HALF ECHO LEARN / REFERENCE COLLECTION
v262.2NEWS
Git blob: b0d42b491c5b5bbc5168dc8d19d77308851d87df
Lines 4001–5000 of 23,379
* There's now a per-user counterpart of /var/lib/machines/ defined as
~/.local/state/machines/. Various tools such as systemd-nspawn +
systemd-vmspawn now will search this directory when looking for a
disk image, when invoked in unprivileged user context.
systemd-dissect's --discover command may now be combined with --user
or --system to choose in which of the directory scopes to look for
images.
* systemd-dissect gained a new --all switch. If specified the tool will
not just discover DDIs (i.e. disk images) but also images stored in
regular directories.
* systemd-dissect gained a new "--shift" switch for recursively
re-chown()ing a directory tree from one set of UID/GIDs to another.
This may be used to shift a tree from the base-0-UID range to the
foreign UID range or back.
* systemd-dissect gained new --usr-hash= and --usr-hash-sig= options,
that are similar to the existing --root-hash=/--root-hash-sig=
options, but for the /usr/ partition. This allows the root hash of
the /usr/ Verity volume and its signature to be specified.
* When dissecting/mounting a DDI disk image, and no Verity root hash or
signature is provided, suitable values are now automatically
discovered from the image itself.
* systemd-gpt-auto-generator now understands root=dissect and
mount.usr=dissect as kernel command line options that explicitly
request the full blown DDI dissector to be used to discover the root
and /usr/ file system, including automatic Verity root hash and
signature discovery, automatic handling of versioning, image policy
enforcement and filtering and so on.
* The DDI dissection logic now understands a concept of partition
"filtering". A partition filter is simply a per-designator globbing
pattern to match the partition labels against. This may be used
to support parallel installations of multiple operating systems on the
same disk, where each OS names its partitions with a specific prefix
or similar. systemd-dissect gained a new --image-filter= switch to
configure this filter. The new "dissect_image" udev plugin and
systemd-gpt-auto-generator now understand the new
systemd.image_filter= kernel command line switch configuring this
filter for the system.
systemd-importd & importctl:
* systemd-pull/importctl now supports ASCII armored (*.asc) GPG signatures.
* The systemd.pull= and rd.systemd.pull= kernel command line switches
(which may be used to automatically download a VM, container, confext,
or sysext at boot) now understand a new flag "blockdev". When
specified the downloaded image is attached to a loopback block device
after download. This may be used to boot directly into a disk image
downloaded via HTTP via a kernel command line like this:
rd.systemd.pull=raw,machine,verify=no,blockdev:image:https://192.168.100.1:8081/image.raw \
root=/dev/disk/by-loop-ref/image.raw-part2
* systemd.pull=/rd.systemd.pull= also gained support for a new flag
"bootorigin". If specified and if the system was network booted
through systemd-stub (which now sets the LoaderDeviceURL EFI
variable, see above), the URL to boot from is now automatically
formed from the UKI network boot URL with a new suffix. Example:
rd.systemd.pull=raw,machine,verify=no,blockdev,bootorigin:rootdisk:image.raw.xz \
root=/dev/disk/by-loop-ref/rootdisk.raw-part2
* The systemd.pull=/rd.systemd.pull= switches now also support a new
flag "runtime=", taking a boolean argument. If true the downloaded
image is placed below the /run/ hierarchy instead of /var/. It
defaults to true for rd.systemd.pull= (i.e. for downloads made in the
initrd), and false for systemd.pull= (i.e. for those made after the
initrd→host transition).
* New generic target units imports-pre.target and imports.target have
been introduced that are ordered before and after all downloads.
* systemd-importd gained support for downloading images compressed with
zstd now, too. (In addition to .xz, .gz and .bz2.)
Factory Reset:
* A new tool systemd-factory-reset has been added that may be used to
request or cancel a factory reset request for the next reboot. It is
also accessible via its own Varlink API.
* A new target unit factory-reset-now.target has been added that
executes an immediate factory reset. (Previously factory-reset.target
existed already that requested it for next reboot).
* A new kernel command line option systemd.factory_reset= has been
added for explicitly requesting a factory reset. (Implemented via a
new systemd-factory-reset-generator)
* A new document explaining the factory reset logic in detail has been
added. It is available online here:
https://systemd.io/FACTORY_RESET
systemd-repart:
* systemd-repart gained a new switch --join-signature= for supporting
offline Verity signing.
* systemd-repart gained a new switch --append-fstab= for controlling
how to write or append automatically generated /etc/fstab entries.
* CopyFiles= lines can now contain an "fsverity=copy" flag to preserve
the fs-verity status of the source files when populating the
filesystem.
* systemd-repart has been updated to automatically generate the
extended attributes systemd-validatefs@.service understands (see
below), for all partitions it recognizes. Controllable via the
AddValidateFS= partition setting (which defaults to true).
* repart.d/ drop-ins gained a new setting FileSystemSectorSize= which
allows configuring the sector size that file systems for newly
formatted file systems explicitly.
* systemd-repart will now enforce a minimum size for ESP/XBOOTLDR
partitions of 100M (on 512b sector drives) or 260M (on 4K sector
drives), in accordance to the requirements for these kind of
partitions.
* The Format= setting in repart.d/ files gained support for a special
value "empty". This is a shortcut to set up an empty partition and
set the partition label to "_empty", and set the "NoAuto" GPT
flag. The former is useful as systemd-sysupdate recognizes empty
partitions that way, the latter is useful to ensure that the
partition is not automatically made used of as is, on any OS that
supports GPT.
systemd-analyze:
* systemd-analyze gained a new "chid" verb, which shows the "Computer
Hardware IDs" (CHIDs) of the local system. This is useful for
preparing CHID-to-DeviceTree mappings when building UKIs.
* systemd-analyze gained a new "transient-settings" verb, which shows
all unit settings one can configure dynamically via the
"--property="/"-p" switch when invoking transient units.
* systemd-analyze gained a new "unit-shell" verb that invokes an
interactive shell inside the namespaces of the main process
of a specified unit. This is useful for debugging unit sandboxes, and
getting an idea how things look like from the "inside" of a service.
* systemd-analyze gained a new "unit-gdb" verb to attach a debugger
to a unit.
Other:
* systemd-ask-password now provides a small Varlink API to
interactively query the user for a password using the usual agent
logic. This makes it easier for external programs (for example
daemons) to query for boot-time passwords and similar, using
systemd's infrastructure.
* The logging logic in systemd's codebase now implements the
DEBUG_INVOCATION= interface added to service management in v257. Or
in other words: the RestartMode=debug setting may now be added for
any of systemd's own service and has the intended effect of enabling
debug logging if it gets automatically restarted.
* The "package note" specification ELF binaries has been extended to
cover PE binaries (i.e. UEFI binaries), too.
* New kernel command line parameters systemd.break= and
rd.systemd.break= have been introduced that insert interactive (as
in: shell prompt) "breakpoints" into the boot process at various
locations, in order to simplify debugging. For now four breakpoints
are defined: "pre-udev", "pre-basic", "pre-mount",
"pre-switch-root". Similar functionality has previously existed in
the Dracut initrd generator, but is generalized with this new
concept, and extended to the post-switch-root boot phases.
* The systemd-path tool now learnt new paths for the per-system and
per-user credential store.
* A new tool systemd-pty-forward has been added that allocates a pseudo
TTY ("PTY") and invokes a process on it, forwarding any output to the
TTY it is invoked on. It can optionally apply background coloring and
suchlike, and is mostly just a separate tool that makes the PTY
forwarding logic used in systemd-nspawn, systemd-vmspawn, run0
available separately.
* systemd-oomd can now reload its configuration at runtime, following
the usual protocols.
* systemd-detect & ConditionVirtualization= now recognize the "Arm
Confidential Compute Architecture" (cca) confidential virtualization.
* systemd-detect-virt now correctly distinguishes between bare-metal
and virtualized machines in Google Compute Engine, and will not
report the former as virtualized.
* systemd-sysusers now generates Linux audit records when it adds
system users.
* systemd-firstboot's interactive prompts for locale or keymaps now
support tab completion.
* systemd-mount gained support for a new --canonicalize= switch that
may be used to turn off client-side path canonicalization before
trying to unmount some path.
* systemd-notify gained a new --fork switch which inverts the role that
systemd-notify plays in the sd_notify() protocol: instead of sending
out notification messages, it will listen for them, forking off a
command that is expected to send them. Once READY=1 is received
systemd-notify will exit, leaving the child running. This is useful
for correctly forking off processes that implement the sd_notify()
protocol from shell scripts.
* systemd-fstab-generator now supports a root=bind:… syntax for
creating bind mounts for the root file system. This is useful for
booting into tarballs downloaded at boot. As an example, consider a
kernel command line like this:
rd.systemd.pull=tar,machine,verify=no:root:http://192.168.100.1:8081/image.tar root=bind:/run/machines/root ip=any
* libapparmor is now loaded via dlopen() instead of using direct shared
library linking. This allows downstream distributions to provide AA
support as a runtime option instead of making the AA userspace a
mandatory dependency.
* A new generic remote-integritysetup.target unit has been added that
matches remote-veritysetup.target and remote-cryptsetup.target's role
for remote block devices, but for dm-integrity devices.
* A new document about finding boot components and the root disk of the
OS has been added. It's available online here:
https://systemd.io/ROOTFS_DISCOVERY
* Whenever any systemd tool begins or ends a new TTY context (i.e. takes
over a TTY for some time) a new OSC sequence is now emitted, with
various details about the context. This new OSC sequence can be
interpreted by terminal emulators to visualize the context/source TTY
output originates from or to show various kinds of metadata for
it. The OSC sequence is specified in this document:
https://systemd.io/OSC_CONTEXT
Contexts are generated for systemd-nspawn/systemd-vmspawn boots, for
run0 or systemd-run sessions, whenever PAM TTY sessions start or end,
and when shell command executions start and end. Metadata sent along
contains hostname, machine ID, boot ID, exit status, unit information
and more.
* If PID 1 makes up a suitable $TERM for a TTY it activates a service
on (in case there are no other hints on how to choose it) it will now
also set $COLORTERM=truecolor. Moreover, if $COLORTERM or $NO_COLOR
are set on the kernel cmdline we'll now import them into PID1's
environment block, just like $TERM itself. Moreover, systemd-nspawn
and run0 will now propagate $COLORTERM and $NO_COLOR from the calling
to the target environment, if set, just like $TERM is already
handled. Or to say this with different words: the triplet of $TERM,
$COLORTERM, $NO_COLOR is now processed jointly and in similar ways,
wherever appropriate.
* systemd-update-done gained a new --root= switch to operate in
"offline" mode on a specific file system tree.
* A new template service systemd-validatefs@.service has been added
that can validate usage of file systems. Specifically, it will look
for certain extended attributes stored on the top-level directory
inode of the mount, which may encode various constraints on use of
the file system. For example, it may encode a directory path the file
system must be mounted to, a GPT type UUID that must be used for the
partition the file system is located in and more. This provides
protection in case GPT auto-discovery is used to discover the mounts,
but essential metadata outside of the file system itself has been
tampered with. This operates under the assumption that the extended
attributes on the root inode of the file system are protected by
dm-verity or dm-crypt/dm-integrity, even if the GPT metadata has no
equivalent cryptographic protection. If a file system carries these
extended attributes but they do not match the current use and
location of the file system an immediate reboot is triggered.
* systemd-gpt-auto-generator now understands a new mount option
x-systemd.validatefs for /etc/fstab entries. If specified an instance
of systemd-validatefs@.service is automatically pulled in by the
relevant mount.
* systemd-fstab-auto-generator and systemd-gpt-auto-generator now
understand root=off on the kernel command line which may be used to
turn off any automatic or non-automatic mounting of the root file
system. This is useful in scenarios where a boot process shall never
transition from initrd context into host context.
* systemd-ssh-proxy now supports an alternative syntax for connecting
to SSH-over-AF_VSOCK, in order to support scp and rsync better: "scp
foo.txt vsock%4711:" should work now. (The pre-existing syntax used
"/" instead of "%" as separator, which is ambiguous in scp/rsync
context even if not for ssh itself.)
* "systemctl start" and related verbs now support a new --verbose
mode. If specified the live log output of the units operated on is
shown as long as the operation lasts.
* sd-bus: a new API call sd_bus_message_dump_json() returns a JSON
representation of a D-Bus message.
* sd-daemon: a new call sd_pidfd_get_inode_id() has been added
for acquiring the unique inode ID of a pidfd, coupling the
$MAINPIDFDID/$MANAGERPIDFDID and session/machine leader pidfd IDs
exposed as described above.
* systemd-coredump will now attach a new COREDUMP_DUMPABLE= journal
field to all coredumps indicating the "dumpable" per-process flag (as
settable via PR_SET_DUMPABLE) at the moment the coredump took
place. It will also add a new journal field COREDUMP_BY_PIDFD= that
indicates whether the coredump was acquired via a stable pidfd to the
process.
* systemd-sysext (and portable services with sysexts applied) will now
take the os-release "ID_LIKE=" field into account when validating that
a sysext images is compatible with the underlying image. Previously
it would only check "ID=".
* A new UID range has been defined for "greeters", i.e. graphical login
prompt UIs that shall be security isolated from each other. This is
supposed to be used by graphical display managers (specifically:
gdm), to ensure that it is harder to exploit the UI sessions used to
prompt the user for login credentials, in order to gain access to the
prompts of other users.
* systemd-socket-activate gained a new --now switch which ensures the
specified binary is immediately invoked, and not delayed until a
connection comes in.
* systemd-ssh-generator will now generate the AF_VSOCK ssh listener
.socket unit, so that a tiny new helper "systemd-ssh-issue" is
invoked when the socket is bound, that generates a drop-in file
/run/issue.d/50-ssh-vsock.issue that is shown by "login" and other
subsystems at login time. The file reports the AF_VSOCK CID of the
system, along with very brief information how to connect to the
system via ssh-over-AF_VSOCK. Or in other words: if the system is
booted up in an AF_VSOCK capable VM the console login screen shown
once boot-up is complete will tell you how to connect to the system
via SSH, if that's available.
* systemd-fsck gained fsck.mode and fsck.repair credentials support to
control the execution mode of fsck.
* systemd-quotacheck gained quotacheck.mode credential support to
control the execution mode of quotacheck.
Contributions from: 16mc1r, A. Wilcox, Aaron Rogers,
Abderrahim Kitouni, Adam Williamson, Adrian Vovk, Ahmad Fatoum,
Alberto Planas, Alex, Alex Xu (Hello71), Alexander Bruy,
Alexander Krabler, Alexander Kurtz, Alexander Shopov,
Alexander Stepchenko, Allison Karlitskaya, Aman Verma,
Américo Monteiro, Andika Triwidada, AndreFerreiraMsc,
Andreas Henriksson, Andreas Schneider, Andreas Stührk, Andres Beltran,
Andrew Sayers, Andrii Chubatiuk, André Monteiro, Andy Shevchenko,
Ani Sinha, Anthony Avina, Anthony Messina, Anton Ryzhov, Anton Tiurin,
Antonio Alvarez Feijoo, Arian van Putten, Arkadiusz Bokowy, Arnaudv6,
AsciiWolf, Avram Dorfman, Bastien Nocera, Beniamino Galvani,
Brett Holman, Busayo Dada, ButterflyOfFire, Carlo Teubner, Chris Grant,
Chris Hofstaedtler, Chris Mayo, Christian Glombek, Christian Hesse,
Christopher Head, Colin Foster, Cosima Neidahl, Craig McLure,
Daan De Meyer, Dai MIKURUBE, Dan McGregor, Dan Streetman,
Daniel Foster, Daniel Rusek, Daniil, David C. Manuelda, David Härdeman,
David Rheinsberg, David Tardon, DeKoile, Debarshi Ray, Deli Zhang,
Devilish Spirits, Dimitri John Ledkov, Duncan Overbruck, Dusty Mabe,
Eaterminer, Eisuke Kawashima, Emilio Sepulveda, Emir SARI,
Emmanuel Ferdman, Enrico Tagliavini, Erik Larsson, Erin Shepherd,
Ettore Atalan, Fabian Möller, Fabian Vogt, Fco. Javier F. Serrador,
Federico Giovanardi, Felix Pehla, Fleuria, Florian Schmaus, Franck Bui,
Frantisek Sumsal, Frede Braendstrup, Gabríel Arthúr Pétursson,
Gavin Li, George Tsiamasiotis, Graham Clinch, Grimmauld, H A, Hang Li,
Harrison Vanderbyl, Hendrik Wolff, Henri Aunin, Igor Opaniuk, Itxaka,
Ivan Kruglov, Ivan Trubach, Jack Wu, Jacob McNamee, James Hilliard,
Jan Engelhardt, Jan Fooken, Jan Kalabza, Jan Macku, Jan Vaclav,
Jan Čermák, Jared Baur, Jaroslav Škarvada, Jasmine Andrever-Wright,
Javier Francisco, Jelle van der Waa, Jeremy Linton, Jesper Nilsson,
Jesse Guo, Jim Spentzos, Joaquim Monteiro, Joey Holtzman,
John Rinehart, Jonas Gorski, Jordan Petridis, Jose Ortuno,
Josh Triplett, Jules Lamur, Justinas Kairys, Jörg Behrmann,
Kamil Páral, Katariina Lounento, Kevin P. Fleming, Khem Raj, KidGrimes,
Kurt Borja, Lennart Poettering, Li Tian, Lin Jian, Linus Heckemann,
Lorenzo Arena, Louis Sautier, LuK1337, Luca Boccassi,
Lucas Adriano Salles, Luke Yeager, Lukáš Nykrýn, Luna Jernberg,
Léane GRASSER, Marco Trevisan (Treviño), Marcos Alano,
Mario Limonciello, Markus Kurz, Martin Homuth-Rosemann,
Martin Hundebøll, Martin Srebotnjak, Martin Wilck, Mate Kukri,
Matteo Croce, Matthew Schwartz, Matthias Gerstner, Matthias Lisin,
Matthieu Baerts (NGI0), Matthieu LAURENT, MaxHearnden,
Michael Catanzaro, Michael Ferrari, Michael Limiero, Michael Olbrich,
Michal Koutný, Michal Sekletár, Michał Moczulski, Mike Yuan,
Miroslav Lichvar, Morten Hauke Solvang,
Muhammad Nuzaihan Bin Kamal Luddin, Myrrh Periwinkle, Nathan,
NetSysFire, Nick Labich, Nick Owens, Nick Rosbrook, Nils K,
Noel Georgi, Nuno Sá, Oliver Schramm, Paul Fertser, Pavithra Barithaya,
Philip Freeman, Philip Withnall, Piotr Drąg, Pontus Lundkvist, Raura,
Ricky Tigg, RocketDev, Ronan Pigott, Rostislav Lastochkin,
Rudi Heitbaum, Ryan Blue, Ryan Thompson, Ryan Wilson, Salim B,
Salvatore Cocuzza, Sam James, Sam Leonard, Samuel Dionne-Riel,
Sea-Eun Lee, Septatrix, Sergey A, Shubhendra Kushwaha, Silvio Knizek,
Solar Designer, SoloSaravanan, Sonia Zorba, Soumyadeep Ghosh,
Stefan Hansson, Stefan Herbrechtsmeier, Steve Ramage,
Temuri Doghonadze, TheHillBright, Thomas Hebb, Thorsten Kukuk,
Tim Crawford, Tim Small, Tim Vangehugten, Tobias Heider,
Tobias Klauser, Todd C. Miller, Tommi Rantala, Tommy Unger, Trollimpo,
Ubuntu, Valentin David, Valentin Hăloiu, Vasiliy Kovalev,
Vishal Chillara Srinivas, Vishwanath Chandapur, Vitaly Kuznetsov,
Volodymyr Shkriabets, Vyacheslav Yurkov, Werner Sembach, Y T,
Yaping Li, Yu Watanabe, ZIHCO, Zbigniew Jędrzejewski-Szmek,
andrejpodzimek, anonymix007, anthisfan, cvlc12, damnkiwi6120, davjav,
fishears, hanjinpeng, haxibami, herbrechtsmeier, honjow, hsu zangmen,
igo95862, jane400, jinyaoguo, joo es, kanitha chim, keentux, kmeaw,
luc-salles, madroach, maia x., msizanoen, naly zzwd, nkraetzschmar,
nl6720, novenary, peelz, persmule, richfifeg, ssoss, tim tom, tuxmainy,
tytan652, val4oss, ver4a, victor-ok, vlefebvre, wrvsrx, wtmpx,
xinpeng wang, z z, Дамјан Георгиевски, наб, 铝箔, 김인수
— Edinburgh, 2025/09/17
CHANGES WITH 257:
Incompatible changes:
* The --purge switch of systemd-tmpfiles (which was added in v256) has
been reworked: it will now only apply to tmpfiles.d/ lines marked
with the new "$" flag. This is an incompatible change, and means any
tmpfiles.d/ files which shall be used together with --purge need to
be updated accordingly. This change has been made to make it harder
to accidentally delete too many files when using --purge incorrectly.
* The systemd-creds 'cat' verb now expects base64-encoded encrypted
credentials as input, for consistency with the 'decrypt' verb and the
LoadCredentialEncrypted= service setting. Previously it could only
read raw, unencoded binary data.
* Support for automatic flushing of the nscd user/group database caches
has been dropped.
* The FileDescriptorName= setting for socket units is now honored by
Accept=yes sockets too, where it was previously silently ignored and
"connection" was used unconditionally.
* systemd-logind now always obeys block inhibitor locks, where previously
it ignored locks taken by the caller or when the caller was root. A
privileged caller can always close the other sessions, remove the
inhibitor locks, or use --force or --check-inhibitors=no to ignore the
inhibitors. This change thus doesn't affect security, since everything
that was possible before at a given privilege level is still possible,
but it should make the inhibitor logic easier to use and understand,
and also help avoiding accidental reboots and shutdowns. New 'block-weak'
inhibitor modes were added, if taken they will make the inhibitor lock
work as in the previous versions. Inhibitor locks can also be taken by
remote users (subject to polkit policy).
* systemd-nspawn will now mount the unified cgroup hierarchy into a
container if no systemd installation is found in a container's root
filesystem. $SYSTEMD_NSPAWN_UNIFIED_HIERARCHY=0 can be used to override
this behavior.
* /dev/disk/by-id/nvme-* block device symlinks without an NVMe
namespace identifier are now fixed to namespace 1 of the device. If
no namespace 1 exists for a device no such symlink is
created. Previously, these symlinks would point to an unspecified
namespace, and thus not be strictly stable references to
multi-namespace NVMe devices. These un-namespaced symlinks are mostly
obsolete, users and applications should always use the ones with
encoded namespace information instead. This change should not affect
too many systems, because most NVMe devices only know a namespace 1
by default.
* Support for cgroup v1 ('legacy' and 'hybrid' hierarchies) is now
considered obsolete and systemd by default will ignore configuration
that enables them. To forcibly reenable cgroup v1 support,
SYSTEMD_CGROUP_ENABLE_LEGACY_FORCE=1 must additionally be set on the
kernel command line.
Announcements of Future Feature Removals:
* The D-Bus method org.freedesktop.systemd1.StartAuxiliaryScope() is
deprecated because accounting data and such cannot be reasonably
migrated between cgroups. It is likely to be fully removed in a
future release (reach out if you have use cases).
* The recommended kernel baseline version has been bumped to v5.4
(released in 2019). Expect limited testing on older kernel versions,
where "old-kernel" taint flag would also be set. Support for them
will be phased out in a future release in 2025, i.e. we expect to bump
the minimum baseline to v5.4 then too.
* The complete removal of support for cgroup v1 ('legacy' and 'hybrid'
hierarchies) is scheduled for v258.
* Support for System V service scripts is deprecated and will be
removed in v258. Please make sure to update your software
*now* to include a native systemd unit file instead of a legacy
System V script to retain compatibility with future systemd releases.
* To work around limitations of X11's keyboard handling systemd's
keyboard mapping hardware database (hwdb.d/60-keyboard.hwdb) so far
mapped the microphone mute and touchpad on/off/toggle keys to the
function keys F20, F21, F22, F23 instead of their correct key codes.
This key code mangling will be removed in the next systemd release.
To maintain compatibility with X11 applications that rely on the old
function key code mappings, this mangling has now been moved to the
relevant X11 keyboard driver modules. In order to ensure these keys
continue to work, update to xf86-input-evdev >= 2.11.0 and
xf86-input-libinput >= 1.5.0 before updating to systemd >= 258.
* Support for the SystemdOptions EFI variable is deprecated.
'bootctl systemd-efi-options' will emit a warning when used. It seems
that this feature is little-used and it is better to use alternative
approaches like credentials and confexts. The plan is to drop support
altogether at a later point, but this might be revisited based on
user feedback.
* systemd-run's switch --expand-environment= which currently is disabled
by default when combined with --scope, will be changed in a future
release to be enabled by default.
libsystemd:
* systemd's JSON API is now available as public interface of
libsystemd, under the name "sd-json". The purpose of the library is
to allow structures to be conveniently created in C code and
serialized to JSON, and for JSON to be conveniently deserialized into
in-memory structures, using callbacks to handle specific
keys. Various data types like integers, floats, booleans, strings,
UUIDs, base64-encoded and hex-encoded binary data, and arrays are
supported natively. The library has been part of systemd for a while
as internal component, and is now made publicly available. One major
user of sd-json is sd-varlink (see below). Note that the
documentation of sd-json is very much incomplete for now, but the
systemd codebase provides plenty real-life code examples.
* systemd's Varlink IPC API is now available as part of libsystemd,
under the name "sd-varlink". This library is a C implementation of
the Varlink IPC system (https://varlink.org/) that has been adopted
by systemd for various interfaces. It relies on the sd-json JSON
component, see above. Note that the documentation of sd-varlink is
very much incomplete for now, but the systemd codebase provides
plenty real-life code examples.
* sd-bus gained a new call sd_bus_pending_method_calls() which returns
the number of currently open asynchronous method calls initiated on
this connection towards peers.
* sd-device gained a new call sd_device_monitor_is_running() that
returns whether the specified monitor object is already running. It
also gained sd_device_monitor_get_fd(),
sd_device_monitor_get_events(), sd_device_monitor_get_timeout() and
sd_device_monitor_receive() to permit sd-device to run on top of a
foreign event loop implementation. It also gained
sd_device_get_driver_subsystem() which returns the subsystem of
driver objects. The new sd_device_get_device_id() call returns a
short string identifying the device record.
System and Service Management:
* The environment variable $REMOTE_ADDR is now set when using
per-connection socket activation for AF_UNIX stream sockets. It
contains the AF_UNIX peer address of the connection. (Previously the
environment variable was only set for IP sockets.)
* Multipath TCP (MPTCP) is now supported as a socket protocol for
.socket units.
* A new /etc/fstab option x-systemd.wants= creates "Wants="
dependencies. (This is similar to the previously available
x-systemd.requires=.)
* The initialization of the system clock during boot and updates has
been simplified: both PID 1 or systemd-timesyncd will pick the latest
minimum time as indicated by the compiled-in epoch,
/usr/lib/clock-epoch, and /var/lib/systemd/timesync/clock. See
systemd(1) for an detailed updated description.
* The kernel's Ctrl-Alt-Delete handling is re-enabled during late
shutdown, so that the user may use it to initiate a reboot if the
system freezes otherwise.
* The new value "identity" for the unit setting PrivateUsers= may be
used to request a user namespace with an identity mapping for the
first 65536 UIDs/GIDs. This is analogous to the systemd-nspawn's
--private-users=identity.
* The new value "disconnected" for the unit setting PrivateTmp= may be
used to specify that a separate tmpfs instance should be used for
/tmp/ and /var/tmp/ for the unit.
* The server manager (and various other tools too) use pidfds in more
places to refer to processes.
* A build option -D link-executor-shared=false can be used to build
the systemd-executor binary (added in a previous release) in a way
where it does not link to shared libsystemd-shared-….so library.
PID1 holds a reference to the executor binary that was on disk when
the manager was started or restarted, but the shared libraries it is
linked to are not loaded until the executor binary needs to be used.
This partial static linking is a workaround for the issue where,
during upgrades, the old libsystemd-shared-….so may have already
been removed and the pinned executor binary will just fail to
execute.
* The systemd.machine_id= kernel command line parameter interpreted by
PID 1 now supports an additional special value: if set to "firmware"
the machine ID is initialized from the SMBIOS/DeviceTree system
UUID. (Previously this was already done automatically in VM
environments, this extends the concept to any system, but only on
explicit request via this option.)
* The ImportCredential= setting in service unit files now permits
renaming of credentials as they are imported.
* The RestartMode= setting gained a new "debug" value. If specified and
the service fails so that it shall be restarted it is invoked in
"debugging mode". Debugging mode means that the $DEBUG_INVOCATION
environment variable will be set to "1" for the new
invocation. Moreover, any setting LogLevelMax= will be temporarily
changed to "debug" for the next invocation. This mode is useful to
automatically repeat invocation of tools in case they fail – but with
additional logging or testing routines enabled.
* A new service setting BindLogSockets= has been added that
controls whether the AF_UNIX sockets required for logging shall be
bind mounted to the mount sandbox allocated for the service.
* At early boot, PID 1 will now optionally load a policy for the new
Linux IPE LSM.
* Transient services (as invoked by the StartTransientUnit() D-Bus
method) may now receive additional, arbitrary file descriptors to
pass to executed service processes during activation using the new
ExtraFileDescriptor= unit property.
* Calendar .timer units gained a new boolean DeferReactivation=
option. If enabled and the repetitive calendar timer elapses again
while the service the timer activates is still running, immediate
reactivation of the service once it finishes is skipped, and the
timer has to elapse again before the service is reactivated.
* Generator processes invoked by the service manager will now receive a
new environment variable $SYSTEMD_SOFT_REBOOTS_COUNT that indicates
how many times the system has been soft-rebooted since the kernel
initialized.
* A new service property ManagedOOMMemoryPressureDurationSec= has been
added that complements the existing
ManagedOOMMemoryPressureDurationLimit= and specifies the PSI
measurement interval for the specific unit.
* The sd_notify() protocol has been extended to allow changing the main
PID of a process by providing a pidfd of the new main process, or by
specifying the pidfd inode number. Previously this was only supported
by specifying the classic UNIX PID, which of course is racy.
* The SocketUser=/SocketGroup= settings of .socket units are now also
applied to POSIX message queues.
* The ProtectControlGroups= unit file setting now supports two
additional values: if set to "private" a new cgroup namespace is
allocated for the service and cgroupfs mounted accordingly; if set to
"strict" a new cgroup namespace is allocated for the service, and
cgroupfs is mounted read-only for the service.
* The StateDirectory=, RuntimeDirectory=, CacheDirectory=,
LogsDirectory=, and ConfigurationDirectory= settings gained support
for configuring the respective directories as read-only, via a ':ro'
flag that can be appended to each setting's value.
* When DynamicUser= is combined with
StateDirectory=/RuntimeDirectory=/CacheDirectory=/LogsDirectory= and
ID mapped mounts are available on the referenced path, the data in
there is now preferably made available by establishing ID mapped from
the "nobody" user to the dynamic user, rather than via recursive
chown()ing.
* A new service property PrivatePIDs= has been added that runs executed
processes as PID 1 - the init process - within their own PID
namespace. PrivatePIDs= also mounts /proc/ so only processes within
the new PID namespace are visible.
systemd-udevd:
* udev rules now set 'uaccess' for /dev/udmabuf, giving locally
logged-in users access to the hardware. This is useful in order to
support IPMI cameras with libcamera.
* Serial port devices will no longer show up as systemd units, unless
they have an IO port or memory assigned to them. This means that only
serial ports that actually exist should show up as .device units now.
* mtd devices (i.e. certain kinds of flash memory devices) will now
show up as .device units in systemd.
* The firmware_node/sun sysfs attribute will now be used (if available)
for naming slot-based network interfaces, i.e. ID_NET_NAME_SLOT.
Moreover the interface aliases specified in DeviceTree are now
searched for both on the interface's parent device (as before) and
the device itself (new).
* Various USB hardware wallets are now recognized by udev via a .hwdb
file, and get the ID_HARDWARE_WALLET= property set, which enables
"uaccess" for them, i.e. direct unprivileged access.
* udevadm info will now output the device ID string in lines prefixed
with "J:", and the driver subsystem in lines prefixed with "B:".
* udev rules files now support case-insensitive attribute matching
(e.g. ATTR{foo}==i"abcd")
systemd-logind:
* New DesignatedMaintenanceTime= configuration option allows shutdowns
to be automatically scheduled at the specified time.
* logind now reacts to Ctrl-Alt-Shift-Esc being pressed. It will send
out a org.freedesktop.login1.SecureAttentionKey signal, indicating a
request by the user for the system to display a secure login dialog.
The handling of SAK can be suppressed in logind configuration.
* logind now supports handing off session-managed access to hidraw
devices via its D-Bus APIs, the same way it already supports that for
DRM and evdev input devices. This permits unprivileged clients to get
hidraw fds for a device, that are automatically suspended when the
session switches away.
* systemd-logind now exposes two D-Bus properties CanLock and CanIdle
for all sessions. These properties indicate whether the session's
class supports screen locking and idleness detection.
* systemd-inhibit now allows interactive polkit authorization. It
gained a --no-ask-password option to suppress it.
systemd-machined:
* Unprivileged clients are now allowed to register VMs and containers.
Machines started via the systemd-vmspawn@.service unit will now be
registered with systemd-machined.
* systemd-machined gained a pretty complete set of Varlink APIs
exposing its functionality. This is an alternative to the
pre-existing D-Bus interface.
systemd-resolved and resolvectl:
* The resolvconf command now supports '-p' switch. If specified, the
interface will not be used as the default route for domain name
lookups.
* resolvectl now enables interactive polkit authorization. It gained a
--no-ask-password option to suppress it.
* systemd-resolved now implements continuous mDNS querying as per
RFC6762 §5.2. Clients can subscribe to the notification stream using
varlink.
systemd-networkd and networkctl:
* IPv6 address labels can be also configured in a new [IPv6AddressLabel]
section with Prefix= and Label= settings in networkd.conf. Please see
networkd.conf(5) for more details.
* 'networkctl edit' can now read the new file contents from standard
input with the new --stdin option.
* 'networkctl edit' and 'cat' now support editing/showing .netdev files
by link. 'networkctl cat' can also list all configuration files
associated with an interface at once with ':all'.
* networkctl gained a --no-ask-password option to suppress interactive
polkit authorization.
* "mac" has been added to the default AlternativeNamesPolicy= setting
for network links (via 99-default.link). This means "enx*" interface
names will now be added to the list of alternative interface names by
default, for all interfaces that have a MAC address assigned
by hardware.
* networkd .netdev bridge devices gained a new setting FDBMaxLearned=
for setting a limit on the number of dynamically learned FDB entries.
* networkd .network files for bridge devices now support Layer 2 (in
addition to the pre-existing Layer 3) MDB entries, via
MulticastGroupAddress=.
* systemd-networkd will now log when per-network sysctls belonging to
network interfaces managed by it are changed outside of networkd,
thus highlighting conflict of ownership/management of these knobs.
* systemd-networkd will now make RFC9463 DNR fields available to
systemd-resolved, for automatic DNS DoT configuration, and similar.
* The "dhcp" and "dhcp-on-stop" values for KeepConfiguration= setting in
.network file are replaced with "dynamic" and "dynamic-on-stop",
respectively. When specified, systemd-networkd will preserve all
dynamic configurations via DHCPv4, DHCPv6, NDISC, and IPv4LL with
ACD, while previously only DHCPv4 configurations were kept. Also,
when systemd-networkd is restarted, regardless of the setting, these
dynamic configurations are unconditionally kept. So, systemd-networkd
can be restarted without disturbing ongoing connections.
* systemd-networkd now updates traffic control configuration without
clearing existing settings. Thus, those settings can be updated by
editing relevant .network files and triggering 'networkctl reload'.
* systemd-networkd now gracefully updates netdev settings specified in
.netdev files when 'networkctl reload' is called. Previously, if the
relevant interfaces existed, new settings would not be applied. Now,
new settings will be applied if possible. Some settings cannot be
updated after a netdev is configured, e.g. VLAN ID can be only
specified on creation. To change such settings, user needs to remove
existing interfaces, and invoke 'networkctl reload' or restart
systemd-networkd.
systemd-boot, systemd-stub, and related tools:
* The EFI stub now supports loading of .ucode sections with microcode
from PE add-on files. It also now supports loading .initrd sections
from PE add-on files.
* A new .profile PE section type is now documented and supported in
systemd-measure, ukify, systemd-stub and systemd-boot. These new
sections allow multiple "profiles" to be stored together in the UKI,
where each .profile section creates groupings of sections in the UKI,
allowing some sections to be shared and other sections like .cmdline
or .initrd unique to the profile. This may be used to provide a
single UKI that synthesizes multiple menu items in the boot menu (for
example, a regular one to boot, plus a debugging one, or a factory
reset one, and so on – which only differ in kernel command line, but
nothing else).
* New .dtbauto and .hwids sections are now documented and supported in
systemd-measure, ukify, systemd-stub, and systemd-boot. A single UKI
can contain multiple .dtbauto sections, and the 'compatible' string
therein will be compared with the equivalent field in the DTB
provided by the firmware, if present. If absent, SMBIOS will be used
to calculate hardware IDs (CHIDs) and look them up in the content of
.hwids, hopefully revealing an fallback 'compatible' string. This
allows including multiple DTBs in a single UKI, with systemd-stub
automatically loading the correct one for the current hardware.
* ukify gained an --extend switch to import an existing UKI to
be extended, and a --measure-base= switch to support measurement
of multi-profile UKIs.
* ukify gained a --certificate-provider switch to use an OpenSSL
provider to load the certificate used to sign artifacts, instead of
having to provide the path to a file on disk.
* bootctl, systemd-keyutil, systemd-measure, systemd-repart, and
systemd-sbsign gained a new --certificate-source switch that allows
loading the X.509 certificate from an OpenSSL provider instead of a
file system path.
* systemd-boot's menu will now react to volume up/down rocker presses
the same way as to arrow up/down presses: they move the menu item up
or down. This is useful on device form factors that have only a
volume rocker but no arrow keys (e.g. phones).
* systemd-stub will report the partition UUID and image identifier its
UKI executable is placed on separately from the data systemd-boot
provides about where to find its own executable, via EFI
variables. This is useful when systemd-boot and UKIs are placed on
distinct partitions (i.e. ESP and XBOOTLDR).
* bootctl gained new switches --print-loader-path and --print-stub-path
that output the path to the boot loader or UKI used for the current
boot.
* bootctl kernel-identify now recognizes EFI add-ons.
* bootctl gained a --random-seed=yes|no option to control provisioning
of the random seed file in the ESP. (This is useful when producing an
image that will be used in multiple instances.)
* bootctl now optionally supports installing UEFI Secure Boot databases
(i.e. db/dbx/… databases in ESL format) for systemd-boot to pick up
and automatically enroll if the system is booted in Setup Mode. This
is controlled via bootctl's new --secure-boot-auto-enroll=yes switch
(and some auxiliary ones). A certificate can be provided in DER
format, and is automatically converted into an ESL, as needed.
* bootctl, systemd-measure, systemd-repart when referencing signing
keys on OpenSSL engines may now query for PINs and similar via
systemd's native systemd-ask-password logic (and take benefit of its
caching and UI).
* A new systemd-sbsign tool has been added, that can be used to sign
EFI binaries (PE) for Secure Boot. This tool supports OpenSSL engines
and providers, with pin caching support for PKCS11. ukify supports it
as an alternative to sbsigntool and pesign.
* A new systemd-keyutil tool has been added, that can be used to perform
various operations on private keys and X.509 certificates.
The journal:
* journalctl can now list invocations of a unit with the
--list-invocation options and show logs for a specific invocation
with the new --invocation/-I option. (This is analogous to the
--list-boots/--boot/-b options.)
systemd-sysupdate and related tools:
* systemd-sysupdated has been added as system service, allowing
unprivileged clients to update the system via D-Bus calls. Note that
for now the systemd-sysupdated API is considered experimental, and is
not considered stable yet.
A new updatectl command-line tool can be used to control the
service.
* systemd-sysupdate gained a new --offline option to force it to
operate locally. This is useful when listing locally installed
versions.
* systemd-sysupdate gained a new --transfer-source= option to set the
directory to which transfer sources configured with
PathRelativeTo=explicit will be interpreted.
* systemd-sysupdate now reports download progress via sd_notify().
* systemd-sysupdate now supports output in JSON mode for all commands.
* systemd-sysupdate definitions may now carry references to ChangeLog
and AppStream metadata.
* Transfer definitions for systemd-sysupdate are supposed to carry the
".transfer" suffix now, changing from ".conf". The latter remains
supported for compatibility, but it's recommended to rename all files
reflecting this suffix change.
* systemd-sysupdate now supports new ".feature" files that may be
used in conjunction with ".transfer" files to group them together, and
allow them to be turned off or on, individually per group.
TPM & systemd-cryptsetup:
* The 'has-tpm2' verb which reports whether TPM2 functionality is
available has been moved from systemd-creds to systemd-analyze.
* systemd-tpm2-setup will gracefully handle TPMs that have a PIN set on
the TPM, and not attempt to automatically set up a Storage Root Key
(SRK) in that case.
* New crypttab option password-cache=yes|no|read-only can be used to
customize password caching.
* New crypttab options fido2-pin=, fido2-up=, fido2-uv= can be used to
enable/disable the PIN query, User Presence check, and User
Verification.
* systemd-cryptenroll gained new options --fido2-salt-file= and
--fido2-parameters-in-header= to simplify manual enrollment of FIDO2
tokens.
* systemd-cryptenroll, systemd-repart, and systemd-storagetm gained a
new --list-devices option to list appropriate candidate block
devices.
* systemd-cryptenroll/systemd-cryptsetup now support combined signed
PCR policies and local systemd-pcrlock policies for unlocking a
disk. Or in other words, it's now possible to bind unlocking of a
local disk to a specific OS vendor *and* a locally managed set of
measurements describing the local system.
varlinkctl:
* varlinkctl gained a new verb 'list-methods' to show a list of
methods implemented by a service.
* varlinkctl gained a --quiet/-q option to suppress method call
replies.
* varlinkctl gained a --graceful= option to suppress specific Varlink
errors, and treat them as success.
* varlinkctl gained a --timeout= option to limit how long the
invocation can take.
* varlinkctl allows remote invocations over ssh, via the new
"ssh-exec:" address specification. It'll make an ssh connection,
start the specified executable on the remote side, and communicate
with the remote process using the Varlink protocol.
The "ssh:" address specification has been renamed to "ssh-unix:"
(reflecting the fact it is used to connect to a remote AF_UNIX socket
via SSH). The old syntax is still supported for backwards
compatibility.
* varlinkctl's 'introspect' verb no longer requires specification of an
interface name. If none is specified all interfaces exposed by the
service are shown. Moreover, more than one interface name may be
specified now, in which case all specified ones are displayed.
systemd-repart: