but not avoid
6075 them entirely) starting with systemd-udevd 247 the udev "tags"
6076 concept (which is a concept for marking and filtering devices during
6077 enumeration and monitoring) has been reworked: udev tags are now
6078 "sticky", meaning that once a tag is assigned to a device it will not
6079 be removed from the device again until the device itself is removed
6080 (i.e. unplugged). This makes sure that any application monitoring
6081 devices that match a specific tag is guaranteed to both see uevents
6082 where the device starts being relevant, and those where it stops
6083 being relevant (the latter now regularly happening due to the new
6084 "unbind" uevent type). The udev tags concept is hence now a concept
6085 tied to a *device* instead of a device *event* — unlike for example
6086 udev properties whose lifecycle (as before) is generally tied to a
6087 device event, meaning that the previously determined properties are
6088 forgotten whenever a new uevent is processed.
6090 With the newly redefined udev tags concept, sometimes it's necessary
6091 to determine which tags are the ones applied by the most recent
6092 uevent/database update, in order to discern them from those
6093 originating from earlier uevents/database updates of the same
6094 device. To accommodate for this a new automatic property CURRENT_TAGS
6095 has been added that works similar to the existing TAGS property but
6096 only lists tags set by the most recent uevent/database
6097 update. Similarly, the libudev/sd-device API has been updated with
6098 new functions to enumerate these 'current' tags, in addition to the
6099 existing APIs that now enumerate the 'sticky' ones.
6101 To properly handle "bind"/"unbind" on Linux 4.14 and newer it is
6102 essential that all udev rules files and applications are updated to
6103 handle the new events. Specifically:
6105 • All rule files that currently use a header guard similar to
6106 ACTION!="add|change",GOTO="xyz_end" should be updated to use
6107 ACTION=="remove",GOTO="xyz_end" instead, so that the
6108 properties/tags they add are also applied whenever "bind" (or
6109 "unbind") is seen. (This is most important for all physical device
6110 types — those for which "bind" and "unbind" are currently
6111 generated, for all other device types this change is still
6112 recommended but not as important — but certainly prepares for
6113 future kernel uevent type additions).
6115 • Similarly, all code monitoring devices that contains an 'if' branch
6116 discerning the "add" + "change" uevent actions from all other
6117 uevents actions (i.e. considering devices only relevant after "add"
6118 or "change", and irrelevant on all other events) should be reworked
6119 to instead negatively check for "remove" only (i.e. considering
6120 devices relevant after all event types, except for "remove", which
6121 invalidates the device). Note that this also means that devices
6122 should be considered relevant on "unbind", even though conceptually
6123 this — in some form — invalidates the device. Since the precise
6124 effect of "unbind" is not generically defined, devices should be
6125 considered relevant even after "unbind", however I/O errors
6126 accessing the device should then be handled gracefully.
6128 • Any code that uses device tags for deciding whether a device is
6129 relevant or not most likely needs to be updated to use the new
6130 udev_device_has_current_tag() API (or sd_device_has_current_tag()
6131 in case sd-device is used), to check whether the tag is set at the
6132 moment an uevent is seen (as opposed to the existing
6133 udev_device_has_tag() API which checks if the tag ever existed on
6134 the device, following the API concept redefinition explained
6137 We are very sorry for this breakage and the requirement to update
6138 packages using these interfaces. We'd again like to underline that
6139 this is not caused by systemd/udev changes, but result of a kernel
6142 * UPCOMING INCOMPATIBILITY: So far most downstream distribution
6143 packages have not retriggered devices once the udev package (or any
6144 auxiliary package installing additional udev rules) is updated. We
6145 intend to work with major distributions to change this, so that
6146 "udevadm trigger -c change" is issued on such upgrades, ensuring that
6147 the updated ruleset is applied to the devices already discovered, so
6148 that (asynchronously) after the upgrade completed the udev database
6149 is consistent with the updated rule set. This means udev rules must
6150 be ready to be retriggered with a "change" action any time, and
6151 result in correct and complete udev database entries. While the
6152 majority of udev rule files known to us currently get this right,
6153 some don't. Specifically, there are udev rules files included in
6154 various packages that only set udev properties on the "add" action,
6155 but do not handle the "change" action. If a device matching those
6156 rules is retriggered with the "change" action (as is intended here)
6157 it would suddenly lose the relevant properties. This always has been
6158 problematic, but as soon as all udev devices are triggered on relevant
6159 package upgrades this will become particularly so. It is strongly
6160 recommended to fix offending rules so that they can handle a "change"
6161 action at any time, and acquire all necessary udev properties even
6162 then. Or in other words: the header guard mentioned above
6163 (ACTION=="remove",GOTO="xyz_end") is the correct approach to handle
6164 this, as it makes sure rules are rerun on "change" correctly, and
6165 accumulate the correct and complete set of udev properties. udev rule
6166 definitions that cannot handle "change" events being triggered at
6167 arbitrary times should be considered buggy.
6169 * The MountAPIVFS= service file setting now defaults to on if
6170 RootImage= and RootDirectory= are used, which means that with those
6171 two settings /proc/, /sys/ and /dev/ are automatically properly set
6172 up for services. Previous behaviour may be restored by explicitly
6173 setting MountAPIVFS=off.
6175 * Since PAM 1.2.0 (2015) configuration snippets may be placed in
6176 /usr/lib/pam.d/ in addition to /etc/pam.d/. If a file exists in the
6177 latter it takes precedence over the former, similar to how most of
6178 systemd's own configuration is handled. Given that PAM stack
6179 definitions are primarily put together by OS vendors/distributions
6180 (though possibly overridden by users), this systemd release moves its
6181 own PAM stack configuration for the "systemd-user" PAM service (i.e.
6182 for the PAM session invoked by the per-user user@.service instance)
6183 from /etc/pam.d/ to /usr/lib/pam.d/. We recommend moving all
6184 packages' vendor versions of their PAM stack definitions from
6185 /etc/pam.d/ to /usr/lib/pam.d/, but if such OS-wide migration is not
6186 desired the location to which systemd installs its PAM stack
6187 configuration may be changed via the -Dpamconfdir Meson option.
6189 * The runtime dependencies on libqrencode, libpcre2, libidn/libidn2,
6190 libpwquality and libcryptsetup have been changed to be based on
6191 dlopen(): instead of regular dynamic library dependencies declared in
6192 the binary ELF headers, these libraries are now loaded on demand
6193 only, if they are available. If the libraries cannot be found the
6194 relevant operations will fail gracefully, or a suitable fallback
6195 logic is chosen. This is supposed to be useful for general purpose
6196 distributions, as it allows minimizing the list of dependencies the
6197 systemd packages pull in, permitting building of more minimal OS
6198 images, while still making use of these "weak" dependencies should
6199 they be installed. Since many package managers automatically
6200 synthesize package dependencies from ELF shared library dependencies,
6201 some additional manual packaging work has to be done now to replace
6202 those (slightly downgraded from "required" to "recommended" or
6203 whatever is conceptually suitable for the package manager). Note that
6204 this change does not alter build-time behaviour: as before the
6205 build-time dependencies have to be installed during build, even if
6206 they now are optional during runtime.
6208 * sd-event.h gained a new call sd_event_add_time_relative() for
6209 installing timers relative to the current time. This is mostly a
6210 convenience wrapper around the pre-existing sd_event_add_time() call
6211 which installs absolute timers.
6213 * sd-event event sources may now be placed in a new "exit-on-failure"
6214 mode, which may be controlled via the new
6215 sd_event_source_get_exit_on_failure() and
6216 sd_event_source_set_exit_on_failure() functions. If enabled, any
6217 failure returned by the event source handler functions will result in
6218 exiting the event loop (unlike the default behaviour of just
6219 disabling the event source but continuing with the event loop). This
6220 feature is useful to set for all event sources that define "primary"
6221 program behaviour (where failure should be fatal) in contrast to
6222 "auxiliary" behaviour (where failure should remain local).
6224 * Most event source types sd-event supports now accept a NULL handler
6225 function, in which case the event loop is exited once the event
6226 source is to be dispatched, using the userdata pointer — converted to
6227 a signed integer — as exit code of the event loop. Previously this
6228 was supported for IO and signal event sources already. Exit event
6229 sources still do not support this (simply because it makes little
6230 sense there, as the event loop is already exiting when they are
6233 * A new per-unit setting RootImageOptions= has been added which allows
6234 tweaking the mount options for any file system mounted as effect of
6235 the RootImage= setting.
6237 * Another new per-unit setting MountImages= has been added, that allows
6238 mounting additional disk images into the file system tree accessible
6241 * Timer units gained a new FixedRandomDelay= boolean setting. If
6242 enabled, the random delay configured with RandomizedDelaySec= is
6243 selected in a way that is stable on a given system (though still
6244 different for different units).
6246 * Socket units gained a new setting Timestamping= that takes "us", "ns"
6247 or "off". This controls the SO_TIMESTAMP/SO_TIMESTAMPNS socket
6250 * systemd-repart now generates JSON output when requested with the new
6253 * systemd-machined's OpenMachineShell() bus call will now pass
6254 additional policy metadata data fields to the PolicyKit
6255 authentication request.
6257 * systemd-tmpfiles gained a new -E switch, which is equivalent to
6258 --exclude-prefix=/dev --exclude-prefix=/proc --exclude=/run
6259 --exclude=/sys. It's particularly useful in combination with --root=,
6260 when operating on OS trees that do not have any of these four runtime
6261 directories mounted, as this means no files below these subtrees are
6262 created or modified, since those mount points should probably remain
6265 * systemd-tmpfiles gained a new --image= switch which is like --root=,
6266 but takes a disk image instead of a directory as argument. The
6267 specified disk image is mounted inside a temporary mount namespace
6268 and the tmpfiles.d/ drop-ins stored in the image are executed and
6269 applied to the image. systemd-sysusers similarly gained a new
6270 --image= switch, that allows the sysusers.d/ drop-ins stored in the
6271 image to be applied onto the image.
6273 * Similarly, the journalctl command also gained an --image= switch,
6274 which is a quick one-step solution to look at the log data included
6277 * journalctl's --output=cat option (which outputs the log content
6278 without any metadata, just the pure text messages) will now make use
6279 of terminal colors when run on a suitable terminal, similarly to the
6282 * JSON group records now support a "description" string that may be
6283 used to add a human-readable textual description to such groups. This
6284 is supposed to match the user's GECOS field which traditionally
6285 didn't have a counterpart for group records.
6287 * The "systemd-dissect" tool that may be used to inspect OS disk images
6288 and that was previously installed to /usr/lib/systemd/ has now been
6289 moved to /usr/bin/, reflecting its updated status of an officially
6290 supported tool with a stable interface. It gained support for a new
6291 --mkdir switch which when combined with --mount has the effect of
6292 creating the directory to mount the image to if it is missing
6293 first. It also gained two new commands --copy-from and --copy-to for
6294 copying files and directories in and out of an OS image without the
6295 need to manually mount it. It also acquired support for a new option
6296 --json= to generate JSON output when inspecting an OS image.
6298 * The cgroup2 file system is now mounted with the
6299 "memory_recursiveprot" mount option, supported since kernel 5.7. This
6300 means that the MemoryLow= and MemoryMin= unit file settings now apply
6301 recursively to whole subtrees.
6303 * systemd-homed now defaults to using the btrfs file system — if
6304 available — when creating home directories in LUKS volumes. This may
6305 be changed with the DefaultFileSystemType= setting in homed.conf.
6306 It's now the default file system in various major distributions and
6307 has the major benefit for homed that it can be grown and shrunk while
6308 mounted, unlike the other contenders ext4 and xfs, which can both be
6309 grown online, but not shrunk (in fact xfs is the technically most
6310 limited option here, as it cannot be shrunk at all).
6312 * JSON user records managed by systemd-homed gained support for
6313 "recovery keys". These are basically secondary passphrases that can
6314 unlock user accounts/home directories. They are computer-generated
6315 rather than user-chosen, and typically have greater entropy.
6316 homectl's --recovery-key= option may be used to add a recovery key to
6317 a user account. The generated recovery key is displayed as a QR code,
6318 so that it can be scanned to be kept in a safe place. This feature is
6319 particularly useful in combination with systemd-homed's support for
6320 FIDO2 or PKCS#11 authentication, as a secure fallback in case the
6321 security tokens are lost. Recovery keys may be entered wherever the
6322 system asks for a password.
6324 * systemd-homed now maintains a "dirty" flag for each LUKS encrypted
6325 home directory which indicates that a home directory has not been
6326 deactivated cleanly when offline. This flag is useful to identify
6327 home directories for which the offline discard logic did not run when
6328 offlining, and where it would be a good idea to log in again to catch
6331 * systemctl gained a new parameter --timestamp= which may be used to
6332 change the style in which timestamps are output, i.e. whether to show
6333 them in local timezone or UTC, or whether to show µs granularity.
6335 * Alibaba's "pouch" container manager is now detected by
6336 systemd-detect-virt, ConditionVirtualization= and similar
6337 constructs. Similar, they now also recognize IBM PowerVM machine
6340 * systemd-nspawn has been reworked to use the /run/host/incoming/ as
6341 place to use for propagating external mounts into the
6342 container. Similarly /run/host/notify is now used as the socket path
6343 for container payloads to communicate with the container manager
6344 using sd_notify(). The container manager now uses the
6345 /run/host/inaccessible/ directory to place "inaccessible" file nodes
6346 of all relevant types which may be used by the container payload as
6347 bind mount source to over-mount inodes to make them inaccessible.
6348 /run/host/container-manager will now be initialized with the same
6349 string as the $container environment variable passed to the
6350 container's PID 1. /run/host/container-uuid will be initialized with
6351 the same string as $container_uuid. This means the /run/host/
6352 hierarchy is now the primary way to make host resources available to
6353 the container. The Container Interface documents these new files and
6356 https://systemd.io/CONTAINER_INTERFACE
6358 * Support for the "ConditionNull=" unit file condition has been
6359 deprecated and undocumented for 6 years. systemd started to warn
6360 about its use 1.5 years ago. It has now been removed entirely.
6362 * sd-bus.h gained a new API call sd_bus_error_has_names(), which takes
6363 a sd_bus_error struct and a list of error names, and checks if the
6364 error matches one of these names. It's a convenience wrapper that is
6365 useful in cases where multiple errors shall be handled the same way.
6367 * A new system call filter list "@known" has been added, that contains
6368 all system calls known at the time systemd was built.
6370 * Behaviour of system call filter allow lists has changed slightly:
6371 system calls that are contained in @known will result in EPERM by
6372 default, while those not contained in it result in ENOSYS. This
6373 should improve compatibility because known system calls will thus be
6374 communicated as prohibited, while unknown (and thus newer ones) will
6375 be communicated as not implemented, which hopefully has the greatest
6376 chance of triggering the right fallback code paths in client
6379 * "systemd-analyze syscall-filter" will now show two separate sections
6380 at the bottom of the output: system calls known during systemd build
6381 time but not included in any of the filter groups shown above, and
6382 system calls defined on the local kernel but known during systemd
6385 * If the $SYSTEMD_LOG_SECCOMP=1 environment variable is set for
6386 systemd-nspawn all system call filter violations will be logged by
6387 the kernel (audit). This is useful for tracking down system calls
6388 invoked by container payloads that are prohibited by the container's
6389 system call filter policy.
6391 * If the $SYSTEMD_SECCOMP=0 environment variable is set for
6392 systemd-nspawn (and other programs that use seccomp) all seccomp
6393 filtering is turned off.
6395 * Two new unit file settings ProtectProc= and ProcSubset= have been
6396 added that expose the hidepid= and subset= mount options of procfs.
6397 All processes of the unit will only see processes in /proc that are
6398 are owned by the unit's user. This is an important new sandboxing
6399 option that is recommended to be set on all system services. All
6400 long-running system services that are included in systemd itself set
6401 this option now. This option is only supported on kernel 5.8 and
6402 above, since the hidepid= option supported on older kernels was not a
6403 per-mount option but actually applied to the whole PID namespace.
6405 * Socket units gained a new boolean setting FlushPending=. If enabled
6406 all pending socket data/connections are flushed whenever the socket
6407 unit enters the "listening" state, i.e. after the associated service
6410 * The unit file setting NUMAMask= gained a new "all" value: when used,
6411 all existing NUMA nodes are added to the NUMA mask.
6413 * A new "credentials" logic has been added to system services. This is
6414 a simple mechanism to pass privileged data to services in a safe and
6415 secure way. It's supposed to be used to pass per-service secret data
6416 such as passwords or cryptographic keys but also associated less
6417 private information such as user names, certificates, and similar to
6418 system services. Each credential is identified by a short user-chosen
6419 name and may contain arbitrary binary data. Two new unit file
6420 settings have been added: SetCredential= and LoadCredential=. The
6421 former allows setting a credential to a literal string, the latter
6422 sets a credential to the contents of a file (or data read from a
6423 user-chosen AF_UNIX stream socket). Credentials are passed to the
6424 service via a special credentials directory, one file for each
6425 credential. The path to the credentials directory is passed in a new
6426 $CREDENTIALS_DIRECTORY environment variable. Since the credentials
6427 are passed in the file system they may be easily referenced in
6428 ExecStart= command lines too, thus no explicit support for the
6429 credentials logic in daemons is required (though ideally daemons
6430 would look for the bits they need in $CREDENTIALS_DIRECTORY
6431 themselves automatically, if set). The $CREDENTIALS_DIRECTORY is
6432 backed by unswappable memory if privileges allow it, immutable if
6433 privileges allow it, is accessible only to the service's UID, and is
6434 automatically destroyed when the service stops.
6436 * systemd-nspawn supports the same credentials logic. It can both
6437 consume credentials passed to it via the aforementioned
6438 $CREDENTIALS_DIRECTORY protocol as well as pass these credentials on
6439 to its payload. The service manager/PID 1 has been updated to match
6440 this: it can also accept credentials from the container manager that
6441 invokes it (in fact: any process that invokes it), and passes them on
6442 to its services. Thus, credentials can be propagated recursively down
6443 the tree: from a system's service manager to a systemd-nspawn
6444 service, to the service manager that runs as container payload and to
6445 the service it runs below. Credentials may also be added on the
6446 systemd-nspawn command line, using new --set-credential= and
6447 --load-credential= command line switches that match the
6448 aforementioned service settings.
6450 * systemd-repart gained new settings Format=, Encrypt=, CopyFiles= in
6451 the partition drop-ins which may be used to format/LUKS
6452 encrypt/populate any created partitions. The partitions are
6453 encrypted/formatted/populated before they are registered in the
6454 partition table, so that they appear atomically: either the
6455 partitions do not exist yet or they exist fully encrypted, formatted,
6456 and populated — there is no time window where they are
6457 "half-initialized". Thus the system is robust to abrupt shutdown: if
6458 the tool is terminated half-way during its operations on next boot it
6459 will start from the beginning.
6461 * systemd-repart's --size= operation gained a new "auto" value. If
6462 specified, and operating on a loopback file it is automatically sized
6463 to the minimal size the size constraints permit. This is useful to
6464 use "systemd-repart" as an image builder for minimally sized images.
6466 * systemd-resolved now gained a third IPC interface for requesting name
6467 resolution: besides D-Bus and local DNS to 127.0.0.53 a Varlink
6468 interface is now supported. The nss-resolve NSS module has been
6469 modified to use this new interface instead of D-Bus. Using Varlink
6470 has a major benefit over D-Bus: it works without a broker service,
6471 and thus already during earliest boot, before the dbus daemon has
6472 been started. This means name resolution via systemd-resolved now
6473 works at the same time systemd-networkd operates: from earliest boot
6474 on, including in the initrd.
6476 * systemd-resolved gained support for a new DNSStubListenerExtra=
6477 configuration file setting which may be used to specify additional IP
6478 addresses the built-in DNS stub shall listen on, in addition to the
6479 main one on 127.0.0.53:53.
6481 * Name lookups issued via systemd-resolved's D-Bus and Varlink
6482 interfaces (and thus also via glibc NSS if nss-resolve is used) will
6483 now honour a trailing dot in the hostname: if specified the search
6484 path logic is turned off. Thus "resolvectl query foo." is now
6485 equivalent to "resolvectl query --search=off foo.".
6487 * systemd-resolved gained a new D-Bus property "ResolvConfMode" that
6488 exposes how /etc/resolv.conf is currently managed: by resolved (and
6489 in which mode if so) or another subsystem. "resolvctl" will display
6490 this property in its status output.
6492 * The resolv.conf snippets systemd-resolved provides will now set "."
6493 as the search domain if no other search domain is known. This turns
6494 off the derivation of an implicit search domain by nss-dns for the
6495 hostname, when the hostname is set to an FQDN. This change is done to
6496 make nss-dns using resolv.conf provided by systemd-resolved behave
6497 more similarly to nss-resolve.
6499 * systemd-tmpfiles' file "aging" logic (i.e. the automatic clean-up of
6500 /tmp/ and /var/tmp/ based on file timestamps) now looks at the
6501 "birth" time (btime) of a file in addition to the atime, mtime, and
6504 * systemd-analyze gained a new verb "capability" that lists all known
6505 capabilities by the systemd build and by the kernel.
6507 * If a file /usr/lib/clock-epoch exists, PID 1 will read its mtime and
6508 advance the system clock to it at boot if it is noticed to be before
6509 that time. Previously, PID 1 would only advance the time to an epoch
6510 time that is set during build-time. With this new file OS builders
6511 can change this epoch timestamp on individual OS images without
6512 having to rebuild systemd.
6514 * systemd-logind will now listen to the KEY_RESTART key from the Linux
6515 input layer and reboot the system if it is pressed, similarly to how
6516 it already handles KEY_POWER, KEY_SUSPEND or KEY_SLEEP. KEY_RESTART
6517 was originally defined in the Multimedia context (to restart playback
6518 of a song or film), but is now primarily used in various embedded
6519 devices for "Reboot" buttons. Accordingly, systemd-logind will now
6520 honour it as such. This may configured in more detail via the new
6521 HandleRebootKey= and RebootKeyIgnoreInhibited=.
6523 * systemd-nspawn/systemd-machined will now reconstruct hardlinks when
6524 copying OS trees, for example in "systemd-nspawn --ephemeral",
6525 "systemd-nspawn --template=", "machinectl clone" and similar. This is
6526 useful when operating with OSTree images, which use hardlinks heavily
6527 throughout, and where such copies previously resulting in "exploding"
6530 * systemd-nspawn's --console= setting gained support for a new
6531 "autopipe" value, which is identical to "interactive" when invoked on
6532 a TTY, and "pipe" otherwise.
6534 * systemd-networkd's .network files gained support for explicitly
6535 configuring the multicast membership entries of bridge devices in the
6536 [BridgeMDB] section. It also gained support for the PIE queuing
6537 discipline in the [FlowQueuePIE] sections.
6539 * systemd-networkd's .netdev files may now be used to create "BareUDP"
6540 tunnels, configured in the new [BareUDP] setting.
6542 * systemd-networkd's Gateway= setting in .network files now accepts the
6543 special values "_dhcp4" and "_ipv6ra" to configure additional,
6544 locally defined, explicit routes to the gateway acquired via DHCP or
6545 IPv6 Router Advertisements. The old setting "_dhcp" is deprecated,
6546 but still accepted for backwards compatibility.
6548 * systemd-networkd's [IPv6PrefixDelegation] section and
6549 IPv6PrefixDelegation= options have been renamed as [IPv6SendRA] and
6550 IPv6SendRA= (the old names are still accepted for backwards
6553 * systemd-networkd's .network files gained the DHCPv6PrefixDelegation=
6554 boolean setting in [Network] section. If enabled, the delegated prefix
6555 gained by another link will be configured, and an address within the
6556 prefix will be assigned.
6558 * systemd-networkd's .network files gained the Announce= boolean setting
6559 in [DHCPv6PrefixDelegation] section. When enabled, the delegated
6560 prefix will be announced through IPv6 router advertisement (IPv6 RA).
6561 The setting is enabled by default.
6563 * VXLAN tunnels may now be marked as independent of any underlying
6564 network interface via the new Independent= boolean setting.
6566 * systemctl gained support for two new verbs: "service-log-level" and
6567 "service-log-target" may be used on services that implement the
6568 generic org.freedesktop.LogControl1 D-Bus interface to dynamically
6569 adjust the log level and target. All of systemd's long-running
6570 services support this now, but ideally all system services would
6571 implement this interface to make the system more uniformly
6574 * The SystemCallErrorNumber= unit file setting now accepts the new
6575 "kill" and "log" actions, in addition to arbitrary error number
6576 specifications as before. If "kill" the processes are killed on the
6577 event, if "log" the offending system call is audit logged.
6579 * A new SystemCallLog= unit file setting has been added that accepts a
6580 list of system calls that shall be logged about (audit).
6582 * The OS image dissection logic (as used by RootImage= in unit files or
6583 systemd-nspawn's --image= switch) has gained support for identifying
6584 and mounting explicit /usr/ partitions, which are now defined in the
6585 discoverable partition specification. This should be useful for
6586 environments where the root file system is
6587 generated/formatted/populated dynamically on first boot and combined
6588 with an immutable /usr/ tree that is supplied by the vendor.
6590 * In the final phase of shutdown, within the systemd-shutdown binary
6591 we'll now try to detach MD devices (i.e software RAID) in addition to
6592 loopback block devices and DM devices as before. This is supposed to
6593 be a safety net only, in order to increase robustness if things go
6594 wrong. Storage subsystems are expected to properly detach their
6595 storage volumes during regular shutdown already (or in case of
6596 storage backing the root file system: in the initrd hook we return to
6599 * If the SYSTEMD_LOG_TID environment variable is set all systemd tools
6600 will now log the thread ID in their log output. This is useful when
6601 working with heavily threaded programs.
6603 * If the SYSTEMD_RDRAND environment variable is set to "0", systemd will
6604 not use the RDRAND CPU instruction. This is useful in environments
6605 such as replay debuggers where non-deterministic behaviour is not
6608 * The autopaging logic in systemd's various tools (such as systemctl)
6609 has been updated to turn on "secure" mode in "less"
6610 (i.e. $LESSECURE=1) if execution in a "sudo" environment is
6611 detected. This disables invoking external programs from the pager,
6612 via the pipe logic. This behaviour may be overridden via the new
6613 $SYSTEMD_PAGERSECURE environment variable.
6615 * Units which have resource limits (.service, .mount, .swap, .slice,
6616 .socket, and .slice) gained new configuration settings
6617 ManagedOOMSwap=, ManagedOOMMemoryPressure=, and
6618 ManagedOOMMemoryPressureLimitPercent= that specify resource pressure
6619 limits and optional action taken by systemd-oomd.
6621 * A new service systemd-oomd has been added. It monitors resource
6622 contention for selected parts of the unit hierarchy using the PSI
6623 information reported by the kernel, and kills processes when memory
6624 or swap pressure is above configured limits. This service is only
6625 enabled by default in developer mode (see below) and should be
6626 considered a preview in this release. Behaviour details and option
6627 names are subject to change without the usual backwards-compatibility
6630 * A new helper oomctl has been added to introspect systemd-oomd state.
6631 It is only enabled by default in developer mode and should be
6632 considered a preview without the usual backwards-compatibility
6635 * New meson option -Dcompat-mutable-uid-boundaries= has been added. If
6636 enabled, systemd reads the system UID boundaries from /etc/login.defs
6637 at runtime, instead of using the built-in values selected during
6638 build. This is an option to improve compatibility for upgrades from
6639 old systems. It's strongly recommended not to make use of this
6640 functionality on new systems (or even enable it during build), as it
6641 makes something runtime-configurable that is mostly an implementation
6642 detail of the OS, and permits avoidable differences in deployments
6643 that create all kinds of problems in the long run.
6645 * New meson option '-Dmode=developer|release' has been added. When
6646 'developer', additional checks and features are enabled that are
6647 relevant during upstream development, e.g. verification that
6648 semi-automatically-generated documentation has been properly updated
6649 following API changes. Those checks are considered hints for
6650 developers and are not actionable in downstream builds. In addition,
6651 extra features that are not ready for general consumption may be
6652 enabled in developer mode. It is thus recommended to set
6653 '-Dmode=release' in end-user and distro builds.
6655 * systemd-cryptsetup gained support for processing detached LUKS
6656 headers specified on the kernel command line via the header=
6657 parameter of the luks.options= kernel command line option. The same
6658 device/path syntax as for key files is supported for header files
6661 * The "net_id" built-in of udev has been updated to ignore ACPI _SUN
6662 slot index data for devices that are connected through a PCI bridge
6663 where the _SUN index is associated with the bridge instead of the
6664 network device itself. Previously this would create ambiguous device
6665 naming if multiple network interfaces were connected to the same PCI
6666 bridge. Since this is a naming scheme incompatibility on systems that
6667 possess hardware like this it has been introduced as new naming
6668 scheme "v247". The previous scheme can be selected via the
6669 "net.naming_scheme=v245" kernel command line parameter.
6671 * ConditionFirstBoot= semantics have been modified to be safe towards
6672 abnormal system power-off during first boot. Specifically, the
6673 "systemd-machine-id-commit.service" service now acts as boot
6674 milestone indicating when the first boot process is sufficiently
6675 complete in order to not consider the next following boot also a
6676 first boot. If the system is reset before this unit is reached the
6677 first time, the next boot will still be considered a first boot; once
6678 it has been reached, no further boots will be considered a first
6679 boot. The "first-boot-complete.target" unit now acts as official hook
6680 point to order against this. If a service shall be run on every boot
6681 until the first boot fully succeeds it may thus be ordered before
6682 this target unit (and pull it in) and carry ConditionFirstBoot=
6685 * bootctl's set-default and set-oneshot commands now accept the three
6686 special strings "@default", "@oneshot", "@current" in place of a boot
6687 entry id. These strings are resolved to the current default and
6688 oneshot boot loader entry, as well as the currently booted one. Thus
6689 a command "bootctl set-default @current" may be used to make the
6690 currently boot menu item the new default for all subsequent boots.
6692 * "systemctl edit" has been updated to show the original effective unit
6693 contents in commented form in the text editor.
6695 * Units in user mode are now segregated into three new slices:
6696 session.slice (units that form the core of graphical session),
6697 app.slice ("normal" user applications), and background.slice
6698 (low-priority tasks). Unless otherwise configured, user units are
6699 placed in app.slice. The plan is to add resource limits and
6700 protections for the different slices in the future.
6702 * New GPT partition types for RISCV32/64 for the root and /usr
6703 partitions, and their associated Verity partitions have been defined,
6704 and are now understood by systemd-gpt-auto-generator, and the OS
6705 image dissection logic.
6707 Contributions from: Adolfo Jayme Barrientos, afg, Alec Moskvin, Alyssa
6708 Ross, Amitanand Chikorde, Andrew Hangsleben, Anita Zhang, Ansgar
6709 Burchardt, Arian van Putten, Aurelien Jarno, Axel Rasmussen, bauen1,
6710 Beniamino Galvani, Benjamin Berg, Bjørn Mork, brainrom, Chandradeep
6711 Dey, Charles Lee, Chris Down, Christian Göttsche, Christof Efkemann,
6712 Christoph Ruegge, Clemens Gruber, Daan De Meyer, Daniele Medri, Daniel
6713 Mack, Daniel Rusek, Dan Streetman, David Tardon, Dimitri John Ledkov,
6714 Dmitry Borodaenko, Elias Probst, Elisei Roca, ErrantSpore, Etienne
6715 Doms, Fabrice Fontaine, fangxiuning, Felix Riemann, Florian Klink,
6716 Franck Bui, Frantisek Sumsal, fwSmit, George Rawlinson, germanztz,
6717 Gibeom Gwon, Glen Whitney, Gogo Gogsi, Göran Uddeborg, Grant Mathews,
6718 Hans de Goede, Hans Ulrich Niedermann, Haochen Tong, Harald Seiler,
6719 huangyong, Hubert Kario, igo95862, Ikey Doherty, Insun Pyo, Jan Chren,
6720 Jan Schlüter, Jérémy Nouhaud, Jian-Hong Pan, Joerg Behrmann, Jonathan
6721 Lebon, Jörg Thalheim, Josh Brobst, Juergen Hoetzel, Julien Humbert,
6722 Kai-Chuan Hsieh, Kairui Song, Kamil Dudka, Kir Kolyshkin, Kristijan
6723 Gjoshev, Kyle Huey, Kyle Russell, Lee Whalen, Lennart Poettering,
6724 lichangze, Luca Boccassi, Lucas Werkmeister, Luca Weiss, Marc
6725 Kleine-Budde, Marco Wang, Martin Wilck, Marti Raudsepp, masmullin2000,
6726 Máté Pozsgay, Matt Fenwick, Michael Biebl, Michael Scherer, Michal
6727 Koutný, Michal Sekletár, Michal Suchanek, Mikael Szreder, Milo
6728 Casagrande, mirabilos, Mitsuha_QuQ, mog422, Muhammet Kara, Nazar
6729 Vinnichuk, Nicholas Narsing, Nicolas Fella, Njibhu, nl6720, Oğuz Ersen,
6730 Olivier Le Moal, Ondrej Kozina, onlybugreports, Pass Automated Testing
6731 Suite, Pat Coulthard, Pavel Sapezhko, Pedro Ruiz, perry_yuan, Peter
6732 Hutterer, Phaedrus Leeds, PhoenixDiscord, Piotr Drąg, Plan C,
6733 Purushottam choudhary, Rasmus Villemoes, Renaud Métrich, Robert Marko,
6734 Roman Beranek, Ronan Pigott, Roy Chen (陳彥廷), RussianNeuroMancer,
6735 Samanta Navarro, Samuel BF, scootergrisen, Sorin Ionescu, Steve Dodd,
6736 Susant Sahani, Timo Rothenpieler, Tobias Hunger, Tobias Kaufmann, Topi
6737 Miettinen, vanou, Vito Caputo, Weblate, Wen Yang, Whired Planck,
6738 williamvds, Yu, Li-Yu, Yuri Chornoivan, Yu Watanabe, Zbigniew
6739 Jędrzejewski-Szmek, Zmicer Turok, Дамјан Георгиевски
6741 – Warsaw, 2020-11-26
6745 * The service manager gained basic support for cgroup v2 freezer. Units
6746 can now be suspended or resumed either using new systemctl verbs,
6747 freeze and thaw respectively, or via D-Bus.
6749 * PID 1 may now automatically load pre-compiled AppArmor policies from
6750 /etc/apparmor/earlypolicy during early boot.
6752 * The CPUAffinity= setting in service unit files now supports a new
6753 special value "numa" that causes the CPU affinity masked to be set
6754 based on the NUMA mask.
6756 * systemd will now log about all left-over processes remaining in a
6757 unit when the unit is stopped. It will now warn about services using
6758 KillMode=none, as this is generally an unsafe thing to make use of.
6760 * Two new unit file settings
6761 ConditionPathIsEncrypted=/AssertPathIsEncrypted= have been
6762 added. They may be used to check whether a specific file system path
6763 resides on a block device that is encrypted on the block level
6764 (i.e. using dm-crypt/LUKS).
6766 * Another pair of new settings ConditionEnvironment=/AssertEnvironment=
6767 has been added that may be used for simple environment checks. This
6768 is particularly useful when passing in environment variables from a
6769 container manager (or from PAM in case of the systemd --user
6772 * .service unit files now accept a new setting CoredumpFilter= which
6773 allows configuration of the memory sections coredumps of the
6774 service's processes shall include.
6776 * .mount units gained a new ReadWriteOnly= boolean option. If set
6777 it will not be attempted to mount a file system read-only if mounting
6778 in read-write mode doesn't succeed. An option x-systemd.rw-only is
6779 available in /etc/fstab to control the same.
6781 * .socket units gained a new boolean setting PassPacketInfo=. If
6782 enabled, the kernel will attach additional per-packet metadata to all
6783 packets read from the socket, as an ancillary message. This controls
6784 the IP_PKTINFO, IPV6_RECVPKTINFO, NETLINK_PKTINFO socket options,
6785 depending on socket type.
6787 * .service units gained a new setting RootHash= which may be used to
6788 specify the root hash for verity enabled disk images which are
6789 specified in RootImage=. RootVerity= may be used to specify a path to
6790 the Verity data matching a RootImage= file system. (The latter is
6791 only useful for images that do not contain the Verity data embedded
6792 into the same image that carries a GPT partition table following the
6793 Discoverable Partition Specification). Similarly, systemd-nspawn
6794 gained a new switch --verity-data= that takes a path to a file with
6795 the verity data of the disk image supplied in --image=, if the image
6796 doesn't contain the verity data itself.
6798 * .service units gained a new setting RootHashSignature= which takes
6799 either a base64 encoded PKCS#7 signature of the root hash specified
6800 with RootHash=, or a path to a file to read the signature from. This
6801 allows validation of the root hash against public keys available in
6802 the kernel keyring, and is only supported on recent kernels
6803 (>= 5.4)/libcryptsetup (>= 2.30). A similar switch has been added to
6804 systemd-nspawn and systemd-dissect (--root-hash-sig=). Support for
6805 this mechanism has also been added to systemd-veritysetup.
6807 * .service unit files gained two new options
6808 TimeoutStartFailureMode=/TimeoutStopFailureMode= that may be used to
6809 tune behaviour if a start or stop timeout is hit, i.e. whether to
6810 terminate the service with SIGTERM, SIGABRT or SIGKILL.
6812 * Most options in systemd that accept hexadecimal values prefixed with
6813 0x in additional to the usual decimal notation now also support octal
6814 notation when the 0o prefix is used and binary notation if the 0b
6817 * Various command line parameters and configuration file settings that
6818 configure key or certificate files now optionally take paths to
6819 AF_UNIX sockets in the file system. If configured that way a stream
6820 connection is made to the socket and the required data read from
6821 it. This is a simple and natural extension to the existing regular
6822 file logic, and permits other software to provide keys or
6823 certificates via simple IPC services, for example when unencrypted
6824 storage on disk is not desired. Specifically, systemd-networkd's
6825 Wireguard and MACSEC key file settings as well as
6826 systemd-journal-gatewayd's and systemd-journal-remote's PEM
6827 key/certificate parameters support this now.
6829 * Unit files, tmpfiles.d/ snippets, sysusers.d/ snippets and other
6830 configuration files that support specifier expansion learnt six new
6831 specifiers: %a resolves to the current architecture, %o/%w/%B/%W
6832 resolve to the various ID fields from /etc/os-release, %l resolves to
6833 the "short" hostname of the system, i.e. the hostname configured in
6834 the kernel truncated at the first dot.
6836 * Support for the .include syntax in unit files has been removed. The
6837 concept has been obsolete for 6 years and we started warning about
6838 its pending removal 2 years ago (also see NEWS file below). It's
6841 * StandardError= and StandardOutput= in unit files no longer support
6842 the "syslog" and "syslog-console" switches. They were long removed
6843 from the documentation, but will now result in warnings when used,
6844 and be converted to "journal" and "journal+console" automatically.
6846 * If the service setting User= is set to the "nobody" user, a warning
6847 message is now written to the logs (but the value is nonetheless
6848 accepted). Setting User=nobody is unsafe, since the primary purpose
6849 of the "nobody" user is to own all files whose owner cannot be mapped
6850 locally. It's in particular used by the NFS subsystem and in user
6851 namespacing. By running a service under this user's UID it might get
6852 read and even write access to all these otherwise unmappable files,
6853 which is quite likely a major security problem.
6855 * tmpfs mounts automatically created by systemd (/tmp, /run, /dev/shm,
6856 and others) now have a size and inode limits applied (50% of RAM for
6857 /tmp and /dev/shm, 10% of RAM for other mounts, etc.). Please note
6858 that the implicit kernel default is 50% too, so there is no change
6859 in the size limit for /tmp and /dev/shm.
6861 * nss-mymachines lost support for resolution of users and groups, and
6862 now only does resolution of hostnames. This functionality is now
6863 provided by nss-systemd. Thus, the 'mymachines' entry should be
6864 removed from the 'passwd:' and 'group:' lines in /etc/nsswitch.conf
6865 (and 'systemd' added if it is not already there).
6867 * A new kernel command line option systemd.hostname= has been added
6868 that allows controlling the hostname that is initialized early during
6871 * A kernel command line option "udev.blockdev_read_only" has been
6872 added. If specified all hardware block devices that show up are
6873 immediately marked as read-only by udev. This option is useful for
6874 making sure that a specific boot under no circumstances modifies data
6875 on disk. Use "blockdev --setrw" to undo the effect of this, per
6878 * A new boolean kernel command line option systemd.swap= has been
6879 added, which may be used to turn off automatic activation of swap
6880 devices listed in /etc/fstab.
6882 * New kernel command line options systemd.condition_needs_update= and
6883 systemd.condition_first_boot= have been added, which override the
6884 result of the ConditionNeedsUpdate= and ConditionFirstBoot=
6887 * A new kernel command line option systemd.clock_usec= has been added
6888 that allows setting the system clock to the specified time in µs
6889 since Jan 1st, 1970 early during boot. This is in particular useful
6890 in order to make test cases more reliable.
6892 * The fs.suid_dumpable sysctl is set to 2 / "suidsafe". This allows
6893 systemd-coredump to save core files for suid processes. When saving
6894 the core file, systemd-coredump will use the effective uid and gid of
6895 the process that faulted.
6897 * The /sys/module/kernel/parameters/crash_kexec_post_notifiers file is
6898 now automatically set to "Y" at boot, in order to enable pstore
6899 generation for collection with systemd-pstore.
6901 * We provide a set of udev rules to enable auto-suspend on PCI and USB
6902 devices that were tested to correctly support it. Previously, this
6903 was distributed as a set of udev rules, but has now been replaced by
6904 by a set of hwdb entries (and a much shorter udev rule to take action
6905 if the device modalias matches one of the new hwdb entries).
6907 As before, entries are periodically imported from the database
6908 maintained by the ChromiumOS project. If you have a device that
6909 supports auto-suspend correctly and where it should be enabled by
6910 default, please submit a patch that adds it to the database (see
6911 /usr/lib/udev/hwdb.d/60-autosuspend.hwdb).
6913 * systemd-udevd gained the new configuration option timeout_signal= as well
6914 as a corresponding kernel command line option udev.timeout_signal=.
6915 The option can be used to configure the UNIX signal that the main
6916 daemon sends to the worker processes on timeout. Setting the signal
6917 to SIGABRT is useful for debugging.
6919 * .link files managed by systemd-udevd gained options RxFlowControl=,
6920 TxFlowControl=, AutoNegotiationFlowControl= in the [Link] section, in
6921 order to configure various flow control parameters. They also gained
6922 RxMiniBufferSize= and RxJumboBufferSize= in order to configure jumbo
6923 frame ring buffer sizes.
6925 * networkd.conf gained a new boolean setting ManageForeignRoutes=. If
6926 enabled systemd-networkd manages all routes configured by other tools.
6928 * .network files managed by systemd-networkd gained a new section
6929 [SR-IOV], in order to configure SR-IOV capable network devices.
6931 * systemd-networkd's [IPv6Prefix] section in .network files gained a
6932 new boolean setting Assign=. If enabled an address from the prefix is
6933 automatically assigned to the interface.
6935 * systemd-networkd gained a new section [DHCPv6PrefixDelegation] which
6936 controls delegated prefixes assigned by DHCPv6 client. The section
6937 has three settings: SubnetID=, Assign=, and Token=. The setting
6938 SubnetID= allows explicit configuration of the preferred subnet that
6939 systemd-networkd's Prefix Delegation logic assigns to interfaces. If
6940 Assign= is enabled (which is the default) an address from any acquired
6941 delegated prefix is automatically chosen and assigned to the
6942 interface. The setting Token= specifies an optional address generation
6945 * systemd-networkd's [Network] section gained a new setting
6946 IPv4AcceptLocal=. If enabled the interface accepts packets with local
6949 * systemd-networkd gained support for configuring the HTB queuing
6950 discipline in the [HierarchyTokenBucket] and
6951 [HierarchyTokenBucketClass] sections. Similar the "pfifo" qdisc may
6952 be configured in the [PFIFO] section, "GRED" in
6953 [GenericRandomEarlyDetection], "SFB" in [StochasticFairBlue], "cake"
6954 in [CAKE], "PIE" in [PIE], "DRR" in [DeficitRoundRobinScheduler] and
6955 [DeficitRoundRobinSchedulerClass], "BFIFO" in [BFIFO],
6956 "PFIFOHeadDrop" in [PFIFOHeadDrop], "PFIFOFast" in [PFIFOFast], "HHF"
6957 in [HeavyHitterFilter], "ETS" in [EnhancedTransmissionSelection] and
6958 "QFQ" in [QuickFairQueueing] and [QuickFairQueueingClass].
6960 * systemd-networkd gained support for a new Termination= setting in the
6961 [CAN] section for configuring the termination resistor. It also
6962 gained a new ListenOnly= setting for controlling whether to only
6963 listen on CAN interfaces, without interfering with traffic otherwise
6964 (which is useful for debugging/monitoring CAN network
6965 traffic). DataBitRate=, DataSamplePoint=, FDMode=, FDNonISO= have
6966 been added to configure various CAN-FD aspects.
6968 * systemd-networkd's [DHCPv6] section gained a new option WithoutRA=.
6969 When enabled, DHCPv6 will be attempted right-away without requiring an
6970 Router Advertisement packet suggesting it first (i.e. without the 'M'
6971 or 'O' flags set). The [IPv6AcceptRA] section gained a boolean option
6972 DHCPv6Client= that may be used to turn off the DHCPv6 client even if
6973 the RA packets suggest it.
6975 * systemd-networkd's [DHCPv4] section gained a new setting UseGateway=
6976 which may be used to turn off use of the gateway information provided
6977 by the DHCP lease. A new FallbackLeaseLifetimeSec= setting may be
6978 used to configure how to process leases that lack a lifetime option.
6980 * systemd-networkd's [DHCPv4] and [DHCPServer] sections gained a new
6981 setting SendVendorOption= allowing configuration of additional vendor
6982 options to send in the DHCP requests/responses. The [DHCPv6] section
6983 gained a new SendOption= setting for sending arbitrary DHCP
6984 options. RequestOptions= has been added to request arbitrary options
6985 from the server. UserClass= has been added to set the DHCP user class
6988 * systemd-networkd's [DHCPServer] section gained a new set of options
6989 EmitPOP3=/POP3=, EmitSMTP=/SMTP=, EmitLPR=/LPR= for including server
6990 information about these three protocols in the DHCP lease. It also
6991 gained support for including "MUD" URLs ("Manufacturer Usage
6992 Description"). Support for "MUD" URLs was also added to the LLDP
6993 stack, configurable in the [LLDP] section in .network files.
6995 * The Mode= settings in [MACVLAN] and [MACVTAP] now support 'source'
6996 mode. Also, the sections now support a new setting SourceMACAddress=.
6998 * systemd-networkd's .netdev files now support a new setting
6999 VLANProtocol= in the [Bridge] section that allows configuration of
7000 the VLAN protocol to use.
7002 * systemd-networkd supports a new Group= setting in the [Link] section
7003 of the .network files, to control the link group.
7005 * systemd-networkd's [Network] section gained a new
7006 IPv6LinkLocalAddressGenerationMode= setting, which specifies how IPv6
7007 link local address is generated.
7009 * A new default .network file is now shipped that matches TUN/TAP
7010 devices that begin with "vt-" in their name. Such interfaces will
7011 have IP routing onto the host links set up automatically. This is
7012 supposed to be used by VM managers to trivially acquire a network
7013 interface which is fully set up for host communication, simply by
7014 carefully picking an interface name to use.
7016 * systemd-networkd's [DHCPv6] section gained a new setting RouteMetric=
7017 which sets the route priority for routes specified by the DHCP server.
7019 * systemd-networkd's [DHCPv6] section gained a new setting VendorClass=
7020 which configures the vendor class information sent to DHCP server.
7022 * The BlackList= settings in .network files' [DHCPv4] and
7023 [IPv6AcceptRA] sections have been renamed DenyList=. The old names
7024 are still understood to provide compatibility.
7026 * networkctl gained the new "forcerenew" command for forcing all DHCP
7027 server clients to renew their lease. The interface "status" output
7028 will now show numerous additional fields of information about an
7029 interface. There are new "up" and "down" commands to bring specific
7030 interfaces up or down.
7032 * systemd-resolved's DNS= configuration option now optionally accepts a
7033 port number (after ":") and a host name (after "#"). When the host
7034 name is specified, the DNS-over-TLS certificate is validated to match
7035 the specified hostname. Additionally, in case of IPv6 addresses, an
7036 interface may be specified (after "%").
7038 * systemd-resolved may be configured to forward single-label DNS names.
7039 This is not standard-conformant, but may make sense in setups where
7040 public DNS servers are not used.
7042 * systemd-resolved's DNS-over-TLS support gained SNI validation.
7044 * systemd-nspawn's --resolv-conf= switch gained a number of new
7045 supported values. Specifically, options starting with "replace-" are
7046 like those prefixed "copy-" but replace any existing resolv.conf
7047 file. And options ending in "-uplink" and "-stub" can now be used to
7048 propagate other flavours of resolv.conf into the container (as
7049 defined by systemd-resolved).
7051 * The various programs included in systemd can now optionally output
7052 their log messages on stderr prefixed with a timestamp, controlled by
7053 the $SYSTEMD_LOG_TIME environment variable.
7055 * systemctl gained a new "-P" switch that is a shortcut for "--value
7058 * "systemctl list-units" and "systemctl list-machines" no longer hide
7059 their first output column with --no-legend. To hide the first column,
7062 * "systemctl reboot" takes the option "--reboot-argument=".
7063 The optional positional argument to "systemctl reboot" is now
7064 being deprecated in favor of this option.
7066 * systemd-run gained a new switch --slice-inherit. If specified the
7067 unit it generates is placed in the same slice as the systemd-run
7070 * systemd-journald gained support for zstd compression of large fields
7071 in journal files. The hash tables in journal files have been hardened
7072 against hash collisions. This is an incompatible change and means
7073 that journal files created with new systemd versions are not readable
7074 with old versions. If the $SYSTEMD_JOURNAL_KEYED_HASH boolean
7075 environment variable for systemd-journald.service is set to 0 this
7076 new hardening functionality may be turned off, so that generated
7077 journal files remain compatible with older journalctl
7080 * journalctl will now include a clickable link in the default output for
7081 each log message for which a URL with further documentation is
7082 known. This is only supported on terminal emulators that support
7083 clickable hyperlinks, and is turned off if a pager is used (since
7084 "less" still doesn't support hyperlinks,
7085 unfortunately). Documentation URLs may be included in log messages
7086 either by including a DOCUMENTATION= journal field in it, or by
7087 associating a journal message catalog entry with the log message's
7088 MESSAGE_ID, which then carries a "Documentation:" tag.
7090 * journald.conf gained a new boolean setting Audit= that may be used to
7091 control whether systemd-journald will enable audit during
7094 * when systemd-journald's log stream is broken up into multiple lines
7095 because the PID of the sender changed this is indicated in the
7096 generated log records via the _LINE_BREAK=pid-change field.
7098 * journalctl's "-o cat" output mode will now show one or more journal
7099 fields specified with --output-fields= instead of unconditionally
7100 MESSAGE=. This is useful to retrieve a very specific set of fields
7101 without any decoration.
7103 * The sd-journal.h API gained two new functions:
7104 sd_journal_enumerate_available_unique() and
7105 sd_journal_enumerate_available_data() that operate like their
7106 counterparts that lack the _available_ in the name, but skip items
7107 that cannot be read and processed by the local implementation
7108 (i.e. are compressed in an unsupported format or such),
7110 * coredumpctl gained a new --file= switch, matching the same one in
7111 journalctl: a specific journal file may be specified to read the
7114 * coredumps collected by systemd-coredump may now be compressed using
7117 * systemd-binfmt gained a new switch --unregister for unregistering all
7118 registered entries at once. This is now invoked automatically at
7119 shutdown, so that binary formats registered with the "F" flag will
7120 not block clean file system unmounting.
7122 * systemd-notify's --pid= switch gained new values: "parent", "self",
7123 "auto" for controlling which PID to send to the service manager: the
7124 systemd-notify process' PID, or the one of the process invoking it.
7126 * systemd-logind's Session bus object learnt a new method call
7127 SetType() for temporarily updating the session type of an already
7128 allocated session. This is useful for upgrading tty sessions to
7129 graphical ones once a compositor is invoked.
7131 * systemd-socket-proxy gained a new switch --exit-idle-time= for
7132 configuring an exit-on-idle time.
7134 * systemd-repart's --empty= setting gained a new value "create". If
7135 specified a new empty regular disk image file is created under the
7136 specified name. Its size may be specified with the new --size=
7137 option. The latter is also supported without the "create" mode, in
7138 order to grow existing disk image files to the specified size. These
7139 two new options are useful when creating or manipulating disk images
7140 instead of operating on actual block devices.
7142 * systemd-repart drop-ins now support a new UUID= setting to control
7143 the UUID to assign to a newly created partition.
7145 * systemd-repart's SizeMin= per-partition parameter now defaults to 10M
7148 * systemd-repart's Label= setting now support the usual, simple
7149 specifier expansion.
7151 * systemd-homed's LUKS backend gained the ability to discard empty file
7152 system blocks automatically when the user logs out. This is enabled
7153 by default to ensure that home directories take minimal space when
7154 logged out but get full size guarantees when logged in. This may be
7155 controlled with the new --luks-offline-discard= switch to homectl.
7157 * If systemd-homed detects that /home/ is encrypted as a whole it will
7158 now default to the directory or subvolume backends instead of the
7159 LUKS backend, in order to avoid double encryption. The default
7160 storage and file system may now be configured explicitly, too, via
7161 the new /etc/systemd/homed.conf configuration file.
7163 * systemd-homed now supports unlocking home directories with FIDO2
7164 security tokens that support the 'hmac-secret' extension, in addition
7165 to the existing support for PKCS#11 security token unlocking
7166 support. Note that many recent hardware security tokens support both
7167 interfaces. The FIDO2 support is accessible via homectl's
7168 --fido2-device= option.
7170 * homectl's --pkcs11-uri= setting now accepts two special parameters:
7171 if "auto" is specified and only one suitable PKCS#11 security token
7172 is plugged in, its URL is automatically determined and enrolled for
7173 unlocking the home directory. If "list" is specified a brief table of
7174 suitable PKCS#11 security tokens is shown. Similar, the new
7175 --fido2-device= option also supports these two special values, for
7176 automatically selecting and listing suitable FIDO2 devices.
7178 * The /etc/crypttab tmp option now optionally takes an argument
7179 selecting the file system to use. Moreover, the default is now
7180 changed from ext2 to ext4.
7182 * There's a new /etc/crypttab option "keyfile-erase". If specified the
7183 key file listed in the same line is removed after use, regardless if
7184 volume activation was successful or not. This is useful if the key
7185 file is only acquired transiently at runtime and shall be erased
7186 before the system continues to boot.
7188 * There's also a new /etc/crypttab option "try-empty-password". If
7189 specified, before asking the user for a password it is attempted to
7190 unlock the volume with an empty password. This is useful for
7191 installing encrypted images whose password shall be set on first boot
7192 instead of at installation time.
7194 * systemd-cryptsetup will now attempt to load the keys to unlock
7195 volumes with automatically from files in
7196 /etc/cryptsetup-keys.d/<volume>.key and
7197 /run/cryptsetup-keys.d/<volume>.key, if any of these files exist.
7199 * systemd-cryptsetup may now activate Microsoft BitLocker volumes via
7200 /etc/crypttab, during boot.
7202 * logind.conf gained a new RuntimeDirectoryInodesMax= setting to
7203 control the inode limit for the per-user $XDG_RUNTIME_DIR tmpfs
7206 * A new generator systemd-xdg-autostart-generator has been added. It
7207 generates systemd unit files from XDG autostart .desktop files, and
7208 may be used to let the systemd user instance manage services that are
7209 started automatically as part of the desktop session.
7211 * "bootctl" gained a new verb "reboot-to-firmware" that may be used
7212 to query and change the firmware's 'Reboot Into Firmware Interface'
7215 * systemd-firstboot gained a new switch --kernel-command-line= that may
7216 be used to initialize the /etc/kernel/cmdline file of the image. It
7217 also gained a new switch --root-password-hashed= which is like
7218 --root-password= but accepts a pre-hashed UNIX password as
7219 argument. The new option --delete-root-password may be used to unset
7220 any password for the root user (dangerous!). The --root-shell= switch
7221 may be used to control the shell to use for the root account. A new
7222 --force option may be used to override any already set settings with
7223 the parameters specified on the command line (by default, the tool
7224 will not override what has already been set before, i.e. is purely
7227 * systemd-firstboot gained support for a new --image= switch, which is
7228 similar to --root= but accepts the path to a disk image file, on
7229 which it then operates.
7231 * A new sd-path.h API has been added to libsystemd. It provides a
7232 simple API for retrieving various search paths and primary
7233 directories for various resources.
7235 * A new call sd_notify_barrier() has been added to the sd-daemon.h
7236 API. The call will block until all previously sent sd_notify()
7237 messages have been processed by the service manager. This is useful
7238 to remove races caused by a process already having disappeared at the
7239 time a notification message is processed by the service manager,
7240 making correct attribution impossible. The systemd-notify tool will
7241 now make use of this call implicitly, but this can be turned off again
7242 via the new --no-block switch.
7244 * When sending a file descriptor (fd) to the service manager to keep
7245 track of, using the sd_notify() mechanism, a new parameter FDPOLL=0
7246 may be specified. If passed the service manager will refrain from
7247 poll()ing on the file descriptor. Traditionally (and when the
7248 parameter is not specified), the service manager will poll it for
7249 POLLHUP or POLLERR events, and immediately close the fds in that
7252 * The service manager (PID1) gained a new D-Bus method call
7253 SetShowStatus() which may be used to control whether it shall show
7254 boot-time status output on the console. This method has a similar
7255 effect to sending SIGRTMIN+20/SIGRTMIN+21 to PID 1.
7257 * The sd-bus API gained a number of convenience functions that take
7258 va_list arguments rather than "...". For example, there's now
7259 sd_bus_call_methodv() to match sd_bus_call_method(). Those calls make
7260 it easier to build wrappers that accept variadic arguments and want
7261 to pass a ready va_list structure to sd-bus.
7263 * sd-bus vtable entries can have a new SD_BUS_VTABLE_ABSOLUTE_OFFSET
7264 flag which alters how the userdata pointer to pass to the callbacks
7265 is determined. When the flag is set, the offset field is converted
7266 as-is into a pointer, without adding it to the object pointer the
7267 vtable is associated with.
7269 * sd-bus now exposes four new functions:
7270 sd_bus_interface_name_is_valid() + sd_bus_service_name_is_valid() +
7271 sd_bus_member_name_is_valid() + sd_bus_object_path_is_valid() will
7272 validate strings to check if they qualify as various D-Bus concepts.
7274 * The sd-bus API gained the SD_BUS_METHOD_WITH_ARGS(),
7275 SD_BUS_METHOD_WITH_ARGS_OFFSET() and SD_BUS_SIGNAL_WITH_ARGS() macros
7276 that simplify adding argument names to D-Bus methods and signals.
7278 * The man pages for the sd-bus and sd-hwdb APIs have been completed.
7280 * Various D-Bus APIs of systemd daemons now have man pages that
7281 document the methods, signals and properties.
7283 * The expectations on user/group name syntax are now documented in
7284 detail; documentation on how classic home directories may be
7285 converted into home directories managed by homed has been added;
7286 documentation regarding integration of homed/userdb functionality in
7287 desktops has been added:
7289 https://systemd.io/USER_NAMES
7290 https://systemd.io/CONVERTING_TO_HOMED
7291 https://systemd.io/USERDB_AND_DESKTOPS
7293 * Documentation for the on-disk Journal file format has been updated
7294 and has now moved to:
7296 https://systemd.io/JOURNAL_FILE_FORMAT
7298 * The interface for containers (https://systemd.io/CONTAINER_INTERFACE)
7299 has been extended by a set of environment variables that expose
7300 select fields from the host's os-release file to the container
7301 payload. Similarly, host's os-release files can be mounted into the
7302 container underneath /run/host. Together, those mechanisms provide a
7303 standardized way to expose information about the host to the
7304 container payload. Both interfaces are implemented in systemd-nspawn.
7306 * All D-Bus services shipped in systemd now implement the generic
7307 LogControl1 D-Bus API which allows clients to change log level +
7308 target of the service during runtime.
7310 * Only relevant for developers: the mkosi.default symlink has been
7311 dropped from version control. Please create a symlink to one of the
7312 distribution-specific defaults in .mkosi/ based on your preference.
7314 Contributions from: 24bisquitz, Adam Nielsen, Alan Perry, Alexander
7315 Malafeev, Amitanand.Chikorde, Alin Popa, Alvin Šipraga, Amos Bird,
7316 Andreas Rammhold, AndreRH, Andrew Doran, Anita Zhang, Ankit Jain,
7317 antznin, Arnaud Ferraris, Arthur Moraes do Lago, Arusekk, Balaji
7318 Punnuru, Balint Reczey, Bastien Nocera, bemarek, Benjamin Berg,
7319 Benjamin Dahlhoff, Benjamin Robin, Chris Down, Chris Kerr, Christian
7320 Göttsche, Christian Hesse, Christian Oder, Ciprian Hacman, Clinton Roy,
7321 codicodi, Corey Hinshaw, Daan De Meyer, Dana Olson, Dan Callaghan,
7322 Daniel Fullmer, Daniel Rusek, Dan Streetman, Dave Reisner, David
7323 Edmundson, David Wood, Denis Pronin, Diego Escalante Urrelo, Dimitri
7324 John Ledkov, dolphrundgren, duguxy, Einsler Lee, Elisei Roca, Emmanuel
7325 Garette, Eric Anderson, Eric DeVolder, Evgeny Vereshchagin,
7326 ExtinctFire, fangxiuning, Ferran Pallarès Roca, Filipe Brandenburger,
7327 Filippo Falezza, Finn, Florian Klink, Florian Mayer, Franck Bui,
7328 Frantisek Sumsal, gaurav, Georg Müller, Gergely Polonkai, Giedrius
7329 Statkevičius, Gigadoc2, gogogogi, Gaurav Singh, gzjsgdsb, Hans de
7330 Goede, Haochen Tong, ianhi, ignapk, Jakov Smolic, James T. Lee, Jan
7331 Janssen, Jan Klötzke, Jan Palus, Jay Burger, Jeremy Cline, Jérémy
7332 Rosen, Jian-Hong Pan, Jiri Slaby, Joel Shapiro, Joerg Behrmann, Jörg
7333 Thalheim, Jouke Witteveen, Kai-Heng Feng, Kenny Levinsen, Kevin
7334 Kuehler, Kumar Kartikeya Dwivedi, layderv, laydervus, Lénaïc Huard,
7335 Lennart Poettering, Lidong Zhong, Luca Boccassi, Luca BRUNO, Lucas
7336 Werkmeister, Lukas Klingsbo, Lukáš Nykrýn, Łukasz Stelmach, Maciej
7337 S. Szmigiero, MadMcCrow, Marc-André Lureau, Marcel Holtmann, Marc
7338 Kleine-Budde, Martin Hundebøll, Matthew Leeds, Matt Ranostay, Maxim
7339 Fomin, MaxVerevkin, Michael Biebl, Michael Chapman, Michael Gubbels,
7340 Michael Marley, Michał Bartoszkiewicz, Michal Koutný, Michal Sekletár,
7341 Mike Gilbert, Mike Kazantsev, Mikhail Novosyolov, ml, Motiejus Jakštys,
7342 nabijaczleweli, nerdopolis, Niccolò Maggioni, Niklas Hambüchen, Norbert
7343 Lange, Paul Cercueil, pelzvieh, Peter Hutterer, Piero La Terza, Pieter
7344 Lexis, Piotr Drąg, Rafael Fontenelle, Richard Petri, Ronan Pigott, Ross
7345 Lagerwall, Rubens Figueiredo, satmandu, Sean-StarLabs, Sebastian
7346 Jennen, sterlinghughes, Surhud More, Susant Sahani, szb512, Thomas
7347 Haller, Tobias Hunger, Tom, Tomáš Pospíšek, Tomer Shechner, Tom Hughes,
7348 Topi Miettinen, Tudor Roman, Uwe Kleine-König, Valery0xff, Vito Caputo,
7349 Vladimir Panteleev, Vladyslav Tronko, Wen Yang, Yegor Vialov, Yigal
7350 Korman, Yi Gao, YmrDtnJu, Yuri Chornoivan, Yu Watanabe, Zbigniew
7351 Jędrzejewski-Szmek, Zhu Li, Дамјан Георгиевски, наб
7353 – Warsaw, 2020-07-30
7357 * A new tool "systemd-repart" has been added, that operates as an
7358 idempotent declarative repartitioner for GPT partition tables.
7359 Specifically, a set of partitions that must or may exist can be
7360 configured via drop-in files, and during every boot the partition
7361 table on disk is compared with these files, creating missing
7362 partitions or growing existing ones based on configurable relative
7363 and absolute size constraints. The tool is strictly incremental,
7364 i.e. does not delete, shrink or move partitions, but only adds and
7365 grows them. The primary use-case is OS images that ship in minimized
7366 form, that on first boot are grown to the size of the underlying
7367 block device or augmented with additional partitions. For example,
7368 the root partition could be extended to cover the whole disk, or a
7369 swap or /home partitions could be added on first boot. It can also be
7370 used for systems that use an A/B update scheme but ship images with
7371 just the A partition, with B added on first boot. The tool is
7372 primarily intended to be run in the initrd, shortly before
7373 transitioning into the host OS, but can also be run after the
7374 transition took place. It automatically discovers the disk backing
7375 the root file system, and should hence not require any additional
7376 configuration besides the partition definition drop-ins. If no
7377 configuration drop-ins are present, no action is taken.
7379 * A new component "userdb" has been added, along with a small daemon
7380 "systemd-userdbd.service" and a client tool "userdbctl". The framework
7381 allows defining rich user and group records in a JSON format,
7382 extending on the classic "struct passwd" and "struct group"
7383 structures. Various components in systemd have been updated to
7384 process records in this format, including systemd-logind and
7385 pam-systemd. The user records are intended to be extensible, and
7386 allow setting various resource management, security and runtime
7387 parameters that shall be applied to processes and sessions of the
7388 user as they log in. This facility is intended to allow associating
7389 such metadata directly with user/group records so that they can be
7390 produced, extended and consumed in unified form. We hope that
7391 eventually frameworks such as sssd will generate records this way, so
7392 that for the first time resource management and various other
7393 per-user settings can be configured in LDAP directories and then
7394 provided to systemd (specifically to systemd-logind and pam-system)
7395 to apply on login. For further details see:
7397 https://systemd.io/USER_RECORD
7398 https://systemd.io/GROUP_RECORD
7399 https://systemd.io/USER_GROUP_API
7401 * A small new service systemd-homed.service has been added, that may be
7402 used to securely manage home directories with built-in encryption.
7403 The complete user record data is unified with the home directory,
7404 thus making home directories naturally migratable. Its primary
7405 back-end is based on LUKS volumes, but fscrypt, plain directories,
7406 and other storage schemes are also supported. This solves a couple of
7407 problems we saw with traditional ways to manage home directories, in
7408 particular when it comes to encryption. For further discussion of
7409 this, see the video of Lennart's talk at AllSystemsGo! 2019:
7411 https://media.ccc.de/v/ASG2019-164-reinventing-home-directories
7413 For further details about the format and expectations on home
7414 directories this new daemon makes, see:
7416 https://systemd.io/HOME_DIRECTORY
7418 * systemd-journald is now multi-instantiable. In addition to the main
7419 instance systemd-journald.service there's now a template unit
7420 systemd-journald@.service, with each instance defining a new named
7421 log 'namespace' (whose name is specified via the instance part of the
7422 unit name). A new unit file setting LogNamespace= has been added,
7423 taking such a namespace name, that assigns services to the specified
7424 log namespaces. As each log namespace is serviced by its own
7425 independent journal daemon, this functionality may be used to improve
7426 performance and increase isolation of applications, at the price of
7427 losing global message ordering. Each instance of journald has a
7428 separate set of configuration files, with possibly different disk
7429 usage limitations and other settings.
7431 journalctl now takes a new option --namespace= to show logs from a
7432 specific log namespace. The sd-journal.h API gained
7433 sd_journal_open_namespace() for opening the log stream of a specific
7434 log namespace. systemd-journald also gained the ability to exit on
7435 idle, which is useful in the context of log namespaces, as this means
7436 log daemons for log namespaces can be activated automatically on
7437 demand and will stop automatically when no longer used, minimizing
7440 * When systemd-tmpfiles copies a file tree using the 'C' line type it
7441 will now label every copied file according to the SELinux database.
7443 * When systemd/PID 1 detects it is used in the initrd it will now boot
7444 into initrd.target rather than default.target by default. This should
7445 make it simpler to build initrds with systemd as for many cases the
7446 only difference between a host OS image and an initrd image now is
7447 the presence of the /etc/initrd-release file.
7449 * A new kernel command line option systemd.cpu_affinity= is now
7450 understood. It's equivalent to the CPUAffinity= option in
7451 /etc/systemd/system.conf and allows setting the CPU mask for PID 1
7452 itself and the default for all other processes.
7454 * When systemd/PID 1 is reloaded (with systemctl daemon-reload or
7455 equivalent), the SELinux database is now reloaded, ensuring that
7456 sockets and other file system objects are generated taking the new
7457 database into account.
7459 * systemd/PID 1 accepts a new "systemd.show-status=error" setting, and
7460 "quiet" has been changed to imply that instead of
7461 "systemd.show-status=auto". In this mode, only messages about errors
7462 and significant delays in boot are shown on the console.
7464 * The sd-event.h API gained native support for the new Linux "pidfd"
7465 concept. This permits watching processes using file descriptors
7466 instead of PID numbers, which fixes a number of races and makes
7467 process supervision more robust and efficient. All of systemd's
7468 components will now use pidfds if the kernel supports it for process
7469 watching, with the exception of PID 1 itself, unfortunately. We hope
7470 to move PID 1 to exclusively using pidfds too eventually, but this
7471 requires some more kernel work first. (Background: PID 1 watches
7472 processes using waitid() with the P_ALL flag, and that does not play
7473 together nicely with pidfds yet.)
7475 * Closely related to this, the sd-event.h API gained two new calls
7476 sd_event_source_send_child_signal() (for sending a signal to a
7477 watched process) and sd_event_source_get_child_process_own() (for
7478 marking a process so that it is killed automatically whenever the
7479 event source watching it is freed).
7481 * systemd-networkd gained support for configuring Token Bucket Filter
7482 (TBF) parameters in its qdisc configuration support. Similarly,
7483 support for Stochastic Fairness Queuing (SFQ), Controlled-Delay
7484 Active Queue Management (CoDel), and Fair Queue (FQ) has been added.
7486 * systemd-networkd gained support for Intermediate Functional Block
7487 (IFB) network devices.
7489 * systemd-networkd gained support for configuring multi-path IP routes,
7490 using the new MultiPathRoute= setting in the [Route] section.
7492 * systemd-networkd's DHCPv4 client has been updated to support a new
7493 SendDecline= option. If enabled, duplicate address detection is done
7494 after a DHCP offer is received from the server. If a conflict is
7495 detected, the address is declined. The DHCPv4 client also gained
7496 support for a new RouteMTUBytes= setting that allows to configure the
7497 MTU size to be used for routes generated from DHCPv4 leases.
7499 * The PrefixRoute= setting in systemd-networkd's [Address] section of
7500 .network files has been deprecated, and replaced by AddPrefixRoute=,
7501 with its sense inverted.
7503 * The Gateway= setting of [Route] sections of .network files gained
7504 support for a special new value "_dhcp". If set, the configured
7505 static route uses the gateway host configured via DHCP.
7507 * New User= and SuppressPrefixLength= settings have been implemented
7508 for the [RoutingPolicyRule] section of .network files to configure
7509 source routing based on UID ranges and prefix length, respectively.
7511 * The Type= match property of .link files has been generalized to
7512 always match the device type shown by 'networkctl status', even for
7513 devices where udev does not set DEVTYPE=. This allows e.g. Type=ether
7516 * sd-bus gained a new API call sd_bus_message_sensitive() that marks a
7517 D-Bus message object as "sensitive". Those objects are erased from
7518 memory when they are freed. This concept is intended to be used for
7519 messages that contain security sensitive data. A new flag
7520 SD_BUS_VTABLE_SENSITIVE has been introduced as well to mark methods
7521 in sd-bus vtables, causing any incoming and outgoing messages of
7522 those methods to be implicitly marked as "sensitive".
7524 * sd-bus gained a new API call sd_bus_message_dump() for dumping the
7525 contents of a message (or parts thereof) to standard output for
7528 * systemd-sysusers gained support for creating users with the primary
7529 group named differently than the user.
7531 * systemd-growfs (i.e. the x-systemd.growfs mount option in /etc/fstab)
7532 gained support for growing XFS partitions. Previously it supported
7533 only ext4 and btrfs partitions.
7535 * The support for /etc/crypttab gained a new x-initrd.attach option. If
7536 set, the specified encrypted volume is unlocked already in the
7537 initrd. This concept corresponds to the x-initrd.mount option in
7540 * systemd-cryptsetup gained native support for unlocking encrypted
7541 volumes utilizing PKCS#11 smartcards, i.e. for example to bind
7542 encryption of volumes to YubiKeys. This is exposed in the new
7543 pkcs11-uri= option in /etc/crypttab.
7545 * The /etc/fstab support in systemd now supports two new mount options
7546 x-systemd.{required,wanted}-by=, for explicitly configuring the units
7547 that the specified mount shall be pulled in by, in place of
7548 the usual local-fs.target/remote-fs.target.
7550 * The https://systemd.io/ web site has been relaunched, directly
7551 populated with most of the documentation included in the systemd
7552 repository. systemd also acquired a new logo, thanks to Tobias
7555 * systemd-udevd gained support for managing "alternative" network
7556 interface names, as supported by new Linux kernels. For the first
7557 time this permits assigning multiple (and longer!) names to a network
7558 interface. systemd-udevd will now by default assign the names
7559 generated via all supported naming schemes to each interface. This
7560 may be further tweaked with .link files and the AlternativeName= and
7561 AlternativeNamesPolicy= settings. Other components of systemd have
7562 been updated to support the new alternative names wherever
7563 appropriate. For example, systemd-nspawn will now generate
7564 alternative interface names for the host-facing side of container
7565 veth links based on the full container name without truncation.
7567 * systemd-nspawn interface naming logic has been updated in another way
7568 too: if the main interface name (i.e. as opposed to new-style
7569 "alternative" names) based on the container name is truncated, a
7570 simple hashing scheme is used to give different interface names to
7571 multiple containers whose names all begin with the same prefix. Since
7572 this changes the primary interface names pointing to containers if
7573 truncation happens, the old scheme may still be requested by
7574 selecting an older naming scheme, via the net.naming_scheme= kernel
7575 command line option.
7577 * PrivateUsers= in service files now works in services run by the
7578 systemd --user per-user instance of the service manager.
7580 * A new per-service sandboxing option ProtectClock= has been added that
7581 locks down write access to the system clock. It takes away device
7582 node access to /dev/rtc as well as the system calls that set the
7583 system clock and the CAP_SYS_TIME and CAP_WAKE_ALARM capabilities.
7584 Note that this option does not affect access to auxiliary services
7585 that allow changing the clock, for example access to
7588 * The systemd-id128 tool gained a new "show" verb for listing or
7589 resolving a number of well-known UUIDs/128-bit IDs, currently mostly
7590 GPT partition table types.
7592 * The Discoverable Partitions Specification has been updated to support
7593 /var and /var/tmp partition discovery. Support for this has been
7594 added to systemd-gpt-auto-generator. For details see:
7596 https://systemd.io/DISCOVERABLE_PARTITIONS
7598 * "systemctl list-unit-files" has been updated to show a new column
7599 with the suggested enablement state based on the vendor preset files
7600 for the respective units.
7602 * "systemctl" gained a new option "--with-dependencies". If specified
7603 commands such as "systemctl status" or "systemctl cat" will now show
7604 all specified units along with all units they depend on.
7606 * networkctl gained support for showing per-interface logs in its
7609 * systemd-networkd-wait-online gained support for specifying the maximum
7610 operational state to wait for, and to wait for interfaces to
7613 * The [Match] section of .link and .network files now supports a new
7614 option PermanentMACAddress= which may be used to check against the
7615 permanent MAC address of a network device even if a randomized MAC
7618 * The [TrafficControlQueueingDiscipline] section in .network files has
7619 been renamed to [NetworkEmulator] with the "NetworkEmulator" prefix
7620 dropped from the individual setting names.
7622 * Any .link and .network files that have an empty [Match] section (this
7623 also includes empty and commented-out files) will now be
7624 rejected. systemd-udev and systemd-networkd started warning about
7625 such files in version 243.
7627 * systemd-logind will now validate access to the operation of changing
7628 the virtual terminal via a polkit action. By default, only users
7629 with at least one session on a local VT are granted permission.
7631 * When systemd sets up PAM sessions that invoked service processes
7632 shall run in, the pam_setcred() API is now invoked, thus permitting
7633 PAM modules to set additional credentials for the processes.
7635 * portablectl attach/detach verbs now accept --now and --enable options
7636 to combine attachment with enablement and invocation, or detachment
7637 with stopping and disablement.
7639 * UPGRADE ISSUE: a bug where some jobs were trimmed as redundant was
7640 fixed, which in turn exposed bugs in unit configuration of services
7641 which have Type=oneshot and should only run once, but do not have
7642 RemainAfterExit=yes set. Without RemainAfterExit=yes, a one-shot
7643 service may be started again after exiting successfully, for example
7644 as a dependency in another transaction. Affected services included
7645 some internal systemd services (most notably
7646 systemd-vconsole-setup.service, which was updated to have
7647 RemainAfterExit=yes), and plymouth-start.service. Please ensure that
7648 plymouth has been suitably updated or patched before upgrading to
7649 this systemd release. See
7650 https://bugzilla.redhat.com/show_bug.cgi?id=1807771 for some
7651 additional discussion.
7653 Contributions from: AJ Bagwell, Alin Popa, Andreas Rammhold, Anita
7654 Zhang, Ansgar Burchardt, Antonio Russo, Arian van Putten, Ashley Davis,
7655 Balint Reczey, Bart Willems, Bastien Nocera, Benjamin Dahlhoff, Charles
7656 (Chas) Williams, cheese1, Chris Down, Chris Murphy, Christian Ehrhardt,
7657 Christian Göttsche, cvoinf, Daan De Meyer, Daniele Medri, Daniel Rusek,
7658 Daniel Shahaf, Dann Frazier, Dan Streetman, Dariusz Gadomski, David
7659 Michael, Dimitri John Ledkov, Emmanuel Bourg, Evgeny Vereshchagin,
7660 ezst036, Felipe Sateler, Filipe Brandenburger, Florian Klink, Franck
7661 Bui, Fran Dieguez, Frantisek Sumsal, Greg "GothAck" Miell, Guilhem
7662 Lettron, Guillaume Douézan-Grard, Hans de Goede, HATAYAMA Daisuke, Iain
7663 Lane, James Buren, Jan Alexander Steffens (heftig), Jérémy Rosen, Jin
7664 Park, Jun'ichi Nomura, Kai Krakow, Kevin Kuehler, Kevin P. Fleming,
7665 Lennart Poettering, Leonid Bloch, Leonid Evdokimov, lothrond, Luca
7666 Boccassi, Lukas K, Lynn Kirby, Mario Limonciello, Mark Deneen, Matthew
7667 Leeds, Michael Biebl, Michal Koutný, Michal Sekletár, Mike Auty, Mike
7668 Gilbert, mtron, nabijaczleweli, Naïm Favier, Nate Jones, Norbert Lange,
7669 Oliver Giles, Paul Davey, Paul Menzel, Peter Hutterer, Piotr Drąg, Rafa
7670 Couto, Raphael, rhn, Robert Scheck, Rocka, Romain Naour, Ryan Attard,
7671 Sascha Dewald, Shengjing Zhu, Slava Kardakov, Spencer Michaels, Sylvain
7672 Plantefeve, Stanislav Angelovič, Susant Sahani, Thomas Haller, Thomas
7673 Schmitt, Timo Schlüßler, Timo Wilken, Tobias Bernard, Tobias Klauser,
7674 Tobias Stoeckmann, Topi Miettinen, tsia, WataruMatsuoka, Wieland
7675 Hoffmann, Wilhelm Schuster, Will Fleming, xduugu, Yong Cong Sin, Yuri
7676 Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zeyu
7679 – Warsaw, 2020-03-06
7683 * Support for the cpuset cgroups v2 controller has been added.
7684 Processes may be restricted to specific CPUs using the new
7685 AllowedCPUs= setting, and to specific memory NUMA nodes using the new
7686 AllowedMemoryNodes= setting.
7688 * The signal used in restart jobs (as opposed to e.g. stop jobs) may
7689 now be configured using a new RestartKillSignal= setting. This
7690 allows units which signals to request termination to implement
7691 different behaviour when stopping in preparation for a restart.
7693 * "systemctl clean" may now be used also for socket, mount, and swap
7696 * systemd will also read configuration options from the EFI variable
7697 SystemdOptions. This may be used to configure systemd behaviour when
7698 modifying the kernel command line is inconvenient, but configuration
7699 on disk is read too late, for example for the options related to
7700 cgroup hierarchy setup. 'bootctl systemd-efi-options' may be used to
7701 set the EFI variable.
7703 * systemd will now disable printk ratelimits in early boot. This should
7704 allow us to capture more logs from the early boot phase where normal
7705 storage is not available and the kernel ring buffer is used for
7706 logging. Configuration on the kernel command line has higher priority
7707 and overrides the systemd setting.
7709 systemd programs which log to /dev/kmsg directly use internal
7710 ratelimits to prevent runaway logging. (Normally this is only used
7711 during early boot, so in practice this change has very little
7714 * Unit files now support top level dropin directories of the form
7715 <unit_type>.d/ (e.g. service.d/) that may be used to add configuration
7716 that affects all corresponding unit files.
7718 * systemctl gained support for 'stop --job-mode=triggering' which will
7719 stop the specified unit and any units which could trigger it.
7721 * Unit status display now includes units triggering and triggered by
7722 the unit being shown.
7724 * The RuntimeMaxSec= setting is now supported by scopes, not just
7725 .service units. This is particularly useful for PAM sessions which
7726 create a scope unit for the user login. systemd.runtime_max_sec=
7727 setting may used with the pam_systemd module to limit the duration
7728 of the PAM session, for example for time-limited logins.
7730 * A new @pkey system call group is now defined to make it easier to
7731 allow-list memory protection syscalls for containers and services
7732 which need to use them.
7734 * systemd-udevd: removed the 30s timeout for killing stale workers on
7735 exit. systemd-udevd now waits for workers to finish. The hard-coded
7736 exit timeout of 30s was too short for some large installations, where
7737 driver initialization could be prematurely interrupted during initrd
7738 processing if the root file system had been mounted and init was
7739 preparing to switch root. If udevd is run without systemd and workers
7740 are hanging while udevd receives an exit signal, udevd will now exit
7741 when udev.event_timeout is reached for the last hanging worker. With
7742 systemd, the exit timeout can additionally be configured using
7743 TimeoutStopSec= in systemd-udevd.service.
7745 * udev now provides a program (fido_id) that identifies FIDO CTAP1
7746 ("U2F")/CTAP2 security tokens based on the usage declared in their
7747 report and descriptor and outputs suitable environment variables.
7748 This replaces the externally maintained allow lists of all known
7749 security tokens that were used previously.
7751 * Automatically generated autosuspend udev rules for allow-listed
7752 devices have been imported from the Chromium OS project. This should
7753 improve power saving with many more devices.
7755 * udev gained a new "CONST{key}=value" setting that allows matching
7756 against system-wide constants without forking a helper binary.
7757 Currently "arch" and "virt" keys are supported.
7759 * udev now opens CDROMs in non-exclusive mode when querying their
7760 capabilities. This should fix issues where other programs trying to
7761 use the CDROM cannot gain access to it, but carries a risk of
7762 interfering with programs writing to the disk, if they did not open
7763 the device in exclusive mode as they should.
7765 * systemd-networkd does not create a default route for IPv4 link local
7766 addressing anymore. The creation of the route was unexpected and was
7767 breaking routing in various cases, but people who rely on it being
7768 created implicitly will need to adjust. Such a route may be requested
7769 with DefaultRouteOnDevice=yes.
7771 Similarly, systemd-networkd will not assign a link-local IPv6 address
7772 when IPv6 link-local routing is not enabled.
7774 * Receive and transmit buffers may now be configured on links with
7775 the new RxBufferSize= and TxBufferSize= settings.
7777 * systemd-networkd may now advertise additional IPv6 routes. A new
7778 [IPv6RoutePrefix] section with Route= and LifetimeSec= options is
7781 * systemd-networkd may now configure "next hop" routes using the
7782 [NextHop] section and Gateway= and Id= settings.
7784 * systemd-networkd will now retain DHCP config on restarts by default
7785 (but this may be overridden using the KeepConfiguration= setting).
7786 The default for SendRelease= has been changed to true.
7788 * The DHCPv4 client now uses the OPTION_INFORMATION_REFRESH_TIME option
7789 received from the server.
7791 The client will use the received SIP server list if UseSIP=yes is
7794 The client may be configured to request specific options from the
7795 server using a new RequestOptions= setting.
7797 The client may be configured to send arbitrary options to the server
7798 using a new SendOption= setting.
7800 A new IPServiceType= setting has been added to configure the "IP
7801 service type" value used by the client.
7803 * The DHCPv6 client learnt a new PrefixDelegationHint= option to
7804 request prefix hints in the DHCPv6 solicitation.
7806 * The DHCPv4 server may be configured to send arbitrary options using
7807 a new SendOption= setting.
7809 * The DHCPv4 server may now be configured to emit SIP server list using
7810 the new EmitSIP= and SIP= settings.
7812 * systemd-networkd and networkctl may now renew DHCP leases on demand.
7813 networkctl has a new 'networkctl renew' verb.
7815 * systemd-networkd may now reconfigure links on demand. networkctl
7816 gained two new verbs: "reload" will reload the configuration, and
7817 "reconfigure DEVICE…" will reconfigure one or more devices.
7819 * .network files may now match on SSID and BSSID of a wireless network,
7820 i.e. the access point name and hardware address using the new SSID=
7821 and BSSID= options. networkctl will display the current SSID and
7822 BSSID for wireless links.
7824 .network files may also match on the wireless network type using the
7825 new WLANInterfaceType= option.
7827 * systemd-networkd now includes default configuration that enables
7828 link-local addressing when connected to an ad-hoc wireless network.
7830 * systemd-networkd may configure the Traffic Control queueing
7831 disciplines in the kernel using the new
7832 [TrafficControlQueueingDiscipline] section and Parent=,
7833 NetworkEmulatorDelaySec=, NetworkEmulatorDelayJitterSec=,
7834 NetworkEmulatorPacketLimit=, NetworkEmulatorLossRate=,
7835 NetworkEmulatorDuplicateRate= settings.
7837 * systemd-tmpfiles gained a new w+ setting to append to files.
7839 * systemd-analyze dump will now report when the memory configuration in
7840 the kernel does not match what systemd has configured (usually,
7841 because some external program has modified the kernel configuration
7844 * systemd-analyze gained a new --base-time= switch instructs the
7845 'calendar' verb to resolve times relative to that timestamp instead
7846 of the present time.
7848 * journalctl --update-catalog now produces deterministic output (making
7849 reproducible image builds easier).
7851 * A new devicetree-overlay setting is now documented in the Boot Loader
7854 * The default value of the WatchdogSec= setting used in systemd
7855 services (the ones bundled with the project itself) may be set at
7856 configuration time using the -Dservice-watchdog= setting. If set to
7857 empty, the watchdogs will be disabled.
7859 * systemd-resolved validates IP addresses in certificates now when GnuTLS
7862 * libcryptsetup >= 2.0.1 is now required.
7864 * A configuration option -Duser-path= may be used to override the $PATH
7865 used by the user service manager. The default is again to use the same
7866 path as the system manager.
7868 * The systemd-id128 tool gained a new switch "-u" (or "--uuid") for
7869 outputting the 128-bit IDs in UUID format (i.e. in the "canonical
7872 * Service units gained a new sandboxing option ProtectKernelLogs= which
7873 makes sure the program cannot get direct access to the kernel log
7874 buffer anymore, i.e. the syslog() system call (not to be confused
7875 with the API of the same name in libc, which is not affected), the
7876 /proc/kmsg and /dev/kmsg nodes and the CAP_SYSLOG capability are made
7877 inaccessible to the service. It's recommended to enable this setting
7878 for all services that should not be able to read from or write to the
7879 kernel log buffer, which are probably almost all.
7881 Contributions from: Aaron Plattner, Alcaro, Anita Zhang, Balint Reczey,
7882 Bastien Nocera, Baybal Ni, Benjamin Bouvier, Benjamin Gilbert, Carlo
7883 Teubner, cbzxt, Chen Qi, Chris Down, Christian Rebischke, Claudio
7884 Zumbo, ClydeByrdIII, crashfistfight, Cyprien Laplace, Daniel Edgecumbe,
7885 Daniel Gorbea, Daniel Rusek, Daniel Stuart, Dan Streetman, David
7886 Pedersen, David Tardon, Dimitri John Ledkov, Dominique Martinet, Donald
7887 A. Cupp Jr, Evgeny Vereshchagin, Fabian Henneke, Filipe Brandenburger,
7888 Franck Bui, Frantisek Sumsal, Georg Müller, Hans de Goede, Haochen
7889 Tong, HATAYAMA Daisuke, Iwan Timmer, Jan Janssen, Jan Kundrát, Jan
7890 Synacek, Jan Tojnar, Jay Strict, Jérémy Rosen, Jóhann B. Guðmundsson,
7891 Jonas Jelten, Jonas Thelemann, Justin Trudell, J. Xing, Kai-Heng Feng,
7892 Kenneth D'souza, Kevin Becker, Kevin Kuehler, Lennart Poettering,
7893 Léonard Gérard, Lorenz Bauer, Luca Boccassi, Maciej Stanczew, Mario
7894 Limonciello, Marko Myllynen, Mark Stosberg, Martin Wilck, matthiasroos,
7895 Michael Biebl, Michael Olbrich, Michael Tretter, Michal Sekletar,
7896 Michal Sekletár, Michal Suchanek, Mike Gilbert, Mike Kazantsev, Nicolas
7897 Douma, nikolas, Norbert Lange, pan93412, Pascal de Bruijn, Paul Menzel,
7898 Pavel Hrdina, Peter Wu, Philip Withnall, Piotr Drąg, Rafael Fontenelle,
7899 Renaud Métrich, Riccardo Schirone, RoadrunnerWMC, Ronan Pigott, Ryan
7900 Attard, Sebastian Wick, Serge, Siddharth Chandrasekara, Steve Ramage,
7901 Steve Traylen, Susant Sahani, Thibault Nélis, Tim Teichmann, Tom
7902 Fitzhenry, Tommy J, Torsten Hilbrich, Vito Caputo, ypf791, Yu Watanabe,
7903 Zach Smith, Zbigniew Jędrzejewski-Szmek
7905 – Warsaw, 2019-11-29
7909 * This release enables unprivileged programs (i.e. requiring neither
7910 setuid nor file capabilities) to send ICMP Echo (i.e. ping) requests
7911 by turning on the "net.ipv4.ping_group_range" sysctl of the Linux
7912 kernel for the whole UNIX group range, i.e. all processes. This
7913 change should be reasonably safe, as the kernel support for it was
7914 specifically implemented to allow safe access to ICMP Echo for
7915 processes lacking any privileges. If this is not desirable, it can be
7916 disabled again by setting the parameter to "1 0".
7918 * Previously, filters defined with SystemCallFilter= would have the
7919 effect that any calling of an offending system call would terminate
7920 the calling thread. This behaviour never made much sense, since
7921 killing individual threads of unsuspecting processes is likely to
7922 create more problems than it solves. With this release the default
7923 action changed from killing the thread to killing the whole
7924 process. For this to work correctly both a kernel version (>= 4.14)
7925 and a libseccomp version (>= 2.4.0) supporting this new seccomp
7926 action is required. If an older kernel or libseccomp is used the old
7927 behaviour continues to be used. This change does not affect any
7928 services that have no system call filters defined, or that use
7929 SystemCallErrorNumber= (and thus see EPERM or another error instead
7930 of being killed when calling an offending system call). Note that
7931 systemd documentation always claimed that the whole process is
7932 killed. With this change behaviour is thus adjusted to match the
7935 * On 64 bit systems, the "kernel.pid_max" sysctl is now bumped to
7936 4194304 by default, i.e. the full 22bit range the kernel allows, up
7937 from the old 16-bit range. This should improve security and
7938 robustness, as PID collisions are made less likely (though certainly
7939 still possible). There are rumours this might create compatibility
7940 problems, though at this moment no practical ones are known to
7941 us. Downstream distributions are hence advised to undo this change in
7942 their builds if they are concerned about maximum compatibility, but
7943 for everybody else we recommend leaving the value bumped. Besides
7944 improving security and robustness this should also simplify things as
7945 the maximum number of allowed concurrent tasks was previously bounded
7946 by both "kernel.pid_max" and "kernel.threads-max" and now effectively
7947 only a single knob is left ("kernel.threads-max"). There have been
7948 concerns that usability is affected by this change because larger PID
7949 numbers are harder to type, but we believe the change from 5 digits
7950 to 7 digits doesn't hamper usability.
7952 * MemoryLow= and MemoryMin= gained hierarchy-aware counterparts,
7953 DefaultMemoryLow= and DefaultMemoryMin=, which can be used to
7954 hierarchically set default memory protection values for a particular
7955 subtree of the unit hierarchy.
7957 * Memory protection directives can now take a value of zero, allowing
7958 explicit opting out of a default value propagated by an ancestor.
7960 * systemd now defaults to the "unified" cgroup hierarchy setup during
7961 build-time, i.e. -Ddefault-hierarchy=unified is now the build-time
7962 default. Previously, -Ddefault-hierarchy=hybrid was the default. This
7963 change reflects the fact that cgroupsv2 support has matured
7964 substantially in both systemd and in the kernel, and is clearly the
7965 way forward. Downstream production distributions might want to
7966 continue to use -Ddefault-hierarchy=hybrid (or even =legacy) for
7967 their builds as unfortunately the popular container managers have not
7968 caught up with the kernel API changes.
7970 * Man pages are not built by default anymore (html pages were already
7971 disabled by default), to make development builds quicker. When
7972 building systemd for a full installation with documentation, meson
7973 should be called with -Dman=true and/or -Dhtml=true as appropriate.
7974 The default was changed based on the assumption that quick one-off or
7975 repeated development builds are much more common than full optimized
7976 builds for installation, and people need to pass various other
7977 options to when doing "proper" builds anyway, so the gain from making
7978 development builds quicker is bigger than the one time disruption for
7981 Two scripts are created in the *build* directory to generate and
7982 preview man and html pages on demand, e.g.:
7984 build/man/man systemctl
7985 build/man/html systemd.index
7987 * libidn2 is used by default if both libidn2 and libidn are installed.
7988 Please use -Dlibidn=true if libidn is preferred.
7990 * The D-Bus "wire format" of the CPUAffinity= attribute is changed on
7991 big-endian machines. Before, bytes were written and read in native
7992 machine order as exposed by the native libc __cpu_mask interface.
7993 Now, little-endian order is always used (CPUs 0–7 are described by
7994 bits 0–7 in byte 0, CPUs 8–15 are described by byte 1, and so on).
7995 This change fixes D-Bus calls that cross endianness boundary.
7997 The presentation format used for CPUAffinity= by "systemctl show" and
7998 "systemd-analyze dump" is changed to present CPU indices instead of
7999 the raw __cpu_mask bitmask. For example, CPUAffinity=0-1 would be
8000 shown as CPUAffinity=03000000000000000000000000000… (on
8001 little-endian) or CPUAffinity=00000000000000300000000000000… (on
8002 64-bit big-endian), and is now shown as CPUAffinity=0-1, matching the
8003 input format. The maximum integer that will be printed in the new
8004 format is 8191 (four digits), while the old format always used a very
8005 long number (with the length varying by architecture), so they can be
8006 unambiguously distinguished.
8008 * /usr/sbin/halt.local is no longer supported. Implementation in
8009 distributions was inconsistent and it seems this functionality was
8012 To replace this functionality, users should:
8013 - either define a new unit and make it a dependency of final.target
8014 (systemctl add-wants final.target my-halt-local.service)
8015 - or move the shutdown script to /usr/lib/systemd/system-shutdown/
8016 and ensure that it accepts "halt", "poweroff", "reboot", and
8017 "kexec" as an argument, see the description in systemd-shutdown(8).
8019 * When a [Match] section in .link or .network file is empty (contains
8020 no match patterns), a warning will be emitted. Please add any "match
8021 all" pattern instead, e.g. OriginalName=* or Name=* in case all
8022 interfaces should really be matched.
8024 * A new setting NUMAPolicy= may be used to set process memory
8025 allocation policy. This setting can be specified in
8026 /etc/systemd/system.conf and hence will set the default policy for
8027 PID1. The default policy can be overridden on a per-service
8028 basis. The related setting NUMAMask= is used to specify NUMA node
8029 mask that should be associated with the selected policy.
8031 * PID 1 will now listen to Out-Of-Memory (OOM) events the kernel
8032 generates when processes it manages are reaching their memory limits,
8033 and will place their units in a special state, and optionally kill or
8034 stop the whole unit.
8036 * The service manager will now expose bus properties for the IO
8037 resources used by units. This information is also shown in "systemctl
8038 status" now (for services that have IOAccounting=yes set). Moreover,
8039 the IO accounting data is included in the resource log message
8040 generated whenever a unit stops.
8042 * Units may now configure an explicit timeout to wait for when killed
8043 with SIGABRT, for example when a service watchdog is hit. Previously,
8044 the regular TimeoutStopSec= timeout was applied in this case too —
8045 now a separate timeout may be set using TimeoutAbortSec=.
8047 * Services may now send a special WATCHDOG=trigger message with
8048 sd_notify() to trigger an immediate "watchdog missed" event, and thus
8049 trigger service termination. This is useful both for testing watchdog
8050 handling, but also for defining error paths in services, that shall
8051 be handled the same way as watchdog events.
8053 * There are two new per-unit settings IPIngressFilterPath= and
8054 IPEgressFilterPath= which allow configuration of a BPF program
8055 (usually by specifying a path to a program uploaded to /sys/fs/bpf/)
8056 to apply to the IP packet ingress/egress path of all processes of a
8057 unit. This is useful to allow running systemd services with BPF
8058 programs set up externally.
8060 * systemctl gained a new "clean" verb for removing the state, cache,
8061 runtime or logs directories of a service while it is terminated. The
8062 new verb may also be used to remove the state maintained on disk for
8063 timer units that have Persistent= configured.
8065 * During the last phase of shutdown systemd will now automatically
8066 increase the log level configured in the "kernel.printk" sysctl so
8067 that any relevant loggable events happening during late shutdown are
8068 made visible. Previously, loggable events happening so late during
8069 shutdown were generally lost if the "kernel.printk" sysctl was set to
8070 high thresholds, as regular logging daemons are terminated at that
8071 time and thus nothing is written to disk.
8073 * If processes terminated during the last phase of shutdown do not exit
8074 quickly systemd will now show their names after a short time, to make
8075 debugging easier. After a longer timeout they are forcibly killed,
8078 * journalctl (and the other tools that display logs) will now highlight
8079 warnings in yellow (previously, both LOG_NOTICE and LOG_WARNING where
8080 shown in bright bold, now only LOG_NOTICE is). Moreover, audit logs
8081 are now shown in blue color, to separate them visually from regular
8082 logs. References to configuration files are now turned into clickable
8083 links on terminals that support that.
8085 * systemd-journald will now stop logging to /var/log/journal during
8086 shutdown when /var/ is on a separate mount, so that it can be
8087 unmounted safely during shutdown.
8089 * systemd-resolved gained support for a new 'strict' DNS-over-TLS mode.
8091 * systemd-resolved "Cache=" configuration option in resolved.conf has
8092 been extended to also accept the 'no-negative' value. Previously,
8093 only a boolean option was allowed (yes/no), having yes as the
8094 default. If this option is set to 'no-negative', negative answers are
8095 not cached while the old cache heuristics are used positive answers.
8096 The default remains unchanged.
8098 * The predictable naming scheme for network devices now supports
8099 generating predictable names for "netdevsim" devices.
8101 Moreover, the "en" prefix was dropped from the ID_NET_NAME_ONBOARD
8104 Those two changes form a new net.naming_scheme= entry. Distributions
8105 which want to preserve naming stability may want to set the
8106 -Ddefault-net-naming-scheme= configuration option.
8108 * systemd-networkd now supports MACsec, nlmon, IPVTAP and Xfrm
8109 interfaces natively.
8111 * systemd-networkd's bridge FDB support now allows configuration of a
8112 destination address for each entry (Destination=), as well as the
8113 VXLAN VNI (VNI=), as well as an option to declare what an entry is
8114 associated with (AssociatedWith=).
8116 * systemd-networkd's DHCPv4 support now understands a new MaxAttempts=
8117 option for configuring the maximum number of DHCP lease requests. It
8118 also learnt a new BlackList= option for deny-listing DHCP servers (a
8119 similar setting has also been added to the IPv6 RA client), as well
8120 as a SendRelease= option for configuring whether to send a DHCP
8121 RELEASE message when terminating.
8123 * systemd-networkd's DHCPv4 and DHCPv6 stacks can now be configured
8124 separately in the [DHCPv4] and [DHCPv6] sections.
8126 * systemd-networkd's DHCP support will now optionally create an
8127 implicit host route to the DNS server specified in the DHCP lease, in
8128 addition to the routes listed explicitly in the lease. This should
8129 ensure that in multi-homed systems DNS traffic leaves the systems on
8130 the interface that acquired the DNS server information even if other
8131 routes such as default routes exist. This behaviour may be turned on
8132 with the new RoutesToDNS= option.
8134 * systemd-networkd's VXLAN support gained a new option
8135 GenericProtocolExtension= for enabling VXLAN Generic Protocol
8136 Extension support, as well as IPDoNotFragment= for setting the IP
8137 "Don't fragment" bit on outgoing packets. A similar option has been
8138 added to the GENEVE support.
8140 * In systemd-networkd's [Route] section you may now configure
8141 FastOpenNoCookie= for configuring per-route TCP fast-open support, as
8142 well as TTLPropagate= for configuring Label Switched Path (LSP) TTL
8143 propagation. The Type= setting now supports local, broadcast,
8144 anycast, multicast, any, xresolve routes, too.
8146 * systemd-networkd's [Network] section learnt a new option
8147 DefaultRouteOnDevice= for automatically configuring a default route
8148 onto the network device.
8150 * systemd-networkd's bridging support gained two new options ProxyARP=
8151 and ProxyARPWifi= for configuring proxy ARP behaviour as well as
8152 MulticastRouter= for configuring multicast routing behaviour. A new
8153 option MulticastIGMPVersion= may be used to change bridge's multicast
8154 Internet Group Management Protocol (IGMP) version.
8156 * systemd-networkd's FooOverUDP support gained the ability to configure
8157 local and peer IP addresses via Local= and Peer=. A new option
8158 PeerPort= may be used to configure the peer's IP port.
8160 * systemd-networkd's TUN support gained a new setting VnetHeader= for
8161 tweaking Generic Segment Offload support.
8163 * The address family for policy rules may be specified using the new
8164 Family= option in the [RoutingPolicyRule] section.
8166 * networkctl gained a new "delete" command for removing virtual network
8167 devices, as well as a new "--stats" switch for showing device
8170 * networkd.conf gained a new setting SpeedMeter= and
8171 SpeedMeterIntervalSec=, to measure bitrate of network interfaces. The
8172 measured speed may be shown by 'networkctl status'.
8174 * "networkctl status" now displays MTU and queue lengths, and more
8175 detailed information about VXLAN and bridge devices.
8177 * systemd-networkd's .network and .link files gained a new Property=
8178 setting in the [Match] section, to match against devices with
8179 specific udev properties.
8181 * systemd-networkd's tunnel support gained a new option
8182 AssignToLoopback= for selecting whether to use the loopback device
8183 "lo" as underlying device.
8185 * systemd-networkd's MACAddress= setting in the [Neighbor] section has
8186 been renamed to LinkLayerAddress=, and it now allows configuration of
8189 * systemd-networkd's handling of the kernel's disable_ipv6 sysctl is
8190 simplified: systemd-networkd will disable the sysctl (enable IPv6) if
8191 IPv6 configuration (static or DHCPv6) was found for a given
8192 interface. It will not touch the sysctl otherwise.
8194 * The order of entries is $PATH used by the user manager instance was
8195 changed to put bin/ entries before the corresponding sbin/ entries.
8196 It is recommended to not rely on this order, and only ever have one
8197 binary with a given name in the system paths under /usr.
8199 * A new tool systemd-network-generator has been added that may generate
8200 .network, .netdev and .link files from IP configuration specified on
8201 the kernel command line in the format used by Dracut.
8203 * The CriticalConnection= setting in .network files is now deprecated,
8204 and replaced by a new KeepConfiguration= setting which allows more
8205 detailed configuration of the IP configuration to keep in place.
8207 * systemd-analyze gained a few new verbs:
8209 - "systemd-analyze timestamp" parses and converts timestamps. This is
8210 similar to the existing "systemd-analyze calendar" command which
8211 does the same for recurring calendar events.
8213 - "systemd-analyze timespan" parses and converts timespans (i.e.
8214 durations as opposed to points in time).
8216 - "systemd-analyze condition" will parse and test ConditionXYZ=
8219 - "systemd-analyze exit-status" will parse and convert exit status
8220 codes to their names and back.
8222 - "systemd-analyze unit-files" will print a list of all unit
8223 file paths and unit aliases.
8225 * SuccessExitStatus=, RestartPreventExitStatus=, and
8226 RestartForceExitStatus= now accept exit status names (e.g. "DATAERR"
8227 is equivalent to "65"). Those exit status name mappings may be
8228 displayed with the systemd-analyze exit-status verb describe above.
8230 * systemd-logind now exposes a per-session SetBrightness() bus call,
8231 which may be used to securely change the brightness of a kernel
8232 brightness device, if it belongs to the session's seat. By using this
8233 call unprivileged clients can make changes to "backlight" and "leds"
8234 devices securely with strict requirements on session membership.
8235 Desktop environments may use this to generically make brightness
8236 changes to such devices without shipping private SUID binaries or
8237 udev rules for that purpose.
8239 * "udevadm info" gained a --wait-for-initialization switch to wait for
8240 a device to be initialized.
8242 * systemd-hibernate-resume-generator will now look for resumeflags= on
8243 the kernel command line, which is similar to rootflags= and may be
8244 used to configure device timeout for the hibernation device.
8246 * sd-event learnt a new API call sd_event_source_disable_unref() for
8247 disabling and unref'ing an event source in a single function. A
8248 related call sd_event_source_disable_unrefp() has been added for use
8249 with gcc's cleanup extension.
8251 * The sd-id128.h public API gained a new definition
8252 SD_ID128_UUID_FORMAT_STR for formatting a 128-bit ID in UUID format
8255 * "busctl introspect" gained a new switch --xml-interface for dumping
8256 XML introspection data unmodified.
8258 * PID 1 may now show the unit name instead of the unit description
8259 string in its status output during boot. This may be configured in
8260 the StatusUnitFormat= setting in /etc/systemd/system.conf or the
8261 kernel command line option systemd.status_unit_format=.
8263 * PID 1 now understands a new option KExecWatchdogSec= in
8264 /etc/systemd/system.conf to set a watchdog timeout for kexec reboots.
8265 Previously watchdog functionality was only available for regular
8266 reboots. The new setting defaults to off, because we don't know in
8267 the general case if the watchdog will be reset after kexec (some
8268 drivers do reset it, but not all), and the new userspace might not be
8269 configured to handle the watchdog.
8271 Moreover, the old ShutdownWatchdogSec= setting has been renamed to
8272 RebootWatchdogSec= to more clearly communicate what it is about. The
8273 old name is still accepted for compatibility.
8275 * The systemd.debug_shell kernel command line option now optionally
8276 takes a tty name to spawn the debug shell on, which allows a
8277 different tty to be selected than the built-in default.
8279 * Service units gained a new ExecCondition= setting which will run
8280 before ExecStartPre= and either continue execution of the unit (for
8281 clean exit codes), stop execution without marking the unit failed
8282 (for exit codes 1 through 254), or stop execution and fail the unit
8283 (for exit code 255 or abnormal termination).
8285 * A new service systemd-pstore.service has been added that pulls data
8286 from /sys/fs/pstore/ and saves it to /var/lib/pstore for later
8289 * timedatectl gained new verbs for configuring per-interface NTP
8290 service configuration for systemd-timesyncd.
8292 * "localectl list-locales" won't list non-UTF-8 locales anymore. It's
8293 2019. (You can set non-UTF-8 locales though, if you know their name.)
8295 * If variable assignments in sysctl.d/ files are prefixed with "-" any
8296 failures to apply them are now ignored.
8298 * systemd-random-seed.service now optionally credits entropy when
8299 applying the seed to the system. Set $SYSTEMD_RANDOM_SEED_CREDIT to
8300 true for the service to enable this behaviour, but please consult the
8301 documentation first, since this comes with a couple of caveats.
8303 * systemd-random-seed.service is now a synchronization point for full
8304 initialization of the kernel's entropy pool. Services that require
8305 /dev/urandom to be correctly initialized should be ordered after this
8308 * The systemd-boot boot loader has been updated to optionally maintain
8309 a random seed file in the EFI System Partition (ESP). During the boot
8310 phase, this random seed is read and updated with a new seed
8311 cryptographically derived from it. Another derived seed is passed to
8312 the OS. The latter seed is then credited to the kernel's entropy pool
8313 very early during userspace initialization (from PID 1). This allows
8314 systems to boot up with a fully initialized kernel entropy pool from
8315 earliest boot on, and thus entirely removes all entropy pool
8316 initialization delays from systems using systemd-boot. Special care
8317 is taken to ensure different seeds are derived on system images
8318 replicated to multiple systems. "bootctl status" will show whether
8319 a seed was received from the boot loader.
8321 * bootctl gained two new verbs:
8323 - "bootctl random-seed" will generate the file in ESP and an EFI
8324 variable to allow a random seed to be passed to the OS as described
8327 - "bootctl is-installed" checks whether systemd-boot is currently
8330 * bootctl will warn if it detects that boot entries are misconfigured
8331 (for example if the kernel image was removed without purging the
8334 * A new document has been added describing systemd's use and support
8335 for the kernel's entropy pool subsystem:
8337 https://systemd.io/RANDOM_SEEDS
8339 * When the system is hibernated the swap device to write the
8340 hibernation image to is now automatically picked from all available
8341 swap devices, preferring the swap device with the highest configured
8342 priority over all others, and picking the device with the most free
8343 space if there are multiple devices with the highest priority.
8345 * /etc/crypttab support has learnt a new keyfile-timeout= per-device
8346 option that permits selecting the timeout how long to wait for a
8347 device with an encryption key before asking for the password.
8349 * IOWeight= has learnt to properly set the IO weight when using the
8350 BFQ scheduler officially found in kernels 5.0+.
8352 * A new mailing list has been created for reporting of security issues:
8353 systemd-security@redhat.com. For mode details, see
8354 https://systemd.io/CONTRIBUTING#security-vulnerability-reports.
8356 Contributions from: Aaron Barany, Adrian Bunk, Alan Jenkins, Albrecht
8357 Lohofener, Andrej Valek, Anita Zhang, Arian van Putten, Balint Reczey,
8358 Bastien Nocera, Ben Boeckel, Benjamin Robin, camoz, Chen Qi, Chris
8359 Chiu, Chris Down, Christian Göttsche, Christian Kellner, Clinton Roy,
8360 Connor Reeder, Daniel Black, Daniel Lublin, Daniele Medri, Dan
8361 Streetman, Dave Reisner, Dave Ross, David Art, David Tardon, Debarshi
8362 Ray, Dimitri John Ledkov, Dominick Grift, Donald Buczek, Douglas
8363 Christman, Eric DeVolder, EtherGraf, Evgeny Vereshchagin, Feldwor,
8364 Felix Riemann, Florian Dollinger, Francesco Pennica, Franck Bui,
8365 Frantisek Sumsal, Franz Pletz, frederik, Hans de Goede, Iago López
8366 Galeiras, Insun Pyo, Ivan Shapovalov, Iwan Timmer, Jack, Jakob
8367 Unterwurzacher, Jan Chren, Jan Klötzke, Jan Losinski, Jan Pokorný, Jan
8368 Synacek, Jan-Michael Brummer, Jeka Pats, Jeremy Soller, Jérémy Rosen,
8369 Jiri Pirko, Joe Lin, Joerg Behrmann, Joe Richey, Jóhann B. Guðmundsson,
8370 Johannes Christ, Johannes Schmitz, Jonathan Rouleau, Jorge Niedbalski,
8371 Jörg Thalheim, Kai Krakow, Kai Lüke, Karel Zak, Kashyap Chamarthy,
8372 Krayushkin Konstantin, Lennart Poettering, Lubomir Rintel, Luca
8373 Boccassi, Luís Ferreira, Marc-André Lureau, Markus Felten, Martin Pitt,
8374 Matthew Leeds, Mattias Jernberg, Michael Biebl, Michael Olbrich,
8375 Michael Prokop, Michael Stapelberg, Michael Zhivich, Michal Koutný,
8376 Michal Sekletar, Mike Gilbert, Milan Broz, Miroslav Lichvar, mpe85,
8377 Mr-Foo, Network Silence, Oliver Harley, pan93412, Paul Menzel, pEJipE,
8378 Peter A. Bigot, Philip Withnall, Piotr Drąg, Rafael Fontenelle, Robert
8379 Scheck, Roberto Santalla, Ronan Pigott, root, RussianNeuroMancer,
8380 Sebastian Jennen, shinygold, Shreyas Behera, Simon Schricker, Susant
8381 Sahani, Thadeu Lima de Souza Cascardo, Theo Ouzhinski, Thiebaud
8382 Weksteen, Thomas Haller, Thomas Weißschuh, Tomas Mraz, Tommi Rantala,
8383 Topi Miettinen, VD-Lycos, ven, Vladimir Yerilov, Wieland Hoffmann,
8384 William A. Kennington III, William Wold, Xi Ruoyao, Yuri Chornoivan,
8385 Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zhang Xianwei
8387 – Camerino, 2019-09-03
8391 * In .link files, MACAddressPolicy=persistent (the default) is changed
8392 to cover more devices. For devices like bridges, tun, tap, bond, and
8393 similar interfaces that do not have other identifying information,
8394 the interface name is used as the basis for persistent seed for MAC
8395 and IPv4LL addresses. The way that devices that were handled
8396 previously is not changed, and this change is about covering more
8397 devices then previously by the "persistent" policy.
8399 MACAddressPolicy=random may be used to force randomized MACs and
8400 IPv4LL addresses for a device if desired.
8402 Hint: the log output from udev (at debug level) was enhanced to
8403 clarify what policy is followed and which attributes are used.
8404 `SYSTEMD_LOG_LEVEL=debug udevadm test-builtin net_setup_link /sys/class/net/<name>`
8405 may be used to view this.
8407 Hint: if a bridge interface is created without any slaves, and gains
8408 a slave later, then now the bridge does not inherit slave's MAC.
8409 To inherit slave's MAC, for example, create the following file:
8411 # /etc/systemd/network/98-bridge-inherit-mac.link
8416 MACAddressPolicy=none
8419 * The .device units generated by systemd-fstab-generator and other
8420 generators do not automatically pull in the corresponding .mount unit
8421 as a Wants= dependency. This means that simply plugging in the device
8422 will not cause the mount unit to be started automatically. But please
8423 note that the mount unit may be started for other reasons, in
8424 particular if it is part of local-fs.target, and any unit which
8425 (transitively) depends on local-fs.target is started.
8427 * networkctl list/status/lldp now accept globbing wildcards for network
8428 interface names to match against all existing interfaces.
8430 * The $PIDFILE environment variable is set to point the absolute path
8431 configured with PIDFile= for processes of that service.
8433 * The fallback DNS server list was augmented with Cloudflare public DNS
8434 servers. Use `-Ddns-servers=` to set a different fallback.
8436 * A new special target usb-gadget.target will be started automatically
8437 when a USB Device Controller is detected (which means that the system
8438 is a USB peripheral).
8440 * A new unit setting CPUQuotaPeriodSec= assigns the time period
8441 relatively to which the CPU time quota specified by CPUQuota= is
8444 * A new unit setting ProtectHostname= may be used to prevent services
8445 from modifying hostname information (even if they otherwise would
8446 have privileges to do so).
8448 * A new unit setting NetworkNamespacePath= may be used to specify a
8449 namespace for service or socket units through a path referring to a
8450 Linux network namespace pseudo-file.
8452 * The PrivateNetwork= setting and JoinsNamespaceOf= dependencies now
8453 have an effect on .socket units: when used the listening socket is
8454 created within the configured network namespace instead of the host
8457 * ExecStart= command lines in unit files may now be prefixed with ':'
8458 in which case environment variable substitution is
8459 disabled. (Supported for the other ExecXYZ= settings, too.)
8461 * .timer units gained two new boolean settings OnClockChange= and
8462 OnTimezoneChange= which may be used to also trigger a unit when the
8463 system clock is changed or the local timezone is
8464 modified. systemd-run has been updated to make these options easily
8465 accessible from the command line for transient timers.
8467 * Two new conditions for units have been added: ConditionMemory= may be
8468 used to conditionalize a unit based on installed system
8469 RAM. ConditionCPUs= may be used to conditionalize a unit based on
8470 installed CPU cores.
8472 * The @default system call filter group understood by SystemCallFilter=
8473 has been updated to include the new rseq() system call introduced in
8476 * A new time-set.target has been added that indicates that the system
8477 time has been set from a local source (possibly imprecise). The
8478 existing time-sync.target is stronger and indicates that the time has
8479 been synchronized with a precise external source. Services where
8480 approximate time is sufficient should use the new target.
8482 * "systemctl start" (and related commands) learnt a new
8483 --show-transaction option. If specified brief information about all
8484 jobs queued because of the requested operation is shown.
8486 * systemd-networkd recognizes a new operation state 'enslaved', used
8487 (instead of 'degraded' or 'carrier') for interfaces which form a
8488 bridge, bond, or similar, and an new 'degraded-carrier' operational
8489 state used for the bond or bridge master interface when one of the
8490 enslaved devices is not operational.
8492 * .network files learnt the new IgnoreCarrierLoss= option for leaving
8493 networks configured even if the carrier is lost.
8495 * The RequiredForOnline= setting in .network files may now specify a
8496 minimum operational state required for the interface to be considered
8497 "online" by systemd-networkd-wait-online. Related to this
8498 systemd-networkd-wait-online gained a new option --operational-state=
8499 to configure the same, and its --interface= option was updated to
8500 optionally also take an operational state specific for an interface.
8502 * systemd-networkd-wait-online gained a new setting --any for waiting
8503 for only one of the requested interfaces instead of all of them.
8505 * systemd-networkd now implements L2TP tunnels.
8507 * Two new .network settings UseAutonomousPrefix= and UseOnLinkPrefix=
8508 may be used to cause autonomous and onlink prefixes received in IPv6
8509 Router Advertisements to be ignored.
8511 * New MulticastFlood=, NeighborSuppression=, and Learning= .network
8512 file settings may be used to tweak bridge behaviour.
8514 * The new TripleSampling= option in .network files may be used to
8515 configure CAN triple sampling.
8517 * A new .netdev settings PrivateKeyFile= and PresharedKeyFile= may be
8518 used to point to private or preshared key for a WireGuard interface.
8520 * /etc/crypttab now supports the same-cpu-crypt and
8521 submit-from-crypt-cpus options to tweak encryption work scheduling
8524 * systemd-tmpfiles will now take a BSD file lock before operating on a
8525 contents of directory. This may be used to temporarily exclude
8526 directories from aging by taking the same lock (useful for example
8527 when extracting a tarball into /tmp or /var/tmp as a privileged user,
8528 which might create files with really old timestamps, which
8529 nevertheless should not be deleted). For further details, see:
8531 https://systemd.io/TEMPORARY_DIRECTORIES
8533 * systemd-tmpfiles' h line type gained support for the
8534 FS_PROJINHERIT_FL ('P') file attribute (introduced in kernel 4.5),
8535 controlling project quota inheritance.
8537 * sd-boot and bootctl now implement support for an Extended Boot Loader
8538 (XBOOTLDR) partition, that is intended to be mounted to /boot, in
8539 addition to the ESP partition mounted to /efi or /boot/efi.
8540 Configuration file fragments, kernels, initrds and other EFI images
8541 to boot will be loaded from both the ESP and XBOOTLDR partitions.
8542 The XBOOTLDR partition was previously described by the Boot Loader
8543 Specification, but implementation was missing in sd-boot. Support for
8544 this concept allows using the sd-boot boot loader in more
8545 conservative scenarios where the boot loader itself is placed in the
8546 ESP but the kernels to boot (and their metadata) in a separate
8549 * A system may now be booted with systemd.volatile=overlay on the
8550 kernel command line, which causes the root file system to be set up
8551 an overlayfs mount combining the root-only root directory with a
8552 writable tmpfs. In this setup, the underlying root device is not
8553 modified, and any changes are lost at reboot.
8555 * Similar, systemd-nspawn can now boot containers with a volatile
8556 overlayfs root with the new --volatile=overlay switch.
8558 * systemd-nspawn can now consume OCI runtime bundles using a new
8559 --oci-bundle= option. This implementation is fully usable, with most
8560 features in the specification implemented, but since this a lot of
8561 new code and functionality, this feature should most likely not
8562 be used in production yet.
8564 * systemd-nspawn now supports various options described by the OCI
8565 runtime specification on the command-line and in .nspawn files:
8566 --inaccessible=/Inaccessible= may be used to mask parts of the file
8567 system tree, --console=/--pipe may be used to configure how standard
8568 input, output, and error are set up.
8570 * busctl learned the `emit` verb to generate D-Bus signals.
8572 * systemd-analyze cat-config may be used to gather and display
8573 configuration spread over multiple files, for example system and user
8574 presets, tmpfiles.d, sysusers.d, udev rules, etc.
8576 * systemd-analyze calendar now takes an optional new parameter
8577 --iterations= which may be used to show a maximum number of iterations
8578 the specified expression will elapse next.
8580 * The sd-bus C API gained support for naming method parameters in the
8583 * systemd-logind gained D-Bus APIs to specify the "reboot parameter"
8584 the reboot() system call expects.
8586 * journalctl learnt a new --cursor-file= option that points to a file
8587 from which a cursor should be loaded in the beginning and to which
8588 the updated cursor should be stored at the end.
8590 * ACRN hypervisor and Windows Subsystem for Linux (WSL) are now
8591 detected by systemd-detect-virt (and may also be used in
8592 ConditionVirtualization=).
8594 * The behaviour of systemd-logind may now be modified with environment
8595 variables $SYSTEMD_REBOOT_TO_FIRMWARE_SETUP,
8596 $SYSTEMD_REBOOT_TO_BOOT_LOADER_MENU, and
8597 $SYSTEMD_REBOOT_TO_BOOT_LOADER_ENTRY. They cause logind to either
8598 skip the relevant operation completely (when set to false), or to
8599 create a flag file in /run/systemd (when set to true), instead of
8600 actually commencing the real operation when requested. The presence
8601 of /run/systemd/reboot-to-firmware-setup,
8602 /run/systemd/reboot-to-boot-loader-menu, and
8603 /run/systemd/reboot-to-boot-loader-entry, may be used by alternative
8604 boot loader implementations to replace some steps logind performs
8605 during reboot with their own operations.
8607 * systemctl can be used to request a reboot into the boot loader menu
8608 or a specific boot loader entry with the new --boot-load-menu= and
8609 --boot-loader-entry= options to a reboot command. (This requires a
8610 boot loader that supports this, for example sd-boot.)
8612 * kernel-install will no longer unconditionally create the output
8613 directory (e.g. /efi/<machine-id>/<kernel-version>) for boot loader
8614 snippets, but will do only if the machine-specific parent directory
8615 (i.e. /efi/<machine-id>/) already exists. bootctl has been modified
8616 to create this parent directory during sd-boot installation.
8618 This makes it easier to use kernel-install with plugins which support
8619 a different layout of the bootloader partitions (for example grub2).
8621 * During package installation (with `ninja install`), we would create
8622 symlinks for getty@tty1.service, systemd-networkd.service,
8623 systemd-networkd.socket, systemd-resolved.service,
8624 remote-cryptsetup.target, remote-fs.target,
8625 systemd-networkd-wait-online.service, and systemd-timesyncd.service
8626 in /etc, as if `systemctl enable` was called for those units, to make
8627 the system usable immediately after installation. Now this is not
8628 done anymore, and instead calling `systemctl preset-all` is
8629 recommended after the first installation of systemd.
8631 * A new boolean sandboxing option RestrictSUIDSGID= has been added that
8632 is built on seccomp. When turned on creation of SUID/SGID files is
8635 * The NoNewPrivileges= and the new RestrictSUIDSGID= options are now
8636 implied if DynamicUser= is turned on for a service. This hardens
8637 these services, so that they neither can benefit from nor create
8638 SUID/SGID executables. This is a minor compatibility breakage, given
8639 that when DynamicUser= was first introduced SUID/SGID behaviour was
8640 unaffected. However, the security benefit of these two options is
8641 substantial, and the setting is still relatively new, hence we opted
8642 to make it mandatory for services with dynamic users.
8644 Contributions from: Adam Jackson, Alexander Tsoy, Andrey Yashkin,
8645 Andrzej Pietrasiewicz, Anita Zhang, Balint Reczey, Beniamino Galvani,
8646 Ben Iofel, Benjamin Berg, Benjamin Dahlhoff, Chris, Chris Morin,
8647 Christopher Wong, Claudius Ellsel, Clemens Gruber, dana, Daniel Black,
8648 Davide Cavalca, David Michael, David Rheinsberg, emersion, Evgeny
8649 Vereshchagin, Filipe Brandenburger, Franck Bui, Frantisek Sumsal,
8650 Giacinto Cifelli, Hans de Goede, Hugo Kindel, Ignat Korchagin, Insun
8651 Pyo, Jan Engelhardt, Jonas Dorel, Jonathan Lebon, Jonathon Kowalski,
8652 Jörg Sommer, Jörg Thalheim, Jussi Pakkanen, Kai-Heng Feng, Lennart
8653 Poettering, Lubomir Rintel, Luís Ferreira, Martin Pitt, Matthias
8654 Klumpp, Michael Biebl, Michael Niewöhner, Michael Olbrich, Michal
8655 Sekletar, Mike Lothian, Paul Menzel, Piotr Drąg, Riccardo Schirone,
8656 Robin Elvedi, Roman Kulikov, Ronald Tschalär, Ross Burton, Ryan
8657 Gonzalez, Sebastian Krzyszkowiak, Stephane Chazelas, StKob, Susant
8658 Sahani, Sylvain Plantefève, Szabolcs Fruhwald, Taro Yamada, Theo
8659 Ouzhinski, Thomas Haller, Tobias Jungel, Tom Yan, Tony Asleson, Topi
8660 Miettinen, unixsysadmin, Van Laser, Vesa Jääskeläinen, Yu, Li-Yu,
8661 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
8663 — Warsaw, 2019-04-11
8667 * The default locale can now be configured at compile time. Otherwise,
8668 a suitable default will be selected automatically (one of C.UTF-8,
8669 en_US.UTF-8, and C).
8671 * The version string shown by systemd and other tools now includes the
8672 git commit hash when built from git. An override may be specified
8673 during compilation, which is intended to be used by distributions to
8674 include the package release information.
8676 * systemd-cat can now filter standard input and standard error streams
8677 for different syslog priorities using the new --stderr-priority=
8680 * systemd-journald and systemd-journal-remote reject entries which
8681 contain too many fields (CVE-2018-16865) and set limits on the
8682 process' command line length (CVE-2018-16864).
8684 * $DBUS_SESSION_BUS_ADDRESS environment variable is set by pam_systemd
8687 * A new network device NamePolicy "keep" is implemented for link files,
8688 and used by default in 99-default.link (the fallback configuration
8689 provided by systemd). With this policy, if the network device name
8690 was already set by userspace, the device will not be renamed again.
8691 This matches the naming scheme that was implemented before
8692 systemd-240. If naming-scheme < 240 is specified, the "keep" policy
8693 is also enabled by default, even if not specified. Effectively, this
8694 means that if naming-scheme >= 240 is specified, network devices will
8695 be renamed according to the configuration, even if they have been
8696 renamed already, if "keep" is not specified as the naming policy in
8697 the .link file. The 99-default.link file provided by systemd includes
8698 "keep" for backwards compatibility, but it is recommended for user
8699 installed .link files to *not* include it.
8701 The "kernel" policy, which keeps kernel names declared to be
8702 "persistent", now works again as documented.
8704 * kernel-install script now optionally takes the paths to one or more
8705 initrd files, and passes them to all plugins.
8707 * The mincore() system call has been dropped from the @system-service
8708 system call filter group, as it is pretty exotic and may potentially
8709 used for side-channel attacks.
8711 * -fPIE is dropped from compiler and linker options. Please specify
8712 -Db_pie=true option to meson to build position-independent
8713 executables. Note that the meson option is supported since meson-0.49.
8715 * The fs.protected_regular and fs.protected_fifos sysctls, which were
8716 added in Linux 4.19 to make some data spoofing attacks harder, are
8717 now enabled by default. While this will hopefully improve the
8718 security of most installations, it is technically a backwards
8719 incompatible change; to disable these sysctls again, place the
8720 following lines in /etc/sysctl.d/60-protected.conf or a similar file:
8722 fs.protected_regular = 0
8723 fs.protected_fifos = 0
8725 Note that the similar hardlink and symlink protection has been
8726 enabled since v199, and may be disabled likewise.
8728 * The files read from the EnvironmentFile= setting in unit files now
8729 parse backslashes inside quotes literally, matching the behaviour of
8732 * udevadm trigger, udevadm control, udevadm settle and udevadm monitor
8733 now automatically become NOPs when run in a chroot() environment.
8735 * The tmpfiles.d/ "C" line type will now copy directory trees not only
8736 when the destination is so far missing, but also if it already exists
8737 as a directory and is empty. This is useful to cater for systems
8738 where directory trees are put together from multiple separate mount
8739 points but otherwise empty.
8741 * A new function sd_bus_close_unref() (and the associated
8742 sd_bus_close_unrefp()) has been added to libsystemd, that combines
8743 sd_bus_close() and sd_bus_unref() in one.
8745 * udevadm control learnt a new option for --ping for testing whether a
8746 systemd-udevd instance is running and reacting.
8748 * udevadm trigger learnt a new option for --wait-daemon for waiting
8749 systemd-udevd daemon to be initialized.
8751 Contributions from: Aaron Plattner, Alberts Muktupāvels, Alex Mayer,
8752 Ayman Bagabas, Beniamino Galvani, Burt P, Chris Down, Chris Lamb, Chris
8753 Morin, Christian Hesse, Claudius Ellsel, dana, Daniel Axtens, Daniele
8754 Medri, Dave Reisner, David Santamaría Rogado, Diego Canuhe, Dimitri
8755 John Ledkov, Evgeny Vereshchagin, Fabrice Fontaine, Filipe
8756 Brandenburger, Franck Bui, Frantisek Sumsal, govwin, Hans de Goede,
8757 James Hilliard, Jan Engelhardt, Jani Uusitalo, Jan Janssen, Jan
8758 Synacek, Jonathan McDowell, Jonathan Roemer, Jonathon Kowalski, Joost
8759 Heitbrink, Jörg Thalheim, Lance, Lennart Poettering, Louis Taylor,
8760 Lucas Werkmeister, Mantas Mikulėnas, Marc-Antoine Perennou,
8761 marvelousblack, Michael Biebl, Michael Sloan, Michal Sekletar, Mike
8762 Auty, Mike Gilbert, Mikhail Kasimov, Neil Brown, Niklas Hambüchen,
8763 Patrick Williams, Paul Seyfert, Peter Hutterer, Philip Withnall, Roger
8764 James, Ronnie P. Thomas, Ryan Gonzalez, Sam Morris, Stephan Edel,
8765 Stephan Gerhold, Susant Sahani, Taro Yamada, Thomas Haller, Topi
8766 Miettinen, YiFei Zhu, YmrDtnJu, YunQiang Su, Yu Watanabe, Zbigniew
8767 Jędrzejewski-Szmek, zsergeant77, Дамјан Георгиевски
8769 — Berlin, 2019-02-14
8773 * NoNewPrivileges=yes has been set for all long-running services
8774 implemented by systemd. Previously, this was problematic due to
8775 SELinux (as this would also prohibit the transition from PID1's label
8776 to the service's label). This restriction has since been lifted, but
8777 an SELinux policy update is required.
8778 (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.)
8780 * DynamicUser=yes is dropped from systemd-networkd.service,
8781 systemd-resolved.service and systemd-timesyncd.service, which was
8782 enabled in v239 for systemd-networkd.service and systemd-resolved.service,
8783 and since v236 for systemd-timesyncd.service. The users and groups
8784 systemd-network, systemd-resolve and systemd-timesync are created
8785 by systemd-sysusers again. Distributors or system administrators
8786 may need to create these users and groups if they not exist (or need
8787 to re-enable DynamicUser= for those units) while upgrading systemd.
8788 Also, the clock file for systemd-timesyncd may need to move from
8789 /var/lib/private/systemd/timesync/clock to /var/lib/systemd/timesync/clock.
8791 * When unit files are loaded from disk, previously systemd would
8792 sometimes (depending on the unit loading order) load units from the
8793 target path of symlinks in .wants/ or .requires/ directories of other
8794 units. This meant that unit could be loaded from different paths
8795 depending on whether the unit was requested explicitly or as a
8796 dependency of another unit, not honouring the priority of directories
8797 in search path. It also meant that it was possible to successfully
8798 load and start units which are not found in the unit search path, as
8799 long as they were requested as a dependency and linked to from
8800 .wants/ or .requires/. The target paths of those symlinks are not
8801 used for loading units anymore and the unit file must be found in
8804 * A new service type has been added: Type=exec. It's very similar to
8805 Type=simple but ensures the service manager will wait for both fork()
8806 and execve() of the main service binary to complete before proceeding
8807 with follow-up units. This is primarily useful so that the manager
8808 propagates any errors in the preparation phase of service execution
8809 back to the job that requested the unit to be started. For example,
8810 consider a service that has ExecStart= set to a file system binary
8811 that doesn't exist. With Type=simple starting the unit would be
8812 considered instantly successful, as only fork() has to complete
8813 successfully and the manager does not wait for execve(), and hence
8814 its failure is seen "too late". With the new Type=exec service type
8815 starting the unit will fail, as the manager will wait for the
8816 execve() and notice its failure, which is then propagated back to the
8819 NOTE: with the next release 241 of systemd we intend to change the
8820 systemd-run tool to default to Type=exec for transient services
8821 started by it. This should be mostly safe, but in specific corner
8822 cases might result in problems, as the systemd-run tool will then
8823 block on NSS calls (such as user name look-ups due to User=) done
8824 between the fork() and execve(), which under specific circumstances
8825 might cause problems. It is recommended to specify "-p Type=simple"
8826 explicitly in the few cases where this applies. For regular,
8827 non-transient services (i.e. those defined with unit files on disk)
8828 we will continue to default to Type=simple.
8830 * The Linux kernel's current default RLIMIT_NOFILE resource limit for
8831 userspace processes is set to 1024 (soft) and 4096
8832 (hard). Previously, systemd passed this on unmodified to all
8833 processes it forked off. With this systemd release the hard limit
8834 systemd passes on is increased to 512K, overriding the kernel's
8835 defaults and substantially increasing the number of simultaneous file
8836 descriptors unprivileged userspace processes can allocate. Note that
8837 the soft limit remains at 1024 for compatibility reasons: the
8838 traditional UNIX select() call cannot deal with file descriptors >=
8839 1024 and increasing the soft limit globally might thus result in
8840 programs unexpectedly allocating a high file descriptor and thus
8841 failing abnormally when attempting to use it with select() (of
8842 course, programs shouldn't use select() anymore, and prefer
8843 poll()/epoll, but the call unfortunately remains undeservedly popular
8844 at this time). This change reflects the fact that file descriptor
8845 handling in the Linux kernel has been optimized in more recent
8846 kernels and allocating large numbers of them should be much cheaper
8847 both in memory and in performance than it used to be. Programs that
8848 want to take benefit of the increased limit have to "opt-in" into
8849 high file descriptors explicitly by raising their soft limit. Of
8850 course, when they do that they must acknowledge that they cannot use
8851 select() anymore (and neither can any shared library they use — or
8852 any shared library used by any shared library they use and so on).
8853 Which default hard limit is most appropriate is of course hard to
8854 decide. However, given reports that ~300K file descriptors are used
8855 in real-life applications we believe 512K is sufficiently high as new
8856 default for now. Note that there are also reports that using very
8857 high hard limits (e.g. 1G) is problematic: some software allocates
8858 large arrays with one element for each potential file descriptor
8859 (Java, …) — a high hard limit thus triggers excessively large memory
8860 allocations in these applications. Hopefully, the new default of 512K
8861 is a good middle ground: higher than what real-life applications
8862 currently need, and low enough for avoid triggering excessively large
8863 allocations in problematic software. (And yes, somebody should fix
8866 * The fs.nr_open and fs.file-max sysctls are now automatically bumped
8867 to the highest possible values, as separate accounting of file
8868 descriptors is no longer necessary, as memcg tracks them correctly as
8869 part of the memory accounting anyway. Thus, from the four limits on
8870 file descriptors currently enforced (fs.file-max, fs.nr_open,
8871 RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two,
8872 and keep only the latter two. A set of build-time options
8873 (-Dbump-proc-sys-fs-file-max=false and -Dbump-proc-sys-fs-nr-open=false)
8874 has been added to revert this change in behaviour, which might be
8875 an option for systems that turn off memcg in the kernel.
8877 * When no /etc/locale.conf file exists (and hence no locale settings
8878 are in place), systemd will now use the "C.UTF-8" locale by default,
8879 and set LANG= to it. This locale is supported by various
8880 distributions including Fedora, with clear indications that upstream
8881 glibc is going to make it available too. This locale enables UTF-8
8882 mode by default, which appears appropriate for 2018.
8884 * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by
8885 default. This effectively switches the RFC3704 Reverse Path filtering
8886 from Strict mode to Loose mode. This is more appropriate for hosts
8887 that have multiple links with routes to the same networks (e.g.
8888 a client with a Wi-Fi and Ethernet both connected to the internet).
8890 Consult the kernel documentation for details on this sysctl:
8891 https://docs.kernel.org/networking/ip-sysctl.html
8893 * The v239 change to turn on "net.ipv4.tcp_ecn" by default has been
8896 * CPUAccounting=yes no longer enables the CPU controller when using
8897 kernel 4.15+ and the unified cgroup hierarchy, as required accounting
8898 statistics are now provided independently from the CPU controller.
8900 * Support for disabling a particular cgroup controller within a sub-tree
8901 has been added through the DisableControllers= directive.
8903 * cgroup_no_v1=all on the kernel command line now also implies
8904 using the unified cgroup hierarchy, unless one explicitly passes
8905 systemd.unified_cgroup_hierarchy=0 on the kernel command line.
8907 * The new "MemoryMin=" unit file property may now be used to set the
8908 memory usage protection limit of processes invoked by the unit. This
8909 controls the cgroup v2 memory.min attribute. Similarly, the new
8910 "IODeviceLatencyTargetSec=" property has been added, wrapping the new
8911 cgroup v2 io.latency cgroup property for configuring per-service I/O
8914 * systemd now supports the cgroup v2 devices BPF logic, as counterpart
8915 to the cgroup v1 "devices" cgroup controller.
8917 * systemd-escape now is able to combine --unescape with --template. It
8918 also learnt a new option --instance for extracting and unescaping the
8919 instance part of a unit name.
8921 * sd-bus now provides the sd_bus_message_readv() which is similar to
8922 sd_bus_message_read() but takes a va_list object. The pair
8923 sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout()
8924 has been added for configuring the default method call timeout to
8925 use. sd_bus_error_move() may be used to efficiently move the contents
8926 from one sd_bus_error structure to another, invalidating the
8927 source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may
8928 be used to control whether a bus connection object is automatically
8929 flushed when an sd-event loop is exited.
8931 * When processing classic BSD syslog log messages, journald will now
8932 save the original time-stamp string supplied in the new
8933 SYSLOG_TIMESTAMP= journal field. This permits consumers to
8934 reconstruct the original BSD syslog message more correctly.
8936 * StandardOutput=/StandardError= in service files gained support for
8937 new "append:…" parameters, for connecting STDOUT/STDERR of a service
8938 to a file, and appending to it.
8940 * The signal to use as last step of killing of unit processes is now
8941 configurable. Previously it was hard-coded to SIGKILL, which may now
8942 be overridden with the new KillSignal= setting. Note that this is the
8943 signal used when regular termination (i.e. SIGTERM) does not suffice.
8944 Similarly, the signal used when aborting a program in case of a
8945 watchdog timeout may now be configured too (WatchdogSignal=).
8947 * The XDG_SESSION_DESKTOP environment variable may now be configured in
8948 the pam_systemd argument line, using the new desktop= switch. This is
8949 useful to initialize it properly from a display manager without
8950 having to touch C code.
8952 * Most configuration options that previously accepted percentage values
8953 now also accept permille values with the '‰' suffix (instead of '%').
8955 * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for
8958 * systemd-resolved's configuration file resolved.conf gained a new
8959 option ReadEtcHosts= which may be used to turn off processing and
8960 honoring /etc/hosts entries.
8962 * The "--wait" switch may now be passed to "systemctl
8963 is-system-running", in which case the tool will synchronously wait
8964 until the system finished start-up.
8966 * hostnamed gained a new bus call to determine the DMI product UUID.
8968 * On x86-64 systemd will now prefer using the RDRAND processor
8969 instruction over /dev/urandom whenever it requires randomness that
8970 neither has to be crypto-grade nor should be reproducible. This
8971 should substantially reduce the amount of entropy systemd requests
8972 from the kernel during initialization on such systems, though not
8973 reduce it to zero. (Why not zero? systemd still needs to allocate
8974 UUIDs and such uniquely, which require high-quality randomness.)
8976 * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP
8977 tunnels. It also gained a new option ForceDHCPv6PDOtherInformation=
8978 for forcing the "Other Information" bit in IPv6 RA messages. The
8979 bonding logic gained four new options AdActorSystemPriority=,
8980 AdUserPortKey=, AdActorSystem= for configuring various 802.3ad
8981 aspects, and DynamicTransmitLoadBalancing= for enabling dynamic
8982 shuffling of flows. The tunnel logic gained a new
8983 IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid
8984 Deployment. The policy rule logic gained four new options IPProtocol=,
8985 SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained
8986 support for the MulticastToUnicast= option. networkd also gained
8987 support for configuring static IPv4 ARP or IPv6 neighbor entries.
8989 * .preset files (as read by 'systemctl preset') may now be used to
8990 instantiate services.
8992 * /etc/crypttab now understands the sector-size= option to configure
8993 the sector size for an encrypted partition.
8995 * Key material for encrypted disks may now be placed on a formatted
8996 medium, and referenced from /etc/crypttab by the UUID of the file
8997 system, followed by "=" suffixed by the path to the key file.
8999 * The "collect" udev component has been removed without replacement, as
9000 it is neither used nor maintained.
9002 * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=,
9003 LogsDirectory=, ConfigurationDirectory= settings are used in a
9004 service the executed processes will now receive a set of environment
9005 variables containing the full paths of these directories.
9006 Specifically, RUNTIME_DIRECTORY=, STATE_DIRECTORY, CACHE_DIRECTORY,
9007 LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options
9008 are used. Note that these options may be used multiple times per
9009 service in which case the resulting paths will be concatenated and
9010 separated by colons.
9012 * Predictable interface naming has been extended to cover InfiniBand
9013 NICs. They will be exposed with an "ib" prefix.
9015 * tmpfiles.d/ line types may now be suffixed with a '-' character, in
9016 which case the respective line failing is ignored.
9018 * .link files may now be used to configure the equivalent to the
9019 "ethtool advertise" commands.
9021 * The sd-device.h and sd-hwdb.h APIs are now exported, as an
9022 alternative to libudev.h. Previously, the latter was just an internal
9023 wrapper around the former, but now these two APIs are exposed
9026 * sd-id128.h gained a new function sd_id128_get_boot_app_specific()
9027 which calculates an app-specific boot ID similar to how
9028 sd_id128_get_machine_app_specific() generates an app-specific machine
9031 * A new tool systemd-id128 has been added that can be used to determine
9032 and generate various 128-bit IDs.
9034 * /etc/os-release gained two new standardized fields DOCUMENTATION_URL=
9037 * systemd-hibernate-resume-generator will now honor the "noresume"
9038 kernel command line option, in which case it will bypass resuming
9039 from any hibernated image.
9041 * The systemd-sleep.conf configuration file gained new options
9042 AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=,
9043 AllowHybridSleep= for prohibiting specific sleep modes even if the
9044 kernel exports them.
9046 * portablectl is now officially supported and has thus moved to
9049 * bootctl learnt the two new commands "set-default" and "set-oneshot"
9050 for setting the default boot loader item to boot to (either
9051 persistently or only for the next boot). This is currently only
9052 compatible with sd-boot, but may be implemented on other boot loaders
9053 too, that follow the boot loader interface. The updated interface is
9054 now documented here:
9056 https://systemd.io/BOOT_LOADER_INTERFACE
9058 * A new kernel command line option systemd.early_core_pattern= is now
9059 understood which may be used to influence the core_pattern PID 1
9060 installs during early boot.
9062 * busctl learnt two new options -j and --json= for outputting method
9063 call replies, properties and monitoring output in JSON.
9065 * journalctl's JSON output now supports simple ANSI coloring as well as
9066 a new "json-seq" mode for generating RFC7464 output.
9068 * Unit files now support the %g/%G specifiers that resolve to the UNIX
9069 group/GID of the service manager runs as, similar to the existing
9070 %u/%U specifiers that resolve to the UNIX user/UID.
9072 * systemd-logind learnt a new global configuration option
9073 UserStopDelaySec= that may be set in logind.conf. It specifies how
9074 long the systemd --user instance shall remain started after a user
9075 logs out. This is useful to speed up repetitive re-connections of the
9076 same user, as it means the user's service manager doesn't have to be
9077 stopped/restarted on each iteration, but can be reused between
9078 subsequent options. This setting defaults to 10s. systemd-logind also
9079 exports two new properties on its Manager D-Bus objects indicating
9080 whether the system's lid is currently closed, and whether the system
9083 * systemd gained support for a generic boot counting logic, which
9084 generically permits automatic reverting to older boot loader entries
9085 if newer updated ones don't work. The boot loader side is implemented
9086 in sd-boot, but is kept open for other boot loaders too. For details
9089 https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT
9091 * The SuccessAction=/FailureAction= unit file settings now learnt two
9092 new parameters: "exit" and "exit-force", which result in immediate
9093 exiting of the service manager, and are only useful in systemd --user
9094 and container environments.
9096 * Unit files gained support for a pair of options
9097 FailureActionExitStatus=/SuccessActionExitStatus= for configuring the
9098 exit status to use as service manager exit status when
9099 SuccessAction=/FailureAction= is set to exit or exit-force.
9101 * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service
9102 options may now be used to configure the log rate limiting applied by
9103 journald per-service.
9105 * systemd-analyze gained a new verb "timespan" for parsing and
9106 normalizing time span values (i.e. strings like "5min 7s 8us").
9108 * systemd-analyze also gained a new verb "security" for analyzing the
9109 security and sand-boxing settings of services in order to determine an
9110 "exposure level" for them, indicating whether a service would benefit
9111 from more sand-boxing options turned on for them.
9113 * "systemd-analyze syscall-filter" will now also show system calls
9114 supported by the local kernel but not included in any of the defined
9117 * .nspawn files now understand the Ephemeral= setting, matching the
9118 --ephemeral command line switch.
9120 * sd-event gained the new APIs sd_event_source_get_floating() and
9121 sd_event_source_set_floating() for controlling whether a specific
9122 event source is "floating", i.e. destroyed along with the even loop
9125 * Unit objects on D-Bus gained a new "Refs" property that lists all
9126 clients that currently have a reference on the unit (to ensure it is
9129 * The JoinControllers= option in system.conf is no longer supported, as
9130 it didn't work correctly, is hard to support properly, is legacy (as
9131 the concept only exists on cgroup v1) and apparently wasn't used.
9133 * Journal messages that are generated whenever a unit enters the failed
9134 state are now tagged with a unique MESSAGE_ID. Similarly, messages
9135 generated whenever a service process exits are now made recognizable,
9136 too. A tagged message is also emitted whenever a unit enters the
9137 "dead" state on success.
9139 * systemd-run gained a new switch --working-directory= for configuring
9140 the working directory of the service to start. A shortcut -d is
9141 equivalent, setting the working directory of the service to the
9142 current working directory of the invoking program. The new --shell
9143 (or just -S) option has been added for invoking the $SHELL of the
9144 caller as a service, and implies --pty --same-dir --wait --collect
9145 --service-type=exec. Or in other words, "systemd-run -S" is now the
9146 quickest way to quickly get an interactive in a fully clean and
9147 well-defined system service context.
9149 * machinectl gained a new verb "import-fs" for importing an OS tree
9150 from a directory. Moreover, when a directory or tarball is imported
9151 and single top-level directory found with the OS itself below the OS
9152 tree is automatically mangled and moved one level up.
9154 * systemd-importd will no longer set up an implicit btrfs loop-back
9155 file system on /var/lib/machines. If one is already set up, it will
9156 continue to be used.
9158 * A new generator "systemd-run-generator" has been added. It will
9159 synthesize a unit from one or more program command lines included in
9160 the kernel command line. This is very useful in container managers
9163 # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"'
9165 This will run "systemd-nspawn" on an image, invoke the specified
9166 command line and immediately shut down the container again, returning
9167 the command line's exit code.
9169 * The block device locking logic is now documented:
9171 https://systemd.io/BLOCK_DEVICE_LOCKING
9173 * loginctl and machinectl now optionally output the various tables in
9174 JSON using the --output= switch. It is our intention to add similar
9175 support to systemctl and all other commands.
9177 * udevadm's query and trigger verb now optionally take a .device unit
9180 * systemd-udevd's network naming logic now understands a new
9181 net.naming_scheme= kernel command line switch, which may be used to
9182 pick a specific version of the naming scheme. This helps stabilizing
9183 interface names even as systemd/udev are updated and the naming logic
9186 * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and
9187 SD_ID128_ALLF to test if a 128-bit ID is set to all 0xFF bytes, and to
9188 initialize one to all 0xFF.
9190 * After loading the SELinux policy systemd will now recursively relabel
9191 all files and directories listed in
9192 /run/systemd/relabel-extra.d/*.relabel (which should be simple
9193 newline separated lists of paths) in addition to the ones it already
9194 implicitly relabels in /run, /dev and /sys. After the relabelling is
9195 completed the *.relabel files (and /run/systemd/relabel-extra.d/) are
9196 removed. This is useful to permit initrds (i.e. code running before
9197 the SELinux policy is in effect) to generate files in the host
9198 filesystem safely and ensure that the correct label is applied during
9199 the transition to the host OS.
9201 * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding
9202 mknod() handling in user namespaces. Previously mknod() would always
9203 fail with EPERM in user namespaces. Since 4.18 mknod() will succeed
9204 but device nodes generated that way cannot be opened, and attempts to
9205 open them result in EPERM. This breaks the "graceful fallback" logic
9206 in systemd's PrivateDevices= sand-boxing option. This option is
9207 implemented defensively, so that when systemd detects it runs in a
9208 restricted environment (such as a user namespace, or an environment
9209 where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD)
9210 where device nodes cannot be created the effect of PrivateDevices= is
9211 bypassed (following the logic that 2nd-level sand-boxing is not
9212 essential if the system systemd runs in is itself already sand-boxed
9213 as a whole). This logic breaks with 4.18 in container managers where
9214 user namespacing is used: suddenly PrivateDevices= succeeds setting
9215 up a private /dev/ file system containing devices nodes — but when
9216 these are opened they don't work.
9218 At this point it is recommended that container managers utilizing
9219 user namespaces that intend to run systemd in the payload explicitly
9220 block mknod() with seccomp or similar, so that the graceful fallback
9223 We are very sorry for the breakage and the requirement to change
9224 container configurations for newer kernels. It's purely caused by an
9225 incompatible kernel change. The relevant kernel developers have been
9226 notified about this userspace breakage quickly, but they chose to
9229 * PermissionsStartOnly= setting is deprecated (but is still supported
9230 for backwards compatibility). The same functionality is provided by
9231 the more flexible "+", "!", and "!!" prefixes to ExecStart= and other
9234 * $DBUS_SESSION_BUS_ADDRESS environment variable is not set by
9235 pam_systemd anymore.
9237 * The naming scheme for network devices was changed to always rename
9238 devices, even if they were already renamed by userspace. The "kernel"
9239 policy was changed to only apply as a fallback, if no other naming
9242 * The requirements to build systemd is bumped to meson-0.46 and
9245 Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander
9246 Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson,
9247 Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov,
9248 asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt
9249 Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen
9250 Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius
9251 Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn
9252 Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner,
9253 David Anderson, Davide Cavalca, David Leeds, David Malcolm, David
9254 Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus,
9255 Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters,
9256 Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad,
9257 Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank
9258 Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe
9259 Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit
9260 Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan
9261 Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld,
9262 javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi
9263 Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens,
9264 Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi,
9265 Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry
9266 Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz
9267 Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier,
9268 Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin
9269 Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko
9270 Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck,
9271 Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich,
9272 Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal
9273 Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal
9274 Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby,
9275 Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł
9276 Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller,
9277 Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez,
9278 Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam
9279 Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher,
9280 Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee
9281 (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen
9282 Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim,
9283 Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas
9284 Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias
9285 Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore
9286 Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech
9287 Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward,
9288 Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe,
9289 Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein
9291 — Warsaw, 2018-12-21
9295 * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
9296 builtin will name network interfaces differently than in previous
9297 versions for virtual network interfaces created with SR-IOV and NPAR
9298 and for devices where the PCI network controller device does not have
9299 a slot number associated.
9301 SR-IOV virtual devices are now named based on the name of the parent
9302 interface, with a suffix of "v<N>", where <N> is the virtual device
9303 number. Previously those virtual devices were named as if completely
9306 The ninth and later NPAR virtual devices will be named following the
9307 scheme used for the first eight NPAR partitions. Previously those
9308 devices were not renamed and the kernel default (eth<n>) was used.
9310 "net_id" will also generate names for PCI devices where the PCI
9311 network controller device does not have an associated slot number
9312 itself, but one of its parents does. Previously those devices were
9313 not renamed and the kernel default (eth<n>) was used.
9315 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
9316 systemd-logind.service. Since v235, IPAddressDeny=any has been set to
9317 the unit. So, it is expected that the default behavior of
9318 systemd-logind is not changed. However, if distribution packagers or
9319 administrators disabled or modified IPAddressDeny= setting by a
9320 drop-in config file, then it may be necessary to update the file to
9321 re-enable AF_INET and AF_INET6 to support network user name services,
9324 * When the RestrictNamespaces= unit property is specified multiple
9325 times, then the specified types are merged now. Previously, only the
9326 last assignment was used. So, if distribution packagers or
9327 administrators modified the setting by a drop-in config file, then it
9328 may be necessary to update the file.
9330 * When OnFailure= is used in combination with Restart= on a service
9331 unit, then the specified units will no longer be triggered on
9332 failures that result in restarting. Previously, the specified units
9333 would be activated each time the unit failed, even when the unit was
9334 going to be restarted automatically. This behaviour contradicted the
9335 documentation. With this release the code is adjusted to match the
9338 * systemd-tmpfiles will now print a notice whenever it encounters
9339 tmpfiles.d/ lines referencing the /var/run/ directory. It will
9340 recommend reworking them to use the /run/ directory instead (for
9341 which /var/run/ is simply a symlinked compatibility alias). This way
9342 systemd-tmpfiles can properly detect line conflicts and merge lines
9343 referencing the same file by two paths, without having to access
9346 * systemctl disable/unmask/preset/preset-all cannot be used with
9347 --runtime. Previously this was allowed, but resulted in unintuitive
9348 behaviour that wasn't useful. systemctl disable/unmask will now undo
9349 both runtime and persistent enablement/masking, i.e. it will remove
9350 any relevant symlinks both in /run and /etc.
9352 * Note that all long-running system services shipped with systemd will
9353 now default to a system call allow list (rather than a deny list, as
9354 before). In particular, systemd-udevd will now enforce one too. For
9355 most cases this should be safe, however downstream distributions
9356 which disabled sandboxing of systemd-udevd (specifically the
9357 MountFlags= setting), might want to disable this security feature
9358 too, as the default allow-listing will prohibit all mount, swap,
9359 reboot and clock changing operations from udev rules.
9361 * sd-boot acquired new loader configuration settings to optionally turn
9362 off Windows and MacOS boot partition discovery as well as
9363 reboot-into-firmware menu items. It is also able to pick a better
9364 screen resolution for HiDPI systems, and now provides loader
9365 configuration settings to change the resolution explicitly.
9367 * systemd-resolved now supports DNS-over-TLS. It's still
9368 turned off by default, use DNSOverTLS=opportunistic to turn it on in
9369 resolved.conf. We intend to make this the default as soon as couple
9370 of additional techniques for optimizing the initial latency caused by
9371 establishing a TLS/TCP connection are implemented.
9373 * systemd-resolved.service and systemd-networkd.service now set
9374 DynamicUser=yes. The users systemd-resolve and systemd-network are
9375 not created by systemd-sysusers anymore.
9377 NOTE: This has a chance of breaking nss-ldap and similar NSS modules
9378 that embed a network facing module into any process using getpwuid()
9379 or related call: the dynamic allocation of the user ID for
9380 systemd-resolved.service means the service manager has to check NSS
9381 if the user name is already taken when forking off the service. Since
9382 the user in the common case won't be defined in /etc/passwd the
9383 lookup is likely to trigger nss-ldap which in turn might use NSS to
9384 ask systemd-resolved for hostname lookups. This will hence result in
9385 a deadlock: a user name lookup in order to start
9386 systemd-resolved.service will result in a hostname lookup for which
9387 systemd-resolved.service needs to be started already. There are
9388 multiple ways to work around this problem: pre-allocate the
9389 "systemd-resolve" user on such systems, so that nss-ldap won't be
9390 triggered; or use a different NSS package that doesn't do networking
9391 in-process but provides a local asynchronous name cache; or configure
9392 the NSS package to avoid lookups for UIDs in the range `pkg-config
9393 systemd --variable=dynamicuidmin` … `pkg-config systemd
9394 --variable=dynamicuidmax`, so that it does not consider itself
9395 authoritative for the same UID range systemd allocates dynamic users
9398 * The systemd-resolve tool has been renamed to resolvectl (it also
9399 remains available under the old name, for compatibility), and its
9400 interface is now verb-based, similar in style to the other <xyz>ctl
9401 tools, such as systemctl or loginctl.
9403 * The resolvectl/systemd-resolve tool also provides 'resolvconf'
9404 compatibility. It may be symlinked under the 'resolvconf' name, in
9405 which case it will take arguments and input compatible with the
9406 Debian and FreeBSD resolvconf tool.
9408 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
9409 where the system initially suspends, and after a timeout resumes and
9412 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
9413 set the client will only send a DUID as client identifier. (EDIT: the
9414 option was broken, and was dropped in v255.)
9416 * The nss-systemd glibc NSS module will now enumerate dynamic users and
9417 groups in effect. Previously, it could resolve UIDs/GIDs to user
9418 names/groups and vice versa, but did not support enumeration.
9420 * journald's Compress= configuration setting now optionally accepts a
9421 byte threshold value. All journal objects larger than this threshold
9422 will be compressed, smaller ones will not. Previously this threshold
9423 was not configurable and set to 512.
9425 * A new system.conf setting NoNewPrivileges= is now available which may
9426 be used to turn off acquisition of new privileges system-wide
9427 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
9428 for all its children). Note that turning this option on means setuid
9429 binaries and file system capabilities lose their special powers.
9430 While turning on this option is a big step towards a more secure
9431 system, doing so is likely to break numerous pre-existing UNIX tools,
9432 in particular su and sudo.
9434 * A new service systemd-time-sync-wait.service has been added. If
9435 enabled it will delay the time-sync.target unit at boot until time
9436 synchronization has been received from the network. This
9437 functionality is useful on systems lacking a local RTC or where it is
9438 acceptable that the boot process shall be delayed by external network
9441 * When hibernating, systemd will now inform the kernel of the image
9442 write offset, on kernels new enough to support this. This means swap
9443 files should work for hibernation now.
9445 * When loading unit files, systemd will now look for drop-in unit files
9446 extensions in additional places. Previously, for a unit file name
9447 "foo-bar-baz.service" it would look for dropin files in
9448 "foo-bar-baz.service.d/*.conf". Now, it will also look in
9449 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
9450 service name truncated after all inner dashes. This scheme allows
9451 writing drop-ins easily that apply to a whole set of unit files at
9452 once. It's particularly useful for mount and slice units (as their
9453 naming is prefix based), but is also useful for service and other
9454 units, for packages that install multiple unit files at once,
9455 following a strict naming regime of beginning the unit file name with
9456 the package's name. Two new specifiers are now supported in unit
9457 files to match this: %j and %J are replaced by the part of the unit
9458 name following the last dash.
9460 * Unit files and other configuration files that support specifier
9461 expansion now understand another three new specifiers: %T and %V will
9462 resolve to /tmp and /var/tmp respectively, or whatever temporary
9463 directory has been set for the calling user. %E will expand to either
9464 /etc (for system units) or $XDG_CONFIG_HOME (for user units).
9466 * The ExecStart= lines of unit files are no longer required to
9467 reference absolute paths. If non-absolute paths are specified the
9468 specified binary name is searched within the service manager's
9469 built-in $PATH, which may be queried with 'systemd-path
9470 search-binaries-default'. It's generally recommended to continue to
9471 use absolute paths for all binaries specified in unit files.
9473 * Units gained a new load state "bad-setting", which is used when a
9474 unit file was loaded, but contained fatal errors which prevent it
9475 from being started (for example, a service unit has been defined
9476 lacking both ExecStart= and ExecStop= lines).
9478 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
9479 support alternative debuggers, for example lldb. The old name
9480 continues to be available however, for compatibility reasons. Use the
9481 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
9482 to pick an alternative debugger instead of the default gdb.
9484 * systemctl and the other tools will now output escape sequences that
9485 generate proper clickable hyperlinks in various terminal emulators
9486 where useful (for example, in the "systemctl status" output you can
9487 now click on the unit file name to quickly open it in the
9488 editor/viewer of your choice). Note that not all terminal emulators
9489 support this functionality yet, but many do. Unfortunately, the
9490 "less" pager doesn't support this yet, hence this functionality is
9491 currently automatically turned off when a pager is started (which
9492 happens quite often due to auto-paging). We hope to remove this
9493 limitation as soon as "less" learns these escape sequences. This new
9494 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
9495 environment variable. For details on these escape sequences see:
9496 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
9498 * networkd's .network files now support a new IPv6MTUBytes= option for
9499 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
9500 option in the [Route] section to configure the MTU to use for
9501 specific routes. It also gained support for configuration of the DHCP
9502 "UserClass" option through the new UserClass= setting. It gained
9503 three new options in the new [CAN] section for configuring CAN
9504 networks. The MULTICAST and ALLMULTI interface flags may now be
9505 controlled explicitly with the new Multicast= and AllMulticast=
9508 * networkd will now automatically make use of the kernel's route
9509 expiration feature, if it is available.
9511 * udevd's .link files now support setting the number of receive and
9512 transmit channels, using the RxChannels=, TxChannels=,
9513 OtherChannels=, CombinedChannels= settings.
9515 * Support for UDPSegmentationOffload= has been removed, given its
9516 limited support in hardware, and waning software support.
9518 * networkd's .netdev files now support creating "netdevsim" interfaces.
9520 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
9521 to query the unit belonging to a specific kernel control group.
9523 * systemd-analyze gained a new verb "cat-config", which may be used to
9524 dump the contents of any configuration file, with all its matching
9525 drop-in files added in, and honouring the usual search and masking
9526 logic applied to systemd configuration files. For example use
9527 "systemd-analyze cat-config systemd/system.conf" to get the complete
9528 system configuration file of systemd how it would be loaded by PID 1
9529 itself. Similar to this, various tools such as systemd-tmpfiles or
9530 systemd-sysusers, gained a new option "--cat-config", which does the
9531 corresponding operation for their own configuration settings. For
9532 example, "systemd-tmpfiles --cat-config" will now output the full
9533 list of tmpfiles.d/ lines in place.
9535 * timedatectl gained three new verbs: "show" shows bus properties of
9536 systemd-timedated, "timesync-status" shows the current NTP
9537 synchronization state of systemd-timesyncd, and "show-timesync"
9538 shows bus properties of systemd-timesyncd.
9540 * systemd-timesyncd gained a bus interface on which it exposes details
9543 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
9544 understood by systemd-timedated. It takes a colon-separated list of
9545 unit names of NTP client services. The list is used by
9546 "timedatectl set-ntp".
9548 * systemd-nspawn gained a new --rlimit= switch for setting initial
9549 resource limits for the container payload. There's a new switch
9550 --hostname= to explicitly override the container's hostname. A new
9551 --no-new-privileges= switch may be used to control the
9552 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
9553 --oom-score-adjust= switch controls the OOM scoring adjustment value
9554 for the payload. The new --cpu-affinity= switch controls the CPU
9555 affinity of the container payload. The new --resolv-conf= switch
9556 allows more detailed control of /etc/resolv.conf handling of the
9557 container. Similarly, the new --timezone= switch allows more detailed
9558 control of /etc/localtime handling of the container.
9560 * systemd-detect-virt gained a new --list switch, which will print a
9561 list of all currently known VM and container environments.
9563 * Support for "Portable Services" has been added, see
9564 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
9565 experimental, but this is expected to change soon. Reflecting this
9566 experimental state, the "portablectl" binary is not installed into
9567 /usr/bin yet. The binary has to be called with the full path
9568 /usr/lib/systemd/portablectl instead.
9570 * journalctl's and systemctl's -o switch now knows a new log output
9571 mode "with-unit". The output it generates is very similar to the
9572 regular "short" mode, but displays the unit name instead of the
9573 syslog tag for each log line. Also, the date is shown with timezone
9574 information. This mode is probably more useful than the classic
9575 "short" output mode for most purposes, except where pixel-perfect
9576 compatibility with classic /var/log/messages formatting is required.
9578 * A new --dump-bus-properties switch has been added to the systemd
9579 binary, which may be used to dump all supported D-Bus properties.
9580 (Options which are still supported, but are deprecated, are *not*
9583 * sd-bus gained a set of new calls:
9584 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
9585 enable/disable the "floating" state of a bus slot object,
9586 i.e. whether the slot object pins the bus it is allocated for into
9587 memory or if the bus slot object gets disconnected when the bus goes
9588 away. sd_bus_open_with_description(),
9589 sd_bus_open_user_with_description(),
9590 sd_bus_open_system_with_description() may be used to allocate bus
9591 objects and set their description string already during allocation.
9593 * sd-event gained support for watching inotify events from the event
9594 loop, in an efficient way, sharing inotify handles between multiple
9595 users. For this a new function sd_event_add_inotify() has been added.
9597 * sd-event and sd-bus gained support for calling special user-supplied
9598 destructor functions for userdata pointers associated with
9599 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
9600 functions sd_bus_slot_set_destroy_callback,
9601 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
9602 sd_bus_track_get_destroy_callback,
9603 sd_event_source_set_destroy_callback,
9604 sd_event_source_get_destroy_callback have been added.
9606 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
9608 * PID 1 will now automatically reschedule .timer units whenever the
9609 local timezone changes. (They previously got rescheduled
9610 automatically when the system clock changed.)
9612 * New documentation has been added to document cgroups delegation,
9613 portable services and the various code quality tools we have set up:
9615 https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md
9616 https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md
9617 https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md
9619 * The Boot Loader Specification has been added to the source tree.
9621 https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md
9623 While moving it into our source tree we have updated it and further
9624 changes are now accepted through the usual github PR workflow.
9626 * pam_systemd will now look for PAM userdata fields systemd.memory_max,
9627 systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
9628 earlier PAM modules. The data in these fields is used to initialize
9629 the session scope's resource properties. Thus external PAM modules
9630 may now configure per-session limits, for example sourced from
9631 external user databases.
9633 * socket units with Accept=yes will now maintain a "refused" counter in
9634 addition to the existing "accepted" counter, counting connections
9635 refused due to the enforced limits.
9637 * The "systemd-path search-binaries-default" command may now be use to
9638 query the default, built-in $PATH PID 1 will pass to the services it
9641 * A new unit file setting PrivateMounts= has been added. It's a boolean
9642 option. If enabled the unit's processes are invoked in their own file
9643 system namespace. Note that this behaviour is also implied if any
9644 other file system namespacing options (such as PrivateTmp=,
9645 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
9646 primarily useful for services that do not use any of the other file
9647 system namespacing options. One such service is systemd-udevd.service
9648 where this is now used by default.
9650 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
9651 when the system is booted in UEFI "secure mode".
9653 * A new unit "system-update-pre.target" is added, which defines an
9654 optional synchronization point for offline system updates, as
9655 implemented by the pre-existing "system-update.target" unit. It
9656 allows ordering services before the service that executes the actual
9657 update process in a generic way.
9659 * Systemd now emits warnings whenever .include syntax is used.
9661 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
9662 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
9663 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
9664 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
9665 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
9666 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
9667 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
9668 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
9669 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
9670 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
9671 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
9672 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
9673 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
9674 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
9675 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
9676 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
9677 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
9678 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
9679 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
9680 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
9681 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
9682 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
9683 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
9684 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
9685 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
9686 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
9687 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
9688 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
9689 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
9691 — Berlin, 2018-06-22
9695 * The MemoryAccounting= unit property now defaults to on. After
9696 discussions with the upstream control group maintainers we learnt
9697 that the negative impact of cgroup memory accounting on current
9698 kernels is finally relatively minimal, so that it should be safe to
9699 enable this by default without affecting system performance. Besides
9700 memory accounting only task accounting is turned on by default, all
9701 other forms of resource accounting (CPU, IO, IP) remain off for now,
9702 because it's not clear yet that their impact is small enough to move
9703 from opt-in to opt-out. We recommend downstreams to leave memory
9704 accounting on by default if kernel 4.14 or higher is primarily
9705 used. On very resource constrained systems or when support for old
9706 kernels is a necessity, -Dmemory-accounting-default=false can be used
9707 to revert this change.
9709 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
9710 %udev_rules_update) and the journal catalog (%journal_catalog_update)
9711 from the upgrade scriptlets of individual packages now do nothing.
9712 Transfiletriggers have been added which will perform those updates
9713 once at the end of the transaction.
9715 Similar transfiletriggers have been added to execute any sysctl.d
9716 and binfmt.d rules. Thus, it should be unnecessary to provide any
9717 scriptlets to execute this configuration from package installation
9720 * systemd-sysusers gained a mode where the configuration to execute is
9721 specified on the command line, but this configuration is not executed
9722 directly, but instead it is merged with the configuration on disk,
9723 and the result is executed. This is useful for package installation
9724 scripts which want to create the user before installing any files on
9725 disk (in case some of those files are owned by that user), while
9726 still allowing local admin overrides.
9728 This functionality is exposed to rpm scriptlets through a new
9729 %sysusers_create_package macro. Old %sysusers_create and
9730 %sysusers_create_inline macros are deprecated.
9732 A transfiletrigger for sysusers.d configuration is now installed,
9733 which means that it should be unnecessary to call systemd-sysusers from
9734 package installation scripts, unless the package installs any files
9735 owned by those newly-created users, in which case
9736 %sysusers_create_package should be used.
9738 * Analogous change has been done for systemd-tmpfiles: it gained a mode
9739 where the command-line configuration is merged with the configuration
9740 on disk. This is exposed as the new %tmpfiles_create_package macro,
9741 and %tmpfiles_create is deprecated. A transfiletrigger is installed
9742 for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
9743 from package installation scripts.
9745 * sysusers.d configuration for a user may now also specify the group
9746 number, in addition to the user number ("u username 123:456"), or
9747 without the user number ("u username -:456").
9749 * Configution items for systemd-sysusers can now be specified as
9750 positional arguments when the new --inline switch is used.
9752 * The login shell of users created through sysusers.d may now be
9753 specified (previously, it was always /bin/sh for root and
9754 /sbin/nologin for other users).
9756 * systemd-analyze gained a new --global switch to look at global user
9757 configuration. It also gained a unit-paths verb to list the unit load
9758 paths that are compiled into systemd (which can be used with
9759 --systemd, --user, or --global).
9761 * udevadm trigger gained a new --settle/-w option to wait for any
9762 triggered events to finish (but just those, and not any other events
9763 which are triggered meanwhile).
9765 * The action that systemd-logind takes when the lid is closed and the
9766 machine is connected to external power can now be configured using
9767 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
9768 was determined by HandleLidSwitch=, and, for backwards compatibility,
9769 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
9771 * journalctl will periodically call sd_journal_process() to make it
9772 resilient against inotify queue overruns when journal files are
9773 rotated very quickly.
9775 * Two new functions in libsystemd — sd_bus_get_n_queued_read and
9776 sd_bus_get_n_queued_write — may be used to check the number of
9777 pending bus messages.
9779 * systemd gained a new
9780 org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
9781 which can be used to migrate foreign processes to scope and service
9782 units. The primary user for this new API is systemd itself: the
9783 systemd --user instance uses this call of the systemd --system
9784 instance to migrate processes if it itself gets the request to
9785 migrate processes and the kernel refuses this due to access
9786 restrictions. Thanks to this "systemd-run --scope --user …" works
9787 again in pure cgroup v2 environments when invoked from the user
9790 * A new TemporaryFileSystem= setting can be used to mask out part of
9791 the real file system tree with tmpfs mounts. This may be combined
9792 with BindPaths= and BindReadOnlyPaths= to hide files or directories
9793 not relevant to the unit, while still allowing some paths lower in
9794 the tree to be accessed.
9796 ProtectHome=tmpfs may now be used to hide user home and runtime
9797 directories from units, in a way that is mostly equivalent to
9798 "TemporaryFileSystem=/home /run/user /root".
9800 * Non-service units are now started with KeyringMode=shared by default.
9801 This means that mount and swapon and other mount tools have access
9802 to keys in the main keyring.
9804 * /sys/fs/bpf is now mounted automatically.
9806 * QNX virtualization is now detected by systemd-detect-virt and may
9807 be used in ConditionVirtualization=.
9809 * IPAccounting= may now be enabled also for slice units.
9811 * A new -Dsplit-bin= build configuration switch may be used to specify
9812 whether bin and sbin directories are merged, or if they should be
9813 included separately in $PATH and various listings of executable
9814 directories. The build configuration scripts will try to autodetect
9815 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
9816 system, but distributions are encouraged to configure this
9819 * A new -Dok-color= build configuration switch may be used to change
9820 the colour of "OK" status messages.
9822 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
9823 PrivateNetwork=yes was buggy in previous versions of systemd. This
9824 means that after the upgrade and daemon-reexec, any such units must
9827 * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
9828 will not exclude read-only files owned by root from cleanup.
9830 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
9831 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
9832 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
9833 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
9834 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
9835 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
9836 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
9837 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
9838 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
9839 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
9840 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
9841 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
9842 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
9843 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
9844 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
9845 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
9847 — Warsaw, 2018-03-05
9851 * Some keyboards come with a zoom see-saw or rocker which until now got
9852 mapped to the Linux "zoomin/out" keys in hwdb. However, these
9853 keycodes are not recognized by any major desktop. They now produce
9854 Up/Down key events so that they can be used for scrolling.
9856 * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
9857 slightly: previously, if an argument was specified for lines of this
9858 type (i.e. the right-most column was set) this string was appended to
9859 existing files each time systemd-tmpfiles was run. This behaviour was
9860 different from what the documentation said, and not particularly
9861 useful, as repeated systemd-tmpfiles invocations would not be
9862 idempotent and grow such files without bounds. With this release
9863 behaviour has been altered to match what the documentation says:
9864 lines of this type only have an effect if the indicated files don't
9865 exist yet, and only then the argument string is written to the file.
9867 * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
9868 systemd-tmpfiles behaviour: previously, read-only files owned by root
9869 were always excluded from the file "aging" algorithm (i.e. the
9870 automatic clean-up of directories like /tmp based on
9871 atime/mtime/ctime). We intend to drop this restriction, and age files
9872 by default even when owned by root and read-only. This behaviour was
9873 inherited from older tools, but there have been requests to remove
9874 it, and it's not obvious why this restriction was made in the first
9875 place. Please speak up now, if you are aware of software that requires
9876 this behaviour, otherwise we'll remove the restriction in v238.
9878 * A new environment variable $SYSTEMD_OFFLINE is now understood by
9879 systemctl. It takes a boolean argument. If on, systemctl assumes it
9880 operates on an "offline" OS tree, and will not attempt to talk to the
9881 service manager. Previously, this mode was implicitly enabled if a
9882 chroot() environment was detected, and this new environment variable
9883 now provides explicit control.
9885 * .path and .socket units may now be created transiently, too.
9886 Previously only service, mount, automount and timer units were
9887 supported as transient units. The systemd-run tool has been updated
9888 to expose this new functionality, you may hence use it now to bind
9889 arbitrary commands to path or socket activation on-the-fly from the
9890 command line. Moreover, almost all properties are now exposed for the
9891 unit types that already supported transient operation.
9893 * The systemd-mount command gained support for a new --owner= parameter
9894 which takes a user name, which is then resolved and included in uid=
9895 and gid= mount options string of the file system to mount.
9897 * A new unit condition ConditionControlGroupController= has been added
9898 that checks whether a specific cgroup controller is available.
9900 * Unit files, udev's .link files, and systemd-networkd's .netdev and
9901 .network files all gained support for a new condition
9902 ConditionKernelVersion= for checking against specific kernel
9905 * In systemd-networkd, the [IPVLAN] section in .netdev files gained
9906 support for configuring device flags in the Flags= setting. In the
9907 same files, the [Tunnel] section gained support for configuring
9908 AllowLocalRemote=. The [Route] section in .network files gained
9909 support for configuring InitialCongestionWindow=,
9910 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
9911 understands RapidCommit=.
9913 * systemd-networkd's DHCPv6 support gained support for Prefix
9916 * sd-bus gained support for a new "watch-bind" feature. When this
9917 feature is enabled, an sd_bus connection may be set up to connect to
9918 an AF_UNIX socket in the file system as soon as it is created. This
9919 functionality is useful for writing early-boot services that
9920 automatically connect to the system bus as soon as it is started,
9921 without ugly time-based polling. systemd-networkd and
9922 systemd-resolved have been updated to make use of this
9923 functionality. busctl exposes this functionality in a new
9924 --watch-bind= command line switch.
9926 * sd-bus will now optionally synthesize a local "Connected" signal as
9927 soon as a D-Bus connection is set up fully. This message mirrors the
9928 already existing "Disconnected" signal which is synthesized when the
9929 connection is terminated. This signal is generally useful but
9930 particularly handy in combination with the "watch-bind" feature
9931 described above. Synthesizing of this message has to be requested
9932 explicitly through the new API call sd_bus_set_connected_signal(). In
9933 addition a new call sd_bus_is_ready() has been added that checks
9934 whether a connection is fully set up (i.e. between the "Connected" and
9935 "Disconnected" signals).
9937 * sd-bus gained two new calls sd_bus_request_name_async() and
9938 sd_bus_release_name_async() for asynchronously registering bus
9939 names. Similar, there is now sd_bus_add_match_async() for installing
9940 a signal match asynchronously. All of systemd's own services have
9941 been updated to make use of these calls. Doing these operations
9942 asynchronously has two benefits: it reduces the risk of deadlocks in
9943 case of cyclic dependencies between bus services, and it speeds up
9944 service initialization since synchronization points for bus
9945 round-trips are removed.
9947 * sd-bus gained two new calls sd_bus_match_signal() and
9948 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
9949 and sd_bus_add_match_async() but instead of taking a D-Bus match
9950 string take match fields as normal function parameters.
9952 * sd-bus gained two new calls sd_bus_set_sender() and
9953 sd_bus_message_set_sender() for setting the sender name of outgoing
9954 messages (either for all outgoing messages or for just one specific
9955 one). These calls are only useful in direct connections as on
9956 brokered connections the broker fills in the sender anyway,
9957 overwriting whatever the client filled in.
9959 * sd-event gained a new pseudo-handle that may be specified on all API
9960 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
9961 used this refers to the default event loop object of the calling
9962 thread. Note however that this does not implicitly allocate one —
9963 which has to be done prior by using sd_event_default(). Similarly
9964 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
9965 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
9966 to the default bus of the specified type of the calling thread. Here
9967 too this does not implicitly allocate bus connection objects, this
9968 has to be done prior with sd_bus_default() and friends.
9970 * sd-event gained a new call pair
9971 sd_event_source_{get|set}_io_fd_own(). This may be used to request
9972 automatic closure of the file descriptor an IO event source watches
9973 when the event source is destroyed.
9975 * systemd-networkd gained support for natively configuring WireGuard
9978 * In previous versions systemd synthesized user records both for the
9979 "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
9980 internally. In order to simplify distribution-wide renames of the
9981 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
9982 new transitional flag file has been added: if
9983 /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
9984 user and group record within the systemd codebase is disabled.
9986 * systemd-notify gained a new --uid= option for selecting the source
9987 user/UID to use for notification messages sent to the service
9990 * journalctl gained a new --grep= option to list only entries in which
9991 the message matches a certain pattern. By default matching is case
9992 insensitive if the pattern is lowercase, and case sensitive
9993 otherwise. Option --case-sensitive=yes|no can be used to override
9994 this an specify case sensitivity or case insensitivity.
9996 * There's now a "systemd-analyze service-watchdogs" command for printing
9997 the current state of the service runtime watchdog, and optionally
9998 enabling or disabling the per-service watchdogs system-wide if given a
9999 boolean argument (i.e. the concept you configure in WatchdogSec=), for
10000 debugging purposes. There's also a kernel command line option
10001 systemd.service_watchdogs= for controlling the same.
10003 * Two new "log-level" and "log-target" options for systemd-analyze were
10004 added that merge the now deprecated get-log-level, set-log-level and
10005 get-log-target, set-log-target pairs. The deprecated options are still
10006 understood for backwards compatibility. The two new options print the
10007 current value when no arguments are given, and set them when a
10008 level/target is given as an argument.
10010 * sysusers.d's "u" lines now optionally accept both a UID and a GID
10011 specification, separated by a ":" character, in order to create users
10012 where UID and GID do not match.
10014 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
10015 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
10016 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
10017 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
10018 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
10019 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
10020 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
10021 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
10022 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
10023 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
10024 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
10025 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
10026 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
10027 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
10028 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
10029 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
10030 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
10031 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
10032 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
10033 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
10040 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
10041 in v235 has been extended to also set the dummy.ko module option
10042 numdummies=0, preventing the kernel from automatically creating
10043 dummy0. All dummy interfaces must now be explicitly created.
10045 * Unknown '%' specifiers in configuration files are now rejected. This
10046 applies to units and tmpfiles.d configuration. Any percent characters
10047 that are followed by a letter or digit that are not supposed to be
10048 interpreted as the beginning of a specifier should be escaped by
10049 doubling ("%%"). (So "size=5%" is still accepted, as well as
10050 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
10051 valid specifiers today.)
10053 * systemd-resolved now maintains a new dynamic
10054 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
10055 recommended to make /etc/resolv.conf a symlink to it. This file
10056 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
10057 includes dynamically acquired search domains, achieving more correct
10058 DNS resolution by software that bypasses local DNS APIs such as NSS.
10060 * The "uaccess" udev tag has been dropped from /dev/kvm and
10061 /dev/dri/renderD*. These devices now have the 0666 permissions by
10062 default (but this may be changed at build-time). /dev/dri/renderD*
10063 will now be owned by the "render" group along with /dev/kfd.
10065 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
10066 systemd-journal-gatewayd.service and
10067 systemd-journal-upload.service. This means "nss-systemd" must be
10068 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
10069 services are resolved properly.
10071 * In /etc/fstab two new mount options are now understood:
10072 x-systemd.makefs and x-systemd.growfs. The former has the effect that
10073 the configured file system is formatted before it is mounted, the
10074 latter that the file system is resized to the full block device size
10075 after it is mounted (i.e. if the file system is smaller than the
10076 partition it resides on, it's grown). This is similar to the fsck
10077 logic in /etc/fstab, and pulls in systemd-makefs@.service and
10078 systemd-growfs@.service as necessary, similar to
10079 systemd-fsck@.service. Resizing is currently only supported on ext4
10082 * In systemd-networkd, the IPv6 RA logic now optionally may announce
10083 DNS server and domain information.
10085 * Support for the LUKS2 on-disk format for encrypted partitions has
10086 been added. This requires libcryptsetup2 during compilation and
10089 * The systemd --user instance will now signal "readiness" when its
10090 basic.target unit has been reached, instead of when the run queue ran
10091 empty for the first time.
10093 * Tmpfiles.d with user configuration are now also supported.
10094 systemd-tmpfiles gained a new --user switch, and snippets placed in
10095 ~/.config/user-tmpfiles.d/ and corresponding directories will be
10096 executed by systemd-tmpfiles --user running in the new
10097 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
10098 running in the user session.
10100 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
10101 %S resolves to the top-level state directory (/var/lib for the system
10102 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
10103 top-level cache directory (/var/cache for the system instance,
10104 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
10105 logs directory (/var/log for the system instance,
10106 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
10107 existing %t specifier, that resolves to the top-level runtime
10108 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
10111 * journalctl learnt a new parameter --output-fields= for limiting the
10112 set of journal fields to output in verbose and JSON output modes.
10114 * systemd-timesyncd's configuration file gained a new option
10115 RootDistanceMaxSec= for setting the maximum root distance of servers
10116 it'll use, as well as the new options PollIntervalMinSec= and
10117 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
10119 * bootctl gained a new command "list" for listing all available boot
10120 menu items on systems that follow the boot loader specification.
10122 * systemctl gained a new --dry-run switch that shows what would be done
10123 instead of doing it, and is currently supported by the shutdown and
10126 * ConditionSecurity= can now detect the TOMOYO security module.
10128 * Unit file [Install] sections are now also respected in unit drop-in
10129 files. This is intended to be used by drop-ins under /usr/lib/.
10131 * systemd-firstboot may now also set the initial keyboard mapping.
10133 * Udev "changed" events for devices which are exposed as systemd
10134 .device units are now propagated to units specified in
10135 ReloadPropagatedFrom= as reload requests.
10137 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
10138 unit template name (i.e. a name in the form of 'foobar@.service',
10139 without the instance component between the '@' and - the '.'), then
10140 the escaped sysfs path of the device is automatically used as the
10143 * SystemCallFilter= in unit files has been extended so that an "errno"
10144 can be specified individually for each system call. Example:
10145 SystemCallFilter=~uname:EILSEQ.
10147 * The cgroup delegation logic has been substantially updated. Delegate=
10148 now optionally takes a list of controllers (instead of a boolean, as
10149 before), which lists the controllers to delegate at least.
10151 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
10153 * A new LogLevelMax= setting configures the maximum log level any
10154 process of the service may log at (i.e. anything with a lesser
10155 priority than what is specified is automatically dropped). A new
10156 LogExtraFields= setting allows configuration of additional journal
10157 fields to attach to all log records generated by any of the unit's
10160 * New StandardInputData= and StandardInputText= settings along with the
10161 new option StandardInput=data may be used to configure textual or
10162 binary data that shall be passed to the executed service process via
10163 standard input, encoded in-line in the unit file.
10165 * StandardInput=, StandardOutput= and StandardError= may now be used to
10166 connect stdin/stdout/stderr of executed processes directly with a
10167 file or AF_UNIX socket in the file system, using the new "file:" option.
10169 * A new unit file option CollectMode= has been added, that allows
10170 tweaking the garbage collection logic for units. It may be used to
10171 tell systemd to garbage collect units that have failed automatically
10172 (normally it only GCs units that exited successfully). systemd-run
10173 and systemd-mount expose this new functionality with a new -G option.
10175 * "machinectl bind" may now be used to bind mount non-directories
10176 (i.e. regularfiles, devices, fifos, sockets).
10178 * systemd-analyze gained a new verb "calendar" for validating and
10179 testing calendar time specifications to use for OnCalendar= in timer
10180 units. Besides validating the expression it will calculate the next
10181 time the specified expression would elapse.
10183 * In addition to the pre-existing FailureAction= unit file setting
10184 there's now SuccessAction=, for configuring a shutdown action to
10185 execute when a unit completes successfully. This is useful in
10186 particular inside containers that shall terminate after some workload
10187 has been completed. Also, both options are now supported for all unit
10188 types, not just services.
10190 * networkds's IP rule support gained two new options
10191 IncomingInterface= and OutgoingInterface= for configuring the incoming
10192 and outgoing interfaces of configured rules. systemd-networkd also
10193 gained support for "vxcan" network devices.
10195 * networkd gained a new setting RequiredForOnline=, taking a
10196 boolean. If set, systemd-wait-online will take it into consideration
10197 when determining that the system is up, otherwise it will ignore the
10198 interface for this purpose.
10200 * The sd_notify() protocol gained support for a new operation: with
10201 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
10202 store again, ahead of POLLHUP or POLLERR when they are removed
10205 * A new document doc/UIDS-GIDS.md has been added to the source tree,
10206 that documents the UID/GID range and assignment assumptions and
10207 requirements of systemd.
10209 * The watchdog device PID 1 will ping may now be configured through the
10210 WatchdogDevice= configuration file setting, or by setting the
10211 systemd.watchdog_service= kernel command line option.
10213 * systemd-resolved's gained support for registering DNS-SD services on
10214 the local network using MulticastDNS. Services may either be
10215 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
10216 the same dir below /run, /usr/lib), or through its D-Bus API.
10218 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
10219 extend the effective start, runtime, and stop time. The service must
10220 continue to send EXTEND_TIMEOUT_USEC within the period specified to
10221 prevent the service manager from making the service as timedout.
10223 * systemd-resolved's DNSSEC support gained support for RFC 8080
10224 (Ed25519 keys and signatures).
10226 * The systemd-resolve command line tool gained a new set of options
10227 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
10228 --set-nta= and --revert to configure per-interface DNS configuration
10229 dynamically during runtime. It's useful for pushing DNS information
10230 into systemd-resolved from DNS hook scripts that various interface
10231 managing software supports (such as pppd).
10233 * systemd-nspawn gained a new --network-namespace-path= command line
10234 option, which may be used to make a container join an existing
10235 network namespace, by specifying a path to a "netns" file.
10237 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
10238 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
10239 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
10240 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
10241 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
10242 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
10243 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
10244 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
10245 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
10246 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
10247 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
10248 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
10249 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
10250 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
10251 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
10252 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
10253 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
10254 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
10255 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
10256 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
10257 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
10258 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
10259 Jędrzejewski-Szmek, Zeal Jagannatha
10261 — Berlin, 2017-12-14
10265 * INCOMPATIBILITY: systemd-logind.service and other long-running
10266 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
10267 communication with the outside. This generally improves security of
10268 the system, and is in almost all cases a safe and good choice, as
10269 these services do not and should not provide any network-facing
10270 functionality. However, systemd-logind uses the glibc NSS API to
10271 query the user database. This creates problems on systems where NSS
10272 is set up to directly consult network services for user database
10273 lookups. In particular, this creates incompatibilities with the
10274 "nss-nis" module, which attempts to directly contact the NIS/YP
10275 network servers it is configured for, and will now consistently
10276 fail. In such cases, it is possible to turn off IP sandboxing for
10277 systemd-logind.service (set IPAddressDeny= in its [Service] section
10278 to the empty string, via a .d/ unit file drop-in). Downstream
10279 distributions might want to update their nss-nis packaging to include
10280 such a drop-in snippet, accordingly, to hide this incompatibility
10281 from the user. Another option is to make use of glibc's nscd service
10282 to proxy such network requests through a privilege-separated, minimal
10283 local caching daemon, or to switch to more modern technologies such
10284 sssd, whose NSS hook-ups generally do not involve direct network
10285 access. In general, we think it's definitely time to question the
10286 implementation choices of nss-nis, i.e. whether it's a good idea
10287 today to embed a network-facing loadable module into all local
10288 processes that need to query the user database, including the most
10289 trivial and benign ones, such as "ls". For more details about
10290 IPAddressDeny= see below.
10292 * A new modprobe.d drop-in is now shipped by default that sets the
10293 bonding module option max_bonds=0. This overrides the kernel default,
10294 to avoid conflicts and ambiguity as to whether or not bond0 should be
10295 managed by systemd-networkd or not. This resolves multiple issues
10296 with bond0 properties not being applied, when bond0 is configured
10297 with systemd-networkd. Distributors may choose to not package this,
10298 however in that case users will be prevented from correctly managing
10299 bond0 interface using systemd-networkd.
10301 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
10302 which print the logging level and target of the system manager. They
10303 complement the existing "set-log-level" and "set-log-target" verbs
10304 used to change those values.
10306 * journald.conf gained a new boolean setting ReadKMsg= which defaults
10307 to on. If turned off kernel log messages will not be read by
10308 systemd-journald or included in the logs. It also gained a new
10309 setting LineMax= for configuring the maximum line length in
10310 STDOUT/STDERR log streams. The new default for this value is 48K, up
10311 from the previous hardcoded 2048.
10313 * A new unit setting RuntimeDirectoryPreserve= has been added, which
10314 allows more detailed control of what to do with a runtime directory
10315 configured with RuntimeDirectory= (i.e. a directory below /run or
10316 $XDG_RUNTIME_DIR) after a unit is stopped.
10318 * The RuntimeDirectory= setting for units gained support for creating
10319 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
10320 one top-level directory.
10322 * Units gained new options StateDirectory=, CacheDirectory=,
10323 LogsDirectory= and ConfigurationDirectory= which are closely related
10324 to RuntimeDirectory= but manage per-service directories below
10325 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
10326 possible to write unit files which when activated automatically gain
10327 properly owned service specific directories in these locations, thus
10328 making unit files self-contained and increasing compatibility with
10329 stateless systems and factory reset where /etc or /var are
10330 unpopulated at boot. Matching these new settings there's also
10331 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
10332 ConfigurationDirectoryMode= for configuring the access mode of these
10333 directories. These settings are particularly useful in combination
10334 with DynamicUser=yes as they provide secure, properly-owned,
10335 writable, and stateful locations for storage, excluded from the
10336 sandbox that such services live in otherwise.
10338 * Automake support has been removed from this release. systemd is now
10341 * systemd-journald will now aggressively cache client metadata during
10342 runtime, speeding up log write performance under pressure. This comes
10343 at a small price though: as much of the metadata is read
10344 asynchronously from /proc/ (and isn't implicitly attached to log
10345 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
10346 metadata stored alongside a log entry might be slightly
10347 out-of-date. Previously it could only be slightly newer than the log
10348 message. The time window is small however, and given that the kernel
10349 is unlikely to be improved anytime soon in this regard, this appears
10352 * nss-myhostname/systemd-resolved will now by default synthesize an
10353 A/AAAA resource record for the "_gateway" hostname, pointing to the
10354 current default IP gateway. Previously it did that for the "gateway"
10355 name, hampering adoption, as some distributions wanted to leave that
10356 hostname open for local use. The old behaviour may still be
10357 requested at build time.
10359 * systemd-networkd's [Address] section in .network files gained a new
10360 Scope= setting for configuring the IP address scope. The [Network]
10361 section gained a new boolean setting ConfigureWithoutCarrier= that
10362 tells systemd-networkd to ignore link sensing when configuring the
10363 device. The [DHCP] section gained a new Anonymize= boolean option for
10364 turning on a number of options suggested in RFC 7844. A new
10365 [RoutingPolicyRule] section has been added for configuring the IP
10366 routing policy. The [Route] section has gained support for a new
10367 Type= setting which permits configuring
10368 blackhole/unreachable/prohibit routes.
10370 * The [VRF] section in .netdev files gained a new Table= setting for
10371 configuring the routing table to use. The [Tunnel] section gained a
10372 new Independent= boolean field for configuring tunnels independent of
10373 an underlying network interface. The [Bridge] section gained a new
10374 GroupForwardMask= option for configuration of propagation of link
10375 local frames between bridge ports.
10377 * The WakeOnLan= setting in .link files gained support for a number of
10378 new modes. A new TCP6SegmentationOffload= setting has been added for
10379 configuring TCP/IPv6 hardware segmentation offload.
10381 * The IPv6 RA sender implementation may now optionally send out RDNSS
10382 and RDNSSL records to supply DNS configuration to peers.
10384 * systemd-nspawn gained support for a new --system-call-filter= command
10385 line option for adding and removing entries in the default system
10386 call filter it applies. Moreover systemd-nspawn has been changed to
10387 implement a system call allow list instead of a deny list.
10389 * systemd-run gained support for a new --pipe command line option. If
10390 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
10391 are directly passed on to the activated transient service
10392 executable. This allows invoking arbitrary processes as systemd
10393 services (for example to take benefit of dependency management,
10394 accounting management, resource management or log management that is
10395 done automatically for services) — while still allowing them to be
10396 integrated in a classic UNIX shell pipeline.
10398 * When a service sends RELOAD=1 via sd_notify() and reload propagation
10399 using ReloadPropagationTo= is configured, a reload is now propagated
10400 to configured units. (Previously this was only done on explicitly
10401 requested reloads, using "systemctl reload" or an equivalent
10404 * For each service unit a restart counter is now kept: it is increased
10405 each time the service is restarted due to Restart=, and may be
10406 queried using "systemctl show -p NRestarts …".
10408 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
10409 @signal and @timer have been added, for usage with SystemCallFilter=
10410 in unit files and the new --system-call-filter= command line option
10411 of systemd-nspawn (see above).
10413 * ExecStart= lines in unit files gained two new modifiers: when a
10414 command line is prefixed with "!" the command will be executed as
10415 configured, except for the credentials applied by
10416 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
10417 "+", but does still apply namespacing options unlike "+". There's
10418 also "!!" now, which is mostly identical, but becomes a NOP on
10419 systems that support ambient capabilities. This is useful to write
10420 unit files that work with ambient capabilities where possible but
10421 automatically fall back to traditional privilege dropping mechanisms
10422 on systems where this is not supported.
10424 * ListenNetlink= settings in socket units now support RDMA netlink
10427 * A new unit file setting LockPersonality= has been added which permits
10428 locking down the chosen execution domain ("personality") of a service
10431 * A new special target "getty-pre.target" has been added, which is
10432 ordered before all text logins, and may be used to order services
10433 before textual logins acquire access to the console.
10435 * systemd will now attempt to load the virtio-rng.ko kernel module very
10436 early on if a VM environment supporting this is detected. This should
10437 improve entropy during early boot in virtualized environments.
10439 * A _netdev option is now supported in /etc/crypttab that operates in a
10440 similar way as the same option in /etc/fstab: it permits configuring
10441 encrypted devices that need to be ordered after the network is up.
10442 Following this logic, two new special targets
10443 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
10444 added that are to cryptsetup.target what remote-fs.target and
10445 remote-fs-pre.target are to local-fs.target.
10447 * Service units gained a new UnsetEnvironment= setting which permits
10448 unsetting specific environment variables for services that are
10449 normally passed to it (for example in order to mask out locale
10450 settings for specific services that can't deal with it).
10452 * Units acquired a new boolean option IPAccounting=. When turned on, IP
10453 traffic accounting (packet count as well as byte count) is done for
10454 the service, and shown as part of "systemctl status" or "systemd-run
10457 * Service units acquired two new options IPAddressAllow= and
10458 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
10459 for configuring a simple IP access control list for all sockets of
10460 the unit. These options are available also on .slice and .socket
10461 units, permitting flexible access list configuration for individual
10462 services as well as groups of services (as defined by a slice unit),
10463 including system-wide. Note that IP ACLs configured this way are
10464 enforced on every single IPv4 and IPv6 socket created by any process
10465 of the service unit, and apply to ingress as well as egress traffic.
10467 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
10468 structured log message is generated each time the unit is stopped,
10469 containing information about the consumed resources of this
10472 * A new setting KeyringMode= has been added to unit files, which may be
10473 used to control how the kernel keyring is set up for executed
10476 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
10477 "systemctl kexec" and "systemctl exit" are now always asynchronous in
10478 behaviour (that is: these commands return immediately after the
10479 operation was enqueued instead of waiting for the operation to
10480 complete). Previously, "systemctl poweroff" and "systemctl reboot"
10481 were asynchronous on systems using systemd-logind (i.e. almost
10482 always, and like they were on sysvinit), and the other three commands
10483 were unconditionally synchronous. With this release this is cleaned
10484 up, and callers will see the same asynchronous behaviour on all
10485 systems for all five operations.
10487 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
10490 * .timer units now accept calendar specifications in other timezones
10491 than UTC or the local timezone.
10493 * The tmpfiles snippet var.conf has been changed to create
10494 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
10495 the "utmp" group already, and it appears to be generally understood
10496 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
10497 databases. Previously this was implemented correctly for all these
10498 databases excepts btmp, which has been opened up like this now
10499 too. Note that while the other databases are world-readable
10500 (i.e. 0644), btmp is not and remains more restrictive.
10502 * The systemd-resolve tool gained a new --reset-server-features
10503 switch. When invoked like this systemd-resolved will forget
10504 everything it learnt about the features supported by the configured
10505 upstream DNS servers, and restarts the feature probing logic on the
10506 next resolver look-up for them at the highest feature level
10509 * The status dump systemd-resolved sends to the logs upon receiving
10510 SIGUSR1 now also includes information about all DNS servers it is
10511 configured to use, and the features levels it probed for them.
10513 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
10514 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
10515 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
10516 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
10517 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
10518 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
10519 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
10520 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
10521 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
10522 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
10523 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
10524 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
10525 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
10526 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
10527 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
10528 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
10529 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
10530 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
10531 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
10532 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
10534 — Berlin, 2017-10-06
10538 * Meson is now supported as build system in addition to Automake. It is
10539 our plan to remove Automake in one of our next releases, so that
10540 Meson becomes our exclusive build system. Hence, please start using
10541 the Meson build system in your downstream packaging. There's plenty
10542 of documentation around how to use Meson, the extremely brief
10545 ./autogen.sh && ./configure && make && sudo make install
10549 meson build && ninja -C build && sudo ninja -C build install
10551 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
10552 which permits configuring a timeout on the time a job is
10553 running. This is particularly useful for setting timeouts on jobs for
10556 * Unit files gained two new options ConditionUser= and ConditionGroup=
10557 for conditionalizing units based on the identity of the user/group
10558 running a systemd user instance.
10560 * systemd-networkd now understands a new FlowLabel= setting in the
10561 [VXLAN] section of .network files, as well as a Priority= in
10562 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
10563 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
10564 gained support for configuration of GENEVE links, and IPv6 address
10565 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
10567 * .link files now understand a new Port= setting.
10569 * systemd-networkd's DHCP support gained support for DHCP option 119
10570 (domain search list).
10572 * systemd-networkd gained support for serving IPv6 address ranges using
10573 the Router Advertisement protocol. The new .network configuration
10574 section [IPv6Prefix] may be used to configure the ranges to
10575 serve. This is implemented based on a new, minimal, native server
10576 implementation of RA.
10578 * journalctl's --output= switch gained support for a new parameter
10579 "short-iso-precise" for a mode where timestamps are shown as precise
10582 * systemd-udevd's "net_id" builtin may now generate stable network
10583 interface names from IBM PowerVM VIO devices as well as ACPI platform
10586 * MulticastDNS support in systemd-resolved may now be explicitly
10587 enabled/disabled using the new MulticastDNS= configuration file
10590 * systemd-resolved may now optionally use libidn2 instead of the libidn
10591 for processing internationalized domain names. Support for libidn2
10592 should be considered experimental and should not be enabled by
10595 * "machinectl pull-tar" and related call may now do verification of
10596 downloaded images using SUSE-style .sha256 checksum files in addition
10597 to the already existing support for validating using Ubuntu-style
10600 * sd-bus gained support for a new sd_bus_message_appendv() call which
10601 is va_list equivalent of sd_bus_message_append().
10603 * sd-boot gained support for validating images using SHIM/MOK.
10605 * The SMACK code learnt support for "onlycap".
10607 * systemd-mount --umount is now much smarter in figuring out how to
10608 properly unmount a device given its mount or device path.
10610 * The code to call libnss_dns as a fallback from libnss_resolve when
10611 the communication with systemd-resolved fails was removed. This
10612 fallback was redundant and interfered with the [!UNAVAIL=return]
10613 suffix. See nss-resolve(8) for the recommended configuration.
10615 * systemd-logind may now be restarted without losing state. It stores
10616 the file descriptors for devices it manages in the system manager
10617 using the FDSTORE= mechanism. Please note that further changes in
10618 other components may be required to make use of this (for example
10619 Xorg has code to listen for stops of systemd-logind and terminate
10620 itself when logind is stopped or restarted, in order to avoid using
10621 stale file descriptors for graphical devices, which is now
10622 counterproductive and must be reverted in order for restarts of
10623 systemd-logind to be safe. See
10624 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
10626 * All kernel-install plugins are called with the environment variable
10627 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
10628 /etc/machine-id. If the machine ID could not be determined,
10629 $KERNEL_INSTALL_MACHINE_ID will be empty. Plugins should not put
10630 anything in the entry directory (passed as the second argument) if
10631 $KERNEL_INSTALL_MACHINE_ID is empty. For backwards compatibility, a
10632 temporary directory is passed as the entry directory and removed
10633 after all the plugins exit.
10635 * If KERNEL_INSTALL_MACHINE_ID is set in /etc/machine-info, kernel-install
10636 will now use its value as the machine ID instead of the machine ID
10637 from /etc/machine-id. If KERNEL_INSTALL_MACHINE_ID isn't set in
10638 /etc/machine-info and no machine ID is set in /etc/machine-id,
10639 kernel-install will try to store the current machine ID there as
10640 KERNEL_INSTALL_MACHINE_ID. If there is no machine ID, kernel-install
10641 will generate a new UUID, store it in /etc/machine-info as
10642 KERNEL_INSTALL_MACHINE_ID and use it as the machine ID.
10644 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
10645 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
10646 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
10647 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
10648 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
10649 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
10650 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
10651 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
10652 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
10653 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
10654 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
10655 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
10656 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
10657 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
10658 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
10659 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
10660 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
10661 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
10662 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
10663 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
10664 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
10665 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
10666 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
10667 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
10668 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
10669 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
10670 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
10671 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
10672 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
10675 — Berlin, 2017-07-12
10679 * The "hybrid" control group mode has been modified to improve
10680 compatibility with "legacy" cgroups-v1 setups. Specifically, the
10681 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
10682 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
10683 cgroups-v1 hierarchy), the only externally visible change being that
10684 the cgroups-v2 hierarchy is also mounted, to
10685 /sys/fs/cgroup/unified. This should provide a large degree of
10686 compatibility with "legacy" cgroups-v1, while taking benefit of the
10687 better management capabilities of cgroups-v2.
10689 * The default control group setup mode may be selected both a boot-time
10690 via a set of kernel command line parameters (specifically:
10691 systemd.unified_cgroup_hierarchy= and
10692 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
10693 default selected on the configure command line
10694 (--with-default-hierarchy=). The upstream default is "hybrid"
10695 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
10696 this will change in a future systemd version to be "unified" (pure
10697 cgroups-v2 mode). The third option for the compile time option is
10698 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
10699 distributions to default to "hybrid" mode for release distributions,
10700 starting with v233. We recommend "unified" for development
10701 distributions (specifically: distributions such as Fedora's rawhide)
10702 as that's where things are headed in the long run. Use "legacy" for
10703 greatest stability and compatibility only.
10705 * Note one current limitation of "unified" and "hybrid" control group
10706 setup modes: the kernel currently does not permit the systemd --user
10707 instance (i.e. unprivileged code) to migrate processes between two
10708 disconnected cgroup subtrees, even if both are managed and owned by
10709 the user. This effectively means "systemd-run --user --scope" doesn't
10710 work when invoked from outside of any "systemd --user" service or
10711 scope. Specifically, it is not supported from session scopes. We are
10712 working on fixing this in a future systemd version. (See #3388 for
10713 further details about this.)
10715 * DBus policy files are now installed into /usr rather than /etc. Make
10716 sure your system has dbus >= 1.9.18 running before upgrading to this
10717 version, or override the install path with --with-dbuspolicydir= .
10719 * All python scripts shipped with systemd (specifically: the various
10720 tests written in Python) now require Python 3.
10722 * systemd unit tests can now run standalone (without the source or
10723 build directories), and can be installed into /usr/lib/systemd/tests/
10724 with 'make install-tests'.
10726 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
10727 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
10730 * Support for the %c, %r, %R specifiers in unit files has been
10731 removed. Specifiers are not supposed to be dependent on configuration
10732 in the unit file itself (so that they resolve the same regardless
10733 where used in the unit files), but these specifiers were influenced
10734 by the Slice= option.
10736 * The shell invoked by debug-shell.service now defaults to /bin/sh in
10737 all cases. If distributions want to use a different shell for this
10738 purpose (for example Fedora's /sbin/sushell) they need to specify
10739 this explicitly at configure time using --with-debug-shell=.
10741 * The confirmation spawn prompt has been reworked to offer the
10744 (c)ontinue, proceed without asking anymore
10745 (D)ump, show the state of the unit
10746 (f)ail, don't execute the command and pretend it failed
10748 (i)nfo, show a short summary of the unit
10749 (j)obs, show jobs that are in progress
10750 (s)kip, don't execute the command and pretend it succeeded
10751 (y)es, execute the command
10753 The 'n' choice for the confirmation spawn prompt has been removed,
10754 because its meaning was confusing.
10756 The prompt may now also be redirected to an alternative console by
10757 specifying the console as parameter to systemd.confirm_spawn=.
10759 * Services of Type=notify require a READY=1 notification to be sent
10760 during startup. If no such message is sent, the service now fails,
10761 even if the main process exited with a successful exit code.
10763 * Services that fail to start up correctly now always have their
10764 ExecStopPost= commands executed. Previously, they'd enter "failed"
10765 state directly, without executing these commands.
10767 * The option MulticastDNS= of network configuration files has acquired
10768 an actual implementation. With MulticastDNS=yes a host can resolve
10769 names of remote hosts and reply to mDNS A and AAAA requests.
10771 * When units are about to be started an additional check is now done to
10772 ensure that all dependencies of type BindsTo= (when used in
10773 combination with After=) have been started.
10775 * systemd-analyze gained a new verb "syscall-filter" which shows which
10776 system call groups are defined for the SystemCallFilter= unit file
10777 setting, and which system calls they contain.
10779 * A new system call filter group "@filesystem" has been added,
10780 consisting of various file system related system calls. Group
10781 "@reboot" has been added, covering reboot, kexec and shutdown related
10782 calls. Finally, group "@swap" has been added covering swap
10783 configuration related calls.
10785 * A new unit file option RestrictNamespaces= has been added that may be
10786 used to restrict access to the various process namespace types the
10787 Linux kernel provides. Specifically, it may be used to take away the
10788 right for a service unit to create additional file system, network,
10789 user, and other namespaces. This sandboxing option is particularly
10790 relevant due to the high amount of recently discovered namespacing
10791 related vulnerabilities in the kernel.
10793 * systemd-udev's .link files gained support for a new AutoNegotiation=
10794 setting for configuring Ethernet auto-negotiation.
10796 * systemd-networkd's .network files gained support for a new
10797 ListenPort= setting in the [DHCP] section to explicitly configure the
10798 UDP client port the DHCP client shall listen on.
10800 * .network files gained a new Unmanaged= boolean setting for explicitly
10801 excluding one or more interfaces from management by systemd-networkd.
10803 * The systemd-networkd ProxyARP= option has been renamed to
10804 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
10805 renamed to ReduceARPProxy=. The old names continue to be available
10808 * systemd-networkd gained support for configuring IPv6 Proxy NDP
10809 addresses via the new IPv6ProxyNDPAddress= .network file setting.
10811 * systemd-networkd's bonding device support gained support for two new
10812 configuration options ActiveSlave= and PrimarySlave=.
10814 * The various options in the [Match] section of .network files gained
10815 support for negative matching.
10817 * New systemd-specific mount options are now understood in /etc/fstab:
10819 x-systemd.mount-timeout= may be used to configure the maximum
10820 permitted runtime of the mount command.
10822 x-systemd.device-bound may be set to bind a mount point to its
10823 backing device unit, in order to automatically remove a mount point
10824 if its backing device is unplugged. This option may also be
10825 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
10826 on the block device, which is now automatically set for all CDROM
10827 drives, so that mounted CDs are automatically unmounted when they are
10828 removed from the drive.
10830 x-systemd.after= and x-systemd.before= may be used to explicitly
10831 order a mount after or before another unit or mount point.
10833 * Enqueued start jobs for device units are now automatically garbage
10834 collected if there are no jobs waiting for them anymore.
10836 * systemctl list-jobs gained two new switches: with --after, for every
10837 queued job the jobs it's waiting for are shown; with --before the
10838 jobs which it's blocking are shown.
10840 * systemd-nspawn gained support for ephemeral boots from disk images
10841 (or in other words: --ephemeral and --image= may now be
10842 combined). Moreover, ephemeral boots are now supported for normal
10843 directories, even if the backing file system is not btrfs. Of course,
10844 if the file system does not support file system snapshots or
10845 reflinks, the initial copy operation will be relatively expensive, but
10846 this should still be suitable for many use cases.
10848 * Calendar time specifications in .timer units now support
10849 specifications relative to the end of a month by using "~" instead of
10850 "-" as separator between month and day. For example, "*-02~03" means
10851 "the third last day in February". In addition a new syntax for
10852 repeated events has been added using the "/" character. For example,
10853 "9..17/2:00" means "every two hours from 9am to 5pm".
10855 * systemd-socket-proxyd gained a new parameter --connections-max= for
10856 configuring the maximum number of concurrent connections.
10858 * sd-id128 gained a new API for generating unique IDs for the host in a
10859 way that does not leak the machine ID. Specifically,
10860 sd_id128_get_machine_app_specific() derives an ID based on the
10861 machine ID in a well-defined, non-reversible, stable way. This is
10862 useful whenever an identifier for the host is needed but where the
10863 identifier shall not be useful to identify the system beyond the
10864 scope of the application itself. (Internally this uses HMAC-SHA256 as
10865 keyed hash function using the machine ID as input.)
10867 * NotifyAccess= gained a new supported value "exec". When set
10868 notifications are accepted from all processes systemd itself invoked,
10869 including all control processes.
10871 * .nspawn files gained support for defining overlay mounts using the
10872 Overlay= and OverlayReadOnly= options. Previously this functionality
10873 was only available on the systemd-nspawn command line.
10875 * systemd-nspawn's --bind= and --overlay= options gained support for
10876 bind/overlay mounts whose source lies within the container tree by
10877 prefixing the source path with "+".
10879 * systemd-nspawn's --bind= and --overlay= options gained support for
10880 automatically allocating a temporary source directory in /var/tmp
10881 that is removed when the container dies. Specifically, if the source
10882 directory is specified as empty string this mechanism is selected. An
10883 example usage is --overlay=+/var::/var, which creates an overlay
10884 mount based on the original /var contained in the image, overlaid
10885 with a temporary directory in the host's /var/tmp. This way changes
10886 to /var are automatically flushed when the container shuts down.
10888 * systemd-nspawn --image= option does now permit raw file system block
10889 devices (in addition to images containing partition tables, as
10892 * The disk image dissection logic in systemd-nspawn gained support for
10893 automatically setting up LUKS encrypted as well as Verity protected
10894 partitions. When a container is booted from an encrypted image the
10895 passphrase is queried at start-up time. When a container with Verity
10896 data is started, the root hash is search in a ".roothash" file
10897 accompanying the disk image (alternatively, pass the root hash via
10898 the new --root-hash= command line option).
10900 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
10901 be used to dissect disk images the same way as systemd-nspawn does
10902 it, following the Bootable Partition Specification. It may even be
10903 used to mount disk images with complex partition setups (including
10904 LUKS and Verity partitions) to a local host directory, in order to
10905 inspect them. This tool is not considered public API (yet), and is
10906 thus not installed into /usr/bin. Please do not rely on its
10907 existence, since it might go away or be changed in later systemd
10910 * A new generator "systemd-verity-generator" has been added, similar in
10911 style to "systemd-cryptsetup-generator", permitting automatic setup of
10912 Verity root partitions when systemd boots up. In order to make use of
10913 this your partition setup should follow the Discoverable Partitions
10914 Specification, and the GPT partition ID of the root file system
10915 partition should be identical to the upper 128-bit of the Verity root
10916 hash. The GPT partition ID of the Verity partition protecting it
10917 should be the lower 128-bit of the Verity root hash. If the partition
10918 image follows this model it is sufficient to specify a single
10919 "roothash=" kernel command line argument to both configure which root
10920 image and verity partition to use as well as the root hash for
10921 it. Note that systemd-nspawn's Verity support follows the same
10922 semantics, meaning that disk images with proper Verity data in place
10923 may be booted in containers with systemd-nspawn as well as on
10924 physical systems via the verity generator. Also note that the "mkosi"
10925 tool available at https://github.com/systemd/mkosi has been updated
10926 to generate Verity protected disk images following this scheme. In
10927 fact, it has been updated to generate disk images that optionally
10928 implement a complete UEFI SecureBoot trust chain, involving a signed
10929 kernel and initrd image that incorporates such a root hash as well as
10930 a Verity-enabled root partition.
10932 * The hardware database (hwdb) udev supports has been updated to carry
10933 accelerometer quirks.
10935 * All system services are now run with a fresh kernel keyring set up
10936 for them. The invocation ID is stored by default in it, thus
10937 providing a safe, non-overridable way to determine the invocation
10938 ID of each service.
10940 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
10941 options for bind mounting arbitrary paths in a service-specific
10942 way. When these options are used, arbitrary host or service files and
10943 directories may be mounted to arbitrary locations in the service's
10946 * Documentation has been added that lists all of systemd's low-level
10947 environment variables:
10949 https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md
10951 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
10952 whether a specific socket file descriptor matches a specified socket
10955 * systemd-firstboot has been updated to check for the
10956 systemd.firstboot= kernel command line option. It accepts a boolean
10957 and when set to false the first boot questions are skipped.
10959 * systemd-fstab-generator has been updated to check for the
10960 systemd.volatile= kernel command line option, which either takes an
10961 optional boolean parameter or the special value "state". If used the
10962 system may be booted in a "volatile" boot mode. Specifically,
10963 "systemd.volatile" is used, the root directory will be mounted as
10964 tmpfs, and only /usr is mounted from the actual root file system. If
10965 "systemd.volatile=state" is used, the root directory will be mounted
10966 as usual, but /var is mounted as tmpfs. This concept provides similar
10967 functionality as systemd-nspawn's --volatile= option, but provides it
10968 on physical boots. Use this option for implementing stateless
10969 systems, or testing systems with all state and/or configuration reset
10970 to the defaults. (Note though that many distributions are not
10971 prepared to boot up without a populated /etc or /var, though.)
10973 * systemd-gpt-auto-generator gained support for LUKS encrypted root
10974 partitions. Previously it only supported LUKS encrypted partitions
10975 for all other uses, except for the root partition itself.
10977 * Socket units gained support for listening on AF_VSOCK sockets for
10978 communication in virtualized QEMU environments.
10980 * The "configure" script gained a new option --with-fallback-hostname=
10981 for specifying the fallback hostname to use if none is configured in
10982 /etc/hostname. For example, by specifying
10983 --with-fallback-hostname=fedora it is possible to default to a
10984 hostname of "fedora" on pristine installations.
10986 * systemd-cgls gained support for a new --unit= switch for listing only
10987 the control groups of a specific unit. Similar --user-unit= has been
10988 added for listing only the control groups of a specific user unit.
10990 * systemd-mount gained a new --umount switch for unmounting a mount or
10991 automount point (and all mount/automount points below it).
10993 * systemd will now refuse full configuration reloads (via systemctl
10994 daemon-reload and related calls) unless at least 16MiB of free space
10995 are available in /run. This is a safety precaution in order to ensure
10996 that generators can safely operate after the reload completed.
10998 * A new unit file option RootImage= has been added, which has a similar
10999 effect as RootDirectory= but mounts the service's root directory from
11000 a disk image instead of plain directory. This logic reuses the same
11001 image dissection and mount logic that systemd-nspawn already uses,
11002 and hence supports any disk images systemd-nspawn supports, including
11003 those following the Discoverable Partition Specification, as well as
11004 Verity enabled images. This option enables systemd to run system
11005 services directly off disk images acting as resource bundles,
11006 possibly even including full integrity data.
11008 * A new MountAPIVFS= unit file option has been added, taking a boolean
11009 argument. If enabled /proc, /sys and /dev (collectively called the
11010 "API VFS") will be mounted for the service. This is only relevant if
11011 RootDirectory= or RootImage= is used for the service, as these mounts
11012 are of course in place in the host mount namespace anyway.
11014 * systemd-nspawn gained support for a new --pivot-root= switch. If
11015 specified the root directory within the container image is pivoted to
11016 the specified mount point, while the original root disk is moved to a
11017 different place. This option enables booting of ostree images
11018 directly with systemd-nspawn.
11020 * The systemd build scripts will no longer complain if the NTP server
11021 addresses are not changed from the defaults. Google now supports
11022 these NTP servers officially. We still recommend downstreams to
11023 properly register an NTP pool with the NTP pool project though.
11025 * coredumpctl gained a new "--reverse" option for printing the list
11026 of coredumps in reverse order.
11028 * coredumpctl will now show additional information about truncated and
11029 inaccessible coredumps, as well as coredumps that are still being
11030 processed. It also gained a new --quiet switch for suppressing
11031 additional informational message in its output.
11033 * coredumpctl gained support for only showing coredumps newer and/or
11034 older than specific timestamps, using the new --since= and --until=
11035 options, reminiscent of journalctl's options by the same name.
11037 * The systemd-coredump logic has been improved so that it may be reused
11038 to collect backtraces in non-compiled languages, for example in
11039 scripting languages such as Python.
11041 * machinectl will now show the UID shift of local containers, if user
11042 namespacing is enabled for them.
11044 * systemd will now optionally run "environment generator" binaries at
11045 configuration load time. They may be used to add environment
11046 variables to the environment block passed to services invoked. One
11047 user environment generator is shipped by default that sets up
11048 environment variables based on files dropped into /etc/environment.d
11049 and ~/.config/environment.d/.
11051 * systemd-resolved now includes the new, recently published 2017 DNSSEC
11054 * hostnamed has been updated to report a new chassis type of
11055 "convertible" to cover "foldable" laptops that can both act as a
11056 tablet and as a laptop, such as various Lenovo Yoga devices.
11058 Contributions from: Adrián López, Alexander Galanin, Alexander
11059 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
11060 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
11061 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
11062 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
11063 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
11064 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
11065 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
11066 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
11067 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
11068 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
11069 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
11070 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
11071 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
11072 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
11073 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
11074 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
11075 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
11076 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
11077 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
11078 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
11079 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
11080 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
11081 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
11082 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
11083 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
11084 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
11085 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
11088 — Berlin, 2017-03-01
11092 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
11093 RestrictAddressFamilies= enabled. These sandboxing options should
11094 generally be compatible with the various external udev call-out
11095 binaries we are aware of, however there may be exceptions, in
11096 particular when exotic languages for these call-outs are used. In
11097 this case, consider turning off these settings locally.
11099 * The new RemoveIPC= option can be used to remove IPC objects owned by
11100 the user or group of a service when that service exits.
11102 * The new ProtectKernelModules= option can be used to disable explicit
11103 load and unload operations of kernel modules by a service. In
11104 addition access to /usr/lib/modules is removed if this option is set.
11106 * ProtectSystem= option gained a new value "strict", which causes the
11107 whole file system tree with the exception of /dev, /proc, and /sys,
11108 to be remounted read-only for a service.
11110 * The new ProtectKernelTunables= option can be used to disable
11111 modification of configuration files in /sys and /proc by a service.
11112 Various directories and files are remounted read-only, so access is
11113 restricted even if the file permissions would allow it.
11115 * The new ProtectControlGroups= option can be used to disable write
11116 access by a service to /sys/fs/cgroup.
11118 * Various systemd services have been hardened with
11119 ProtectKernelTunables=yes, ProtectControlGroups=yes,
11120 RestrictAddressFamilies=.
11122 * Support for dynamically creating users for the lifetime of a service
11123 has been added. If DynamicUser=yes is specified, user and group IDs
11124 will be allocated from the range 61184…65519 for the lifetime of the
11125 service. They can be resolved using the new nss-systemd.so NSS
11126 module. The module must be enabled in /etc/nsswitch.conf. Services
11127 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
11128 any resources allocated by the service will be cleaned up when the
11129 service exits. They also have ProtectHome=read-only and
11130 ProtectSystem=strict enabled, so they are not able to make any
11131 permanent modifications to the system.
11133 * The nss-systemd module also always resolves root and nobody, making
11134 it possible to have no /etc/passwd or /etc/group files in minimal
11135 container or chroot environments.
11137 * Services may be started with their own user namespace using the new
11138 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
11139 under which the service is running are mapped. All other users are
11142 * Support for the cgroup namespace has been added to systemd-nspawn. If
11143 supported by kernel, the container system started by systemd-nspawn
11144 will have its own view of the cgroup hierarchy. This new behaviour
11145 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
11147 * The new MemorySwapMax= option can be used to limit the maximum swap
11148 usage under the unified cgroup hierarchy.
11150 * Support for the CPU controller in the unified cgroup hierarchy has
11151 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
11152 options. This controller requires out-of-tree patches for the kernel
11153 and the support is provisional.
11155 * Mount and automount units may now be created transiently
11156 (i.e. dynamically at runtime via the bus API, instead of requiring
11157 unit files in the file system).
11159 * systemd-mount is a new tool which may mount file systems – much like
11160 mount(8), optionally pulling in additional dependencies through
11161 transient .mount and .automount units. For example, this tool
11162 automatically runs fsck on a backing block device before mounting,
11163 and allows the automount logic to be used dynamically from the
11164 command line for establishing mount points. This tool is particularly
11165 useful when dealing with removable media, as it will ensure fsck is
11166 run – if necessary – before the first access and that the file system
11167 is quickly unmounted after each access by utilizing the automount
11168 logic. This maximizes the chance that the file system on the
11169 removable media stays in a clean state, and if it isn't in a clean
11170 state is fixed automatically.
11172 * LazyUnmount=yes option for mount units has been added to expose the
11173 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
11176 * /efi will be used as the mount point of the EFI boot partition, if
11177 the directory is present, and the mount point was not configured
11178 through other means (e.g. fstab). If /efi directory does not exist,
11179 /boot will be used as before. This makes it easier to automatically
11180 mount the EFI partition on systems where /boot is used for something
11183 * When operating on GPT disk images for containers, systemd-nspawn will
11184 now mount the ESP to /boot or /efi according to the same rules as PID
11185 1 running on a host. This allows tools like "bootctl" to operate
11186 correctly within such containers, in order to make container images
11187 bootable on physical systems.
11189 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
11191 * Two new user session targets have been added to support running
11192 graphical sessions under the systemd --user instance:
11193 graphical-session.target and graphical-session-pre.target. See
11194 systemd.special(7) for a description of how those targets should be
11197 * The vconsole initialization code has been significantly reworked to
11198 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
11199 support unicode keymaps. Font and keymap configuration will now be
11200 copied to all allocated virtual consoles.
11202 * FreeBSD's bhyve virtualization is now detected.
11204 * Information recorded in the journal for core dumps now includes the
11205 contents of /proc/mountinfo and the command line of the process at
11206 the top of the process hierarchy (which is usually the init process
11209 * systemd-journal-gatewayd learned the --directory= option to serve
11210 files from the specified location.
11212 * journalctl --root=… can be used to peruse the journal in the
11213 /var/log/ directories inside of a container tree. This is similar to
11214 the existing --machine= option, but does not require the container to
11217 * The hardware database has been extended to support
11218 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
11221 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
11222 specify the click rate for mice which include a horizontal wheel with
11223 a click rate that is different than the one for the vertical wheel.
11225 * systemd-run gained a new --wait option that makes service execution
11226 synchronous. (Specifically, the command will not return until the
11227 specified service binary exited.)
11229 * systemctl gained a new --wait option that causes the start command to
11230 wait until the units being started have terminated again.
11232 * A new journal output mode "short-full" has been added which displays
11233 timestamps with abbreviated English day names and adds a timezone
11234 suffix. Those timestamps include more information than the default
11235 "short" output mode, and can be passed directly to journalctl's
11236 --since= and --until= options.
11238 * /etc/resolv.conf will be bind-mounted into containers started by
11239 systemd-nspawn, if possible, so any changes to resolv.conf contents
11240 are automatically propagated to the container.
11242 * The number of instances for socket-activated services originating
11243 from a single IP address can be limited with
11244 MaxConnectionsPerSource=, extending the existing setting of
11247 * systemd-networkd gained support for vcan ("Virtual CAN") interface
11250 * .netdev and .network configuration can now be extended through
11253 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
11254 Segmentation Offload, Generic Receive Offload, Large Receive Offload
11255 can be enabled and disabled using the new UDPSegmentationOffload=,
11256 TCPSegmentationOffload=, GenericSegmentationOffload=,
11257 GenericReceiveOffload=, LargeReceiveOffload= options in the
11258 [Link] section of .link files.
11260 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
11261 Port VLAN ID can be configured for bridge devices using the new STP=,
11262 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
11263 section of .netdev files.
11265 * The route table to which routes received over DHCP or RA should be
11266 added can be configured with the new RouteTable= option in the [DHCP]
11267 and [IPv6AcceptRA] sections of .network files.
11269 * The Address Resolution Protocol can be disabled on links managed by
11270 systemd-networkd using the ARP=no setting in the [Link] section of
11273 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
11274 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
11275 encode information about the result and exit codes of the current
11276 service runtime cycle.
11278 * systemd-sysctl will now configure kernel parameters in the order
11279 they occur in the configuration files. This matches what sysctl
11280 has been traditionally doing.
11282 * kernel-install "plugins" that are executed to perform various
11283 tasks after a new kernel is added and before an old one is removed
11284 can now return a special value to terminate the procedure and
11285 prevent any later plugins from running.
11287 * Journald's SplitMode=login setting has been deprecated. It has been
11288 removed from documentation, and its use is discouraged. In a future
11289 release it will be completely removed, and made equivalent to current
11290 default of SplitMode=uid.
11292 * Storage=both option setting in /etc/systemd/coredump.conf has been
11293 removed. With fast LZ4 compression storing the core dump twice is not
11296 * The --share-system systemd-nspawn option has been replaced with an
11297 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
11298 this functionality is discouraged. In addition the variables
11299 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
11300 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
11301 individual namespaces.
11303 * "machinectl list" now shows the IP address of running containers in
11304 the output, as well as OS release information.
11306 * "loginctl list" now shows the TTY of each session in the output.
11308 * sd-bus gained new API calls sd_bus_track_set_recursive(),
11309 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
11310 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
11311 tracking objects in a "recursive" mode, where a single client can be
11312 counted multiple times, if it takes multiple references.
11314 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
11315 sd_bus_get_exit_on_disconnect(). They may be used to make a
11316 process using sd-bus automatically exit if the bus connection is
11319 * Bus clients of the service manager may now "pin" loaded units into
11320 memory, by taking an explicit reference on them. This is useful to
11321 ensure the client can retrieve runtime data about the service even
11322 after the service completed execution. Taking such a reference is
11323 available only for privileged clients and should be helpful to watch
11324 running services in a race-free manner, and in particular collect
11325 information about exit statuses and results.
11327 * The nss-resolve module has been changed to strictly return UNAVAIL
11328 when communication via D-Bus with resolved failed, and NOTFOUND when
11329 a lookup completed but was negative. This means it is now possible to
11330 neatly configure fallbacks using nsswitch.conf result checking
11331 expressions. Taking benefit of this, the new recommended
11332 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
11334 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
11336 * A new setting CtrlAltDelBurstAction= has been added to
11337 /etc/systemd/system.conf which may be used to configure the precise
11338 behaviour if the user on the console presses Ctrl-Alt-Del more often
11339 than 7 times in 2s. Previously this would unconditionally result in
11340 an expedited, immediate reboot. With this new setting the precise
11341 operation may be configured in more detail, and also turned off
11344 * In .netdev files two new settings RemoteChecksumTx= and
11345 RemoteChecksumRx= are now understood that permit configuring the
11346 remote checksumming logic for VXLAN networks.
11348 * The service manager learnt a new "invocation ID" concept for invoked
11349 services. Each runtime cycle of a service will get a new invocation
11350 ID (a 128-bit random UUID) assigned that identifies the current
11351 run of the service uniquely and globally. A new invocation ID
11352 is generated each time a service starts up. The journal will store
11353 the invocation ID of a service along with any logged messages, thus
11354 making the invocation ID useful for matching the online runtime of a
11355 service with the offline log data it generated in a safe way without
11356 relying on synchronized timestamps. In many ways this new service
11357 invocation ID concept is similar to the kernel's boot ID concept that
11358 uniquely and globally identifies the runtime of each boot. The
11359 invocation ID of a service is passed to the service itself via an
11360 environment variable ($INVOCATION_ID). A new bus call
11361 GetUnitByInvocationID() has been added that is similar to GetUnit()
11362 but instead of retrieving the bus path for a unit by its name
11363 retrieves it by its invocation ID. The returned path is valid only as
11364 long as the passed invocation ID is current.
11366 * systemd-resolved gained a new "DNSStubListener" setting in
11367 resolved.conf. It either takes a boolean value or the special values
11368 "udp" and "tcp", and configures whether to enable the stub DNS
11369 listener on 127.0.0.53:53.
11371 * IP addresses configured via networkd may now carry additional
11372 configuration settings supported by the kernel. New options include:
11373 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
11374 PrefixRoute=, AutoJoin=.
11376 * The PAM configuration fragment file for "user@.service" shipped with
11377 systemd (i.e. the --user instance of systemd) has been stripped to
11378 the minimum necessary to make the system boot. Previously, it
11379 contained Fedora-specific stanzas that did not apply to other
11380 distributions. It is expected that downstream distributions add
11381 additional configuration lines, matching their needs to this file,
11382 using it only as rough template of what systemd itself needs. Note
11383 that this reduced fragment does not even include an invocation of
11384 pam_limits which most distributions probably want to add, even though
11385 systemd itself does not need it. (There's also the new build time
11386 option --with-pamconfdir=no to disable installation of the PAM
11387 fragment entirely.)
11389 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
11390 capability is now also dropped from its set (in addition to
11391 CAP_SYS_MKNOD as before).
11393 * In service unit files it is now possible to connect a specific named
11394 file descriptor with stdin/stdout/stdout of an executed service. The
11395 name may be specified in matching .socket units using the
11396 FileDescriptorName= setting.
11398 * A number of journal settings may now be configured on the kernel
11399 command line. Specifically, the following options are now understood:
11400 systemd.journald.max_level_console=,
11401 systemd.journald.max_level_store=,
11402 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
11403 systemd.journald.max_level_wall=.
11405 * "systemctl is-enabled --full" will now show by which symlinks a unit
11406 file is enabled in the unit dependency tree.
11408 * Support for VeraCrypt encrypted partitions has been added to the
11409 "cryptsetup" logic and /etc/crypttab.
11411 * systemd-detect-virt gained support for a new --private-users switch
11412 that checks whether the invoking processes are running inside a user
11413 namespace. Similar, a new special value "private-users" for the
11414 existing ConditionVirtualization= setting has been added, permitting
11415 skipping of specific units in user namespace environments.
11417 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
11418 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
11419 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
11420 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
11421 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
11422 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
11423 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
11424 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
11425 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
11426 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
11427 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
11428 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
11429 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
11430 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
11431 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
11432 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
11433 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
11434 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
11435 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
11436 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
11437 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
11438 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
11439 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
11440 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
11441 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
11442 Jędrzejewski-Szmek, Zeal Jagannatha
11444 — Santa Fe, 2016-11-03
11448 * In service units the various ExecXYZ= settings have been extended
11449 with an additional special character as first argument of the
11450 assigned value: if the character '+' is used the specified command
11451 line it will be run with full privileges, regardless of User=,
11452 Group=, CapabilityBoundingSet= and similar options. The effect is
11453 similar to the existing PermissionsStartOnly= option, but allows
11454 configuration of this concept for each executed command line
11457 * Services may now alter the service watchdog timeout at runtime by
11458 sending a WATCHDOG_USEC= message via sd_notify().
11460 * MemoryLimit= and related unit settings now optionally take percentage
11461 specifications. The percentage is taken relative to the amount of
11462 physical memory in the system (or in case of containers, the assigned
11463 amount of memory). This allows scaling service resources neatly with
11464 the amount of RAM available on the system. Similarly, systemd-logind's
11465 RuntimeDirectorySize= option now also optionally takes percentage
11468 * In similar fashion TasksMax= takes percentage values now, too. The
11469 value is taken relative to the configured maximum number of processes
11470 on the system. The per-service task maximum has been changed to 15%
11471 using this functionality. (Effectively this is an increase of 512 →
11472 4915 for service units, given the kernel's default pid_max setting.)
11474 * Calendar time specifications in .timer units now understand a ".."
11475 syntax for time ranges. Example: "4..7:10" may now be used for
11476 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
11479 * The InaccessableDirectories=, ReadOnlyDirectories= and
11480 ReadWriteDirectories= unit file settings have been renamed to
11481 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
11482 applied to all kinds of file nodes, and not just directories, with
11483 the exception of symlinks. Specifically these settings may now be
11484 used on block and character device nodes, UNIX sockets and FIFOS as
11485 well as regular files. The old names of these settings remain
11486 available for compatibility.
11488 * systemd will now log about all service processes it kills forcibly
11489 (using SIGKILL) because they remained after the clean shutdown phase
11490 of the service completed. This should help identifying services that
11491 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
11492 systemd-logind's configuration a similar log message is generated for
11493 processes killed at the end of each session due to this setting.
11495 * systemd will now set the $JOURNAL_STREAM environment variable for all
11496 services whose stdout/stderr are connected to the Journal (which
11497 effectively means by default: all services). The variable contains
11498 the device and inode number of the file descriptor used for
11499 stdout/stderr. This may be used by invoked programs to detect whether
11500 their stdout/stderr is connected to the Journal, in which case they
11501 can switch over to direct Journal communication, thus being able to
11502 pass extended, structured metadata along with their log messages. As
11503 one example, this is now used by glib's logging primitives.
11505 * When using systemd's default tmp.mount unit for /tmp, the mount point
11506 will now be established with the "nosuid" and "nodev" options. This
11507 avoids privilege escalation attacks that put traps and exploits into
11508 /tmp. However, this might cause problems if you e.g. put container
11509 images or overlays into /tmp; if you need this, override tmp.mount's
11510 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
11513 * systemd now supports the "memory" cgroup controller also on
11516 * The systemd-cgtop tool now optionally takes a control group path as
11517 command line argument. If specified, the control group list shown is
11518 limited to subgroups of that group.
11520 * The SystemCallFilter= unit file setting gained support for
11521 pre-defined, named system call filter sets. For example
11522 SystemCallFilter=@clock is now an effective way to make all clock
11523 changing-related system calls unavailable to a service. A number of
11524 similar pre-defined groups are defined. Writing system call filters
11525 for system services is simplified substantially with this new
11526 concept. Accordingly, all of systemd's own, long-running services now
11527 enable system call filtering based on this, by default.
11529 * A new service setting MemoryDenyWriteExecute= has been added, taking
11530 a boolean value. If turned on, a service may no longer create memory
11531 mappings that are writable and executable at the same time. This
11532 enhances security for services where this is enabled as it becomes
11533 harder to dynamically write and then execute memory in exploited
11534 service processes. This option has been enabled for all of systemd's
11535 own long-running services.
11537 * A new RestrictRealtime= service setting has been added, taking a
11538 boolean argument. If set the service's processes may no longer
11539 acquire realtime scheduling. This improves security as realtime
11540 scheduling may otherwise be used to easily freeze the system.
11542 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
11543 value. This may be used for requesting that the system manager inside
11544 of the container reports start-up completion to nspawn which then
11545 propagates this notification further to the service manager
11546 supervising nspawn itself. A related option NotifyReady= in .nspawn
11547 files has been added too. This functionality allows ordering of the
11548 start-up of multiple containers using the usual systemd ordering
11551 * machinectl gained a new command "stop" that is an alias for
11554 * systemd-resolved gained support for contacting DNS servers on
11555 link-local IPv6 addresses.
11557 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
11558 its caches. A method call for requesting the same operation has been
11559 added to the bus API too, and is made available via "systemd-resolve
11562 * systemd-resolve gained a new --status switch. If passed a brief
11563 summary of the used DNS configuration with per-interface information
11566 * resolved.conf gained a new Cache= boolean option, defaulting to
11567 on. If turned off local DNS caching is disabled. This comes with a
11568 performance penalty in particular when DNSSEC is enabled. Note that
11569 resolved disables its internal caching implicitly anyway, when the
11570 configured DNS server is on a host-local IP address such as ::1 or
11571 127.0.0.1, thus automatically avoiding double local caching.
11573 * systemd-resolved now listens on the local IP address 127.0.0.53:53
11574 for DNS requests. This improves compatibility with local programs
11575 that do not use the libc NSS or systemd-resolved's bus APIs for name
11576 resolution. This minimal DNS service is only available to local
11577 programs and does not implement the full DNS protocol, but enough to
11578 cover local DNS clients. A new, static resolv.conf file, listing just
11579 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
11580 now recommended to make /etc/resolv.conf a symlink to this file in
11581 order to route all DNS lookups to systemd-resolved, regardless if
11582 done via NSS, the bus API or raw DNS packets. Note that this local
11583 DNS service is not as fully featured as the libc NSS or
11584 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
11585 used to deliver link-local address information (as this implies
11586 sending a local interface index along), LLMNR/mDNS support via this
11587 interface is severely restricted. It is thus strongly recommended for
11588 all applications to use the libc NSS API or native systemd-resolved
11591 * systemd-networkd's bridge support learned a new setting
11592 VLANFiltering= for controlling VLAN filtering. Moreover a new section
11593 in .network files has been added for configuring VLAN bridging in
11594 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
11596 * systemd-networkd's IPv6 Router Advertisement code now makes use of
11597 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
11598 now be acquired without relying on DHCPv6. Two new options
11599 UseDomains= and UseDNS= have been added to configure this behaviour.
11601 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
11602 renamed IPv6AcceptRA=, without altering its behaviour. The old
11603 setting name remains available for compatibility reasons.
11605 * The systemd-networkd VTI/VTI6 tunneling support gained new options
11606 Key=, InputKey= and OutputKey=.
11608 * systemd-networkd gained support for VRF ("Virtual Routing Function")
11609 interface configuration.
11611 * "systemctl edit" may now be used to create new unit files by
11612 specifying the --force switch.
11614 * sd-event gained a new function sd_event_get_iteration() for
11615 requesting the current iteration counter of the event loop. It starts
11616 at zero and is increased by one with each event loop iteration.
11618 * A new rpm macro %systemd_ordering is provided by the macros.systemd
11619 file. It can be used in lieu of %systemd_requires in packages which
11620 don't use any systemd functionality and are intended to be installed
11621 in minimal containers without systemd present. This macro provides
11622 ordering dependencies to ensure that if the package is installed in
11623 the same rpm transaction as systemd, systemd will be installed before
11624 the scriptlets for the package are executed, allowing unit presets
11627 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
11628 been added to simplify packaging of generators.
11630 * The os-release file gained VERSION_CODENAME field for the
11631 distribution nickname (e.g. VERSION_CODENAME=woody).
11633 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
11634 can be set to disable parsing of metadata and the creation
11635 of persistent symlinks for that device.
11637 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
11638 to make them available to logged-in users has been reverted.
11640 * Much of the common code of the various systemd components is now
11641 built into an internal shared library libsystemd-shared-231.so
11642 (incorporating the systemd version number in the name, to be updated
11643 with future releases) that the components link to. This should
11644 decrease systemd footprint both in memory during runtime and on
11645 disk. Note that the shared library is not for public use, and is
11646 neither API nor ABI stable, but is likely to change with every new
11647 released update. Packagers need to make sure that binaries
11648 linking to libsystemd-shared.so are updated in step with the
11651 * Configuration for "mkosi" is now part of the systemd
11652 repository. mkosi is a tool to easily build legacy-free OS images,
11653 and is available on github: https://github.com/systemd/mkosi. If
11654 "mkosi" is invoked in the build tree a new raw OS image is generated
11655 incorporating the systemd sources currently being worked on and a
11656 clean, fresh distribution installation. The generated OS image may be
11657 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
11658 UEFI PC. This functionality is particularly useful to easily test
11659 local changes made to systemd in a pristine, defined environment. See
11660 doc/HACKING for details.
11662 * configure learned the --with-support-url= option to specify the
11663 distribution's bugtracker.
11665 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
11666 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
11667 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
11668 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
11669 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
11670 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
11671 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
11672 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
11673 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
11674 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
11675 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
11676 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
11677 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
11678 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
11679 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
11680 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
11681 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
11682 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
11683 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
11685 — Berlin, 2016-07-25
11689 * DNSSEC is now turned on by default in systemd-resolved (in
11690 "allow-downgrade" mode), but may be turned off during compile time by
11691 passing "--with-default-dnssec=no" to "configure" (and of course,
11692 during runtime with DNSSEC= in resolved.conf). We recommend
11693 downstreams to leave this on at least during development cycles and
11694 report any issues with the DNSSEC logic upstream. We are very
11695 interested in collecting feedback about the DNSSEC validator and its
11696 limitations in the wild. Note however, that DNSSEC support is
11697 probably nothing downstreams should turn on in stable distros just
11698 yet, as it might create incompatibilities with a few DNS servers and
11699 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
11700 automatically whenever we detect such incompatible setups, but there
11701 might be systems we do not cover yet. Hence: please help us testing
11702 the DNSSEC code, leave this on where you can, report back, but then
11703 again don't consider turning this on in your stable, LTS or
11704 production release just yet. (Note that you have to enable
11705 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
11706 and its DNSSEC mode for hostname resolution from local
11709 * systemd-resolve conveniently resolves DANE records with the --tlsa
11710 option and OPENPGPKEY records with the --openpgp option. It also
11711 supports dumping raw DNS record data via the new --raw= switch.
11713 * systemd-logind will now by default terminate user processes that are
11714 part of the user session scope unit (session-XX.scope) when the user
11715 logs out. This behavior is controlled by the KillUserProcesses=
11716 setting in logind.conf, and the previous default of "no" is now
11717 changed to "yes". This means that user sessions will be properly
11718 cleaned up after, but additional steps are necessary to allow
11719 intentionally long-running processes to survive logout.
11721 While the user is logged in at least once, user@.service is running,
11722 and any service that should survive the end of any individual login
11723 session can be started at a user service or scope using systemd-run.
11724 systemd-run(1) man page has been extended with an example which shows
11725 how to run screen in a scope unit underneath user@.service. The same
11726 command works for tmux.
11728 After the user logs out of all sessions, user@.service will be
11729 terminated too, by default, unless the user has "lingering" enabled.
11730 To effectively allow users to run long-term tasks even if they are
11731 logged out, lingering must be enabled for them. See loginctl(1) for
11732 details. The default polkit policy was modified to allow users to
11733 set lingering for themselves without authentication.
11735 Previous defaults can be restored at compile time by the
11736 --without-kill-user-processes option to "configure".
11738 * systemd-logind gained new configuration settings SessionsMax= and
11739 InhibitorsMax=, both with a default of 8192. It will not register new
11740 user sessions or inhibitors above this limit.
11742 * systemd-logind will now reload configuration on SIGHUP.
11744 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
11745 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
11746 enable. Also, support for the "io" cgroup controller in the unified
11747 hierarchy has been added, so that the "memory", "pids" and "io" are
11748 now the controllers that are supported on the unified hierarchy.
11750 WARNING: it is not possible to use previous systemd versions with
11751 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
11752 is necessary to also update systemd in the initramfs if using the
11753 unified hierarchy. An updated SELinux policy is also required.
11755 * LLDP support has been extended, and both passive (receive-only) and
11756 active (sender) modes are supported. Passive mode ("routers-only") is
11757 enabled by default in systemd-networkd. Active LLDP mode is enabled
11758 by default for containers on the internal network. The "networkctl
11759 lldp" command may be used to list information gathered. "networkctl
11760 status" will also show basic LLDP information on connected peers now.
11762 * The IAID and DUID unique identifier sent in DHCP requests may now be
11763 configured for the system and each .network file managed by
11764 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
11766 * systemd-networkd gained support for configuring proxy ARP support for
11767 each interface, via the ProxyArp= setting in .network files. It also
11768 gained support for configuring the multicast querier feature of
11769 bridge devices, via the new MulticastQuerier= setting in .netdev
11770 files. Similarly, snooping on the IGMP traffic can be controlled
11771 via the new setting MulticastSnooping=.
11773 A new setting PreferredLifetime= has been added for addresses
11774 configured in .network file to configure the lifetime intended for an
11777 The systemd-networkd DHCP server gained the option EmitRouter=, which
11778 defaults to yes, to configure whether the DHCP Option 3 (Router)
11781 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
11782 systemd-socket-activate and installed into /usr/bin. It is now fully
11785 * systemd-journald now uses separate threads to flush changes to disk
11786 when closing journal files, thus reducing impact of slow disk I/O on
11787 logging performance.
11789 * The sd-journal API gained two new calls
11790 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
11791 can be used to open journal files using file descriptors instead of
11792 file or directory paths. sd_journal_open_container() has been
11793 deprecated, sd_journal_open_directory_fd() should be used instead
11794 with the flag SD_JOURNAL_OS_ROOT.
11796 * journalctl learned a new output mode "-o short-unix" that outputs log
11797 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
11798 UTC). It also gained support for a new --no-hostname setting to
11799 suppress the hostname column in the family of "short" output modes.
11801 * systemd-ask-password now optionally skips printing of the password to
11802 stdout with --no-output which can be useful in scripts.
11804 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
11805 (devices tagged with ID_MAKER_TOOL) are now tagged with
11806 "uaccess" and are available to logged in users.
11808 * The DeviceAllow= unit setting now supports specifiers (with "%").
11810 * "systemctl show" gained a new --value switch, which allows print a
11811 only the contents of a specific unit property, without also printing
11812 the property's name. Similar support was added to "show*" verbs
11813 of loginctl and machinectl that output "key=value" lists.
11815 * A new unit type "generated" was added for files dynamically generated
11816 by generator tools. Similarly, a new unit type "transient" is used
11817 for unit files created using the runtime API. "systemctl enable" will
11818 refuse to operate on such files.
11820 * A new command "systemctl revert" has been added that may be used to
11821 revert to the vendor version of a unit file, in case local changes
11822 have been made by adding drop-ins or overriding the unit file.
11824 * "machinectl clean" gained a new verb to automatically remove all or
11825 just hidden container images.
11827 * systemd-tmpfiles gained support for a new line type "e" for emptying
11828 directories, if they exist, without creating them if they don't.
11830 * systemd-nspawn gained support for automatically patching the UID/GIDs
11831 of the owners and the ACLs of all files and directories in a
11832 container tree to match the UID/GID user namespacing range selected
11833 for the container invocation. This mode is enabled via the new
11834 --private-users-chown switch. It also gained support for
11835 automatically choosing a free, previously unused UID/GID range when
11836 starting a container, via the new --private-users=pick setting (which
11837 implies --private-users-chown). Together, these options for the first
11838 time make user namespacing for nspawn containers fully automatic and
11839 thus deployable. The systemd-nspawn@.service template unit file has
11840 been changed to use this functionality by default.
11842 * systemd-nspawn gained a new --network-zone= switch, that allows
11843 creating ad-hoc virtual Ethernet links between multiple containers,
11844 that only exist as long as at least one container referencing them is
11845 running. This allows easy connecting of multiple containers with a
11846 common link that implements an Ethernet broadcast domain. Each of
11847 these network "zones" may be named relatively freely by the user, and
11848 may be referenced by any number of containers, but each container may
11849 only reference one of these "zones". On the lower level, this is
11850 implemented by an automatically managed bridge network interface for
11851 each zone, that is created when the first container referencing its
11852 zone is created and removed when the last one referencing its zone
11855 * The default start timeout may now be configured on the kernel command
11856 line via systemd.default_timeout_start_sec=. It was already
11857 configurable via the DefaultTimeoutStartSec= option in
11858 /etc/systemd/system.conf.
11860 * Socket units gained a new TriggerLimitIntervalSec= and
11861 TriggerLimitBurst= setting to configure a limit on the activation
11862 rate of the socket unit.
11864 * The LimitNICE= setting now optionally takes normal UNIX nice values
11865 in addition to the raw integer limit value. If the specified
11866 parameter is prefixed with "+" or "-" and is in the range -20…19 the
11867 value is understood as UNIX nice value. If not prefixed like this it
11868 is understood as raw RLIMIT_NICE limit.
11870 * Note that the effect of the PrivateDevices= unit file setting changed
11871 slightly with this release: the per-device /dev file system will be
11872 mounted read-only from this version on, and will have "noexec"
11873 set. This (minor) change of behavior might cause some (exceptional)
11874 legacy software to break, when PrivateDevices=yes is set for its
11875 service. Please leave PrivateDevices= off if you run into problems
11878 * systemd-bootchart has been split out to a separate repository:
11879 https://github.com/systemd/systemd-bootchart
11881 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
11882 merged into the kernel in its current form.
11884 * The compatibility libraries libsystemd-daemon.so,
11885 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
11886 which have been deprecated since systemd-209 have been removed along
11887 with the corresponding pkg-config files. All symbols provided by
11888 those libraries are provided by libsystemd.so.
11890 * The Capabilities= unit file setting has been removed (it is ignored
11891 for backwards compatibility). AmbientCapabilities= and
11892 CapabilityBoundingSet= should be used instead.
11894 * A new special target has been added, initrd-root-device.target,
11895 which creates a synchronization point for dependencies of the root
11896 device in early userspace. Initramfs builders must ensure that this
11897 target is now included in early userspace.
11899 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
11900 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
11901 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
11902 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
11903 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
11904 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
11905 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
11906 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
11907 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
11908 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
11909 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
11910 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
11911 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
11912 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
11913 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
11914 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
11915 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
11916 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
11917 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
11918 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
11919 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
11920 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
11921 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
11922 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
11923 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
11926 — Fairfax, 2016-05-21
11930 * The systemd-resolved DNS resolver service has gained a substantial
11931 set of new features, most prominently it may now act as a DNSSEC
11932 validating stub resolver. DNSSEC mode is currently turned off by
11933 default, but is expected to be turned on by default in one of the
11934 next releases. For now, we invite everybody to test the DNSSEC logic
11935 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
11936 service also gained a full set of D-Bus interfaces, including calls
11937 to configure DNS and DNSSEC settings per link (for use by external
11938 network management software). systemd-resolved and systemd-networkd
11939 now distinguish between "search" and "routing" domains. The former
11940 are used to qualify single-label names, the latter are used purely
11941 for routing lookups within certain domains to specific links.
11942 resolved now also synthesizes RRs for all entries from /etc/hosts.
11944 * The systemd-resolve tool (which is a client utility for
11945 systemd-resolved) has been improved considerably and is now fully
11946 supported and documented. Hence it has moved from /usr/lib/systemd to
11949 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
11952 * The coredump collection logic has been reworked: when a coredump is
11953 collected it is now written to disk, compressed and processed
11954 (including stacktrace extraction) from a new instantiated service
11955 systemd-coredump@.service, instead of directly from the
11956 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
11957 processing large coredumps can take up a substantial amount of
11958 resources and time, and this previously happened entirely outside of
11959 systemd's service supervision. With the new logic the core_pattern
11960 hook only does minimal metadata collection before passing off control
11961 to the new instantiated service, which is configured with a time
11962 limit, a nice level and other settings to minimize negative impact on
11963 the rest of the system. Also note that the new logic will honour the
11964 RLIMIT_CORE setting of the crashed process, which now allows users
11965 and processes to turn off coredumping for their processes by setting
11968 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
11969 and all forked processes by default. Previously, PID 1 would leave
11970 the setting at "0" for all processes, as set by the kernel. Note that
11971 the resource limit traditionally has no effect on the generated
11972 coredumps on the system if the /proc/sys/kernel/core_pattern hook
11973 logic is used. Since the limit is now honoured (see above) its
11974 default has been changed so that the coredumping logic is enabled by
11975 default for all processes, while allowing specific opt-out.
11977 * When the stacktrace is extracted from processes of system users, this
11978 is now done as "systemd-coredump" user, in order to sandbox this
11979 potentially security sensitive parsing operation. (Note that when
11980 processing coredumps of normal users this is done under the user ID
11981 of process that crashed, as before.) Packagers should take notice
11982 that it is now necessary to create the "systemd-coredump" system user
11983 and group at package installation time.
11985 * The systemd-activate socket activation testing tool gained support
11986 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
11987 and --seqpacket switches. It also has been extended to support both
11988 new-style and inetd-style file descriptor passing. Use the new
11989 --inetd switch to request inetd-style file descriptor passing.
11991 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
11992 variable, which takes a boolean value. If set to false, ANSI color
11993 output is disabled in the tools even when run on a terminal that
11996 * The VXLAN support in networkd now supports two new settings
11997 DestinationPort= and PortRange=.
11999 * A new systemd.machine_id= kernel command line switch has been added,
12000 that may be used to set the machine ID in /etc/machine-id if it is
12001 not initialized yet. This command line option has no effect if the
12002 file is already initialized.
12004 * systemd-nspawn gained a new --as-pid2 switch that invokes any
12005 specified command line as PID 2 rather than PID 1 in the
12006 container. In this mode PID 1 is a minimal stub init process that
12007 implements the special POSIX and Linux semantics of PID 1 regarding
12008 signal and child process management. Note that this stub init process
12009 is implemented in nspawn itself and requires no support from the
12010 container image. This new logic is useful to support running
12011 arbitrary commands in the container, as normal processes are
12012 generally not prepared to run as PID 1.
12014 * systemd-nspawn gained a new --chdir= switch for setting the current
12015 working directory for the process started in the container.
12017 * "journalctl /dev/sda" will now output all kernel log messages for
12018 specified device from the current boot, in addition to all devices
12019 that are parents of it. This should make log output about devices
12020 pretty useful, as long as kernel drivers attach enough metadata to
12021 the log messages. (The usual SATA drivers do.)
12023 * The sd-journal API gained two new calls
12024 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
12025 that report whether log data from /run or /var has been found.
12027 * journalctl gained a new switch "--fields" that prints all journal
12028 record field names currently in use in the journal. This is backed
12029 by two new sd-journal API calls sd_journal_enumerate_fields() and
12030 sd_journal_restart_fields().
12032 * Most configurable timeouts in systemd now expect an argument of
12033 "infinity" to turn them off, instead of "0" as before. The semantics
12034 from now on is that a timeout of "0" means "now", and "infinity"
12035 means "never". To maintain backwards compatibility, "0" continues to
12036 turn off previously existing timeout settings.
12038 * "systemctl reload-or-try-restart" has been renamed to "systemctl
12039 try-reload-or-restart" to clarify what it actually does: the "try"
12040 logic applies to both reloading and restarting, not just restarting.
12041 The old name continues to be accepted for compatibility.
12043 * On boot-up, when PID 1 detects that the system clock is behind the
12044 release date of the systemd version in use, the clock is now set
12045 to the latter. Previously, this was already done in timesyncd, in order
12046 to avoid running with clocks set to the various clock epochs such as
12047 1902, 1938 or 1970. With this change the logic is now done in PID 1
12048 in addition to timesyncd during early boot-up, so that it is enforced
12049 before the first process is spawned by systemd. Note that the logic
12050 in timesyncd remains, as it is more comprehensive and ensures
12051 clock monotonicity by maintaining a persistent timestamp file in
12052 /var. Since /var is generally not available in earliest boot or the
12053 initrd, this part of the logic remains in timesyncd, and is not done
12056 * Support for tweaking details in net_cls.class_id through the
12057 NetClass= configuration directive has been removed, as the kernel
12058 people have decided to deprecate that controller in cgroup v2.
12059 Userspace tools such as nftables are moving over to setting rules
12060 that are specific to the full cgroup path of a task, which obsoletes
12061 these controllers anyway. The NetClass= directive is kept around for
12062 legacy compatibility reasons. For a more in-depth description of the
12063 kernel change, please refer to the respective upstream commit:
12065 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
12067 * A new service setting RuntimeMaxSec= has been added that may be used
12068 to specify a maximum runtime for a service. If the timeout is hit, the
12069 service is terminated and put into a failure state.
12071 * A new service setting AmbientCapabilities= has been added. It allows
12072 configuration of additional Linux process capabilities that are
12073 passed to the activated processes. This is only available on very
12076 * The process resource limit settings in service units may now be used
12077 to configure hard and soft limits individually.
12079 * The various libsystemd APIs such as sd-bus or sd-event now publicly
12080 expose support for gcc's __attribute__((cleanup())) C extension.
12081 Specifically, for many object destructor functions alternative
12082 versions have been added that have names suffixed with "p" and take a
12083 pointer to a pointer to the object to destroy, instead of just a
12084 pointer to the object itself. This is useful because these destructor
12085 functions may be used directly as parameters to the cleanup
12086 construct. Internally, systemd has been a heavy user of this GCC
12087 extension for a long time, and with this change similar support is
12088 now available to consumers of the library outside of systemd. Note
12089 that by using this extension in your sources compatibility with old
12090 and strictly ANSI compatible C compilers is lost. However, all gcc or
12091 LLVM versions of recent years support this extension.
12093 * Timer units gained support for a new setting RandomizedDelaySec= that
12094 allows configuring some additional randomized delay to the configured
12095 time. This is useful to spread out timer events to avoid load peaks in
12096 clusters or larger setups.
12098 * Calendar time specifications now support sub-second accuracy.
12100 * Socket units now support listening on SCTP and UDP-lite protocol
12103 * The sd-event API now comes with a full set of man pages.
12105 * Older versions of systemd contained experimental support for
12106 compressing journal files and coredumps with the LZ4 compressor that
12107 was not compatible with the lz4 binary (due to API limitations of the
12108 lz4 library). This support has been removed; only support for files
12109 compatible with the lz4 binary remains. This LZ4 logic is now
12110 officially supported and no longer considered experimental.
12112 * The dkr image import logic has been removed again from importd. dkr's
12113 micro-services focus doesn't fit into the machine image focus of
12114 importd, and quickly got out of date with the upstream dkr API.
12116 * Creation of the /run/lock/lockdev/ directory was dropped from
12117 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
12118 been available for many years. If you still need this, you need to
12119 create your own tmpfiles.d config file with:
12121 d /run/lock/lockdev 0775 root lock -
12123 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
12124 and RebootArgument= have been moved from the [Service] section of
12125 unit files to [Unit], and they are now supported on all unit types,
12126 not just service units. Of course, systemd will continue to
12127 understand these settings also at the old location, in order to
12128 maintain compatibility.
12130 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
12131 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
12132 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
12133 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
12134 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
12135 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
12136 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
12137 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
12138 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
12139 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
12140 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
12141 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
12142 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
12143 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
12144 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
12145 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
12146 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
12147 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
12148 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
12150 — Berlin, 2016-02-11
12154 * A number of properties previously only settable in unit
12155 files are now also available as properties to set when
12156 creating transient units programmatically via the bus, as it
12157 is exposed with systemd-run's --property=
12158 setting. Specifically, these are: SyslogIdentifier=,
12159 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
12160 EnvironmentFile=, ReadWriteDirectories=,
12161 ReadOnlyDirectories=, InaccessibleDirectories=,
12162 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
12164 * When creating transient services via the bus API it is now
12165 possible to pass in a set of file descriptors to use as
12166 STDIN/STDOUT/STDERR for the invoked process.
12168 * Slice units may now be created transiently via the bus APIs,
12169 similar to the way service and scope units may already be
12170 created transiently.
12172 * Wherever systemd expects a calendar timestamp specification
12173 (like in journalctl's --since= and --until= switches) UTC
12174 timestamps are now supported. Timestamps suffixed with "UTC"
12175 are now considered to be in Universal Time Coordinated
12176 instead of the local timezone. Also, timestamps may now
12177 optionally be specified with sub-second accuracy. Both of
12178 these additions also apply to recurring calendar event
12179 specification, such as OnCalendar= in timer units.
12181 * journalctl gained a new "--sync" switch that asks the
12182 journal daemon to write all so far unwritten log messages to
12183 disk and sync the files, before returning.
12185 * systemd-tmpfiles learned two new line types "q" and "Q" that
12186 operate like "v", but also set up a basic btrfs quota
12187 hierarchy when used on a btrfs file system with quota
12190 * tmpfiles' "v", "q" and "Q" will now create a plain directory
12191 instead of a subvolume (even on a btrfs file system) if the
12192 root directory is a plain directory, and not a
12193 subvolume. This should simplify things with certain chroot()
12194 environments which are not aware of the concept of btrfs
12197 * systemd-detect-virt gained a new --chroot switch to detect
12198 whether execution takes place in a chroot() environment.
12200 * CPUAffinity= now takes CPU index ranges in addition to
12201 individual indexes.
12203 * The various memory-related resource limit settings (such as
12204 LimitAS=) now understand the usual K, M, G, … suffixes to
12205 the base of 1024 (IEC). Similar, the time-related resource
12206 limit settings understand the usual min, h, day, … suffixes
12209 * There's a new system.conf setting DefaultTasksMax= to
12210 control the default TasksMax= setting for services and
12211 scopes running on the system. (TasksMax= is the primary
12212 setting that exposes the "pids" cgroup controller on systemd
12213 and was introduced in the previous systemd release.) The
12214 setting now defaults to 512, which means services that are
12215 not explicitly configured otherwise will only be able to
12216 create 512 processes or threads at maximum, from this
12217 version on. Note that this means that thread- or
12218 process-heavy services might need to be reconfigured to set
12219 TasksMax= to a higher value. It is sufficient to set
12220 TasksMax= in these specific unit files to a higher value, or
12221 even "infinity". Similar, there's now a logind.conf setting
12222 UserTasksMax= that defaults to 4096 and limits the total
12223 number of processes or tasks each user may own
12224 concurrently. nspawn containers also have the TasksMax=
12225 value set by default now, to 8192. Note that all of this
12226 only has an effect if the "pids" cgroup controller is
12227 enabled in the kernel. The general benefit of these changes
12228 should be a more robust and safer system, that provides a
12229 certain amount of per-service fork() bomb protection.
12231 * systemd-nspawn gained the new --network-veth-extra= switch
12232 to define additional and arbitrarily-named virtual Ethernet
12233 links between the host and the container.
12235 * A new service execution setting PassEnvironment= has been
12236 added that allows importing select environment variables
12237 from PID1's environment block into the environment block of
12240 * Timer units gained support for a new RemainAfterElapse=
12241 setting which takes a boolean argument. It defaults to on,
12242 exposing behaviour unchanged to previous releases. If set to
12243 off, timer units are unloaded after they elapsed if they
12244 cannot elapse again. This is particularly useful for
12245 transient timer units, which shall not stay around longer
12246 than until they first elapse.
12248 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
12249 default now (the kernel default is 16). This is beneficial
12250 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
12251 allows substantially larger numbers of queued
12252 datagrams. This should increase the capability of systemd to
12253 parallelize boot-up, as logging and sd_notify() are unlikely
12254 to stall execution anymore. If you need to change the value
12255 from the new defaults, use the usual sysctl.d/ snippets.
12257 * The compression framing format used by the journal or
12258 coredump processing has changed to be in line with what the
12259 official LZ4 tools generate. LZ4 compression support in
12260 systemd was considered unsupported previously, as the format
12261 was not compatible with the normal tools. With this release
12262 this has changed now, and it is hence safe for downstream
12263 distributions to turn it on. While not compressing as well
12264 as the XZ, LZ4 is substantially faster, which makes
12265 it a good default choice for the compression logic in the
12266 journal and in coredump handling.
12268 * Any reference to /etc/mtab has been dropped from
12269 systemd. The file has been obsolete since a while, but
12270 systemd refused to work on systems where it was incorrectly
12271 set up (it should be a symlink or non-existent). Please make
12272 sure to update to util-linux 2.27.1 or newer in conjunction
12273 with this systemd release, which also drops any reference to
12274 /etc/mtab. If you maintain a distribution make sure that no
12275 software you package still references it, as this is a
12276 likely source of bugs. There's also a glibc bug pending,
12277 asking for removal of any reference to this obsolete file:
12279 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
12281 Note that only util-linux versions built with
12282 --enable-libmount-force-mountinfo are supported.
12284 * Support for the ".snapshot" unit type has been removed. This
12285 feature turned out to be little useful and little used, and
12286 has now been removed from the core and from systemctl.
12288 * The dependency types RequiresOverridable= and
12289 RequisiteOverridable= have been removed from systemd. They
12290 have been used only very sparingly to our knowledge and
12291 other options that provide a similar effect (such as
12292 systemctl --mode=ignore-dependencies) are much more useful
12293 and commonly used. Moreover, they were only half-way
12294 implemented as the option to control behaviour regarding
12295 these dependencies was never added to systemctl. By removing
12296 these dependency types the execution engine becomes a bit
12297 simpler. Unit files that use these dependencies should be
12298 changed to use the non-Overridable dependency types
12299 instead. In fact, when parsing unit files with these
12300 options, that's what systemd will automatically convert them
12301 too, but it will also warn, asking users to fix the unit
12302 files accordingly. Removal of these dependency types should
12303 only affect a negligible number of unit files in the wild.
12305 * Behaviour of networkd's IPForward= option changed
12306 (again). It will no longer maintain a per-interface setting,
12307 but propagate one way from interfaces where this is enabled
12308 to the global kernel setting. The global setting will be
12309 enabled when requested by a network that is set up, but
12310 never be disabled again. This change was made to make sure
12311 IPv4 and IPv6 behaviour regarding packet forwarding is
12312 similar (as the Linux IPv6 stack does not support
12313 per-interface control of this setting) and to minimize
12316 * In unit files the behaviour of %u, %U, %h, %s has
12317 changed. These specifiers will now unconditionally resolve
12318 to the various user database fields of the user that the
12319 systemd instance is running as, instead of the user
12320 configured in the specific unit via User=. Note that this
12321 effectively doesn't change much, as resolving of these
12322 specifiers was already turned off in the --system instance
12323 of systemd, as we cannot do NSS lookups from PID 1. In the
12324 --user instance of systemd these specifiers where correctly
12325 resolved, but hardly made any sense, since the user instance
12326 lacks privileges to do user switches anyway, and User= is
12327 hence useless. Moreover, even in the --user instance of
12328 systemd behaviour was awkward as it would only take settings
12329 from User= assignment placed before the specifier into
12330 account. In order to unify and simplify the logic around
12331 this the specifiers will now always resolve to the
12332 credentials of the user invoking the manager (which in case
12333 of PID 1 is the root user).
12335 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
12336 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
12337 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
12338 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
12339 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
12340 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
12341 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
12342 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
12343 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
12344 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
12345 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
12346 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
12347 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
12348 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
12351 — Berlin, 2015-11-18
12355 * systemd now depends on util-linux v2.27. More specifically,
12356 the newly added mount monitor feature in libmount now
12357 replaces systemd's former own implementation.
12359 * libmount mandates /etc/mtab not to be regular file, and
12360 systemd now enforces this condition at early boot.
12361 /etc/mtab has been deprecated and warned about for a very
12362 long time, so systems running systemd should already have
12363 stopped having this file around as anything else than a
12364 symlink to /proc/self/mounts.
12366 * Support for the "pids" cgroup controller has been added. It
12367 allows accounting the number of tasks in a cgroup and
12368 enforcing limits on it. This adds two new setting
12369 TasksAccounting= and TasksMax= to each unit, as well as a
12370 global option DefaultTasksAccounting=.
12372 * Support for the "net_cls" cgroup controller has been added.
12373 It allows assigning a net class ID to each task in the
12374 cgroup, which can then be used in firewall rules and traffic
12375 shaping configurations. Note that the kernel netfilter net
12376 class code does not currently work reliably for ingress
12377 packets on unestablished sockets.
12379 This adds a new config directive called NetClass= to CGroup
12380 enabled units. Allowed values are positive numbers for fixed
12381 assignments and "auto" for picking a free value
12384 * 'systemctl is-system-running' now returns 'offline' if the
12385 system is not booted with systemd. This command can now be
12386 used as a substitute for 'systemd-notify --booted'.
12388 * Watchdog timeouts have been increased to 3 minutes for all
12389 in-tree service files. Apparently, disk IO issues are more
12390 frequent than we hoped, and user reported >1 minute waiting
12393 * 'machine-id-commit' functionality has been merged into
12394 'machine-id-setup --commit'. The separate binary has been
12397 * The WorkingDirectory= directive in unit files may now be set
12398 to the special value '~'. In this case, the working
12399 directory is set to the home directory of the user
12400 configured in User=.
12402 * "machinectl shell" will now open the shell in the home
12403 directory of the selected user by default.
12405 * The CrashChVT= configuration file setting is renamed to
12406 CrashChangeVT=, following our usual logic of not
12407 abbreviating unnecessarily. The old directive is still
12408 supported for compat reasons. Also, this directive now takes
12409 an integer value between 1 and 63, or a boolean value. The
12410 formerly supported '-1' value for disabling stays around for
12413 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
12414 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
12415 RootDirectory= properties can now be set for transient
12418 * The systemd-analyze tool gained a new "set-log-target" verb
12419 to change the logging target the system manager logs to
12420 dynamically during runtime. This is similar to how
12421 "systemd-analyze set-log-level" already changes the log
12424 * In nspawn /sys is now mounted as tmpfs, with only a selected
12425 set of subdirectories mounted in from the real sysfs. This
12426 enhances security slightly, and is useful for ensuring user
12427 namespaces work correctly.
12429 * Support for USB FunctionFS activation has been added. This
12430 allows implementation of USB gadget services that are
12431 activated as soon as they are requested, so that they don't
12432 have to run continuously, similar to classic socket
12435 * The "systemctl exit" command now optionally takes an
12436 additional parameter that sets the exit code to return from
12437 the systemd manager when exiting. This is only relevant when
12438 running the systemd user instance, or when running the
12439 system instance in a container.
12441 * sd-bus gained the new API calls sd_bus_path_encode_many()
12442 and sd_bus_path_decode_many() that allow easy encoding and
12443 decoding of multiple identifier strings inside a D-Bus
12444 object path. Another new call sd_bus_default_flush_close()
12445 has been added to flush and close per-thread default
12448 * systemd-cgtop gained support for a -M/--machine= switch to
12449 show the control groups within a certain container only.
12451 * "systemctl kill" gained support for an optional --fail
12452 switch. If specified the requested operation will fail of no
12453 processes have been killed, because the unit had no
12454 processes attached, or similar.
12456 * A new systemd.crash_reboot=1 kernel command line option has
12457 been added that triggers a reboot after crashing. This can
12458 also be set through CrashReboot= in systemd.conf.
12460 * The RuntimeDirectory= setting now understands unit
12461 specifiers like %i or %f.
12463 * A new (still internal) library API sd-ipv4acd has been added,
12464 that implements address conflict detection for IPv4. It's
12465 based on code from sd-ipv4ll, and will be useful for
12466 detecting DHCP address conflicts.
12468 * File descriptors passed during socket activation may now be
12469 named. A new API sd_listen_fds_with_names() is added to
12470 access the names. The default names may be overridden,
12471 either in the .socket file using the FileDescriptorName=
12472 parameter, or by passing FDNAME= when storing the file
12473 descriptors using sd_notify().
12475 * systemd-networkd gained support for:
12477 - Setting the IPv6 Router Advertisement settings via
12478 IPv6AcceptRouterAdvertisements= in .network files.
12480 - Configuring the HelloTimeSec=, MaxAgeSec= and
12481 ForwardDelaySec= bridge parameters in .netdev files.
12483 - Configuring PreferredSource= for static routes in
12486 * The "ask-password" framework used to query for LUKS harddisk
12487 passwords or SSL passwords during boot gained support for
12488 caching passwords in the kernel keyring, if it is
12489 available. This makes sure that the user only has to type in
12490 a passphrase once if there are multiple objects to unlock
12491 with the same one. Previously, such password caching was
12492 available only when Plymouth was used; this moves the
12493 caching logic into the systemd codebase itself. The
12494 "systemd-ask-password" utility gained a new --keyname=
12495 switch to control which kernel keyring key to use for
12496 caching a password in. This functionality is also useful for
12497 enabling display managers such as gdm to automatically
12498 unlock the user's GNOME keyring if its passphrase, the
12499 user's password and the harddisk password are the same, if
12500 gdm-autologin is used.
12502 * When downloading tar or raw images using "machinectl
12503 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
12504 file is now also downloaded, if it is available and stored
12505 next to the image file.
12507 * Units of type ".socket" gained a new boolean setting
12508 Writable= which is only useful in conjunction with
12509 ListenSpecial=. If true, enables opening the specified
12510 special file in O_RDWR mode rather than O_RDONLY mode.
12512 * systemd-rfkill has been reworked to become a singleton
12513 service that is activated through /dev/rfkill on each rfkill
12514 state change and saves the settings to disk. This way,
12515 systemd-rfkill is now compatible with devices that exist
12516 only intermittendly, and even restores state if the previous
12517 system shutdown was abrupt rather than clean.
12519 * The journal daemon gained support for vacuuming old journal
12520 files controlled by the number of files that shall remain,
12521 in addition to the already existing control by size and by
12522 date. This is useful as journal interleaving performance
12523 degrades with too many separate journal files, and allows
12524 putting an effective limit on them. The new setting defaults
12525 to 100, but this may be changed by setting SystemMaxFiles=
12526 and RuntimeMaxFiles= in journald.conf. Also, the
12527 "journalctl" tool gained the new --vacuum-files= switch to
12528 manually vacuum journal files to leave only the specified
12529 number of files in place.
12531 * udev will now create /dev/disk/by-path links for ATA devices
12532 on kernels where that is supported.
12534 * Galician, Serbian, Turkish and Korean translations were added.
12536 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
12537 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
12538 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
12539 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
12540 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
12541 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
12542 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
12543 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
12544 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
12545 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
12546 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
12547 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
12548 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
12549 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
12550 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
12551 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
12552 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
12553 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
12555 — Berlin, 2015-10-07
12559 * The DHCP implementation of systemd-networkd gained a set of
12562 - The DHCP server now supports emitting DNS and NTP
12563 information. It may be enabled and configured via
12564 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
12565 and NTP information is enabled, but no servers are
12566 configured, the corresponding uplink information (if there
12567 is any) is propagated.
12569 - Server and client now support transmission and reception
12570 of timezone information. It can be configured via the
12571 newly introduced network options UseTimezone=,
12572 EmitTimezone=, and Timezone=. Transmission of timezone
12573 information is enabled between host and containers by
12574 default now: the container will change its local timezone
12575 to what the host has set.
12577 - Lease timeouts can now be configured via
12578 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
12580 - The DHCP server improved on the stability of
12581 leases. Clients are more likely to get the same lease
12582 information back, even if the server loses state.
12584 - The DHCP server supports two new configuration options to
12585 control the lease address pool metrics, PoolOffset= and
12588 * The encapsulation limit of tunnels in systemd-networkd may
12589 now be configured via 'EncapsulationLimit='. It allows
12590 modifying the maximum additional levels of encapsulation
12591 that are permitted to be prepended to a packet.
12593 * systemd now supports the concept of user buses replacing
12594 session buses, if used with dbus-1.10 (and enabled via dbus
12595 --enable-user-session). It previously only supported this on
12596 kdbus-enabled systems, and this release expands this to
12597 'dbus-daemon' systems.
12599 * systemd-networkd now supports predictable interface names
12600 for virtio devices.
12602 * systemd now optionally supports the new Linux kernel
12603 "unified" control group hierarchy. If enabled via the kernel
12604 command-line option 'systemd.unified_cgroup_hierarchy=1',
12605 systemd will try to mount the unified cgroup hierarchy
12606 directly on /sys/fs/cgroup. If not enabled, or not
12607 available, systemd will fall back to the legacy cgroup
12608 hierarchy setup, as before. Host system and containers can
12609 mix and match legacy and unified hierarchies as they
12610 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
12611 environment variable to individually select the hierarchy to
12612 use for executed containers. By default, nspawn will use the
12613 unified hierarchy for the containers if the host uses the
12614 unified hierarchy, and the legacy hierarchy otherwise.
12615 Please note that at this point the unified hierarchy is an
12616 experimental kernel feature and is likely to change in one
12617 of the next kernel releases. Therefore, it should not be
12618 enabled by default in downstream distributions yet. The
12619 minimum required kernel version for the unified hierarchy to
12620 work is 4.2. Note that when the unified hierarchy is used
12621 for the first time delegated access to controllers is
12622 safe. Because of this systemd-nspawn containers will get
12623 access to controllers now, as will systemd user
12624 sessions. This means containers and user sessions may now
12625 manage their own resources, partitioning up what the system
12628 * A new special scope unit "init.scope" has been introduced
12629 that encapsulates PID 1 of the system. It may be used to
12630 determine resource usage and enforce resource limits on PID
12631 1 itself. PID 1 hence moved out of the root of the control
12634 * The cgtop tool gained support for filtering out kernel
12635 threads when counting tasks in a control group. Also, the
12636 count of processes is now recursively summed up by
12637 default. Two options -k and --recursive= have been added to
12638 revert to old behaviour. The tool has also been updated to
12639 work correctly in containers now.
12641 * systemd-nspawn's --bind= and --bind-ro= options have been
12642 extended to allow creation of non-recursive bind mounts.
12644 * libsystemd gained two new calls sd_pid_get_cgroup() and
12645 sd_peer_get_cgroup() which return the control group path of
12646 a process or peer of a connected AF_UNIX socket. This
12647 function call is particularly useful when implementing
12648 delegated subtrees support in the control group hierarchy.
12650 * The "sd-event" event loop API of libsystemd now supports
12651 correct dequeuing of real-time signals, without losing
12654 * When systemd requests a polkit decision when managing units it
12655 will now add additional fields to the request, including unit
12656 name and desired operation. This enables more powerful polkit
12657 policies, that make decisions depending on these parameters.
12659 * nspawn learnt support for .nspawn settings files, that may
12660 accompany the image files or directories of containers, and
12661 may contain additional settings for the container. This is
12662 an alternative to configuring container parameters via the
12663 nspawn command line.
12665 Contributions from: Cristian Rodríguez, Daniel Mack, David
12666 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
12667 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
12668 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
12669 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
12670 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
12671 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
12672 Andersen, Tom Gundersen, Torstein Husebø
12674 — Berlin, 2015-09-08
12678 * machinectl gained a new verb 'shell' which opens a fresh
12679 shell on the target container or the host. It is similar to
12680 the existing 'login' command of machinectl, but spawns the
12681 shell directly without prompting for username or
12682 password. The pseudo machine '.host' now refers to the local
12683 host and is used by default. Hence, 'machinectl shell' can
12684 be used as replacement for 'su -' which spawns a session as
12685 a fresh systemd unit in a way that is fully isolated from
12686 the originating session.
12688 * systemd-networkd learned to cope with private-zone DHCP
12689 options and allows other programs to query the values.
12691 * SELinux access control when enabling/disabling units is no
12692 longer enforced with this release. The previous implementation
12693 was incorrect, and a new corrected implementation is not yet
12694 available. As unit file operations are still protected via
12695 polkit and D-Bus policy this is not a security problem. Yet,
12696 distributions which care about optimal SELinux support should
12697 probably not stabilize on this release.
12699 * sd-bus gained support for matches of type "arg0has=", that
12700 test for membership of strings in string arrays sent in bus
12703 * systemd-resolved now dumps the contents of its DNS and LLMNR
12704 caches to the logs on reception of the SIGUSR1 signal. This
12705 is useful to debug DNS behaviour.
12707 * The coredumpctl tool gained a new --directory= option to
12708 operate on journal files in a specific directory.
12710 * "systemctl reboot" and related commands gained a new
12711 "--message=" option which may be used to set a free-text
12712 wall message when shutting down or rebooting the
12713 system. This message is also logged, which is useful for
12714 figuring out the reason for a reboot or shutdown a
12717 * The "systemd-resolve-host" tool's -i switch now takes
12718 network interface numbers as alternative to interface names.
12720 * A new unit file setting for services has been introduced:
12721 UtmpMode= allows configuration of how precisely systemd
12722 handles utmp and wtmp entries for the service if this is
12723 enabled. This allows writing services that appear similar to
12724 user sessions in the output of the "w", "who", "last" and
12727 * systemd-resolved will now locally synthesize DNS resource
12728 records for the "localhost" and "gateway" domains as well as
12729 the local hostname. This should ensure that clients querying
12730 RRs via resolved will get similar results as those going via
12731 NSS, if nss-myhostname is enabled.
12733 Contributions from: Alastair Hughes, Alex Crawford, Daniel
12734 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
12735 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
12736 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
12737 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
12738 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
12739 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
12740 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
12741 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
12742 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
12743 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
12744 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
12746 — Berlin, 2015-08-27
12750 * The systemd-efi-boot-generator functionality was merged into
12751 systemd-gpt-auto-generator.
12753 * systemd-networkd now supports Group Policy for vxlan
12754 devices. It can be enabled via the new boolean configuration
12755 option called 'GroupPolicyExtension='.
12757 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
12758 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
12759 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
12761 — Berlin, 2015-07-31
12765 * The python-systemd code has been removed from the systemd repository.
12766 A new repository has been created which accommodates the code from
12767 now on, and we kindly ask distributions to create a separate package
12768 for this: https://github.com/systemd/python-systemd
12770 * The systemd daemon will now reload its main configuration
12771 (/etc/systemd/system.conf) on daemon-reload.
12773 * sd-dhcp now exposes vendor specific extensions via
12774 sd_dhcp_lease_get_vendor_specific().
12776 * systemd-networkd gained a number of new configuration options.
12778 - A new boolean configuration option for TAP devices called
12779 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
12780 device, thus allowing to send and receive GSO packets.
12782 - A new tunnel configuration option called 'CopyDSCP='.
12783 If enabled, the DSCP field of ip6 tunnels is copied into the
12784 decapsulated packet.
12786 - A set of boolean bridge configuration options were added.
12787 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
12788 and 'UnicastFlood=' are now parsed by networkd and applied to the
12789 respective bridge link device via the respective IFLA_BRPORT_*
12792 - A new string configuration option to override the hostname sent
12793 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
12794 is true, networkd will use the configured hostname instead of the
12795 system hostname when sending DHCP requests.
12797 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
12798 networkd will configure the IPv6 flow-label of the tunnel device
12799 according to RFC2460.
12801 - The 'macvtap' virtual network devices are now supported, similar to
12802 the already supported 'macvlan' devices.
12804 * systemd-resolved now implements RFC5452 to improve resilience against
12805 cache poisoning. Additionally, source port randomization is enabled
12806 by default to further protect against DNS spoofing attacks.
12808 * nss-mymachines now supports translating UIDs and GIDs of running
12809 containers with user-namespaces enabled. If a container 'foo'
12810 translates a host uid 'UID' to the container uid 'TUID', then
12811 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
12812 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
12813 mapped as 'vg-foo-TGID'.
12815 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
12816 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
12817 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
12818 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
12819 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
12820 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
12821 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
12822 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
12823 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
12824 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
12826 — Berlin, 2015-07-29
12830 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
12831 There are no known issues with current sysfs, and udev does not need
12832 or should be used to work around such bugs.
12834 * udev does no longer enable USB HID power management. Several reports
12835 indicate, that some devices cannot handle that setting.
12837 * The udev accelerometer helper was removed. The functionality
12838 is now fully included in iio-sensor-proxy. But this means,
12839 older iio-sensor-proxy versions will no longer provide
12840 accelerometer/orientation data with this systemd version.
12841 Please upgrade iio-sensor-proxy to version 1.0.
12843 * networkd gained a new configuration option IPv6PrivacyExtensions=
12844 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
12845 for Stateless Address") on selected networks.
12847 * For the sake of fewer build-time dependencies and less code in the
12848 main repository, the python bindings are about to be removed in the
12849 next release. A new repository has been created which accommodates
12850 the code from now on, and we kindly ask distributions to create a
12851 separate package for this. The removal will take place in v223.
12853 https://github.com/systemd/python-systemd
12855 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
12856 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
12857 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
12858 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
12859 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
12860 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
12861 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
12862 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
12863 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
12864 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
12866 — Berlin, 2015-07-07
12870 * The sd-bus.h and sd-event.h APIs have now been declared
12871 stable and have been added to the official interface of
12872 libsystemd.so. sd-bus implements an alternative D-Bus client
12873 library, that is relatively easy to use, very efficient and
12874 supports both classic D-Bus as well as kdbus as transport
12875 backend. sd-event is a generic event loop abstraction that
12876 is built around Linux epoll, but adds features such as event
12877 prioritization or efficient timer handling. Both APIs are good
12878 choices for C programs looking for a bus and/or event loop
12879 implementation that is minimal and does not have to be
12880 portable to other kernels.
12882 * kdbus support is no longer compile-time optional. It is now
12883 always built-in. However, it can still be disabled at
12884 runtime using the kdbus=0 kernel command line setting, and
12885 that setting may be changed to default to off, by specifying
12886 --disable-kdbus at build-time. Note though that the kernel
12887 command line setting has no effect if the kdbus.ko kernel
12888 module is not installed, in which case kdbus is (obviously)
12889 also disabled. We encourage all downstream distributions to
12890 begin testing kdbus by adding it to the kernel images in the
12891 development distributions, and leaving kdbus support in
12894 * The minimal required util-linux version has been bumped to
12897 * Support for chkconfig (--enable-chkconfig) was removed in
12898 favor of calling an abstraction tool
12899 /lib/systemd/systemd-sysv-install. This needs to be
12900 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
12901 in README for details.
12903 * If there's a systemd unit and a SysV init script for the
12904 same service name, and the user executes "systemctl enable"
12905 for it (or a related call), then this will now enable both
12906 (or execute the related operation on both), not just the
12909 * The libudev API documentation has been converted from gtkdoc
12912 * gudev has been removed from the systemd tree, it is now an
12915 * The systemd-cgtop tool learnt a new --raw switch to generate
12916 "raw" (machine parsable) output.
12918 * networkd's IPForwarding= .network file setting learnt the
12919 new setting "kernel", which ensures that networkd does not
12920 change the IP forwarding sysctl from the default kernel
12923 * The systemd-logind bus API now exposes a new boolean
12924 property "Docked" that reports whether logind considers the
12925 system "docked", i.e. connected to a docking station or not.
12927 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
12928 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
12929 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
12930 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
12931 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
12932 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
12933 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
12934 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
12935 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
12936 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
12937 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
12938 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
12939 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
12940 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
12941 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
12942 Fink, Zbigniew Jędrzejewski-Szmek
12944 — Berlin, 2015-06-19
12948 * The gudev library has been extracted into a separate repository
12949 available at: https://git.gnome.org/browse/libgudev/
12950 It is now managed as part of the Gnome project. Distributions
12951 are recommended to pass --disable-gudev to systemd and use
12952 gudev from the Gnome project instead. gudev is still included
12953 in systemd, for now. It will be removed soon, though. Please
12954 also see the announcement-thread on systemd-devel:
12955 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
12957 * systemd now exposes a CPUUsageNSec= property for each
12958 service unit on the bus, that contains the overall consumed
12959 CPU time of a service (the sum of what each process of the
12960 service consumed). This value is only available if
12961 CPUAccounting= is turned on for a service, and is then shown
12962 in the "systemctl status" output.
12964 * Support for configuring alternative mappings of the old SysV
12965 runlevels to systemd targets has been removed. They are now
12966 hardcoded in a way that runlevels 2, 3, 4 all map to
12967 multi-user.target and 5 to graphical.target (which
12968 previously was already the default behaviour).
12970 * The auto-mounter logic gained support for mount point
12971 expiry, using a new TimeoutIdleSec= setting in .automount
12972 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
12974 * The EFI System Partition (ESP) as mounted to /boot by
12975 systemd-efi-boot-generator will now be unmounted
12976 automatically after 2 minutes of not being used. This should
12977 minimize the risk of ESP corruptions.
12979 * New /etc/fstab options x-systemd.requires= and
12980 x-systemd.requires-mounts-for= are now supported to express
12981 additional dependencies for mounts. This is useful for
12982 journaling file systems that support external journal
12983 devices or overlay file systems that require underlying file
12984 systems to be mounted.
12986 * systemd does not support direct live-upgrades (via systemctl
12987 daemon-reexec) from versions older than v44 anymore. As no
12988 distribution we are aware of shipped such old versions in a
12989 stable release this should not be problematic.
12991 * When systemd forks off a new per-connection service instance
12992 it will now set the $REMOTE_ADDR environment variable to the
12993 remote IP address, and $REMOTE_PORT environment variable to
12994 the remote IP port. This behaviour is similar to the
12995 corresponding environment variables defined by CGI.
12997 * systemd-networkd gained support for uplink failure
12998 detection. The BindCarrier= option allows binding interface
12999 configuration dynamically to the link sense of other
13000 interfaces. This is useful to achieve behaviour like in
13003 * systemd-networkd gained support for configuring the DHCP
13004 client identifier to use when requesting leases.
13006 * systemd-networkd now has a per-network UseNTP= option to
13007 configure whether NTP server information acquired via DHCP
13008 is passed on to services like systemd-timesyncd.
13010 * systemd-networkd gained support for vti6 tunnels.
13012 * Note that systemd-networkd manages the sysctl variable
13013 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
13014 it is configured for since v219. The variable controls IP
13015 forwarding, and is a per-interface alternative to the global
13016 /proc/sys/net/ipv[46]/ip_forward. This setting is
13017 configurable in the IPForward= option, which defaults to
13018 "no". This means if networkd is used for an interface it is
13019 no longer sufficient to set the global sysctl option to turn
13020 on IP forwarding! Instead, the .network file option
13021 IPForward= needs to be turned on! Note that the
13022 implementation of this behaviour was broken in v219 and has
13023 been fixed in v220.
13025 * Many bonding and vxlan options are now configurable in
13028 * systemd-nspawn gained a new --property= setting to set unit
13029 properties for the container scope. This is useful for
13030 setting resource parameters (e.g. "CPUShares=500") on
13031 containers started from the command line.
13033 * systemd-nspawn gained a new --private-users= switch to make
13034 use of user namespacing available on recent Linux kernels.
13036 * systemd-nspawn may now be called as part of a shell pipeline
13037 in which case the pipes used for stdin and stdout are passed
13038 directly to the process invoked in the container, without
13039 indirection via a pseudo tty.
13041 * systemd-nspawn gained a new switch to control the UNIX
13042 signal to use when killing the init process of the container
13043 when shutting down.
13045 * systemd-nspawn gained a new --overlay= switch for mounting
13046 overlay file systems into the container using the new kernel
13049 * When a container image is imported via systemd-importd and
13050 the host file system is not btrfs, a loopback block device
13051 file is created in /var/lib/machines.raw with a btrfs file
13052 system inside. It is then mounted to /var/lib/machines to
13053 enable btrfs features for container management. The loopback
13054 file and btrfs file system is grown as needed when container
13055 images are imported via systemd-importd.
13057 * systemd-machined/systemd-importd gained support for btrfs
13058 quota, to enforce container disk space limits on disk. This
13059 is exposed in "machinectl set-limit".
13061 * systemd-importd now can import containers from local .tar,
13062 .raw and .qcow2 images, and export them to .tar and .raw. It
13063 can also import dkr v2 images now from the network (on top
13066 * systemd-importd gained support for verifying downloaded
13067 images with gpg2 (previously only gpg1 was supported).
13069 * systemd-machined, systemd-logind, systemd: most bus calls are
13070 now accessible to unprivileged processes via polkit. Also,
13071 systemd-logind will now allow users to kill their own sessions
13072 without further privileges or authorization.
13074 * systemd-shutdownd has been removed. This service was
13075 previously responsible for implementing scheduled shutdowns
13076 as exposed in /usr/bin/shutdown's time parameter. This
13077 functionality has now been moved into systemd-logind and is
13078 accessible via a bus interface.
13080 * "systemctl reboot" gained a new switch --firmware-setup that
13081 can be used to reboot into the EFI firmware setup, if that
13082 is available. systemd-logind now exposes an API on the bus
13083 to trigger such reboots, in case graphical desktop UIs want
13084 to cover this functionality.
13086 * "systemctl enable", "systemctl disable" and "systemctl mask"
13087 now support a new "--now" switch. If specified the units
13088 that are enabled will also be started, and the ones
13089 disabled/masked also stopped.
13091 * The Gummiboot EFI boot loader tool has been merged into
13092 systemd, and renamed to "systemd-boot". The bootctl tool has been
13093 updated to support systemd-boot.
13095 * An EFI kernel stub has been added that may be used to create
13096 kernel EFI binaries that contain not only the actual kernel,
13097 but also an initrd, boot splash, command line and OS release
13098 information. This combined binary can then be signed as a
13099 single image, so that the firmware can verify it all in one
13100 step. systemd-boot has special support for EFI binaries created
13101 like this and can extract OS release information from them
13102 and show them in the boot menu. This functionality is useful
13103 to implement cryptographically verified boot schemes.
13105 * Optional support has been added to systemd-fsck to pass
13106 fsck's progress report to an AF_UNIX socket in the file
13109 * udev will no longer create device symlinks for all block devices by
13110 default. A deny list for excluding special block devices from this
13111 logic has been turned into an allow list that requires picking block
13112 devices explicitly that require device symlinks.
13114 * A new (currently still internal) API sd-device.h has been
13115 added to libsystemd. This modernized API is supposed to
13116 replace libudev eventually. In fact, already much of libudev
13117 is now just a wrapper around sd-device.h.
13119 * A new hwdb database for storing metadata about pointing
13120 stick devices has been added.
13122 * systemd-tmpfiles gained support for setting file attributes
13123 similar to the "chattr" tool with new 'h' and 'H' lines.
13125 * systemd-journald will no longer unconditionally set the
13126 btrfs NOCOW flag on new journal files. This is instead done
13127 with tmpfiles snippet using the new 'h' line type. This
13128 allows easy disabling of this logic, by masking the
13129 journal-nocow.conf tmpfiles file.
13131 * systemd-journald will now translate audit message types to
13132 human readable identifiers when writing them to the
13133 journal. This should improve readability of audit messages.
13135 * The LUKS logic gained support for the offset= and skip=
13136 options in /etc/crypttab, as previously implemented by
13139 * /usr/lib/os-release gained a new optional field VARIANT= for
13140 distributions that support multiple variants (such as a
13141 desktop edition, a server edition, …)
13143 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
13144 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
13145 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
13146 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
13147 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
13148 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
13149 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
13150 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
13151 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
13152 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
13153 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
13154 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
13155 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
13156 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
13157 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
13158 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
13159 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
13160 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
13161 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
13162 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
13163 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
13164 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
13165 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
13166 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
13167 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
13168 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
13169 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
13171 — Berlin, 2015-05-22
13175 * Introduce a new API "sd-hwdb.h" for querying the hardware
13176 metadata database. With this minimal interface one can query
13177 and enumerate the udev hwdb, decoupled from the old libudev
13178 library. libudev's interface for this is now only a wrapper
13179 around sd-hwdb. A new tool systemd-hwdb has been added to
13180 interface with and update the database.
13182 * When any of systemd's tools copies files (for example due to
13183 tmpfiles' C lines) a btrfs reflink will attempted first,
13184 before bytewise copying is done.
13186 * systemd-nspawn gained a new --ephemeral switch. When
13187 specified a btrfs snapshot is taken of the container's root
13188 directory, and immediately removed when the container
13189 terminates again. Thus, a container can be started whose
13190 changes never alter the container's root directory, and are
13191 lost on container termination. This switch can also be used
13192 for starting a container off the root file system of the
13193 host without affecting the host OS. This switch is only
13194 available on btrfs file systems.
13196 * systemd-nspawn gained a new --template= switch. It takes the
13197 path to a container tree to use as template for the tree
13198 specified via --directory=, should that directory be
13199 missing. This allows instantiating containers dynamically,
13200 on first run. This switch is only available on btrfs file
13203 * When a .mount unit refers to a mount point on which multiple
13204 mounts are stacked, and the .mount unit is stopped all of
13205 the stacked mount points will now be unmounted until no
13206 mount point remains.
13208 * systemd now has an explicit notion of supported and
13209 unsupported unit types. Jobs enqueued for unsupported unit
13210 types will now fail with an "unsupported" error code. More
13211 specifically .swap, .automount and .device units are not
13212 supported in containers, .busname units are not supported on
13213 non-kdbus systems. .swap and .automount are also not
13214 supported if their respective kernel compile time options
13217 * machinectl gained support for two new "copy-from" and
13218 "copy-to" commands for copying files from a running
13219 container to the host or vice versa.
13221 * machinectl gained support for a new "bind" command to bind
13222 mount host directories into local containers. This is
13223 currently only supported for nspawn containers.
13225 * networkd gained support for configuring bridge forwarding
13226 database entries (fdb) from .network files.
13228 * A new tiny daemon "systemd-importd" has been added that can
13229 download container images in tar, raw, qcow2 or dkr formats,
13230 and make them available locally in /var/lib/machines, so
13231 that they can run as nspawn containers. The daemon can GPG
13232 verify the downloads (not supported for dkr, since it has no
13233 provisions for verifying downloads). It will transparently
13234 decompress bz2, xz, gzip compressed downloads if necessary,
13235 and restore sparse files on disk. The daemon uses privilege
13236 separation to ensure the actual download logic runs with
13237 fewer privileges than the daemon itself. machinectl has
13238 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
13239 make the functionality of importd available to the
13240 user. With this in place the Fedora and Ubuntu "Cloud"
13241 images can be downloaded and booted as containers unmodified
13242 (the Fedora images lack the appropriate GPG signature files
13243 currently, so they cannot be verified, but this will change
13244 soon, hopefully). Note that downloading images is currently
13245 only fully supported on btrfs.
13247 * machinectl is now able to list container images found in
13248 /var/lib/machines, along with some metadata about sizes of
13249 disk and similar. If the directory is located on btrfs and
13250 quota is enabled, this includes quota display. A new command
13251 "image-status" has been added that shows additional
13252 information about images.
13254 * machinectl is now able to clone container images
13255 efficiently, if the underlying file system (btrfs) supports
13256 it, with the new "machinectl clone" command. It also
13257 gained commands for renaming and removing images, as well as
13258 marking them read-only or read-write (supported also on
13259 legacy file systems).
13261 * networkd gained support for collecting LLDP network
13262 announcements, from hardware that supports this. This is
13263 shown in networkctl output.
13265 * systemd-run gained support for a new -t (--pty) switch for
13266 invoking a binary on a pty whose input and output is
13267 connected to the invoking terminal. This allows executing
13268 processes as system services while interactively
13269 communicating with them via the terminal. Most interestingly
13270 this is supported across container boundaries. Invoking
13271 "systemd-run -t /bin/bash" is an alternative to running a
13272 full login session, the difference being that the former
13273 will not register a session, nor go through the PAM session
13276 * tmpfiles gained support for a new "v" line type for creating
13277 btrfs subvolumes. If the underlying file system is a legacy
13278 file system, this automatically degrades to creating a
13279 normal directory. Among others /var/lib/machines is now
13280 created like this at boot, should it be missing.
13282 * The directory /var/lib/containers/ has been deprecated and
13283 been replaced by /var/lib/machines. The term "machines" has
13284 been used in the systemd context as generic term for both
13285 VMs and containers, and hence appears more appropriate for
13286 this, as the directory can also contain raw images bootable
13289 * systemd-nspawn when invoked with -M but without --directory=
13290 or --image= is now capable of searching for the container
13291 root directory, subvolume or disk image automatically, in
13292 /var/lib/machines. systemd-nspawn@.service has been updated
13293 to make use of this, thus allowing it to be used for raw
13296 * A new machines.target unit has been introduced that is
13297 supposed to group all containers/VMs invoked as services on
13298 the system. systemd-nspawn@.service has been updated to
13299 integrate with that.
13301 * machinectl gained a new "start" command, for invoking a
13302 container as a service. "machinectl start foo" is mostly
13303 equivalent to "systemctl start systemd-nspawn@foo.service",
13304 but handles escaping in a nicer way.
13306 * systemd-nspawn will now mount most of the cgroupfs tree
13307 read-only into each container, with the exception of the
13308 container's own subtree in the name=systemd hierarchy.
13310 * journald now sets the special FS_NOCOW file flag for its
13311 journal files. This should improve performance on btrfs, by
13312 avoiding heavy fragmentation when journald's write-pattern
13313 is used on COW file systems. It degrades btrfs' data
13314 integrity guarantees for the files to the same levels as for
13315 ext3/ext4 however. This should be OK though as journald does
13316 its own data integrity checks and all its objects are
13317 checksummed on disk. Also, journald should handle btrfs disk
13318 full events a lot more gracefully now, by processing SIGBUS
13319 errors, and not relying on fallocate() anymore.
13321 * When journald detects that journal files it is writing to
13322 have been deleted it will immediately start new journal
13325 * systemd now provides a way to store file descriptors
13326 per-service in PID 1. This is useful for daemons to ensure
13327 that fds they require are not lost during a daemon
13328 restart. The fds are passed to the daemon on the next
13329 invocation in the same way socket activation fds are
13330 passed. This is now used by journald to ensure that the
13331 various sockets connected to all the system's stdout/stderr
13332 are not lost when journald is restarted. File descriptors
13333 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
13334 an extension to sd_notify(). Note that a limit is enforced
13335 on the number of fds a service can store in PID 1, and it
13336 defaults to 0, so that no fds may be stored, unless this is
13337 explicitly turned on.
13339 * The default TERM variable to use for units connected to a
13340 terminal, when no other value is explicitly is set is now
13341 vt220 rather than vt102. This should be fairly safe still,
13342 but allows PgUp/PgDn work.
13344 * The /etc/crypttab option header= as known from Debian is now
13347 * "loginctl user-status" and "loginctl session-status" will
13348 now show the last 10 lines of log messages of the
13349 user/session following the status output. Similar,
13350 "machinectl status" will show the last 10 log lines
13351 associated with a virtual machine or container
13352 service. (Note that this is usually not the log messages
13353 done in the VM/container itself, but simply what the
13354 container manager logs. For nspawn this includes all console
13357 * "loginctl session-status" without further argument will now
13358 show the status of the session of the caller. Similar,
13359 "lock-session", "unlock-session", "activate",
13360 "enable-linger", "disable-linger" may now be called without
13361 session/user parameter in which case they apply to the
13362 caller's session/user.
13364 * An X11 session scriptlet is now shipped that uploads
13365 $DISPLAY and $XAUTHORITY into the environment of the systemd
13366 --user daemon if a session begins. This should improve
13367 compatibility with X11 enabled applications run as systemd
13370 * Generators are now subject to masking via /etc and /run, the
13371 same way as unit files.
13373 * networkd .network files gained support for configuring
13374 per-link IPv4/IPv6 packet forwarding as well as IPv4
13375 masquerading. This is by default turned on for veth links to
13376 containers, as registered by systemd-nspawn. This means that
13377 nspawn containers run with --network-veth will now get
13378 automatic routed access to the host's networks without any
13379 further configuration or setup, as long as networkd runs on
13382 * systemd-nspawn gained the --port= (-p) switch to expose TCP
13383 or UDP posts of a container on the host. With this in place
13384 it is possible to run containers with private veth links
13385 (--network-veth), and have their functionality exposed on
13386 the host as if their services were running directly on the
13389 * systemd-nspawn's --network-veth switch now gained a short
13390 version "-n", since with the changes above it is now truly
13391 useful out-of-the-box. The systemd-nspawn@.service has been
13392 updated to make use of it too by default.
13394 * systemd-nspawn will now maintain a per-image R/W lock, to
13395 ensure that the same image is not started more than once
13396 writable. (It's OK to run an image multiple times
13397 simultaneously in read-only mode.)
13399 * systemd-nspawn's --image= option is now capable of
13400 dissecting and booting MBR and GPT disk images that contain
13401 only a single active Linux partition. Previously it
13402 supported only GPT disk images with proper GPT type
13403 IDs. This allows running cloud images from major
13404 distributions directly with systemd-nspawn, without
13407 * In addition to collecting mouse dpi data in the udev
13408 hardware database, there's now support for collecting angle
13409 information for mouse scroll wheels. The database is
13410 supposed to guarantee similar scrolling behavior on mice
13411 that it knows about. There's also support for collecting
13412 information about Touchpad types.
13414 * udev's input_id built-in will now also collect touch screen
13415 dimension data and attach it to probed devices.
13417 * /etc/os-release gained support for a Distribution Privacy
13420 * networkd gained support for creating "ipvlan", "gretap",
13421 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
13423 * systemd-tmpfiles gained support for "a" lines for setting
13426 * systemd-nspawn will now mount /tmp in the container to
13427 tmpfs, automatically.
13429 * systemd now exposes the memory.usage_in_bytes cgroup
13430 attribute and shows it for each service in the "systemctl
13431 status" output, if available.
13433 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
13434 immediate reboot is triggered. This useful if shutdown is
13435 hung and is unable to complete, to expedite the
13436 operation. Note that this kind of reboot will still unmount
13437 all file systems, and hence should not result in fsck being
13438 run on next reboot.
13440 * A .device unit for an optical block device will now be
13441 considered active only when a medium is in the drive. Also,
13442 mount units are now bound to their backing devices thus
13443 triggering automatic unmounting when devices become
13444 unavailable. With this in place systemd will now
13445 automatically unmount left-over mounts when a CD-ROM is
13446 ejected or a USB stick is yanked from the system.
13448 * networkd-wait-online now has support for waiting for
13449 specific interfaces only (with globbing), and for giving up
13450 after a configurable timeout.
13452 * networkd now exits when idle. It will be automatically
13453 restarted as soon as interfaces show up, are removed or
13454 change state. networkd will stay around as long as there is
13455 at least one DHCP state machine or similar around, that keep
13458 * networkd may now configure IPv6 link-local addressing in
13459 addition to IPv4 link-local addressing.
13461 * The IPv6 "token" for use in SLAAC may now be configured for
13462 each .network interface in networkd.
13464 * Routes configured with networkd may now be assigned a scope
13467 * networkd's [Match] sections now support globbing and lists
13468 of multiple space-separated matches per item.
13470 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
13471 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
13472 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
13473 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
13474 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
13475 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
13476 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
13477 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
13478 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
13479 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
13480 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
13481 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
13482 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
13483 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
13484 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
13485 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
13486 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
13487 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
13488 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
13489 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
13490 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
13491 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
13492 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
13493 Hoffmann, Zbigniew Jędrzejewski-Szmek
13495 — Berlin, 2015-02-16
13499 * When querying unit file enablement status (for example via
13500 "systemctl is-enabled"), a new state "indirect" is now known
13501 which indicates that a unit might not be enabled itself, but
13502 another unit listed in its Also= setting might be.
13504 * Similar to the various existing ConditionXYZ= settings for
13505 units, there are now matching AssertXYZ= settings. While
13506 failing conditions cause a unit to be skipped, but its job
13507 to succeed, failing assertions declared like this will cause
13508 a unit start operation and its job to fail.
13510 * hostnamed now knows a new chassis type "embedded".
13512 * systemctl gained a new "edit" command. When used on a unit
13513 file, this allows extending unit files with .d/ drop-in
13514 configuration snippets or editing the full file (after
13515 copying it from /usr/lib to /etc). This will invoke the
13516 user's editor (as configured with $EDITOR), and reload the
13517 modified configuration after editing.
13519 * "systemctl status" now shows the suggested enablement state
13520 for a unit, as declared in the (usually vendor-supplied)
13521 system preset files.
13523 * nss-myhostname will now resolve the single-label hostname
13524 "gateway" to the locally configured default IP routing
13525 gateways, ordered by their metrics. This assigns a stable
13526 name to the used gateways, regardless which ones are
13527 currently configured. Note that the name will only be
13528 resolved after all other name sources (if nss-myhostname is
13529 configured properly) and should hence not negatively impact
13530 systems that use the single-label hostname "gateway" in
13533 * systemd-inhibit now allows filtering by mode when listing
13536 * Scope and service units gained a new "Delegate" boolean
13537 property, which, when set, allows processes running inside the
13538 unit to further partition resources. This is primarily
13539 useful for systemd user instances as well as container
13542 * journald will now pick up audit messages directly from
13543 the kernel, and log them like any other log message. The
13544 audit fields are split up and fully indexed. This means that
13545 journalctl in many ways is now a (nicer!) alternative to
13546 ausearch, the traditional audit client. Note that this
13547 implements only a minimal audit client. If you want the
13548 special audit modes like reboot-on-log-overflow, please use
13549 the traditional auditd instead, which can be used in
13550 parallel to journald.
13552 * The ConditionSecurity= unit file option now understands the
13553 special string "audit" to check whether auditing is
13556 * journalctl gained two new commands --vacuum-size= and
13557 --vacuum-time= to delete old journal files until the
13558 remaining ones take up no more than the specified size on disk,
13559 or are not older than the specified time.
13561 * A new, native PPPoE library has been added to sd-network,
13562 systemd's library of light-weight networking protocols. This
13563 library will be used in a future version of networkd to
13564 enable PPPoE communication without an external pppd daemon.
13566 * The busctl tool now understands a new "capture" verb that
13567 works similar to "monitor", but writes a packet capture
13568 trace to STDOUT that can be redirected to a file which is
13569 compatible with libcap's capture file format. This can then
13570 be loaded in Wireshark and similar tools to inspect bus
13573 * The busctl tool now understands a new "tree" verb that shows
13574 the object trees of a specific service on the bus, or of all
13577 * The busctl tool now understands a new "introspect" verb that
13578 shows all interfaces and members of objects on the bus,
13579 including their signature and values. This is particularly
13580 useful to get more information about bus objects shown by
13581 the new "busctl tree" command.
13583 * The busctl tool now understands new verbs "call",
13584 "set-property" and "get-property" for invoking bus method
13585 calls, setting and getting bus object properties in a
13588 * busctl gained a new --augment-creds= argument that controls
13589 whether the tool shall augment credential information it
13590 gets from the bus with data from /proc, in a possibly
13593 * nspawn's --link-journal= switch gained two new values
13594 "try-guest" and "try-host" that work like "guest" and
13595 "host", but do not fail if the host has no persistent
13596 journaling enabled. -j is now equivalent to
13597 --link-journal=try-guest.
13599 * macvlan network devices created by nspawn will now have
13600 stable MAC addresses.
13602 * A new SmackProcessLabel= unit setting has been added, which
13603 controls the SMACK security label processes forked off by
13604 the respective unit shall use.
13606 * If compiled with --enable-xkbcommon, systemd-localed will
13607 verify x11 keymap settings by compiling the given keymap. It
13608 will spew out warnings if the compilation fails. This
13609 requires libxkbcommon to be installed.
13611 * When a coredump is collected, a larger number of metadata
13612 fields is now collected and included in the journal records
13613 created for it. More specifically, control group membership,
13614 environment variables, memory maps, working directory,
13615 chroot directory, /proc/$PID/status, and a list of open file
13616 descriptors is now stored in the log entry.
13618 * The udev hwdb now contains DPI information for mice. For
13621 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
13623 * All systemd programs that read standalone configuration
13624 files in /etc now also support a corresponding series of
13625 .conf.d configuration directories in /etc/, /run/,
13626 /usr/local/lib/, /usr/lib/, and (if configured with
13627 --enable-split-usr) /lib/. In particular, the following
13628 configuration files now have corresponding configuration
13629 directories: system.conf user.conf, logind.conf,
13630 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
13631 resolved.conf, timesyncd.conf, journal-remote.conf, and
13632 journal-upload.conf. Note that distributions should use the
13633 configuration directories in /usr/lib/; the directories in
13634 /etc/ are reserved for the system administrator.
13636 * systemd-rfkill will no longer take the rfkill device name
13637 into account when storing rfkill state on disk, as the name
13638 might be dynamically assigned and not stable. Instead, the
13639 ID_PATH udev variable combined with the rfkill type (wlan,
13640 bluetooth, …) is used.
13642 * A new service systemd-machine-id-commit.service has been
13643 added. When used on systems where /etc is read-only during
13644 boot, and /etc/machine-id is not initialized (but an empty
13645 file), this service will copy the temporary machine ID
13646 created as replacement into /etc after the system is fully
13647 booted up. This is useful for systems that are freshly
13648 installed with a non-initialized machine ID, but should get
13649 a fixed machine ID for subsequent boots.
13651 * networkd's .netdev files now provide a large set of
13652 configuration parameters for VXLAN devices. Similarly, the
13653 bridge port cost parameter is now configurable in .network
13654 files. There's also new support for configuring IP source
13655 routing. networkd .link files gained support for a new
13656 OriginalName= match that is useful to match against the
13657 original interface name the kernel assigned. .network files
13658 may include MTU= and MACAddress= fields for altering the MTU
13659 and MAC address while being connected to a specific network
13662 * The LUKS logic gained supported for configuring
13663 UUID-specific key files. There's also new support for naming
13664 LUKS device from the kernel command line, using the new
13665 luks.name= argument.
13667 * Timer units may now be transiently created via the bus API
13668 (this was previously already available for scope and service
13669 units). In addition it is now possible to create multiple
13670 transient units at the same time with a single bus call. The
13671 "systemd-run" tool has been updated to make use of this for
13672 running commands on a specified time, in at(1)-style.
13674 * tmpfiles gained support for "t" lines, for assigning
13675 extended attributes to files. Among other uses this may be
13676 used to assign SMACK labels to files.
13678 Contributions from: Alin Rauta, Alison Chaiken, Andrej
13679 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
13680 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
13681 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
13682 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
13683 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
13684 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
13685 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
13686 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
13687 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
13688 Michael Biebl, Michael Chapman, Michael Marineau, Michal
13689 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
13690 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
13691 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
13692 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
13693 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
13694 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
13695 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
13697 — Berlin, 2014-12-10
13701 * journalctl gained the new options -t/--identifier= to match
13702 on the syslog identifier (aka "tag"), as well as --utc to
13703 show log timestamps in the UTC timezone. journalctl now also
13704 accepts -n/--lines=all to disable line capping in a pager.
13706 * journalctl gained a new switch, --flush, that synchronously
13707 flushes logs from /run/log/journal to /var/log/journal if
13708 persistent storage is enabled. systemd-journal-flush.service
13709 now waits until the operation is complete.
13711 * Services can notify the manager before they start a reload
13712 (by sending RELOADING=1) or shutdown (by sending
13713 STOPPING=1). This allows the manager to track and show the
13714 internal state of daemons and closes a race condition when
13715 the process is still running but has closed its D-Bus
13718 * Services with Type=oneshot do not have to have any ExecStart
13721 * User units are now loaded also from
13722 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
13723 /run/systemd/user directory that was already previously
13724 supported, but is under the control of the user.
13726 * Job timeouts (i.e. timeouts on the time a job that is
13727 queued stays in the run queue) can now optionally result in
13728 immediate reboot or power-off actions (JobTimeoutAction= and
13729 JobTimeoutRebootArgument=). This is useful on ".target"
13730 units, to limit the maximum time a target remains
13731 undispatched in the run queue, and to trigger an emergency
13732 operation in such a case. This is now used by default to
13733 turn off the system if boot-up (as defined by everything in
13734 basic.target) hangs and does not complete for at least
13735 15min. Also, if power-off or reboot hang for at least 30min
13736 an immediate power-off/reboot operation is triggered. This
13737 functionality is particularly useful to increase reliability
13738 on embedded devices, but also on laptops which might
13739 accidentally get powered on when carried in a backpack and
13740 whose boot stays stuck in a hard disk encryption passphrase
13743 * systemd-logind can be configured to also handle lid switch
13744 events even when the machine is docked or multiple displays
13745 are attached (HandleLidSwitchDocked= option).
13747 * A helper binary and a service have been added which can be
13748 used to resume from hibernation in the initramfs. A
13749 generator will parse the resume= option on the kernel
13750 command line to trigger resume.
13752 * A user console daemon systemd-consoled has been
13753 added. Currently, it is a preview, and will so far open a
13754 single terminal on each session of the user marked as
13755 Desktop=systemd-console.
13757 * Route metrics can be specified for DHCP routes added by
13760 * The SELinux context of socket-activated services can be set
13761 from the information provided by the networking stack
13762 (SELinuxContextFromNet= option).
13764 * Userspace firmware loading support has been removed and
13765 the minimum supported kernel version is thus bumped to 3.7.
13767 * Timeout for udev workers has been increased from 1 to 3
13768 minutes, but a warning will be printed after 1 minute to
13769 help diagnose kernel modules that take a long time to load.
13771 * Udev rules can now remove tags on devices with TAG-="foobar".
13773 * systemd's readahead implementation has been removed. In many
13774 circumstances it didn't give expected benefits even for
13775 rotational disk drives and was becoming less relevant in the
13776 age of SSDs. As none of the developers has been using
13777 rotating media anymore, and nobody stepped up to actively
13778 maintain this component of systemd it has now been removed.
13780 * Swap units can use Options= to specify discard options.
13781 Discard options specified for swaps in /etc/fstab are now
13784 * Docker containers are now detected as a separate type of
13787 * The Password Agent protocol gained support for queries where
13788 the user input is shown, useful e.g. for user names.
13789 systemd-ask-password gained a new --echo option to turn that
13792 * The default sysctl.d/ snippets will now set:
13794 net.core.default_qdisc = fq_codel
13796 This selects Fair Queuing Controlled Delay as the default
13797 queuing discipline for network interfaces. fq_codel helps
13798 fight the network bufferbloat problem. It is believed to be
13799 a good default with no tuning required for most workloads.
13800 Downstream distributions may override this choice. On 10Gbit
13801 servers that do not do forwarding, "fq" may perform better.
13802 Systems without a good clocksource should use "pfifo_fast".
13804 * If kdbus is enabled during build a new option BusPolicy= is
13805 available for service units, that allows locking all service
13806 processes into a stricter bus policy, in order to limit
13807 access to various bus services, or even hide most of them
13808 from the service's view entirely.
13810 * networkctl will now show the .network and .link file
13811 networkd has applied to a specific interface.
13813 * sd-login gained a new API call sd_session_get_desktop() to
13814 query which desktop environment has been selected for a
13817 * UNIX utmp support is now compile-time optional to support
13818 legacy-free systems.
13820 * systemctl gained two new commands "add-wants" and
13821 "add-requires" for pulling in units from specific targets
13824 * If the word "rescue" is specified on the kernel command line
13825 the system will now boot into rescue mode (aka
13826 rescue.target), which was previously available only by
13827 specifying "1" or "systemd.unit=rescue.target" on the kernel
13828 command line. This new kernel command line option nicely
13829 mirrors the already existing "emergency" kernel command line
13832 * New kernel command line options mount.usr=, mount.usrflags=,
13833 mount.usrfstype= have been added that match root=, rootflags=,
13834 rootfstype= but allow mounting a specific file system to
13837 * The $NOTIFY_SOCKET is now also passed to control processes of
13838 services, not only the main process.
13840 * This version reenables support for fsck's -l switch. This
13841 means at least version v2.25 of util-linux is required for
13842 operation, otherwise dead-locks on device nodes may
13843 occur. Again: you need to update util-linux to at least
13844 v2.25 when updating systemd to v217.
13846 * The "multi-seat-x" tool has been removed from systemd, as
13847 its functionality has been integrated into X servers 1.16,
13848 and the tool is hence redundant. It is recommended to update
13849 display managers invoking this tool to simply invoke X
13850 directly from now on, again.
13852 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
13853 message flag has been added for all of systemd's polkit
13854 authenticated method calls has been added. In particular this
13855 now allows optional interactive authorization via polkit for
13856 many of PID1's privileged operations such as unit file
13857 enabling and disabling.
13859 * "udevadm hwdb --update" learnt a new switch "--usr" for
13860 placing the rebuilt hardware database in /usr instead of
13861 /etc. When used only hardware database entries stored in
13862 /usr will be used, and any user database entries in /etc are
13863 ignored. This functionality is useful for vendors to ship a
13864 pre-built database on systems where local configuration is
13865 unnecessary or unlikely.
13867 * Calendar time specifications in .timer units now also
13868 understand the strings "semi-annually", "quarterly" and
13869 "minutely" as shortcuts (in addition to the preexisting
13870 "annually", "hourly", …).
13872 * systemd-tmpfiles will now correctly create files in /dev
13873 at boot which are marked for creation only at boot. It is
13874 recommended to always create static device nodes with 'c!'
13875 and 'b!', so that they are created only at boot and not
13876 overwritten at runtime.
13878 * When the watchdog logic is used for a service (WatchdogSec=)
13879 and the watchdog timeout is hit the service will now be
13880 terminated with SIGABRT (instead of just SIGTERM), in order
13881 to make sure a proper coredump and backtrace is
13882 generated. This ensures that hanging services will result in
13883 similar coredump/backtrace behaviour as services that hit a
13884 segmentation fault.
13886 Contributions from: Andreas Henriksson, Andrei Borzenkov,
13887 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
13888 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
13889 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
13890 Herrmann, David Sommerseth, David Strauss, Emil Renner
13891 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
13892 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
13893 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
13894 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
13895 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
13896 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
13897 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
13898 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
13899 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
13900 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
13901 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
13902 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
13903 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
13904 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
13905 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
13906 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
13909 — Berlin, 2014-10-28
13913 * timedated no longer reads NTP implementation unit names from
13914 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
13915 implementations should add a
13917 Conflicts=systemd-timesyncd.service
13919 to their unit files to take over and replace systemd's NTP
13920 default functionality.
13922 * systemd-sysusers gained a new line type "r" for configuring
13923 which UID/GID ranges to allocate system users/groups
13924 from. Lines of type "u" may now add an additional column
13925 that specifies the home directory for the system user to be
13926 created. Also, systemd-sysusers may now optionally read user
13927 information from STDIN instead of a file. This is useful for
13928 invoking it from RPM preinst scriptlets that need to create
13929 users before the first RPM file is installed since these
13930 files might need to be owned by them. A new
13931 %sysusers_create_inline RPM macro has been introduced to do
13932 just that. systemd-sysusers now updates the shadow files as
13933 well as the user/group databases, which should enhance
13934 compatibility with certain tools like grpck.
13936 * A number of bus APIs of PID 1 now optionally consult polkit to
13937 permit access for otherwise unprivileged clients under certain
13938 conditions. Note that this currently doesn't support
13939 interactive authentication yet, but this is expected to be
13940 added eventually, too.
13942 * /etc/machine-info now has new fields for configuring the
13943 deployment environment of the machine, as well as the
13944 location of the machine. hostnamectl has been updated with
13945 new command to update these fields.
13947 * systemd-timesyncd has been updated to automatically acquire
13948 NTP server information from systemd-networkd, which might
13949 have been discovered via DHCP.
13951 * systemd-resolved now includes a caching DNS stub resolver
13952 and a complete LLMNR name resolution implementation. A new
13953 NSS module "nss-resolve" has been added which can be used
13954 instead of glibc's own "nss-dns" to resolve hostnames via
13955 systemd-resolved. Hostnames, addresses and arbitrary RRs may
13956 be resolved via systemd-resolved D-Bus APIs. In contrast to
13957 the glibc internal resolver systemd-resolved is aware of
13958 multi-homed system, and keeps DNS server and caches separate
13959 and per-interface. Queries are sent simultaneously on all
13960 interfaces that have DNS servers configured, in order to
13961 properly handle VPNs and local LANs which might resolve
13962 separate sets of domain names. systemd-resolved may acquire
13963 DNS server information from systemd-networkd automatically,
13964 which in turn might have discovered them via DHCP. A tool
13965 "systemd-resolve-host" has been added that may be used to
13966 query the DNS logic in resolved. systemd-resolved implements
13967 IDNA and automatically uses IDNA or UTF-8 encoding depending
13968 on whether classic DNS or LLMNR is used as transport. In the
13969 next releases we intend to add a DNSSEC and mDNS/DNS-SD
13970 implementation to systemd-resolved.
13972 * A new NSS module nss-mymachines has been added, that
13973 automatically resolves the names of all local registered
13974 containers to their respective IP addresses.
13976 * A new client tool "networkctl" for systemd-networkd has been
13977 added. It currently is entirely passive and will query
13978 networking configuration from udev, rtnetlink and networkd,
13979 and present it to the user in a very friendly
13980 way. Eventually, we hope to extend it to become a full
13981 control utility for networkd.
13983 * .socket units gained a new DeferAcceptSec= setting that
13984 controls the kernels' TCP_DEFER_ACCEPT sockopt for
13985 TCP. Similarly, support for controlling TCP keep-alive
13986 settings has been added (KeepAliveTimeSec=,
13987 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
13988 turning off Nagle's algorithm on TCP has been added
13991 * logind learned a new session type "web", for use in projects
13992 like Cockpit which register web clients as PAM sessions.
13994 * timer units with at least one OnCalendar= setting will now
13995 be started only after time-sync.target has been
13996 reached. This way they will not elapse before the system
13997 clock has been corrected by a local NTP client or
13998 similar. This is particular useful on RTC-less embedded
13999 machines, that come up with an invalid system clock.
14001 * systemd-nspawn's --network-veth= switch should now result in
14002 stable MAC addresses for both the outer and the inner side
14005 * systemd-nspawn gained a new --volatile= switch for running
14006 container instances with /etc or /var unpopulated.
14008 * The kdbus client code has been updated to use the new Linux
14009 3.17 memfd subsystem instead of the old kdbus-specific one.
14011 * systemd-networkd's DHCP client and server now support
14012 FORCERENEW. There are also new configuration options to
14013 configure the vendor client identifier and broadcast mode
14016 * systemd will no longer inform the kernel about the current
14017 timezone, as this is necessarily incorrect and racy as the
14018 kernel has no understanding of DST and similar
14019 concepts. This hence means FAT timestamps will be always
14020 considered UTC, similar to what Android is already
14021 doing. Also, when the RTC is configured to the local time
14022 (rather than UTC) systemd will never synchronize back to it,
14023 as this might confuse Windows at a later boot.
14025 * systemd-analyze gained a new command "verify" for offline
14026 validation of unit files.
14028 * systemd-networkd gained support for a couple of additional
14029 settings for bonding networking setups. Also, the metric for
14030 statically configured routes may now be configured. For
14031 network interfaces where this is appropriate the peer IP
14032 address may now be configured.
14034 * systemd-networkd's DHCP client will no longer request
14035 broadcasting by default, as this tripped up some networks.
14036 For hardware where broadcast is required the feature should
14037 be switched back on using RequestBroadcast=yes.
14039 * systemd-networkd will now set up IPv4LL addresses (when
14040 enabled) even if DHCP is configured successfully.
14042 * udev will now default to respect network device names given
14043 by the kernel when the kernel indicates that these are
14044 predictable. This behavior can be tweaked by changing
14045 NamePolicy= in the relevant .link file.
14047 * A new library systemd-terminal has been added that
14048 implements full TTY stream parsing and rendering. This
14049 library is supposed to be used later on for implementing a
14050 full userspace VT subsystem, replacing the current kernel
14053 * A new tool systemd-journal-upload has been added to push
14054 journal data to a remote system running
14055 systemd-journal-remote.
14057 * journald will no longer forward all local data to another
14058 running syslog daemon. This change has been made because
14059 rsyslog (which appears to be the most commonly used syslog
14060 implementation these days) no longer makes use of this, and
14061 instead pulls the data out of the journal on its own. Since
14062 forwarding the messages to a non-existent syslog server is
14063 more expensive than we assumed we have now turned this
14064 off. If you run a syslog server that is not a recent rsyslog
14065 version, you have to turn this option on again
14066 (ForwardToSyslog= in journald.conf).
14068 * journald now optionally supports the LZ4 compressor for
14069 larger journal fields. This compressor should perform much
14070 better than XZ which was the previous default.
14072 * machinectl now shows the IP addresses of local containers,
14073 if it knows them, plus the interface name of the container.
14075 * A new tool "systemd-escape" has been added that makes it
14076 easy to escape strings to build unit names and similar.
14078 * sd_notify() messages may now include a new ERRNO= field
14079 which is parsed and collected by systemd and shown among the
14080 "systemctl status" output for a service.
14082 * A new component "systemd-firstboot" has been added that
14083 queries the most basic systemd information (timezone,
14084 hostname, root password) interactively on first
14085 boot. Alternatively it may also be used to provision these
14086 things offline on OS images installed into directories.
14088 * The default sysctl.d/ snippets will now set
14090 net.ipv4.conf.default.promote_secondaries=1
14092 This has the benefit of no flushing secondary IP addresses
14093 when primary addresses are removed.
14095 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
14096 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
14097 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
14098 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
14099 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
14100 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
14101 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
14102 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
14103 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
14104 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
14105 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
14106 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
14107 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
14108 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
14109 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
14111 — Berlin, 2014-08-19
14115 * A new tool systemd-sysusers has been added. This tool
14116 creates system users and groups in /etc/passwd and
14117 /etc/group, based on static declarative system user/group
14118 definitions in /usr/lib/sysusers.d/. This is useful to
14119 enable factory resets and volatile systems that boot up with
14120 an empty /etc directory, and thus need system users and
14121 groups created during early boot. systemd now also ships
14122 with two default sysusers.d/ files for the most basic
14123 users and groups systemd and the core operating system
14126 * A new tmpfiles snippet has been added that rebuilds the
14127 essential files in /etc on boot, should they be missing.
14129 * A directive for ensuring automatic clean-up of
14130 /var/cache/man/ has been removed from the default
14131 configuration. This line should now be shipped by the man
14132 implementation. The necessary change has been made to the
14133 man-db implementation. Note that you need to update your man
14134 implementation to one that ships this line, otherwise no
14135 automatic clean-up of /var/cache/man will take place.
14137 * A new condition ConditionNeedsUpdate= has been added that
14138 may conditionalize services to only run when /etc or /var
14139 are "older" than the vendor operating system resources in
14140 /usr. This is useful for reconstructing or updating /etc
14141 after an offline update of /usr or a factory reset, on the
14142 next reboot. Services that want to run once after such an
14143 update or reset should use this condition and order
14144 themselves before the new systemd-update-done.service, which
14145 will mark the two directories as fully updated. A number of
14146 service files have been added making use of this, to rebuild
14147 the udev hardware database, the journald message catalog and
14148 dynamic loader cache (ldconfig). The systemd-sysusers tool
14149 described above also makes use of this now. With this in
14150 place it is now possible to start up a minimal operating
14151 system with /etc empty cleanly. For more information on the
14152 concepts involved see this recent blog story:
14154 https://0pointer.de/blog/projects/stateless.html
14156 * A new system group "input" has been introduced, and all
14157 input device nodes get this group assigned. This is useful
14158 for system-level software to get access to input devices. It
14159 complements what is already done for "audio" and "video".
14161 * systemd-networkd learnt minimal DHCPv4 server support in
14162 addition to the existing DHCPv4 client support. It also
14163 learnt DHCPv6 client and IPv6 Router Solicitation client
14164 support. The DHCPv4 client gained support for static routes
14165 passed in from the server. Note that the [DHCPv4] section
14166 known in older systemd-networkd versions has been renamed to
14167 [DHCP] and is now also used by the DHCPv6 client. Existing
14168 .network files using settings of this section should be
14169 updated, though compatibility is maintained. Optionally, the
14170 client hostname may now be sent to the DHCP server.
14172 * networkd gained support for vxlan virtual networks as well
14173 as tun/tap and dummy devices.
14175 * networkd gained support for automatic allocation of address
14176 ranges for interfaces from a system-wide pool of
14177 addresses. This is useful for dynamically managing a large
14178 number of interfaces with a single network configuration
14179 file. In particular this is useful to easily assign
14180 appropriate IP addresses to the veth links of a large number
14181 of nspawn instances.
14183 * RPM macros for processing sysusers, sysctl and binfmt
14184 drop-in snippets at package installation time have been
14187 * The /etc/os-release file should now be placed in
14188 /usr/lib/os-release. The old location is automatically
14189 created as symlink. /usr/lib is the more appropriate
14190 location of this file, since it shall actually describe the
14191 vendor operating system shipped in /usr, and not the
14192 configuration stored in /etc.
14194 * .mount units gained a new boolean SloppyOptions= setting
14195 that maps to mount(8)'s -s option which enables permissive
14196 parsing of unknown mount options.
14198 * tmpfiles learnt a new "L+" directive which creates a symlink
14199 but (unlike "L") deletes a pre-existing file first, should
14200 it already exist and not already be the correct
14201 symlink. Similarly, "b+", "c+" and "p+" directives have been
14202 added as well, which create block and character devices, as
14203 well as fifos in the filesystem, possibly removing any
14204 pre-existing files of different types.
14206 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
14207 'argument' field (which so far specified the source to
14208 symlink/copy the files from) is now optional. If omitted the
14209 same file os copied from /usr/share/factory/ suffixed by the
14210 full destination path. This is useful for populating /etc
14211 with essential files, by copying them from vendor defaults
14212 shipped in /usr/share/factory/etc.
14214 * A new command "systemctl preset-all" has been added that
14215 applies the service preset settings to all installed unit
14216 files. A new switch --preset-mode= has been added that
14217 controls whether only enable or only disable operations
14220 * A new command "systemctl is-system-running" has been added
14221 that allows checking the overall state of the system, for
14222 example whether it is fully up and running.
14224 * When the system boots up with an empty /etc, the equivalent
14225 to "systemctl preset-all" is executed during early boot, to
14226 make sure all default services are enabled after a factory
14229 * systemd now contains a minimal preset file that enables the
14230 most basic services systemd ships by default.
14232 * Unit files' [Install] section gained a new DefaultInstance=
14233 field for defining the default instance to create if a
14234 template unit is enabled with no instance specified.
14236 * A new passive target cryptsetup-pre.target has been added
14237 that may be used by services that need to make they run and
14238 finish before the first LUKS cryptographic device is set up.
14240 * The /dev/loop-control and /dev/btrfs-control device nodes
14241 are now owned by the "disk" group by default, opening up
14242 access to this group.
14244 * systemd-coredump will now automatically generate a
14245 stack trace of all core dumps taking place on the system,
14246 based on elfutils' libdw library. This stack trace is logged
14249 * systemd-coredump may now optionally store coredumps directly
14250 on disk (in /var/lib/systemd/coredump, possibly compressed),
14251 instead of storing them unconditionally in the journal. This
14252 mode is the new default. A new configuration file
14253 /etc/systemd/coredump.conf has been added to configure this
14254 and other parameters of systemd-coredump.
14256 * coredumpctl gained a new "info" verb to show details about a
14257 specific coredump. A new switch "-1" has also been added
14258 that makes sure to only show information about the most
14259 recent entry instead of all entries. Also, as the tool is
14260 generally useful now the "systemd-" prefix of the binary
14261 name has been removed. Distributions that want to maintain
14262 compatibility with the old name should add a symlink from
14263 the old name to the new name.
14265 * journald's SplitMode= now defaults to "uid". This makes sure
14266 that unprivileged users can access their own coredumps with
14267 coredumpctl without restrictions.
14269 * New kernel command line options "systemd.wants=" (for
14270 pulling an additional unit during boot), "systemd.mask="
14271 (for masking a specific unit for the boot), and
14272 "systemd.debug-shell" (for enabling the debug shell on tty9)
14273 have been added. This is implemented in the new generator
14274 "systemd-debug-generator".
14276 * systemd-nspawn will now by default filter a couple of
14277 syscalls for containers, among them those required for
14278 kernel module loading, direct x86 IO port access, swap
14279 management, and kexec. Most importantly though
14280 open_by_handle_at() is now prohibited for containers,
14281 closing a hole similar to a recently discussed vulnerability
14282 in docker regarding access to files on file hierarchies the
14283 container should normally not have access to. Note that, for
14284 nspawn, we generally make no security claims anyway (and
14285 this is explicitly documented in the man page), so this is
14286 just a fix for one of the most obvious problems.
14288 * A new man page file-hierarchy(7) has been added that
14289 contains a minimized, modernized version of the file system
14290 layout systemd expects, similar in style to the FHS
14291 specification or hier(5). A new tool systemd-path(1) has
14292 been added to query many of these paths for the local
14295 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
14296 longer done. Since the directory now has a per-user size
14297 limit, and is cleaned on logout this appears unnecessary,
14298 in particular since this now brings the lifecycle of this
14299 directory closer in line with how IPC objects are handled.
14301 * systemd.pc now exports a number of additional directories,
14302 including $libdir (which is useful to identify the library
14303 path for the primary architecture of the system), and a
14304 couple of drop-in directories.
14306 * udev's predictable network interface names now use the dev_port
14307 sysfs attribute, introduced in linux 3.15 instead of dev_id to
14308 distinguish between ports of the same PCI function. dev_id should
14309 only be used for ports using the same HW address, hence the need
14312 * machined has been updated to export the OS version of a
14313 container (read from /etc/os-release and
14314 /usr/lib/os-release) on the bus. This is now shown in
14315 "machinectl status" for a machine.
14317 * A new service setting RestartForceExitStatus= has been
14318 added. If configured to a set of exit signals or process
14319 return values, the service will be restarted when the main
14320 daemon process exits with any of them, regardless of the
14323 * systemctl's -H switch for connecting to remote systemd
14324 machines has been extended so that it may be used to
14325 directly connect to a specific container on the
14326 host. "systemctl -H root@foobar:waldi" will now connect as
14327 user "root" to host "foobar", and then proceed directly to
14328 the container named "waldi". Note that currently you have to
14329 authenticate as user "root" for this to work, as entering
14330 containers is a privileged operation.
14332 Contributions from: Andreas Henriksson, Benjamin Steinwender,
14333 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
14334 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
14335 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
14336 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
14337 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
14338 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
14339 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
14340 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
14341 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
14342 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
14343 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
14345 — Berlin, 2014-07-03
14349 * As an experimental feature, udev now tries to lock the
14350 disk device node (flock(LOCK_SH|LOCK_NB)) while it
14351 executes events for the disk or any of its partitions.
14352 Applications like partitioning programs can lock the
14353 disk device node (flock(LOCK_EX)) and claim temporary
14354 device ownership that way; udev will entirely skip all event
14355 handling for this disk and its partitions. If the disk
14356 was opened for writing, the close will trigger a partition
14357 table rescan in udev's "watch" facility, and if needed
14358 synthesize "change" events for the disk and all its partitions.
14359 This is now unconditionally enabled, and if it turns out to
14360 cause major problems, we might turn it on only for specific
14361 devices, or might need to disable it entirely. Device Mapper
14362 devices are excluded from this logic.
14364 * We temporarily dropped the "-l" switch for fsck invocations,
14365 since they collide with the flock() logic above. util-linux
14366 upstream has been changed already to avoid this conflict,
14367 and we will re-add "-l" as soon as util-linux with this
14368 change has been released.
14370 * The dependency on libattr has been removed. Since a long
14371 time, the extended attribute calls have moved to glibc, and
14372 libattr is thus unnecessary.
14374 * Virtualization detection works without privileges now. This
14375 means the systemd-detect-virt binary no longer requires
14376 CAP_SYS_PTRACE file capabilities, and our daemons can run
14377 with fewer privileges.
14379 * systemd-networkd now runs under its own "systemd-network"
14380 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
14381 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
14382 loses the ability to write to files owned by root this way.
14384 * Similarly, systemd-resolved now runs under its own
14385 "systemd-resolve" user with no capabilities remaining.
14387 * Similarly, systemd-bus-proxyd now runs under its own
14388 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
14390 * systemd-networkd gained support for setting up "veth"
14391 virtual Ethernet devices for container connectivity, as well
14392 as GRE and VTI tunnels.
14394 * systemd-networkd will no longer automatically attempt to
14395 manually load kernel modules necessary for certain tunnel
14396 transports. Instead, it is assumed the kernel loads them
14397 automatically when required. This only works correctly on
14398 very new kernels. On older kernels, please consider adding
14399 the kernel modules to /etc/modules-load.d/ as a work-around.
14401 * The resolv.conf file systemd-resolved generates has been
14402 moved to /run/systemd/resolve/. If you have a symlink from
14403 /etc/resolv.conf, it might be necessary to correct it.
14405 * Two new service settings, ProtectHome= and ProtectSystem=,
14406 have been added. When enabled, they will make the user data
14407 (such as /home) inaccessible or read-only and the system
14408 (such as /usr) read-only, for specific services. This allows
14409 very light-weight per-service sandboxing to avoid
14410 modifications of user data or system files from
14411 services. These two new switches have been enabled for all
14412 of systemd's long-running services, where appropriate.
14414 * Socket units gained new SocketUser= and SocketGroup=
14415 settings to set the owner user and group of AF_UNIX sockets
14416 and FIFOs in the file system.
14418 * Socket units gained a new RemoveOnStop= setting. If enabled,
14419 all FIFOS and sockets in the file system will be removed
14420 when the specific socket unit is stopped.
14422 * Socket units gained a new Symlinks= setting. It takes a list
14423 of symlinks to create to file system sockets or FIFOs
14424 created by the specific Unix sockets. This is useful to
14425 manage symlinks to socket nodes with the same lifecycle as
14428 * The /dev/log socket and /dev/initctl FIFO have been moved to
14429 /run, and have been replaced by symlinks. This allows
14430 connecting to these facilities even if PrivateDevices=yes is
14431 used for a service (which makes /dev/log itself unavailable,
14432 but /run is left). This also has the benefit of ensuring
14433 that /dev only contains device nodes, directories and
14434 symlinks, and nothing else.
14436 * sd-daemon gained two new calls sd_pid_notify() and
14437 sd_pid_notifyf(). They are similar to sd_notify() and
14438 sd_notifyf(), but allow overriding of the source PID of
14439 notification messages if permissions permit this. This is
14440 useful to send notify messages on behalf of a different
14441 process (for example, the parent process). The
14442 systemd-notify tool has been updated to make use of this
14443 when sending messages (so that notification messages now
14444 originate from the shell script invoking systemd-notify and
14445 not the systemd-notify process itself. This should minimize
14446 a race where systemd fails to associate notification
14447 messages to services when the originating process already
14450 * A new "on-abnormal" setting for Restart= has been added. If
14451 set, it will result in automatic restarts on all "abnormal"
14452 reasons for a process to exit, which includes unclean
14453 signals, core dumps, timeouts and watchdog timeouts, but
14454 does not include clean and unclean exit codes or clean
14455 signals. Restart=on-abnormal is an alternative for
14456 Restart=on-failure for services that shall be able to
14457 terminate and avoid restarts on certain errors, by
14458 indicating so with an unclean exit code. Restart=on-failure
14459 or Restart=on-abnormal is now the recommended setting for
14460 all long-running services.
14462 * If the InaccessibleDirectories= service setting points to a
14463 mount point (or if there are any submounts contained within
14464 it), it is now attempted to completely unmount it, to make
14465 the file systems truly unavailable for the respective
14468 * The ReadOnlyDirectories= service setting and
14469 systemd-nspawn's --read-only parameter are now recursively
14470 applied to all submounts, too.
14472 * Mount units may now be created transiently via the bus APIs.
14474 * The support for SysV and LSB init scripts has been removed
14475 from the systemd daemon itself. Instead, it is now
14476 implemented as a generator that creates native systemd units
14477 from these scripts when needed. This enables us to remove a
14478 substantial amount of legacy code from PID 1, following the
14479 fact that many distributions only ship a very small number
14480 of LSB/SysV init scripts nowadays.
14482 * Privileged Xen (dom0) domains are not considered
14483 virtualization anymore by the virtualization detection
14484 logic. After all, they generally have unrestricted access to
14485 the hardware and usually are used to manage the unprivileged
14488 * systemd-tmpfiles gained a new "C" line type, for copying
14489 files or entire directories.
14491 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
14492 lines. So far, they have been non-globbing versions of the
14493 latter, and have thus been redundant. In future, it is
14494 recommended to only use "z". "m" has hence been removed
14495 from the documentation, even though it stays supported.
14497 * A tmpfiles snippet to recreate the most basic structure in
14498 /var has been added. This is enough to create the /var/run →
14499 /run symlink and create a couple of structural
14500 directories. This allows systems to boot up with an empty or
14501 volatile /var. Of course, while with this change, the core OS
14502 now is capable with dealing with a volatile /var, not all
14503 user services are ready for it. However, we hope that sooner
14504 or later, many service daemons will be changed upstream so
14505 that they are able to automatically create their necessary
14506 directories in /var at boot, should they be missing. This is
14507 the first step to allow state-less systems that only require
14508 the vendor image for /usr to boot.
14510 * systemd-nspawn has gained a new --tmpfs= switch to mount an
14511 empty tmpfs instance to a specific directory. This is
14512 particularly useful for making use of the automatic
14513 reconstruction of /var (see above), by passing --tmpfs=/var.
14515 * Access modes specified in tmpfiles snippets may now be
14516 prefixed with "~", which indicates that they shall be masked
14517 by whether the existing file or directory is currently
14518 writable, readable or executable at all. Also, if specified,
14519 the sgid/suid/sticky bits will be masked for all
14522 * A new passive target unit "network-pre.target" has been
14523 added which is useful for services that shall run before any
14524 network is configured, for example firewall scripts.
14526 * The "floppy" group that previously owned the /dev/fd*
14527 devices is no longer used. The "disk" group is now used
14528 instead. Distributions should probably deprecate usage of
14531 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
14532 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
14533 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
14534 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
14535 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
14536 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
14539 — Berlin, 2014-06-11
14543 * A new "systemd-timesyncd" daemon has been added for
14544 synchronizing the system clock across the network. It
14545 implements an SNTP client. In contrast to NTP
14546 implementations such as chrony or the NTP reference server,
14547 this only implements a client side, and does not bother with
14548 the full NTP complexity, focusing only on querying time from
14549 one remote server and synchronizing the local clock to
14550 it. Unless you intend to serve NTP to networked clients or
14551 want to connect to local hardware clocks, this simple NTP
14552 client should be more than appropriate for most
14553 installations. The daemon runs with minimal privileges, and
14554 has been hooked up with networkd to only operate when
14555 network connectivity is available. The daemon saves the
14556 current clock to disk every time a new NTP sync has been
14557 acquired, and uses this to possibly correct the system clock
14558 early at bootup, in order to accommodate for systems that
14559 lack an RTC such as the Raspberry Pi and embedded devices,
14560 and to make sure that time monotonically progresses on these
14561 systems, even if it is not always correct. To make use of
14562 this daemon, a new system user and group "systemd-timesync"
14563 needs to be created on installation of systemd.
14565 * The queue "seqnum" interface of libudev has been disabled, as
14566 it was generally incompatible with device namespacing as
14567 sequence numbers of devices go "missing" if the devices are
14568 part of a different namespace.
14570 * "systemctl list-timers" and "systemctl list-sockets" gained
14571 a --recursive switch for showing units of these types also
14572 for all local containers, similar in style to the already
14573 supported --recursive switch for "systemctl list-units".
14575 * A new RebootArgument= setting has been added for service
14576 units, which may be used to specify a kernel reboot argument
14577 to use when triggering reboots with StartLimitAction=.
14579 * A new FailureAction= setting has been added for service
14580 units which may be used to specify an operation to trigger
14581 when a service fails. This works similarly to
14582 StartLimitAction=, but unlike it, controls what is done
14583 immediately rather than only after several attempts to
14584 restart the service in question.
14586 * hostnamed got updated to also expose the kernel name,
14587 release, and version on the bus. This is useful for
14588 executing commands like hostnamectl with the -H switch.
14589 systemd-analyze makes use of this to properly display
14590 details when running non-locally.
14592 * The bootchart tool can now show cgroup information in the
14593 graphs it generates.
14595 * The CFS CPU quota cgroup attribute is now exposed for
14596 services. The new CPUQuota= switch has been added for this
14597 which takes a percentage value. Setting this will have the
14598 result that a service may never get more CPU time than the
14599 specified percentage, even if the machine is otherwise idle.
14601 * systemd-networkd learned IPIP and SIT tunnel support.
14603 * LSB init scripts exposing a dependency on $network will now
14604 get a dependency on network-online.target rather than simply
14605 network.target. This should bring LSB handling closer to
14606 what it was on SysV systems.
14608 * A new fsck.repair= kernel option has been added to control
14609 how fsck shall deal with unclean file systems at boot.
14611 * The (.ini) configuration file parser will now silently ignore
14612 sections whose names begin with "X-". This may be used to maintain
14613 application-specific extension sections in unit files.
14615 * machined gained a new API to query the IP addresses of
14616 registered containers. "machinectl status" has been updated
14617 to show these addresses in its output.
14619 * A new call sd_uid_get_display() has been added to the
14620 sd-login APIs for querying the "primary" session of a
14621 user. The "primary" session of the user is elected from the
14622 user's sessions and generally a graphical session is
14623 preferred over a text one.
14625 * A minimal systemd-resolved daemon has been added. It
14626 currently simply acts as a companion to systemd-networkd and
14627 manages resolv.conf based on per-interface DNS
14628 configuration, possibly supplied via DHCP. In the long run
14629 we hope to extend this into a local DNSSEC enabled DNS and
14632 * The systemd-networkd-wait-online tool is now enabled by
14633 default. It will delay network-online.target until a network
14634 connection has been configured. The tool primarily integrates
14635 with networkd, but will also make a best effort to make sense
14636 of network configuration performed in some other way.
14638 * Two new service options StartupCPUShares= and
14639 StartupBlockIOWeight= have been added that work similarly to
14640 CPUShares= and BlockIOWeight= however only apply during
14641 system startup. This is useful to prioritize certain services
14642 differently during bootup than during normal runtime.
14644 * hostnamed has been changed to prefer the statically
14645 configured hostname in /etc/hostname (unless set to
14646 'localhost' or empty) over any dynamic one supplied by
14647 dhcp. With this change, the rules for picking the hostname
14648 match more closely the rules of other configuration settings
14649 where the local administrator's configuration in /etc always
14650 overrides any other settings.
14652 Contributions from: Ali H. Caliskan, Alison Chaiken, Bas van
14653 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
14654 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
14655 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
14656 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
14657 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
14658 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
14659 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
14660 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
14661 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
14662 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
14663 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
14664 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
14665 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
14666 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
14667 Lindskog, WaLyong Cho, Will Woods, Zbigniew
14670 — Beijing, 2014-05-28
14674 * When restoring the screen brightness at boot, stay away from
14675 the darkest setting or from the lowest 5% of the available
14676 range, depending on which is the larger value of both. This
14677 should effectively protect the user from rebooting into a
14678 black screen, should the brightness have been set to minimum
14681 * sd-login gained a new sd_machine_get_class() call to
14682 determine the class ("vm" or "container") of a machine
14683 registered with machined.
14685 * sd-login gained new calls
14686 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
14687 to query the identity of the peer of a local AF_UNIX
14688 connection. They operate similarly to their sd_pid_get_xyz()
14691 * PID 1 will now maintain a system-wide system state engine
14692 with the states "starting", "running", "degraded",
14693 "maintenance", "stopping". These states are bound to system
14694 startup, normal runtime, runtime with at least one failed
14695 service, rescue/emergency mode and system shutdown. This
14696 state is shown in the "systemctl status" output when no unit
14697 name is passed. It is useful to determine system state, in
14698 particularly when doing so for many systems or containers at
14701 * A new command "list-machines" has been added to "systemctl"
14702 that lists all local OS containers and shows their system
14703 state (see above), if systemd runs inside of them.
14705 * systemctl gained a new "-r" switch to recursively enumerate
14706 units on all local containers, when used with the
14707 "list-unit" command (which is the default one that is
14708 executed when no parameters are specified).
14710 * The GPT automatic partition discovery logic will now honour
14711 two GPT partition flags: one may be set on a partition to
14712 cause it to be mounted read-only, and the other may be set
14713 on a partition to ignore it during automatic discovery.
14715 * Two new GPT type UUIDs have been added for automatic root
14716 partition discovery, for 32-bit and 64-bit ARM. This is not
14717 particularly useful for discovering the root directory on
14718 these architectures during bare-metal boots (since UEFI is
14719 not common there), but still very useful to allow booting of
14720 ARM disk images in nspawn with the -i option.
14722 * MAC addresses of interfaces created with nspawn's
14723 --network-interface= switch will now be generated from the
14724 machine name, and thus be stable between multiple invocations
14727 * logind will now automatically remove all IPC objects owned
14728 by a user if she or he fully logs out. This makes sure that
14729 users who are logged out cannot continue to consume IPC
14730 resources. This covers SysV memory, semaphores and message
14731 queues as well as POSIX shared memory and message
14732 queues. Traditionally, SysV and POSIX IPC had no lifecycle
14733 limits. With this functionality, that is corrected. This may
14734 be turned off by using the RemoveIPC= switch of logind.conf.
14736 * The systemd-machine-id-setup and tmpfiles tools gained a
14737 --root= switch to operate on a specific root directory,
14740 * journald can now forward logged messages to the TTYs of all
14741 logged in users ("wall"). This is the default for all
14742 emergency messages now.
14744 * A new tool systemd-journal-remote has been added to stream
14745 journal log messages across the network.
14747 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
14748 controller trees are mounted into it. Note that the
14749 directories mounted beneath it are not read-only. This is a
14750 security measure and is particularly useful because glibc
14751 actually includes a search logic to pick any tmpfs it can
14752 find to implement shm_open() if /dev/shm is not available
14753 (which it might very well be in namespaced setups).
14755 * machinectl gained a new "poweroff" command to cleanly power
14756 down a local OS container.
14758 * The PrivateDevices= unit file setting will now also drop the
14759 CAP_MKNOD capability from the capability bound set, and
14760 imply DevicePolicy=closed.
14762 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
14763 comprehensively on all long-running systemd services where
14764 this is appropriate.
14766 * systemd-udevd will now run in a disassociated mount
14767 namespace. To mount directories from udev rules, make sure to
14768 pull in mount units via SYSTEMD_WANTS properties.
14770 * The kdbus support gained support for uploading policy into
14771 the kernel. sd-bus gained support for creating "monitoring"
14772 connections that can eavesdrop into all bus communication
14773 for debugging purposes.
14775 * Timestamps may now be specified in seconds since the UNIX
14776 epoch Jan 1st, 1970 by specifying "@" followed by the value
14779 * Native tcpwrap support in systemd has been removed. tcpwrap
14780 is old code, not really maintained anymore and has serious
14781 shortcomings, and better options such as firewalls
14782 exist. For setups that require tcpwrap usage, please
14783 consider invoking your socket-activated service via tcpd,
14784 like on traditional inetd.
14786 * A new system.conf configuration option
14787 DefaultTimerAccuracySec= has been added that controls the
14788 default AccuracySec= setting of .timer units.
14790 * Timer units gained a new WakeSystem= switch. If enabled,
14791 timers configured this way will cause the system to resume
14792 from system suspend (if the system supports that, which most
14795 * Timer units gained a new Persistent= switch. If enabled,
14796 timers configured this way will save to disk when they have
14797 been last triggered. This information is then used on next
14798 reboot to possible execute overdue timer events, that
14799 could not take place because the system was powered off.
14800 This enables simple anacron-like behaviour for timer units.
14802 * systemctl's "list-timers" will now also list the time a
14803 timer unit was last triggered in addition to the next time
14804 it will be triggered.
14806 * systemd-networkd will now assign predictable IPv4LL
14807 addresses to its local interfaces.
14809 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
14810 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
14811 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
14812 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
14813 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
14814 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
14815 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
14816 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
14819 — Berlin, 2014-03-25
14823 * A new unit file setting RestrictAddressFamilies= has been
14824 added to restrict which socket address families unit
14825 processes gain access to. This takes address family names
14826 like "AF_INET" or "AF_UNIX", and is useful to minimize the
14827 attack surface of services via exotic protocol stacks. This
14828 is built on seccomp system call filters.
14830 * Two new unit file settings RuntimeDirectory= and
14831 RuntimeDirectoryMode= have been added that may be used to
14832 manage a per-daemon runtime directories below /run. This is
14833 an alternative for setting up directory permissions with
14834 tmpfiles snippets, and has the advantage that the runtime
14835 directory's lifetime is bound to the daemon runtime and that
14836 the daemon starts up with an empty directory each time. This
14837 is particularly useful when writing services that drop
14838 privileges using the User= or Group= setting.
14840 * The DeviceAllow= unit setting now supports globbing for
14841 matching against device group names.
14843 * The systemd configuration file system.conf gained new
14844 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
14845 DefaultMemoryAccounting= to globally turn on/off accounting
14846 for specific resources (cgroups) for all units. These
14847 settings may still be overridden individually in each unit
14850 * systemd-gpt-auto-generator is now able to discover /srv and
14851 root partitions in addition to /home and swap partitions. It
14852 also supports LUKS-encrypted partitions now. With this in
14853 place, automatic discovery of partitions to mount following
14854 the Discoverable Partitions Specification
14855 (https://systemd.io/DISCOVERABLE_PARTITIONS/)
14856 is now a lot more complete. This allows booting without
14857 /etc/fstab and without root= on the kernel command line on
14858 systems prepared appropriately.
14860 * systemd-nspawn gained a new --image= switch which allows
14861 booting up disk images and Linux installations on any block
14862 device that follow the Discoverable Partitions Specification
14863 (see above). This means that installations made with
14864 appropriately updated installers may now be started and
14865 deployed using container managers, completely
14866 unmodified. (We hope that libvirt-lxc will add support for
14867 this feature soon, too.)
14869 * systemd-nspawn gained a new --network-macvlan= setting to
14870 set up a private macvlan interface for the
14871 container. Similarly, systemd-networkd gained a new
14872 Kind=macvlan setting in .netdev files.
14874 * systemd-networkd now supports configuring local addresses
14877 * A new tool systemd-network-wait-online has been added to
14878 synchronously wait for network connectivity using
14881 * The sd-bus.h bus API gained a new sd_bus_track object for
14882 tracking the lifecycle of bus peers. Note that sd-bus.h is
14883 still not a public API though (unless you specify
14884 --enable-kdbus on the configure command line, which however
14885 voids your warranty and you get no API stability guarantee).
14887 * The $XDG_RUNTIME_DIR runtime directories for each user are
14888 now individual tmpfs instances, which has the benefit of
14889 introducing separate pools for each user, with individual
14890 size limits, and thus making sure that unprivileged clients
14891 can no longer negatively impact the system or other users by
14892 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
14893 RuntimeDirectorySize= has been introduced that allows
14894 controlling the default size limit for all users. It
14895 defaults to 10% of the available physical memory. This is no
14896 replacement for quotas on tmpfs though (which the kernel
14897 still does not support), as /dev/shm and /tmp are still
14898 shared resources used by both the system and unprivileged
14901 * logind will now automatically turn off automatic suspending
14902 on laptop lid close when more than one display is
14903 connected. This was previously expected to be implemented
14904 individually in desktop environments (such as GNOME),
14905 however has been added to logind now, in order to fix a
14906 boot-time race where a desktop environment might not have
14907 been started yet and thus not been able to take an inhibitor
14908 lock at the time where logind already suspends the system
14909 due to a closed lid.
14911 * logind will now wait at least 30s after each system
14912 suspend/resume cycle, and 3min after system boot before
14913 suspending the system due to a closed laptop lid. This
14914 should give USB docking stations and similar enough time to
14915 be probed and configured after system resume and boot in
14916 order to then act as suspend blocker.
14918 * systemd-run gained a new --property= setting which allows
14919 initialization of resource control properties (and others)
14920 for the created scope or service unit. Example: "systemd-run
14921 --property=BlockIOWeight=10 updatedb" may be used to run
14922 updatedb at a low block IO scheduling weight.
14924 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
14925 now also work in --scope mode.
14927 * When systemd is compiled with kdbus support, basic support
14928 for enforced policies is now in place. (Note that enabling
14929 kdbus still voids your warranty and no API compatibility
14930 promises are made.)
14932 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
14933 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
14934 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
14935 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
14936 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
14937 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
14938 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
14939 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
14940 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
14941 Zbigniew Jędrzejewski-Szmek
14943 — Berlin, 2014-03-12
14947 * systemd will now relabel /dev after loading the SMACK policy
14948 according to SMACK rules.
14950 * A new unit file option AppArmorProfile= has been added to
14951 set the AppArmor profile for the processes of a unit.
14953 * A new condition check ConditionArchitecture= has been added
14954 to conditionalize units based on the system architecture, as
14955 reported by uname()'s "machine" field.
14957 * systemd-networkd now supports matching on the system
14958 virtualization, architecture, kernel command line, hostname
14961 * logind is now a lot more aggressive when suspending the
14962 machine due to a closed laptop lid. Instead of acting only
14963 on the lid close action, it will continuously watch the lid
14964 status and act on it. This is useful for laptops where the
14965 power button is on the outside of the chassis so that it can
14966 be reached without opening the lid (such as the Lenovo
14967 Yoga). On those machines, logind will now immediately
14968 re-suspend the machine if the power button has been
14969 accidentally pressed while the laptop was suspended and in a
14970 backpack or similar.
14972 * logind will now watch SW_DOCK switches and inhibit reaction
14973 to the lid switch if it is pressed. This means that logind
14974 will not suspend the machine anymore if the lid is closed
14975 and the system is docked, if the laptop supports SW_DOCK
14976 notifications via the input layer. Note that ACPI docking
14977 stations do not generate this currently. Also note that this
14978 logic is usually not fully sufficient and Desktop
14979 Environments should take a lid switch inhibitor lock when an
14980 external display is connected, as systemd will not watch
14983 * nspawn will now make use of the devices cgroup controller by
14984 default, and only permit creation of and access to the usual
14985 API device nodes like /dev/null or /dev/random, as well as
14986 access to (but not creation of) the pty devices.
14988 * We will now ship a default .network file for
14989 systemd-networkd that automatically configures DHCP for
14990 network interfaces created by nspawn's --network-veth or
14991 --network-bridge= switches.
14993 * systemd will now understand the usual M, K, G, T suffixes
14994 according to SI conventions (i.e. to the base 1000) when
14995 referring to throughput and hardware metrics. It will stay
14996 with IEC conventions (i.e. to the base 1024) for software
14997 metrics, according to what is customary according to
14998 Wikipedia. We explicitly document which base applies for
14999 each configuration option.
15001 * The DeviceAllow= setting in unit files now supports a syntax to
15002 allow-list an entire group of devices node majors at once, based on
15003 the /proc/devices listing. For example, with the string "char-pts",
15004 it is now possible to allow-list all current and future pseudo-TTYs
15007 * sd-event learned a new "post" event source. Event sources of
15008 this type are triggered by the dispatching of any event
15009 source of a type that is not "post". This is useful for
15010 implementing clean-up and check event sources that are
15011 triggered by other work being done in the program.
15013 * systemd-networkd is no longer statically enabled, but uses
15014 the usual [Install] sections so that it can be
15015 enabled/disabled using systemctl. It still is enabled by
15018 * When creating a veth interface pair with systemd-nspawn, the
15019 host side will now be prefixed with "vb-" if
15020 --network-bridge= is used, and with "ve-" if --network-veth
15021 is used. This way, it is easy to distinguish these cases on
15022 the host, for example to apply different configuration to
15023 them with systemd-networkd.
15025 * The compatibility libraries for libsystemd-journal.so,
15026 libsystem-id128.so, libsystemd-login.so and
15027 libsystemd-daemon.so do not make use of IFUNC
15028 anymore. Instead, we now build libsystemd.so multiple times
15029 under these alternative names. This means that the footprint
15030 is drastically increased, but given that these are
15031 transitional compatibility libraries, this should not matter
15032 much. This change has been made necessary to support the ARM
15033 platform for these compatibility libraries, as the ARM
15034 toolchain is not really at the same level as the toolchain
15035 for other architectures like x86 and does not support
15036 IFUNC. Please make sure to use --enable-compat-libs only
15037 during a transitional period!
15039 * The .include syntax has been deprecated and is not documented
15040 anymore. Drop-in files in .d directories should be used instead.
15042 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
15043 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
15044 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
15045 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
15046 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
15047 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
15048 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
15049 Zbigniew Jędrzejewski-Szmek
15051 — Berlin, 2014-02-24
15055 * A new component "systemd-networkd" has been added that can
15056 be used to configure local network interfaces statically or
15057 via DHCP. It is capable of bringing up bridges, VLANs, and
15058 bonding. Currently, no hook-ups for interactive network
15059 configuration are provided. Use this for your initrd,
15060 container, embedded, or server setup if you need a simple,
15061 yet powerful, network configuration solution. This
15062 configuration subsystem is quite nifty, as it allows wildcard
15063 hotplug matching in interfaces. For example, with a single
15064 configuration snippet, you can configure that all Ethernet
15065 interfaces showing up are automatically added to a bridge,
15066 or similar. It supports link-sensing and more.
15068 * A new tool "systemd-socket-proxyd" has been added which can
15069 act as a bidirectional proxy for TCP sockets. This is
15070 useful for adding socket activation support to services that
15071 do not actually support socket activation, including virtual
15072 machines and the like.
15074 * Add a new tool to save/restore rfkill state on
15077 * Save/restore state of keyboard backlights in addition to
15078 display backlights on shutdown/boot.
15080 * udev learned a new SECLABEL{} construct to label device
15081 nodes with a specific security label when they appear. For
15082 now, only SECLABEL{selinux} is supported, but the syntax is
15083 prepared for additional security frameworks.
15085 * udev gained a new scheme to configure link-level attributes
15086 from files in /etc/systemd/network/*.link. These files can
15087 match against MAC address, device path, driver name and type,
15088 and will apply attributes like the naming policy, link speed,
15089 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
15090 address assignment policy (randomized, …).
15092 * The configuration of network interface naming rules for
15093 "permanent interface names" has changed: a new NamePolicy=
15094 setting in the [Link] section of .link files determines the
15095 priority of possible naming schemes (onboard, slot, MAC,
15096 path). The default value of this setting is determined by
15097 /usr/lib/net/links/99-default.link. Old
15098 80-net-name-slot.rules udev configuration file has been
15099 removed, so local configuration overriding this file should
15100 be adapted to override 99-default.link instead.
15102 * When the User= switch is used in a unit file, also
15103 initialize $SHELL= based on the user database entry.
15105 * systemd no longer depends on libdbus. All communication is
15106 now done with sd-bus, systemd's low-level bus library
15109 * kdbus support has been added to PID 1 itself. When kdbus is
15110 enabled, this causes PID 1 to set up the system bus and
15111 enable support for a new ".busname" unit type that
15112 encapsulates bus name activation on kdbus. It works a little
15113 bit like ".socket" units, except for bus names. A new
15114 generator has been added that converts classic dbus1 service
15115 activation files automatically into native systemd .busname
15116 and .service units.
15118 * sd-bus: add a light-weight vtable implementation that allows
15119 defining objects on the bus with a simple static const
15120 vtable array of its methods, signals and properties.
15122 * systemd will not generate or install static dbus
15123 introspection data anymore to /usr/share/dbus-1/interfaces,
15124 as the precise format of these files is unclear, and
15125 nothing makes use of it.
15127 * A proxy daemon is now provided to proxy clients connecting
15128 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
15129 compatibility with classic D-Bus.
15131 * A bus driver implementation has been added that supports the
15132 classic D-Bus bus driver calls on kdbus, also for
15133 compatibility purposes.
15135 * A new API "sd-event.h" has been added that implements a
15136 minimal event loop API built around epoll. It provides a
15137 couple of features that direct epoll usage is lacking:
15138 prioritization of events, scales to large numbers of timer
15139 events, per-event timer slack (accuracy), system-wide
15140 coalescing of timer events, exit handlers, watchdog
15141 supervision support using systemd's sd_notify() API, child
15144 * A new API "sd-rntl.h" has been added that provides an API
15145 around the route netlink interface of the kernel, similar in
15146 style to "sd-bus.h".
15148 * A new API "sd-dhcp-client.h" has been added that provides a
15149 small DHCPv4 client-side implementation. This is used by
15150 "systemd-networkd".
15152 * There is a new kernel command line option
15153 "systemd.restore_state=0|1". When set to "0", none of the
15154 systemd tools will restore saved runtime state to hardware
15155 devices. More specifically, the rfkill and backlight states
15158 * The FsckPassNo= compatibility option in mount/service units
15159 has been removed. The fstab generator will now add the
15160 necessary dependencies automatically, and does not require
15161 PID1's support for that anymore.
15163 * journalctl gained a new switch, --list-boots, that lists
15164 recent boots with their times and boot IDs.
15166 * The various tools like systemctl, loginctl, timedatectl,
15167 busctl, systemd-run, … have gained a new switch "-M" to
15168 connect to a specific, local OS container (as direct
15169 connection, without requiring SSH). This works on any
15170 container that is registered with machined, such as those
15171 created by libvirt-lxc or nspawn.
15173 * systemd-run and systemd-analyze also gained support for "-H"
15174 to connect to remote hosts via SSH. This is particularly
15175 useful for systemd-run because it enables queuing of jobs
15176 onto remote systems.
15178 * machinectl gained a new command "login" to open a getty
15179 login in any local container. This works with any container
15180 that is registered with machined (such as those created by
15181 libvirt-lxc or nspawn), and which runs systemd inside.
15183 * machinectl gained a new "reboot" command that may be used to
15184 trigger a reboot on a specific container that is registered
15185 with machined. This works on any container that runs an init
15186 system of some kind.
15188 * systemctl gained a new "list-timers" command to print a nice
15189 listing of installed timer units with the times they elapse
15192 * Alternative reboot() parameters may now be specified on the
15193 "systemctl reboot" command line and are passed to the
15194 reboot() system call.
15196 * systemctl gained a new --job-mode= switch to configure the
15197 mode to queue a job with. This is a more generic version of
15198 --fail, --irreversible, and --ignore-dependencies, which are
15199 still available but not advertised anymore.
15201 * /etc/systemd/system.conf gained new settings to configure
15202 various default timeouts of units, as well as the default
15203 start limit interval and burst. These may still be overridden
15206 * PID1 will now export on the bus profile data of the security
15207 policy upload process (such as the SELinux policy upload to
15210 * journald: when forwarding logs to the console, include
15211 timestamps (following the setting in
15212 /sys/module/printk/parameters/time).
15214 * OnCalendar= in timer units now understands the special
15215 strings "yearly" and "annually". (Both are equivalent)
15217 * The accuracy of timer units is now configurable with the new
15218 AccuracySec= setting. It defaults to 1min.
15220 * A new dependency type JoinsNamespaceOf= has been added that
15221 allows running two services within the same /tmp and network
15222 namespace, if PrivateNetwork= or PrivateTmp= are used.
15224 * A new command "cat" has been added to systemctl. It outputs
15225 the original unit file of a unit, and concatenates the
15226 contents of additional "drop-in" unit file snippets, so that
15227 the full configuration is shown.
15229 * systemctl now supports globbing on the various "list-xyz"
15230 commands, like "list-units" or "list-sockets", as well as on
15231 those commands which take multiple unit names.
15233 * journalctl's --unit= switch gained support for globbing.
15235 * All systemd daemons now make use of the watchdog logic so
15236 that systemd automatically notices when they hang.
15238 * If the $container_ttys environment variable is set,
15239 getty-generator will automatically spawn a getty for each
15240 listed tty. This is useful for container managers to request
15241 login gettys to be spawned on as many ttys as needed.
15243 * %h, %s, %U specifier support is not available anymore when
15244 used in unit files for PID 1. This is because NSS calls are
15245 not safe from PID 1. They stay available for --user
15246 instances of systemd, and as special case for the root user.
15248 * loginctl gained a new "--no-legend" switch to turn off output
15249 of the legend text.
15251 * The "sd-login.h" API gained three new calls:
15252 sd_session_is_remote(), sd_session_get_remote_user(),
15253 sd_session_get_remote_host() to query information about
15256 * The udev hardware database now also carries vendor/product
15257 information of SDIO devices.
15259 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
15260 determine whether watchdog notifications are requested by
15261 the system manager.
15263 * Socket-activated per-connection services now include a
15264 short description of the connection parameters in the
15267 * tmpfiles gained a new "--boot" option. When this is not used,
15268 only lines where the command character is not suffixed with
15269 "!" are executed. When this option is specified, those
15270 options are executed too. This partitions tmpfiles
15271 directives into those that can be safely executed at any
15272 time, and those which should be run only at boot (for
15273 example, a line that creates /run/nologin).
15275 * A new API "sd-resolve.h" has been added which provides a simple
15276 asynchronous wrapper around glibc NSS hostname resolution
15277 calls, such as getaddrinfo(). In contrast to glibc's
15278 getaddrinfo_a(), it does not use signals. In contrast to most
15279 other asynchronous name resolution libraries, this one does
15280 not reimplement DNS, but reuses NSS, so that alternate
15281 hostname resolution systems continue to work, such as mDNS,
15282 LDAP, etc. This API is based on libasyncns, but it has been
15283 cleaned up for inclusion in systemd.
15285 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
15286 "sd-daemon.h" are no longer found in individual libraries
15287 libsystemd-journal.so, libsystemd-login.so,
15288 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
15289 merged them into a single library, libsystemd.so, which
15290 provides all symbols. The reason for this is cyclic
15291 dependencies, as these libraries tend to use each other's
15292 symbols. So far, we have managed to workaround that by linking
15293 a copy of a good part of our code into each of these
15294 libraries again and again, which, however, makes certain
15295 things hard to do, like sharing static variables. Also, it
15296 substantially increases footprint. With this change, there
15297 is only one library for the basic APIs systemd
15298 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
15299 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
15300 library as well, however are subject to the --enable-kdbus
15301 switch (see below). Note that "sd-dhcp-client.h" is not part
15302 of this library (this is because it only consumes, never
15303 provides, services of/to other APIs). To make the transition
15304 easy from the separate libraries to the unified one, we
15305 provide the --enable-compat-libs compile-time switch which
15306 will generate stub libraries that are compatible with the
15307 old ones but redirect all calls to the new one.
15309 * All of the kdbus logic and the new APIs "sd-bus.h",
15310 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
15311 and "sd-utf8.h" are compile-time optional via the
15312 "--enable-kdbus" switch, and they are not compiled in by
15313 default. To make use of kdbus, you have to explicitly enable
15314 the switch. Note however, that neither the kernel nor the
15315 userspace API for all of this is considered stable yet. We
15316 want to maintain the freedom to still change the APIs for
15317 now. By specifying this build-time switch, you acknowledge
15318 that you are aware of the instability of the current
15321 * Also, note that while kdbus is pretty much complete,
15322 it lacks one thing: proper policy support. This means you
15323 can build a fully working system with all features; however,
15324 it will be highly insecure. Policy support will be added in
15325 one of the next releases, at the same time that we will
15326 declare the APIs stable.
15328 * When the kernel command line argument "kdbus" is specified,
15329 systemd will automatically load the kdbus.ko kernel module. At
15330 this stage of development, it is only useful for testing kdbus
15331 and should not be used in production. Note: if "--enable-kdbus"
15332 is specified, and the kdbus.ko kernel module is available, and
15333 "kdbus" is added to the kernel command line, the entire system
15334 runs with kdbus instead of dbus-daemon, with the above mentioned
15335 problem of missing the system policy enforcement. Also a future
15336 version of kdbus.ko or a newer systemd will not be compatible with
15337 each other, and will unlikely be able to boot the machine if only
15338 one of them is updated.
15340 * systemctl gained a new "import-environment" command which
15341 uploads the caller's environment (or parts thereof) into the
15342 service manager so that it is inherited by services started
15343 by the manager. This is useful to upload variables like
15344 $DISPLAY into the user service manager.
15346 * A new PrivateDevices= switch has been added to service units
15347 which allows running a service with a namespaced /dev
15348 directory that does not contain any device nodes for
15349 physical devices. More specifically, it only includes devices
15350 such as /dev/null, /dev/urandom, and /dev/zero which are API
15353 * logind has been extended to support behaviour like VT
15354 switching on seats that do not support a VT. This makes
15355 multi-session available on seats that are not the first seat
15356 (seat0), and on systems where kernel support for VTs has
15357 been disabled at compile-time.
15359 * If a process holds a delay lock for system sleep or shutdown
15360 and fails to release it in time, we will now log its
15361 identity. This makes it easier to identify processes that
15362 cause slow suspends or power-offs.
15364 * When parsing /etc/crypttab, support for a new key-slot=
15365 option as supported by Debian is added. It allows indicating
15366 which LUKS slot to use on disk, speeding up key loading.
15368 * The sd_journal_sendv() API call has been checked and
15369 officially declared to be async-signal-safe so that it may
15370 be invoked from signal handlers for logging purposes.
15372 * Boot-time status output is now enabled automatically after a
15373 short timeout if boot does not progress, in order to give
15374 the user an indication what she or he is waiting for.
15376 * The boot-time output has been improved to show how much time
15377 remains until jobs expire.
15379 * The KillMode= switch in service units gained a new possible
15380 value "mixed". If set, and the unit is shut down, then the
15381 initial SIGTERM signal is sent only to the main daemon
15382 process, while the following SIGKILL signal is sent to
15383 all remaining processes of the service.
15385 * When a scope unit is registered, a new property "Controller"
15386 may be set. If set to a valid bus name, systemd will send a
15387 RequestStop() signal to this name when it would like to shut
15388 down the scope. This may be used to hook manager logic into
15389 the shutdown logic of scope units. Also, scope units may now
15390 be put in a special "abandoned" state, in which case the
15391 manager process which created them takes no further
15392 responsibilities for it.
15394 * When reading unit files, systemd will now verify
15395 the access mode of these files, and warn about certain
15396 suspicious combinations. This has been added to make it
15397 easier to track down packaging bugs where unit files are
15398 marked executable or world-writable.
15400 * systemd-nspawn gained a new "--setenv=" switch to set
15401 container-wide environment variables. The similar option in
15402 systemd-activate was renamed from "--environment=" to
15403 "--setenv=" for consistency.
15405 * systemd-nspawn has been updated to create a new kdbus domain
15406 for each container that is invoked, thus allowing each
15407 container to have its own set of system and user buses,
15408 independent of the host.
15410 * systemd-nspawn gained a new --drop-capability= switch to run
15411 the container with less capabilities than the default. Both
15412 --drop-capability= and --capability= now take the special
15413 string "all" for dropping or keeping all capabilities.
15415 * systemd-nspawn gained new switches for executing containers
15416 with specific SELinux labels set.
15418 * systemd-nspawn gained a new --quiet switch to not generate
15419 any additional output but the container's own console
15422 * systemd-nspawn gained a new --share-system switch to run a
15423 container without PID namespacing enabled.
15425 * systemd-nspawn gained a new --register= switch to control
15426 whether the container is registered with systemd-machined or
15427 not. This is useful for containers that do not run full
15428 OS images, but only specific apps.
15430 * systemd-nspawn gained a new --keep-unit which may be used
15431 when invoked as the only program from a service unit, and
15432 results in registration of the unit service itself in
15433 systemd-machined, instead of a newly opened scope unit.
15435 * systemd-nspawn gained a new --network-interface= switch for
15436 moving arbitrary interfaces to the container. The new
15437 --network-veth switch creates a virtual Ethernet connection
15438 between host and container. The new --network-bridge=
15439 switch then allows assigning the host side of this virtual
15440 Ethernet connection to a bridge device.
15442 * systemd-nspawn gained a new --personality= switch for
15443 setting the kernel personality for the container. This is
15444 useful when running a 32-bit container on a 64-bit host. A
15445 similar option Personality= is now also available for service
15448 * logind will now also track a "Desktop" identifier for each
15449 session which encodes the desktop environment of it. This is
15450 useful for desktop environments that want to identify
15451 multiple running sessions of itself easily.
15453 * A new SELinuxContext= setting for service units has been
15454 added that allows setting a specific SELinux execution
15455 context for a service.
15457 * Most systemd client tools will now honour $SYSTEMD_LESS for
15458 settings of the "less" pager. By default, these tools will
15459 override $LESS to allow certain operations to work, such as
15460 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
15461 influence this logic.
15463 * systemd's "seccomp" hook-up has been changed to make use of
15464 the libseccomp library instead of using its own
15465 implementation. This has benefits for portability among
15468 * For usage together with SystemCallFilter=, a new
15469 SystemCallErrorNumber= setting has been introduced that
15470 allows configuration of a system error number to be returned
15471 on filtered system calls, instead of immediately killing the
15472 process. Also, SystemCallArchitectures= has been added to
15473 limit access to system calls of a particular architecture
15474 (in order to turn off support for unused secondary
15475 architectures). There is also a global
15476 SystemCallArchitectures= setting in system.conf now to turn
15477 off support for non-native system calls system-wide.
15479 * systemd requires a kernel with a working name_to_handle_at(),
15480 please see the kernel config requirements in the README file.
15482 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
15483 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
15484 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
15485 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
15486 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
15487 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
15488 Elia Pinto, Florian Weimer, George McCollister, Goffredo
15489 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
15490 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
15491 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
15492 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
15493 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
15494 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
15495 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
15496 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
15497 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
15498 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
15499 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
15500 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
15501 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
15502 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
15503 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
15504 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
15505 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
15507 — Berlin, 2014-02-20
15511 * logind has gained support for facilitating privileged input
15512 and drm device access for unprivileged clients. This work is
15513 useful to allow Wayland display servers (and similar
15514 programs, such as kmscon) to run under the user's ID and
15515 access input and drm devices which are normally
15516 protected. When this is used (and the kernel is new enough)
15517 logind will "mute" IO on the file descriptors passed to
15518 Wayland as long as it is in the background and "unmute" it
15519 if it returns into the foreground. This allows secure
15520 session switching without allowing background sessions to
15521 eavesdrop on input and display data. This also introduces
15522 session switching support if VT support is turned off in the
15523 kernel, and on seats that are not seat0.
15525 * A new kernel command line option luks.options= is understood
15526 now which allows specifying LUKS options for usage for LUKS
15527 encrypted partitions specified with luks.uuid=.
15529 * tmpfiles.d(5) snippets may now use specifier expansion in
15530 path names. More specifically %m, %b, %H, %v, are now
15531 replaced by the local machine id, boot id, hostname, and
15532 kernel version number.
15534 * A new tmpfiles.d(5) command "m" has been introduced which
15535 may be used to change the owner/group/access mode of a file
15536 or directory if it exists, but do nothing if it does not.
15538 * This release removes high-level support for the
15539 MemorySoftLimit= cgroup setting. The underlying kernel
15540 cgroup attribute memory.soft_limit= is currently badly
15541 designed and likely to be removed from the kernel API in its
15542 current form, hence we should not expose it for now.
15544 * The memory.use_hierarchy cgroup attribute is now enabled for
15545 all cgroups systemd creates in the memory cgroup
15546 hierarchy. This option is likely to be come the built-in
15547 default in the kernel anyway, and the non-hierarchical mode
15548 never made much sense in the intrinsically hierarchical
15551 * A new field _SYSTEMD_SLICE= is logged along with all journal
15552 messages containing the slice a message was generated
15553 from. This is useful to allow easy per-customer filtering of
15554 logs among other things.
15556 * systemd-journald will no longer adjust the group of journal
15557 files it creates to the "systemd-journal" group. Instead we
15558 rely on the journal directory to be owned by the
15559 "systemd-journal" group, and its setgid bit set, so that the
15560 kernel file system layer will automatically enforce that
15561 journal files inherit this group assignment. The reason for
15562 this change is that we cannot allow NSS look-ups from
15563 journald which would be necessary to resolve
15564 "systemd-journal" to a numeric GID, because this might
15565 create deadlocks if NSS involves synchronous queries to
15566 other daemons (such as nscd, or sssd) which in turn are
15567 logging clients of journald and might block on it, which
15568 would then dead lock. A tmpfiles.d(5) snippet included in
15569 systemd will make sure the setgid bit and group are
15570 properly set on the journal directory if it exists on every
15571 boot. However, we recommend adjusting it manually after
15572 upgrades too (or from RPM scriptlets), so that the change is
15573 not delayed until next reboot.
15575 * Backlight and random seed files in /var/lib/ have moved into
15576 the /var/lib/systemd/ directory, in order to centralize all
15577 systemd generated files in one directory.
15579 * Boot time performance measurements (as displayed by
15580 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
15581 performance information if that's available to determine how
15582 much time BIOS and boot loader initialization required. With
15583 a sufficiently new BIOS you hence no longer need to boot
15584 with Gummiboot to get access to such information.
15586 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
15587 Cristian Rodríguez, Dave Reisner, David Herrmann, David
15588 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
15589 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
15590 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
15591 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
15592 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
15594 — Berlin, 2013-10-02
15598 * The Restart= option for services now understands a new
15599 on-watchdog setting, which will restart the service
15600 automatically if the service stops sending out watchdog keep
15601 alive messages (as configured with WatchdogSec=).
15603 * The getty generator (which is responsible for bringing up a
15604 getty on configured serial consoles) will no longer only
15605 start a getty on the primary kernel console but on all
15606 others, too. This makes the order in which console= is
15607 specified on the kernel command line less important.
15609 * libsystemd-logind gained a new sd_session_get_vt() call to
15610 retrieve the VT number of a session.
15612 * If the option "tries=0" is set for an entry of /etc/crypttab
15613 its passphrase is queried indefinitely instead of any
15614 maximum number of tries.
15616 * If a service with a configure PID file terminates its PID
15617 file will now be removed automatically if it still exists
15618 afterwards. This should put an end to stale PID files.
15620 * systemd-run will now also take relative binary path names
15621 for execution and no longer insists on absolute paths.
15623 * InaccessibleDirectories= and ReadOnlyDirectories= now take
15624 paths that are optionally prefixed with "-" to indicate that
15625 it should not be considered a failure if they do not exist.
15627 * journalctl -o (and similar commands) now understands a new
15628 output mode "short-precise", it is similar to "short" but
15629 shows timestamps with usec accuracy.
15631 * The option "discard" (as known from Debian) is now
15632 synonymous to "allow-discards" in /etc/crypttab. In fact,
15633 "discard" is preferred now (since it is easier to remember
15636 * Some licensing clean-ups were made, so that more code is now
15637 LGPL-2.1 licensed than before.
15639 * A minimal tool to save/restore the display backlight
15640 brightness across reboots has been added. It will store the
15641 backlight setting as late as possible at shutdown, and
15642 restore it as early as possible during reboot.
15644 * A logic to automatically discover and enable home and swap
15645 partitions on GPT disks has been added. With this in place
15646 /etc/fstab becomes optional for many setups as systemd can
15647 discover certain partitions located on the root disk
15648 automatically. Home partitions are recognized under their
15649 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
15650 partitions are recognized under their GPT type ID
15651 0657fd6da4ab43c484e50933c84b4f4f.
15653 * systemd will no longer pass any environment from the kernel
15654 or initrd to system services. If you want to set an
15655 environment for all services, do so via the kernel command
15656 line systemd.setenv= assignment.
15658 * The systemd-sysctl tool no longer natively reads the file
15659 /etc/sysctl.conf. If desired, the file should be symlinked
15660 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
15661 legacy support by a symlink rather than built-in code, it
15662 also makes the otherwise hidden order of application of the
15663 different files visible. (Note that this partly reverts to a
15664 pre-198 application order of sysctl knobs!)
15666 * The "systemctl set-log-level" and "systemctl dump" commands
15667 have been moved to systemd-analyze.
15669 * systemd-run learned the new --remain-after-exit switch,
15670 which causes the scope unit not to be cleaned up
15671 automatically after the process terminated.
15673 * tmpfiles learned a new --exclude-prefix= switch to exclude
15674 certain paths from operation.
15676 * journald will now automatically flush all messages to disk
15677 as soon as a message at the log level CRIT, ALERT or EMERG
15680 Contributions from: Andrew Cook, Brandon Philips, Christian
15681 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
15682 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
15683 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
15684 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
15685 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
15686 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
15687 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
15688 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
15689 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
15690 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
15691 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
15692 William Giokas, Zbigniew Jędrzejewski-Szmek
15694 — Berlin, 2013-09-13
15698 * The documentation has been updated to cover the various new
15699 concepts introduced with 205.
15701 * Unit files now understand the new %v specifier which
15702 resolves to the kernel version string as returned by "uname
15705 * systemctl now supports filtering the unit list output by
15706 load state, active state and sub state, using the new
15707 --state= parameter.
15709 * "systemctl status" will now show the results of the
15710 condition checks (like ConditionPathExists= and similar) of
15711 the last start attempts of the unit. They are also logged to
15714 * "journalctl -b" may now be used to look for boot output of a
15715 specific boot. Try "journalctl -b -1" for the previous boot,
15716 but the syntax is substantially more powerful.
15718 * "journalctl --show-cursor" has been added which prints the
15719 cursor string the last shown log line. This may then be used
15720 with the new "journalctl --after-cursor=" switch to continue
15721 browsing logs from that point on.
15723 * "journalctl --force" may now be used to force regeneration
15726 * Creation of "dead" device nodes has been moved from udev
15727 into kmod and tmpfiles. Previously, udev would read the kmod
15728 databases to pre-generate dead device nodes based on meta
15729 information contained in kernel modules, so that these would
15730 be auto-loaded on access rather then at boot. As this
15731 does not really have much to do with the exposing actual
15732 kernel devices to userspace this has always been slightly
15733 alien in the udev codebase. Following the new scheme kmod
15734 will now generate a runtime snippet for tmpfiles from the
15735 module meta information and it now is tmpfiles' job to the
15736 create the nodes. This also allows overriding access and
15737 other parameters for the nodes using the usual tmpfiles
15738 facilities. As side effect this allows us to remove the
15739 CAP_SYS_MKNOD capability bit from udevd entirely.
15741 * logind's device ACLs may now be applied to these "dead"
15742 devices nodes too, thus finally allowing managed access to
15743 devices such as /dev/snd/sequencer without loading the
15744 backing module right-away.
15746 * A new RPM macro has been added that may be used to apply
15747 tmpfiles configuration during package installation.
15749 * systemd-detect-virt and ConditionVirtualization= now can
15750 detect User-Mode-Linux machines (UML).
15752 * journald will now implicitly log the effective capabilities
15753 set of processes in the message metadata.
15755 * systemd-cryptsetup has gained support for TrueCrypt volumes.
15757 * The initrd interface has been simplified (more specifically,
15758 support for passing performance data via environment
15759 variables and fsck results via files in /run has been
15760 removed). These features were non-essential, and are
15761 nowadays available in a much nicer way by having systemd in
15762 the initrd serialize its state and have the hosts systemd
15763 deserialize it again.
15765 * The udev "keymap" data files and tools to apply keyboard
15766 specific mappings of scan to key codes, and force-release
15767 scan code lists have been entirely replaced by a udev
15768 "keyboard" builtin and a hwdb data file.
15770 * systemd will now honour the kernel's "quiet" command line
15771 argument also during late shutdown, resulting in a
15772 completely silent shutdown when used.
15774 * There's now an option to control the SO_REUSEPORT socket
15775 option in .socket units.
15777 * Instance units will now automatically get a per-template
15778 subslice of system.slice unless something else is explicitly
15779 configured. For example, instances of sshd@.service will now
15780 implicitly be placed in system-sshd.slice rather than
15781 system.slice as before.
15783 * Test coverage support may now be enabled at build time.
15785 Contributions from: Dave Reisner, Frederic Crozat, Harald
15786 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
15787 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
15788 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
15789 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
15790 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
15791 Giokas, Zbigniew Jędrzejewski-Szmek
15793 — Berlin, 2013-07-23
15797 * Two new unit types have been introduced:
15799 Scope units are very similar to service units, however, are
15800 created out of pre-existing processes — instead of PID 1
15801 forking off the processes. By using scope units it is
15802 possible for system services and applications to group their
15803 own child processes (worker processes) in a powerful way
15804 which then maybe used to organize them, or kill them
15805 together, or apply resource limits on them.
15807 Slice units may be used to partition system resources in an
15808 hierarchical fashion and then assign other units to them. By
15809 default there are now three slices: system.slice (for all
15810 system services), user.slice (for all user sessions),
15811 machine.slice (for VMs and containers).
15813 Slices and scopes have been introduced primarily in
15814 context of the work to move cgroup handling to a
15815 single-writer scheme, where only PID 1
15816 creates/removes/manages cgroups.
15818 * There's a new concept of "transient" units. In contrast to
15819 normal units these units are created via an API at runtime,
15820 not from configuration from disk. More specifically this
15821 means it is now possible to run arbitrary programs as
15822 independent services, with all execution parameters passed
15823 in via bus APIs rather than read from disk. Transient units
15824 make systemd substantially more dynamic then it ever was,
15825 and useful as a general batch manager.
15827 * logind has been updated to make use of scope and slice units
15828 for managing user sessions. As a user logs in he will get
15829 his own private slice unit, to which all sessions are added
15830 as scope units. We also added support for automatically
15831 adding an instance of user@.service for the user into the
15832 slice. Effectively logind will no longer create cgroup
15833 hierarchies on its own now, it will defer entirely to PID 1
15834 for this by means of scope, service and slice units. Since
15835 user sessions this way become entities managed by PID 1
15836 the output of "systemctl" is now a lot more comprehensive.
15838 * A new mini-daemon "systemd-machined" has been added which
15839 may be used by virtualization managers to register local
15840 VMs/containers. nspawn has been updated accordingly, and
15841 libvirt will be updated shortly. machined will collect a bit
15842 of meta information about the VMs/containers, and assign
15843 them their own scope unit (see above). The collected
15844 meta-data is then made available via the "machinectl" tool,
15845 and exposed in "ps" and similar tools. machined/machinectl
15846 is compile-time optional.
15848 * As discussed earlier, the low-level cgroup configuration
15849 options ControlGroup=, ControlGroupModify=,
15850 ControlGroupPersistent=, ControlGroupAttribute= have been
15851 removed. Please use high-level attribute settings instead as
15852 well as slice units.
15854 * A new bus call SetUnitProperties() has been added to alter
15855 various runtime parameters of a unit. This is primarily
15856 useful to alter cgroup parameters dynamically in a nice way,
15857 but will be extended later on to make more properties
15858 modifiable at runtime. systemctl gained a new set-properties
15859 command that wraps this call.
15861 * A new tool "systemd-run" has been added which can be used to
15862 run arbitrary command lines as transient services or scopes,
15863 while configuring a number of settings via the command
15864 line. This tool is currently very basic, however already
15865 very useful. We plan to extend this tool to even allow
15866 queuing of execution jobs with time triggers from the
15867 command line, similar in fashion to "at".
15869 * nspawn will now inform the user explicitly that kernels with
15870 audit enabled break containers, and suggest the user to turn
15873 * Support for detecting the IMA and AppArmor security
15874 frameworks with ConditionSecurity= has been added.
15876 * journalctl gained a new "-k" switch for showing only kernel
15877 messages, mimicking dmesg output; in addition to "--user"
15878 and "--system" switches for showing only user's own logs
15881 * systemd-delta can now show information about drop-in
15882 snippets extending unit files.
15884 * libsystemd-bus has been substantially updated but is still
15885 not available as public API.
15887 * systemd will now look for the "debug" argument on the kernel
15888 command line and enable debug logging, similar to what
15889 "systemd.log_level=debug" already did before.
15891 * "systemctl set-default", "systemctl get-default" has been
15892 added to configure the default.target symlink, which
15893 controls what to boot into by default.
15895 * "systemctl set-log-level" has been added as a convenient
15896 way to raise and lower systemd logging threshold.
15898 * "systemd-analyze plot" will now show the time the various
15899 generators needed for execution, as well as information
15900 about the unit file loading.
15902 * libsystemd-journal gained a new sd_journal_open_files() call
15903 for opening specific journal files. journactl also gained a
15904 new switch to expose this new functionality. Previously we
15905 only supported opening all files from a directory, or all
15906 files from the system, as opening individual files only is
15907 racy due to journal file rotation.
15909 * systemd gained the new DefaultEnvironment= setting in
15910 /etc/systemd/system.conf to set environment variables for
15913 * If a privileged process logs a journal message with the
15914 OBJECT_PID= field set, then journald will automatically
15915 augment this with additional OBJECT_UID=, OBJECT_GID=,
15916 OBJECT_COMM=, OBJECT_EXE=, … fields. This is useful if
15917 system services want to log events about specific client
15918 processes. journactl/systemctl has been updated to make use
15919 of this information if all log messages regarding a specific
15922 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
15923 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
15924 Reisner, David Coppa, David King, David Strauss, Eelco
15925 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
15926 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
15927 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
15928 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
15929 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
15930 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
15931 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
15932 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
15933 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
15934 Łukasz Stelmach, 장동준
15938 * The Python bindings gained some minimal support for the APIs
15939 exposed by libsystemd-logind.
15941 * ConditionSecurity= gained support for detecting SMACK. Since
15942 this condition already supports SELinux and AppArmor we only
15943 miss IMA for this. Patches welcome!
15945 Contributions from: Karol Lewandowski, Lennart Poettering,
15946 Zbigniew Jędrzejewski-Szmek
15950 * systemd-nspawn will now create /etc/resolv.conf if
15951 necessary, before bind-mounting the host's file onto it.
15953 * systemd-nspawn will now store meta information about a
15954 container on the container's cgroup as extended attribute
15955 fields, including the root directory.
15957 * The cgroup hierarchy has been reworked in many ways. All
15958 objects any of the components systemd creates in the cgroup
15959 tree are now suffixed. More specifically, user sessions are
15960 now placed in cgroups suffixed with ".session", users in
15961 cgroups suffixed with ".user", and nspawn containers in
15962 cgroups suffixed with ".nspawn". Furthermore, all cgroup
15963 names are now escaped in a simple scheme to avoid collision
15964 of userspace object names with kernel filenames. This work
15965 is preparation for making these objects relocatable in the
15966 cgroup tree, in order to allow easy resource partitioning of
15967 these objects without causing naming conflicts.
15969 * systemctl list-dependencies gained the new switches
15970 --plain, --reverse, --after and --before.
15972 * systemd-inhibit now shows the process name of processes that
15973 have taken an inhibitor lock.
15975 * nss-myhostname will now also resolve "localhost"
15976 implicitly. This makes /etc/hosts an optional file and
15977 nicely handles that on IPv6 ::1 maps to both "localhost" and
15978 the local hostname.
15980 * libsystemd-logind.so gained a new call
15981 sd_get_machine_names() to enumerate running containers and
15982 VMs (currently only supported by very new libvirt and
15983 nspawn). sd_login_monitor can now be used to watch
15984 VMs/containers coming and going.
15986 * .include is not allowed recursively anymore, and only in
15987 unit files. Usually it is better to use drop-in snippets in
15988 .d/*.conf anyway, as introduced with systemd 198.
15990 * systemd-analyze gained a new "critical-chain" command that
15991 determines the slowest chain of units run during system
15992 boot-up. It is very useful for tracking down where
15993 optimizing boot time is the most beneficial.
15995 * systemd will no longer allow manipulating service paths in
15996 the name=systemd:/system cgroup tree using ControlGroup= in
15997 units. (But is still fine with it in all other dirs.)
15999 * There's a new systemd-nspawn@.service service file that may
16000 be used to easily run nspawn containers as system
16001 services. With the container's root directory in
16002 /var/lib/container/foobar it is now sufficient to run
16003 "systemctl start systemd-nspawn@foobar.service" to boot it.
16005 * systemd-cgls gained a new parameter "--machine" to list only
16006 the processes within a certain container.
16008 * ConditionSecurity= now can check for "apparmor". We still
16009 are lacking checks for SMACK and IMA for this condition
16010 check though. Patches welcome!
16012 * A new configuration file /etc/systemd/sleep.conf has been
16013 added that may be used to configure which kernel operation
16014 systemd is supposed to execute when "suspend", "hibernate"
16015 or "hybrid-sleep" is requested. This makes the new kernel
16016 "freeze" state accessible to the user.
16018 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
16019 the passed argument if applicable.
16021 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
16022 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
16023 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
16024 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
16025 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
16026 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
16027 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
16032 * The output of 'systemctl list-jobs' got some polishing. The
16033 '--type=' argument may now be passed more than once. A new
16034 command 'systemctl list-sockets' has been added which shows
16035 a list of kernel sockets systemd is listening on with the
16036 socket units they belong to, plus the units these socket
16039 * The experimental libsystemd-bus library got substantial
16040 updates to work in conjunction with the (also experimental)
16041 kdbus kernel project. It works well enough to exchange
16042 messages with some sophistication. Note that kdbus is not
16043 ready yet, and the library is mostly an elaborate test case
16044 for now, and not installable.
16046 * systemd gained a new unit 'systemd-static-nodes.service'
16047 that generates static device nodes earlier during boot, and
16048 can run in conjunction with udev.
16050 * libsystemd-login gained a new call sd_pid_get_user_unit()
16051 to retrieve the user systemd unit a process is running
16052 in. This is useful for systems where systemd is used as
16055 * systemd-nspawn now places all containers in the new /machine
16056 top-level cgroup directory in the name=systemd
16057 hierarchy. libvirt will soon do the same, so that we get a
16058 uniform separation of /system, /user and /machine for system
16059 services, user processes and containers/virtual
16060 machines. This new cgroup hierarchy is also useful to stick
16061 stable names to specific container instances, which can be
16062 recognized later this way (this name may be controlled
16063 via systemd-nspawn's new -M switch). libsystemd-login also
16064 gained a new call sd_pid_get_machine_name() to retrieve the
16065 name of the container/VM a specific process belongs to.
16067 * bootchart can now store its data in the journal.
16069 * libsystemd-journal gained a new call
16070 sd_journal_add_conjunction() for AND expressions to the
16071 matching logic. This can be used to express more complex
16072 logical expressions.
16074 * journactl can now take multiple --unit= and --user-unit=
16077 * The cryptsetup logic now understands the "luks.key=" kernel
16078 command line switch for specifying a file to read the
16079 decryption key from. Also, if a configured key file is not
16080 found the tool will now automatically fall back to prompting
16083 * Python systemd.journal module was updated to wrap recently
16084 added functions from libsystemd-journal. The interface was
16085 changed to bring the low level interface in s.j._Reader
16086 closer to the C API, and the high level interface in
16087 s.j.Reader was updated to wrap and convert all data about
16090 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
16091 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
16092 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
16093 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
16094 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
16095 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
16099 * journalctl --update-catalog now understands a new --root=
16100 option to operate on catalogs found in a different root
16103 * During shutdown after systemd has terminated all running
16104 services a final killing loop kills all remaining left-over
16105 processes. We will now print the name of these processes
16106 when we send SIGKILL to them, since this usually indicates a
16109 * If /etc/crypttab refers to password files stored on
16110 configured mount points automatic dependencies will now be
16111 generated to ensure the specific mount is established first
16112 before the key file is attempted to be read.
16114 * 'systemctl status' will now show information about the
16115 network sockets a socket unit is listening on.
16117 * 'systemctl status' will also shown information about any
16118 drop-in configuration file for units. (Drop-In configuration
16119 files in this context are files such as
16120 /etc/systemd/system/foobar.service.d/*.conf)
16122 * systemd-cgtop now optionally shows summed up CPU times of
16123 cgroups. Press '%' while running cgtop to switch between
16124 percentage and absolute mode. This is useful to determine
16125 which cgroups use up the most CPU time over the entire
16126 runtime of the system. systemd-cgtop has also been updated
16127 to be 'pipeable' for processing with further shell tools.
16129 * 'hostnamectl set-hostname' will now allow setting of FQDN
16132 * The formatting and parsing of time span values has been
16133 changed. The parser now understands fractional expressions
16134 such as "5.5h". The formatter will now output fractional
16135 expressions for all time spans under 1min, i.e. "5.123456s"
16136 rather than "5s 123ms 456us". For time spans under 1s
16137 millisecond values are shown, for those under 1ms
16138 microsecond values are shown. This should greatly improve
16139 all time-related output of systemd.
16141 * libsystemd-login and libsystemd-journal gained new
16142 functions for querying the poll() events mask and poll()
16143 timeout value for integration into arbitrary event
16146 * localectl gained the ability to list available X11 keymaps
16147 (models, layouts, variants, options).
16149 * 'systemd-analyze dot' gained the ability to filter for
16150 specific units via shell-style globs, to create smaller,
16151 more useful graphs. I.e. it is now possible to create simple
16152 graphs of all the dependencies between only target units, or
16153 of all units that Avahi has dependencies with.
16155 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
16156 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
16157 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
16158 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
16159 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
16160 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
16161 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
16165 * The boot-time readahead implementation for rotating media
16166 will now read the read-ahead data in multiple passes which
16167 consist of all read requests made in equidistant time
16168 intervals. This means instead of strictly reading read-ahead
16169 data in its physical order on disk we now try to find a
16170 middle ground between physical and access time order.
16172 * /etc/os-release files gained a new BUILD_ID= field for usage
16173 on operating systems that provide continuous builds of OS
16176 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
16177 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
16178 William Douglas, Zbigniew Jędrzejewski-Szmek
16182 * systemd-python gained an API exposing libsystemd-daemon.
16184 * The SMACK setup logic gained support for uploading CIPSO
16187 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
16188 ReadOnlyDirectories= and InaccessibleDirectories= has
16189 changed. The private /tmp and /var/tmp directories are now
16190 shared by all processes of a service (which means
16191 ExecStartPre= may now leave data in /tmp that ExecStart= of
16192 the same service can still access). When a service is
16193 stopped its temporary directories are immediately deleted
16194 (normal clean-up with tmpfiles is still done in addition to
16197 * By default, systemd will now set a couple of sysctl
16198 variables in the kernel: the safe sysrq options are turned
16199 on, IP route verification is turned on, and source routing
16200 disabled. The recently added hardlink and softlink
16201 protection of the kernel is turned on. These settings should
16202 be reasonably safe, and good defaults for all new systems.
16204 * The predictable network naming logic may now be turned off
16205 with a new kernel command line switch: net.ifnames=0.
16207 * A new libsystemd-bus module has been added that implements a
16208 pretty complete D-Bus client library. For details see:
16210 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
16212 * journald will now explicitly flush the journal files to disk
16213 at the latest 5min after each write. The file will then also
16214 be marked offline until the next write. This should increase
16215 reliability in case of a crash. The synchronization delay
16216 can be configured via SyncIntervalSec= in journald.conf.
16218 * There's a new remote-fs-setup.target unit that can be used
16219 to pull in specific services when at least one remote file
16220 system is to be mounted.
16222 * There are new targets timers.target and paths.target as
16223 canonical targets to pull user timer and path units in
16224 from. This complements sockets.target with a similar
16225 purpose for socket units.
16227 * libudev gained a new call udev_device_set_attribute_value()
16228 to set sysfs attributes of a device.
16230 * The udev daemon now sets the default number of worker
16231 processes executed in parallel based on the number of available
16232 CPUs instead of the amount of available RAM. This is supposed
16233 to provide a more reliable default and limit a too aggressive
16234 parallelism for setups with 1000s of devices connected.
16236 Contributions from: Auke Kok, Colin Walters, Cristian
16237 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
16238 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
16239 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
16240 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
16241 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
16242 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
16243 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
16244 Zbigniew Jędrzejewski-Szmek
16248 * Configuration of unit files may now be extended via drop-in
16249 files without having to edit/override the unit files
16250 themselves. More specifically, if the administrator wants to
16251 change one value for a service file foobar.service he can
16252 now do so by dropping in a configuration snippet into
16253 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
16254 will load all these snippets and apply them on top of the
16255 main unit configuration file, possibly extending or
16256 overriding its settings. Using these drop-in snippets is
16257 generally nicer than the two earlier options for changing
16258 unit files locally: copying the files from
16259 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
16260 them there; or creating a new file in /etc/systemd/system/
16261 that incorporates the original one via ".include". Drop-in
16262 snippets into these .d/ directories can be placed in any
16263 directory systemd looks for units in, and the usual
16264 overriding semantics between /usr/lib, /etc and /run apply
16267 * Most unit file settings which take lists of items can now be
16268 reset by assigning the empty string to them. For example,
16269 normally, settings such as Environment=FOO=BAR append a new
16270 environment variable assignment to the environment block,
16271 each time they are used. By assigning Environment= the empty
16272 string the environment block can be reset to empty. This is
16273 particularly useful with the .d/*.conf drop-in snippets
16274 mentioned above, since this adds the ability to reset list
16275 settings from vendor unit files via these drop-ins.
16277 * systemctl gained a new "list-dependencies" command for
16278 listing the dependencies of a unit recursively.
16280 * Inhibitors are now honored and listed by "systemctl
16281 suspend", "systemctl poweroff" (and similar) too, not only
16282 GNOME. These commands will also list active sessions by
16285 * Resource limits (as exposed by the various control group
16286 controllers) can now be controlled dynamically at runtime
16287 for all units. More specifically, you can now use a command
16288 like "systemctl set-cgroup-attr foobar.service cpu.shares
16289 2000" to alter the CPU shares a specific service gets. These
16290 settings are stored persistently on disk, and thus allow the
16291 administrator to easily adjust the resource usage of
16292 services with a few simple commands. This dynamic resource
16293 management logic is also available to other programs via the
16294 bus. Almost any kernel cgroup attribute and controller is
16297 * systemd-vconsole-setup will now copy all font settings to
16298 all allocated VTs, where it previously applied them only to
16301 * libsystemd-login gained the new sd_session_get_tty() API
16304 * This release drops support for a few legacy or
16305 distribution-specific LSB facility names when parsing init
16306 scripts: $x-display-manager, $mail-transfer-agent,
16307 $mail-transport-agent, $mail-transfer-agent, $smtp,
16308 $null. Also, the mail-transfer-agent.target unit backing
16309 this has been removed. Distributions which want to retain
16310 compatibility with this should carry the burden for
16311 supporting this themselves and patch support for these back
16312 in, if they really need to. Also, the facilities $syslog and
16313 $local_fs are now ignored, since systemd does not support
16314 early-boot LSB init scripts anymore, and these facilities
16315 are implied anyway for normal services. syslog.target has
16318 * There are new bus calls on PID1's Manager object for
16319 cancelling jobs, and removing snapshot units. Previously,
16320 both calls were only available on the Job and Snapshot
16321 objects themselves.
16323 * systemd-journal-gatewayd gained SSL support.
16325 * The various "environment" files, such as /etc/locale.conf
16326 now support continuation lines with a backslash ("\") as
16327 last character in the line, similarly in style (but different)
16328 to how this is supported in shells.
16330 * For normal user processes the _SYSTEMD_USER_UNIT= field is
16331 now implicitly appended to every log entry logged. systemctl
16332 has been updated to filter by this field when operating on a
16333 user systemd instance.
16335 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
16336 CAP_AUDIT_CONTROL capabilities to the capabilities set for
16337 the container. This makes it easier to boot unmodified
16338 Fedora systems in a container, which however still requires
16339 audit=0 to be passed on the kernel command line. Auditing in
16340 kernel and userspace is unfortunately still too broken in
16341 context of containers, hence we recommend compiling it out
16342 of the kernel or using audit=0. Hopefully this will be fixed
16343 one day for good in the kernel.
16345 * nspawn gained the new --bind= and --bind-ro= parameters to
16346 bind mount specific directories from the host into the
16349 * nspawn will now mount its own devpts file system instance
16350 into the container, in order not to leak pty devices from
16351 the host into the container.
16353 * systemd will now read the firmware boot time performance
16354 information from the EFI variables, if the used boot loader
16355 supports this, and takes it into account for boot performance
16356 analysis via "systemd-analyze". This is currently supported
16357 only in conjunction with Gummiboot, but could be supported
16358 by other boot loaders too. For details see:
16360 https://systemd.io/BOOT_LOADER_INTERFACE
16362 * A new generator has been added that automatically mounts the
16363 EFI System Partition (ESP) to /boot, if that directory
16364 exists, is empty, and no other file system has been
16365 configured to be mounted there.
16367 * logind will now send out PrepareForSleep(false) out
16368 unconditionally, after coming back from suspend. This may be
16369 used by applications as asynchronous notification for
16370 system resume events.
16372 * "systemctl unlock-sessions" has been added, that allows
16373 unlocking the screens of all user sessions at once, similar
16374 to how "systemctl lock-sessions" already locked all users
16375 sessions. This is backed by a new D-Bus call UnlockSessions().
16377 * "loginctl seat-status" will now show the master device of a
16378 seat. (i.e. the device of a seat that needs to be around for
16379 the seat to be considered available, usually the graphics
16382 * tmpfiles gained a new "X" line type, that allows
16383 configuration of files and directories (with wildcards) that
16384 shall be excluded from automatic cleanup ("aging").
16386 * udev default rules set the device node permissions now only
16387 at "add" events, and do not change them any longer with a
16388 later "change" event.
16390 * The log messages for lid events and power/sleep keypresses
16391 now carry a message ID.
16393 * We now have a substantially larger unit test suite, but this
16394 continues to be work in progress.
16396 * udevadm hwdb gained a new --root= parameter to change the
16397 root directory to operate relative to.
16399 * logind will now issue a background sync() request to the kernel
16400 early at shutdown, so that dirty buffers are flushed to disk early
16401 instead of at the last moment, in order to optimize shutdown
16404 * A new bootctl tool has been added that is an interface for
16405 certain boot loader operations. This is currently a preview
16406 and is likely to be extended into a small mechanism daemon
16407 like timedated, localed, hostnamed, and can be used by
16408 graphical UIs to enumerate available boot options, and
16409 request boot into firmware operations.
16411 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
16412 the rest of the package. It also has been updated to work
16413 correctly in initrds.
16415 * polkit previously has been runtime optional, and is now also
16416 compile time optional via a configure switch.
16418 * systemd-analyze has been reimplemented in C. Also "systemctl
16419 dot" has moved into systemd-analyze.
16421 * "systemctl status" with no further parameters will now print
16422 the status of all active or failed units.
16424 * Operations such as "systemctl start" can now be executed
16425 with a new mode "--irreversible" which may be used to queue
16426 operations that cannot accidentally be reversed by a later
16427 job queuing. This is by default used to make shutdown
16428 requests more robust.
16430 * The Python API of systemd now gained a new module for
16431 reading journal files.
16433 * A new tool kernel-install has been added that can install
16434 kernel images according to the Boot Loader Specification:
16436 https://systemd.io/BOOT_LOADER_SPECIFICATION
16438 * Boot time console output has been improved to provide
16439 animated boot time output for hanging jobs.
16441 * A new tool systemd-activate has been added which can be used
16442 to test socket activation with, directly from the command
16443 line. This should make it much easier to test and debug
16444 socket activation in daemons.
16446 * journalctl gained a new "--reverse" (or -r) option to show
16447 journal output in reverse order (i.e. newest line first).
16449 * journalctl gained a new "--pager-end" (or -e) option to jump
16450 to immediately jump to the end of the journal in the
16451 pager. This is only supported in conjunction with "less".
16453 * journalctl gained a new "--user-unit=" option, that works
16454 similarly to "--unit=" but filters for user units rather than
16457 * A number of unit files to ease adoption of systemd in
16458 initrds has been added. This moves some minimal logic from
16459 the various initrd implementations into systemd proper.
16461 * The journal files are now owned by a new group
16462 "systemd-journal", which exists specifically to allow access
16463 to the journal, and nothing else. Previously, we used the
16464 "adm" group for that, which however possibly covers more
16465 than just journal/log file access. This new group is now
16466 already used by systemd-journal-gatewayd to ensure this
16467 daemon gets access to the journal files and as little else
16468 as possible. Note that "make install" will also set FS ACLs
16469 up for /var/log/journal to give "adm" and "wheel" read
16470 access to it, in addition to "systemd-journal" which owns
16471 the journal files. We recommend that packaging scripts also
16472 add read access to "adm" + "wheel" to /var/log/journal, and
16473 all existing/future journal files. To normal users and
16474 administrators little changes, however packagers need to
16475 ensure to create the "systemd-journal" system group at
16476 package installation time.
16478 * The systemd-journal-gatewayd now runs as unprivileged user
16479 systemd-journal-gateway:systemd-journal-gateway. Packaging
16480 scripts need to create these system user/group at
16483 * timedated now exposes a new boolean property CanNTP that
16484 indicates whether a local NTP service is available or not.
16486 * systemd-detect-virt will now also detect xen PVs
16488 * The pstore file system is now mounted by default, if it is
16491 * In addition to the SELinux and IMA policies we will now also
16492 load SMACK policies at early boot.
16494 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
16495 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
16496 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
16497 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
16498 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
16499 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
16500 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
16501 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
16502 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
16503 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
16504 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
16505 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
16506 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
16507 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
16511 * Timer units now support calendar time events in addition to
16512 monotonic time events. That means you can now trigger a unit
16513 based on a calendar time specification such as "Thu,Fri
16514 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
16515 or fifth day of any month of the year 2013, given that it is
16516 a Thursday or a Friday. This brings timer event support
16517 considerably closer to cron's capabilities. For details on
16518 the supported calendar time specification language see
16521 * udev now supports a number of different naming policies for
16522 network interfaces for predictable names, and a combination
16523 of these policies is now the default. Please see this wiki
16524 document for details:
16526 https://www.freedesktop.org/software/systemd/man/systemd.net-naming-scheme.html
16528 * Auke Kok's bootchart implementation has been added to the
16529 systemd tree. It is an optional component that can graph the
16530 boot in quite some detail. It is one of the best bootchart
16531 implementations around and minimal in its code and
16534 * nss-myhostname has been integrated into the systemd source
16535 tree. nss-myhostname guarantees that the local hostname
16536 always stays resolvable via NSS. It has been a weak
16537 requirement of systemd-hostnamed since a long time, and
16538 since its code is actually trivial we decided to just
16539 include it in systemd's source tree. It can be turned off
16540 with a configure switch.
16542 * The read-ahead logic is now capable of properly detecting
16543 whether a btrfs file system is on SSD or rotating media, in
16544 order to optimize the read-ahead scheme. Previously, it was
16545 only capable of detecting this on traditional file systems
16548 * In udev, additional device properties are now read from the
16549 IAB in addition to the OUI database. Also, Bluetooth company
16550 identities are attached to the devices as well.
16552 * In service files %U may be used as specifier that is
16553 replaced by the configured user name of the service.
16555 * nspawn may now be invoked without a controlling TTY. This
16556 makes it suitable for invocation as its own service. This
16557 may be used to set up a simple containerized server system
16558 using only core OS tools.
16560 * systemd and nspawn can now accept socket file descriptors
16561 when they are started for socket activation. This enables
16562 implementation of socket activated nspawn
16563 containers. i.e. think about autospawning an entire OS image
16564 when the first SSH or HTTP connection is received. We expect
16565 that similar functionality will also be added to libvirt-lxc
16568 * journalctl will now suppress ANSI color codes when
16569 presenting log data.
16571 * systemctl will no longer show control group information for
16572 a unit if the control group is empty anyway.
16574 * logind can now automatically suspend/hibernate/shutdown the
16577 * /etc/machine-info and hostnamed now also expose the chassis
16578 type of the system. This can be used to determine whether
16579 the local system is a laptop, desktop, handset or
16580 tablet. This information may either be configured by the
16581 user/vendor or is automatically determined from ACPI and DMI
16582 information if possible.
16584 * A number of polkit actions are now bound together with "imply"
16585 rules. This should simplify creating UIs because many actions
16586 will now authenticate similar ones as well.
16588 * Unit files learnt a new condition ConditionACPower= which
16589 may be used to conditionalize a unit depending on whether an
16590 AC power source is connected or not, of whether the system
16591 is running on battery power.
16593 * systemctl gained a new "is-failed" verb that may be used in
16594 shell scripts and suchlike to check whether a specific unit
16595 is in the "failed" state.
16597 * The EnvironmentFile= setting in unit files now supports file
16598 globbing, and can hence be used to easily read a number of
16599 environment files at once.
16601 * systemd will no longer detect and recognize specific
16602 distributions. All distribution-specific #ifdeffery has been
16603 removed, systemd is now fully generic and
16604 distribution-agnostic. Effectively, not too much is lost as
16605 a lot of the code is still accessible via explicit configure
16606 switches. However, support for some distribution specific
16607 legacy configuration file formats has been dropped. We
16608 recommend distributions to simply adopt the configuration
16609 files everybody else uses now and convert the old
16610 configuration from packaging scripts. Most distributions
16611 already did that. If that's not possible or desirable,
16612 distributions are welcome to forward port the specific
16613 pieces of code locally from the git history.
16615 * When logging a message about a unit systemd will now always
16616 log the unit name in the message meta data.
16618 * localectl will now also discover system locale data that is
16619 not stored in locale archives, but directly unpacked.
16621 * logind will no longer unconditionally use framebuffer
16622 devices as seat masters, i.e. as devices that are required
16623 to be existing before a seat is considered preset. Instead,
16624 it will now look for all devices that are tagged as
16625 "seat-master" in udev. By default, framebuffer devices will
16626 be marked as such, but depending on local systems, other
16627 devices might be marked as well. This may be used to
16628 integrate graphics cards using closed source drivers (such
16629 as NVidia ones) more nicely into logind. Note however, that
16630 we recommend using the open source NVidia drivers instead,
16631 and no udev rules for the closed-source drivers will be
16632 shipped from us upstream.
16634 Contributions from: Adam Williamson, Alessandro Crismani, Auke
16635 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
16636 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
16637 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
16638 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
16639 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
16640 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
16641 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
16642 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
16643 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
16644 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
16645 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
16650 * udev gained support for loading additional device properties
16651 from an indexed database that is keyed by vendor/product IDs
16652 and similar device identifiers. For the beginning this
16653 "hwdb" is populated with data from the well-known PCI and
16654 USB database, but also includes PNP, ACPI and OID data. In
16655 the longer run this indexed database shall grow into
16656 becoming the one central database for non-essential
16657 userspace device metadata. Previously, data from the PCI/USB
16658 database was only attached to select devices, since the
16659 lookup was a relatively expensive operation due to O(n) time
16660 complexity (with n being the number of entries in the
16661 database). Since this is now O(1), we decided to add in this
16662 data for all devices where this is available, by
16663 default. Note that the indexed database needs to be rebuilt
16664 when new data files are installed. To achieve this you need
16665 to update your packaging scripts to invoke "udevadm hwdb
16666 --update" after installation of hwdb data files. For
16667 RPM-based distributions we introduced the new
16668 %udev_hwdb_update macro for this purpose.
16670 * The Journal gained support for the "Message Catalog", an
16671 indexed database to link up additional information with
16672 journal entries. For further details please check:
16674 https://www.freedesktop.org/wiki/Software/systemd/catalog
16676 The indexed message catalog database also needs to be
16677 rebuilt after installation of message catalog files. Use
16678 "journalctl --update-catalog" for this. For RPM-based
16679 distributions we introduced the %journal_catalog_update
16680 macro for this purpose.
16682 * The Python Journal bindings gained support for the standard
16683 Python logging framework.
16685 * The Journal API gained new functions for checking whether
16686 the underlying file system of a journal file is capable of
16687 properly reporting file change notifications, or whether
16688 applications that want to reflect journal changes "live"
16689 need to recheck journal files continuously in appropriate
16692 * It is now possible to set the "age" field for tmpfiles
16693 entries to 0, indicating that files matching this entry
16694 shall always be removed when the directories are cleaned up.
16696 * coredumpctl gained a new "gdb" verb which invokes gdb
16697 right-away on the selected coredump.
16699 * There's now support for "hybrid sleep" on kernels that
16700 support this, in addition to "suspend" and "hibernate". Use
16701 "systemctl hybrid-sleep" to make use of this.
16703 * logind's HandleSuspendKey= setting (and related settings)
16704 now gained support for a new "lock" setting to simply
16705 request the screen lock on all local sessions, instead of
16706 actually executing a suspend or hibernation.
16708 * systemd will now mount the EFI variables file system by
16711 * Socket units now gained support for configuration of the
16712 SMACK security label.
16714 * timedatectl will now output the time of the last and next
16715 daylight saving change.
16717 * We dropped support for various legacy and distro-specific
16718 concepts, such as insserv, early-boot SysV services
16719 (i.e. those for non-standard runlevels such as 'b' or 'S')
16720 or ArchLinux /etc/rc.conf support. We recommend the
16721 distributions who still need support this to either continue
16722 to maintain the necessary patches downstream, or find a
16723 different solution. (Talk to us if you have questions!)
16725 * Various systemd components will now bypass polkit checks for
16726 root and otherwise handle properly if polkit is not found to
16727 be around. This should fix most issues for polkit-less
16728 systems. Quite frankly this should have been this way since
16729 day one. It is absolutely our intention to make systemd work
16730 fine on polkit-less systems, and we consider it a bug if
16731 something does not work as it should if polkit is not around.
16733 * For embedded systems it is now possible to build udev and
16734 systemd without blkid and/or kmod support.
16736 * "systemctl switch-root" is now capable of switching root
16737 more than once. I.e. in addition to transitions from the
16738 initrd to the host OS it is now possible to transition to
16739 further OS images from the host. This is useful to implement
16740 offline updating tools.
16742 * Various other additions have been made to the RPM macros
16743 shipped with systemd. Use %udev_rules_update() after
16744 installing new udev rules files. %_udevhwdbdir,
16745 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
16746 %_sysctldir are now available which resolve to the right
16747 directories for packages to place various data files in.
16749 * journalctl gained the new --full switch (in addition to
16750 --all, to disable ellipsation for long messages.
16752 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
16753 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
16754 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
16755 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
16756 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
16757 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
16758 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
16759 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
16760 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
16764 * journalctl gained new --since= and --until= switches to
16765 filter by time. It also now supports nice filtering for
16766 units via --unit=/-u.
16768 * Type=oneshot services may use ExecReload= and do the
16771 * The journal daemon now supports time-based rotation and
16772 vacuuming, in addition to the usual disk-space based
16775 * The journal will now index the available field values for
16776 each field name. This enables clients to show pretty drop
16777 downs of available match values when filtering. The bash
16778 completion of journalctl has been updated
16779 accordingly. journalctl gained a new switch -F to list all
16780 values a certain field takes in the journal database.
16782 * More service events are now written as structured messages
16783 to the journal, and made recognizable via message IDs.
16785 * The timedated, localed and hostnamed mini-services which
16786 previously only provided support for changing time, locale
16787 and hostname settings from graphical DEs such as GNOME now
16788 also have a minimal (but very useful) text-based client
16789 utility each. This is probably the nicest way to changing
16790 these settings from the command line now, especially since
16791 it lists available options and is fully integrated with bash
16794 * There's now a new tool "systemd-coredumpctl" to list and
16795 extract coredumps from the journal.
16797 * We now install a README each in /var/log/ and
16798 /etc/rc.d/init.d explaining where the system logs and init
16799 scripts went. This hopefully should help folks who go to
16800 that dirs and look into the otherwise now empty void and
16801 scratch their heads.
16803 * When user-services are invoked (by systemd --user) the
16804 $MANAGERPID env var is set to the PID of systemd.
16806 * SIGRTMIN+24 when sent to a --user instance will now result
16807 in immediate termination of systemd.
16809 * gatewayd received numerous feature additions such as a
16810 "follow" mode, for live syncing and filtering.
16812 * browse.html now allows filtering and showing detailed
16813 information on specific entries. Keyboard navigation and
16814 mouse screen support has been added.
16816 * gatewayd/journalctl now supports HTML5/JSON
16817 Server-Sent-Events as output.
16819 * The SysV init script compatibility logic will now
16820 heuristically determine whether a script supports the
16821 "reload" verb, and only then make this available as
16822 "systemctl reload".
16824 * "systemctl status --follow" has been removed, use "journalctl
16827 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
16828 have been removed since they are hardly useful to be
16831 * And I'd like to take the opportunity to specifically mention
16832 Zbigniew for his great contributions. Zbigniew, you rock!
16834 Contributions from: Andrew Eikum, Christian Hesse, Colin
16835 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
16836 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
16837 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
16838 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
16839 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
16840 Jędrzejewski-Szmek, Сковорода Никита Андреевич
16844 * If /etc/vconsole.conf is non-existent or empty we will no
16845 longer load any console font or key map at boot by
16846 default. Instead the kernel defaults will be left
16847 intact. This is definitely the right thing to do, as no
16848 configuration should mean no configuration, and hard-coding
16849 font names that are different on all archs is probably a bad
16850 idea. Also, the kernel default key map and font should be
16851 good enough for most cases anyway, and mostly identical to
16852 the userspace fonts/key maps we previously overloaded them
16853 with. If distributions want to continue to default to a
16854 non-kernel font or key map they should ship a default
16855 /etc/vconsole.conf with the appropriate contents.
16857 Contributions from: Colin Walters, Daniel J Walsh, Dave
16858 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
16859 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
16863 * journalctl gained a new --cursor= switch to show entries
16864 starting from the specified location in the journal.
16866 * We now enforce a size limit on journal entry fields exported
16867 with "-o json" in journalctl. Fields larger than 4K will be
16868 assigned null. This can be turned off with --all.
16870 * An (optional) journal gateway daemon is now available as
16871 "systemd-journal-gatewayd.service". This service provides
16872 access to the journal via HTTP and JSON. This functionality
16873 will be used to implement live log synchronization in both
16874 pull and push modes, but has various other users too, such
16875 as easy log access for debugging of embedded devices. Right
16876 now it is already useful to retrieve the journal via HTTP:
16878 # systemctl start systemd-journal-gatewayd.service
16879 # wget http://localhost:19531/entries
16881 This will download the journal contents in a
16882 /var/log/messages compatible format. The same as JSON:
16884 # curl -H"Accept: application/json" http://localhost:19531/entries
16886 This service is also accessible via a web browser where a
16887 single static HTML5 app is served that uses the JSON logic
16888 to enable the user to do some basic browsing of the
16889 journal. This will be extended later on. Here's an example
16890 screenshot of this app in its current state:
16892 https://0pointer.de/public/journal-gatewayd
16894 Contributions from: Kay Sievers, Lennart Poettering, Robert
16895 Milasan, Tom Gundersen
16899 * The bash completion logic is now available for journalctl
16902 * We do not mount the "cpuset" controller anymore together with
16903 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
16904 started if no parameters are assigned to it. "cpuset" hence
16905 broke code that assumed it could create "cpu" groups and
16908 * journalctl -f will now subscribe to terminal size changes,
16909 and line break accordingly.
16911 Contributions from: Dave Reisner, Kay Sievers, Lennart
16912 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
16916 * nspawn will now create a symlink /etc/localtime in the
16917 container environment, copying the host's timezone
16918 setting. Previously this has been done via a bind mount, but
16919 since symlinks cannot be bind mounted this has now been
16920 changed to create/update the appropriate symlink.
16922 * journalctl -n's line number argument is now optional, and
16923 will default to 10 if omitted.
16925 * journald will now log the maximum size the journal files may
16926 take up on disk. This is particularly useful if the default
16927 built-in logic of determining this parameter from the file
16928 system size is used. Use "systemctl status
16929 systemd-journald.service" to see this information.
16931 * The multi-seat X wrapper tool has been stripped down. As X
16932 is now capable of enumerating graphics devices via udev in a
16933 seat-aware way the wrapper is not strictly necessary
16934 anymore. A stripped down temporary stop-gap is still shipped
16935 until the upstream display managers have been updated to
16936 fully support the new X logic. Expect this wrapper to be
16937 removed entirely in one of the next releases.
16939 * HandleSleepKey= in logind.conf has been split up into
16940 HandleSuspendKey= and HandleHibernateKey=. The old setting
16941 is not available anymore. X11 and the kernel are
16942 distinguishing between these keys and we should too. This
16943 also means the inhibition lock for these keys has been split
16946 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
16947 Poettering, Lukas Nykryn, Václav Pavlín
16951 * Whenever a unit changes state we will now log this to the
16952 journal and show along the unit's own log output in
16953 "systemctl status".
16955 * ConditionPathIsMountPoint= can now properly detect bind
16956 mount points too. (Previously, a bind mount of one file
16957 system to another place in the same file system could not be
16958 detected as mount, since they shared struct stat's st_dev
16961 * We will now mount the cgroup controllers cpu, cpuacct,
16962 cpuset and the controllers net_cls, net_prio together by
16965 * nspawn containers will now have a virtualized boot
16966 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
16967 over with a randomized ID at container initialization). This
16968 has the effect of making "journalctl -b" do the right thing
16971 * The JSON output journal serialization has been updated not
16972 to generate "endless" list objects anymore, but rather one
16973 JSON object per line. This is more in line how most JSON
16974 parsers expect JSON objects. The new output mode
16975 "json-pretty" has been added to provide similar output, but
16976 neatly aligned for readability by humans.
16978 * We dropped all explicit sync() invocations in the shutdown
16979 code. The kernel does this implicitly anyway in the kernel
16980 reboot() syscall. halt(8)'s -n option is now a compatibility
16983 * We now support virtualized reboot() in containers, as
16984 supported by newer kernels. We will fall back to exit() if
16985 CAP_SYS_REBOOT is not available to the container. Also,
16986 nspawn makes use of this now and will actually reboot the
16987 container if the containerized OS asks for that.
16989 * journalctl will only show local log output by default
16990 now. Use --merge (-m) to show remote log output, too.
16992 * libsystemd-journal gained the new sd_journal_get_usage()
16993 call to determine the current disk usage of all journal
16994 files. This is exposed in the new "journalctl --disk-usage"
16997 * journald gained a new configuration setting SplitMode= in
16998 journald.conf which may be used to control how user journals
16999 are split off. See journald.conf(5) for details.
17001 * A new condition type ConditionFileNotEmpty= has been added.
17003 * tmpfiles' "w" lines now support file globbing, to write
17004 multiple files at once.
17006 * We added Python bindings for the journal submission
17007 APIs. More Python APIs for a number of selected APIs will
17008 likely follow. Note that we intend to add native bindings
17009 only for the Python language, as we consider it common
17010 enough to deserve bindings shipped within systemd. There are
17011 various projects outside of systemd that provide bindings
17012 for languages such as PHP or Lua.
17014 * Many conditions will now resolve specifiers such as %i. In
17015 addition, PathChanged= and related directives of .path units
17016 now support specifiers as well.
17018 * There's now a new RPM macro definition for the system preset
17021 * journald will now warn if it ca not forward a message to the
17022 syslog daemon because its socket is full.
17024 * timedated will no longer write or process /etc/timezone,
17025 except on Debian. As we do not support late mounted /usr
17026 anymore /etc/localtime always being a symlink is now safe,
17027 and hence the information in /etc/timezone is not necessary
17030 * logind will now always reserve one VT for a text getty (VT6
17031 by default). Previously if more than 6 X sessions where
17032 started they took up all the VTs with auto-spawned gettys,
17033 so that no text gettys were available anymore.
17035 * udev will now automatically inform the btrfs kernel logic
17036 about btrfs RAID components showing up. This should make
17037 simple hotplug based btrfs RAID assembly work.
17039 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
17040 (but not for its children which will stay at the kernel
17041 default). This should allow setups with a lot more listening
17044 * systemd will now always pass the configured timezone to the
17045 kernel at boot. timedated will do the same when the timezone
17048 * logind's inhibition logic has been updated. By default,
17049 logind will now handle the lid switch, the power and sleep
17050 keys all the time, even in graphical sessions. If DEs want
17051 to handle these events on their own they should take the new
17052 handle-power-key, handle-sleep-key and handle-lid-switch
17053 inhibitors during their runtime. A simple way to achieve
17054 that is to invoke the DE wrapped in an invocation of:
17056 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch …
17058 * Access to unit operations is now checked via SELinux taking
17059 the unit file label and client process label into account.
17061 * systemd will now notify the administrator in the journal
17062 when he over-mounts a non-empty directory.
17064 * There are new specifiers that are resolved in unit files,
17065 for the hostname (%H), the machine ID (%m) and the boot ID
17068 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
17069 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
17070 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
17071 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
17072 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
17073 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
17074 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
17078 * Support for reading structured kernel messages from
17079 /dev/kmsg has now been added and is enabled by default.
17081 * Support for reading kernel messages from /proc/kmsg has now
17082 been removed. If you want kernel messages in the journal
17083 make sure to run a recent kernel (>= 3.5) that supports
17084 reading structured messages from /dev/kmsg (see
17085 above). /proc/kmsg is now exclusive property of classic
17086 syslog daemons again.
17088 * The libudev API gained the new
17089 udev_device_new_from_device_id() call.
17091 * The logic for file system namespace (ReadOnlyDirectory=,
17092 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
17093 require pivot_root() anymore. This means fewer temporary
17094 directories are created below /tmp for this feature.
17096 * nspawn containers will now see and receive all submounts
17097 made on the host OS below the root file system of the
17100 * Forward Secure Sealing is now supported for Journal files,
17101 which provide cryptographical sealing of journal files so
17102 that attackers cannot alter log history anymore without this
17103 being detectable. Lennart will soon post a blog story about
17104 this explaining it in more detail.
17106 * There are two new service settings RestartPreventExitStatus=
17107 and SuccessExitStatus= which allow configuration of exit
17108 status (exit code or signal) which will be excepted from the
17109 restart logic, resp. consider successful.
17111 * journalctl gained the new --verify switch that can be used
17112 to check the integrity of the structure of journal files and
17113 (if Forward Secure Sealing is enabled) the contents of
17116 * nspawn containers will now be run with /dev/stdin, /dev/fd/
17117 and similar symlinks pre-created. This makes running shells
17118 as container init process a lot more fun.
17120 * The fstab support can now handle PARTUUID= and PARTLABEL=
17123 * A new ConditionHost= condition has been added to match
17124 against the hostname (with globs) and machine ID. This is
17125 useful for clusters where a single OS image is used to
17126 provision a large number of hosts which shall run slightly
17127 different sets of services.
17129 * Services which hit the restart limit will now be placed in a
17132 Contributions from: Bertram Poettering, Dave Reisner, Huang
17133 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
17134 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
17138 * When running in --user mode systemd will now become a
17139 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
17140 tree a lot more organized.
17142 * A new PartOf= unit dependency type has been introduced that
17143 may be used to group services in a natural way.
17145 * "systemctl enable" may now be used to enable instances of
17148 * journalctl now prints error log levels in red, and
17149 warning/notice log levels in bright white. It also supports
17150 filtering by log level now.
17152 * cgtop gained a new -n switch (similar to top), to configure
17153 the maximum number of iterations to run for. It also gained
17154 -b, to run in batch mode (accepting no input).
17156 * The suffix ".service" may now be omitted on most systemctl
17157 command lines involving service unit names.
17159 * There's a new bus call in logind to lock all sessions, as
17160 well as a loginctl verb for it "lock-sessions".
17162 * libsystemd-logind.so gained a new call sd_journal_perror()
17163 that works similar to libc perror() but logs to the journal
17164 and encodes structured information about the error number.
17166 * /etc/crypttab entries now understand the new keyfile-size=
17169 * shutdown(8) now can send a (configurable) wall message when
17170 a shutdown is cancelled.
17172 * The mount propagation mode for the root file system will now
17173 default to "shared", which is useful to make containers work
17174 nicely out-of-the-box so that they receive new mounts from
17175 the host. This can be undone locally by running "mount
17176 --make-rprivate /" if needed.
17178 * The prefdm.service file has been removed. Distributions
17179 should maintain this unit downstream if they intend to keep
17180 it around. However, we recommend writing normal unit files
17181 for display managers instead.
17183 * Since systemd is a crucial part of the OS we will now
17184 default to a number of compiler switches that improve
17185 security (hardening) such as read-only relocations, stack
17186 protection, and suchlike.
17188 * The TimeoutSec= setting for services is now split into
17189 TimeoutStartSec= and TimeoutStopSec= to allow configuration
17190 of individual time outs for the start and the stop phase of
17193 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
17194 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
17195 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
17196 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
17197 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
17198 Gundersen, Zbigniew Jędrzejewski-Szmek
17202 * The journal and id128 C APIs are now fully documented as man
17205 * Extra safety checks have been added when transitioning from
17206 the initial RAM disk to the main system to avoid accidental
17209 * /etc/crypttab entries now understand the new keyfile-offset=
17212 * systemctl -t can now be used to filter by unit load state.
17214 * The journal C API gained the new sd_journal_wait() call to
17215 make writing synchronous journal clients easier.
17217 * journalctl gained the new -D switch to show journals from a
17218 specific directory.
17220 * journalctl now displays a special marker between log
17221 messages of two different boots.
17223 * The journal is now explicitly flushed to /var via a service
17224 systemd-journal-flush.service, rather than implicitly simply
17225 by seeing /var/log/journal to be writable.
17227 * journalctl (and the journal C APIs) can now match for much
17228 more complex expressions, with alternatives and
17231 * When transitioning from the initial RAM disk to the main
17232 system we will now kill all processes in a killing spree to
17233 ensure no processes stay around by accident.
17235 * Three new specifiers may be used in unit files: %u, %h, %s
17236 resolve to the user name, user home directory resp. user
17237 shell. This is useful for running systemd user instances.
17239 * We now automatically rotate journal files if their data
17240 object hash table gets a fill level > 75%. We also size the
17241 hash table based on the configured maximum file size. This
17242 together should lower hash collisions drastically and thus
17243 speed things up a bit.
17245 * journalctl gained the new "--header" switch to introspect
17246 header data of journal files.
17248 * A new setting SystemCallFilters= has been added to services which may
17249 be used to apply deny lists or allow lists to system calls. This is
17250 based on SECCOMP Mode 2 of Linux 3.5.
17252 * nspawn gained a new --link-journal= switch (and quicker: -j)
17253 to link the container journal with the host. This makes it
17254 very easy to centralize log viewing on the host for all
17255 guests while still keeping the journal files separated.
17257 * Many bugfixes and optimizations
17259 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
17260 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
17261 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
17266 * Several tools now understand kernel command line arguments,
17267 which are only read when run in an initial RAM disk. They
17268 usually follow closely their normal counterparts, but are
17271 * There's a new tool to analyze the readahead files that are
17272 automatically generated at boot. Use:
17274 /usr/lib/systemd/systemd-readahead analyze /.readahead
17276 * We now provide an early debug shell on tty9 if this enabled. Use:
17278 systemctl enable debug-shell.service
17280 * All plymouth related units have been moved into the Plymouth
17281 package. Please make sure to upgrade your Plymouth version
17284 * systemd-tmpfiles now supports getting passed the basename of
17285 a configuration file only, in which case it will look for it
17286 in all appropriate directories automatically.
17288 * udevadm info now takes a /dev or /sys path as argument, and
17289 does the right thing. Example:
17291 udevadm info /dev/sda
17292 udevadm info /sys/class/block/sda
17294 * systemctl now prints a warning if a unit is stopped but a
17295 unit that might trigger it continues to run. Example: a
17296 service is stopped but the socket that activates it is left
17299 * "systemctl status" will now mention if the log output was
17300 shortened due to rotation since a service has been started.
17302 * The journal API now exposes functions to determine the
17303 "cutoff" times due to rotation.
17305 * journald now understands SIGUSR1 and SIGUSR2 for triggering
17306 immediately flushing of runtime logs to /var if possible,
17307 resp. for triggering immediate rotation of the journal
17310 * It is now considered an error if a service is attempted to
17311 be stopped that is not loaded.
17313 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
17315 * systemd-analyze now supports Python 3
17317 * tmpfiles now supports cleaning up directories via aging
17318 where the first level dirs are always kept around but
17319 directories beneath it automatically aged. This is enabled
17320 by prefixing the age field with '~'.
17322 * Seat objects now expose CanGraphical, CanTTY properties
17323 which is required to deal with very fast bootups where the
17324 display manager might be running before the graphics drivers
17325 completed initialization.
17327 * Seat objects now expose a State property.
17329 * We now include RPM macros for service enabling/disabling
17330 based on the preset logic. We recommend RPM based
17331 distributions to make use of these macros if possible. This
17332 makes it simpler to reuse RPM spec files across
17335 * We now make sure that the collected systemd unit name is
17336 always valid when services log to the journal via
17339 * There's a new man page kernel-command-line(7) detailing all
17340 command line options we understand.
17342 * The fstab generator may now be disabled at boot by passing
17343 fstab=0 on the kernel command line.
17345 * A new kernel command line option modules-load= is now understood
17346 to load a specific kernel module statically, early at boot.
17348 * Unit names specified on the systemctl command line are now
17349 automatically escaped as needed. Also, if file system or
17350 device paths are specified they are automatically turned
17351 into the appropriate mount or device unit names. Example:
17353 systemctl status /home
17354 systemctl status /dev/sda
17356 * The SysVConsole= configuration option has been removed from
17357 system.conf parsing.
17359 * The SysV search path is no longer exported on the D-Bus
17362 * The Names= option has been removed from unit file parsing.
17364 * There's a new man page bootup(7) detailing the boot process.
17366 * Every unit and every generator we ship with systemd now
17367 comes with full documentation. The self-explanatory boot is
17370 * A couple of services gained "systemd-" prefixes in their
17371 name if they wrap systemd code, rather than only external
17372 code. Among them fsck@.service which is now
17373 systemd-fsck@.service.
17375 * The HaveWatchdog property has been removed from the D-Bus
17378 * systemd.confirm_spawn= on the kernel command line should now
17381 * There's a new man page crypttab(5) which details all options
17382 we actually understand.
17384 * systemd-nspawn gained a new --capability= switch to pass
17385 additional capabilities to the container.
17387 * timedated will now read known NTP implementation unit names
17388 from /usr/lib/systemd/ntp-units.d/*.list,
17389 systemd-timedated-ntp.target has been removed.
17391 * journalctl gained a new switch "-b" that lists log data of
17392 the current boot only.
17394 * The notify socket is in the abstract namespace again, in
17395 order to support daemons which chroot() at start-up.
17397 * There is a new Storage= configuration option for journald
17398 which allows configuration of where log data should go. This
17399 also provides a way to disable journal logging entirely, so
17400 that data collected is only forwarded to the console, the
17401 kernel log buffer or another syslog implementation.
17403 * Many bugfixes and optimizations
17405 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
17406 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
17407 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
17408 Shawn Landden, Tom Gundersen
17412 * "systemctl help <unit>" now shows the man page if one is
17415 * Several new man pages have been added.
17417 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
17418 MaxLevelConsole= can now be specified in
17419 journald.conf. These options allow reducing the amount of
17420 data stored on disk or forwarded by the log level.
17422 * TimerSlackNSec= can now be specified in system.conf for
17423 PID1. This allows system-wide power savings.
17425 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
17426 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
17431 * logind is now capable of (optionally) handling power and
17432 sleep keys as well as the lid switch.
17434 * journalctl now understands the syntax "journalctl
17435 /usr/bin/avahi-daemon" to get all log output of a specific
17438 * CapabilityBoundingSet= in system.conf now also influences
17439 the capability bound set of usermode helpers of the kernel.
17441 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
17442 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
17443 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
17444 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
17448 * Note that we skipped 139 releases here in order to set the
17449 new version to something that is greater than both udev's
17450 and systemd's most recent version number.
17452 * udev: all udev sources are merged into the systemd source tree now.
17453 All future udev development will happen in the systemd tree. It
17454 is still fully supported to use the udev daemon and tools without
17455 systemd running, like in initramfs or other init systems. Building
17456 udev though, will require the *build* of the systemd tree, but
17457 udev can be properly *run* without systemd.
17459 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
17460 should be used to create dead device nodes as workarounds for broken
17463 * udev: RUN+="socket:…" and udev_monitor_new_from_socket() is
17464 no longer supported. udev_monitor_new_from_netlink() needs to be
17465 used to subscribe to events.
17467 * udev: when udevd is started by systemd, processes which are left
17468 behind by forking them off of udev rules, are unconditionally cleaned
17469 up and killed now after the event handling has finished. Services or
17470 daemons must be started as systemd services. Services can be
17471 pulled-in by udev to get started, but they can no longer be directly
17472 forked by udev rules.
17474 * udev: the daemon binary is called systemd-udevd now and installed
17475 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
17476 to adapt to that, create symlink, or rename the binary after building
17479 * libudev no longer provides these symbols:
17480 udev_monitor_from_socket()
17481 udev_queue_get_failed_list_entry()
17482 udev_get_{dev,sys,run}_path()
17483 The versions number was bumped and symbol versioning introduced.
17485 * systemd-loginctl and systemd-journalctl have been renamed
17486 to loginctl and journalctl to match systemctl.
17488 * The config files: /etc/systemd/systemd-logind.conf and
17489 /etc/systemd/systemd-journald.conf have been renamed to
17490 logind.conf and journald.conf. Package updates should rename
17491 the files to the new names on upgrade.
17493 * For almost all files the license is now LGPL2.1+, changed
17494 from the previous GPL2.0+. Exceptions are some minor stuff
17495 of udev (which will be changed to LGPL2.1 eventually, too),
17496 and the MIT licensed sd-daemon.[ch] library that is suitable
17497 to be used as drop-in files.
17499 * systemd and logind now handle system sleep states, in
17500 particular suspending and hibernating.
17502 * logind now implements a sleep/shutdown/idle inhibiting logic
17503 suitable for a variety of uses. Soonishly Lennart will blog
17504 about this in more detail.
17506 * var-run.mount and var-lock.mount are no longer provided
17507 (which previously bind mounted these directories to their new
17508 places). Distributions which have not converted these
17509 directories to symlinks should consider stealing these files
17510 from git history and add them downstream.
17512 * We introduced the Documentation= field for units and added
17513 this to all our shipped units. This is useful to make it
17514 easier to explore the boot and the purpose of the various
17517 * All smaller setup units (such as
17518 systemd-vconsole-setup.service) now detect properly if they
17519 are run in a container and are skipped when
17520 appropriate. This guarantees an entirely noise-free boot in
17521 Linux container environments such as systemd-nspawn.
17523 * A framework for implementing offline system updates is now
17524 integrated, for details see:
17525 https://www.freedesktop.org/software/systemd/man/systemd.offline-updates.html
17527 * A new service type Type=idle is available now which helps us
17528 avoiding ugly interleaving of getty output and boot status
17531 * There's now a system-wide CapabilityBoundingSet= option to
17532 globally reduce the set of capabilities for the
17533 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
17534 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
17535 even CAP_NET_ADMIN system-wide for secure systems.
17537 * There are now system-wide DefaultLimitXXX= options to
17538 globally change the defaults of the various resource limits
17539 for all units started by PID 1.
17541 * Harald Hoyer's systemd test suite has been integrated into
17542 systemd which allows easy testing of systemd builds in qemu
17543 and nspawn. (This is really awesome! Ask us for details!)
17545 * The fstab parser is now implemented as generator, not inside
17548 * systemctl will now warn you if .mount units generated from
17549 /etc/fstab are out of date due to changes in fstab that
17550 have not been read by systemd yet.
17552 * systemd is now suitable for usage in initrds. Dracut has
17553 already been updated to make use of this. With this in place
17554 initrds get a slight bit faster but primarily are much
17555 easier to introspect and debug since "systemctl status" in
17556 the host system can be used to introspect initrd services,
17557 and the journal from the initrd is kept around too.
17559 * systemd-delta has been added, a tool to explore differences
17560 between user/admin configuration and vendor defaults.
17562 * PrivateTmp= now affects both /tmp and /var/tmp.
17564 * Boot time status messages are now much prettier and feature
17565 proper english language. Booting up systemd has never been
17568 * Read-ahead pack files now include the inode number of all
17569 files to pre-cache. When the inode changes the pre-caching
17570 is not attempted. This should be nicer to deal with updated
17571 packages which might result in changes of read-ahead
17574 * We now temporaritly lower the kernel's read_ahead_kb variable
17575 when collecting read-ahead data to ensure the kernel's
17576 built-in read-ahead does not add noise to our measurements
17577 of necessary blocks to pre-cache.
17579 * There's now RequiresMountsFor= to add automatic dependencies
17580 for all mounts necessary for a specific file system path.
17582 * MountAuto= and SwapAuto= have been removed from
17583 system.conf. Mounting file systems at boot has to take place
17586 * nspawn now learned a new switch --uuid= to set the machine
17587 ID on the command line.
17589 * nspawn now learned the -b switch to automatically search
17590 for an init system.
17592 * vt102 is now the default TERM for serial TTYs, upgraded from
17595 * systemd-logind now works on VT-less systems.
17597 * The build tree has been reorganized. The individual
17598 components now have directories of their own.
17600 * A new condition type ConditionPathIsReadWrite= is now available.
17602 * nspawn learned the new -C switch to create cgroups for the
17603 container in other hierarchies.
17605 * We now have support for hardware watchdogs, configurable in
17608 * The scheduled shutdown logic now has a public API.
17610 * We now mount /tmp as tmpfs by default, but this can be
17611 masked and /etc/fstab can override it.
17613 * Since udisks does not make use of /media anymore we are not
17614 mounting a tmpfs on it anymore.
17616 * journalctl gained a new --local switch to only interleave
17617 locally generated journal files.
17619 * We can now load the IMA policy at boot automatically.
17621 * The GTK tools have been split off into a systemd-ui.
17623 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
17624 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
17625 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
17626 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
17627 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
17628 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
17629 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
17630 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
17631 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
17636 * This is mostly a bugfix release
17638 * Support optional initialization of the machine ID from the
17639 KVM or container configured UUID.
17641 * Support immediate reboots with "systemctl reboot -ff"
17643 * Show /etc/os-release data in systemd-analyze output
17645 * Many bugfixes for the journal, including endianness fixes and
17646 ensuring that disk space enforcement works
17648 * sd-login.h is C++ compatible again
17650 * Extend the /etc/os-release format on request of the Debian
17653 * We now refuse non-UTF8 strings used in various configuration
17654 and unit files. This is done to ensure we do not pass invalid
17655 data over D-Bus or expose it elsewhere.
17657 * Register Mimo USB Screens as suitable for automatic seat
17660 * Read SELinux client context from journal clients in a race
17663 * Reorder configuration file lookup order. /etc now always
17664 overrides /run in order to allow the administrator to always
17665 and unconditionally override vendor-supplied or
17666 automatically generated data.
17668 * The various user visible bits of the journal now have man
17669 pages. We still lack man pages for the journal API calls
17672 * We now ship all man pages in HTML format again in the
17675 Contributions from: Dave Reisner, Dirk Eibach, Frederic
17676 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
17677 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
17682 * This is mostly a bugfix release
17684 * systems lacking /etc/os-release are no longer supported.
17686 * Various functionality updates to libsystemd-login.so
17688 * Track class of PAM logins to distinguish greeters from
17689 normal user logins.
17691 Contributions from: Kay Sievers, Lennart Poettering, Michael
17696 * This is an important bugfix release for v41.
17698 * Building man pages is now optional which should be useful
17699 for those building systemd from git but unwilling to install
17702 * Watchdog support for supervising services is now usable. In
17703 a future release support for hardware watchdogs
17704 (i.e. /dev/watchdog) will be added building on this.
17706 * Service start rate limiting is now configurable and can be
17707 turned off per service. When a start rate limit is hit a
17708 reboot can automatically be triggered.
17710 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
17712 Contributions from: Benjamin Franzke, Bill Nottingham,
17713 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
17714 Schmidt, Michał Górny, Piotr Drąg
17718 * The systemd binary is installed /usr/lib/systemd/systemd now;
17719 An existing /sbin/init symlink needs to be adapted with the
17722 * The code that loads kernel modules has been ported to invoke
17723 libkmod directly, instead of modprobe. This means we do not
17724 support systems with module-init-tools anymore.
17726 * Watchdog support is now already useful, but still not
17729 * A new kernel command line option systemd.setenv= is
17730 understood to set system wide environment variables
17731 dynamically at boot.
17733 * We now limit the set of capabilities of systemd-journald.
17735 * We now set SIGPIPE to ignore by default, since it only is
17736 useful in shell pipelines, and has little use in general
17737 code. This can be disabled with IgnoreSIPIPE=no in unit
17740 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
17741 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
17746 * This is mostly a bugfix release
17748 * We now expose the reason why a service failed in the
17749 "Result" D-Bus property.
17751 * Rudimentary service watchdog support (will be completed over
17752 the next few releases.)
17754 * When systemd forks off in order execute some service we will
17755 now immediately changes its argv[0] to reflect which process
17756 it will execute. This is useful to minimize the time window
17757 with a generic argv[0], which makes bootcharts more useful
17759 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
17760 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
17761 Mike Kazantsev, Ray Strode
17765 * This is mostly a test release, but incorporates many
17768 * New systemd-cgtop tool to show control groups by their
17771 * Linking against libacl for ACLs is optional again. If
17772 disabled, support tracking device access for active logins
17773 goes becomes unavailable, and so does access to the user
17774 journals by the respective users.
17776 * If a group "adm" exists, journal files are automatically
17777 owned by them, thus allow members of this group full access
17778 to the system journal as well as all user journals.
17780 * The journal now stores the SELinux context of the logging
17781 client for all entries.
17783 * Add C++ inclusion guards to all public headers
17785 * New output mode "cat" in the journal to print only text
17786 messages, without any meta data like date or time.
17788 * Include tiny X server wrapper as a temporary stop-gap to
17789 teach XOrg udev display enumeration. This is used by display
17790 managers such as gdm, and will go away as soon as XOrg
17791 learned native udev hotplugging for display devices.
17793 * Add new systemd-cat tool for executing arbitrary programs
17794 with STDERR/STDOUT connected to the journal. Can also act as
17795 BSD logger replacement, and does so by default.
17797 * Optionally store all locally generated coredumps in the
17798 journal along with meta data.
17800 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
17801 writing short strings to files (for usage for /sys), and for
17802 creating symlinks, character and block device nodes.
17804 * New unit file option ControlGroupPersistent= to make cgroups
17805 persistent, following the mechanisms outlined in
17806 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
17808 * Support multiple local RTCs in a sane way
17810 * No longer monopolize IO when replaying readahead data on
17811 rotating disks, since we might starve non-file-system IO to
17812 death, since fanotify() will not see accesses done by blkid,
17815 * Do not show kernel threads in systemd-cgls anymore, unless
17816 requested with new -k switch.
17818 Contributions from: Dan Horák, Kay Sievers, Lennart
17819 Poettering, Michal Schmidt
17823 * This is mostly a test release, but incorporates many
17826 * The git repository moved to:
17827 git://anongit.freedesktop.org/systemd/systemd
17828 ssh://git.freedesktop.org/git/systemd/systemd
17830 * First release with the journal
17831 https://0pointer.de/blog/projects/the-journal.html
17833 * The journal replaces both systemd-kmsg-syslogd and
17834 systemd-stdout-bridge.
17836 * New sd_pid_get_unit() API call in libsystemd-logind
17838 * Many systemadm clean-ups
17840 * Introduce remote-fs-pre.target which is ordered before all
17841 remote mounts and may be used to start services before all
17844 * Added Mageia support
17846 * Add bash completion for systemd-loginctl
17848 * Actively monitor PID file creation for daemons which exit in
17849 the parent process before having finished writing the PID
17850 file in the daemon process. Daemons which do this need to be
17851 fixed (i.e. PID file creation must have finished before the
17852 parent exits), but we now react a bit more gracefully to them.
17854 * Add colourful boot output, mimicking the well-known output
17855 of existing distributions.
17857 * New option PassCredentials= for socket units, for
17858 compatibility with a recent kernel ABI breakage.
17860 * /etc/rc.local is now hooked in via a generator binary, and
17861 thus will no longer act as synchronization point during
17864 * systemctl list-unit-files now supports --root=.
17866 * systemd-tmpfiles now understands two new commands: z, Z for
17867 relabelling files according to the SELinux database. This is
17868 useful to apply SELinux labels to specific files in /sys,
17869 among other things.
17871 * Output of SysV services is now forwarded to both the console
17872 and the journal by default, not only just the console.
17874 * New man pages for all APIs from libsystemd-login.
17876 * The build tree got reorganized and the build system is a
17877 lot more modular allowing embedded setups to specifically
17878 select the components of systemd they are interested in.
17880 * Support for Linux systems lacking the kernel VT subsystem is
17883 * configure's --with-rootdir= got renamed to
17884 --with-rootprefix= to follow the naming used by udev and
17887 * Unless specified otherwise we will now install to /usr instead
17888 of /usr/local by default.
17890 * Processes with '@' in argv[0][0] are now excluded from the
17891 final shut-down killing spree, following the logic explained
17893 https://systemd.io/ROOT_STORAGE_DAEMONS/
17895 * All processes remaining in a service cgroup when we enter
17896 the START or START_PRE states are now killed with
17897 SIGKILL. That means it is no longer possible to spawn
17898 background processes from ExecStart= lines (which was never
17899 supported anyway, and bad style).
17901 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
17902 reloading of units together.
17904 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
17905 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
17906 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
17907 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
17908 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek