1 systemd System and Service Manager
3 CHANGES WITH 255 in spe:
5 Announcements of Future Feature Removals and Incompatible Changes:
9 Announcements of Future Feature Removals and Incompatible Changes:
11 * Support for split-usr (/usr/ mounted separately during late boot,
12 instead of being mounted by the initrd before switching to the rootfs)
13 and unmerged-usr (parallel directories /bin/ and /usr/bin/, /lib/ and
14 /usr/lib/, …) has been removed. For more details, see:
15 https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html
17 * We intend to remove cgroup v1 support from a systemd release after
18 the end of 2023. If you run services that make explicit use of
19 cgroup v1 features (i.e. the "legacy hierarchy" with separate
20 hierarchies for each controller), please implement compatibility with
21 cgroup v2 (i.e. the "unified hierarchy") sooner rather than later.
22 Most of Linux userspace has been ported over already.
24 * Support for System V service scripts is now deprecated and will be
25 removed in a future release. Please make sure to update your software
26 *now* to include a native systemd unit file instead of a legacy
27 System V script to retain compatibility with future systemd releases.
29 * Support for the SystemdOptions EFI variable is deprecated.
30 'bootctl systemd-efi-options' will emit a warning when used. It seems
31 that this feature is little-used and it is better to use alternative
32 approaches like credentials and confexts. The plan is to drop support
33 altogether at a later point, but this might be revisited based on
36 * systemd-run's switch --expand-environment= which currently is disabled
37 by default when combined with --scope, will be changed in a future
38 release to be enabled by default.
40 * "systemctl switch-root" is now restricted to initrd transitions only.
42 Transitions between real systems should be done with
43 "systemctl soft-reboot" instead.
45 * The "ip=off" and "ip=none" kernel command line options interpreted by
46 systemd-network-generator will now result in IPv6RA + link-local
47 addressing being disabled, too. Previously DHCP was turned off, but
48 IPv6RA and IPv6 link-local addressing was left enabled.
50 * The NAMING_BRIDGE_MULTIFUNCTION_SLOT naming scheme has been deprecated
53 * SuspendMode=, HibernateState= and HybridSleepState= in the [Sleep]
54 section of systemd-sleep.conf are now deprecated and have no effect.
55 They did not (and could not) take any value other than the respective
56 default. HybridSleepMode= is also deprecated, and will now always use
57 the 'suspend' disk mode.
61 * The way services are spawned has been overhauled. Previously, a
62 process was forked that shared all of the manager's memory (via
63 copy-on-write) while doing all the required setup (e.g.: mount
64 namespaces, CGroup configuration, etc.) before exec'ing the target
65 executable. This was problematic for various reasons: several glibc
66 APIs were called that are not supposed to be used after a fork but
67 before an exec, copy-on-write meant that if either process (the
68 manager or the child) touched a memory page a copy was triggered, and
69 also the memory footprint of the child process was that of the
70 manager, but with the memory limits of the service. From this version
71 onward, the new process is spawned using CLONE_VM and CLONE_VFORK
72 semantics via posix_spawn(3), and it immediately execs a new internal
73 binary, systemd-executor, that receives the configuration to apply
74 via memfd, and sets up the process before exec'ing the target
75 executable. The systemd-executor binary is pinned by file descriptor
76 by each manager instance (system and users), and the reference is
77 updated on daemon-reexec - it is thus important to reexec all running
78 manager instances when the systemd-executor and/or libsystemd*
79 libraries are updated on the filesystem.
81 * Most of the internal process tracking is being changed to use PIDFDs
82 instead of PIDs when the kernel supports it, to improve robustness
85 * A new option SurviveFinalKillSignal= can be used to configure the
86 unit to be skipped in the final SIGTERM/SIGKILL spree on shutdown.
87 This is part of the required configuration to let a unit's processes
88 survive a soft-reboot operation.
90 * System extension images (sysext) can now set
91 EXTENSION_RELOAD_MANAGER=1 in their extension-release files to
92 automatically reload the service manager (PID 1) when
93 merging/refreshing/unmerging on boot. Generally, while this can be
94 used to ship services in system extension images it's recommended to
95 do that via portable services instead.
97 * The ExtensionImages= and ExtensionDirectories= options now support
98 confexts images/directories.
100 * A new option NFTSet= provides a method for integrating dynamic cgroup
101 IDs into firewall rules with NFT sets. The benefit of using this
102 setting is to be able to use control group as a selector in firewall
103 rules easily and this in turn allows more fine grained filtering.
104 Also, NFT rules for cgroup matching use numeric cgroup IDs, which
105 change every time a service is restarted, making them hard to use in
108 * A new option CoredumpReceive= can be set for service and scope units,
109 together with Delegate=yes, to make systemd-coredump on the host
110 forward core files from processes crashing inside the delegated
111 CGroup subtree to systemd-coredump running in the container. This new
112 option is by default used by systemd-nspawn containers that use the
115 * A new ConditionSecurity=measured-uki option is now available, to ensure
116 a unit can only run when the system has been booted from a measured UKI.
118 * MemoryAvailable= now considers physical memory if there are no CGroup
119 memory limits set anywhere in the tree.
121 * The $USER environment variable is now always set for services, while
122 previously it was only set if User= was specified. A new option
123 SetLoginEnvironment= is now supported to determine whether to also set
124 $HOME, $LOGNAME, and $SHELL.
126 * Socket units now support a new pair of
127 PollLimitBurst=/PollLimitInterval= options to configure a limit on
128 how often polling events on the file descriptors backing this unit
129 will be considered within a time window.
131 * Scope units can now be created using PIDFDs instead of PIDs to select
132 the processes they should include.
134 * Sending SIGRTMIN+18 with 0x500 as sigqueue() value will now cause the
135 manager to dump the list of currently pending jobs.
137 * If the kernel supports MOVE_MOUNT_BENEATH, the systemctl and
138 machinectl bind and mount-image verbs will now cause the new mount to
139 replace the old mount (if any), instead of overmounting it.
141 * Units now have MemoryPeak, MemorySwapPeak, MemorySwapCurrent and
142 MemoryZSwapCurrent properties, which respectively contain the values
143 of the cgroup v2's memory.peak, memory.swap.peak, memory.swap.current
144 and memory.zswap.current properties. This information is also shown in
145 "systemctl status" output, if available.
147 TPM2 Support + Disk Encryption & Authentication:
149 * systemd-cryptenroll now allows specifying a PCR bank and explicit hash
150 value in the --tpm2-pcrs= option.
152 * systemd-cryptenroll now allows specifying a TPM2 key handle (nv
153 index) to be used instead of the default SRK via the new
154 --tpm2-seal-key-handle= option.
156 * systemd-cryptenroll now allows TPM2 enrollment using only a TPM2
157 public key (in TPM2B_PUBLIC format) – without access to the TPM2
158 device itself – which enables offline sealing of LUKS images for a
159 specific TPM2 chip, as long as the SRK public key is known. Pass the
160 public to the tool via the new --tpm2-device-key= switch.
162 * systemd-cryptsetup is now installed in /usr/bin/ and is no longer an
163 internal-only executable.
165 * The TPM2 Storage Root Key will now be set up, if not already present,
166 by a new systemd-tpm2-setup.service early boot service. The SRK will
167 be stored in PEM format and TPM2_PUBLIC format (the latter is useful
168 for systemd-cryptenroll --tpm2-device-key=, as mentioned above) for
169 easier access. A new "srk" verb has been added to systemd-analyze to
170 allow extracting it on demand if it is already set up.
172 * The internal systemd-pcrphase executable has been renamed to
175 * The systemd-pcrextend tool gained a new --pcr= switch to override
176 which PCR to measure into.
178 * systemd-pcrextend now exposes a Varlink interface at
179 io.systemd.PCRExtend that can be used to do measurements and event
182 * TPM measurements are now also written to an event log at
183 /run/log/systemd/tpm2-measure.log, using a derivative of the TCG
184 Canonical Event Log format. Previously we'd only log them to the
185 journal, where they however were subject to rotation and similar.
187 * A new component "systemd-pcrlock" has been added that allows managing
188 local TPM2 PCR policies for PCRs 0-7 and similar, which are hard to
189 predict by the OS vendor because of the inherently local nature of
190 what measurements they contain, such as firmware versions of the
191 system and extension cards and suchlike. pcrlock can predict PCR
192 measurements ahead of time based on various inputs, such as the local
193 TPM2 event log, GPT partition tables, PE binaries, UKI kernels, and
194 various other things. It can then pre-calculate a TPM2 policy from
195 this, which it stores in an TPM2 NV index. TPM2 objects (such as disk
196 encryption keys) can be locked against this NV index, so that they
197 are locked against a specific combination of system firmware and
198 state. Alternatives for each component are supported to allowlist
199 multiple kernel versions or boot loader version simultaneously
200 without losing access to the disk encryption keys. The tool can also
201 be used to analyze and validate the local TPM2 event log.
202 systemd-cryptsetup, systemd-cryptenroll, systemd-repart have all been
203 updated to support such policies. There's currently no support for
204 locking the system's root disk against a pcrlock policy, this will be
205 added soon. Moreover, it is currently not possible to combine a
206 pcrlock policy with a signed PCR policy. This component is
207 experimental and its public interface is subject to change.
209 systemd-boot, systemd-stub, ukify, bootctl, kernel-install:
211 * bootctl will now show whether the system was booted from a UKI in its
214 * systemd-boot and systemd-stub now use different project keys in their
215 respective SBAT sections, so that they can be revoked individually if
218 * systemd-boot will no longer load unverified Devicetree blobs when UEFI
219 SecureBoot is enabled. For more details see:
220 https://github.com/systemd/systemd/security/advisories/GHSA-6m6p-rjcq-334c
222 * systemd-boot gained new hotkeys to reboot and power off the system
223 from the boot menu ("B" and "O"). If the "auto-poweroff" and
224 "auto-reboot" options in loader.conf are set these entries are also
225 shown as menu items (which is useful on devices lacking a regular
228 * systemd-boot gained a new configuration value "menu-disabled" for the
229 set-timeout option, to allow completely disabling the boot menu,
230 including the hotkey.
232 * systemd-boot will now measure the content of loader.conf in TPM2
235 * systemd-stub will now concatenate the content of all kernel
236 command-line addons before measuring them in TPM2 PCR 12, in a single
237 measurement, instead of measuring them individually.
239 * systemd-stub will now measure and load Devicetree Blob addons, which
240 are searched and loaded following the same model as the existing
241 kernel command-line addons.
243 * systemd-stub will now ignore unauthenticated kernel command line options
244 passed from systemd-boot when running inside Confidential VMs with UEFI
247 * systemd-stub will now load a Devicetree blob even if the firmware did
248 not load any beforehand (e.g.: for ACPI systems).
250 * ukify is no longer considered experimental, and now ships in /usr/bin/.
252 * ukify gained a new verb inspect to describe the sections of a UKI and
253 print the contents of the well-known sections.
255 * ukify gained a new verb genkey to generate a set of key pairs for
256 signing UKIs and their PCR data.
258 * The 90-loaderentry kernel-install hook now supports installing device
261 * kernel-install now supports the --json=, --root=, --image=, and
262 --image-policy= options for the inspect verb.
264 * kernel-install now supports new list and add-all verbs. The former
265 lists all installed kernel images (if those are available in
266 /usr/lib/modules/). The latter will install all the kernels it can
271 * A new option --copy-from= has been added that synthesizes partition
272 definitions from the given image, which are then applied by the
273 systemd-repart algorithm.
275 * A new option --copy-source= has been added, which can be used to specify
276 a directory to which CopyFiles= is considered relative to.
278 * New --make-ddi=confext, --make-ddi=sysext, and --make-ddi=portable
279 options have been added to make it easier to generate these types of
280 DDIs, without having to provide repart.d definitions for them.
282 * The dm-verity salt and UUID will now be derived from the specified
285 * New VerityDataBlockSizeBytes= and VerityHashBlockSizeBytes= can now be
286 configured in repart.d/ configuration files.
288 * A new Subvolumes= setting is now supported in repart.d/ configuration
289 files, to indicate which directories in the target partition should be
292 * A new --tpm2-device-key= option can be used to lock a disk against a
293 specific TPM2 public key. This matches the same switch the
294 systemd-cryptenroll tool now supports (see above).
298 * The journalctl --lines= parameter now accepts +N to show the oldest N
299 entries instead of the newest.
301 * journald now ensures that sealing happens once per epoch, and sets a
302 new compatibility flag to distinguish old journal files that were
303 created before this change, for backward compatibility.
307 * udev will now create symlinks to loopback block devices in the
308 /dev/disk/by-loop-ref/ directory that are based on the .lo_file_name
309 string field selected during allocation. The systemd-dissect tool and
310 the util-linux losetup command now supports a complementing new switch
311 --loop-ref= for selecting the string. This means a loopback block
312 device may now be allocated under a caller-chosen reference and can
313 subsequently be referenced without first having to look up the block
314 device name the caller ended up with.
316 * udev also creates symlinks to loopback block devices in the
317 /dev/disk/by-loop-inode/ directory based on the .st_dev/st_ino fields
318 of the inode attached to the loopback block device. This means that
319 attaching a file to a loopback device will implicitly make a handle
320 available to be found via that file's inode information.
322 * udevadm info gained support for JSON output via a new --json= flag, and
323 for filtering output using the same mechanism that udevadm trigger
326 * The predictable network interface naming logic is extended to include
327 the SR-IOV-R "representor" information in network interface names.
328 This feature was intended for v254, but even though the code was
329 merged, the part that actually enabled the feature was forgotten.
330 It is now enabled by default and is part of the new "v255" naming
333 * A new hwdb/rules file has been added that sets the
334 ID_NET_AUTO_LINK_LOCAL_ONLY=1 udev property on all network interfaces
335 that should usually only be configured with link-local addressing
336 (IPv4LL + IPv6LL), i.e. for PC-to-PC cables ("laplink") or
337 Thunderbolt networking. systemd-networkd and NetworkManager (soon)
338 will make use of this information to apply an appropriate network
339 configuration by default.
341 * The ID_NET_DRIVER property on network interfaces is now set
342 relatively early in the udev rule set so that other rules may rely on
343 its use. This is implemented in a new "net-driver" udev built-in.
347 * The "duid-only" option for DHCPv4 client's ClientIdentifier= setting
348 is now dropped, as it never worked, hence it should not be used by
351 * The 'prefixstable' ipv6 address generation mode now considers the SSID
352 when generating stable addresses, so that a different stable address
353 is used when roaming between wireless networks. If you already use
354 'prefixstable' addresses with wireless networks, the stable address
355 will be changed by the update.
357 * The DHCPv4 client gained a RapidCommit option, true by default, which
358 enables RFC4039 Rapid Commit behavior to obtain a lease in a
359 simplified 2-message exchange instead of the typical 4-message
360 exchange, if also supported by the DHCP server.
362 * The DHCPv4 client gained new InitialCongestionWindow= and
363 InitialAdvertisedReceiveWindow= options for route configurations.
365 * The DHCPv4 client gained a new RequestAddress= option that allows
366 to send a preferred IP address in the initial DHCPDISCOVER message.
368 * The DHCPv4 server and client gained support for IPv6-only mode
371 * The SendHostname= and Hostname= options are now available for the
372 DHCPv6 client, independently of the DHCPv4= option, so that these
373 configuration values can be set independently for each client.
375 * The DHCPv4 and DHCPv6 client state can now be queried via D-Bus,
376 including lease information.
378 * The DHCPv6 client can now be configured to use a custom DUID type.
380 * .network files gained a new IPv4ReversePathFilter= setting in the
381 [Network] section, to control sysctl's rp_filter setting.
383 * .network files gaiend a new HopLimit= setting in the [Route] section,
384 to configure a per-route hop limit.
386 * .network files gained a new TCPRetransmissionTimeoutSec= setting in
387 the [Route] section, to configure a per-route TCP retransmission
390 * A new directive NFTSet= provides a method for integrating network
391 configuration into firewall rules with NFT sets. The benefit of using
392 this setting is that static network configuration or dynamically
393 obtained network addresses can be used in firewall rules with the
394 indirection of NFT set types.
396 * The [IPv6AcceptRA] section supports the following new options:
397 UsePREF64=, UseHopLimit=, UseICMP6RateLimit=, and NFTSet=.
399 * The [IPv6SendRA] section supports the following new options:
400 RetransmitSec=, HopLimit=, HomeAgent=, HomeAgentLifetimeSec=, and
401 HomeAgentPreference=.
403 * A new [IPv6PREF64Prefix] set of options, containing Prefix= and
404 LifetimeSec=, has been introduced to append pref64 options in router
405 advertisements (RFC8781).
407 * The network generator now configures the interfaces with only
408 link-local addressing if "ip=link-local" is specified on the kernel
411 * The prefix of the configuration files generated by the network
412 generator from the kernel command line is now prefixed with '70-',
413 to make them have higher precedence over the default configuration
416 * Added a new -Ddefault-network=BOOL meson option, that causes more
417 .network files to be installed as enabled by default. These configuration
418 files will which match generic setups, e.g. 89-ethernet.network matches
419 all Ethernet interfaces and enables both DHCPv4 and DHCPv6 clients.
421 * If a ID_NET_MANAGED_BY= udev property is set on a network device and
422 it is any other string than "io.systemd.Network" then networkd will
423 not manage this device. This may be used to allow multiple network
424 management services to run in parallel and assign ownership of
425 specific devices explicitly. NetworkManager will soon implement a
430 * systemctl is-failed now checks the system state if no unit is
433 * systemctl will now automatically soft-reboot if a new root file system
434 is found under /run/nextroot/ when a reboot operation is invoked.
438 * Wall messages now work even when utmp support is disabled, using
439 systemd-logind to query the necessary information.
441 * systemd-logind now sends a new PrepareForShutdownWithMetadata D-Bus
442 signal before shutdown/reboot/soft-reboot that includes additional
443 information compared to the PrepareForShutdown signal. Currently the
444 additional information is the type of operation that is about to be
447 Hibernation & Suspend:
449 * The kernel and OS versions will no longer be checked on resume from
452 * Hibernation into swap files backed by btrfs are now
453 supported. (Previously this was supported only for other file
458 * A new systemd-vmspawn tool has been added, that aims to provide for VMs
459 the same interfaces and functionality that systemd-nspawn provides for
460 containers. For now it supports QEMU as a backend, and exposes some of
461 its options to the user. This component is experimental and its public
462 interface is subject to change.
464 * "systemd-analyze plot" has gained tooltips on each unit name with
465 related-unit information in its svg output, such as Before=,
466 Requires=, and similar properties.
468 * A new varlinkctl tool has been added to allow interfacing with
469 Varlink services, and introspection has been added to all such
470 services. This component is experimental and its public interface is
473 * systemd-sysext and systemd-confext now expose a Varlink service
474 at io.systemd.sysext.
476 * portable services now accept confexts as extensions.
478 * systemd-sysupdate now accepts directories in the MatchPattern= option.
480 * systemd-run will now output the invocation ID of the launched
481 transient unit and its peak memory usage.
483 * systemd-analyze, systemd-tmpfiles, systemd-sysusers, systemd-sysctl,
484 and systemd-binfmt gained a new --tldr option that can be used instead
485 of --cat-config to suppress uninteresting configuration lines, such as
486 comments and whitespace.
488 * resolvectl gained a new "show-server-state" command that shows
489 current statistics of the resolver. This is backed by a new
490 DumpStatistics() Varlink method provided by systemd-resolved.
492 * systemd-timesyncd will now emit a D-Bus signal when the LinkNTPServers
495 * vconsole now supports KEYMAP=@kernel for preserving the kernel keymap
498 * seccomp now supports the LoongArch64 architecture.
500 * seccomp may now be enabled for services running as a non-root User=
501 without NoNewPrivileges=yes.
503 * systemd-id128 now supports a new -P option to show only values. The
504 combination of -P and --app options is also supported.
506 * A new pam_systemd_loadkey.so PAM module is now available, which will
507 automatically fetch the passphrase used by cryptsetup to unlock the
508 root file system and set it as the PAM authtok. This enables, among
509 other things, configuring auto-unlock of the GNOME Keyring / KDE
510 Wallet when autologin is configured.
512 * Many meson options now use the 'feature' type, which means they
513 take enabled/disabled/auto as values.
515 * A new meson option -Dconfigfiledir= can be used to change where
516 configuration files with default values are installed to.
518 * Options and verbs in man pages are now tagged with the version they
519 were first introduced in.
521 * A new component "systemd-storagetm" has been added, which exposes all
522 local block devices as NVMe-TCP devices, fully automatically. It's
523 hooked into a new target unit storage-target-mode.target that is
524 suppsoed to be booted into via
525 rd.systemd.unit=storage-target-mode.target on the kernel command
526 line. This is intended to be used for installers and debugging to
527 quickly get access to the local disk. It's inspired by MacOS "target
528 disk mode". This component is experimental and its public interface is
531 * A new component "systemd-bsod" has been added, which can show logged
532 error messages full screen, if they have a log level of LOG_EMERG log
533 level. This component is experimental and its public interface is
536 * The systemd-dissect tool's --with command will now set the
537 $SYSTEMD_DISSECT_DEVICE environment variable to the block device it
538 operates on for the invoked process.
540 * The systemd-mount tool gained a new --tmpfs switch for mounting a new
541 'tmpfs' instance. This is useful since it does so via .mount units
542 and thus can be executed remotely or in containers.
544 * The various tools in systemd that take "verbs" (such as systemctl,
545 loginctl, machinectl, …) now will suggest a close verb name in case
546 the user specified an unrecognized one.
548 * libsystemd now exports a new function sd_id128_get_app_specific()
549 that generates "app-specific" 128bit IDs from any ID. It's similar to
550 sd_id128_get_machine_app_specific() and
551 sd_id128_get_boot_app_specific() but takes the ID to base calculation
552 on as input. This new functionality is also exposed in the
553 "systemd-id128" tool where you can now combine --app= with `show`.
555 * All tools that parse timestamps now can also parse RFC3339 style
556 timestamps that include the "T" and Z" characters.
558 * New documentation has been added:
560 https://systemd.io/FILE_DESCRIPTOR_STORE
561 https://systemd.io/TPM2_PCR_MEASUREMENTS
562 https://systemd.io/MOUNT_REQUIREMENTS
564 * The codebase now recognizes the suffix .confext.raw and .sysext.raw
565 as alternative to the .raw suffix generally accepted for DDIs. It is
566 recommended to name configuration extensions and system extensions
567 with such suffixes, to indicate their purpose in the name.
569 * The sd-device API gained a new function
570 sd_device_enumerator_add_match_property_required() which allows
571 configuring matches on properties that are strictly required. This is
572 different from the existing sd_device_enumerator_add_match_property()
573 matches of which one one needs to apply.
575 * The MAC address the veth side of an nspawn container shall get
576 assigned may now be controlled via the $SYSTEMD_NSPAWN_NETWORK_MAC
577 environment variable.
579 * The libiptc dependency is now implemented via dlopen(), so that tools
580 such as networkd and nspawn no longer have a hard dependency on the
581 shared library when compiled with support for libiptc.
583 * New rpm macros have been added: %systemd_user_daemon_reexec does
584 daemon-reexec for all user managers, and %systemd_postun_with_reload
585 and %systemd_user_postun_with_reload do a reload for system and user
588 * coredumpctl now propagates SIGTERM to the debugger process.
590 Contributions from: 김인수, Abderrahim Kitouni, Adam Goldman,
591 Adam Williamson, Alexandre Peixoto Ferreira, Alex Hudspith,
592 Alvin Alvarado, André Paiusco, Antonio Alvarez Feijoo,
593 Anton Lundin, Arian van Putten, Arseny Maslennikov, Arthur Shau,
594 Balázs Úr, beh_10257, Benjamin Peterson, Bertrand Jacquin,
595 Brian Norris, Charles Lee, Cheng-Chia Tseng, Chris Patterson,
596 Christian Hergert, Christian Hesse, Christian Kirbach,
597 Clayton Craft, commondservice, cunshunxia, Curtis Klein, cvlc12,
598 Daan De Meyer, Daniele Medri, Daniel P. Berrangé, Daniel Rusek,
599 Daniel Thompson, Dan Nicholson, Dan Streetman, David Rheinsberg,
600 David Santamaría Rogado, David Tardon, dependabot[bot],
601 Diego Viola, Dmitry V. Levin, Emanuele Giuseppe Esposito,
602 Emil Renner Berthing, Emil Velikov, Etienne Dechamps, Fabian Vogt,
603 felixdoerre, Felix Dörre, Florian Schmaus, Franck Bui,
604 Frantisek Sumsal, G2-Games, Gioele Barabucci, Hugo Carvalho,
605 huyubiao, Iago López Galeiras, IllusionMan1212, Jade Lovelace,
606 janana, Jan Janssen, Jan Kuparinen, Jan Macku, Jeremy Fleischman,
607 Jin Liu, jjimbo137, Joerg Behrmann, Johannes Segitz, Jordan Rome,
608 Jordan Williams, Julien Malka, Juno Computers, Khem Raj, khm,
609 Kingbom Dou, Kiran Vemula, Krzesimir Nowak, Laszlo Gombos,
610 Lennart Poettering, linuxlion, Luca Boccassi, Lucas Adriano Salles,
611 Lukas, Lukáš Nykrýn, Maanya Goenka, Maarten, Malte Poll,
612 Marc Pervaz Boocha, Martin Beneš, Martin Joerg, Martin Wilck,
613 Mathieu Tortuyaux, Matthias Schiffer, Maxim Mikityanskiy,
614 Max Kellermann, Michael A Cassaniti, Michael Biebl, Michael Kuhn,
615 Michael Vasseur, Michal Koutný, Michal Sekletár, Mike Yuan,
616 Milton D. Miller II, mordner, msizanoen, NAHO, Nandakumar Raghavan,
617 Neil Wilson, Nick Rosbrook, Nils K, NRK, Oğuz Ersen,
618 Omojola Joshua, onenowy, Paul Meyer, Paymon MARANDI, pelaufer,
619 Peter Hutterer, PhylLu, Pierre GRASSER, Piotr Drąg, Priit Laes,
620 Rahil Bhimjiani, Raito Bezarius, Raul Cheleguini, Reto Schneider,
621 Richard Maw, Robby Red, RoepLuke, Roland Hieber, Roland Singer,
622 Ronan Pigott, Sam James, Sam Leonard, Sergey A, Susant Sahani,
623 Sven Joachim, Tad Fisher, Takashi Sakamoto, Thorsten Kukuk, Tj,
624 Tomasz Świątek, Topi Miettinen, Valentin David,
625 Valentin Lefebvre, Victor Westerhuis, Vincent Haupert,
626 Vishal Chillara Srinivas, Vito Caputo, Warren, Weblate,
627 Xiaotian Wu, xinpeng wang, Yaron Shahrabani, Yo-Jung Lin,
628 Yu Watanabe, Zbigniew Jędrzejewski-Szmek, zeroskyx,
629 Дамјан Георгиевски, наб
631 — Edinburgh, 2023-12-06
635 Announcements of Future Feature Removals and Incompatible Changes:
637 * The next release (v255) will remove support for split-usr (/usr/
638 mounted separately during late boot, instead of being mounted by the
639 initrd before switching to the rootfs) and unmerged-usr (parallel
640 directories /bin/ and /usr/bin/, /lib/ and /usr/lib/, …). For more
642 https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html
644 * We intend to remove cgroup v1 support from a systemd release after
645 the end of 2023. If you run services that make explicit use of
646 cgroup v1 features (i.e. the "legacy hierarchy" with separate
647 hierarchies for each controller), please implement compatibility with
648 cgroup v2 (i.e. the "unified hierarchy") sooner rather than later.
649 Most of Linux userspace has been ported over already.
651 * Support for System V service scripts is now deprecated and will be
652 removed in a future release. Please make sure to update your software
653 *now* to include a native systemd unit file instead of a legacy
654 System V script to retain compatibility with future systemd releases.
656 * Support for the SystemdOptions EFI variable is deprecated.
657 'bootctl systemd-efi-options' will emit a warning when used. It seems
658 that this feature is little-used and it is better to use alternative
659 approaches like credentials and confexts. The plan is to drop support
660 altogether at a later point, but this might be revisited based on
663 * EnvironmentFile= now treats the line following a comment line
664 trailing with escape as a non comment line. For details, see:
665 https://github.com/systemd/systemd/issues/27975
667 * PrivateNetwork=yes and NetworkNamespacePath= now imply
668 PrivateMounts=yes unless PrivateMounts=no is explicitly specified.
670 * Behaviour of sandboxing options for the per-user service manager
671 units has changed. They now imply PrivateUsers=yes, which means user
672 namespaces will be implicitly enabled when a sandboxing option is
673 enabled in a user unit. Enabling user namespaces has the drawback
674 that system users will no longer be visible (and processes/files will
675 appear as owned by 'nobody') in the user unit.
677 By definition a sandboxed user unit should run with reduced
678 privileges, so impact should be small. This will remove a great
679 source of confusion that has been reported by users over the years,
680 due to how these options require an extra setting to be manually
681 enabled when used in the per-user service manager, which is not
682 needed in the system service manager. For more details, see:
683 https://lists.freedesktop.org/archives/systemd-devel/2022-December/048682.html
685 * systemd-run's switch --expand-environment= which currently is disabled
686 by default when combined with --scope, will be changed in a future
687 release to be enabled by default.
689 Security Relevant Changes:
691 * pam_systemd will now by default pass the CAP_WAKE_ALARM ambient
692 process capability to invoked session processes of regular users on
693 local seats (as well as to systemd --user), unless configured
694 otherwise via data from JSON user records, or via the PAM module's
695 parameter list. This is useful in order allow desktop tools such as
696 GNOME's Alarm Clock application to set a timer for
697 CLOCK_REALTIME_ALARM that wakes up the system when it elapses. A
698 per-user service unit file may thus use AmbientCapability= to pass
699 the capability to invoked processes. Note that this capability is
700 relatively narrow in focus (in particular compared to other process
701 capabilities such as CAP_SYS_ADMIN) and we already — by default —
702 permit more impactful operations such as system suspend to local
707 * Memory limits that apply while the unit is activating are now
708 supported. Previously IO and CPU settings were already supported via
709 StartupCPUWeight= and similar. The same logic has been added for the
710 various manager and unit memory settings (DefaultStartupMemoryLow=,
711 StartupMemoryLow=, StartupMemoryHigh=, StartupMemoryMax=,
712 StartupMemorySwapMax=, StartupMemoryZSwapMax=).
714 * The service manager gained support for enqueuing POSIX signals to
715 services that carry an additional integer value, exposing the
716 sigqueue() system call. This is accessible via new D-Bus calls
717 org.freedesktop.systemd1.Manager.QueueSignalUnit() and
718 org.freedesktop.systemd1.Unit.QueueSignal(), as well as in systemctl
719 via the new --kill-value= option.
721 * systemctl gained a new "list-paths" verb, which shows all currently
722 active .path units, similarly to how "systemctl list-timers" shows
723 active timers, and "systemctl list-sockets" shows active sockets.
725 * systemctl gained a new --when= switch which is honoured by the various
726 forms of shutdown (i.e. reboot, kexec, poweroff, halt) and allows
727 scheduling these operations by time, similar in fashion to how this
728 has been supported by SysV shutdown.
730 * If MemoryDenyWriteExecute= is enabled for a service and the kernel
731 supports the new PR_SET_MDWE prctl() call, it is used instead of the
732 seccomp()-based system call filter to achieve the same effect.
734 * A new set of kernel command line options is now understood:
735 systemd.tty.term.<name>=, systemd.tty.rows.<name>=,
736 systemd.tty.columns.<name>= allow configuring the TTY type and
737 dimensions for the tty specified via <name>. When systemd invokes a
738 service on a tty (via TTYName=) it will look for these and configure
739 the TTY accordingly. This is particularly useful in VM environments
740 to propagate host terminal settings into the appropriate TTYs of the
743 * A new RootEphemeral= setting is now understood in service units. It
744 takes a boolean argument. If enabled for services that use RootImage=
745 or RootDirectory= an ephemeral copy of the disk image or directory
746 tree is made when the service is started. It is removed automatically
747 when the service is stopped. That ephemeral copy is made using
748 btrfs/xfs reflinks or btrfs snapshots, if available.
750 * The service activation logic gained new settings RestartSteps= and
751 RestartMaxDelaySec= which allow exponentially-growing restart
752 intervals for Restart=.
754 * The service activation logic gained a new setting RestartMode= which
755 can be set to 'direct' to skip the inactive/failed states when
756 restarting, so that dependent units are not notified until the service
757 converges to a final (successful or failed) state. For example, this
758 means that OnSuccess=/OnFailure= units will not be triggered until the
759 service state has converged.
761 * PID 1 will now automatically load the virtio_console kernel module
762 during early initialization if running in a suitable VM. This is done
763 so that early-boot logging can be written to the console if available.
765 * Similarly, virtio-vsock support is loaded early in suitable VM
766 environments. PID 1 will send sd_notify() notifications via AF_VSOCK
767 to the VMM if configured, thus loading this early is beneficial.
769 * A new verb "fdstore" has been added to systemd-analyze to show the
770 current contents of the file descriptor store of a unit. This is
771 backed by a new D-Bus call DumpUnitFileDescriptorStore() provided by
774 * The service manager will now set a new $FDSTORE environment variable
775 when invoking processes for services that have the file descriptor
778 * A new service option FileDescriptorStorePreserve= has been added that
779 allows tuning the lifecycle of the per-service file descriptor store.
780 If set to "yes", the entries in the fd store are retained even after
781 the service has been fully stopped.
783 * The "systemctl clean" command may now be used to clear the fdstore of
786 * Unit *.preset files gained a new directive "ignore", in addition to
787 the existing "enable" and "disable". As the name suggests, matching
788 units are left unchanged, i.e. neither enabled nor disabled.
790 * Service units gained a new setting DelegateSubgroup=. It takes the
791 name of a sub-cgroup to place any processes the service manager forks
792 off in. Previously, the service manager would place all service
793 processes directly in the top-level cgroup it created for the
794 service. This usually meant that main process in a service with
795 delegation enabled would first have to create a subgroup and move
796 itself down into it, in order to not conflict with the "no processes
797 in inner cgroups" rule of cgroup v2. With this option, this step is
798 now handled by PID 1.
800 * The service manager will now look for .upholds/ directories,
801 similarly to the existing support for .wants/ and .requires/
802 directories. Symlinks in this directory result in Upholds=
805 The [Install] section of unit files gained support for a new
806 UpheldBy= directive to generate .upholds/ symlinks automatically when
809 * The service manager now supports a new kernel command line option
810 systemd.default_device_timeout_sec=, which may be used to override
811 the default timeout for .device units.
813 * A new "soft-reboot" mechanism has been added to the service manager.
814 A "soft reboot" is similar to a regular reboot, except that it
815 affects userspace only: the service manager shuts down any running
816 services and other units, then optionally switches into a new root
817 file system (mounted to /run/nextroot/), and then passes control to a
818 systemd instance in the new file system which then starts the system
819 up again. The kernel is not rebooted and neither is the hardware,
820 firmware or boot loader. This provides a fast, lightweight mechanism
821 to quickly reset or update userspace, without the latency that a full
822 system reset involves. Moreover, open file descriptors may be passed
823 across the soft reboot into the new system where they will be passed
824 back to the originating services. This allows pinning resources
825 across the reboot, thus minimizing grey-out time further. This new
826 reboot mechanism is accessible via the new "systemctl soft-reboot"
829 * Services using RootDirectory= or RootImage= will now have read-only
830 access to a copy of the host's os-release file under
831 /run/host/os-release, which will be kept up-to-date on 'soft-reboot'.
832 This was already the case for Portable Services, and the feature has
833 now been extended to all services that do not run off the host's
836 * A new service setting MemoryKSM= has been added to enable kernel
837 same-page merging individually for services.
839 * A new service setting ImportCredentials= has been added that augments
840 LoadCredential= and LoadCredentialEncrypted= and searches for
841 credentials to import from the system, and supports globbing.
843 * A new job mode "restart-dependencies" has been added to the service
844 manager (exposed via systemctl --job-mode=). It is only valid when
845 used with "start" jobs, and has the effect that the "start" job will
846 be propagated as "restart" jobs to currently running units that have
847 a BindsTo= or Requires= dependency on the started unit.
849 * A new verb "whoami" has been added to "systemctl" which determines as
850 part of which unit the command is being invoked. It writes the unit
851 name to standard output. If one or more PIDs are specified reports
852 the unit names the processes referenced by the PIDs belong to.
854 * The system and service credential logic has been improved: there's
855 now a clearly defined place where system provisioning tools running
856 in the initrd can place credentials that will be imported into the
857 system's set of credentials during the initrd → host transition: the
858 /run/credentials/@initrd/ directory. Once the credentials placed
859 there are imported into the system credential set they are deleted
860 from this directory, and the directory itself is deleted afterwards
863 * A new kernel command line option systemd.set_credential_binary= has
864 been added, that is similar to the pre-existing
865 systemd.set_credential= but accepts arbitrary binary credential data,
866 encoded in Base64. Note that the kernel command line is not a
867 recommend way to transfer credentials into a system, since it is
868 world-readable from userspace.
870 * The default machine ID to use may now be configured via the
871 system.machine_id system credential. It will only be used if no
872 machine ID was set yet on the host.
874 * On Linux kernel 6.4 and newer system and service credentials will now
875 be placed in a tmpfs instance that has the "noswap" mount option
876 set. Previously, a "ramfs" instance was used. By switching to tmpfs
877 ACL support and overall size limits can now be enforced, without
878 compromising on security, as the memory is never paged out either
881 * The service manager now can detect when it is running in a
882 'Confidential Virtual Machine', and a corresponding 'cvm' value is now
883 accepted by ConditionSecurity= for units that want to conditionalize
884 themselves on this. systemd-detect-virt gained new 'cvm' and
885 '--list-cvm' switches to respectively perform the detection or list
886 all known flavours of confidential VM, depending on the vendor. The
887 manager will publish a 'ConfidentialVirtualization' D-Bus property,
888 and will also set a SYSTEMD_CONFIDENTIAL_VIRTUALIZATION= environment
889 variable for unit generators. Finally, udev rules can match on a new
890 'cvm' key that will be set when in a confidential VM.
891 Additionally, when running in a 'Confidential Virtual Machine', SMBIOS
892 strings and QEMU's fw_cfg protocol will not be used to import
893 credentials and kernel command line parameters by the system manager,
894 systemd-boot and systemd-stub, because the hypervisor is considered
895 untrusted in this particular setting.
899 * The sd-journal API gained a new call sd_journal_get_seqnum() to
900 retrieve the current log record's sequence number and sequence number
901 ID, which allows applications to order records the same way as
902 journal does internally. The sequence number is now also exported in
903 the JSON and "export" output of the journal.
905 * journalctl gained a new switch --truncate-newline. If specified
906 multi-line log records will be truncated at the first newline,
907 i.e. only the first line of each log message will be shown.
909 * systemd-journal-upload gained support for --namespace=, similar to
910 the switch of the same name of journalctl.
914 * systemd-repart's drop-in files gained a new ExcludeFiles= option which
915 may be used to exclude certain files from the effect of CopyFiles=.
917 * systemd-repart's Verity support now implements the Minimize= setting
918 to minimize the size of the resulting partition.
920 * systemd-repart gained a new --offline= switch, which may be used to
921 control whether images shall be built "online" or "offline",
922 i.e. whether to make use of kernel facilities such as loopback block
923 devices and device mapper or not.
925 * If systemd-repart is told to populate a newly created ESP or XBOOTLDR
926 partition with some files, it will now default to VFAT rather than
929 * systemd-repart gained a new --architecture= switch. If specified, the
930 per-architecture GPT partition types (i.e. the root and /usr/
931 partitions) configured in the partition drop-in files are
932 automatically adjusted to match the specified CPU architecture, in
933 order to simplify cross-architecture DDI building.
935 * systemd-repart will now default to a minimum size of 300MB for XFS
936 filesystems if no size parameter is specified. This matches what the
937 XFS tools (xfsprogs) can support.
939 systemd-boot, systemd-stub, ukify, bootctl, kernel-install:
941 * gnu-efi is no longer required to build systemd-boot and systemd-stub.
942 Instead, pyelftools is now needed, and it will be used to perform the
943 ELF -> PE relocations at build time.
945 * bootctl gained a new switch --print-root-device/-R that prints the
946 block device the root file system is backed by. If specified twice,
947 it returns the whole disk block device (as opposed to partition block
948 device) the root file system is on. It's useful for invocations such
949 as "cfdisk $(bootctl -RR)" to quickly show the partition table of the
952 * systemd-stub will now look for the SMBIOS Type 1 field
953 "io.systemd.stub.kernel-cmdline-extra" and append its value to the
954 kernel command line it invokes. This is useful for VMMs such as qemu
955 to pass additional kernel command lines into the system even when
956 booting via full UEFI. The contents of the field are measured into
959 * The KERNEL_INSTALL_LAYOUT= setting for kernel-install gained a new
960 value "auto". With this value, a kernel will be automatically
961 analyzed, and if it qualifies as UKI, it will be installed as if the
962 setting was to set to "uki", otherwise as "bls".
964 * systemd-stub can now optionally load UEFI PE "add-on" images that may
965 contain additional kernel command line information. These "add-ons"
966 superficially look like a regular UEFI executable, and are expected
967 to be signed via SecureBoot/shim. However, they do not actually
968 contain code, but instead a subset of the PE sections that UKIs
969 support. They are supposed to provide a way to extend UKIs with
970 additional resources in a secure and authenticated way. Currently,
971 only the .cmdline PE section may be used in add-ons, in which case
972 any specified string is appended to the command line embedded into
973 the UKI itself. A new 'addon<EFI-ARCH>.efi.stub' is now provided that
974 can be used to trivially create addons, via 'ukify' or 'objcopy'. In
975 the future we expect other sections to be made extensible like this as
978 * ukify has been updated to allow building these UEFI PE "add-on"
979 images, using the new 'addon<EFI-ARCH>.efi.stub'.
981 * ukify now accepts SBAT information to place in the .sbat PE section
982 of UKIs and addons. If a UKI is built the SBAT information from the
983 inner kernel is merged with any SBAT information associated with
984 systemd-stub and the SBAT data specified on the ukify command line.
986 * The kernel-install script has been rewritten in C, and reuses much of
987 the infrastructure of existing tools such as bootctl. It also gained
988 --esp-path= and --boot-path= options to override the path to the ESP,
989 and the $BOOT partition. Options --make-entry-directory= and
990 --entry-token= have been added as well, similar to bootctl's options
993 * A new kernel-install plugin 60-ukify has been added which will
994 combine kernel/initrd locally into a UKI and optionally sign them
995 with a local key. This may be used to switch to UKI mode even on
996 systems where a local kernel or initrd is used. (Typically UKIs are
997 built and signed by the vendor.)
999 * The ukify tool now supports "pesign" in addition to the pre-existing
1000 "sbsign" for signing UKIs.
1002 * systemd-measure and systemd-stub now look for the .uname PE section
1003 that should contain the kernel's "uname -r" string.
1005 * systemd-measure and ukify now calculate expected PCR hashes for a UKI
1006 "offline", i.e. without access to a TPM (physical or
1009 Memory Pressure & Control:
1011 * The sd-event API gained new calls sd_event_add_memory_pressure(),
1012 sd_event_source_set_memory_pressure_type(),
1013 sd_event_source_set_memory_pressure_period() to create and configure
1014 an event source that is called whenever the OS signals memory
1015 pressure. Another call sd_event_trim_memory() is provided that
1016 compacts the process' memory use by releasing allocated but unused
1017 malloc() memory back to the kernel. Services can also provide their
1018 own custom callback to do memory trimming. This should improve system
1019 behaviour under memory pressure, as on Linux traditionally provided
1020 no mechanism to return process memory back to the kernel if the
1021 kernel was under memory pressure. This makes use of the kernel's PSI
1022 interface. Most long-running services in systemd have been hooked up
1023 with this, and in particular systems with low memory should benefit
1026 * Service units gained new settings MemoryPressureWatch= and
1027 MemoryPressureThresholdSec= to configure the PSI memory pressure
1028 logic individually. If these options are used, the
1029 $MEMORY_PRESSURE_WATCH and $MEMORY_PRESSURE_WRITE environment
1030 variables will be set for the invoked processes to inform them about
1031 the requested memory pressure behaviour. (This is used by the
1032 aforementioned sd-events API additions, if set.)
1034 * systemd-analyze gained a new "malloc" verb that shows the output
1035 generated by glibc's malloc_info() on services that support it. Right
1036 now, only the service manager has been updated accordingly. This
1037 call requires privileges.
1039 User & Session Management:
1041 * The sd-login API gained a new call sd_session_get_username() to
1042 return the user name of the owner of a login session. It also gained
1043 a new call sd_session_get_start_time() to retrieve the time the login
1044 session started. A new call sd_session_get_leader() has been added to
1045 return the PID of the "leader" process of a session. A new call
1046 sd_uid_get_login_time() returns the time since the specified user has
1047 most recently been continuously logged in with at least one session.
1049 * JSON user records gained a new set of fields capabilityAmbientSet and
1050 capabilityBoundingSet which contain a list of POSIX capabilities to
1051 set for the logged in users in the ambient and bounding sets,
1052 respectively. homectl gained the ability to configure these two sets
1053 for users via --capability-bounding-set=/--capability-ambient-set=.
1055 * pam_systemd learnt two new module options
1056 default-capability-bounding-set= and default-capability-ambient-set=,
1057 which configure the default bounding sets for users as they are
1058 logging in, if the JSON user record doesn't specify this explicitly
1059 (see above). The built-in default for the ambient set now contains
1060 the CAP_WAKE_ALARM, thus allowing regular users who may log in
1061 locally to resume from a system suspend via a timer.
1063 * The Session D-Bus objects systemd-logind gained a new SetTTY() method
1064 call to update the TTY of a session after it has been allocated. This
1065 is useful for SSH sessions which are typically allocated first, and
1066 for which a TTY is added later.
1068 * The sd-login API gained a new call sd_pid_notifyf_with_fds() which
1069 combines the various other sd_pid_notify() flavours into one: takes a
1070 format string, an overriding PID, and a set of file descriptors to
1071 send. It also gained a new call sd_pid_notify_barrier() call which is
1072 equivalent to sd_notify_barrier() but allows the originating PID to
1075 * "loginctl list-users" and "loginctl list-sessions" will now show the
1076 state of each logged in user/session in their tabular output. It will
1077 also show the current idle state of sessions.
1081 * systemd-dissect will now show the intended CPU architecture of an
1084 * systemd-dissect will now install itself as mount helper for the "ddi"
1085 pseudo-file system type. This means you may now mount DDIs directly
1086 via /bin/mount or /etc/fstab, making full use of embedded Verity
1087 information and all other DDI features.
1089 Example: mount -t ddi myimage.raw /some/where
1091 * The systemd-dissect tool gained the new switches --attach/--detach to
1092 attach/detach a DDI to a loopback block device without mounting it.
1093 It will automatically derive the right sector size from the image
1094 and set up Verity and similar, but not mount the file systems in it.
1096 * When systemd-gpt-auto-generator or the DDI mounting logic mount an
1097 ESP or XBOOTLDR partition the MS_NOSYMFOLLOW mount option is now
1098 implied. Given that these file systems are typically untrusted, this
1099 should make mounting them automatically have less of a security
1102 * All tools that parse DDIs (such as systemd-nspawn, systemd-dissect,
1103 systemd-tmpfiles, …) now understand a new switch --image-policy= which
1104 takes a string encoding image dissection policy. With this mechanism
1105 automatic discovery and use of specific partition types and the
1106 cryptographic requirements on the partitions (Verity, LUKS, …) can be
1107 restricted, permitting better control of the exposed attack surfaces
1108 when mounting disk images. systemd-gpt-auto-generator will honour such
1109 an image policy too, configurable via the systemd.image_policy= kernel
1110 command line option. Unit files gained the RootImagePolicy=,
1111 MountImagePolicy= and ExtensionImagePolicy= to configure the same for
1112 disk images a service runs off.
1114 * systemd-analyze gained a new verb "image-policy" to validate and
1115 parse image policy strings.
1117 * systemd-dissect gained support for a new --validate switch to
1118 superficially validate DDI structure, and check whether a specific
1119 image policy allows the DDI.
1121 * systemd-dissect gained support for a new --mtree-hash switch to
1122 optionally disable calculating mtree hashes, which can be slow on
1125 * systemd-dissect --copy-to, --copy-from, --list and --mtree switches
1126 are now able to operate on directories too, other than images.
1130 * networkd's GENEVE support as gained a new .network option
1131 InheritInnerProtocol=.
1133 * The [Tunnel] section in .netdev files has gained a new setting
1134 IgnoreDontFragment for controlling the IPv4 "DF" flag of datagrams.
1136 * A new global IPv6PrivacyExtensions= setting has been added that
1137 selects the default value of the per-network setting of the same
1140 * The predictable network interface naming logic was extended to
1141 include SR-IOV-R "representor" information in network interface
1142 names. Unfortunately, this feature was not enabled by default and can
1143 only be enabled at compilation time by setting
1144 -Ddefault-net-naming-scheme=v254.
1146 * The DHCPv4 + DHCPv6 + IPv6 RA logic in networkd gained support for
1147 the RFC8910 captive portal option.
1151 * udevadm gained the new "verify" verb for validating udev rules files
1154 * udev gained a new tool "iocost" that can be used to configure QoS IO
1155 cost data based on hwdb information onto suitable block devices. Also
1156 see https://github.com/iocost-benchmark/iocost-benchmarks.
1158 TPM2 Support + Disk Encryption & Authentication:
1160 * systemd-cryptenroll/systemd-cryptsetup will now install a TPM2 SRK
1161 ("Storage Root Key") as first step in the TPM2, and then use that
1162 for binding FDE to, if TPM2 support is used. This matches
1163 recommendations of TCG (see
1164 https://trustedcomputinggroup.org/wp-content/uploads/TCG-TPM-v2.0-Provisioning-Guidance-Published-v1r1.pdf)
1166 * systemd-cryptenroll and other tools that take TPM2 PCR parameters now
1167 understand textual identifiers for these PCRs.
1169 * systemd-veritysetup + /etc/veritytab gained support for a series of
1170 new options: hash-offset=, superblock=, format=, data-block-size=,
1171 hash-block-size=, data-blocks=, salt=, uuid=, hash=, fec-device=,
1172 fec-offset=, fec-roots= to configure various aspects of a Verity
1175 * systemd-cryptsetup + /etc/crypttab gained support for a new
1176 veracrypt-pim= option for setting the Personal Iteration Multiplier
1177 of veracrypt volumes.
1179 * systemd-integritysetup + /etc/integritytab gained support for a new
1180 mode= setting for controlling the dm-integrity mode (journal, bitmap,
1181 direct) for the volume.
1183 * systemd-analyze gained a new verb "pcrs" that shows the known TPM PCR
1184 registers, their symbolic names and current values.
1188 * The ACL support in tmpfiles.d/ has been updated: if an uppercase "X"
1189 access right is specified this is equivalent to "x" but only if the
1190 inode in question already has the executable bit set for at least
1191 some user/group. Otherwise the "x" bit will be turned off.
1193 * tmpfiles.d/'s C line type now understands a new modifier "+": a line
1194 with C+ will result in a "merge" copy, i.e. all files of the source
1195 tree are copied into the target tree, even if that tree already
1196 exists, resulting in a combined tree of files already present in the
1197 target tree and those copied in.
1199 * systemd-tmpfiles gained a new --graceful switch. If specified lines
1200 with unknown users/groups will silently be skipped.
1204 * systemd-notify gained two new options --fd= and --fdname= for sending
1205 arbitrary file descriptors to the service manager (while specifying an
1206 explicit name for it).
1208 * systemd-notify gained a new --exec switch, which makes it execute the
1209 specified command line after sending the requested messages. This is
1210 useful for sending out READY=1 first, and then continuing invocation
1211 without changing process ID, so that the tool can be nicely used
1212 within an ExecStart= line of a unit file that uses Type=notify.
1214 sd-event + sd-bus APIs:
1216 * The sd-event API gained a new call sd_event_source_leave_ratelimit()
1217 which may be used to explicitly end a rate-limit state an event
1218 source might be in, resetting all rate limiting counters.
1220 * When the sd-bus library is used to make connections to AF_UNIX D-Bus
1221 sockets, it will now encode the "description" set via
1222 sd_bus_set_description() into the source socket address. It will also
1223 look for this information when accepting a connection. This is useful
1224 to track individual D-Bus connections on a D-Bus broker for debug
1229 * systemd-resolved gained a new resolved.conf setting
1230 StateRetentionSec= which may be used to retain cached DNS records
1231 even after their nominal TTL, and use them in case upstream DNS
1232 servers cannot be reached. This can be used to make name resolution
1233 more resilient in case of network problems.
1235 * resolvectl gained a new verb "show-cache" to show the current cache
1236 contents of systemd-resolved. This verb communicates with the
1237 systemd-resolved daemon and requires privileges.
1241 * Meson >= 0.60.0 is now required to build systemd.
1243 * The default keymap to apply may now be chosen at build-time via the
1244 new -Ddefault-keymap= meson option.
1246 * Most of systemd's long-running services now have a generic handler of
1247 the SIGRTMIN+18 signal handler which executes various operations
1248 depending on the sigqueue() parameter sent along. For example, values
1249 0x100…0x107 allow changing the maximum log level of such
1250 services. 0x200…0x203 allow changing the log target of such
1251 services. 0x300 make the services trim their memory similarly to the
1252 automatic PSI-triggered action, see above. 0x301 make the services
1253 output their malloc_info() data to the logs.
1255 * machinectl gained new "edit" and "cat" verbs for editing .nspawn
1256 files, inspired by systemctl's verbs of the same name which edit unit
1257 files. Similarly, networkctl gained the same verbs for editing
1258 .network, .netdev, .link files.
1260 * A new syscall filter group "@sandbox" has been added that contains
1261 syscalls for sandboxing system calls such as those for seccomp and
1264 * New documentation has been added:
1266 https://systemd.io/COREDUMP
1267 https://systemd.io/MEMORY_PRESSURE
1270 * systemd-firstboot gained a new --reset option. If specified, the
1271 settings in /etc/ it knows how to initialize are reset.
1273 * systemd-sysext is now a multi-call binary and is also installed under
1274 the systemd-confext alias name (via a symlink). When invoked that way
1275 it will operate on /etc/ instead of /usr/ + /opt/. It thus becomes a
1276 powerful, atomic, secure configuration management of sorts, that
1277 locally can merge configuration from multiple confext configuration
1278 images into a single immutable tree.
1280 * The --network-macvlan=, --network-ipvlan=, --network-interface=
1281 switches of systemd-nspawn may now optionally take the intended
1282 network interface inside the container.
1284 * All our programs will now send an sd_notify() message with their exit
1285 status in the EXIT_STATUS= field when exiting, using the usual
1286 protocol, including PID 1. This is useful for VMMs and container
1287 managers to collect an exit status from a system as it shuts down, as
1288 set via "systemctl exit …". This is particularly useful in test cases
1289 and similar, as invocations via a VM can now nicely propagate an exit
1290 status to the host, similar to local processes.
1292 * systemd-run gained a new switch --expand-environment=no to disable
1293 server-side environment variable expansion in specified command
1294 lines. Expansion defaults to enabled for all execution types except
1295 --scope, where it defaults to off (and prints a warning) for backward
1296 compatibility reasons. --scope will be flipped to enabled by default
1297 too in a future release. If you are using --scope and passing a '$'
1298 character in the payload you should start explicitly using
1299 --expand-environment=yes/no according to the use case.
1301 * The systemd-system-update-generator has been updated to also look for
1302 the special flag file /etc/system-update in addition to the existing
1303 support for /system-update to decide whether to enter system update
1306 * The /dev/hugepages/ file system is now mounted with nosuid + nodev
1307 mount options by default.
1309 * systemd-fstab-generator now understands two new kernel command line
1310 options systemd.mount-extra= and systemd.swap-extra=, which configure
1311 additional mounts or swaps in a format similar to /etc/fstab. 'fsck'
1312 will be ran on these block devices, like it already happens for
1313 'root='. It also now supports the new fstab.extra and
1314 fstab.extra.initrd credentials that may contain additional /etc/fstab
1315 lines to apply at boot.
1317 * systemd-getty-generator now understands two new credentials
1318 getty.ttys.container and getty.ttys.serial. These credentials may
1319 contain a list of TTY devices – one per line – to instantiate
1320 container-getty@.service and serial-getty@.service on.
1322 * The getty/serial-getty/container-getty units now import the 'agetty.*'
1323 and 'login.*' credentials, which are consumed by the 'login' and
1324 'agetty' programs starting from util-linux v2.40.
1326 * systemd-sysupdate's sysupdate.d/ drop-ins gained a new setting
1327 PathRelativeTo=, which can be set to "esp", "xbootldr", "boot", in
1328 which case the Path= setting is taken relative to the ESP or XBOOTLDR
1329 partitions, rather than the system's root directory /. The relevant
1330 directories are automatically discovered.
1332 * The systemd-ac-power tool gained a new switch --low, which reports
1333 whether the battery charge is considered "low", similar to how the
1334 s2h suspend logic checks this state to decide whether to enter system
1335 suspend or hibernation.
1337 * The /etc/os-release file can now have two new optional fields
1338 VENDOR_NAME= and VENDOR_URL= to carry information about the vendor of
1341 * When the system hibernates, information about the device and offset
1342 used is now written to a non-volatile EFI variable. On next boot the
1343 system will attempt to resume from the location indicated in this EFI
1344 variable. This should make hibernation a lot more robust, while
1345 requiring no manual configuration of the resume location.
1347 * The $XDG_STATE_HOME environment variable (added in more recent
1348 versions of the XDG basedir specification) is now honoured to
1349 implement the StateDirectory= setting in user services.
1351 * A new component "systemd-battery-check" has been added. It may run
1352 during early boot (usually in the initrd), and checks the battery
1353 charge level of the system. In case the charge level is very low the
1354 user is notified (graphically via Plymouth – if available – as well
1355 as in text form on the console), and the system is turned off after a
1356 10s delay. The feature can be disabled by passing
1357 systemd.battery-check=0 through the kernel command line.
1359 * The 'passwdqc' library is now supported as an alternative to the
1360 'pwquality' library and can be selected at build time.
1362 Contributions from: 김인수, 07416, Addison Snelling, Adrian Vovk,
1363 Aidan Dang, Alexander Krabler, Alfred Klomp, Anatoli Babenia,
1364 Andrei Stepanov, Andrew Baxter, Antonio Alvarez Feijoo,
1365 Arian van Putten, Arthur Shau, A S Alam,
1366 Asier Sarasua Garmendia, Balló György, Bastien Nocera,
1367 Benjamin Herrenschmidt, Benjamin Raison, Bill Peterson,
1368 Brad Fitzpatrick, Brett Holman, bri, Chen Qi, Chitoku,
1369 Christian Hesse, Christoph Anton Mitterer, Christopher Gurnee,
1370 Colin Walters, Cornelius Hoffmann, Cristian Rodríguez, cunshunxia,
1371 cvlc12, Cyril Roelandt, Daan De Meyer, Daniele Medri,
1372 Daniel P. Berrangé, Daniel Rusek, Dan Streetman, David Edmundson,
1373 David Schroeder, David Tardon, dependabot[bot],
1374 Dimitri John Ledkov, Dmitrii Fomchenkov, Dmitry V. Levin, dmkUK,
1375 Dominique Martinet, don bright, drosdeck, Edson Juliano Drosdeck,
1376 Egor Ignatov, EinBaum, Emanuele Giuseppe Esposito, Eric Curtin,
1377 Erik Sjölund, Evgeny Vereshchagin, Florian Klink, Franck Bui,
1378 François Rigault, Fran Diéguez, Franklin Yu, Frantisek Sumsal,
1379 Fuminobu TAKEYAMA, Gaël PORTAY, Gerd Hoffmann, Gertalitec,
1380 Gibeom Gwon, Gustavo Noronha Silva, Hannu Lounento,
1381 Hans de Goede, Haochen Tong, HATAYAMA Daisuke, Henrik Holst,
1382 Hoe Hao Cheng, Igor Tsiglyar, Ivan Vecera, James Hilliard,
1383 Jan Engelhardt, Jan Janssen, Jan Luebbe, Jan Macku, Janne Sirén,
1384 jcg, Jeidnx, Joan Bruguera, Joerg Behrmann, jonathanmetzman,
1385 Jordan Rome, Josef Miegl, Joshua Goins, Joyce, Joyce Brum,
1386 Juno Computers, Kai Lueke, Kevin P. Fleming, Kiran Vemula, Klaus,
1387 Klaus Zipfel, Lawrence Thorpe, Lennart Poettering, licunlong,
1388 Lily Foster, Luca Boccassi, Ludwig Nussel, Luna Jernberg,
1389 maanyagoenka, Maanya Goenka, Maksim Kliazovich, Malte Poll,
1390 Marko Korhonen, Masatake YAMATO, Mateusz Poliwczak, Matt Johnston,
1391 Miao Wang, Micah Abbott, Michael A Cassaniti, Michal Koutný,
1392 Michal Sekletár, Mike Yuan, mooo, Morten Linderud, msizanoen,
1393 Nick Rosbrook, nikstur, Olivier Gayot, Omojola Joshua,
1394 Paolo Velati, Paul Barker, Pavel Borecki, Petr Menšík,
1395 Philipp Kern, Philip Withnall, Piotr Drąg, Quintin Hill,
1396 Rene Hollander, Richard Phibel, Robert Meijers, Robert Scheck,
1397 Roger Gammans, Romain Geissler, Ronan Pigott, Russell Harmon,
1398 saikat0511, Samanta Navarro, Sam James, Sam Morris,
1399 Simon Braunschmidt, Sjoerd Simons, Sorah Fukumori,
1400 Stanislaw Gruszka, Stefan Roesch, Steven Luo, Steve Ramage,
1401 Susant Sahani, taniishkaaa, Tanishka, Temuri Doghonadze,
1402 Thierry Martin, Thomas Blume, Thomas Genty, Thomas Weißschuh,
1403 Thorsten Kukuk, Times-Z, Tobias Powalowski, tofylion,
1404 Topi Miettinen, Uwe Kleine-König, Velislav Ivanov,
1405 Vitaly Kuznetsov, Vít Zikmund, Weblate, Will Fancher,
1406 William Roberts, Winterhuman, Wolfgang Müller, Xeonacid,
1407 Xiaotian Wu, Xi Ruoyao, Yuri Chornoivan, Yu Watanabe, Yuxiang Zhu,
1408 Zbigniew Jędrzejewski-Szmek, zhmylove, ZjYwMj,
1409 Дамјан Георгиевски, наб
1411 — Edinburgh, 2023-07-28
1415 Announcements of Future Feature Removals and Incompatible Changes:
1417 * We intend to remove cgroup v1 support from systemd release after the
1418 end of 2023. If you run services that make explicit use of cgroup v1
1419 features (i.e. the "legacy hierarchy" with separate hierarchies for
1420 each controller), please implement compatibility with cgroup v2 (i.e.
1421 the "unified hierarchy") sooner rather than later. Most of Linux
1422 userspace has been ported over already.
1424 * We intend to remove support for split-usr (/usr mounted separately
1425 during boot) and unmerged-usr (parallel directories /bin and
1426 /usr/bin, /lib and /usr/lib, etc). This will happen in the second
1427 half of 2023, in the first release that falls into that time window.
1428 For more details, see:
1429 https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html
1431 * We intend to change behaviour w.r.t. units of the per-user service
1432 manager and sandboxing options, so that they work without having to
1433 manually enable PrivateUsers= as well, which is not required for
1434 system units. To make this work, we will implicitly enable user
1435 namespaces (PrivateUsers=yes) when a sandboxing option is enabled in a
1436 user unit. The drawback is that system users will no longer be visible
1437 (and appear as 'nobody') to the user unit when a sandboxing option is
1438 enabled. By definition a sandboxed user unit should run with reduced
1439 privileges, so impact should be small. This will remove a great source
1440 of confusion that has been reported by users over the years, due to
1441 how these options require an extra setting to be manually enabled when
1442 used in the per-user service manager, as opposed as to the system
1443 service manager. We plan to enable this change in the next release
1444 later this year. For more details, see:
1445 https://lists.freedesktop.org/archives/systemd-devel/2022-December/048682.html
1447 Deprecations and incompatible changes:
1449 * systemctl will now warn when invoked without /proc/ mounted
1450 (e.g. when invoked after chroot() into an directory tree without the
1451 API mount points like /proc/ being set up.) Operation in such an
1452 environment is not fully supported.
1454 * The return value of 'systemctl is-active|is-enabled|is-failed' for
1455 unknown units is changed: previously 1 or 3 were returned, but now 4
1456 (EXIT_PROGRAM_OR_SERVICES_STATUS_UNKNOWN) is used as documented.
1458 * 'udevadm hwdb' subcommand is deprecated and will emit a warning.
1459 systemd-hwdb (added in 2014) should be used instead.
1461 * 'bootctl --json' now outputs a single JSON array, instead of a stream
1462 of newline-separated JSON objects.
1464 * Udev rules in 60-evdev.rules have been changed to load hwdb
1465 properties for all modalias patterns. Previously only the first
1466 matching pattern was used. This could change what properties are
1467 assigned if the user has more and less specific patterns that could
1468 match the same device, but it is expected that the change will have
1469 no effect for most users.
1471 * systemd-networkd-wait-online exits successfully when all interfaces
1472 are ready or unmanaged. Previously, if neither '--any' nor
1473 '--interface=' options were used, at least one interface had to be in
1474 configured state. This change allows the case where systemd-networkd
1475 is enabled, but no interfaces are configured, to be handled
1476 gracefully. It may occur in particular when a different network
1477 manager is also enabled and used.
1479 * Some compatibility helpers were dropped: EmergencyAction= in the user
1480 manager, as well as measuring kernel command line into PCR 8 in
1481 systemd-stub, along with the -Defi-tpm-pcr-compat compile-time
1484 * The '-Dupdate-helper-user-timeout=' build-time option has been
1485 renamed to '-Dupdate-helper-user-timeout-sec=', and now takes an
1486 integer as parameter instead of a string.
1488 * The DDI image dissection logic (which backs RootImage= in service
1489 unit files, the --image= switch in various tools such as
1490 systemd-nspawn, as well as systemd-dissect) will now only mount file
1491 systems of types btrfs, ext4, xfs, erofs, squashfs, vfat. This list
1492 can be overridden via the $SYSTEMD_DISSECT_FILE_SYSTEMS environment
1493 variable. These file systems are fairly well supported and maintained
1494 in current kernels, while others are usually more niche, exotic or
1495 legacy and thus typically do not receive the same level of security
1498 * The default per-link multicast DNS mode is changed to "yes"
1499 (that was previously "no"). As the default global multicast DNS mode
1500 has been "yes" (but can be changed by the build option), now the
1501 multicast DNS is enabled on all links by default. You can disable the
1502 multicast DNS on all links by setting MulticastDNS= in resolved.conf,
1503 or on an interface by calling "resolvectl mdns INTERFACE no".
1507 * A tool 'ukify' tool to build, measure, and sign Unified Kernel Images
1508 (UKIs) has been added. This replaces functionality provided by
1509 'dracut --uefi' and extends it with automatic calculation of PE file
1510 offsets, insertion of signed PCR policies generated by
1511 systemd-measure, support for initrd concatenation, signing of the
1512 embedded Linux image and the combined image with sbsign, and
1513 heuristics to autodetect the kernel uname and verify the splash
1516 Changes in systemd and units:
1518 * A new service type Type=notify-reload is defined. When such a unit is
1519 reloaded a UNIX process signal (typically SIGHUP) is sent to the main
1520 service process. The manager will then wait until it receives a
1521 "RELOADING=1" followed by a "READY=1" notification from the unit as
1522 response (via sd_notify()). Otherwise, this type is the same as
1523 Type=notify. A new setting ReloadSignal= may be used to change the
1524 signal to send from the default of SIGHUP.
1526 user@.service, systemd-networkd.service, systemd-udevd.service, and
1527 systemd-logind have been updated to this type.
1529 * Initrd environments which are not on a pure memory file system (e.g.
1530 overlayfs combination as opposed to tmpfs) are now supported. With
1531 this change, during the initrd → host transition ("switch root")
1532 systemd will erase all files of the initrd only when the initrd is
1533 backed by a memory file system such as tmpfs.
1535 * New per-unit MemoryZSwapMax= option has been added to configure
1536 memory.zswap.max cgroup properties (the maximum amount of zswap
1539 * A new LogFilterPatterns= option has been added for units. It may be
1540 used to specify accept/deny regular expressions for log messages
1541 generated by the unit, that shall be enforced by systemd-journald.
1542 Rejected messages are neither stored in the journal nor forwarded.
1543 This option may be used to suppress noisy or uninteresting messages
1546 * The manager has a new
1547 org.freedesktop.systemd1.Manager.GetUnitByPIDFD() D-Bus method to
1548 query process ownership via a PIDFD, which is more resilient against
1549 PID recycling issues.
1551 * Scope units now support OOMPolicy=. Login session scopes default to
1552 OOMPolicy=continue, allowing login scopes to survive the OOM killer
1553 terminating some processes in the scope.
1555 * systemd-fstab-generator now supports x-systemd.makefs option for
1556 /sysroot/ (in the initrd).
1558 * The maximum rate at which daemon reloads are executed can now be
1559 limited with the new ReloadLimitIntervalSec=/ReloadLimitBurst=
1560 options. (Or the equivalent on the kernel command line:
1561 systemd.reload_limit_interval_sec=/systemd.reload_limit_burst=). In
1562 addition, systemd now logs the originating unit and PID when a reload
1563 request is received over D-Bus.
1565 * When enabling a swap device systemd will now reinitialize the device
1566 when the page size of the swap space does not match the page size of
1567 the running kernel. Note that this requires the 'swapon' utility to
1568 provide the '--fixpgsz' option, as implemented by util-linux, and it
1569 is not supported by busybox at the time of writing.
1571 * systemd now executes generator programs in a mount namespace
1572 "sandbox" with most of the file system read-only and write access
1573 restricted to the output directories, and with a temporary /tmp/
1574 mount provided. This provides a safeguard against programming errors
1575 in the generators, but also fixes here-docs in shells, which
1576 previously didn't work in early boot when /tmp/ wasn't available
1577 yet. (This feature has no security implications, because the code is
1578 still privileged and can trivially exit the sandbox.)
1580 * The system manager will now parse a new "vmm.notify_socket"
1581 system credential, which may be supplied to a VM via SMBIOS. If
1582 found, the manager will send a "READY=1" notification on the
1583 specified socket after boot is complete. This allows readiness
1584 notification to be sent from a VM guest to the VM host over a VSOCK
1587 * The sample PAM configuration file for systemd-user@.service now
1588 includes a call to pam_namespace. This puts children of user@.service
1589 in the expected namespace. (Many distributions replace their file
1590 with something custom, so this change has limited effect.)
1592 * A new environment variable $SYSTEMD_DEFAULT_MOUNT_RATE_LIMIT_BURST
1593 can be used to override the mount units burst late limit for
1594 parsing '/proc/self/mountinfo', which was introduced in v249.
1597 * Drop-ins for init.scope changing control group resource limits are
1598 now applied, while they were previously ignored.
1600 * New build-time configuration options '-Ddefault-timeout-sec=' and
1601 '-Ddefault-user-timeout-sec=' have been added, to let distributions
1602 choose the default timeout for starting/stopping/aborting system and
1603 user units respectively.
1605 * Service units gained a new setting OpenFile= which may be used to
1606 open arbitrary files in the file system (or connect to arbitrary
1607 AF_UNIX sockets in the file system), and pass the open file
1608 descriptor to the invoked process via the usual file descriptor
1609 passing protocol. This is useful to give unprivileged services access
1610 to select files which have restrictive access modes that would
1611 normally not allow this. It's also useful in case RootDirectory= or
1612 RootImage= is used to allow access to files from the host environment
1613 (which is after all not visible from the service if these two options
1618 * The new net naming scheme "v253" has been introduced. In the new
1619 scheme, ID_NET_NAME_PATH is also set for USB devices not connected via
1620 a PCI bus. This extends the coverage of predictable interface names
1621 in some embedded systems.
1623 The "amba" bus path is now included in ID_NET_NAME_PATH, resulting in
1624 a more informative path on some embedded systems.
1626 * Partition block devices will now also get symlinks in
1627 /dev/disk/by-diskseq/<seq>-part<n>, which may be used to reference
1628 block device nodes via the kernel's "diskseq" value. Previously those
1629 symlinks were only created for the main block device.
1631 * A new operator '-=' is supported for SYMLINK variables. This allows
1632 symlinks to be unconfigured even if an earlier rule added them.
1634 * 'udevadm --trigger --settle' now also works for network devices
1635 that are being renamed.
1637 Changes in sd-boot, bootctl, and the Boot Loader Specification:
1639 * systemd-boot now passes its random seed directly to the kernel's RNG
1640 via the LINUX_EFI_RANDOM_SEED_TABLE_GUID configuration table, which
1641 means the RNG gets seeded very early in boot before userspace has
1644 * systemd-boot will pass a disk-backed random seed – even when secure
1645 boot is enabled – if it can additionally get a random seed from EFI
1646 itself (via EFI's RNG protocol), or a prior seed in
1647 LINUX_EFI_RANDOM_SEED_TABLE_GUID from a preceding bootloader.
1649 * systemd-boot-system-token.service was renamed to
1650 systemd-boot-random-seed.service and extended to always save a random
1651 seed to ESP on every boot when a compatible boot loader is used. This
1652 allows a refreshed random seed to be used in the boot loader.
1654 * systemd-boot handles various seed inputs using a domain- and
1655 field-separated hashing scheme.
1657 * systemd-boot's 'random-seed-mode' option has been removed. A system
1658 token is now always required to be present for random seeds to be
1661 * systemd-boot now supports being loaded from other locations than the
1662 ESP, for example for direct kernel boot under QEMU or when embedded
1665 * systemd-boot now parses SMBIOS information to detect
1666 virtualization. This information is used to skip some warnings which
1667 are not useful in a VM and to conditionalize other aspects of
1670 * systemd-boot now supports a new 'if-safe' mode that will perform UEFI
1671 Secure Boot automated certificate enrollment from the ESP only if it
1672 is considered 'safe' to do so. At the moment 'safe' means running in
1675 * systemd-stub now processes random seeds in the same way as
1676 systemd-boot already does, in case a unified kernel image is being
1677 used from a different bootloader than systemd-boot, or without any
1680 * bootctl will now generate a system token on all EFI systems, even
1681 virtualized ones, and is activated in the case that the system token
1682 is missing from either sd-boot and sd-stub booted systems.
1684 * bootctl now implements two new verbs: 'kernel-identify' prints the
1685 type of a kernel image file, and 'kernel-inspect' provides
1686 information about the embedded command line and kernel version of
1689 * bootctl now honours $KERNEL_INSTALL_CONF_ROOT with the same meaning
1690 as for kernel-install.
1692 * The JSON output of "bootctl list" will now contain two more fields:
1693 isDefault and isSelected are boolean fields set to true on the
1694 default and currently booted boot menu entries.
1696 * bootctl gained a new verb "unlink" for removing a boot loader entry
1697 type #1 file from disk in a safe and robust way.
1699 * bootctl also gained a new verb "cleanup" that automatically removes
1700 all files from the ESP's and XBOOTLDR's "entry-token" directory, that
1701 is not referenced anymore by any installed Type #1 boot loader
1702 specification entry. This is particularly useful in environments where
1703 a large number of entries reference the same or partly the same
1704 resources (for example, for snapshot-based setups).
1706 Changes in kernel-install:
1708 * A new "installation layout" can be configured as layout=uki. With
1709 this setting, a Boot Loader Specification Type#1 entry will not be
1710 created. Instead, a new kernel-install plugin 90-uki-copy.install
1711 will copy any .efi files from the staging area into the boot
1712 partition. A plugin to generate the UKI .efi file must be provided
1715 Changes in systemctl:
1717 * 'systemctl reboot' has dropped support for accepting a positional
1718 argument as the argument to the reboot(2) syscall. Please use the
1719 --reboot-argument= option instead.
1721 * 'systemctl disable' will now warn when called on units without
1722 install information. A new --no-warn option has been added that
1723 silences this warning.
1725 * New option '--drop-in=' can be used to tell 'systemctl edit' the name
1726 of the drop-in to edit. (Previously, 'override.conf' was always
1729 * 'systemctl list-dependencies' now respects --type= and --state=.
1731 * 'systemctl kexec' now supports XEN VMM environments.
1733 * 'systemctl edit' will now tell the invoked editor to jump into the
1734 first line with actual unit file data, skipping over synthesized
1737 Changes in systemd-networkd and related tools:
1739 * The [DHCPv4] section in .network file gained new SocketPriority=
1740 setting that assigns the Linux socket priority used by the DHCPv4 raw
1741 socket. This may be used in conjunction with the
1742 EgressQOSMaps=setting in [VLAN] section of .netdev file to send the
1743 desired ethernet 802.1Q frame priority for DHCPv4 initial
1744 packets. This cannot be achieved with netfilter mangle tables because
1745 of the raw socket bypass.
1747 * The [DHCPv4] and [IPv6AcceptRA] sections in .network file gained a
1748 new QuickAck= boolean setting that enables the TCP quick ACK mode for
1749 the routes configured by the acquired DHCPv4 lease or received router
1750 advertisements (RAs).
1752 * The RouteMetric= option (for DHCPv4, DHCPv6, and IPv6 advertised
1753 routes) now accepts three values, for high, medium, and low preference
1754 of the router (which can be set with the RouterPreference=) setting.
1756 * systemd-networkd-wait-online now supports matching via alternative
1759 * The [DHCPv6] section in .network file gained new SendRelease=
1760 setting which enables the DHCPv6 client to send release when
1761 it stops. This is the analog of the [DHCPv4] SendRelease= setting.
1762 It is enabled by default.
1764 * If the Address= setting in [Network] or [Address] sections in .network
1765 specified without its prefix length, then now systemd-networkd assumes
1766 /32 for IPv4 or /128 for IPv6 addresses.
1768 * networkctl shows network and link file dropins in status output.
1770 Changes in systemd-dissect:
1772 * systemd-dissect gained a new option --list, to print the paths of
1773 all files and directories in a DDI.
1775 * systemd-dissect gained a new option --mtree, to generate a file
1776 manifest compatible with BSD mtree(5) of a DDI
1778 * systemd-dissect gained a new option --with, to execute a command with
1779 the specified DDI temporarily mounted and used as working
1780 directory. This is for example useful to convert a DDI to "tar"
1781 simply by running it within a "systemd-dissect --with" invocation.
1783 * systemd-dissect gained a new option --discover, to search for
1784 Discoverable Disk Images (DDIs) in well-known directories of the
1785 system. This will list machine, portable service and system extension
1788 * systemd-dissect now understands 2nd stage initrd images stored as a
1789 Discoverable Disk Image (DDI).
1791 * systemd-dissect will now display the main UUID of GPT DDIs (i.e. the
1792 disk UUID stored in the GPT header) among the other data it can show.
1794 * systemd-dissect gained a new --in-memory switch to operate on an
1795 in-memory copy of the specified DDI file. This is useful to access a
1796 DDI with write access without persisting any changes. It's also
1797 useful for accessing a DDI without keeping the originating file
1800 * The DDI dissection logic will now automatically detect the intended
1801 sector size of disk images stored in files, based on the GPT
1802 partition table arrangement. Loopback block devices for such DDIs
1803 will then be configured automatically for the right sector size. This
1804 is useful to make dealing with modern 4K sector size DDIs fully
1805 automatic. The systemd-dissect tool will now show the detected sector
1806 size among the other DDI information in its output.
1808 Changes in systemd-repart:
1810 * systemd-repart gained new options --include-partitions= and
1811 --exclude-partitions= to filter operation on partitions by type UUID.
1812 This allows systemd-repart to be used to build images in which the
1813 type of one partition is set based on the contents of another
1814 partition (for example when the boot partition shall include a verity
1815 hash of the root partition).
1817 * systemd-repart also gained a --defer-partitions= option that is
1818 similar to --exclude-partitions=, but the size of the partition is
1819 still taken into account when sizing partitions, but without
1822 * systemd-repart gained a new --sector-size= option to specify what
1823 sector size should be used when an image is created.
1825 * systemd-repart now supports generating erofs file systems via
1826 CopyFiles= (a read-only file system similar to squashfs).
1828 * The Minimize= option was extended to accept "best" (which means the
1829 most minimal image possible, but may require multiple attempts) and
1830 "guess" (which means a reasonably small image).
1832 * The systemd-growfs binary now comes with a regular unit file template
1833 systemd-growfs@.service which can be instantiated directly for any
1834 desired file system. (Previously, the unit was generated dynamically
1835 by various generators, but no regular unit file template was
1838 Changes in journal tools:
1840 * Various systemd tools will append extra fields to log messages when
1841 in debug mode, or when SYSTEMD_ENABLE_LOG_CONTEXT=1 is set. Currently
1842 this includes information about D-Bus messages when sd-bus is used,
1843 e.g. DBUS_SENDER=, DBUS_DESTINATION=, and DBUS_PATH=, and information
1844 about devices when sd-device is used, e.g. DEVNAME= and DRIVER=.
1845 Details of what is logged and when are subject to change.
1847 * The systemd-journald-audit.socket can now be disabled via the usual
1848 "systemctl disable" mechanism to stop collection of audit
1849 messages. Please note that it is not enabled statically anymore and
1850 must be handled by the preset/enablement logic in package
1851 installation scripts.
1853 * New options MaxUse=, KeepFree=, MaxFileSize=, and MaxFiles= can
1854 be used to curtail disk use by systemd-journal-remote. This is
1855 similar to the options supported by systemd-journald.
1857 Changes in systemd-cryptenroll, systemd-cryptsetup, and related
1860 * When enrolling new keys systemd-cryptenroll now supports unlocking
1861 via FIDO2 tokens (option --unlock-fido2-device=). Previously, a
1862 password was strictly required to be specified.
1864 * systemd-cryptsetup now supports pre-flight requests for FIDO2 tokens
1865 (except for tokens with user verification, UV) to identify tokens
1866 before authentication. Multiple FIDO2 tokens can now be enrolled at
1867 the same time, and systemd-cryptsetup will automatically select one
1868 that corresponds to one of the available LUKS key slots.
1870 * systemd-cryptsetup now supports new options tpm2-measure-bank= and
1871 tpm2-measure-pcr= in crypttab(5). These allow specifying the TPM2 PCR
1872 bank and number into which the volume key should be measured. This is
1873 automatically enabled for the encrypted root volume discovered and
1874 activated by systemd-gpt-auto-generator.
1876 * systemd-gpt-auto-generator mounts the ESP and XBOOTLDR partitions with
1877 "noexec,nosuid,nodev".
1879 * systemd-gpt-auto-generator will now honour the rootfstype= and
1880 rootflags= kernel command line switches for root file systems it
1881 discovers, to match behaviour in case an explicit root fs is
1882 specified via root=.
1884 * systemd-pcrphase gained new options --machine-id and --file-system=
1885 to measure the machine-id and mount point information into PCR 15.
1886 New service unit files systemd-pcrmachine.service and
1887 systemd-pcrfs@.service have been added that invoke the tool with
1888 these switches during early boot.
1890 * systemd-pcrphase gained a --graceful switch will make it exit cleanly
1891 with a success exit code even if no TPM device is detected.
1893 * systemd-cryptenroll now stores the user-supplied PIN with a salt,
1894 making it harder to brute-force.
1896 Changes in other tools:
1898 * systemd-homed gained support for luksPbkdfForceIterations (the
1899 intended number of iterations for the PBKDF operation on LUKS).
1901 * Environment variables $SYSTEMD_HOME_MKFS_OPTIONS_BTRFS,
1902 $SYSTEMD_HOME_MKFS_OPTIONS_EXT4, and $SYSTEMD_HOME_MKFS_OPTIONS_XFS
1903 may now be used to specify additional arguments for mkfs when
1904 systemd-homed formats a file system.
1906 * systemd-hostnamed now exports the contents of
1907 /sys/class/dmi/id/bios_vendor and /sys/class/dmi/id/bios_date via two
1908 new D-Bus properties: FirmwareVendor and FirmwareDate. This allows
1909 unprivileged code to access those values.
1911 systemd-hostnamed also exports the SUPPORT_END= field from
1912 os-release(5) as OperatingSystemSupportEnd. hostnamectl make uses of
1913 this to show the status of the installed system.
1915 * systemd-measure gained an --append= option to sign multiple phase
1916 paths with different signing keys. This allows secrets to be
1917 accessible only in certain parts of the boot sequence. Note that
1918 'ukify' provides similar functionality in a more accessible form.
1920 * systemd-timesyncd will now write a structured log message with
1921 MESSAGE_ID set to SD_MESSAGE_TIME_BUMP when it bumps the clock based
1922 on a on-disk timestamp, similarly to what it did when reaching
1923 synchronization via NTP.
1925 * systemd-timesyncd will now update the on-disk timestamp file on each
1926 boot at least once, making it more likely that the system time
1927 increases in subsequent boots.
1929 * systemd-vconsole-setup gained support for system/service credentials:
1930 vconsole.keymap/vconsole.keymap_toggle and
1931 vconsole.font/vconsole.font_map/vconsole.font_unimap are analogous
1932 the similarly-named options in vconsole.conf.
1934 * systemd-localed will now save the XKB keyboard configuration to
1935 /etc/vconsole.conf, and also read it from there with a higher
1936 preference than the /etc/X11/xorg.conf.d/00-keyboard.conf config
1937 file. Previously, this information was stored in the former file in
1938 converted form, and only in latter file in the original form. Tools
1939 which want to access keyboard configuration can now do so from a
1942 * systemd-resolved gained support for configuring the nameservers and
1943 search domains via kernel command line (nameserver=, domain=) and
1944 credentials (network.dns, network.search_domains).
1946 * systemd-resolved will now synthesize host names for the DNS stub
1947 addresses it supports. Specifically when "_localdnsstub" is resolved,
1948 127.0.0.53 is returned, and if "_localdnsproxy" is resolved
1949 127.0.0.54 is returned.
1951 * systemd-notify will now send a "RELOADING=1" notification when called
1952 with --reloading, and "STOPPING=1" when called with --stopping. This
1953 can be used to implement notifications from units where it's easier
1954 to call a program than to use the sd-daemon library.
1956 * systemd-analyze's 'plot' command can now output its information in
1957 JSON, controlled via the --json= switch. Also, new --table, and
1958 --no-legend options have been added.
1960 * 'machinectl enable' will now automatically enable machines.target
1961 unit in addition to adding the machine unit to the target.
1963 Similarly, 'machinectl start|stop' gained a --now option to enable or
1964 disable the machine unit when starting or stopping it.
1966 * systemd-sysusers will now create /etc/ if it is missing.
1968 * systemd-sleep 'HibernateDelaySec=' setting is changed back to
1969 pre-v252's behaviour, and a new 'SuspendEstimationSec=' setting is
1970 added to provide the new initial value for the new automated battery
1971 estimation functionality. If 'HibernateDelaySec=' is set to any value,
1972 the automated estimate (and thus the automated hibernation on low
1973 battery to avoid data loss) functionality will be disabled.
1975 * Default tmpfiles.d/ configuration will now automatically create
1976 credentials storage directory '/etc/credstore/' with the appropriate,
1977 secure permissions. If '/run/credstore/' exists, its permissions will
1978 be fixed too in case they are not correct.
1980 Changes in libsystemd and shared code:
1982 * sd-bus gained new convenience functions sd_bus_emit_signal_to(),
1983 sd_bus_emit_signal_tov(), and sd_bus_message_new_signal_to().
1985 * sd-id128 functions now return -EUCLEAN (instead of -EIO) when the
1986 128-bit ID in files such as /etc/machine-id has an invalid
1987 format. They also accept NULL as output parameter in more places,
1988 which is useful when the caller only wants to validate the inputs and
1989 does not need the output value.
1991 * sd-login gained new functions sd_pidfd_get_session(),
1992 sd_pidfd_get_owner_uid(), sd_pidfd_get_unit(),
1993 sd_pidfd_get_user_unit(), sd_pidfd_get_slice(),
1994 sd_pidfd_get_user_slice(), sd_pidfd_get_machine_name(), and
1995 sd_pidfd_get_cgroup(), that are analogous to sd_pid_get_*(),
1996 but accept a PIDFD instead of a PID.
1998 * sd-path (and systemd-path) now export four new paths:
1999 SD_PATH_SYSTEMD_SYSTEM_ENVIRONMENT_GENERATOR,
2000 SD_PATH_SYSTEMD_USER_ENVIRONMENT_GENERATOR,
2001 SD_PATH_SYSTEMD_SEARCH_SYSTEM_ENVIRONMENT_GENERATOR, and
2002 SD_PATH_SYSTEMD_SEARCH_USER_ENVIRONMENT_GENERATOR,
2004 * sd_notify() now supports AF_VSOCK as transport for notification
2005 messages (in addition to the existing AF_UNIX support). This is
2006 enabled if $NOTIFY_SOCKET is set in a "vsock:CID:port" format.
2008 * Detection of chroot() environments now works if /proc/ is not
2009 mounted. This affects systemd-detect-virt --chroot, but also means
2010 that systemd tools will silently skip various operations in such an
2013 * "Lockheed Martin Hardened Security for Intel Processors" (HS SRE)
2014 virtualization is now detected.
2016 Changes in the build system:
2018 * Standalone variants of systemd-repart and systemd-shutdown may now be
2019 built (if -Dstandalone=true).
2021 * systemd-ac-power has been moved from /usr/lib/ to /usr/bin/, to, for
2022 example, allow scripts to conditionalize execution on AC power
2025 * The libp11kit library is now loaded through dlopen(3).
2027 Changes in the documentation:
2029 * Specifications that are not closely tied to systemd have moved to
2030 https://uapi-group.org/specifications/: the Boot Loader Specification
2031 and the Discoverable Partitions Specification.
2033 Contributions from: 김인수, 13r0ck, Aidan Dang, Alberto Planas,
2034 Alvin Šipraga, Andika Triwidada, AndyChi, angus-p, Anita Zhang,
2035 Antonio Alvarez Feijoo, Arsen Arsenović, asavah, Benjamin Fogle,
2036 Benjamin Tissoires, berenddeschouwer, BerndAdameit,
2037 Bernd Steinhauser, blutch112, cake03, Callum Farmer, Carlo Teubner,
2038 Charles Hardin, chris, Christian Brauner, Christian Göttsche,
2039 Cristian Rodríguez, Daan De Meyer, Dan Streetman, DaPigGuy,
2040 Darrell Kavanagh, David Tardon, dependabot[bot], Dirk Su,
2041 Dmitry V. Levin, drosdeck, Edson Juliano Drosdeck, edupont,
2042 Eric DeVolder, Erik Moqvist, Evgeny Vereshchagin, Fabian Gurtner,
2043 Felix Riemann, Franck Bui, Frantisek Sumsal, Geert Lorang,
2044 Gerd Hoffmann, Gio, Hannoskaj, Hans de Goede, Hugo Carvalho,
2045 igo95862, Ilya Leoshkevich, Ivan Shapovalov, Jacek Migacz,
2046 Jade Lovelace, Jan Engelhardt, Jan Janssen, Jan Macku, January,
2047 Jason A. Donenfeld, jcg, Jean-Tiare Le Bigot, Jelle van der Waa,
2048 Jeremy Linton, Jian Zhang, Jiayi Chen, Jia Zhang, Joerg Behrmann,
2049 Jörg Thalheim, Joshua Goins, joshuazivkovic, Joshua Zivkovic,
2050 Kai-Chuan Hsieh, Khem Raj, Koba Ko, Lennart Poettering, lichao,
2051 Li kunyu, Luca Boccassi, Luca BRUNO, Ludwig Nussel,
2052 Łukasz Stelmach, Lycowolf, marcel151, Marcus Schäfer, Marek Vasut,
2053 Mark Laws, Michael Biebl, Michał Kotyla, Michal Koutný,
2054 Michal Sekletár, Mike Gilbert, Mike Yuan, MkfsSion, ml,
2055 msizanoen1, mvzlb, MVZ Ludwigsburg, Neil Moore, Nick Rosbrook,
2056 noodlejetski, Pasha Vorobyev, Peter Cai, p-fpv, Phaedrus Leeds,
2057 Philipp Jungkamp, Quentin Deslandes, Raul Tambre, Ray Strode,
2058 reuben olinsky, Richard E. van der Luit, Richard Phibel,
2059 Ricky Tigg, Robin Humble, rogg, Rudi Heitbaum, Sam James,
2060 Samuel Cabrero, Samuel Thibault, Siddhesh Poyarekar, Simon Brand,
2061 Space Meyer, Spindle Security, Steve Ramage, Takashi Sakamoto,
2062 Thomas Haller, Tonći Galić, Topi Miettinen, Torsten Hilbrich,
2063 Tuetuopay, uerdogan, Ulrich Ölmann, Valentin David,
2064 Vitaly Kuznetsov, Vito Caputo, Waltibaba, Will Fancher,
2065 William Roberts, wouter bolsterlee, Youfu Zhang, Yu Watanabe,
2066 Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски,
2069 — Warsaw, 2023-02-15
2073 Announcements of Future Feature Removals:
2075 * We intend to remove cgroup v1 support from systemd release after the
2076 end of 2023. If you run services that make explicit use of cgroup v1
2077 features (i.e. the "legacy hierarchy" with separate hierarchies for
2078 each controller), please implement compatibility with cgroup v2 (i.e.
2079 the "unified hierarchy") sooner rather than later. Most of Linux
2080 userspace has been ported over already.
2082 * We intend to remove support for split-usr (/usr mounted separately
2083 during boot) and unmerged-usr (parallel directories /bin and
2084 /usr/bin, /lib and /usr/lib, etc). This will happen in the second
2085 half of 2023, in the first release that falls into that time window.
2086 For more details, see:
2087 https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html
2089 Compatibility Breaks:
2091 * ConditionKernelVersion= checks that use the '=' or '!=' operators
2092 will now do simple string comparisons (instead of version comparisons
2093 à la stverscmp()). Version comparisons are still done for the
2094 ordering operators '<', '>', '<=', '>='. Moreover, if no operator is
2095 specified, a shell-style glob match is now done. This creates a minor
2096 incompatibility compared to older systemd versions when the '*', '?',
2097 '[', ']' characters are used, as these will now match as shell globs
2098 instead of literally. Given that kernel version strings typically do
2099 not include these characters we expect little breakage through this
2102 * The service manager will now read the SELinux label used for SELinux
2103 access checks from the unit file at the time it loads the file.
2104 Previously, the label would be read at the moment of the access
2105 check, which was problematic since at that time the unit file might
2106 already have been updated or removed.
2110 * systemd-measure is a new tool for calculating and signing expected
2111 TPM2 PCR values for a given unified kernel image (UKI) booted via
2112 sd-stub. The public key used for the signature and the signed
2113 expected PCR information can be embedded inside the UKI. This
2114 information can be extracted from the UKI by external tools and code
2115 in the image itself and is made available to userspace in the booted
2118 systemd-cryptsetup, systemd-cryptenroll, and systemd-creds have been
2119 updated to make use of this information if available in the booted
2120 kernel: when locking an encrypted volume/credential to the TPM
2121 systemd-cryptenroll/systemd-creds will use the public key to bind the
2122 volume/credential to any kernel that carries PCR information signed
2123 by the same key pair. When unlocking such volumes/credentials
2124 systemd-cryptsetup/systemd-creds will use the signature embedded in
2125 the booted UKI to gain access.
2127 Binding TPM-based disk encryption to public keys/signatures of PCR
2128 values — instead of literal PCR values — addresses the inherent
2129 "brittleness" of traditional PCR-bound TPM disk encryption schemes:
2130 disks remain accessible even if the UKI is updated, without any TPM
2131 specific preparation during the OS update — as long as each UKI
2132 carries the necessary PCR signature information.
2134 Net effect: if you boot a properly prepared kernel, TPM-bound disk
2135 encryption now defaults to be locked to kernels which carry PCR
2136 signatures from the same key pair. Example: if a hypothetical distro
2137 FooOS prepares its UKIs like this, TPM-based disk encryption is now –
2138 by default – bound to only FooOS kernels, and encrypted volumes bound
2139 to the TPM cannot be unlocked on kernels from other sources. (But do
2140 note this behaviour requires preparation/enabling in the UKI, and of
2141 course users can always enroll non-TPM ways to unlock the volume.)
2143 * systemd-pcrphase is a new tool that is invoked at six places during
2144 system runtime, and measures additional words into TPM2 PCR 11, to
2145 mark milestones of the boot process. This allows binding access to
2146 specific TPM2-encrypted secrets to specific phases of the boot
2147 process. (Example: LUKS2 disk encryption key only accessible in the
2148 initrd, but not later.)
2150 Changes in systemd itself, i.e. the manager and units
2152 * The cpu controller is delegated to user manager units by default, and
2153 CPUWeight= settings are applied to the top-level user slice units
2154 (app.slice, background.slice, session.slice). This provides a degree
2155 of resource isolation between different user services competing for
2158 * Systemd can optionally do a full preset in the "first boot" condition
2159 (instead of just enable-only). This behaviour is controlled by the
2160 compile-time option -Dfirst-boot-full-preset. Right now it defaults
2161 to 'false', but the plan is to switch it to 'true' for the subsequent
2164 * Drop-ins are now allowed for transient units too.
2166 * Systemd will set the taint flag 'support-ended' if it detects that
2167 the OS image is past its end-of-support date. This date is declared
2168 in a new /etc/os-release field SUPPORT_END= described below.
2170 * Two new settings ConditionCredential= and AssertCredential= can be
2171 used to skip or fail units if a certain system credential is not
2174 * ConditionMemory= accepts size suffixes (K, M, G, T, …).
2176 * DefaultSmackProcessLabel= can be used in system.conf and user.conf to
2177 specify the SMACK security label to use when not specified in a unit
2180 * DefaultDeviceTimeoutSec= can be used in system.conf and user.conf to
2181 specify the default timeout when waiting for device units to
2184 * C.UTF-8 is used as the default locale if nothing else has been
2187 * [Condition|Assert]Firmware= have been extended to support certain
2188 SMBIOS fields. For example
2190 ConditionFirmware=smbios-field(board_name = "Custom Board")
2192 conditionalizes the unit to run only when
2193 /sys/class/dmi/id/board_name contains "Custom Board" (without the
2196 * ConditionFirstBoot= now correctly evaluates as true only during the
2197 boot phase of the first boot. A unit executed later, after booting
2198 has completed, will no longer evaluate this condition as true.
2200 * Socket units will now create sockets in the SELinuxContext= of the
2201 associated service unit, if any.
2203 * Boot phase transitions (start initrd → exit initrd → boot complete →
2204 shutdown) will be measured into TPM2 PCR 11, so that secrets can be
2205 bound to a specific runtime phase. E.g.: a LUKS encryption key can be
2206 unsealed only in the initrd.
2208 * Service credentials (i.e. SetCredential=/LoadCredential=/…) will now
2209 also be provided to ExecStartPre= processes.
2211 * Various units are now correctly ordered against
2212 initrd-switch-root.target where previously a conflict without
2213 ordering was configured. A stop job for those units would be queued,
2214 but without the ordering it could be executed only after
2215 initrd-switch-root.service, leading to units not being restarted in
2216 the host system as expected.
2218 * In order to fully support the IPMI watchdog driver, which has not yet
2219 been ported to the new common watchdog device interface,
2220 /dev/watchdog0 will be tried first and systemd will silently fallback
2221 to /dev/watchdog if it is not found.
2223 * New watchdog-related D-Bus properties are now published by systemd:
2224 WatchdogDevice, WatchdogLastPingTimestamp,
2225 WatchdogLastPingTimestampMonotonic.
2227 * At shutdown, API virtual files systems (proc, sys, etc.) will be
2230 * At shutdown, systemd will now log about processes blocking unmounting
2233 * A new meson build option 'clock-valid-range-usec-max' was added to
2234 allow disabling system time correction if RTC returns a timestamp far
2237 * Propagated restart jobs will no longer be discarded while a unit is
2240 * PID 1 will now import system credentials from SMBIOS Type 11 fields
2241 ("OEM vendor strings"), in addition to qemu_fwcfg. This provides a
2242 simple, fast and generic path for supplying credentials to a VM,
2243 without involving external tools such as cloud-init/ignition.
2245 * The CPUWeight= setting of unit files now accepts a new special value
2246 "idle", which configures "idle" level scheduling for the unit.
2248 * Service processes that are activated due to a .timer or .path unit
2249 triggering will now receive information about this via environment
2250 variables. Note that this is information is lossy, as activation
2251 might be coalesced and only one of the activating triggers will be
2252 reported. This is hence more suited for debugging or tracing rather
2253 than for behaviour decisions.
2255 * The riscv_flush_icache(2) system call has been added to the list of
2256 system calls allowed by default when SystemCallFilter= is used.
2258 * The selinux context derived from the target executable, instead of
2259 'init_t' used for the manager itself, is now used when creating
2260 listening sockets for units that specify SELinuxContextFromNet=yes.
2262 Changes in sd-boot, bootctl, and the Boot Loader Specification:
2264 * The Boot Loader Specification has been cleaned up and clarified.
2265 Various corner cases in version string comparisons have been fixed
2266 (e.g. comparisons for empty strings). Boot counting is now part of
2267 the main specification.
2269 * New PCRs measurements are performed during boot: PCR 11 for the
2270 kernel+initrd combo, PCR 13 for any sysext images. If a measurement
2271 took place this is now reported to userspace via the new
2272 StubPcrKernelImage and StubPcrInitRDSysExts EFI variables.
2274 * As before, systemd-stub will measure kernel parameters and system
2275 credentials into PCR 12. It will now report this fact via the
2276 StubPcrKernelParameters EFI variable to userspace.
2278 * The UEFI monotonic boot counter is now included in the updated random
2279 seed file maintained by sd-boot, providing some additional entropy.
2281 * sd-stub will use LoadImage/StartImage to execute the kernel, instead
2282 of arranging the image manually and jumping to the kernel entry
2283 point. sd-stub also installs a temporary UEFI SecurityOverride to
2284 allow the (unsigned) nested image to be booted. This is safe because
2285 the outer (signed) stub+kernel binary must have been verified before
2286 the stub was executed.
2288 * Booting in EFI mixed mode (a 64-bit kernel over 32-bit UEFI firmware)
2289 is now supported by sd-boot.
2291 * bootctl gained a bunch of new options: --all-architectures to install
2292 binaries for all supported EFI architectures, --root= and --image=
2293 options to operate on a directory or disk image, and
2294 --install-source= to specify the source for binaries to install,
2295 --efi-boot-option-description= to control the name of the boot entry.
2297 * The sd-boot stub exports a StubFeatures flag, which is used by
2298 bootctl to show features supported by the stub that was used to boot.
2300 * The PE section offsets that are used by tools that assemble unified
2301 kernel images have historically been hard-coded. This may lead to
2302 overlapping PE sections which may break on boot. The UKI will now try
2303 to detect and warn about this.
2305 Any tools that assemble UKIs must update to calculate these offsets
2306 dynamically. Future sd-stub versions may use offsets that will not
2307 work with the currently used set of hard-coded offsets!
2309 * sd-stub now accepts (and passes to the initrd and then to the full
2310 OS) new PE sections '.pcrsig' and '.pcrkey' that can be used to embed
2311 signatures of expected PCR values, to allow sealing secrets via the
2312 TPM2 against pre-calculated PCR measurements.
2314 Changes in the hardware database:
2316 * 'systemd-hwdb query' now supports the --root= option.
2318 Changes in systemctl:
2320 * systemctl now supports --state= and --type= options for the 'show'
2323 * systemctl gained a new verb 'list-automounts' to list automount
2326 * systemctl gained support for a new --image= switch to be able to
2327 operate on the specified disk image (similar to the existing --root=
2328 which operates relative to some directory).
2330 Changes in systemd-networkd:
2332 * networkd can set Linux NetLabel labels for integration with the
2333 network control in security modules via a new NetLabel= option.
2335 * The RapidCommit= is (re-)introduced to enable faster configuration
2336 via DHCPv6 (RFC 3315).
2338 * networkd gained a new option TCPCongestionControlAlgorithm= that
2339 allows setting a per-route TCP algorithm.
2341 * networkd gained a new option KeepFileDescriptor= to allow keeping a
2342 reference (file descriptor) open on TUN/TAP interfaces, which is
2343 useful to avoid link flaps while the underlying service providing the
2344 interface is being serviced.
2346 * RouteTable= now also accepts route table names.
2348 Changes in systemd-nspawn:
2350 * The --bind= and --overlay= options now support relative paths.
2352 * The --bind= option now supports a 'rootidmap' value, which will
2353 use id-mapped mounts to map the root user inside the container to the
2354 owner of the mounted directory on the host.
2356 Changes in systemd-resolved:
2358 * systemd-resolved now persists DNSOverTLS in its state file too. This
2359 fixes a problem when used in combination with NetworkManager, which
2360 sends the setting only once, causing it to be lost if resolved was
2361 restarted at any point.
2363 * systemd-resolved now exposes a Varlink socket at
2364 /run/systemd/resolve/io.systemd.Resolve.Monitor, accessible only for
2365 root. Processed DNS requests in a JSON format will be published to
2366 any clients connected to this socket.
2368 resolvectl gained a 'monitor' verb to make use of this.
2370 * systemd-resolved now treats unsupported DNSSEC algorithms as INSECURE
2371 instead of returning SERVFAIL, as per RFC:
2372 https://datatracker.ietf.org/doc/html/rfc6840#section-5.2
2374 * OpenSSL is the default crypto backend for systemd-resolved. (gnutls
2375 is still supported.)
2377 Changes in libsystemd and other libraries:
2379 * libsystemd now exports sd_bus_error_setfv() (a convenience function
2380 for setting bus errors), sd_id128_string_equal (a convenience
2381 function for 128-bit ID string comparisons), and
2382 sd_bus_message_read_strv_extend() (a function to incrementally read
2385 * libsystemd now exports sd_device_get_child_first()/_next() as a
2386 high-level interface for enumerating child devices. It also supports
2387 sd_device_new_child() for opening a child device given a device
2390 * libsystemd now exports sd_device_monitor_set()/get_description()
2391 which allow setting a custom description that will be used in log
2392 messages by sd_device_monitor*.
2394 * Private shared libraries (libsystemd-shared-nnn.so,
2395 libsystemd-core-nnn.so) are now installed into arch-specific
2396 directories to allow multi-arch installs.
2398 * A new sd-gpt.h header is now published, listing GUIDs from the
2399 Discoverable Partitions specification. For more details see:
2400 https://systemd.io/DISCOVERABLE_PARTITIONS/
2402 * A new function sd_hwdb_new_from_path() has been added to open a hwdb
2403 database given an explicit path to the file.
2405 * The signal number argument to sd_event_add_signal() now can now be
2406 ORed with the SD_EVENT_SIGNAL_PROCMASK flag, causing sigprocmask() to
2407 be automatically invoked to block the specified signal. This is
2408 useful to simplify invocations as the caller doesn't have to do this
2411 * A new convenience call sd_event_set_signal_exit() has been added to
2412 sd-event to set up signal handling so that the event loop
2413 automatically terminates cleanly on SIGTERM/SIGINT.
2415 Changes in other components:
2417 * systemd-sysusers, systemd-tmpfiles, and systemd-sysctl configuration
2418 can now be provided via the credential mechanism.
2420 * systemd-analyze gained a new verb 'compare-versions' that implements
2421 comparisons for versions strings (similarly to 'rpmdev-vercmp' and
2422 'dpkg --compare-versions').
2424 * 'systemd-analyze dump' is extended to accept glob patterns for unit
2425 names to limit the output to matching units.
2427 * tmpfiles.d/ lines can read file contents to write from a credential.
2428 The new modifier char '^' is used to specify that the argument is a
2429 credential name. This mechanism is used to automatically populate
2430 /etc/motd, /etc/issue, and /etc/hosts from credentials.
2432 * tmpfiles.d/ may now be configured to avoid changing uid/gid/mode of
2433 an inode if the specification is prefixed with ':' and the inode
2436 * Default tmpfiles.d/ configuration now carries a line to automatically
2437 use an 'ssh.authorized_keys.root' credential if provided to set up
2438 the SSH authorized_keys file for the root user.
2440 * systemd-tmpfiles will now gracefully handle absent source of "C" copy
2443 * tmpfiles.d/ F/w lines now optionally permit encoding of the payload
2444 in base64. This is useful to write arbitrary binary data into files.
2446 * The pkgconfig and rpm macros files now export the directory for user
2447 units as 'user_tmpfiles_dir' and '%_user_tmpfilesdir'.
2449 * Detection of Apple Virtualization and detection of Parallels and
2450 KubeVirt virtualization on non-x86 archs have been added.
2452 * os-release gained a new field SUPPORT_END=YYYY-MM-DD to inform the
2453 user when their system will become unsupported.
2455 * When performing suspend-then-hibernate, the system will estimate the
2456 discharge rate and use that to set the delay until hibernation and
2457 hibernate immediately instead of suspending when running from a
2458 battery and the capacity is below 5%.
2460 * systemd-sysctl gained a --strict option to fail when a sysctl
2461 setting is unknown to the kernel.
2463 * machinectl supports --force for the 'copy-to' and 'copy-from'
2466 * coredumpctl gained the --root and --image options to look for journal
2467 files under the specified root directory, image, or block device.
2469 * 'journalctl -o' and similar commands now implement a new output mode
2470 "short-delta". It is similar to "short-monotonic", but also shows the
2471 time delta between subsequent messages.
2473 * journalctl now respects the --quiet flag when verifying consistency
2476 * Journal log messages gained a new implicit field _RUNTIME_SCOPE= that
2477 will indicate whether a message was logged in the 'initrd' phase or
2478 in the 'system' phase of the boot process.
2480 * Journal files gained a new compatibility flag
2481 'HEADER_INCOMPATIBLE_COMPACT'. Files with this flag implement changes
2482 to the storage format that allow reducing size on disk. As with other
2483 compatibility flags, older journalctl versions will not be able to
2484 read journal files using this new format. The environment variable
2485 'SYSTEMD_JOURNAL_COMPACT=0' can be passed to systemd-journald to
2486 disable this functionality. It is enabled by default.
2488 * systemd-run's --working-directory= switch now works when used in
2489 combination with --scope.
2491 * portablectl gained a --force flag to skip certain sanity checks. This
2492 is implemented using new flags accepted by systemd-portabled for the
2493 *WithExtensions() D-Bus methods: SD_SYSTEMD_PORTABLE_FORCE_ATTACH
2494 flag now means that the attach/detach checks whether the units are
2495 already present and running will be skipped. Similarly,
2496 SD_SYSTEMD_PORTABLE_FORCE_SYSEXT flag means that the check whether
2497 image name matches the name declared inside of the image will be
2498 skipped. Callers must be sure to do those checks themselves if
2501 * systemd-portabled will now use the original filename to check
2502 extension-release.NAME for correctness, in case it is passed a
2505 * systemd-portabled now uses PrivateTmp=yes in the 'trusted' profile
2508 * sysext's extension-release files now support '_any' as a special
2509 value for the ID= field, to allow distribution-independent extensions
2510 (e.g.: fully statically compiled binaries, scripts). It also gained
2511 support for a new ARCHITECTURE= field that may be used to explicitly
2512 restrict an image to hosts of a specific architecture.
2514 * systemd-repart now supports creating squashfs partitions. This
2515 requires mksquashfs from squashfs-tools.
2517 * systemd-repart gained a --split flag to also generate split
2518 artifacts, i.e. a separate file for each partition. This is useful in
2519 conjunction with systemd-sysupdate or other tools, or to generate
2520 split dm-verity artifacts.
2522 * systemd-repart is now able to generate dm-verity partitions, including
2525 * systemd-repart can now set a partition UUID to zero, allowing it to
2526 be filled in later, such as when using verity partitions.
2528 * systemd-repart now supports drop-ins for its configuration files.
2530 * Package metadata logged by systemd-coredump in the system journal is
2533 * xdg-autostart-service now expands 'tilde' characters in Exec lines.
2535 * systemd-oomd now automatically links against libatomic, if available.
2537 * systemd-oomd now sends out a 'Killed' D-Bus signal when a cgroup is
2540 * scope units now also provide oom-kill status.
2542 * systemd-pstore will now try to load only the efi_pstore kernel module
2543 before running, ensuring that pstore can be used.
2545 * systemd-logind gained a new StopIdleSessionSec= option to stop an idle
2546 session after a preconfigure timeout.
2548 * systemd-homed will now wait up to 30 seconds for workers to terminate,
2549 rather than indefinitely.
2551 * homectl gained a new '--luks-sector-size=' flag that allows users to
2552 select the preferred LUKS sector size. Must be a power of 2 between 512
2553 and 4096. systemd-userdbd records gained a corresponding field.
2555 * systemd-sysusers will now respect the 'SOURCE_DATE_EPOCH' environment
2556 variable when generating the 'sp_lstchg' field, to ensure an image
2557 build can be reproducible.
2559 * 'udevadm wait' will now listen to kernel uevents too when called with
2562 * When naming network devices udev will now consult the Devicetree
2563 "alias" fields for the device.
2565 * systemd-udev will now create infiniband/by-path and
2566 infiniband/by-ibdev links for Infiniband verbs devices.
2568 * systemd-udev-trigger.service will now also prioritize input devices.
2570 * ConditionACPower= and systemd-ac-power will now assume the system is
2571 running on AC power if no battery can be found.
2573 * All features and tools using the TPM2 will now communicate with it
2574 using a bind key. Beforehand, the tpm2 support used encrypted sessions
2575 by creating a primary key that was used to encrypt traffic. This
2576 creates a problem as the key created for encrypting the traffic could
2577 be faked by an active interposer on the bus. In cases when a pin is
2578 used, a bind key will be used. The pin is used as the auth value for
2579 the seal key, aka the disk encryption key, and that auth value will be
2580 used in the session establishment. An attacker would need the pin
2581 value to create the secure session and thus an active interposer
2582 without the pin cannot interpose on TPM2 traffic.
2584 * systemd-growfs no longer requires udev to run.
2586 * systemd-backlight now will better support systems with multiple
2589 * systemd-cryptsetup's keyfile-timeout= option now also works when a
2590 device is used as a keyfile.
2592 * systemd-cryptenroll gained a new --unlock-key-file= option to get the
2593 unlocking key from a key file (instead of prompting the user). Note
2594 that this is the key for unlocking the volume in order to be able to
2595 enroll a new key, but it is not the key that is enrolled.
2597 * systemd-dissect gained a new --umount switch that will safely and
2598 synchronously unmount all partitions of an image previously mounted
2599 with 'systemd-dissect --mount'.
2601 * When using gcrypt, all systemd tools and services will now configure
2602 it to prefer the OS random number generator if present.
2604 * All example code shipped with documentation has been relicensed from CC0
2607 * Unit tests will no longer fail when running on a system without
2610 Experimental features:
2612 * BPF programs can now be compiled with bpf-gcc (requires libbpf >= 1.0
2613 and bpftool >= 7.0).
2615 * sd-boot can automatically enroll SecureBoot keys from files found on
2616 the ESP. This enrollment can be either automatic ('force' mode) or
2617 controlled by the user ('manual' mode). It is sufficient to place the
2618 SecureBoot keys in the right place in the ESP and they will be picked
2619 up by sd-boot and shown in the boot menu.
2621 * The mkosi config in systemd gained support for automatically
2622 compiling a kernel with the configuration appropriate for testing
2623 systemd. This may be useful when developing or testing systemd in
2624 tandem with the kernel.
2626 Contributions from: 김인수, Adam Williamson, adrian5, Aidan Dang,
2627 Akihiko Odaki, Alban Bedel, Albert Mikaelyan, Aleksey Vasenev,
2628 Alexander Graf, Alexander Shopov, Alexander Wilson,
2629 Alper Nebi Yasak, anarcat, Anders Jonsson, Andre Kalb,
2630 Andrew Stone, Andrey Albershteyn, Anita Zhang, Ansgar Burchardt,
2631 Antonio Alvarez Feijoo, Arnaud Ferraris, Aryan singh, asavah,
2632 Avamander, Avram Lubkin, Balázs Meskó, Bastien Nocera,
2633 Benjamin Franzke, BerndAdameit, bin456789, Celeste Liu,
2634 Chih-Hsuan Yen, Christian Brauner, Christian Göttsche,
2635 Christian Hesse, Clyde Byrd III, codefiles, Colin Walters,
2636 Cristian Rodríguez, Daan De Meyer, Daniel Braunwarth,
2637 Daniel Rusek, Dan Streetman, Darsey Litzenberger, David Edmundson,
2638 David Jaša, David Rheinsberg, David Seifert, David Tardon,
2639 dependabot[bot], Devendra Tewari, Dominique Martinet, drosdeck,
2640 Edson Juliano Drosdeck, Eduard Tolosa, eggfly, Einsler Lee,
2641 Elias Probst, Eli Schwartz, Evgeny Vereshchagin, exploide, Fei Li,
2642 Foster Snowhill, Franck Bui, Frank Dana, Frantisek Sumsal,
2643 Gerd Hoffmann, Gio, Goffredo Baroncelli, gtwang01,
2644 Guillaume W. Bres, H A, Hans de Goede, Heinrich Schuchardt,
2645 Hugo Carvalho, i-do-cpp, igo95862, j00512545, Jacek Migacz,
2646 Jade Bilkey, James Hilliard, Jan B, Janis Goldschmidt,
2647 Jan Janssen, Jan Kuparinen, Jan Luebbe, Jan Macku,
2648 Jason A. Donenfeld, Javkhlanbayar Khongorzul, Jeremy Soller,
2649 JeroenHD, jiangchuangang, João Loureiro,
2650 Joaquín Ignacio Aramendía, Jochen Sprickerhof,
2651 Johannes Schauer Marin Rodrigues, Jonas Kümmerlin,
2652 Jonas Witschel, Jonathan Kang, Jonathan Lebon, Joost Heitbrink,
2653 Jörg Thalheim, josh-gordon-fb, Joyce, Kai Lueke, lastkrick,
2654 Lennart Poettering, Leon M. George, licunlong, Li kunyu,
2655 LockBlock-dev, Loïc Collignon, Lubomir Rintel, Luca Boccassi,
2656 Luca BRUNO, Ludwig Nussel, Łukasz Stelmach, Maccraft123,
2657 Marc Kleine-Budde, Marius Vollmer, Martin Wilck, matoro,
2658 Matthias Lisin, Max Gautier, Maxim Mikityanskiy, Michael Biebl,
2659 Michal Koutný, Michal Sekletár, Michal Stanke, Mike Gilbert,
2660 Mitchell Freiderich, msizanoen1, Nick Rosbrook, nl6720, Oğuz Ersen,
2661 Oleg Solovyov, Olga Smirnova, Pablo Ceballos, Pavel Zhukov,
2662 Phaedrus Leeds, Philipp Gortan, Piotr Drąg, Pyfisch,
2663 Quentin Deslandes, Rahil Bhimjiani, Rene Hollander, Richard Huang,
2664 Richard Phibel, Rudi Heitbaum, Sam James, Sarah Brofeldt,
2665 Sean Anderson, Sebastian Scheibner, Shreenidhi Shedi,
2666 Sonali Srivastava, Steve Ramage, Suraj Krishnan, Swapnil Devesh,
2667 Takashi Sakamoto, Ted X. Toth, Temuri Doghonadze, Thomas Blume,
2668 Thomas Haller, Thomas Hebb, Tomáš Hnyk, Tomasz Paweł Gajc,
2669 Topi Miettinen, Ulrich Ölmann, undef, Uriel Corfa,
2670 Victor Westerhuis, Vincent Dagonneau, Vishal Chillara Srinivas,
2671 Vito Caputo, Weblate, Wenchao Hao, William Roberts, williamsumendap,
2672 wineway, xiaoyang, Yuri Chornoivan, Yu Watanabe,
2673 Zbigniew Jędrzejewski-Szmek, Zhaofeng Li, наб
2675 – The Great Beyond, 2022-10-31 👻
2679 Backwards-incompatible changes:
2681 * The minimum kernel version required has been bumped from 3.13 to 4.15,
2682 and CLOCK_BOOTTIME is now assumed to always exist.
2684 * C11 with GNU extensions (aka "gnu11") is now used to build our
2685 components. Public API headers are still restricted to ISO C89.
2687 * In v250, a systemd-networkd feature that automatically configures
2688 routes to addresses specified in AllowedIPs= was added and enabled by
2689 default. However, this causes network connectivity issues in many
2690 existing setups. Hence, it has been disabled by default since
2691 systemd-stable 250.3. The feature can still be used by explicitly
2692 configuring RouteTable= setting in .netdev files.
2694 * Jobs started via StartUnitWithFlags() will no longer return 'skipped'
2695 when a Condition*= check does not succeed, restoring the JobRemoved
2696 signal to the behaviour it had before v250.
2698 * The org.freedesktop.portable1 methods GetMetadataWithExtensions() and
2699 GetImageMetadataWithExtensions() have been fixed to provide an extra
2700 return parameter, containing the actual extension release metadata.
2701 The current implementation was judged to be broken and unusable, and
2702 thus the usual procedure of adding a new set of methods was skipped,
2703 and backward compatibility broken instead on the assumption that
2704 nobody can be affected given the current state of this interface.
2706 * All kernels supported by systemd mix bytes returned by RDRAND (or
2707 similar) into the entropy pool at early boot. This means that on
2708 those systems, even if /dev/urandom is not yet initialized, it still
2709 returns bytes that are of at least RDRAND quality. For that reason,
2710 we no longer have reason to invoke RDRAND from systemd itself, which
2711 has historically been a source of bugs. Furthermore, kernels ≥5.6
2712 provide the getrandom(GRND_INSECURE) interface for returning random
2713 bytes before the entropy pool is initialized without warning into
2714 kmsg, which is what we attempt to use if available. systemd's direct
2715 usage of RDRAND has been removed. x86 systems ≥Broadwell that are
2716 running an older kernel may experience kmsg warnings that were not
2717 seen with 250. For newer kernels, non-x86 systems, or older x86
2718 systems, there should be no visible changes.
2720 * sd-boot will now measure the kernel command line into TPM PCR 12
2721 rather than PCR 8. This improves usefulness of the measurements on
2722 systems where sd-boot is chainloaded from Grub. Grub measures all
2723 commands its executes into PCR 8, which makes it very hard to use
2724 reasonably, hence separate ourselves from that and use PCR 12
2725 instead, which is what certain Ubuntu editions already do. To retain
2726 compatibility with systems running older systemd systems a new meson
2727 option 'efi-tpm-pcr-compat' has been added (which defaults to false).
2728 If enabled, the measurement is done twice: into the new-style PCR 12
2729 *and* the old-style PCR 8. It's strongly advised to migrate all users
2730 to PCR 12 for this purpose in the long run, as we intend to remove
2731 this compatibility feature in two years' time.
2733 * busctl capture now writes output in the newer pcapng format instead
2736 * A udev rule that imported hwdb matches for USB devices with lowercase
2737 hexadecimal vendor/product ID digits was added in systemd 250. This
2738 has been reverted, since uppercase hexadecimal digits are supposed to
2739 be used, and we already had a rule with the appropriate match.
2741 Users might need to adjust their local hwdb entries.
2743 * arch_prctl(2) has been moved to the @default set in the syscall filters
2744 (as exposed via the SystemCallFilter= setting in service unit files).
2745 It is apparently used by the linker now.
2747 * The tmpfiles entries that create the /run/systemd/netif directory and
2748 its subdirectories were moved from tmpfiles.d/systemd.conf to
2749 tmpfiles.d/systemd-network.conf.
2751 Users might need to adjust their files that override tmpfiles.d/systemd.conf
2752 to account for this change.
2754 * The requirement for Portable Services images to contain a well-formed
2755 os-release file (i.e.: contain at least an ID field) is now enforced.
2756 This applies to base images and extensions, and also to systemd-sysext.
2758 Changes in the Boot Loader Specification, kernel-install and sd-boot:
2760 * kernel-install's and bootctl's Boot Loader Specification Type #1
2761 entry generation logic has been reworked. The user may now pick
2762 explicitly by which "token" string to name the installation's boot
2763 entries, via the new /etc/kernel/entry-token file or the new
2764 --entry-token= switch to bootctl. By default — as before — the
2765 entries are named after the local machine ID. However, in "golden
2766 image" environments, where the machine ID shall be initialized on
2767 first boot (as opposed to at installation time before first boot) the
2768 machine ID will not be available at build time. In this case the
2769 --entry-token= switch to bootctl (or the /etc/kernel/entry-token
2770 file) may be used to override the "token" for the entries, for
2771 example the IMAGE_ID= or ID= fields from /etc/os-release. This will
2772 make the OS images independent of any machine ID, and ensure that the
2773 images will not carry any identifiable information before first boot,
2774 but on the other hand means that multiple parallel installations of
2775 the very same image on the same disk cannot be supported.
2777 Summary: if you are building golden images that shall acquire
2778 identity information exclusively on first boot, make sure to both
2779 remove /etc/machine-id *and* to write /etc/kernel/entry-token to the
2780 value of the IMAGE_ID= or ID= field of /etc/os-release or another
2781 suitable identifier before deploying the image.
2783 * The Boot Loader Specification has been extended with
2784 /loader/entries.srel file located in the EFI System Partition (ESP)
2785 that disambiguates the format of the entries in the /loader/entries/
2786 directory (in order to discern them from incompatible uses of this
2787 directory by other projects). For entries that follow the
2788 Specification, the string "type1" is stored in this file.
2790 bootctl will now write this file automatically when installing the
2791 systemd-boot boot loader.
2793 * kernel-install supports a new initrd_generator= setting in
2794 /etc/kernel/install.conf, that is exported as
2795 $KERNEL_INSTALL_INITRD_GENERATOR to kernel-install plugins. This
2796 allows choosing different initrd generators.
2798 * kernel-install will now create a "staging area" (an initially-empty
2799 directory to gather files for a Boot Loader Specification Type #1
2800 entry). The path to this directory is exported as
2801 $KERNEL_INSTALL_STAGING_AREA to kernel-install plugins, which should
2802 drop files there instead of writing them directly to the final
2803 location. kernel-install will move them when all files have been
2804 prepared successfully.
2806 * New option sort-key= has been added to the Boot Loader Specification
2807 to override the sorting order of the entries in the boot menu. It is
2808 read by sd-boot and bootctl, and will be written by kernel-install,
2809 with the default value of IMAGE_ID= or ID= fields from
2810 os-release. Together, this means that on multiboot installations,
2811 entries should be grouped and sorted in a predictable way.
2813 * The sort order of boot entries has been updated: entries which have
2814 the new field sort-key= are sorted by it first, and all entries
2815 without it are ordered later. After that, entries are sorted by
2816 version so that newest entries are towards the beginning of the list.
2818 * The kernel-install tool gained a new 'inspect' verb which shows the
2819 paths and other settings used.
2821 * sd-boot can now optionally beep when the menu is shown and menu
2822 entries are selected, which can be useful on machines without a
2823 working display. (Controllable via a loader.conf setting.)
2825 * The --make-machine-id-directory= switch to bootctl has been replaced
2826 by --make-entry-directory=, given that the entry directory is not
2827 necessarily named after the machine ID, but after some other suitable
2828 ID as selected via --entry-token= described above. The old name of
2829 the option is still understood to maximize compatibility.
2831 * 'bootctl list' gained support for a new --json= switch to output boot
2832 menu entries in JSON format.
2834 * 'bootctl is-installed' now supports the --graceful, and various verbs
2835 omit output with the new option --quiet.
2837 Changes in systemd-homed:
2839 * Starting with v250 systemd-homed uses UID/GID mapping on the mounts
2840 of activated home directories it manages (if the kernel and selected
2841 file systems support it). So far it mapped three UID ranges: the
2842 range from 0…60000, the user's own UID, and the range 60514…65534,
2843 leaving everything else unmapped (in other words, the 16-bit UID range
2844 is mapped almost fully, with the exception of the UID subrange used
2845 for systemd-homed users, with one exception: the user's own UID).
2846 Unmapped UIDs may not be used for file ownership in the home
2847 directory — any chown() attempts with them will fail. With this
2848 release a fourth range is added to these mappings:
2849 524288…1879048191. This range is the UID range intended for container
2852 https://systemd.io/UIDS-GIDS
2854 This range may be used for container managers that place container OS
2855 trees in the home directory (which is a questionable approach, for
2856 quota, permission, SUID handling and network file system
2857 compatibility reasons, but nonetheless apparently commonplace). Note
2858 that this mapping is mapped 1:1 in a pass-through fashion, i.e. the
2859 UID assignments from the range are not managed or mapped by
2860 `systemd-homed`, and must be managed with other mechanisms, in the
2861 context of the local system.
2863 Typically, a better approach to user namespacing in relevant
2864 container managers would be to leave container OS trees on disk at
2865 UID offset 0, but then map them to a dynamically allocated runtime
2866 UID range via another UID mount map at container invocation
2867 time. That way user namespace UID ranges become strictly a runtime
2868 concept, and do not leak into persistent file systems, persistent
2869 user databases or persistent configuration, thus greatly simplifying
2870 handling, and improving compatibility with home directories intended
2871 to be portable like the ones managed by systemd-homed.
2873 Changes in shared libraries:
2875 * A new libsystemd-core-<version>.so private shared library is
2876 installed under /usr/lib/systemd/system, mirroring the existing
2877 libsystemd-shared-<version>.so library. This allows the total
2878 installation size to be reduced by binary code reuse.
2880 * The <version> tag used in the name of libsystemd-shared.so and
2881 libsystemd-core.so can be configured via the meson option
2882 'shared-lib-tag'. Distributions may build subsequent versions of the
2883 systemd package with unique tags (e.g. the full package version),
2884 thus allowing multiple installations of those shared libraries to be
2885 available at the same time. This is intended to fix an issue where
2886 programs that link to those libraries would fail to execute because
2887 they were installed earlier or later than the appropriate version of
2890 * The sd-id128 API gained a new call sd_id128_to_uuid_string() that is
2891 similar to sd_id128_to_string() but formats the ID in RFC 4122 UUID
2892 format instead of as a simple series of hex characters.
2894 * The sd-device API gained two new calls sd_device_new_from_devname()
2895 and sd_device_new_from_path() which permit allocating an sd_device
2896 object from a device node name or file system path.
2898 * sd-device also gained a new call sd_device_open() which will open the
2899 device node associated with a device for which an sd_device object
2900 has been allocated. The call is supposed to address races around
2901 device nodes being removed/recycled due to hotplug events, or media
2902 change events: the call checks internally whether the major/minor of
2903 the device node and the "diskseq" (in case of block devices) match
2904 with the metadata loaded in the sd_device object, thus ensuring that
2905 the device once opened really matches the provided sd_device object.
2907 Changes in PID1, systemctl, and systemd-oomd:
2909 * A new set of service monitor environment variables will be passed to
2910 OnFailure=/OnSuccess= handlers, but only if exactly one unit lists the
2911 handler unit as OnFailure=/OnSuccess=. The variables are:
2912 $MONITOR_SERVICE_RESULT, $MONITOR_EXIT_CODE, $MONITOR_EXIT_STATUS,
2913 $MONITOR_INVOCATION_ID and $MONITOR_UNIT. For cases when a single
2914 handler needs to watch multiple units, use a templated handler.
2916 * A new ExtensionDirectories= setting in service unit files allows
2917 system extensions to be loaded from a directory. (It is similar to
2918 ExtensionImages=, but takes paths to directories, instead of
2921 'portablectl attach --extension=' now also accepts directory paths.
2923 * The user.delegate and user.invocation_id extended attributes on
2924 cgroups are used in addition to trusted.delegate and
2925 trusted.invocation_id. The latter pair requires privileges to set,
2926 but the former doesn't and can be also set by the unprivileged user
2929 (Only supported on kernels ≥5.6.)
2931 * Units that were killed by systemd-oomd will now have a service result
2932 of 'oom-kill'. The number of times a service was killed is tallied
2933 in the 'user.oomd_ooms' extended attribute.
2935 The OOMPolicy= unit file setting is now also honoured by
2938 * In unit files the new %y/%Y specifiers can be used to refer to
2939 normalized unit file path, which is particularly useful for symlinked
2942 The new %q specifier resolves to the pretty hostname
2943 (i.e. PRETTY_HOSTNAME= from /etc/machine-info).
2945 The new %d specifier resolves to the credentials directory of a
2946 service (same as $CREDENTIALS_DIRECTORY).
2948 * The RootDirectory=, MountAPIVFS=, ExtensionDirectories=,
2949 *Capabilities*=, ProtectHome=, *Directory=, TemporaryFileSystem=,
2950 PrivateTmp=, PrivateDevices=, PrivateNetwork=, NetworkNamespacePath=,
2951 PrivateIPC=, IPCNamespacePath=, PrivateUsers=, ProtectClock=,
2952 ProtectKernelTunables=, ProtectKernelModules=, ProtectKernelLogs=,
2953 MountFlags= service settings now also work in unprivileged user
2954 services, i.e. those run by the user's --user service manager, as long
2955 as user namespaces are enabled on the system.
2957 * Services with Restart=always and a failing ExecCondition= will no
2958 longer be restarted, to bring ExecCondition= behaviour in line with
2959 Condition*= settings.
2961 * LoadCredential= now accepts a directory as the argument; all files
2962 from the directory will be loaded as credentials.
2964 * A new D-Bus property ControlGroupId is now exposed on service units,
2965 that encapsulates the service's numeric cgroup ID that newer kernels
2966 assign to each cgroup.
2968 * PID 1 gained support for configuring the "pre-timeout" of watchdog
2969 devices and the associated governor, via the new
2970 RuntimeWatchdogPreSec= and RuntimeWatchdogPreGovernor= configuration
2971 options in /etc/systemd/system.conf.
2973 * systemctl's --timestamp= option gained a new choice "unix", to show
2974 timestamp as unix times, i.e. seconds since 1970, Jan 1st.
2976 * A new "taint" flag named "old-kernel" is introduced which is set when
2977 the kernel systemd runs on is older then the current baseline version
2978 (see above). The flag is shown in "systemctl status" output.
2980 * Two additional taint flags "short-uid-range" and "short-gid-range"
2981 have been added as well, which are set when systemd notices it is run
2982 within a userns namespace that does not define the full 0…65535 UID
2985 * A new "unmerged-usr" taint flag has been added that is set whenever
2986 running on systems where /bin/ + /sbin/ are *not* symlinks to their
2987 counterparts in /usr/, i.e. on systems where the /usr/-merge has not
2990 * Generators invoked by PID 1 will now have a couple of useful
2991 environment variables set describing the execution context a
2992 bit. $SYSTEMD_SCOPE encodes whether the generator is called from the
2993 system service manager, or from the per-user service
2994 manager. $SYSTEMD_IN_INITRD encodes whether the generator is invoked
2995 in initrd context or on the host. $SYSTEMD_FIRST_BOOT encodes whether
2996 systemd considers the current boot to be a "first"
2997 boot. $SYSTEMD_VIRTUALIZATION encode whether virtualization is
2998 detected and which type of hypervisor/container
2999 manager. $SYSTEMD_ARCHITECTURE indicates which architecture the
3000 kernel is built for.
3002 * PID 1 will now automatically pick up system credentials from qemu's
3003 fw_cfg interface, thus allowing passing arbitrary data into VM
3004 systems similar to how this is already supported for passing them
3005 into `systemd-nspawn` containers. Credentials may now also be passed
3006 in via the new kernel command line option `systemd.set_credential=`
3007 (note that kernel command line options are world-readable during
3008 runtime, and only useful for credentials that require no
3009 confidentiality). The credentials that can be passed to unified
3010 kernels that use the `systemd-stub` UEFI stub are now similarly
3011 picked up automatically. Automatic importing of system credentials
3012 this way can be turned off via the new
3013 `systemd.import_credentials=no` kernel command line option.
3015 * LoadCredential= will now automatically look for credentials in the
3016 /etc/credstore/, /run/credstore/, /usr/lib/credstore/ directories if
3017 the argument is not an absolute path. Similarly,
3018 LoadCredentialEncrypted= will check the same directories plus
3019 /etc/credstore.encrypted/, /run/credstore.encrypted/ and
3020 /usr/lib/credstore.encrypted/. The idea is to use those directories
3021 as the system-wide location for credentials that services should pick
3024 * System and service credentials are described in great detail in a new
3027 https://systemd.io/CREDENTIALS
3029 Changes in systemd-journald:
3031 * The journal JSON export format has been added to listed of stable
3032 interfaces (https://systemd.io/PORTABILITY_AND_STABILITY/).
3034 * journalctl --list-boots now supports JSON output and the --reverse option.
3036 * Under docs/: JOURNAL_EXPORT_FORMATS was imported from the wiki and
3037 updated, BUILDING_IMAGES is new:
3039 https://systemd.io/JOURNAL_EXPORT_FORMATS
3040 https://systemd.io/BUILDING_IMAGES
3044 * Two new hwdb files have been added. One lists "handhelds" (PDAs,
3045 calculators, etc.), the other AV production devices (DJ tables,
3046 keypads, etc.) that should accessible to the seat owner user by
3049 * udevadm trigger gained a new --prioritized-subsystem= option to
3050 process certain subsystems (and all their parent devices) earlier.
3052 systemd-udev-trigger.service now uses this new option to trigger
3053 block and TPM devices first, hopefully making the boot a bit faster.
3055 * udevadm trigger now implements --type=all, --initialized-match,
3056 --initialized-nomatch to trigger both subsystems and devices, only
3057 already-initialized devices, and only devices which haven't been
3058 initialized yet, respectively.
3060 * udevadm gained a new "wait" command for safely waiting for a specific
3061 device to show up in the udev device database. This is useful in
3062 scripts that asynchronously allocate a block device (e.g. through
3063 repartitioning, or allocating a loopback device or similar) and need
3064 to synchronize on the creation to complete.
3066 * udevadm gained a new "lock" command for locking one or more block
3067 devices while formatting it or writing a partition table to it. It is
3068 an implementation of https://systemd.io/BLOCK_DEVICE_LOCKING and
3069 usable in scripts dealing with block devices.
3071 * udevadm info will show a couple of additional device fields in its
3072 output, and will not apply a limited set of coloring to line types.
3074 * udevadm info --tree will now show a tree of objects (i.e. devices and
3075 suchlike) in the /sys/ hierarchy.
3077 * Block devices will now get a new set of device symlinks in
3078 /dev/disk/by-diskseq/<nr>, which may be used to reference block
3079 device nodes via the kernel's "diskseq" value. Note that this does
3080 not guarantee that opening a device by a symlink like this will
3081 guarantee that the opened device actually matches the specified
3082 diskseq value. To be safe against races, the actual diskseq value of
3083 the opened device (BLKGETDISKSEQ ioctl()) must still be compred with
3084 the one in the symlink path.
3086 * .link files gained support for setting MDI/MID-X on a link.
3088 * .link files gained support for [Match] Firmware= setting to match on
3089 the device firmware description string. By mistake, it was previously
3090 only supported in .network files.
3092 * .link files gained support for [Link] SR-IOVVirtualFunctions= setting
3093 and [SR-IOV] section to configure SR-IOV virtual functions.
3095 Changes in systemd-networkd:
3097 * The default scope for unicast routes configured through [Route]
3098 section is changed to "link", to make the behavior consistent with
3099 "ip route" command. The manual configuration of [Route] Scope= is
3102 * A new unit systemd-networkd-wait-online@<interface>.service has been
3103 added that can be used to wait for a specific network interface to be
3106 * systemd-networkd gained a new [Bridge] Isolated=true|false setting
3107 that configures the eponymous kernel attribute on the bridge.
3109 * .netdev files now can be used to create virtual WLAN devices, and
3110 configure various settings on them, via the [WLAN] section.
3112 * .link/.network files gained support for [Match] Kind= setting to match
3113 on device kind ("bond", "bridge", "gre", "tun", "veth", etc.)
3115 This value is also shown by 'networkctl status'.
3117 * The Local= setting in .netdev files for various virtual network
3118 devices gained support for specifying, in addition to the network
3119 address, the name of a local interface which must have the specified
3122 * systemd-networkd gained a new [Tunnel] External= setting in .netdev
3123 files, to configure tunnels in external mode (a.k.a. collect metadata
3126 * [Network] L2TP= setting was removed. Please use interface specifier in
3127 Local= setting in .netdev files of corresponding L2TP interface.
3129 * New [DHCPServer] BootServerName=, BootServerAddress=, and
3130 BootFilename= settings can be used to configure the server address,
3131 server name, and file name sent in the DHCP packet (e.g. to configure
3134 Changes in systemd-resolved:
3136 * systemd-resolved is started earlier (in sysinit.target), so it
3137 available earlier and will also be started in the initrd if installed
3140 Changes in disk encryption:
3142 * systemd-cryptenroll can now control whether to require the user to
3143 enter a PIN when using TPM-based unlocking of a volume via the new
3144 --tpm2-with-pin= option.
3146 Option tpm2-pin= can be used in /etc/crypttab.
3148 * When unlocking devices via TPM, TPM2 parameter encryption is now
3149 used, to ensure that communication between CPU and discrete TPM chips
3150 cannot be eavesdropped to acquire disk encryption keys.
3152 * A new switch --fido2-credential-algorithm= has been added to
3153 systemd-cryptenroll allowing selection of the credential algorithm to
3154 use when binding encryption to FIDO2 tokens.
3156 Changes in systemd-hostnamed:
3158 * HARDWARE_VENDOR= and HARDWARE_MODEL= can be set in /etc/machine-info
3159 to override the values gleaned from the hwdb.
3161 * A ID_CHASSIS property can be set in the hwdb (for the DMI device
3162 /sys/class/dmi/id) to override the chassis that is reported by
3165 * hostnamed's D-Bus interface gained a new method GetHardwareSerial()
3166 for reading the hardware serial number, as reportd by DMI. It also
3167 exposes a new method D-Bus property FirmwareVersion that encode the
3168 firmware version of the system.
3170 Changes in other components:
3172 * /etc/locale.conf is now populated through tmpfiles.d factory /etc/
3173 handling with the values that were configured during systemd build
3174 (if /etc/locale.conf has not been created through some other
3175 mechanism). This means that /etc/locale.conf should always have
3176 reasonable contents and we avoid a potential mismatch in defaults.
3178 * The userdbctl tool will now show UID range information as part of the
3179 list of known users.
3181 * A new build-time configuration setting default-user-shell= can be
3182 used to set the default shell for user records and nspawn shell
3183 invocations (instead of the default /bin/bash).
3185 * systemd-timesyncd now provides a D-Bus API for receiving NTP server
3186 information dynamically at runtime via IPC.
3188 * The systemd-creds tool gained a new "has-tpm2" verb, which reports
3189 whether a functioning TPM2 infrastructure is available, i.e. if
3190 firmware, kernel driver and systemd all have TPM2 support enabled and
3193 * The systemd-creds tool gained support for generating encrypted
3194 credentials that are using an empty encryption key. While this
3195 provides no integrity nor confidentiality it's useful to implement
3196 codeflows that work the same on TPM-ful and TPM2-less systems. The
3197 service manager will only accept credentials "encrypted" that way if
3198 a TPM2 device cannot be detected, to ensure that credentials
3199 "encrypted" like that cannot be used to trick TPM2 systems.
3201 * When deciding whether to colorize output, all systemd programs now
3202 also check $COLORTERM (in addition to $NO_COLOR, $SYSTEMD_COLORS, and
3205 * Meson's new install_tag feature is now in use for several components,
3206 allowing to build and install select binaries only: pam, nss, devel
3207 (pkg-config files), systemd-boot, libsystemd, libudev. Example:
3208 $ meson build systemd-boot
3209 $ meson install --tags systemd-boot --no-rebuild
3210 https://mesonbuild.com/Installing.html#installation-tags
3212 * A new build configuration option has been added, to allow selecting the
3213 default compression algorithm used by systemd-journald and systemd-coredump.
3214 This allows to build-in support for decompressing all supported formats,
3215 but choose a specific one for compression. E.g.:
3216 $ meson -Ddefault-compression=xz
3218 Experimental features:
3220 * sd-boot gained a new *experimental* setting "reboot-for-bitlocker" in
3221 loader.conf that implements booting Microsoft Windows from the
3222 sd-boot in a way that first reboots the system, to reset the TPM
3223 PCRs. This improves compatibility with BitLocker's TPM use, as the
3224 PCRs will only record the Windows boot process, and not sd-boot
3225 itself, thus retaining the PCR measurements not involving sd-boot.
3226 Note that this feature is experimental for now, and is likely going
3227 to be generalized and renamed in a future release, without retaining
3228 compatibility with the current implementation.
3230 * A new systemd-sysupdate component has been added that automatically
3231 discovers, downloads, and installs A/B-style updates for the host
3232 installation itself, or container images, portable service images,
3233 and other assets. See the new systemd-sysupdate man page for updates.
3235 Contributions from: 4piu, Adam Williamson, adrian5, Albert Brox,
3236 AlexCatze, Alex Henrie, Alfonso Sánchez-Beato, Alice S,
3237 Alvin Šipraga, amarjargal, Amarjargal, Andrea Pappacoda,
3238 Andreas Rammhold, Andy Chi, Anita Zhang, Antonio Alvarez Feijoo,
3239 Arfrever Frehtes Taifersar Arahesis, ash, Bastien Nocera, Be,
3240 bearhoney, Ben Efros, Benjamin Berg, Benjamin Franzke,
3241 Brett Holman, Christian Brauner, Clyde Byrd III, Curtis Klein,
3242 Daan De Meyer, Daniele Medri, Daniel Mack, Danilo Krummrich,
3243 David, David Bond, Davide Cavalca, David Tardon, davijosw,
3244 dependabot[bot], Donald Chan, Dorian Clay, Eduard Tolosa,
3245 Elias Probst, Eli Schwartz, Erik Sjölund, Evgeny Vereshchagin,
3246 Federico Ceratto, Franck Bui, Frantisek Sumsal, Gaël PORTAY,
3247 Georges Basile Stavracas Neto, Gibeom Gwon, Goffredo Baroncelli,
3248 Grigori Goronzy, Hans de Goede, Heiko Becker, Hugo Carvalho,
3249 Jakob Lell, James Hilliard, Jan Janssen, Jason A. Donenfeld,
3250 Joan Bruguera, Joerie de Gram, Josh Triplett, Julia Kartseva,
3251 Kazuo Moriwaka, Khem Raj, ksa678491784, Lance, Lan Tian,
3252 Laura Barcziova, Lennart Poettering, Leviticoh, licunlong,
3253 Lidong Zhong, lincoln auster, Lubomir Rintel, Luca Boccassi,
3254 Luca BRUNO, lucagoc, Ludwig Nussel, Marcel Hellwig, march1993,
3255 Marco Scardovi, Mario Limonciello, Mariusz Tkaczyk,
3256 Markus Weippert, Martin, Martin Liska, Martin Wilck, Matija Skala,
3257 Matthew Blythe, Matthias Lisin, Matthijs van Duin, Matt Walton,
3258 Max Gautier, Michael Biebl, Michael Olbrich, Michal Koutný,
3259 Michal Sekletár, Mike Gilbert, MkfsSion, Morten Linderud,
3260 Nick Rosbrook, Nikolai Grigoriev, Nikolai Kostrigin,
3261 Nishal Kulkarni, Noel Kuntze, Pablo Ceballos, Peter Hutterer,
3262 Peter Morrow, Pigmy-penguin, Piotr Drąg, prumian, Richard Neill,
3263 Rike-Benjamin Schuppner, rodin-ia, Romain Naour, Ruben Kerkhof,
3264 Ryan Hendrickson, Santa Wiryaman, Sebastian Pucilowski, Seth Falco,
3265 Simon Ellmann, Sonali Srivastava, Stefan Seering,
3266 Stephen Hemminger, tawefogo, techtino, Temuri Doghonadze,
3267 Thomas Batten, Thomas Haller, Thomas Weißschuh, Tobias Stoeckmann,
3268 Tomasz Pala, Tyson Whitehead, Vishal Chillara Srinivas,
3269 Vivien Didelot, w30023233, wangyuhang, Weblate, Xiaotian Wu,
3270 yangmingtai, YmrDtnJu, Yonathan Randolph, Yutsuten, Yu Watanabe,
3271 Zbigniew Jędrzejewski-Szmek, наб
3273 — Edinburgh, 2022-05-21
3277 * Support for encrypted and authenticated credentials has been added.
3278 This extends the credential logic introduced with v247 to support
3279 non-interactive symmetric encryption and authentication, based on a
3280 key that is stored on the /var/ file system or in the TPM2 chip (if
3281 available), or the combination of both (by default if a TPM2 chip
3282 exists the combination is used, otherwise the /var/ key only). The
3283 credentials are automatically decrypted at the moment a service is
3284 started, and are made accessible to the service itself in unencrypted
3285 form. A new tool 'systemd-creds' encrypts credentials for this
3286 purpose, and two new service file settings LoadCredentialEncrypted=
3287 and SetCredentialEncrypted= configure such credentials.
3289 This feature is useful to store sensitive material such as SSL
3290 certificates, passwords and similar securely at rest and only decrypt
3291 them when needed, and in a way that is tied to the local OS
3292 installation or hardware.
3294 * systemd-gpt-auto-generator can now automatically set up discoverable
3295 LUKS2 encrypted swap partitions.
3297 * The GPT Discoverable Partitions Specification has been substantially
3298 extended with support for root and /usr/ partitions for the majority
3299 of architectures systemd supports. This includes platforms that do
3300 not natively support UEFI, because even though GPT is specified under
3301 UEFI umbrella, it is useful on other systems too. Specifically,
3302 systemd-nspawn, systemd-sysext, systemd-gpt-auto-generator and
3303 Portable Services use the concept without requiring UEFI.
3305 * The GPT Discoverable Partitions Specifications has been extended with
3306 a new set of partitions that may carry PKCS#7 signatures for Verity
3307 partitions, encoded in a simple JSON format. This implements a simple
3308 mechanism for building disk images that are fully authenticated and
3309 can be tested against a set of cryptographic certificates. This is
3310 now implemented for the various systemd tools that can operate with
3311 disk images, such as systemd-nspawn, systemd-sysext, systemd-dissect,
3312 Portable services/RootImage=, systemd-tmpfiles, and systemd-sysusers.
3313 The PKCS#7 signatures are passed to the kernel (where they are
3314 checked against certificates from the kernel keyring), or can be
3315 verified against certificates provided in userspace (via a simple
3316 drop-in file mechanism).
3318 * systemd-dissect's inspection logic will now report for which uses a
3319 disk image is intended. Specifically, it will display whether an
3320 image is suitable for booting on UEFI or in a container (using
3321 systemd-nspawn's --image= switch), whether it can be used as portable
3322 service, or attached as system extension.
3324 * The system-extension.d/ drop-in files now support a new field
3325 SYSEXT_SCOPE= that may encode which purpose a system extension image
3326 is for: one of "initrd", "system" or "portable". This is useful to
3327 make images more self-descriptive, and to ensure system extensions
3328 cannot be attached in the wrong contexts.
3330 * The os-release file learnt a new PORTABLE_PREFIXES= field which may
3331 be used in portable service images to indicate which unit prefixes
3334 * The GPT image dissection logic in systemd-nspawn/systemd-dissect/…
3335 now is able to decode images for non-native architectures as well.
3336 This allows systemd-nspawn to boot images of non-native architectures
3337 if the corresponding user mode emulator is installed and
3338 systemd-binfmtd is running.
3340 * systemd-logind gained new settings HandlePowerKeyLongPress=,
3341 HandleRebootKeyLongPress=, HandleSuspendKeyLongPress= and
3342 HandleHibernateKeyLongPress= which may be used to configure actions
3343 when the relevant keys are pressed for more than 5s. This is useful
3344 on devices that only have hardware for a subset of these keys. By
3345 default, if the reboot key is pressed long the poweroff operation is
3346 now triggered, and when the suspend key is pressed long the hibernate
3347 operation is triggered. Long pressing the other two keys currently
3348 does not trigger any operation by default.
3350 * When showing unit status updates on the console during boot and
3351 shutdown, and a service is slow to start so that the cylon animation
3352 is shown, the most recent sd_notify() STATUS= text is now shown as
3353 well. Services may use this to make the boot/shutdown output easier
3354 to understand, and to indicate what precisely a service that is slow
3355 to start or stop is waiting for. In particular, the per-user service
3356 manager instance now reports what it is doing and which service it is
3357 waiting for this way to the system service manager.
3359 * The service manager will now re-execute on reception of the
3360 SIGRTMIN+25 signal. It previously already did that on SIGTERM — but
3361 only when running as PID 1. There was no signal to request this when
3362 running as per-user service manager, i.e. as any other PID than 1.
3363 SIGRTMIN+25 works for both system and user managers.
3365 * The hardware watchdog logic in PID 1 gained support for operating
3366 with the default timeout configured in the hardware, instead of
3367 insisting on re-configuring it. Set RuntimeWatchdogSec=default to
3368 request this behavior.
3370 * A new kernel command line option systemd.watchdog_sec= is now
3371 understood which may be used to override the hardware watchdog
3372 time-out for the boot.
3374 * A new setting DefaultOOMScoreAdjust= is now supported in
3375 /etc/systemd/system.conf and /etc/systemd/user.conf. It may be used
3376 to set the default process OOM score adjustment value for processes
3377 started by the service manager. For per-user service managers this
3378 now defaults to 100, but for per-system service managers is left as
3379 is. This means that by default now services forked off the user
3380 service manager are more likely to be killed by the OOM killer than
3381 system services or the managers themselves.
3383 * A new per-service setting RestrictFileSystems= as been added that
3384 restricts the file systems a service has access to by their type.
3385 This is based on the new BPF LSM of the Linux kernel. It provides an
3386 effective way to make certain API file systems unavailable to
3387 services (and thus minimizing attack surface). A new command
3388 "systemd-analyze filesystems" has been added that lists all known
3389 file system types (and how they are grouped together under useful
3392 * Services now support a new setting RestrictNetworkInterfaces= for
3393 restricting access to specific network interfaces.
3395 * Service unit files gained new settings StartupAllowedCPUs= and
3396 StartupAllowedMemoryNodes=. These are similar to their counterparts
3397 without the "Startup" prefix and apply during the boot process
3398 only. This is useful to improve boot-time behavior of the system and
3399 assign resources differently during boot than during regular
3400 runtime. This is similar to the preexisting StartupCPUWeight=
3403 * Related to this: the various StartupXYZ= settings
3404 (i.e. StartupCPUWeight=, StartupAllowedCPUs=, …) are now also applied
3405 during shutdown. The settings not prefixed with "Startup" hence apply
3406 during regular runtime, and those that are prefixed like that apply
3407 during boot and shutdown.
3409 * A new per-unit set of conditions/asserts
3410 [Condition|Assert][Memory|CPU|IO]Pressure= have been added to make a
3411 unit skip/fail activation if the system's (or a slice's) memory/cpu/io
3412 pressure is above the configured threshold, using the kernel PSI
3413 feature. For more details see systemd.unit(5) and
3414 https://docs.kernel.org/accounting/psi.html
3416 * The combination of ProcSubset=pid and ProtectKernelTunables=yes and/or
3417 ProtectKernelLogs=yes can now be used.
3419 * The default maximum numbers of inodes have been raised from 64k to 1M
3420 for /dev/, and from 400k to 1M for /tmp/.
3422 * The per-user service manager learnt support for communicating with
3423 systemd-oomd to acquire OOM kill information.
3425 * A new service setting ExecSearchPath= has been added that allows
3426 changing the search path for executables for services. It affects
3427 where we look for the binaries specified in ExecStart= and similar,
3428 and the specified directories are also added the $PATH environment
3429 variable passed to invoked processes.
3431 * A new setting RuntimeRandomizedExtraSec= has been added for service
3432 and scope units that allows extending the runtime time-out as
3433 configured by RuntimeMaxSec= with a randomized amount.
3435 * The syntax of the service unit settings RuntimeDirectory=,
3436 StateDirectory=, CacheDirectory=, LogsDirectory= has been extended:
3437 if the specified value is now suffixed with a colon, followed by
3438 another filename, the latter will be created as symbolic link to the
3439 specified directory. This allows creating these service directories
3440 together with alias symlinks to make them available under multiple
3443 * Service unit files gained two new settings TTYRows=/TTYColumns= for
3444 configuring rows/columns of the TTY device passed to
3445 stdin/stdout/stderr of the service. This is useful to propagate TTY
3446 dimensions to a virtual machine.
3448 * A new service unit file setting ExitType= has been added that
3449 specifies when to assume a service has exited. By default systemd
3450 only watches the main process of a service. By setting
3451 ExitType=cgroup it can be told to wait for the last process in a
3454 * Automount unit files gained a new setting ExtraOptions= that can be
3455 used to configure additional mount options to pass to the kernel when
3456 mounting the autofs instance.
3458 * "Urlification" (generation of ESC sequences that generate clickable
3459 hyperlinks in modern terminals) may now be turned off altogether
3462 * Path units gained new TriggerLimitBurst= and TriggerLimitIntervalSec=
3463 settings that default to 200 and 2 s respectively. The ratelimit
3464 ensures that a path unit cannot cause PID1 to busy-loop when it is
3465 trying to trigger a service that is skipped because of a Condition*=
3466 not being satisfied. This matches the configuration and behaviour of
3469 * The TPM2/FIDO2/PKCS11 support in systemd-cryptsetup is now also built
3470 as a plug-in for cryptsetup. This means the plain cryptsetup command
3471 may now be used to unlock volumes set up this way.
3473 * The TPM2 logic in cryptsetup will now automatically detect systems
3474 where the TPM2 chip advertises SHA256 PCR banks but the firmware only
3475 updates the SHA1 banks. In such a case PCR policies will be
3476 automatically bound to the latter, not the former. This makes the PCR
3477 policies reliable, but of course do not provide the same level of
3478 trust as SHA256 banks.
3480 * The TPM2 logic in systemd-cryptsetup/systemd-cryptsetup now supports
3481 RSA primary keys in addition to ECC, improving compatibility with
3482 TPM2 chips that do not support ECC. RSA keys are much slower to use
3483 than ECC, and hence are only used if ECC is not available.
3485 * /etc/crypttab gained support for a new token-timeout= setting for
3486 encrypted volumes that allows configuration of the maximum time to
3487 wait for PKCS#11/FIDO2 tokens to be plugged in. If the time elapses
3488 the logic will query the user for a regular passphrase/recovery key
3491 * Support for activating dm-integrity volumes at boot via a new file
3492 /etc/integritytab and the tool systemd-integritysetup have been
3493 added. This is similar to /etc/crypttab and /etc/veritytab, but deals
3494 with dm-integrity instead of dm-crypt/dm-verity.
3496 * The systemd-veritysetup-generator now understands a new usrhash=
3497 kernel command line option for specifying the Verity root hash for
3498 the partition backing the /usr/ file system. A matching set of
3499 systemd.verity_usr_* kernel command line options has been added as
3500 well. These all work similar to the corresponding options for the
3503 * The sd-device API gained a new API call sd_device_get_diskseq() to
3504 return the DISKSEQ property of a device structure. The "disk
3505 sequence" concept is a new feature recently introduced to the Linux
3506 kernel that allows detecting reuse cycles of block devices, i.e. can
3507 be used to recognize when loopback block devices are reused for a
3508 different purpose or CD-ROM drives get their media changed.
3510 * A new unit systemd-boot-update.service has been added. If enabled
3511 (the default) and the sd-boot loader is detected to be installed, it
3512 is automatically updated to the newest version when out of date. This
3513 is useful to ensure the boot loader remains up-to-date, and updates
3514 automatically propagate from the OS tree in /usr/.
3516 * sd-boot will now build with SBAT by default in order to facilitate
3517 working with recent versions of Shim that require it to be present.
3519 * sd-boot can now parse Microsoft Windows' Boot Configuration Data.
3520 This is used to robustly generate boot entry titles for Windows.
3522 * A new generic target unit factory-reset.target has been added. It is
3523 hooked into systemd-logind similar in fashion to
3524 reboot/poweroff/suspend/hibernate, and is supposed to be used to
3525 initiate a factory reset operation. What precisely this operation
3526 entails is up for the implementer to decide, the primary goal of the
3527 new unit is provide a framework where to plug in the implementation
3528 and how to trigger it.
3530 * A new meson build-time option 'clock-valid-range-usec-max' has been
3531 added which takes a time in µs and defaults to 15 years. If the RTC
3532 time is noticed to be more than the specified time ahead of the
3533 built-in epoch of systemd (which by default is the release timestamp
3534 of systemd) it is assumed that the RTC is not working correctly, and
3535 the RTC is reset to the epoch. (It already is reset to the epoch when
3536 noticed to be before it.) This should increase the chance that time
3537 doesn't accidentally jump too far ahead due to faulty hardware or
3540 * A new setting SaveIntervalSec= has been added to systemd-timesyncd,
3541 which may be used to automatically save the current system time to
3542 disk in regular intervals. This is useful to maintain a roughly
3543 monotonic clock even without RTC hardware and with some robustness
3544 against abnormal system shutdown.
3546 * systemd-analyze verify gained support for a pair of new --image= +
3547 --root= switches for verifying units below a specific root
3548 directory/image instead of on the host.
3550 * systemd-analyze verify gained support for verifying unit files under
3551 an explicitly specified unit name, independently of what the filename
3554 * systemd-analyze verify gained a new switch --recursive-errors= which
3555 controls whether to only fail on errors found in the specified units
3556 or recursively any dependent units.
3558 * systemd-analyze security now supports a new --offline mode for
3559 analyzing unit files stored on disk instead of loaded units. It may
3560 be combined with --root=/--image to analyze unit files under a root
3561 directory or disk image. It also learnt a new --threshold= parameter
3562 for specifying an exposure level threshold: if the exposure level
3563 exceeds the specified value the call will fail. It also gained a new
3564 --security-policy= switch for configuring security policies to
3565 enforce on the units. A policy is a JSON file that lists which tests
3566 shall be weighted how much to determine the overall exposure
3567 level. Altogether these new features are useful for fully automatic
3568 analysis and enforcement of security policies on unit files.
3570 * systemd-analyze security gain a new --json= switch for JSON output.
3572 * systemd-analyze learnt a new --quiet switch for reducing
3573 non-essential output. It's honored by the "dot", "syscall-filter",
3574 "filesystems" commands.
3576 * systemd-analyze security gained a --profile= option that can be used
3577 to take into account a portable profile when analyzing portable
3578 services, since a lot of the security-related settings are enabled
3581 * systemd-analyze learnt a new inspect-elf verb that parses ELF core
3582 files, binaries and executables and prints metadata information,
3583 including the build-id and other info described on:
3584 https://systemd.io/COREDUMP_PACKAGE_METADATA/
3586 * .network files gained a new UplinkInterface= in the [IPv6SendRA]
3587 section, for automatically propagating DNS settings from other
3590 * The static lease DHCP server logic in systemd-networkd may now serve
3591 IP addresses outside of the configured IP pool range for the server.
3593 * CAN support in systemd-networkd gained four new settings Loopback=,
3594 OneShot=, PresumeAck=, ClassicDataLengthCode= for tweaking CAN
3595 control modes. It gained a number of further settings for tweaking
3598 * The [CAN] section in .network file gained new TimeQuantaNSec=,
3599 PropagationSegment=, PhaseBufferSegment1=, PhaseBufferSegment2=,
3600 SyncJumpWidth=, DataTimeQuantaNSec=, DataPropagationSegment=,
3601 DataPhaseBufferSegment1=, DataPhaseBufferSegment2=, and
3602 DataSyncJumpWidth= settings to control bit-timing processed by the
3605 * DHCPv4 client support in systemd-networkd learnt a new Label= option
3606 for configuring the address label to apply to configure IPv4
3609 * The [IPv6AcceptRA] section of .network files gained support for a new
3610 UseMTU= setting that may be used to control whether to apply the
3611 announced MTU settings to the local interface.
3613 * The [DHCPv4] section in .network file gained a new Use6RD= boolean
3614 setting to control whether the DHCPv4 client request and process the
3617 * The [DHCPv6PrefixDelegation] section in .network file is renamed to
3618 [DHCPPrefixDelegation], as now the prefix delegation is also supported
3619 with DHCPv4 protocol by enabling the Use6RD= setting.
3621 * The [DHCPPrefixDelegation] section in .network file gained a new
3622 setting UplinkInterface= to specify the upstream interface.
3624 * The [DHCPv6] section in .network file gained a new setting
3625 UseDelegatedPrefix= to control whether the delegated prefixes will be
3626 propagated to the downstream interfaces.
3628 * The [IPv6AcceptRA] section of .network files now understands two new
3629 settings UseGateway=/UseRoutePrefix= for explicitly configuring
3630 whether to use the relevant fields from the IPv6 Router Advertisement
3633 * The ForceDHCPv6PDOtherInformation= setting in the [DHCPv6] section
3634 has been removed. Please use the WithoutRA= and UseDelegatedPrefix=
3635 settings in the [DHCPv6] section and the DHCPv6Client= setting in the
3636 [IPv6AcceptRA] section to control when the DHCPv6 client is started
3637 and how the delegated prefixes are handled by the DHCPv6 client.
3639 * The IPv6Token= section in the [Network] section is deprecated, and
3640 the [IPv6AcceptRA] section gained the Token= setting for its
3641 replacement. The [IPv6Prefix] section also gained the Token= setting.
3642 The Token= setting gained 'eui64' mode to explicitly configure an
3643 address with the EUI64 algorithm based on the interface MAC address.
3644 The 'prefixstable' mode can now optionally take a secret key. The
3645 Token= setting in the [DHCPPrefixDelegation] section now supports all
3646 algorithms supported by the same settings in the other sections.
3648 * The [RoutingPolicyRule] section of .network file gained a new
3649 SuppressInterfaceGroup= setting.
3651 * The IgnoreCarrierLoss= setting in the [Network] section of .network
3652 files now allows a duration to be specified, controlling how long to
3653 wait before reacting to carrier loss.
3655 * The [DHCPServer] section of .network file gained a new Router=
3656 setting to specify the router address.
3658 * The [CAKE] section of .network files gained various new settings
3659 AutoRateIngress=, CompensationMode=, FlowIsolationMode=, NAT=,
3660 MPUBytes=, PriorityQueueingPreset=, FirewallMark=, Wash=, SplitGSO=,
3661 and UseRawPacketSize= for configuring CAKE.
3663 * systemd-networkd now ships with new default .network files:
3664 80-container-vb.network which matches host-side network bridge device
3665 created by systemd-nspawn's --network-bridge or --network-zone
3666 switch, and 80-6rd-tunnel.network which matches automatically created
3667 sit tunnel with 6rd prefix when the DHCP 6RD option is received.
3669 * systemd-networkd's handling of Endpoint= resolution for WireGuard
3670 interfaces has been improved.
3672 * systemd-networkd will now automatically configure routes to addresses
3673 specified in AllowedIPs=. This feature can be controlled via
3674 RouteTable= and RouteMetric= settings in [WireGuard] or
3675 [WireGuardPeer] sections.
3677 * systemd-networkd will now once again automatically generate persistent
3678 MAC addresses for batadv and bridge interfaces. Users can disable this
3679 by using MACAddress=none in .netdev files.
3681 * systemd-networkd and systemd-udevd now support IP over InfiniBand
3682 interfaces. The Kind= setting in .netdev file accepts "ipoib". And
3683 systemd.netdev files gained the [IPoIB] section.
3685 * systemd-networkd and systemd-udevd now support net.ifname-policy=
3686 option on the kernel command-line. This is implemented through the
3687 systemd-network-generator service that automatically generates
3688 appropriate .link, .network, and .netdev files.
3690 * The various systemd-udevd "ethtool" buffer settings now understand
3691 the special value "max" to configure the buffers to the maximum the
3694 * systemd-udevd's .link files may now configure a large variety of
3695 NIC coalescing settings, plus more hardware offload settings.
3697 * .link files gained a new WakeOnLanPassword= setting in the [Link]
3698 section that allows to specify a WoL "SecureOn" password on hardware
3701 * systemd-nspawn's --setenv= switch now supports an additional syntax:
3702 if only a variable name is specified (i.e. without being suffixed by
3703 a '=' character and a value) the current value of the environment
3704 variable is propagated to the container. e.g. --setenv=FOO will
3705 lookup the current value of $FOO in the environment, and pass it down
3706 to the container. Similar behavior has been added to homectl's,
3707 machinectl's and systemd-run's --setenv= switch.
3709 * systemd-nspawn gained a new switch --suppress-sync= which may be used
3710 to optionally suppress the effect of the sync()/fsync()/fdatasync()
3711 system calls for the container payload. This is useful for build
3712 system environments where safety against abnormal system shutdown is
3713 not essential as all build artifacts can be regenerated any time, but
3714 the performance win is beneficial.
3716 * systemd-nspawn will now raise the RLIMIT_NOFILE hard limit to the
3717 same value that PID 1 uses for most forked off processes.
3719 * systemd-nspawn's --bind=/--bind-ro= switches now optionally take
3720 uidmap/nouidmap options as last parameter. If "uidmap" is used the
3721 bind mounts are created with UID mapping taking place that ensures
3722 the host's file ownerships are mapped 1:1 to container file
3723 ownerships, even if user namespacing is used. This way
3724 files/directories bound into containers will no longer show up as
3725 owned by the nobody user as they typically did if no special care was
3726 taken to shift them manually.
3728 * When discovering Windows installations sd-boot will now attempt to
3729 show the Windows version.
3731 * The color scheme to use in sd-boot may now be configured at
3734 * sd-boot gained the ability to change screen resolution during
3735 boot-time, by hitting the "r" key. This will cycle through available
3736 resolutions and save the last selection.
3738 * sd-boot learnt a new hotkey "f". When pressed the system will enter
3739 firmware setup. This is useful in environments where it is difficult
3740 to hit the right keys early enough to enter the firmware, and works
3741 on any firmware regardless which key it natively uses.
3743 * sd-boot gained support for automatically booting into the menu item
3744 selected on the last boot (using the "@saved" identifier for menu
3747 * sd-boot gained support for automatically loading all EFI drivers
3748 placed in the /EFI/systemd/drivers/ subdirectory of the EFI System
3749 Partition (ESP). These drivers are loaded before the menu entries are
3750 loaded. This is useful e.g. to load additional file system drivers
3751 for the XBOOTLDR partition.
3753 * systemd-boot will now paint the input cursor on its own instead of
3754 relying on the firmware to do so, increasing compatibility with broken
3755 firmware that doesn't make the cursor reasonably visible.
3757 * sd-boot now embeds a .osrel PE section like we expect from Boot
3758 Loader Specification Type #2 Unified Kernels. This means sd-boot
3759 itself may be used in place of a Type #2 Unified Kernel. This is
3760 useful for debugging purposes as it allows chain-loading one a
3761 (development) sd-boot instance from another.
3763 * sd-boot now supports a new "devicetree" field in Boot Loader
3764 Specification Type #1 entries: if configured the specified device
3765 tree file is installed before the kernel is invoked. This is useful
3766 for installing/applying new devicetree files without updating the
3769 * Similarly, sd-stub now can read devicetree data from a PE section
3770 ".dtb" and apply it before invoking the kernel.
3772 * sd-stub (the EFI stub that can be glued in front of a Linux kernel)
3773 gained the ability to pick up credentials and sysext files, wrap them
3774 in a cpio archive, and pass as an additional initrd to the invoked
3775 Linux kernel, in effect placing those files in the /.extra/ directory
3776 of the initrd environment. This is useful to implement trusted initrd
3777 environments which are fully authenticated but still can be extended
3778 (via sysexts) and parameterized (via encrypted/authenticated
3779 credentials, see above).
3781 Credentials can be located next to the kernel image file (credentials
3782 specific to a single boot entry), or in one of the shared directories
3783 (credentials applicable to multiple boot entries).
3785 * sd-stub now comes with a full man page, that explains its feature set
3786 and how to combine a kernel image, an initrd and the stub to build a
3787 complete EFI unified kernel image, implementing Boot Loader
3788 Specification Type #2.
3790 * sd-stub may now provide the initrd to the executed kernel via the
3791 LINUX_EFI_INITRD_MEDIA_GUID EFI protocol, adding compatibility for
3792 non-x86 architectures.
3794 * bootctl learnt new set-timeout and set-timeout-oneshot commands that
3795 may be used to set the boot menu time-out of the boot loader (for all
3796 or just the subsequent boot).
3798 * bootctl and kernel-install will now read variables
3799 KERNEL_INSTALL_LAYOUT= from /etc/machine-info and layout= from
3800 /etc/kernel/install.conf. When set, it specifies the layout to use
3801 for installation directories on the boot partition, so that tools
3802 don't need to guess it based on the already-existing directories. The
3803 only value that is defined natively is "bls", corresponding to the
3805 https://systemd.io/BOOT_LOADER_SPECIFICATION/. Plugins for
3806 kernel-install that implement a different layout can declare other
3807 values for this variable.
3809 'bootctl install' will now write KERNEL_INSTALL_LAYOUT=bls, on the
3810 assumption that if the user installed sd-boot to the ESP, they intend
3811 to use the entry layout understood by sd-boot. It'll also write
3812 KERNEL_INSTALL_MACHINE_ID= if it creates any directories using the ID
3813 (and it wasn't specified in the config file yet). Similarly,
3814 kernel-install will now write KERNEL_INSTALL_MACHINE_ID= (if it
3815 wasn't specified in the config file yet). Effectively, those changes
3816 mean that the machine-id used for boot loader entry installation is
3817 "frozen" upon first use and becomes independent of the actual
3820 Configuring KERNEL_INSTALL_MACHINE_ID fixes the following problem:
3821 images created for distribution ("golden images") are built with no
3822 machine-id, so that a unique machine-id can be created on the first
3823 boot. But those images may contain boot loader entries with the
3824 machine-id used during build included in paths. Using a "frozen"
3825 value allows unambiguously identifying entries that match the
3826 specific installation, while still permitting parallel installations
3829 Configuring KERNEL_INSTALL_LAYOUT obviates the need for
3830 kernel-install to guess the installation layout. This fixes the
3831 problem where a (possibly empty) directory in the boot partition is
3832 created from a different layout causing kernel-install plugins to
3833 assume the wrong layout. A particular example of how this may happen
3834 is the grub2 package in Fedora which includes directories under /boot
3835 directly in its file list. Various other packages pull in grub2 as a
3836 dependency, so it may be installed even if unused, breaking
3837 installations that use the bls layout.
3839 * bootctl and systemd-bless-boot can now be linked statically.
3841 * systemd-sysext now optionally doesn't insist on extension-release.d/
3842 files being placed in the image under the image's file name. If the
3843 file system xattr user.extension-release.strict is set on the
3844 extension release file, it is accepted regardless of its name. This
3845 relaxes security restrictions a bit, as system extension may be
3846 attached under a wrong name this way.
3848 * udevadm's test-builtin command learnt a new --action= switch for
3849 testing the built-in with the specified action (in place of the
3852 * udevadm info gained new switches --property=/--value for showing only
3853 specific udev properties/values instead of all.
3855 * A new hwdb database has been added that contains matches for various
3856 types of signal analyzers (protocol analyzers, logic analyzers,
3857 oscilloscopes, multimeters, bench power supplies, etc.) that should
3858 be accessible to regular users.
3860 * A new hwdb database entry has been added that carries information
3861 about types of cameras (regular or infrared), and in which direction
3862 they point (front or back).
3864 * A new rule to allow console users access to rfkill by default has been
3867 * Device nodes for the Software Guard eXtension enclaves (sgx_vepc) are
3868 now also owned by the system group "sgx".
3870 * A new build-time meson option "extra-net-naming-schemes=" has been
3871 added to define additional naming schemes for udev's network
3872 interface naming logic. This is useful for enterprise distributions
3873 and similar which want to pin the schemes of certain distribution
3874 releases under a specific name and previously had to patch the
3875 sources to introduce new named schemes.
3877 * The predictable naming logic for network interfaces has been extended
3878 to generate stable names from Xen netfront device information.
3880 * hostnamed's chassis property can now be sourced from chassis-type
3881 field encoded in devicetree (in addition to the existing DMI
3884 * systemd-cgls now optionally displays cgroup IDs and extended
3885 attributes for each cgroup. (Controllable via the new --xattr= +
3886 --cgroup-id= switches.)
3888 * coredumpctl gained a new --all switch for operating on all
3889 Journal files instead of just the local ones.
3891 * systemd-coredump will now use libdw/libelf via dlopen() rather than
3892 directly linking, allowing users to easily opt-out of backtrace/metadata
3893 analysis of core files, and reduce image sizes when this is not needed.
3895 * systemd-coredump will now analyze core files with libdw/libelf in a
3896 forked, sandboxed process.
3898 * systemd-homed will now try to unmount an activate home area in
3899 regular intervals once the user logged out fully. Previously this was
3900 attempted exactly once but if the home directory was busy for some
3901 reason it was not tried again.
3903 * systemd-homed's LUKS2 home area backend will now create a BSD file
3904 system lock on the image file while the home area is active
3905 (i.e. mounted). If a home area is found to be locked, logins are
3906 politely refused. This should improve behavior when using home areas
3907 images that are accessible via the network from multiple clients, and
3908 reduce the chance of accidental file system corruption in that case.
3910 * Optionally, systemd-homed will now drop the kernel buffer cache once
3911 a user has fully logged out, configurable via the new --drop-caches=
3914 * systemd-homed now makes use of UID mapped mounts for the home areas.
3915 If the kernel and used file system support it, files are now
3916 internally owned by the "nobody" user (i.e. the user typically used
3917 for indicating "this ownership is not mapped"), and dynamically
3918 mapped to the UID used locally on the system via the UID mapping
3919 mount logic of recent kernels. This makes migrating home areas
3920 between different systems cheaper because recursively chown()ing file
3921 system trees is no longer necessary.
3923 * systemd-homed's CIFS backend now optionally supports CIFS service
3924 names with a directory suffix, in order to place home directories in
3925 a subdirectory of a CIFS share, instead of the top-level directory.
3927 * systemd-homed's CIFS backend gained support for specifying additional
3928 mount options in the JSON user record (cifsExtraMountOptions field,
3929 and --cifs-extra-mount-options= homectl switch). This is for example
3930 useful for configuring mount options such as "noserverino" that some
3931 SMB3 services require (use that to run a homed home directory from a
3932 FritzBox SMB3 share this way).
3934 * systemd-homed will now default to btrfs' zstd compression for home
3935 areas. This is inspired by Fedora's recent decision to switch to zstd
3938 * Additional mount options to use when mounting the file system of
3939 LUKS2 volumes in systemd-homed has been added. Via the
3940 $SYSTEMD_HOME_MOUNT_OPTIONS_BTRFS, $SYSTEMD_HOME_MOUNT_OPTIONS_EXT4,
3941 $SYSTEMD_HOME_MOUNT_OPTIONS_XFS environment variables to
3942 systemd-homed or via the luksExtraMountOptions user record JSON
3943 property. (Exposed via homectl --luks-extra-mount-options)
3945 * homectl's resize command now takes the special size specifications
3946 "min" and "max" to shrink/grow the home area to the minimum/maximum
3947 size possible, taking disk usage/space constraints and file system
3948 limitations into account. Resizing is now generally graceful: the
3949 logic will try to get as close to the specified size as possible, but
3950 not consider it a failure if the request couldn't be fulfilled
3953 * systemd-homed gained the ability to automatically shrink home areas
3954 on logout to their minimal size and grow them again on next
3955 login. This ensures that while inactive, a home area only takes up
3956 the minimal space necessary, but once activated, it provides
3957 sufficient space for the user's needs. This behavior is only
3958 supported if btrfs is used as file system inside the home area
3959 (because only for btrfs online growing/shrinking is implemented in
3960 the kernel). This behavior is now enabled by default, but may be
3961 controlled via the new --auto-resize-mode= setting of homectl.
3963 * systemd-homed gained support for automatically re-balancing free disk
3964 space among active home areas, in case the LUKS2 backends are used,
3965 and no explicit disk size was requested. This way disk space is
3966 automatically managed and home areas resized in regular intervals and
3967 manual resizing when disk space becomes scarce should not be
3968 necessary anymore. This behavior is only supported if btrfs is used
3969 within the home areas (as only then online shrinking and growing is
3970 supported), and may be configured via the new rebalanceWeight JSON
3971 user record field (as exposed via the new --rebalance-weight= homectl
3972 setting). Re-balancing is mostly automatic, but can also be requested
3973 explicitly via "homectl rebalance", which is synchronous, and thus
3974 may be used to wait until the rebalance run is complete.
3976 * userdbctl gained a --json= switch for configured the JSON formatting
3977 to use when outputting user or group records.
3979 * userdbctl gained a new --multiplexer= switch for explicitly
3980 configuring whether to use the systemd-userdbd server side user
3981 record resolution logic.
3983 * userdbctl's ssh-authorized-keys command learnt a new --chain switch,
3984 for chaining up another command to execute after completing the
3985 look-up. Since the OpenSSH's AuthorizedKeysCommand only allows
3986 configuration of a single command to invoke, this maybe used to
3987 invoke multiple: first userdbctl's own implementation, and then any
3988 other also configured in the command line.
3990 * The sd-event API gained a new function sd_event_add_inotify_fd() that
3991 is similar to sd_event_add_inotify() but accepts a file descriptor
3992 instead of a path in the file system for referencing the inode to
3995 * The sd-event API gained a new function
3996 sd_event_source_set_ratelimit_expire_callback() that may be used to
3997 define a callback function that is called whenever an event source
3998 leaves the rate limiting phase.
4000 * New documentation has been added explaining which steps are necessary
4001 to port systemd to a new architecture:
4003 https://systemd.io/PORTING_TO_NEW_ARCHITECTURES
4005 * The x-systemd.makefs option in /etc/fstab now explicitly supports
4006 ext2, ext3, and f2fs file systems.
4008 * Mount units and units generated from /etc/fstab entries with 'noauto'
4009 are now ordered the same as other units. Effectively, they will be
4010 started earlier (if something actually pulled them in) and stopped
4011 later, similarly to normal mount units that are part of
4012 fs-local.target. This change should be invisible to users, but
4013 should prevent those units from being stopped too early during
4016 * The systemd-getty-generator now honors a new kernel command line
4017 argument systemd.getty_auto= and a new environment variable
4018 $SYSTEMD_GETTY_AUTO that allows turning it off at boot. This is for
4019 example useful to turn off gettys inside of containers or similar
4022 * systemd-resolved now listens on a second DNS stub address: 127.0.0.54
4023 (in addition to 127.0.0.53, as before). If DNS requests are sent to
4024 this address they are propagated in "bypass" mode only, i.e. are
4025 almost not processed locally, but mostly forwarded as-is to the
4026 current upstream DNS servers. This provides a stable DNS server
4027 address that proxies all requests dynamically to the right upstream
4028 DNS servers even if these dynamically change. This stub does not do
4029 mDNS/LLMNR resolution. However, it will translate look-ups to
4030 DNS-over-TLS if necessary. This new stub is particularly useful in
4031 container/VM environments, or for tethering setups: use DNAT to
4032 redirect traffic to any IP address to this stub.
4034 * systemd-importd now honors new environment variables
4035 $SYSTEMD_IMPORT_BTRFS_SUBVOL, $SYSTEMD_IMPORT_BTRFS_QUOTA,
4036 $SYSTEMD_IMPORT_SYNC, which may be used disable btrfs subvolume
4037 generation, btrfs quota setup and disk synchronization.
4039 * systemd-importd and systemd-resolved can now be optionally built with
4040 OpenSSL instead of libgcrypt.
4042 * systemd-repart no longer requires OpenSSL.
4044 * systemd-sysusers will no longer create the redundant 'nobody' group
4045 by default, as the 'nobody' user is already created with an
4046 appropriate primary group.
4048 * If a unit uses RuntimeMaxSec, systemctl show will now display it.
4050 * systemctl show-environment gained support for --output=json.
4052 * pam_systemd will now first try to use the X11 abstract socket, and
4053 fallback to the socket file in /tmp/.X11-unix/ only if that does not
4056 * systemd-journald will no longer go back to volatile storage
4057 regardless of configuration when its unit is restarted.
4059 * Initial support for the LoongArch architecture has been added (system
4060 call lists, GPT partition table UUIDs, etc).
4062 * systemd-journald's own logging messages are now also logged to the
4063 journal itself when systemd-journald logs to /dev/kmsg.
4065 * systemd-journald now re-enables COW for archived journal files on
4066 filesystems that support COW. One benefit of this change is that
4067 archived journal files will now get compressed on btrfs filesystems
4068 that have compression enabled.
4070 * systemd-journald now deduplicates fields in a single log message
4071 before adding it to the journal. In archived journal files, it will
4072 also punch holes for unused parts and truncate the file as
4073 appropriate, leading to reductions in disk usage.
4075 * journalctl --verify was extended with more informative error
4078 * More of sd-journal's functions are now resistant against journal file
4081 * The shutdown command learnt a new option --show, to display the
4084 * A LICENSES/ directory is now included in the git tree. It contains a
4085 README.md file that explains the licenses used by source files in
4086 this repository. It also contains the text of all applicable
4087 licenses as they appear on spdx.org.
4089 Contributions from: Aakash Singh, acsfer, Adolfo Jayme Barrientos,
4090 Adrian Vovk, Albert Brox, Alberto Mardegan, Alexander Kanavin,
4091 alexlzhu, Alfonso Sánchez-Beato, Alvin Šipraga, Alyssa Ross,
4092 Amir Omidi, Anatol Pomozov, Andika Triwidada, Andreas Rammhold,
4093 Andreas Valder, Andrej Lajovic, Andrew Soutar, Andrew Stone, Andy Chi,
4094 Anita Zhang, Anssi Hannula, Antonio Alvarez Feijoo,
4095 Antony Deepak Thomas, Arnaud Ferraris, Arvid E. Picciani,
4096 Bastien Nocera, Benjamin Berg, Benjamin Herrenschmidt, Ben Stockett,
4097 Bogdan Seniuc, Boqun Feng, Carl Lei, chlorophyll-zz, Chris Packham,
4098 Christian Brauner, Christian Göttsche, Christian Wehrli,
4099 Christoph Anton Mitterer, Cristian Rodríguez, Daan De Meyer,
4100 Daniel Maixner, Dann Frazier, Dan Streetman, Davide Cavalca,
4101 David Seifert, David Tardon, dependabot[bot], Dimitri John Ledkov,
4102 Dimitri Papadopoulos, Dimitry Ishenko, Dmitry Khlebnikov,
4103 Dominique Martinet, duament, Egor, Egor Ignatov, Emil Renner Berthing,
4104 Emily Gonyer, Ettore Atalan, Evgeny Vereshchagin, Florian Klink,
4105 Franck Bui, Frantisek Sumsal, Geass-LL, Gibeom Gwon, GnunuX,
4106 Gogo Gogsi, gregzuro, Greg Zuro, Gustavo Costa, Hans de Goede,
4107 Hela Basa, Henri Chain, hikigaya58, Hugo Carvalho,
4108 Hugo Osvaldo Barrera, Iago Lopez Galeiras, Iago López Galeiras,
4109 I-dont-need-name, igo95862, Jack Dähn, James Hilliard, Jan Janssen,
4110 Jan Kuparinen, Jan Macku, Jan Palus, Jarkko Sakkinen, Jayce Fayne,
4111 jiangchuangang, jlempen, John Lindgren, Jonas Dreßler, Jonas Jelten,
4112 Jonas Witschel, Joris Hartog, José Expósito, Julia Kartseva,
4113 Kai-Heng Feng, Kai Wohlfahrt, Kay Siver Bø, KennthStailey,
4114 Kevin Kuehler, Kevin Orr, Khem Raj, Kristian Klausen, Kyle Laker,
4115 lainahai, LaserEyess, Lennart Poettering, Lia Lenckowski, longpanda,
4116 Luca Boccassi, Luca BRUNO, Ludwig Nussel, Lukas Senionis,
4117 Maanya Goenka, Maciek Borzecki, Marcel Menzel, Marco Scardovi,
4118 Marcus Harrison, Mark Boudreau, Matthijs van Duin, Mauricio Vásquez,
4119 Maxime de Roucy, Max Resch, MertsA, Michael Biebl, Michael Catanzaro,
4120 Michal Koutný, Michal Sekletár, Miika Karanki, Mike Gilbert,
4121 Milo Turner, ml, monosans, Nacho Barrientos, nassir90, Nishal Kulkarni,
4122 nl6720, Ondrej Kozina, Paulo Neves, Pavel Březina, pedro martelletto,
4123 Peter Hutterer, Peter Morrow, Piotr Drąg, Rasmus Villemoes, ratijas,
4124 Raul Tambre, rene, Riccardo Schirone, Robert-L-Turner, Robert Scheck,
4125 Ross Jennings, saikat0511, Scott Lamb, Scott Worley,
4126 Sergei Trofimovich, Sho Iizuka, Slava Bacherikov, Slimane Selyan Amiri,
4127 StefanBruens, Steven Siloti, svonohr, Taiki Sugawara, Takashi Sakamoto,
4128 Takuro Onoue, Thomas Blume, Thomas Haller, Thomas Mühlbacher,
4129 Tianlu Shao, Toke Høiland-Jørgensen, Tom Yan, Tony Asleson,
4130 Topi Miettinen, Ulrich Ölmann, Urs Ritzmann, Vincent Bernat,
4131 Vito Caputo, Vladimir Panteleev, WANG Xuerui, Wind/owZ, Wu Xiaotian,
4132 xdavidwu, Xiaotian Wu, xujing, yangmingtai, Yao Wei, Yao Wei (魏銘廷),
4133 Yegor Alexeyev, Yu Watanabe, Zbigniew Jędrzejewski-Szmek,
4134 Дамјан Георгиевски, наб
4136 — Warsaw, 2021-12-23
4140 * When operating on disk images via the --image= switch of various
4141 tools (such as systemd-nspawn or systemd-dissect), or when udev finds
4142 no 'root=' parameter on the kernel command line, and multiple
4143 suitable root or /usr/ partitions exist in the image, then a simple
4144 comparison inspired by strverscmp() is done on the GPT partition
4145 label, and the newest partition is picked. This permits a simple and
4146 generic whole-file-system A/B update logic where new operating system
4147 versions are dropped into partitions whose label is then updated with
4148 a matching version identifier.
4150 * systemd-sysusers now supports querying the passwords to set for the
4151 users it creates via the "credentials" logic introduced in v247: the
4152 passwd.hashed-password.<user> and passwd.plaintext-password.<user>
4153 credentials are consulted for the password to use (either in UNIX
4154 hashed form, or literally). By default these credentials are inherited
4155 down from PID1 (which in turn imports it from a container manager if
4156 there is one). This permits easy configuration of user passwords
4157 during first boot. Example:
4159 # systemd-nspawn -i foo.raw --volatile=yes --set-credential=passwd.plaintext-password.root:foo
4161 Note that systemd-sysusers operates in purely additive mode: it
4162 executes no operation if the declared users already exist, and hence
4163 doesn't set any passwords as effect of the command line above if the
4164 specified root user exists already in the image. (Note that
4165 --volatile=yes ensures it doesn't, though.)
4167 * systemd-firstboot now also supports querying various system
4168 parameters via the credential subsystems. Thus, as above this may be
4169 used to initialize important system parameters on first boot of
4170 previously unprovisioned images (i.e. images with a mostly empty
4173 * PID 1 may now show both the unit name and the unit description
4174 strings in its status output during boot. This may be configured with
4175 StatusUnitFormat=combined in system.conf or
4176 systemd.status-unit-format=combined on the kernel command line.
4178 * The systemd-machine-id-setup tool now supports a --image= switch for
4179 provisioning a machine ID file into an OS disk image, similar to how
4180 --root= operates on an OS file tree. This matches the existing switch
4181 of the same name for systemd-tmpfiles, systemd-firstboot, and
4182 systemd-sysusers tools.
4184 * Similarly, systemd-repart gained support for the --image= switch too.
4185 In combination with the existing --size= option, this makes the tool
4186 particularly useful for easily growing disk images in a single
4187 invocation, following the declarative rules included in the image
4190 * systemd-repart's partition configuration files gained support for a
4191 new switch MakeDirectories= which may be used to create arbitrary
4192 directories inside file systems that are created, before registering
4193 them in the partition table. This is useful in particular for root
4194 partitions to create mount point directories for other partitions
4195 included in the image. For example, a disk image that contains a
4196 root, /home/, and /var/ partitions, may set MakeDirectories=yes to
4197 create /home/ and /var/ as empty directories in the root file system
4198 on its creation, so that the resulting image can be mounted
4199 immediately, even in read-only mode.
4201 * systemd-repart's CopyBlocks= setting gained support for the special
4202 value "auto". If used, a suitable matching partition on the booted OS
4203 is found as source to copy blocks from. This is useful when
4204 implementing replicating installers, that are booted from one medium
4205 and then stream their own root partition onto the target medium.
4207 * systemd-repart's partition configuration files gained support for a
4208 Flags=, a ReadOnly= and a NoAuto= setting, allowing control of these
4209 GPT partition flags for the created partitions: this is useful for
4210 marking newly created partitions as read-only, or as not being
4211 subject for automatic mounting from creation on.
4213 * The /etc/os-release file has been extended with two new (optional)
4214 variables IMAGE_VERSION= and IMAGE_ID=, carrying identity and version
4215 information for OS images that are updated comprehensively and
4216 atomically as one image. Two new specifiers %M, %A now resolve to
4217 these two fields in the various configuration options that resolve
4220 * portablectl gained a new switch --extension= for enabling portable
4221 service images with extensions that follow the extension image
4222 concept introduced with v248, and thus allows layering multiple
4223 images when setting up the root filesystem of the service.
4225 * systemd-coredump will now extract ELF build-id information from
4226 processes dumping core and include it in the coredump report.
4227 Moreover, it will look for ELF .note.package sections with
4228 distribution packaging meta-information about the crashing process.
4229 This is useful to directly embed the rpm or deb (or any other)
4230 package name and version in ELF files, making it easy to match
4231 coredump reports with the specific package for which the software was
4232 compiled. This is particularly useful on environments with ELF files
4233 from multiple vendors, different distributions and versions, as is
4234 common today in our containerized and sand-boxed world. For further
4237 https://systemd.io/COREDUMP_PACKAGE_METADATA
4239 * A new udev hardware database has been added for FireWire devices
4242 * The "net_id" built-in of udev has been updated with three
4243 backwards-incompatible changes:
4245 - PCI hotplug slot names on s390 systems are now parsed as
4246 hexadecimal numbers. They were incorrectly parsed as decimal
4247 previously, or ignored if the name was not a valid decimal
4250 - PCI onboard indices up to 65535 are allowed. Previously, numbers
4251 above 16383 were rejected. This primarily impacts s390 systems,
4252 where values up to 65535 are used.
4254 - Invalid characters in interface names are replaced with "_".
4256 The new version of the net naming scheme is "v249". The previous
4257 scheme can be selected via the "net.naming-scheme=v247" kernel
4258 command line parameter.
4260 * sd-bus' sd_bus_is_ready() and sd_bus_is_open() calls now accept a
4261 NULL bus object, for which they will return false. Or in other words,
4262 an unallocated bus connection is neither ready nor open.
4264 * The sd-device API acquired a new API function
4265 sd_device_get_usec_initialized() that returns the monotonic time when
4266 the udev device first appeared in the database.
4268 * sd-device gained a new APIs sd_device_trigger_with_uuid() and
4269 sd_device_get_trigger_uuid(). The former is similar to
4270 sd_device_trigger() but returns a randomly generated UUID that is
4271 associated with the synthetic uevent generated by the call. This UUID
4272 may be read from the sd_device object a monitor eventually receives,
4273 via the sd_device_get_trigger_uuid(). This interface requires kernel
4274 4.13 or above to work, and allows tracking a synthetic uevent through
4275 the entire device management stack. The "udevadm trigger --settle"
4276 logic has been updated to make use of this concept if available to
4277 wait precisely for the uevents it generates. "udevadm trigger" also
4278 gained a new parameter --uuid that prints the UUID for each generated
4281 * sd-device also gained new APIs sd_device_new_from_ifname() and
4282 sd_device_new_from_ifindex() for allocating an sd-device object for
4283 the specified network interface. The former accepts an interface name
4284 (either a primary or an alternative name), the latter an interface
4287 * The native Journal protocol has been documented. Clients may talk
4288 this as alternative to the classic BSD syslog protocol for locally
4289 delivering log records to the Journal. The protocol has been stable
4290 for a long time and in fact been implemented already in a variety
4291 of alternative client libraries. This documentation makes the support
4294 https://systemd.io/JOURNAL_NATIVE_PROTOCOL
4296 * A new BPFProgram= setting has been added to service files. It may be
4297 set to a path to a loaded kernel BPF program, i.e. a path to a bpffs
4298 file, or a bind mount or symlink to one. This may be used to upload
4299 and manage BPF programs externally and then hook arbitrary systemd
4302 * The "home.arpa" domain that has been officially declared as the
4303 choice for domain for local home networks per RFC 8375 has been added
4304 to the default NTA list of resolved, since DNSSEC is generally not
4305 available on private domains.
4307 * The CPUAffinity= setting of unit files now resolves "%" specifiers.
4309 * A new ManageForeignRoutingPolicyRules= setting has been added to
4310 .network files which may be used to exclude foreign-created routing
4311 policy rules from systemd-networkd management.
4313 * systemd-network-wait-online gained two new switches -4 and -6 that
4314 may be used to tweak whether to wait for only IPv4 or only IPv6
4317 * .network files gained a new RequiredFamilyForOnline= setting to
4318 fine-tune whether to require an IPv4 or IPv6 address in order to
4319 consider an interface "online".
4321 * networkctl will now show an over-all "online" state in the per-link
4324 * In .network files a new OutgoingInterface= setting has been added to
4325 specify the output interface in bridge FDB setups.
4327 * In .network files the Multipath group ID may now be configured for
4328 [NextHop] entries, via the new Group= setting.
4330 * The DHCP server logic configured in .network files gained a new
4331 setting RelayTarget= that turns the server into a DHCP server relay.
4332 The RelayAgentCircuitId= and RelayAgentRemoteId= settings may be used
4333 to further tweak the DHCP relay behaviour.
4335 * The DHCP server logic also gained a new ServerAddress= setting in
4336 .network files that explicitly specifies the server IP address to
4337 use. If not specified, the address is determined automatically, as
4340 * The DHCP server logic in systemd-networkd gained support for static
4341 DHCP leases, configurable via the [DHCPServerStaticLease]
4342 section. This allows explicitly mapping specific MAC addresses to
4343 fixed IP addresses and vice versa.
4345 * The RestrictAddressFamilies= setting in service files now supports a
4346 new special value "none". If specified sockets of all address
4347 families will be made unavailable to services configured that way.
4349 * systemd-fstab-generator and systemd-repart have been updated to
4350 support booting from disks that carry only a /usr/ partition but no
4351 root partition yet, and where systemd-repart can add it in on the
4352 first boot. This is useful for implementing systems that ship with a
4353 single /usr/ file system, and whose root file system shall be set up
4354 and formatted on a LUKS-encrypted volume whose key is generated
4355 locally (and possibly enrolled in the TPM) during the first boot.
4357 * The [Address] section of .network files now accepts a new
4358 RouteMetric= setting that configures the routing metric to use for
4359 the prefix route created as effect of the address configuration.
4360 Similarly, the [DHCPv6PrefixDelegation] and [IPv6Prefix] sections
4361 gained matching settings for their prefix routes. (The option of the
4362 same name in the [DHCPv6] section is moved to [IPv6AcceptRA], since
4363 it conceptually belongs there; the old option is still understood for
4366 * The DHCPv6 IAID and DUID are now explicitly configurable in .network
4369 * A new udev property ID_NET_DHCP_BROADCAST on network interface
4370 devices is now honoured by systemd-networkd, controlling whether to
4371 issue DHCP offers via broadcasting. This is used to ensure that s390
4372 layer 3 network interfaces work out-of-the-box with systemd-networkd.
4374 * nss-myhostname and systemd-resolved will now synthesize address
4375 records for a new special hostname "_outbound". The name will always
4376 resolve to the local IP addresses most likely used for outbound
4377 connections towards the default routes. On multi-homed hosts this is
4378 useful to have a stable handle referring to "the" local IP address
4379 that matters most, to the point where this is defined.
4381 * The Discoverable Partition Specification has been updated with a new
4382 GPT partition flag "grow-file-system" defined for its partition
4383 types. Whenever partitions with this flag set are automatically
4384 mounted (i.e. via systemd-gpt-auto-generator or the --image= switch
4385 of systemd-nspawn or other tools; and as opposed to explicit mounting
4386 via /etc/fstab), the file system within the partition is
4387 automatically grown to the full size of the partition. If the file
4388 system size already matches the partition size this flag has no
4389 effect. Previously, this functionality has been available via the
4390 explicit x-systemd.growfs mount option, and this new flag extends
4391 this to automatically discovered mounts. A new GrowFileSystem=
4392 setting has been added to systemd-repart drop-in files that allows
4393 configuring this partition flag. This new flag defaults to on for
4394 partitions automatically created by systemd-repart, except if they
4395 are marked read-only. See the specification for further details:
4397 https://systemd.io/DISCOVERABLE_PARTITIONS
4399 * .network files gained a new setting RoutesToNTP= in the [DHCPv4]
4400 section. If enabled (which is the default), and an NTP server address
4401 is acquired through a DHCP lease on this interface an explicit route
4402 to this address is created on this interface to ensure that NTP
4403 traffic to the NTP server acquired on an interface is also routed
4404 through that interface. The pre-existing RoutesToDNS= setting that
4405 implements the same for DNS servers is now enabled by default.
4407 * A pair of service settings SocketBindAllow= + SocketBindDeny= have
4408 been added that may be used to restrict the network interfaces
4409 sockets created by the service may be bound to. This is implemented
4412 * A new ConditionFirmware= setting has been added to unit files to
4413 conditionalize on certain firmware features. At the moment it may
4414 check whether running on a UEFI system, a device.tree system, or if
4415 the system is compatible with some specified device-tree feature.
4417 * A new ConditionOSRelease= setting has been added to unit files to
4418 check os-release(5) fields. The "=", "!=", "<", "<=", ">=", ">"
4419 operators may be used to check if some field has some specific value
4420 or do an alphanumerical comparison. Equality comparisons are useful
4421 for fields like ID, but relative comparisons for fields like
4422 VERSION_ID or IMAGE_VERSION.
4424 * hostnamed gained a new Describe() D-Bus method that returns a JSON
4425 serialization of the host data it exposes. This is exposed via
4426 "hostnamectl --json=" to acquire a host identity description in JSON.
4427 It's our intention to add a similar features to most services and
4428 objects systemd manages, in order to simplify integration with
4429 program code that can consume JSON.
4431 * Similarly, networkd gained a Describe() method on its Manager and
4432 Link bus objects. This is exposed via "networkctl --json=".
4434 * hostnamectl's various "get-xyz"/"set-xyz" verb pairs
4435 (e.g. "hostnamectl get-hostname", "hostnamectl "set-hostname") have
4436 been replaced by a single "xyz" verb (e.g. "hostnamectl hostname")
4437 that is used both to get the value (when no argument is given), and
4438 to set the value (when an argument is specified). The old names
4439 continue to be supported for compatibility.
4441 * systemd-detect-virt and ConditionVirtualization= are now able to
4442 correctly identify Amazon EC2 environments.
4444 * The LogLevelMax= setting of unit files now applies not only to log
4445 messages generated *by* the service, but also to log messages
4446 generated *about* the service by PID 1. To suppress logs concerning a
4447 specific service comprehensively, set this option to a high log
4450 * bootctl gained support for a new --make-machine-id-directory= switch
4451 that allows precise control on whether to create the top-level
4452 per-machine directory in the boot partition that typically contains
4453 Type 1 boot loader entries.
4455 * During build SBAT data to include in the systemd-boot EFI PE binaries
4456 may be specified now.
4458 * /etc/crypttab learnt a new option "headless". If specified any
4459 requests to query the user interactively for passwords or PINs will
4460 be skipped. This is useful on systems that are headless, i.e. where
4461 an interactive user is generally not present.
4463 * /etc/crypttab also learnt a new option "password-echo=" that allows
4464 configuring whether the encryption password prompt shall echo the
4465 typed password and if so, do so literally or via asterisks. (The
4466 default is the same behaviour as before: provide echo feedback via
4469 * FIDO2 support in systemd-cryptenroll/systemd-cryptsetup and
4470 systemd-homed has been updated to allow explicit configuration of the
4471 "user presence" and "user verification" checks, as well as whether a
4472 PIN is required for authentication, via the new switches
4473 --fido2-with-user-presence=, --fido2-with-user-verification=,
4474 --fido2-with-client-pin= to systemd-cryptenroll and homectl. Which
4475 features are available, and may be enabled or disabled depends on the
4478 * systemd-nspawn's --private-user= switch now accepts the special value
4479 "identity" which configures a user namespacing environment with an
4480 identity mapping of 65535 UIDs. This means the container UID 0 is
4481 mapped to the host UID 0, and the UID 1 to host UID 1. On first look
4482 this doesn't appear to be useful, however it does reduce the attack
4483 surface a bit, since the resulting container will possess process
4484 capabilities only within its namespace and not on the host.
4486 * systemd-nspawn's --private-user-chown switch has been replaced by a
4487 more generic --private-user-ownership= switch that accepts one of
4488 three values: "chown" is equivalent to the old --private-user-chown,
4489 and "off" is equivalent to the absence of the old switch. The value
4490 "map" uses the new UID mapping mounts of Linux 5.12 to map ownership
4491 of files and directories of the underlying image to the chosen UID
4492 range for the container. "auto" is equivalent to "map" if UID mapping
4493 mount are supported, otherwise it is equivalent to "chown". The short
4494 -U switch systemd-nspawn now implies --private-user-ownership=auto
4495 instead of the old --private-user-chown. Effectively this means: if
4496 the backing file system supports UID mapping mounts the feature is
4497 now used by default if -U is used. Generally, it's a good idea to use
4498 UID mapping mounts instead of recursive chown()ing, since it allows
4499 running containers off immutable images (since no modifications of
4500 the images need to take place), and share images between multiple
4501 instances. Moreover, the recursive chown()ing operation is slow and
4502 can be avoided. Conceptually it's also a good thing if transient UID
4503 range uses do not leak into persistent file ownership anymore. TLDR:
4504 finally, the last major drawback of user namespacing has been
4505 removed, and -U should always be used (unless you use btrfs, where
4506 UID mapped mounts do not exist; or your container actually needs
4507 privileges on the host).
4509 * nss-systemd now synthesizes user and group shadow records in addition
4510 to the main user and group records. Thus, hashed passwords managed by
4511 systemd-homed are now accessible via the shadow database.
4513 * The userdb logic (and thus nss-systemd, and so on) now read
4514 additional user/group definitions in JSON format from the drop-in
4515 directories /etc/userdb/, /run/userdb/, /run/host/userdb/ and
4516 /usr/lib/userdb/. This is a simple and powerful mechanism for making
4517 additional users available to the system, with full integration into
4518 NSS including the shadow databases. Since the full JSON user/group
4519 record format is supported this may also be used to define users with
4520 resource management settings and other runtime settings that
4521 pam_systemd and systemd-logind enforce at login.
4523 * The userdbctl tool gained two new switches --with-dropin= and
4524 --with-varlink= which can be used to fine-tune the sources used for
4525 user database lookups.
4527 * systemd-nspawn gained a new switch --bind-user= for binding a host
4528 user account into the container. This does three things: the user's
4529 home directory is bind mounted from the host into the container,
4530 below the /run/userdb/home/ hierarchy. A free UID is picked in the
4531 container, and a user namespacing UID mapping to the host user's UID
4532 installed. And finally, a minimal JSON user and group record (along
4533 with its hashed password) is dropped into /run/host/userdb/. These
4534 records are picked up automatically by the userdb drop-in logic
4535 describe above, and allow the user to login with the same password as
4536 on the host. Effectively this means: if host and container run new
4537 enough systemd versions making a host user available to the container
4538 is trivially simple.
4540 * systemd-journal-gatewayd now supports the switches --user, --system,
4541 --merge, --file= that are equivalent to the same switches of
4542 journalctl, and permit exposing only the specified subset of the
4545 * The OnFailure= dependency between units is now augmented with a
4546 implicit reverse dependency OnFailureOf= (this new dependency cannot
4547 be configured directly it's only created as effect of an OnFailure=
4548 dependency in the reverse order — it's visible in "systemctl show"
4549 however). Similar, Slice= now has an reverse dependency SliceOf=,
4550 that is also not configurable directly, but useful to determine all
4551 units that are members of a slice.
4553 * A pair of new dependency types between units PropagatesStopTo= +
4554 StopPropagatedFrom= has been added, that allows propagation of unit
4555 stop events between two units. It operates similar to the existing
4556 PropagatesReloadTo= + ReloadPropagatedFrom= dependencies.
4558 * A new dependency type OnSuccess= has been added (plus the reverse
4559 dependency OnSuccessOf=, which cannot be configured directly, but
4560 exists only as effect of the reverse OnSuccess=). It is similar to
4561 OnFailure=, but triggers in the opposite case: when a service exits
4562 cleanly. This allows "chaining up" of services where one or more
4563 services are started once another service has successfully completed.
4565 * A new dependency type Upholds= has been added (plus the reverse
4566 dependency UpheldBy=, which cannot be configured directly, but exists
4567 only as effect of Upholds=). This dependency type is a stronger form
4568 of Wants=: if a unit has an UpHolds= dependency on some other unit
4569 and the former is active then the latter is started whenever it is
4570 found inactive (and no job is queued for it). This is an alternative
4571 to Restart= inside service units, but less configurable, and the
4572 request to uphold a unit is not encoded in the unit itself but in
4573 another unit that intends to uphold it.
4575 * The systemd-ask-password tool now also supports reading passwords
4576 from the credentials subsystem, via the new --credential= switch.
4578 * The systemd-ask-password tool learnt a new switch --emoji= which may
4579 be used to explicit control whether the lock and key emoji (🔐) is
4580 shown in the password prompt on suitable TTYs.
4582 * The --echo switch of systemd-ask-password now optionally takes a
4583 parameter that controls character echo. It may either show asterisks
4584 (default, as before), turn echo off entirely, or echo the typed
4585 characters literally.
4587 * The systemd-ask-password tool also gained a new -n switch for
4588 suppressing output of a trailing newline character when writing the
4589 acquired password to standard output, similar to /bin/echo's -n
4592 * New documentation has been added that describes the organization of
4593 the systemd source code tree:
4595 https://systemd.io/ARCHITECTURE
4597 * Units using ConditionNeedsUpdate= will no longer be activated in
4600 * It is now possible to list a template unit in the WantedBy= or
4601 RequiredBy= settings of the [Install] section of another template
4602 unit, which will be instantiated using the same instance name.
4604 * A new MemoryAvailable property is available for units. If the unit,
4605 or the slices it is part of, have a memory limit set via MemoryMax=/
4606 MemoryHigh=, MemoryAvailable will indicate how much more memory the
4607 unit can claim before hitting the limits.
4609 * systemd-coredump will now try to stay below the cgroup memory limit
4610 placed on itself or one of the slices it runs under, if the storage
4611 area for core files (/var/lib/systemd/coredump/) is placed on a tmpfs,
4612 since files written on such filesystems count toward the cgroup memory
4613 limit. If there is not enough available memory in such cases to store
4614 the core file uncompressed, systemd-coredump will skip to compressed
4615 storage directly (if enabled) and it will avoid analyzing the core file
4616 to print backtrace and metadata in the journal.
4618 * tmpfiles.d/ drop-ins gained a new '=' modifier to check if the type
4619 of a path matches the configured expectations, and remove it if not.
4621 * tmpfiles.d/'s 'Age' now accepts an 'age-by' argument, which allows to
4622 specify which of the several available filesystem timestamps (access
4623 time, birth time, change time, modification time) to look at when
4624 deciding whether a path has aged enough to be cleaned.
4626 * A new IPv6StableSecretAddress= setting has been added to .network
4627 files, which takes an IPv6 address to use as secret for IPv6 address
4630 * The [DHCPServer] logic in .network files gained support for a new
4631 UplinkInterface= setting that permits configuration of the uplink
4632 interface name to propagate DHCP lease information from.
4634 * The WakeOnLan= setting in .link files now accepts a list of flags
4635 instead of a single one, to configure multiple wake-on-LAN policies.
4637 * User-space defined tracepoints (USDT) have been added to udev at
4638 strategic locations. This is useful for tracing udev behaviour and
4639 performance with bpftrace and similar tools.
4641 * systemd-journald-upload gained a new NetworkTimeoutSec= option for
4642 setting a network timeout time.
4644 * If a system service is running in a new mount namespace (RootDirectory=
4645 and friends), all file systems will be mounted with MS_NOSUID by
4646 default, unless the system is running with SELinux enabled.
4648 * When enumerating time zones the timedatectl tool will now consult the
4649 'tzdata.zi' file shipped by the IANA time zone database package, in
4650 addition to 'zone1970.tab', as before. This makes sure time zone
4651 aliases are now correctly supported. Some distributions so far did
4652 not install this additional file, most do however. If you
4653 distribution does not install it yet, it might make sense to change
4656 * Intel HID rfkill event is no longer masked, since it's the only
4657 source of rfkill event on newer HP laptops. To have both backward and
4658 forward compatibility, userspace daemon needs to debounce duplicated
4659 events in a short time window.
4661 Contributions from: Aakash Singh, adrian5, Albert Brox,
4662 Alexander Sverdlin, Alexander Tsoy, Alexey Rubtsov, alexlzhu,
4663 Allen Webb, Alvin Šipraga, Alyssa Ross, Anders Wenhaug,
4664 Andrea Pappacoda, Anita Zhang, asavah, Balint Reczey, Bertrand Jacquin,
4665 borna-blazevic, caoxia2008cxx, Carlo Teubner, Christian Göttsche,
4666 Christian Hesse, Daniel Schaefer, Dan Streetman,
4667 David Santamaría Rogado, David Tardon, Deepak Rawat, dgcampea,
4668 Dimitri John Ledkov, ei-ke, Emilio Herrera, Emil Renner Berthing,
4669 Eric Cook, Flos Lonicerae, Franck Bui, Francois Gervais,
4670 Frantisek Sumsal, Gibeom Gwon, gitm0, Hamish Moffatt, Hans de Goede,
4671 Harsh Barsaiyan, Henri Chain, Hristo Venev, Icenowy Zheng, Igor Zhbanov,
4672 imayoda, Jakub Warczarek, James Buren, Jan Janssen, Jan Macku,
4673 Jan Synacek, Jason Francis, Jayanth Ananthapadmanaban, Jeremy Szu,
4674 Jérôme Carretero, Jesse Stricker, jiangchuangang, Joerg Behrmann,
4675 Jóhann B. Guðmundsson, Jörg Deckert, Jörg Thalheim, Juergen Hoetzel,
4676 Julia Kartseva, Kai-Heng Feng, Khem Raj, KoyamaSohei, laineantti,
4677 Lennart Poettering, LetzteInstanz, Luca Adrian L, Luca Boccassi,
4678 Lucas Magasweran, Mantas Mikulėnas, Marco Antonio Mauro, Mark Wielaard,
4679 Masahiro Matsuya, Matt Johnston, Michael Catanzaro, Michal Koutný,
4680 Michal Sekletár, Mike Crowe, Mike Kazantsev, Milan, milaq,
4681 Miroslav Suchý, Morten Linderud, nerdopolis, nl6720, Noah Meyerhans,
4682 Oleg Popov, Olle Lundberg, Ondrej Kozina, Paweł Marciniak, Perry.Yuan,
4683 Peter Hutterer, Peter Kjellerstedt, Peter Morrow, Phaedrus Leeds,
4684 plattrap, qhill, Raul Tambre, Roman Beranek, Roshan Shariff,
4685 Ryan Hendrickson, Samuel BF, scootergrisen, Sebastian Blunt,
4686 Seong-ho Cho, Sergey Bugaev, Sevan Janiyan, Sibo Dong, simmon,
4687 Simon Watts, Srinidhi Kaushik, Štěpán Němec, Steve Bonds, Susant Sahani,
4688 sverdlin, syyhao1994, Takashi Sakamoto, Topi Miettinen, tramsay,
4689 Trent Piepho, Uwe Kleine-König, Viktor Mihajlovski, Vincent Dechenaux,
4690 Vito Caputo, William A. Kennington III, Yangyang Shen, Yegor Alexeyev,
4691 Yi Gao, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, zsien, наб
4693 — Edinburgh, 2021-07-07
4697 * A concept of system extension images is introduced. Such images may
4698 be used to extend the /usr/ and /opt/ directory hierarchies at
4699 runtime with additional files (even if the file system is read-only).
4700 When a system extension image is activated, its /usr/ and /opt/
4701 hierarchies and os-release information are combined via overlayfs
4702 with the file system hierarchy of the host OS.
4704 A new systemd-sysext tool can be used to merge, unmerge, list, and
4705 refresh system extension hierarchies. See
4706 https://www.freedesktop.org/software/systemd/man/systemd-sysext.html.
4708 The systemd-sysext.service automatically merges installed system
4709 extensions during boot (before basic.target, but not in very early
4710 boot, since various file systems have to be mounted first).
4712 The SYSEXT_LEVEL= field in os-release(5) may be used to specify the
4713 supported system extension level.
4715 * A new ExtensionImages= unit setting can be used to apply the same
4716 system extension image concept from systemd-sysext to the namespaced
4717 file hierarchy of specific services, following the same rules and
4720 * Support for a new special "root=tmpfs" kernel command-line option has
4721 been added. When specified, a tmpfs is mounted on /, and mount.usr=
4722 should be used to point to the operating system implementation.
4724 * A new configuration file /etc/veritytab may be used to configure
4725 dm-verity integrity protection for block devices. Each line is in the
4726 format "volume-name data-device hash-device roothash options",
4727 similar to /etc/crypttab.
4729 * A new kernel command-line option systemd.verity.root_options= may be
4730 used to configure dm-verity behaviour for the root device.
4732 * The key file specified in /etc/crypttab (the third field) may now
4733 refer to an AF_UNIX/SOCK_STREAM socket in the file system. The key is
4734 acquired by connecting to that socket and reading from it. This
4735 allows the implementation of a service to provide key information
4736 dynamically, at the moment when it is needed.
4738 * When the hostname is set explicitly to "localhost", systemd-hostnamed
4739 will respect this. Previously such a setting would be mostly silently
4740 ignored. The goal is to honour configuration as specified by the
4743 * The fallback hostname that will be used by the system manager and
4744 systemd-hostnamed can now be configured in two new ways: by setting
4745 DEFAULT_HOSTNAME= in os-release(5), or by setting
4746 $SYSTEMD_DEFAULT_HOSTNAME in the environment block. As before, it can
4747 also be configured during compilation. The environment variable is
4748 intended for testing and local overrides, the os-release(5) field is
4749 intended to allow customization by different variants of a
4750 distribution that share the same compiled packages.
4752 * The environment block of the manager itself may be configured through
4753 a new ManagerEnvironment= setting in system.conf or user.conf. This
4754 complements existing ways to set the environment block (the kernel
4755 command line for the system manager, the inherited environment and
4756 user@.service unit file settings for the user manager).
4758 * systemd-hostnamed now exports the default hostname and the source of
4759 the configured hostname ("static", "transient", or "default") as
4762 * systemd-hostnamed now exports the "HardwareVendor" and
4763 "HardwareModel" D-Bus properties, which are supposed to contain a
4764 pair of cleaned up, human readable strings describing the system's
4765 vendor and model. It's typically sourced from the firmware's DMI
4766 tables, but may be augmented from a new hwdb database. hostnamectl
4767 shows this in the status output.
4769 * Support has been added to systemd-cryptsetup for extracting the
4770 PKCS#11 token URI and encrypted key from the LUKS2 JSON embedded
4771 metadata header. This allows the information how to open the
4772 encrypted device to be embedded directly in the device and obviates
4773 the need for configuration in an external file.
4775 * systemd-cryptsetup gained support for unlocking LUKS2 volumes using
4776 TPM2 hardware, as well as FIDO2 security tokens (in addition to the
4777 pre-existing support for PKCS#11 security tokens).
4779 * systemd-repart may enroll encrypted partitions using TPM2
4780 hardware. This may be useful for example to create an encrypted /var
4781 partition bound to the machine on first boot.
4783 * A new systemd-cryptenroll tool has been added to enroll TPM2, FIDO2
4784 and PKCS#11 security tokens to LUKS volumes, list and destroy
4787 https://0pointer.net/blog/unlocking-luks2-volumes-with-tpm2-fido2-pkcs11-security-hardware-on-systemd-248.html
4789 It also supports enrolling "recovery keys" and regular passphrases.
4791 * The libfido2 dependency is now based on dlopen(), so that the library
4792 is used at runtime when installed, but is not a hard runtime
4795 * systemd-cryptsetup gained support for two new options in
4796 /etc/crypttab: "no-write-workqueue" and "no-read-workqueue" which
4797 request synchronous processing of encryption/decryption IO.
4799 * The manager may be configured at compile time to use the fexecve()
4800 instead of the execve() system call when spawning processes. Using
4801 fexecve() closes a window between checking the security context of an
4802 executable and spawning it, but unfortunately the kernel displays
4803 stale information in the process' "comm" field, which impacts ps
4806 * The configuration option -Dcompat-gateway-hostname has been dropped.
4807 "_gateway" is now the only supported name.
4809 * The ConditionSecurity=tpm2 unit file setting may be used to check if
4810 the system has at least one TPM2 (tpmrm class) device.
4812 * A new ConditionCPUFeature= has been added that may be used to
4813 conditionalize units based on CPU features. For example,
4814 ConditionCPUFeature=rdrand will condition a unit so that it is only
4815 run when the system CPU supports the RDRAND opcode.
4817 * The existing ConditionControlGroupController= setting has been
4818 extended with two new values "v1" and "v2". "v2" means that the
4819 unified v2 cgroup hierarchy is used, and "v1" means that legacy v1
4820 hierarchy or the hybrid hierarchy are used.
4822 * A new PrivateIPC= setting on a unit file allows executed processes to
4823 be moved into a private IPC namespace, with separate System V IPC
4824 identifiers and POSIX message queues.
4826 A new IPCNamespacePath= allows the unit to be joined to an existing
4829 * The tables of system calls in seccomp filters are now automatically
4830 generated from kernel lists exported on
4831 https://fedora.juszkiewicz.com.pl/syscalls.html.
4833 The following architectures should now have complete lists:
4834 alpha, arc, arm64, arm, i386, ia64, m68k, mips64n32, mips64, mipso32,
4835 powerpc, powerpc64, s390, s390x, tilegx, sparc, x86_64, x32.
4837 * The MountAPIVFS= service file setting now additionally mounts a tmpfs
4838 on /run/ if it is not already a mount point. A writable /run/ has
4839 always been a requirement for a functioning system, but this was not
4840 guaranteed when using a read-only image.
4842 Users can always specify BindPaths= or InaccessiblePaths= as
4843 overrides, and they will take precedence. If the host's root mount
4844 point is used, there is no change in behaviour.
4846 * New bind mounts and file system image mounts may be injected into the
4847 mount namespace of a service (without restarting it). This is exposed
4848 respectively as 'systemctl bind <unit> <path>…' and
4849 'systemctl mount-image <unit> <image>…'.
4851 * The StandardOutput= and StandardError= settings can now specify files
4852 to be truncated for output (as "truncate:<path>").
4854 * The ExecPaths= and NoExecPaths= settings may be used to specify
4855 noexec for parts of the file system.
4857 * sd-bus has a new function sd_bus_open_user_machine() to open a
4858 connection to the session bus of a specific user in a local container
4859 or on the local host. This is exposed in the existing -M switch to
4860 systemctl and similar tools:
4862 systemctl --user -M lennart@foobar start foo
4864 This will connect to the user bus of a user "lennart" in container
4865 "foobar". If no container name is specified, the specified user on
4866 the host itself is connected to
4868 systemctl --user -M lennart@ start quux
4870 * sd-bus also gained a convenience function sd_bus_message_send() to
4871 simplify invocations of sd_bus_send(), taking only a single
4872 parameter: the message to send.
4874 * sd-event allows rate limits to be set on event sources, for dealing
4875 with high-priority event sources that might starve out others. See
4876 the new man page sd_event_source_set_ratelimit(3) for details.
4878 * systemd.link files gained a [Link] Promiscuous= switch, which allows
4879 the device to be raised in promiscuous mode.
4881 New [Link] TransmitQueues= and ReceiveQueues= settings allow the
4882 number of TX and RX queues to be configured.
4884 New [Link] TransmitQueueLength= setting allows the size of the TX
4885 queue to be configured.
4887 New [Link] GenericSegmentOffloadMaxBytes= and
4888 GenericSegmentOffloadMaxSegments= allow capping the packet size and
4889 the number of segments accepted in Generic Segment Offload.
4891 * systemd-networkd gained support for the "B.A.T.M.A.N. advanced"
4892 wireless routing protocol that operates on ISO/OSI Layer 2 only and
4893 uses ethernet frames to route/bridge packets. This encompasses a new
4894 "batadv" netdev Type=, a new [BatmanAdvanced] section with a bunch of
4895 new settings in .netdev files, and a new BatmanAdvanced= setting in
4898 * systemd.network files gained a [Network] RouteTable= configuration
4899 switch to select the routing policy table.
4901 systemd.network files gained a [RoutingPolicyRule] Type=
4902 configuration switch (one of "blackhole, "unreachable", "prohibit").
4904 systemd.network files gained a [IPv6AcceptRA] RouteDenyList= and
4905 RouteAllowList= settings to ignore/accept route advertisements from
4906 routers matching specified prefixes. The DenyList= setting has been
4907 renamed to PrefixDenyList= and a new PrefixAllowList= option has been
4910 systemd.network files gained a [DHCPv6] UseAddress= setting to
4911 optionally ignore the address provided in the lease.
4913 systemd.network files gained a [DHCPv6PrefixDelegation]
4914 ManageTemporaryAddress= switch.
4916 systemd.network files gained a new ActivationPolicy= setting which
4917 allows configuring how the UP state of an interface shall be managed,
4918 i.e. whether the interface is always upped, always downed, or may be
4919 upped/downed by the user using "ip link set dev".
4921 * The default for the Broadcast= setting in .network files has slightly
4922 changed: the broadcast address will not be configured for wireguard
4925 * systemd.netdev files gained a [VLAN] Protocol=, IngressQOSMaps=,
4926 EgressQOSMaps=, and [MACVLAN] BroadcastMulticastQueueLength=
4927 configuration options for VLAN packet handling.
4929 * udev rules may now set log_level= option. This allows debug logs to
4930 be enabled for select events, e.g. just for a specific subsystem or
4931 even a single device.
4933 * udev now exports the VOLUME_ID, LOGICAL_VOLUME_ID, VOLUME_SET_ID, and
4934 DATA_PREPARED_ID properties for block devices with ISO9660 file
4937 * udev now exports decoded DMI information about installed memory slots
4938 as device properties under the /sys/class/dmi/id/ pseudo device.
4940 * /dev/ is not mounted noexec anymore. This didn't provide any
4941 significant security benefits and would conflict with the executable
4942 mappings used with /dev/sgx device nodes. The previous behaviour can
4943 be restored for individual services with NoExecPaths=/dev (or by allow-
4944 listing and excluding /dev from ExecPaths=).
4946 * Permissions for /dev/vsock are now set to 0o666, and /dev/vhost-vsock
4947 and /dev/vhost-net are owned by the kvm group.
4949 * The hardware database has been extended with a list of fingerprint
4950 readers that correctly support USB auto-suspend using data from
4953 * systemd-resolved can now answer DNSSEC questions through the stub
4954 resolver interface in a way that allows local clients to do DNSSEC
4955 validation themselves. For a question with DO+CD set, it'll proxy the
4956 DNS query and respond with a mostly unmodified packet received from
4957 the upstream server.
4959 * systemd-resolved learnt a new boolean option CacheFromLocalhost= in
4960 resolved.conf. If true the service will provide caching even for DNS
4961 lookups made to an upstream DNS server on the 127.0.0.1/::1
4962 addresses. By default (and when the option is false) systemd-resolved
4963 will not cache such lookups, in order to avoid duplicate local
4964 caching, under the assumption the local upstream server caches
4967 * systemd-resolved now implements RFC5001 NSID in its local DNS
4968 stub. This may be used by local clients to determine whether they are
4969 talking to the DNS resolver stub or a different DNS server.
4971 * When resolving host names and other records resolvectl will now
4972 report where the data was acquired from (i.e. the local cache, the
4973 network, locally synthesized, …) and whether the network traffic it
4974 effected was encrypted or not. Moreover the tool acquired a number of
4975 new options --cache=, --synthesize=, --network=, --zone=,
4976 --trust-anchor=, --validate= that take booleans and may be used to
4977 tweak a lookup, i.e. whether it may be answered from cached
4978 information, locally synthesized information, information acquired
4979 through the network, the local mDNS/LLMNR zone, the DNSSEC trust
4980 anchor, and whether DNSSEC validation shall be executed for the
4983 * systemd-nspawn gained a new --ambient-capability= setting
4984 (AmbientCapability= in .nspawn files) to configure ambient
4985 capabilities passed to the container payload.
4987 * systemd-nspawn gained the ability to configure the firewall using the
4988 nftables subsystem (in addition to the existing iptables
4989 support). Similarly, systemd-networkd's IPMasquerade= option now
4990 supports nftables as back-end, too. In both cases NAT on IPv6 is now
4991 supported too, in addition to IPv4 (the iptables back-end still is
4994 "IPMasquerade=yes", which was the same as "IPMasquerade=ipv4" before,
4995 retains its meaning, but has been deprecated. Please switch to either
4996 "ivp4" or "both" (if covering IPv6 is desired).
4998 * systemd-importd will now download .verity and .roothash.p7s files
4999 along with the machine image (as exposed via machinectl pull-raw).
5001 * systemd-oomd now gained a new DefaultMemoryPressureDurationSec=
5002 setting to configure the time a unit's cgroup needs to exceed memory
5003 pressure limits before action will be taken, and a new
5004 ManagedOOMPreference=none|avoid|omit setting to avoid killing certain
5007 systemd-oomd is now considered fully supported (the usual
5008 backwards-compatibility promises apply). Swap is not required for
5009 operation, but it is still recommended.
5011 * systemd-timesyncd gained a new ConnectionRetrySec= setting which
5012 configures the retry delay when trying to contact servers.
5014 * systemd-stdio-bridge gained --system/--user options to connect to the
5015 system bus (previous default) or the user session bus.
5017 * systemd-localed may now call locale-gen to generate missing locales
5018 on-demand (UTF-8-only). This improves integration with Debian-based
5019 distributions (Debian/Ubuntu/PureOS/Tanglu/...) and Arch Linux.
5021 * systemctl --check-inhibitors=true may now be used to obey inhibitors
5022 even when invoked non-interactively. The old --ignore-inhibitors
5023 switch is now deprecated and replaced by --check-inhibitors=false.
5025 * systemctl import-environment will now emit a warning when called
5026 without any arguments (i.e. to import the full environment block of
5027 the called program). This command will usually be invoked from a
5028 shell, which means that it'll inherit a bunch of variables which are
5029 specific to that shell, and usually to the TTY the shell is connected
5030 to, and don't have any meaning in the global context of the system or
5031 user service manager. Instead, only specific variables should be
5032 imported into the manager environment block.
5034 Similarly, programs which update the manager environment block by
5035 directly calling the D-Bus API of the manager, should also push
5036 specific variables, and not the full inherited environment.
5038 * systemctl's status output now shows unit state with a more careful
5039 choice of Unicode characters: units in maintenance show a "○" symbol
5040 instead of the usual "●", failed units show "×", and services being
5043 * coredumpctl gained a --debugger-arguments= switch to pass arguments
5044 to the debugger. It also gained support for showing coredump info in
5045 a simple JSON format.
5047 * systemctl/loginctl/machinectl's --signal= option now accept a special
5048 value "list", which may be used to show a brief table with known
5049 process signals and their numbers.
5051 * networkctl now shows the link activation policy in status.
5053 * Various tools gained --pager/--no-pager/--json= switches to
5054 enable/disable the pager and provide JSON output.
5056 * Various tools now accept two new values for the SYSTEMD_COLORS
5057 environment variable: "16" and "256", to configure how many terminal
5058 colors are used in output.
5060 * less 568 or newer is now required for the auto-paging logic of the
5061 various tools. Hyperlink ANSI sequences in terminal output are now
5062 used even if a pager is used, and older versions of less are not able
5063 to display these sequences correctly. SYSTEMD_URLIFY=0 may be used to
5064 disable this output again.
5066 * Builds with support for separate / and /usr/ hierarchies ("split-usr"
5067 builds, non-merged-usr builds) are now officially deprecated. A
5068 warning is emitted during build. Support is slated to be removed in
5069 about a year (when the Debian Bookworm release development starts).
5071 * Systems with the legacy cgroup v1 hierarchy are now marked as
5072 "tainted", to make it clearer that using the legacy hierarchy is not
5075 * systemd-localed will now refuse to configure a keymap which is not
5076 installed in the file system. This is intended as a bug fix, but
5077 could break cases where systemd-localed was used to configure the
5078 keymap in advanced of it being installed. It is necessary to install
5079 the keymap file first.
5081 * The main git development branch has been renamed to 'main'.
5083 * mmcblk[0-9]boot[0-9] devices will no longer be probed automatically
5084 for partitions, as in the vast majority of cases they contain none
5085 and are used internally by the bootloader (eg: uboot).
5087 * systemd will now set the $SYSTEMD_EXEC_PID environment variable for
5088 spawned processes to the PID of the process itself. This may be used
5089 by programs for detecting whether they were forked off by the service
5090 manager itself or are a process forked off further down the tree.
5092 * The sd-device API gained four new calls: sd_device_get_action() to
5093 determine the uevent add/remove/change/… action the device object has
5094 been seen for, sd_device_get_seqno() to determine the uevent sequence
5095 number, sd_device_new_from_stat_rdev() to allocate a new sd_device
5096 object from stat(2) data of a device node, and sd_device_trigger() to
5097 write to the 'uevent' attribute of a device.
5099 * For most tools the --no-legend= switch has been replaced by
5100 --legend=no and --legend=yes, to force whether tables are shown with
5103 * Units acquired a new property "Markers" that takes a list of zero,
5104 one or two of the following strings: "needs-reload" and
5105 "needs-restart". These markers may be set via "systemctl
5106 set-property". Once a marker is set, "systemctl reload-or-restart
5107 --marked" may be invoked to execute the operation the units are
5108 marked for. This is useful for package managers that want to mark
5109 units for restart/reload while updating, but effect the actual
5110 operations at a later step at once.
5112 * The sd_bus_message_read_strv() API call of sd-bus may now also be
5113 used to parse arrays of D-Bus signatures and D-Bus paths, in addition
5116 * bootctl will now report whether the UEFI firmware used a TPM2 device
5117 and measured the boot process into it.
5119 * systemd-tmpfiles learnt support for a new environment variable
5120 $SYSTEMD_TMPFILES_FORCE_SUBVOL which takes a boolean value. If true
5121 the v/q/Q lines in tmpfiles.d/ snippets will create btrfs subvolumes
5122 even if the root fs of the system is not itself a btrfs volume.
5124 * systemd-detect-virt/ConditionVirtualization= will now explicitly
5125 detect Docker/Podman environments where possible. Moreover, they
5126 should be able to generically detect any container manager as long as
5127 it assigns the container a cgroup.
5129 * portablectl gained a new "reattach" verb for detaching/reattaching a
5130 portable service image, useful for updating images on-the-fly.
5132 * Intel SGX enclave device nodes (which expose a security feature of
5133 newer Intel CPUs) will now be owned by a new system group "sgx".
5135 Contributions from: Adam Nielsen, Adrian Vovk, AJ Jordan, Alan Perry,
5136 Alastair Pharo, Alexander Batischev, Ali Abdallah, Andrew Balmos,
5137 Anita Zhang, Annika Wickert, Ansgar Burchardt, Antonio Terceiro,
5138 Antonius Frie, Ardy, Arian van Putten, Ariel Fermani, Arnaud T,
5139 A S Alam, Bastien Nocera, Benjamin Berg, Benjamin Robin, Björn Daase,
5140 caoxia, Carlo Wood, Charles Lee, ChopperRob, chri2, Christian Ehrhardt,
5141 Christian Hesse, Christopher Obbard, clayton craft, corvusnix, cprn,
5142 Daan De Meyer, Daniele Medri, Daniel Rusek, Dan Sanders, Dan Streetman,
5143 Darren Ng, David Edmundson, David Tardon, Deepak Rawat, Devon Pringle,
5144 Dmitry Borodaenko, dropsignal, Einsler Lee, Endre Szabo,
5145 Evgeny Vereshchagin, Fabian Affolter, Fangrui Song, Felipe Borges,
5146 feliperodriguesfr, Felix Stupp, Florian Hülsmann, Florian Klink,
5147 Florian Westphal, Franck Bui, Frantisek Sumsal, Gablegritule,
5148 Gaël PORTAY, Gaurav, Giedrius Statkevičius, Greg Depoire-Ferrer,
5149 Gustavo Costa, Hans de Goede, Hela Basa, heretoenhance, hide,
5150 Iago López Galeiras, igo95862, Ilya Dmitrichenko, Jameer Pathan,
5151 Jan Tojnar, Jiehong, Jinyuan Si, Joerg Behrmann, John Slade,
5152 Jonathan G. Underwood, Jonathan McDowell, Josh Triplett, Joshua Watt,
5153 Julia Cartwright, Julien Humbert, Kairui Song, Karel Zak,
5154 Kevin Backhouse, Kevin P. Fleming, Khem Raj, Konomi, krissgjeng,
5155 l4gfcm, Lajos Veres, Lennart Poettering, Lincoln Ramsay, Luca Boccassi,
5156 Luca BRUNO, Lucas Werkmeister, Luka Kudra, Luna Jernberg,
5157 Marc-André Lureau, Martin Wilck, Matthias Klumpp, Matt Turner,
5158 Michael Gisbers, Michael Marley, Michael Trapp, Michal Fabik,
5159 Michał Kopeć, Michal Koutný, Michal Sekletár, Michele Guerini Rocco,
5160 Mike Gilbert, milovlad, moson-mo, Nick, nihilix-melix, Oğuz Ersen,
5161 Ondrej Mosnacek, pali, Pavel Hrdina, Pavel Sapezhko, Perry Yuan,
5162 Peter Hutterer, Pierre Dubouilh, Piotr Drąg, Pjotr Vertaalt,
5163 Richard Laager, RussianNeuroMancer, Sam Lunt, Sebastiaan van Stijn,
5164 Sergey Bugaev, shenyangyang4, simmon, Simonas Kazlauskas,
5165 Slimane Selyan Amiri, Stefan Agner, Steve Ramage, Susant Sahani,
5166 Sven Mueller, Tad Fisher, Takashi Iwai, Thomas Haller, Tom Shield,
5167 Topi Miettinen, Torsten Hilbrich, tpgxyz, Tyler Hicks, ulf-f,
5168 Ulrich Ölmann, Vincent Pelletier, Vinnie Magro, Vito Caputo, Vlad,
5169 walbit-de, Whired Planck, wouter bolsterlee, Xℹ Ruoyao, Yangyang Shen,
5170 Yuri Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek,
5171 Zmicer Turok, Дамјан Георгиевски
5173 — Berlin, 2021-03-30
5177 * KERNEL API INCOMPATIBILITY: Linux 4.14 introduced two new uevents
5178 "bind" and "unbind" to the Linux device model. When this kernel
5179 change was made, systemd-udevd was only minimally updated to handle
5180 and propagate these new event types. The introduction of these new
5181 uevents (which are typically generated for USB devices and devices
5182 needing a firmware upload before being functional) resulted in a
5183 number of issues which we so far didn't address. We hoped the kernel
5184 maintainers would themselves address these issues in some form, but
5185 that did not happen. To handle them properly, many (if not most) udev
5186 rules files shipped in various packages need updating, and so do many
5187 programs that monitor or enumerate devices with libudev or sd-device,
5188 or otherwise process uevents. Please note that this incompatibility
5189 is not fault of systemd or udev, but caused by an incompatible kernel
5190 change that happened back in Linux 4.14, but is becoming more and
5191 more visible as the new uevents are generated by more kernel drivers.
5193 To minimize issues resulting from this kernel change (but not avoid
5194 them entirely) starting with systemd-udevd 247 the udev "tags"
5195 concept (which is a concept for marking and filtering devices during
5196 enumeration and monitoring) has been reworked: udev tags are now
5197 "sticky", meaning that once a tag is assigned to a device it will not
5198 be removed from the device again until the device itself is removed
5199 (i.e. unplugged). This makes sure that any application monitoring
5200 devices that match a specific tag is guaranteed to both see uevents
5201 where the device starts being relevant, and those where it stops
5202 being relevant (the latter now regularly happening due to the new
5203 "unbind" uevent type). The udev tags concept is hence now a concept
5204 tied to a *device* instead of a device *event* — unlike for example
5205 udev properties whose lifecycle (as before) is generally tied to a
5206 device event, meaning that the previously determined properties are
5207 forgotten whenever a new uevent is processed.
5209 With the newly redefined udev tags concept, sometimes it's necessary
5210 to determine which tags are the ones applied by the most recent
5211 uevent/database update, in order to discern them from those
5212 originating from earlier uevents/database updates of the same
5213 device. To accommodate for this a new automatic property CURRENT_TAGS
5214 has been added that works similar to the existing TAGS property but
5215 only lists tags set by the most recent uevent/database
5216 update. Similarly, the libudev/sd-device API has been updated with
5217 new functions to enumerate these 'current' tags, in addition to the
5218 existing APIs that now enumerate the 'sticky' ones.
5220 To properly handle "bind"/"unbind" on Linux 4.14 and newer it is
5221 essential that all udev rules files and applications are updated to
5222 handle the new events. Specifically:
5224 • All rule files that currently use a header guard similar to
5225 ACTION!="add|change",GOTO="xyz_end" should be updated to use
5226 ACTION=="remove",GOTO="xyz_end" instead, so that the
5227 properties/tags they add are also applied whenever "bind" (or
5228 "unbind") is seen. (This is most important for all physical device
5229 types — those for which "bind" and "unbind" are currently
5230 generated, for all other device types this change is still
5231 recommended but not as important — but certainly prepares for
5232 future kernel uevent type additions).
5234 • Similarly, all code monitoring devices that contains an 'if' branch
5235 discerning the "add" + "change" uevent actions from all other
5236 uevents actions (i.e. considering devices only relevant after "add"
5237 or "change", and irrelevant on all other events) should be reworked
5238 to instead negatively check for "remove" only (i.e. considering
5239 devices relevant after all event types, except for "remove", which
5240 invalidates the device). Note that this also means that devices
5241 should be considered relevant on "unbind", even though conceptually
5242 this — in some form — invalidates the device. Since the precise
5243 effect of "unbind" is not generically defined, devices should be
5244 considered relevant even after "unbind", however I/O errors
5245 accessing the device should then be handled gracefully.
5247 • Any code that uses device tags for deciding whether a device is
5248 relevant or not most likely needs to be updated to use the new
5249 udev_device_has_current_tag() API (or sd_device_has_current_tag()
5250 in case sd-device is used), to check whether the tag is set at the
5251 moment an uevent is seen (as opposed to the existing
5252 udev_device_has_tag() API which checks if the tag ever existed on
5253 the device, following the API concept redefinition explained
5256 We are very sorry for this breakage and the requirement to update
5257 packages using these interfaces. We'd again like to underline that
5258 this is not caused by systemd/udev changes, but result of a kernel
5261 * UPCOMING INCOMPATIBILITY: So far most downstream distribution
5262 packages have not retriggered devices once the udev package (or any
5263 auxiliary package installing additional udev rules) is updated. We
5264 intend to work with major distributions to change this, so that
5265 "udevadm trigger -c change" is issued on such upgrades, ensuring that
5266 the updated ruleset is applied to the devices already discovered, so
5267 that (asynchronously) after the upgrade completed the udev database
5268 is consistent with the updated rule set. This means udev rules must
5269 be ready to be retriggered with a "change" action any time, and
5270 result in correct and complete udev database entries. While the
5271 majority of udev rule files known to us currently get this right,
5272 some don't. Specifically, there are udev rules files included in
5273 various packages that only set udev properties on the "add" action,
5274 but do not handle the "change" action. If a device matching those
5275 rules is retriggered with the "change" action (as is intended here)
5276 it would suddenly lose the relevant properties. This always has been
5277 problematic, but as soon as all udev devices are triggered on relevant
5278 package upgrades this will become particularly so. It is strongly
5279 recommended to fix offending rules so that they can handle a "change"
5280 action at any time, and acquire all necessary udev properties even
5281 then. Or in other words: the header guard mentioned above
5282 (ACTION=="remove",GOTO="xyz_end") is the correct approach to handle
5283 this, as it makes sure rules are rerun on "change" correctly, and
5284 accumulate the correct and complete set of udev properties. udev rule
5285 definitions that cannot handle "change" events being triggered at
5286 arbitrary times should be considered buggy.
5288 * The MountAPIVFS= service file setting now defaults to on if
5289 RootImage= and RootDirectory= are used, which means that with those
5290 two settings /proc/, /sys/ and /dev/ are automatically properly set
5291 up for services. Previous behaviour may be restored by explicitly
5292 setting MountAPIVFS=off.
5294 * Since PAM 1.2.0 (2015) configuration snippets may be placed in
5295 /usr/lib/pam.d/ in addition to /etc/pam.d/. If a file exists in the
5296 latter it takes precedence over the former, similar to how most of
5297 systemd's own configuration is handled. Given that PAM stack
5298 definitions are primarily put together by OS vendors/distributions
5299 (though possibly overridden by users), this systemd release moves its
5300 own PAM stack configuration for the "systemd-user" PAM service (i.e.
5301 for the PAM session invoked by the per-user user@.service instance)
5302 from /etc/pam.d/ to /usr/lib/pam.d/. We recommend moving all
5303 packages' vendor versions of their PAM stack definitions from
5304 /etc/pam.d/ to /usr/lib/pam.d/, but if such OS-wide migration is not
5305 desired the location to which systemd installs its PAM stack
5306 configuration may be changed via the -Dpamconfdir Meson option.
5308 * The runtime dependencies on libqrencode, libpcre2, libidn/libidn2,
5309 libpwquality and libcryptsetup have been changed to be based on
5310 dlopen(): instead of regular dynamic library dependencies declared in
5311 the binary ELF headers, these libraries are now loaded on demand
5312 only, if they are available. If the libraries cannot be found the
5313 relevant operations will fail gracefully, or a suitable fallback
5314 logic is chosen. This is supposed to be useful for general purpose
5315 distributions, as it allows minimizing the list of dependencies the
5316 systemd packages pull in, permitting building of more minimal OS
5317 images, while still making use of these "weak" dependencies should
5318 they be installed. Since many package managers automatically
5319 synthesize package dependencies from ELF shared library dependencies,
5320 some additional manual packaging work has to be done now to replace
5321 those (slightly downgraded from "required" to "recommended" or
5322 whatever is conceptually suitable for the package manager). Note that
5323 this change does not alter build-time behaviour: as before the
5324 build-time dependencies have to be installed during build, even if
5325 they now are optional during runtime.
5327 * sd-event.h gained a new call sd_event_add_time_relative() for
5328 installing timers relative to the current time. This is mostly a
5329 convenience wrapper around the pre-existing sd_event_add_time() call
5330 which installs absolute timers.
5332 * sd-event event sources may now be placed in a new "exit-on-failure"
5333 mode, which may be controlled via the new
5334 sd_event_source_get_exit_on_failure() and
5335 sd_event_source_set_exit_on_failure() functions. If enabled, any
5336 failure returned by the event source handler functions will result in
5337 exiting the event loop (unlike the default behaviour of just
5338 disabling the event source but continuing with the event loop). This
5339 feature is useful to set for all event sources that define "primary"
5340 program behaviour (where failure should be fatal) in contrast to
5341 "auxiliary" behaviour (where failure should remain local).
5343 * Most event source types sd-event supports now accept a NULL handler
5344 function, in which case the event loop is exited once the event
5345 source is to be dispatched, using the userdata pointer — converted to
5346 a signed integer — as exit code of the event loop. Previously this
5347 was supported for IO and signal event sources already. Exit event
5348 sources still do not support this (simply because it makes little
5349 sense there, as the event loop is already exiting when they are
5352 * A new per-unit setting RootImageOptions= has been added which allows
5353 tweaking the mount options for any file system mounted as effect of
5354 the RootImage= setting.
5356 * Another new per-unit setting MountImages= has been added, that allows
5357 mounting additional disk images into the file system tree accessible
5360 * Timer units gained a new FixedRandomDelay= boolean setting. If
5361 enabled, the random delay configured with RandomizedDelaySec= is
5362 selected in a way that is stable on a given system (though still
5363 different for different units).
5365 * Socket units gained a new setting Timestamping= that takes "us", "ns"
5366 or "off". This controls the SO_TIMESTAMP/SO_TIMESTAMPNS socket
5369 * systemd-repart now generates JSON output when requested with the new
5372 * systemd-machined's OpenMachineShell() bus call will now pass
5373 additional policy metadata data fields to the PolicyKit
5374 authentication request.
5376 * systemd-tmpfiles gained a new -E switch, which is equivalent to
5377 --exclude-prefix=/dev --exclude-prefix=/proc --exclude=/run
5378 --exclude=/sys. It's particularly useful in combination with --root=,
5379 when operating on OS trees that do not have any of these four runtime
5380 directories mounted, as this means no files below these subtrees are
5381 created or modified, since those mount points should probably remain
5384 * systemd-tmpfiles gained a new --image= switch which is like --root=,
5385 but takes a disk image instead of a directory as argument. The
5386 specified disk image is mounted inside a temporary mount namespace
5387 and the tmpfiles.d/ drop-ins stored in the image are executed and
5388 applied to the image. systemd-sysusers similarly gained a new
5389 --image= switch, that allows the sysusers.d/ drop-ins stored in the
5390 image to be applied onto the image.
5392 * Similarly, the journalctl command also gained an --image= switch,
5393 which is a quick one-step solution to look at the log data included
5396 * journalctl's --output=cat option (which outputs the log content
5397 without any metadata, just the pure text messages) will now make use
5398 of terminal colors when run on a suitable terminal, similarly to the
5401 * JSON group records now support a "description" string that may be
5402 used to add a human-readable textual description to such groups. This
5403 is supposed to match the user's GECOS field which traditionally
5404 didn't have a counterpart for group records.
5406 * The "systemd-dissect" tool that may be used to inspect OS disk images
5407 and that was previously installed to /usr/lib/systemd/ has now been
5408 moved to /usr/bin/, reflecting its updated status of an officially
5409 supported tool with a stable interface. It gained support for a new
5410 --mkdir switch which when combined with --mount has the effect of
5411 creating the directory to mount the image to if it is missing
5412 first. It also gained two new commands --copy-from and --copy-to for
5413 copying files and directories in and out of an OS image without the
5414 need to manually mount it. It also acquired support for a new option
5415 --json= to generate JSON output when inspecting an OS image.
5417 * The cgroup2 file system is now mounted with the
5418 "memory_recursiveprot" mount option, supported since kernel 5.7. This
5419 means that the MemoryLow= and MemoryMin= unit file settings now apply
5420 recursively to whole subtrees.
5422 * systemd-homed now defaults to using the btrfs file system — if
5423 available — when creating home directories in LUKS volumes. This may
5424 be changed with the DefaultFileSystemType= setting in homed.conf.
5425 It's now the default file system in various major distributions and
5426 has the major benefit for homed that it can be grown and shrunk while
5427 mounted, unlike the other contenders ext4 and xfs, which can both be
5428 grown online, but not shrunk (in fact xfs is the technically most
5429 limited option here, as it cannot be shrunk at all).
5431 * JSON user records managed by systemd-homed gained support for
5432 "recovery keys". These are basically secondary passphrases that can
5433 unlock user accounts/home directories. They are computer-generated
5434 rather than user-chosen, and typically have greater entropy.
5435 homectl's --recovery-key= option may be used to add a recovery key to
5436 a user account. The generated recovery key is displayed as a QR code,
5437 so that it can be scanned to be kept in a safe place. This feature is
5438 particularly useful in combination with systemd-homed's support for
5439 FIDO2 or PKCS#11 authentication, as a secure fallback in case the
5440 security tokens are lost. Recovery keys may be entered wherever the
5441 system asks for a password.
5443 * systemd-homed now maintains a "dirty" flag for each LUKS encrypted
5444 home directory which indicates that a home directory has not been
5445 deactivated cleanly when offline. This flag is useful to identify
5446 home directories for which the offline discard logic did not run when
5447 offlining, and where it would be a good idea to log in again to catch
5450 * systemctl gained a new parameter --timestamp= which may be used to
5451 change the style in which timestamps are output, i.e. whether to show
5452 them in local timezone or UTC, or whether to show µs granularity.
5454 * Alibaba's "pouch" container manager is now detected by
5455 systemd-detect-virt, ConditionVirtualization= and similar
5456 constructs. Similar, they now also recognize IBM PowerVM machine
5459 * systemd-nspawn has been reworked to use the /run/host/incoming/ as
5460 place to use for propagating external mounts into the
5461 container. Similarly /run/host/notify is now used as the socket path
5462 for container payloads to communicate with the container manager
5463 using sd_notify(). The container manager now uses the
5464 /run/host/inaccessible/ directory to place "inaccessible" file nodes
5465 of all relevant types which may be used by the container payload as
5466 bind mount source to over-mount inodes to make them inaccessible.
5467 /run/host/container-manager will now be initialized with the same
5468 string as the $container environment variable passed to the
5469 container's PID 1. /run/host/container-uuid will be initialized with
5470 the same string as $container_uuid. This means the /run/host/
5471 hierarchy is now the primary way to make host resources available to
5472 the container. The Container Interface documents these new files and
5475 https://systemd.io/CONTAINER_INTERFACE
5477 * Support for the "ConditionNull=" unit file condition has been
5478 deprecated and undocumented for 6 years. systemd started to warn
5479 about its use 1.5 years ago. It has now been removed entirely.
5481 * sd-bus.h gained a new API call sd_bus_error_has_names(), which takes
5482 a sd_bus_error struct and a list of error names, and checks if the
5483 error matches one of these names. It's a convenience wrapper that is
5484 useful in cases where multiple errors shall be handled the same way.
5486 * A new system call filter list "@known" has been added, that contains
5487 all system calls known at the time systemd was built.
5489 * Behaviour of system call filter allow lists has changed slightly:
5490 system calls that are contained in @known will result in EPERM by
5491 default, while those not contained in it result in ENOSYS. This
5492 should improve compatibility because known system calls will thus be
5493 communicated as prohibited, while unknown (and thus newer ones) will
5494 be communicated as not implemented, which hopefully has the greatest
5495 chance of triggering the right fallback code paths in client
5498 * "systemd-analyze syscall-filter" will now show two separate sections
5499 at the bottom of the output: system calls known during systemd build
5500 time but not included in any of the filter groups shown above, and
5501 system calls defined on the local kernel but known during systemd
5504 * If the $SYSTEMD_LOG_SECCOMP=1 environment variable is set for
5505 systemd-nspawn all system call filter violations will be logged by
5506 the kernel (audit). This is useful for tracking down system calls
5507 invoked by container payloads that are prohibited by the container's
5508 system call filter policy.
5510 * If the $SYSTEMD_SECCOMP=0 environment variable is set for
5511 systemd-nspawn (and other programs that use seccomp) all seccomp
5512 filtering is turned off.
5514 * Two new unit file settings ProtectProc= and ProcSubset= have been
5515 added that expose the hidepid= and subset= mount options of procfs.
5516 All processes of the unit will only see processes in /proc that are
5517 are owned by the unit's user. This is an important new sandboxing
5518 option that is recommended to be set on all system services. All
5519 long-running system services that are included in systemd itself set
5520 this option now. This option is only supported on kernel 5.8 and
5521 above, since the hidepid= option supported on older kernels was not a
5522 per-mount option but actually applied to the whole PID namespace.
5524 * Socket units gained a new boolean setting FlushPending=. If enabled
5525 all pending socket data/connections are flushed whenever the socket
5526 unit enters the "listening" state, i.e. after the associated service
5529 * The unit file setting NUMAMask= gained a new "all" value: when used,
5530 all existing NUMA nodes are added to the NUMA mask.
5532 * A new "credentials" logic has been added to system services. This is
5533 a simple mechanism to pass privileged data to services in a safe and
5534 secure way. It's supposed to be used to pass per-service secret data
5535 such as passwords or cryptographic keys but also associated less
5536 private information such as user names, certificates, and similar to
5537 system services. Each credential is identified by a short user-chosen
5538 name and may contain arbitrary binary data. Two new unit file
5539 settings have been added: SetCredential= and LoadCredential=. The
5540 former allows setting a credential to a literal string, the latter
5541 sets a credential to the contents of a file (or data read from a
5542 user-chosen AF_UNIX stream socket). Credentials are passed to the
5543 service via a special credentials directory, one file for each
5544 credential. The path to the credentials directory is passed in a new
5545 $CREDENTIALS_DIRECTORY environment variable. Since the credentials
5546 are passed in the file system they may be easily referenced in
5547 ExecStart= command lines too, thus no explicit support for the
5548 credentials logic in daemons is required (though ideally daemons
5549 would look for the bits they need in $CREDENTIALS_DIRECTORY
5550 themselves automatically, if set). The $CREDENTIALS_DIRECTORY is
5551 backed by unswappable memory if privileges allow it, immutable if
5552 privileges allow it, is accessible only to the service's UID, and is
5553 automatically destroyed when the service stops.
5555 * systemd-nspawn supports the same credentials logic. It can both
5556 consume credentials passed to it via the aforementioned
5557 $CREDENTIALS_DIRECTORY protocol as well as pass these credentials on
5558 to its payload. The service manager/PID 1 has been updated to match
5559 this: it can also accept credentials from the container manager that
5560 invokes it (in fact: any process that invokes it), and passes them on
5561 to its services. Thus, credentials can be propagated recursively down
5562 the tree: from a system's service manager to a systemd-nspawn
5563 service, to the service manager that runs as container payload and to
5564 the service it runs below. Credentials may also be added on the
5565 systemd-nspawn command line, using new --set-credential= and
5566 --load-credential= command line switches that match the
5567 aforementioned service settings.
5569 * systemd-repart gained new settings Format=, Encrypt=, CopyFiles= in
5570 the partition drop-ins which may be used to format/LUKS
5571 encrypt/populate any created partitions. The partitions are
5572 encrypted/formatted/populated before they are registered in the
5573 partition table, so that they appear atomically: either the
5574 partitions do not exist yet or they exist fully encrypted, formatted,
5575 and populated — there is no time window where they are
5576 "half-initialized". Thus the system is robust to abrupt shutdown: if
5577 the tool is terminated half-way during its operations on next boot it
5578 will start from the beginning.
5580 * systemd-repart's --size= operation gained a new "auto" value. If
5581 specified, and operating on a loopback file it is automatically sized
5582 to the minimal size the size constraints permit. This is useful to
5583 use "systemd-repart" as an image builder for minimally sized images.
5585 * systemd-resolved now gained a third IPC interface for requesting name
5586 resolution: besides D-Bus and local DNS to 127.0.0.53 a Varlink
5587 interface is now supported. The nss-resolve NSS module has been
5588 modified to use this new interface instead of D-Bus. Using Varlink
5589 has a major benefit over D-Bus: it works without a broker service,
5590 and thus already during earliest boot, before the dbus daemon has
5591 been started. This means name resolution via systemd-resolved now
5592 works at the same time systemd-networkd operates: from earliest boot
5593 on, including in the initrd.
5595 * systemd-resolved gained support for a new DNSStubListenerExtra=
5596 configuration file setting which may be used to specify additional IP
5597 addresses the built-in DNS stub shall listen on, in addition to the
5598 main one on 127.0.0.53:53.
5600 * Name lookups issued via systemd-resolved's D-Bus and Varlink
5601 interfaces (and thus also via glibc NSS if nss-resolve is used) will
5602 now honour a trailing dot in the hostname: if specified the search
5603 path logic is turned off. Thus "resolvectl query foo." is now
5604 equivalent to "resolvectl query --search=off foo.".
5606 * systemd-resolved gained a new D-Bus property "ResolvConfMode" that
5607 exposes how /etc/resolv.conf is currently managed: by resolved (and
5608 in which mode if so) or another subsystem. "resolvctl" will display
5609 this property in its status output.
5611 * The resolv.conf snippets systemd-resolved provides will now set "."
5612 as the search domain if no other search domain is known. This turns
5613 off the derivation of an implicit search domain by nss-dns for the
5614 hostname, when the hostname is set to an FQDN. This change is done to
5615 make nss-dns using resolv.conf provided by systemd-resolved behave
5616 more similarly to nss-resolve.
5618 * systemd-tmpfiles' file "aging" logic (i.e. the automatic clean-up of
5619 /tmp/ and /var/tmp/ based on file timestamps) now looks at the
5620 "birth" time (btime) of a file in addition to the atime, mtime, and
5623 * systemd-analyze gained a new verb "capability" that lists all known
5624 capabilities by the systemd build and by the kernel.
5626 * If a file /usr/lib/clock-epoch exists, PID 1 will read its mtime and
5627 advance the system clock to it at boot if it is noticed to be before
5628 that time. Previously, PID 1 would only advance the time to an epoch
5629 time that is set during build-time. With this new file OS builders
5630 can change this epoch timestamp on individual OS images without
5631 having to rebuild systemd.
5633 * systemd-logind will now listen to the KEY_RESTART key from the Linux
5634 input layer and reboot the system if it is pressed, similarly to how
5635 it already handles KEY_POWER, KEY_SUSPEND or KEY_SLEEP. KEY_RESTART
5636 was originally defined in the Multimedia context (to restart playback
5637 of a song or film), but is now primarily used in various embedded
5638 devices for "Reboot" buttons. Accordingly, systemd-logind will now
5639 honour it as such. This may configured in more detail via the new
5640 HandleRebootKey= and RebootKeyIgnoreInhibited=.
5642 * systemd-nspawn/systemd-machined will now reconstruct hardlinks when
5643 copying OS trees, for example in "systemd-nspawn --ephemeral",
5644 "systemd-nspawn --template=", "machinectl clone" and similar. This is
5645 useful when operating with OSTree images, which use hardlinks heavily
5646 throughout, and where such copies previously resulting in "exploding"
5649 * systemd-nspawn's --console= setting gained support for a new
5650 "autopipe" value, which is identical to "interactive" when invoked on
5651 a TTY, and "pipe" otherwise.
5653 * systemd-networkd's .network files gained support for explicitly
5654 configuring the multicast membership entries of bridge devices in the
5655 [BridgeMDB] section. It also gained support for the PIE queuing
5656 discipline in the [FlowQueuePIE] sections.
5658 * systemd-networkd's .netdev files may now be used to create "BareUDP"
5659 tunnels, configured in the new [BareUDP] setting.
5661 * systemd-networkd's Gateway= setting in .network files now accepts the
5662 special values "_dhcp4" and "_ipv6ra" to configure additional,
5663 locally defined, explicit routes to the gateway acquired via DHCP or
5664 IPv6 Router Advertisements. The old setting "_dhcp" is deprecated,
5665 but still accepted for backwards compatibility.
5667 * systemd-networkd's [IPv6PrefixDelegation] section and
5668 IPv6PrefixDelegation= options have been renamed as [IPv6SendRA] and
5669 IPv6SendRA= (the old names are still accepted for backwards
5672 * systemd-networkd's .network files gained the DHCPv6PrefixDelegation=
5673 boolean setting in [Network] section. If enabled, the delegated prefix
5674 gained by another link will be configured, and an address within the
5675 prefix will be assigned.
5677 * systemd-networkd's .network files gained the Announce= boolean setting
5678 in [DHCPv6PrefixDelegation] section. When enabled, the delegated
5679 prefix will be announced through IPv6 router advertisement (IPv6 RA).
5680 The setting is enabled by default.
5682 * VXLAN tunnels may now be marked as independent of any underlying
5683 network interface via the new Independent= boolean setting.
5685 * systemctl gained support for two new verbs: "service-log-level" and
5686 "service-log-target" may be used on services that implement the
5687 generic org.freedesktop.LogControl1 D-Bus interface to dynamically
5688 adjust the log level and target. All of systemd's long-running
5689 services support this now, but ideally all system services would
5690 implement this interface to make the system more uniformly
5693 * The SystemCallErrorNumber= unit file setting now accepts the new
5694 "kill" and "log" actions, in addition to arbitrary error number
5695 specifications as before. If "kill" the processes are killed on the
5696 event, if "log" the offending system call is audit logged.
5698 * A new SystemCallLog= unit file setting has been added that accepts a
5699 list of system calls that shall be logged about (audit).
5701 * The OS image dissection logic (as used by RootImage= in unit files or
5702 systemd-nspawn's --image= switch) has gained support for identifying
5703 and mounting explicit /usr/ partitions, which are now defined in the
5704 discoverable partition specification. This should be useful for
5705 environments where the root file system is
5706 generated/formatted/populated dynamically on first boot and combined
5707 with an immutable /usr/ tree that is supplied by the vendor.
5709 * In the final phase of shutdown, within the systemd-shutdown binary
5710 we'll now try to detach MD devices (i.e software RAID) in addition to
5711 loopback block devices and DM devices as before. This is supposed to
5712 be a safety net only, in order to increase robustness if things go
5713 wrong. Storage subsystems are expected to properly detach their
5714 storage volumes during regular shutdown already (or in case of
5715 storage backing the root file system: in the initrd hook we return to
5718 * If the SYSTEMD_LOG_TID environment variable is set all systemd tools
5719 will now log the thread ID in their log output. This is useful when
5720 working with heavily threaded programs.
5722 * If the SYSTEMD_RDRAND environment variable is set to "0", systemd will
5723 not use the RDRAND CPU instruction. This is useful in environments
5724 such as replay debuggers where non-deterministic behaviour is not
5727 * The autopaging logic in systemd's various tools (such as systemctl)
5728 has been updated to turn on "secure" mode in "less"
5729 (i.e. $LESSECURE=1) if execution in a "sudo" environment is
5730 detected. This disables invoking external programs from the pager,
5731 via the pipe logic. This behaviour may be overridden via the new
5732 $SYSTEMD_PAGERSECURE environment variable.
5734 * Units which have resource limits (.service, .mount, .swap, .slice,
5735 .socket, and .slice) gained new configuration settings
5736 ManagedOOMSwap=, ManagedOOMMemoryPressure=, and
5737 ManagedOOMMemoryPressureLimitPercent= that specify resource pressure
5738 limits and optional action taken by systemd-oomd.
5740 * A new service systemd-oomd has been added. It monitors resource
5741 contention for selected parts of the unit hierarchy using the PSI
5742 information reported by the kernel, and kills processes when memory
5743 or swap pressure is above configured limits. This service is only
5744 enabled by default in developer mode (see below) and should be
5745 considered a preview in this release. Behaviour details and option
5746 names are subject to change without the usual backwards-compatibility
5749 * A new helper oomctl has been added to introspect systemd-oomd state.
5750 It is only enabled by default in developer mode and should be
5751 considered a preview without the usual backwards-compatibility
5754 * New meson option -Dcompat-mutable-uid-boundaries= has been added. If
5755 enabled, systemd reads the system UID boundaries from /etc/login.defs
5756 at runtime, instead of using the built-in values selected during
5757 build. This is an option to improve compatibility for upgrades from
5758 old systems. It's strongly recommended not to make use of this
5759 functionality on new systems (or even enable it during build), as it
5760 makes something runtime-configurable that is mostly an implementation
5761 detail of the OS, and permits avoidable differences in deployments
5762 that create all kinds of problems in the long run.
5764 * New meson option '-Dmode=developer|release' has been added. When
5765 'developer', additional checks and features are enabled that are
5766 relevant during upstream development, e.g. verification that
5767 semi-automatically-generated documentation has been properly updated
5768 following API changes. Those checks are considered hints for
5769 developers and are not actionable in downstream builds. In addition,
5770 extra features that are not ready for general consumption may be
5771 enabled in developer mode. It is thus recommended to set
5772 '-Dmode=release' in end-user and distro builds.
5774 * systemd-cryptsetup gained support for processing detached LUKS
5775 headers specified on the kernel command line via the header=
5776 parameter of the luks.options= kernel command line option. The same
5777 device/path syntax as for key files is supported for header files
5780 * The "net_id" built-in of udev has been updated to ignore ACPI _SUN
5781 slot index data for devices that are connected through a PCI bridge
5782 where the _SUN index is associated with the bridge instead of the
5783 network device itself. Previously this would create ambiguous device
5784 naming if multiple network interfaces were connected to the same PCI
5785 bridge. Since this is a naming scheme incompatibility on systems that
5786 possess hardware like this it has been introduced as new naming
5787 scheme "v247". The previous scheme can be selected via the
5788 "net.naming-scheme=v245" kernel command line parameter.
5790 * ConditionFirstBoot= semantics have been modified to be safe towards
5791 abnormal system power-off during first boot. Specifically, the
5792 "systemd-machine-id-commit.service" service now acts as boot
5793 milestone indicating when the first boot process is sufficiently
5794 complete in order to not consider the next following boot also a
5795 first boot. If the system is reset before this unit is reached the
5796 first time, the next boot will still be considered a first boot; once
5797 it has been reached, no further boots will be considered a first
5798 boot. The "first-boot-complete.target" unit now acts as official hook
5799 point to order against this. If a service shall be run on every boot
5800 until the first boot fully succeeds it may thus be ordered before
5801 this target unit (and pull it in) and carry ConditionFirstBoot=
5804 * bootctl's set-default and set-oneshot commands now accept the three
5805 special strings "@default", "@oneshot", "@current" in place of a boot
5806 entry id. These strings are resolved to the current default and
5807 oneshot boot loader entry, as well as the currently booted one. Thus
5808 a command "bootctl set-default @current" may be used to make the
5809 currently boot menu item the new default for all subsequent boots.
5811 * "systemctl edit" has been updated to show the original effective unit
5812 contents in commented form in the text editor.
5814 * Units in user mode are now segregated into three new slices:
5815 session.slice (units that form the core of graphical session),
5816 app.slice ("normal" user applications), and background.slice
5817 (low-priority tasks). Unless otherwise configured, user units are
5818 placed in app.slice. The plan is to add resource limits and
5819 protections for the different slices in the future.
5821 * New GPT partition types for RISCV32/64 for the root and /usr
5822 partitions, and their associated Verity partitions have been defined,
5823 and are now understood by systemd-gpt-auto-generator, and the OS
5824 image dissection logic.
5826 Contributions from: Adolfo Jayme Barrientos, afg, Alec Moskvin, Alyssa
5827 Ross, Amitanand Chikorde, Andrew Hangsleben, Anita Zhang, Ansgar
5828 Burchardt, Arian van Putten, Aurelien Jarno, Axel Rasmussen, bauen1,
5829 Beniamino Galvani, Benjamin Berg, Bjørn Mork, brainrom, Chandradeep
5830 Dey, Charles Lee, Chris Down, Christian Göttsche, Christof Efkemann,
5831 Christoph Ruegge, Clemens Gruber, Daan De Meyer, Daniele Medri, Daniel
5832 Mack, Daniel Rusek, Dan Streetman, David Tardon, Dimitri John Ledkov,
5833 Dmitry Borodaenko, Elias Probst, Elisei Roca, ErrantSpore, Etienne
5834 Doms, Fabrice Fontaine, fangxiuning, Felix Riemann, Florian Klink,
5835 Franck Bui, Frantisek Sumsal, fwSmit, George Rawlinson, germanztz,
5836 Gibeom Gwon, Glen Whitney, Gogo Gogsi, Göran Uddeborg, Grant Mathews,
5837 Hans de Goede, Hans Ulrich Niedermann, Haochen Tong, Harald Seiler,
5838 huangyong, Hubert Kario, igo95862, Ikey Doherty, Insun Pyo, Jan Chren,
5839 Jan Schlüter, Jérémy Nouhaud, Jian-Hong Pan, Joerg Behrmann, Jonathan
5840 Lebon, Jörg Thalheim, Josh Brobst, Juergen Hoetzel, Julien Humbert,
5841 Kai-Chuan Hsieh, Kairui Song, Kamil Dudka, Kir Kolyshkin, Kristijan
5842 Gjoshev, Kyle Huey, Kyle Russell, Lee Whalen, Lennart Poettering,
5843 lichangze, Luca Boccassi, Lucas Werkmeister, Luca Weiss, Marc
5844 Kleine-Budde, Marco Wang, Martin Wilck, Marti Raudsepp, masmullin2000,
5845 Máté Pozsgay, Matt Fenwick, Michael Biebl, Michael Scherer, Michal
5846 Koutný, Michal Sekletár, Michal Suchanek, Mikael Szreder, Milo
5847 Casagrande, mirabilos, Mitsuha_QuQ, mog422, Muhammet Kara, Nazar
5848 Vinnichuk, Nicholas Narsing, Nicolas Fella, Njibhu, nl6720, Oğuz Ersen,
5849 Olivier Le Moal, Ondrej Kozina, onlybugreports, Pass Automated Testing
5850 Suite, Pat Coulthard, Pavel Sapezhko, Pedro Ruiz, perry_yuan, Peter
5851 Hutterer, Phaedrus Leeds, PhoenixDiscord, Piotr Drąg, Plan C,
5852 Purushottam choudhary, Rasmus Villemoes, Renaud Métrich, Robert Marko,
5853 Roman Beranek, Ronan Pigott, Roy Chen (陳彥廷), RussianNeuroMancer,
5854 Samanta Navarro, Samuel BF, scootergrisen, Sorin Ionescu, Steve Dodd,
5855 Susant Sahani, Timo Rothenpieler, Tobias Hunger, Tobias Kaufmann, Topi
5856 Miettinen, vanou, Vito Caputo, Weblate, Wen Yang, Whired Planck,
5857 williamvds, Yu, Li-Yu, Yuri Chornoivan, Yu Watanabe, Zbigniew
5858 Jędrzejewski-Szmek, Zmicer Turok, Дамјан Георгиевски
5860 – Warsaw, 2020-11-26
5864 * The service manager gained basic support for cgroup v2 freezer. Units
5865 can now be suspended or resumed either using new systemctl verbs,
5866 freeze and thaw respectively, or via D-Bus.
5868 * PID 1 may now automatically load pre-compiled AppArmor policies from
5869 /etc/apparmor/earlypolicy during early boot.
5871 * The CPUAffinity= setting in service unit files now supports a new
5872 special value "numa" that causes the CPU affinity masked to be set
5873 based on the NUMA mask.
5875 * systemd will now log about all left-over processes remaining in a
5876 unit when the unit is stopped. It will now warn about services using
5877 KillMode=none, as this is generally an unsafe thing to make use of.
5879 * Two new unit file settings
5880 ConditionPathIsEncrypted=/AssertPathIsEncrypted= have been
5881 added. They may be used to check whether a specific file system path
5882 resides on a block device that is encrypted on the block level
5883 (i.e. using dm-crypt/LUKS).
5885 * Another pair of new settings ConditionEnvironment=/AssertEnvironment=
5886 has been added that may be used for simple environment checks. This
5887 is particularly useful when passing in environment variables from a
5888 container manager (or from PAM in case of the systemd --user
5891 * .service unit files now accept a new setting CoredumpFilter= which
5892 allows configuration of the memory sections coredumps of the
5893 service's processes shall include.
5895 * .mount units gained a new ReadWriteOnly= boolean option. If set
5896 it will not be attempted to mount a file system read-only if mounting
5897 in read-write mode doesn't succeed. An option x-systemd.rw-only is
5898 available in /etc/fstab to control the same.
5900 * .socket units gained a new boolean setting PassPacketInfo=. If
5901 enabled, the kernel will attach additional per-packet metadata to all
5902 packets read from the socket, as an ancillary message. This controls
5903 the IP_PKTINFO, IPV6_RECVPKTINFO, NETLINK_PKTINFO socket options,
5904 depending on socket type.
5906 * .service units gained a new setting RootHash= which may be used to
5907 specify the root hash for verity enabled disk images which are
5908 specified in RootImage=. RootVerity= may be used to specify a path to
5909 the Verity data matching a RootImage= file system. (The latter is
5910 only useful for images that do not contain the Verity data embedded
5911 into the same image that carries a GPT partition table following the
5912 Discoverable Partition Specification). Similarly, systemd-nspawn
5913 gained a new switch --verity-data= that takes a path to a file with
5914 the verity data of the disk image supplied in --image=, if the image
5915 doesn't contain the verity data itself.
5917 * .service units gained a new setting RootHashSignature= which takes
5918 either a base64 encoded PKCS#7 signature of the root hash specified
5919 with RootHash=, or a path to a file to read the signature from. This
5920 allows validation of the root hash against public keys available in
5921 the kernel keyring, and is only supported on recent kernels
5922 (>= 5.4)/libcryptsetup (>= 2.30). A similar switch has been added to
5923 systemd-nspawn and systemd-dissect (--root-hash-sig=). Support for
5924 this mechanism has also been added to systemd-veritysetup.
5926 * .service unit files gained two new options
5927 TimeoutStartFailureMode=/TimeoutStopFailureMode= that may be used to
5928 tune behaviour if a start or stop timeout is hit, i.e. whether to
5929 terminate the service with SIGTERM, SIGABRT or SIGKILL.
5931 * Most options in systemd that accept hexadecimal values prefixed with
5932 0x in additional to the usual decimal notation now also support octal
5933 notation when the 0o prefix is used and binary notation if the 0b
5936 * Various command line parameters and configuration file settings that
5937 configure key or certificate files now optionally take paths to
5938 AF_UNIX sockets in the file system. If configured that way a stream
5939 connection is made to the socket and the required data read from
5940 it. This is a simple and natural extension to the existing regular
5941 file logic, and permits other software to provide keys or
5942 certificates via simple IPC services, for example when unencrypted
5943 storage on disk is not desired. Specifically, systemd-networkd's
5944 Wireguard and MACSEC key file settings as well as
5945 systemd-journal-gatewayd's and systemd-journal-remote's PEM
5946 key/certificate parameters support this now.
5948 * Unit files, tmpfiles.d/ snippets, sysusers.d/ snippets and other
5949 configuration files that support specifier expansion learnt six new
5950 specifiers: %a resolves to the current architecture, %o/%w/%B/%W
5951 resolve to the various ID fields from /etc/os-release, %l resolves to
5952 the "short" hostname of the system, i.e. the hostname configured in
5953 the kernel truncated at the first dot.
5955 * Support for the .include syntax in unit files has been removed. The
5956 concept has been obsolete for 6 years and we started warning about
5957 its pending removal 2 years ago (also see NEWS file below). It's
5960 * StandardError= and StandardOutput= in unit files no longer support
5961 the "syslog" and "syslog-console" switches. They were long removed
5962 from the documentation, but will now result in warnings when used,
5963 and be converted to "journal" and "journal+console" automatically.
5965 * If the service setting User= is set to the "nobody" user, a warning
5966 message is now written to the logs (but the value is nonetheless
5967 accepted). Setting User=nobody is unsafe, since the primary purpose
5968 of the "nobody" user is to own all files whose owner cannot be mapped
5969 locally. It's in particular used by the NFS subsystem and in user
5970 namespacing. By running a service under this user's UID it might get
5971 read and even write access to all these otherwise unmappable files,
5972 which is quite likely a major security problem.
5974 * tmpfs mounts automatically created by systemd (/tmp, /run, /dev/shm,
5975 and others) now have a size and inode limits applied (50% of RAM for
5976 /tmp and /dev/shm, 10% of RAM for other mounts, etc.). Please note
5977 that the implicit kernel default is 50% too, so there is no change
5978 in the size limit for /tmp and /dev/shm.
5980 * nss-mymachines lost support for resolution of users and groups, and
5981 now only does resolution of hostnames. This functionality is now
5982 provided by nss-systemd. Thus, the 'mymachines' entry should be
5983 removed from the 'passwd:' and 'group:' lines in /etc/nsswitch.conf
5984 (and 'systemd' added if it is not already there).
5986 * A new kernel command line option systemd.hostname= has been added
5987 that allows controlling the hostname that is initialized early during
5990 * A kernel command line option "udev.blockdev_read_only" has been
5991 added. If specified all hardware block devices that show up are
5992 immediately marked as read-only by udev. This option is useful for
5993 making sure that a specific boot under no circumstances modifies data
5994 on disk. Use "blockdev --setrw" to undo the effect of this, per
5997 * A new boolean kernel command line option systemd.swap= has been
5998 added, which may be used to turn off automatic activation of swap
5999 devices listed in /etc/fstab.
6001 * New kernel command line options systemd.condition-needs-update= and
6002 systemd.condition-first-boot= have been added, which override the
6003 result of the ConditionNeedsUpdate= and ConditionFirstBoot=
6006 * A new kernel command line option systemd.clock-usec= has been added
6007 that allows setting the system clock to the specified time in µs
6008 since Jan 1st, 1970 early during boot. This is in particular useful
6009 in order to make test cases more reliable.
6011 * The fs.suid_dumpable sysctl is set to 2 / "suidsafe". This allows
6012 systemd-coredump to save core files for suid processes. When saving
6013 the core file, systemd-coredump will use the effective uid and gid of
6014 the process that faulted.
6016 * The /sys/module/kernel/parameters/crash_kexec_post_notifiers file is
6017 now automatically set to "Y" at boot, in order to enable pstore
6018 generation for collection with systemd-pstore.
6020 * We provide a set of udev rules to enable auto-suspend on PCI and USB
6021 devices that were tested to correctly support it. Previously, this
6022 was distributed as a set of udev rules, but has now been replaced by
6023 by a set of hwdb entries (and a much shorter udev rule to take action
6024 if the device modalias matches one of the new hwdb entries).
6026 As before, entries are periodically imported from the database
6027 maintained by the ChromiumOS project. If you have a device that
6028 supports auto-suspend correctly and where it should be enabled by
6029 default, please submit a patch that adds it to the database (see
6030 /usr/lib/udev/hwdb.d/60-autosuspend.hwdb).
6032 * systemd-udevd gained the new configuration option timeout_signal= as well
6033 as a corresponding kernel command line option udev.timeout_signal=.
6034 The option can be used to configure the UNIX signal that the main
6035 daemon sends to the worker processes on timeout. Setting the signal
6036 to SIGABRT is useful for debugging.
6038 * .link files managed by systemd-udevd gained options RxFlowControl=,
6039 TxFlowControl=, AutoNegotiationFlowControl= in the [Link] section, in
6040 order to configure various flow control parameters. They also gained
6041 RxMiniBufferSize= and RxJumboBufferSize= in order to configure jumbo
6042 frame ring buffer sizes.
6044 * networkd.conf gained a new boolean setting ManageForeignRoutes=. If
6045 enabled systemd-networkd manages all routes configured by other tools.
6047 * .network files managed by systemd-networkd gained a new section
6048 [SR-IOV], in order to configure SR-IOV capable network devices.
6050 * systemd-networkd's [IPv6Prefix] section in .network files gained a
6051 new boolean setting Assign=. If enabled an address from the prefix is
6052 automatically assigned to the interface.
6054 * systemd-networkd gained a new section [DHCPv6PrefixDelegation] which
6055 controls delegated prefixes assigned by DHCPv6 client. The section
6056 has three settings: SubnetID=, Assign=, and Token=. The setting
6057 SubnetID= allows explicit configuration of the preferred subnet that
6058 systemd-networkd's Prefix Delegation logic assigns to interfaces. If
6059 Assign= is enabled (which is the default) an address from any acquired
6060 delegated prefix is automatically chosen and assigned to the
6061 interface. The setting Token= specifies an optional address generation
6064 * systemd-networkd's [Network] section gained a new setting
6065 IPv4AcceptLocal=. If enabled the interface accepts packets with local
6068 * systemd-networkd gained support for configuring the HTB queuing
6069 discipline in the [HierarchyTokenBucket] and
6070 [HierarchyTokenBucketClass] sections. Similar the "pfifo" qdisc may
6071 be configured in the [PFIFO] section, "GRED" in
6072 [GenericRandomEarlyDetection], "SFB" in [StochasticFairBlue], "cake"
6073 in [CAKE], "PIE" in [PIE], "DRR" in [DeficitRoundRobinScheduler] and
6074 [DeficitRoundRobinSchedulerClass], "BFIFO" in [BFIFO],
6075 "PFIFOHeadDrop" in [PFIFOHeadDrop], "PFIFOFast" in [PFIFOFast], "HHF"
6076 in [HeavyHitterFilter], "ETS" in [EnhancedTransmissionSelection] and
6077 "QFQ" in [QuickFairQueueing] and [QuickFairQueueingClass].
6079 * systemd-networkd gained support for a new Termination= setting in the
6080 [CAN] section for configuring the termination resistor. It also
6081 gained a new ListenOnly= setting for controlling whether to only
6082 listen on CAN interfaces, without interfering with traffic otherwise
6083 (which is useful for debugging/monitoring CAN network
6084 traffic). DataBitRate=, DataSamplePoint=, FDMode=, FDNonISO= have
6085 been added to configure various CAN-FD aspects.
6087 * systemd-networkd's [DHCPv6] section gained a new option WithoutRA=.
6088 When enabled, DHCPv6 will be attempted right-away without requiring an
6089 Router Advertisement packet suggesting it first (i.e. without the 'M'
6090 or 'O' flags set). The [IPv6AcceptRA] section gained a boolean option
6091 DHCPv6Client= that may be used to turn off the DHCPv6 client even if
6092 the RA packets suggest it.
6094 * systemd-networkd's [DHCPv4] section gained a new setting UseGateway=
6095 which may be used to turn off use of the gateway information provided
6096 by the DHCP lease. A new FallbackLeaseLifetimeSec= setting may be
6097 used to configure how to process leases that lack a lifetime option.
6099 * systemd-networkd's [DHCPv4] and [DHCPServer] sections gained a new
6100 setting SendVendorOption= allowing configuration of additional vendor
6101 options to send in the DHCP requests/responses. The [DHCPv6] section
6102 gained a new SendOption= setting for sending arbitrary DHCP
6103 options. RequestOptions= has been added to request arbitrary options
6104 from the server. UserClass= has been added to set the DHCP user class
6107 * systemd-networkd's [DHCPServer] section gained a new set of options
6108 EmitPOP3=/POP3=, EmitSMTP=/SMTP=, EmitLPR=/LPR= for including server
6109 information about these three protocols in the DHCP lease. It also
6110 gained support for including "MUD" URLs ("Manufacturer Usage
6111 Description"). Support for "MUD" URLs was also added to the LLDP
6112 stack, configurable in the [LLDP] section in .network files.
6114 * The Mode= settings in [MACVLAN] and [MACVTAP] now support 'source'
6115 mode. Also, the sections now support a new setting SourceMACAddress=.
6117 * systemd-networkd's .netdev files now support a new setting
6118 VLANProtocol= in the [Bridge] section that allows configuration of
6119 the VLAN protocol to use.
6121 * systemd-networkd supports a new Group= setting in the [Link] section
6122 of the .network files, to control the link group.
6124 * systemd-networkd's [Network] section gained a new
6125 IPv6LinkLocalAddressGenerationMode= setting, which specifies how IPv6
6126 link local address is generated.
6128 * A new default .network file is now shipped that matches TUN/TAP
6129 devices that begin with "vt-" in their name. Such interfaces will
6130 have IP routing onto the host links set up automatically. This is
6131 supposed to be used by VM managers to trivially acquire a network
6132 interface which is fully set up for host communication, simply by
6133 carefully picking an interface name to use.
6135 * systemd-networkd's [DHCPv6] section gained a new setting RouteMetric=
6136 which sets the route priority for routes specified by the DHCP server.
6138 * systemd-networkd's [DHCPv6] section gained a new setting VendorClass=
6139 which configures the vendor class information sent to DHCP server.
6141 * The BlackList= settings in .network files' [DHCPv4] and
6142 [IPv6AcceptRA] sections have been renamed DenyList=. The old names
6143 are still understood to provide compatibility.
6145 * networkctl gained the new "forcerenew" command for forcing all DHCP
6146 server clients to renew their lease. The interface "status" output
6147 will now show numerous additional fields of information about an
6148 interface. There are new "up" and "down" commands to bring specific
6149 interfaces up or down.
6151 * systemd-resolved's DNS= configuration option now optionally accepts a
6152 port number (after ":") and a host name (after "#"). When the host
6153 name is specified, the DNS-over-TLS certificate is validated to match
6154 the specified hostname. Additionally, in case of IPv6 addresses, an
6155 interface may be specified (after "%").
6157 * systemd-resolved may be configured to forward single-label DNS names.
6158 This is not standard-conformant, but may make sense in setups where
6159 public DNS servers are not used.
6161 * systemd-resolved's DNS-over-TLS support gained SNI validation.
6163 * systemd-nspawn's --resolv-conf= switch gained a number of new
6164 supported values. Specifically, options starting with "replace-" are
6165 like those prefixed "copy-" but replace any existing resolv.conf
6166 file. And options ending in "-uplink" and "-stub" can now be used to
6167 propagate other flavours of resolv.conf into the container (as
6168 defined by systemd-resolved).
6170 * The various programs included in systemd can now optionally output
6171 their log messages on stderr prefixed with a timestamp, controlled by
6172 the $SYSTEMD_LOG_TIME environment variable.
6174 * systemctl gained a new "-P" switch that is a shortcut for "--value
6177 * "systemctl list-units" and "systemctl list-machines" no longer hide
6178 their first output column with --no-legend. To hide the first column,
6181 * "systemctl reboot" takes the option "--reboot-argument=".
6182 The optional positional argument to "systemctl reboot" is now
6183 being deprecated in favor of this option.
6185 * systemd-run gained a new switch --slice-inherit. If specified the
6186 unit it generates is placed in the same slice as the systemd-run
6189 * systemd-journald gained support for zstd compression of large fields
6190 in journal files. The hash tables in journal files have been hardened
6191 against hash collisions. This is an incompatible change and means
6192 that journal files created with new systemd versions are not readable
6193 with old versions. If the $SYSTEMD_JOURNAL_KEYED_HASH boolean
6194 environment variable for systemd-journald.service is set to 0 this
6195 new hardening functionality may be turned off, so that generated
6196 journal files remain compatible with older journalctl
6199 * journalctl will now include a clickable link in the default output for
6200 each log message for which a URL with further documentation is
6201 known. This is only supported on terminal emulators that support
6202 clickable hyperlinks, and is turned off if a pager is used (since
6203 "less" still doesn't support hyperlinks,
6204 unfortunately). Documentation URLs may be included in log messages
6205 either by including a DOCUMENTATION= journal field in it, or by
6206 associating a journal message catalog entry with the log message's
6207 MESSAGE_ID, which then carries a "Documentation:" tag.
6209 * journald.conf gained a new boolean setting Audit= that may be used to
6210 control whether systemd-journald will enable audit during
6213 * when systemd-journald's log stream is broken up into multiple lines
6214 because the PID of the sender changed this is indicated in the
6215 generated log records via the _LINE_BREAK=pid-change field.
6217 * journalctl's "-o cat" output mode will now show one or more journal
6218 fields specified with --output-fields= instead of unconditionally
6219 MESSAGE=. This is useful to retrieve a very specific set of fields
6220 without any decoration.
6222 * The sd-journal.h API gained two new functions:
6223 sd_journal_enumerate_available_unique() and
6224 sd_journal_enumerate_available_data() that operate like their
6225 counterparts that lack the _available_ in the name, but skip items
6226 that cannot be read and processed by the local implementation
6227 (i.e. are compressed in an unsupported format or such),
6229 * coredumpctl gained a new --file= switch, matching the same one in
6230 journalctl: a specific journal file may be specified to read the
6233 * coredumps collected by systemd-coredump may now be compressed using
6236 * systemd-binfmt gained a new switch --unregister for unregistering all
6237 registered entries at once. This is now invoked automatically at
6238 shutdown, so that binary formats registered with the "F" flag will
6239 not block clean file system unmounting.
6241 * systemd-notify's --pid= switch gained new values: "parent", "self",
6242 "auto" for controlling which PID to send to the service manager: the
6243 systemd-notify process' PID, or the one of the process invoking it.
6245 * systemd-logind's Session bus object learnt a new method call
6246 SetType() for temporarily updating the session type of an already
6247 allocated session. This is useful for upgrading tty sessions to
6248 graphical ones once a compositor is invoked.
6250 * systemd-socket-proxy gained a new switch --exit-idle-time= for
6251 configuring an exit-on-idle time.
6253 * systemd-repart's --empty= setting gained a new value "create". If
6254 specified a new empty regular disk image file is created under the
6255 specified name. Its size may be specified with the new --size=
6256 option. The latter is also supported without the "create" mode, in
6257 order to grow existing disk image files to the specified size. These
6258 two new options are useful when creating or manipulating disk images
6259 instead of operating on actual block devices.
6261 * systemd-repart drop-ins now support a new UUID= setting to control
6262 the UUID to assign to a newly created partition.
6264 * systemd-repart's SizeMin= per-partition parameter now defaults to 10M
6267 * systemd-repart's Label= setting now support the usual, simple
6268 specifier expansion.
6270 * systemd-homed's LUKS backend gained the ability to discard empty file
6271 system blocks automatically when the user logs out. This is enabled
6272 by default to ensure that home directories take minimal space when
6273 logged out but get full size guarantees when logged in. This may be
6274 controlled with the new --luks-offline-discard= switch to homectl.
6276 * If systemd-homed detects that /home/ is encrypted as a whole it will
6277 now default to the directory or subvolume backends instead of the
6278 LUKS backend, in order to avoid double encryption. The default
6279 storage and file system may now be configured explicitly, too, via
6280 the new /etc/systemd/homed.conf configuration file.
6282 * systemd-homed now supports unlocking home directories with FIDO2
6283 security tokens that support the 'hmac-secret' extension, in addition
6284 to the existing support for PKCS#11 security token unlocking
6285 support. Note that many recent hardware security tokens support both
6286 interfaces. The FIDO2 support is accessible via homectl's
6287 --fido2-device= option.
6289 * homectl's --pkcs11-uri= setting now accepts two special parameters:
6290 if "auto" is specified and only one suitable PKCS#11 security token
6291 is plugged in, its URL is automatically determined and enrolled for
6292 unlocking the home directory. If "list" is specified a brief table of
6293 suitable PKCS#11 security tokens is shown. Similar, the new
6294 --fido2-device= option also supports these two special values, for
6295 automatically selecting and listing suitable FIDO2 devices.
6297 * The /etc/crypttab tmp option now optionally takes an argument
6298 selecting the file system to use. Moreover, the default is now
6299 changed from ext2 to ext4.
6301 * There's a new /etc/crypttab option "keyfile-erase". If specified the
6302 key file listed in the same line is removed after use, regardless if
6303 volume activation was successful or not. This is useful if the key
6304 file is only acquired transiently at runtime and shall be erased
6305 before the system continues to boot.
6307 * There's also a new /etc/crypttab option "try-empty-password". If
6308 specified, before asking the user for a password it is attempted to
6309 unlock the volume with an empty password. This is useful for
6310 installing encrypted images whose password shall be set on first boot
6311 instead of at installation time.
6313 * systemd-cryptsetup will now attempt to load the keys to unlock
6314 volumes with automatically from files in
6315 /etc/cryptsetup-keys.d/<volume>.key and
6316 /run/cryptsetup-keys.d/<volume>.key, if any of these files exist.
6318 * systemd-cryptsetup may now activate Microsoft BitLocker volumes via
6319 /etc/crypttab, during boot.
6321 * logind.conf gained a new RuntimeDirectoryInodesMax= setting to
6322 control the inode limit for the per-user $XDG_RUNTIME_DIR tmpfs
6325 * A new generator systemd-xdg-autostart-generator has been added. It
6326 generates systemd unit files from XDG autostart .desktop files, and
6327 may be used to let the systemd user instance manage services that are
6328 started automatically as part of the desktop session.
6330 * "bootctl" gained a new verb "reboot-to-firmware" that may be used
6331 to query and change the firmware's 'Reboot Into Firmware Interface'
6334 * systemd-firstboot gained a new switch --kernel-command-line= that may
6335 be used to initialize the /etc/kernel/cmdline file of the image. It
6336 also gained a new switch --root-password-hashed= which is like
6337 --root-password= but accepts a pre-hashed UNIX password as
6338 argument. The new option --delete-root-password may be used to unset
6339 any password for the root user (dangerous!). The --root-shell= switch
6340 may be used to control the shell to use for the root account. A new
6341 --force option may be used to override any already set settings with
6342 the parameters specified on the command line (by default, the tool
6343 will not override what has already been set before, i.e. is purely
6346 * systemd-firstboot gained support for a new --image= switch, which is
6347 similar to --root= but accepts the path to a disk image file, on
6348 which it then operates.
6350 * A new sd-path.h API has been added to libsystemd. It provides a
6351 simple API for retrieving various search paths and primary
6352 directories for various resources.
6354 * A new call sd_notify_barrier() has been added to the sd-daemon.h
6355 API. The call will block until all previously sent sd_notify()
6356 messages have been processed by the service manager. This is useful
6357 to remove races caused by a process already having disappeared at the
6358 time a notification message is processed by the service manager,
6359 making correct attribution impossible. The systemd-notify tool will
6360 now make use of this call implicitly, but this can be turned off again
6361 via the new --no-block switch.
6363 * When sending a file descriptor (fd) to the service manager to keep
6364 track of, using the sd_notify() mechanism, a new parameter FDPOLL=0
6365 may be specified. If passed the service manager will refrain from
6366 poll()ing on the file descriptor. Traditionally (and when the
6367 parameter is not specified), the service manager will poll it for
6368 POLLHUP or POLLERR events, and immediately close the fds in that
6371 * The service manager (PID1) gained a new D-Bus method call
6372 SetShowStatus() which may be used to control whether it shall show
6373 boot-time status output on the console. This method has a similar
6374 effect to sending SIGRTMIN+20/SIGRTMIN+21 to PID 1.
6376 * The sd-bus API gained a number of convenience functions that take
6377 va_list arguments rather than "...". For example, there's now
6378 sd_bus_call_methodv() to match sd_bus_call_method(). Those calls make
6379 it easier to build wrappers that accept variadic arguments and want
6380 to pass a ready va_list structure to sd-bus.
6382 * sd-bus vtable entries can have a new SD_BUS_VTABLE_ABSOLUTE_OFFSET
6383 flag which alters how the userdata pointer to pass to the callbacks
6384 is determined. When the flag is set, the offset field is converted
6385 as-is into a pointer, without adding it to the object pointer the
6386 vtable is associated with.
6388 * sd-bus now exposes four new functions:
6389 sd_bus_interface_name_is_valid() + sd_bus_service_name_is_valid() +
6390 sd_bus_member_name_is_valid() + sd_bus_object_path_is_valid() will
6391 validate strings to check if they qualify as various D-Bus concepts.
6393 * The sd-bus API gained the SD_BUS_METHOD_WITH_ARGS(),
6394 SD_BUS_METHOD_WITH_ARGS_OFFSET() and SD_BUS_SIGNAL_WITH_ARGS() macros
6395 that simplify adding argument names to D-Bus methods and signals.
6397 * The man pages for the sd-bus and sd-hwdb APIs have been completed.
6399 * Various D-Bus APIs of systemd daemons now have man pages that
6400 document the methods, signals and properties.
6402 * The expectations on user/group name syntax are now documented in
6403 detail; documentation on how classic home directories may be
6404 converted into home directories managed by homed has been added;
6405 documentation regarding integration of homed/userdb functionality in
6406 desktops has been added:
6408 https://systemd.io/USER_NAMES
6409 https://systemd.io/CONVERTING_TO_HOMED
6410 https://systemd.io/USERDB_AND_DESKTOPS
6412 * Documentation for the on-disk Journal file format has been updated
6413 and has now moved to:
6415 https://systemd.io/JOURNAL_FILE_FORMAT
6417 * The interface for containers (https://systemd.io/CONTAINER_INTERFACE)
6418 has been extended by a set of environment variables that expose
6419 select fields from the host's os-release file to the container
6420 payload. Similarly, host's os-release files can be mounted into the
6421 container underneath /run/host. Together, those mechanisms provide a
6422 standardized way to expose information about the host to the
6423 container payload. Both interfaces are implemented in systemd-nspawn.
6425 * All D-Bus services shipped in systemd now implement the generic
6426 LogControl1 D-Bus API which allows clients to change log level +
6427 target of the service during runtime.
6429 * Only relevant for developers: the mkosi.default symlink has been
6430 dropped from version control. Please create a symlink to one of the
6431 distribution-specific defaults in .mkosi/ based on your preference.
6433 Contributions from: 24bisquitz, Adam Nielsen, Alan Perry, Alexander
6434 Malafeev, Amitanand.Chikorde, Alin Popa, Alvin Šipraga, Amos Bird,
6435 Andreas Rammhold, AndreRH, Andrew Doran, Anita Zhang, Ankit Jain,
6436 antznin, Arnaud Ferraris, Arthur Moraes do Lago, Arusekk, Balaji
6437 Punnuru, Balint Reczey, Bastien Nocera, bemarek, Benjamin Berg,
6438 Benjamin Dahlhoff, Benjamin Robin, Chris Down, Chris Kerr, Christian
6439 Göttsche, Christian Hesse, Christian Oder, Ciprian Hacman, Clinton Roy,
6440 codicodi, Corey Hinshaw, Daan De Meyer, Dana Olson, Dan Callaghan,
6441 Daniel Fullmer, Daniel Rusek, Dan Streetman, Dave Reisner, David
6442 Edmundson, David Wood, Denis Pronin, Diego Escalante Urrelo, Dimitri
6443 John Ledkov, dolphrundgren, duguxy, Einsler Lee, Elisei Roca, Emmanuel
6444 Garette, Eric Anderson, Eric DeVolder, Evgeny Vereshchagin,
6445 ExtinctFire, fangxiuning, Ferran Pallarès Roca, Filipe Brandenburger,
6446 Filippo Falezza, Finn, Florian Klink, Florian Mayer, Franck Bui,
6447 Frantisek Sumsal, gaurav, Georg Müller, Gergely Polonkai, Giedrius
6448 Statkevičius, Gigadoc2, gogogogi, Gaurav Singh, gzjsgdsb, Hans de
6449 Goede, Haochen Tong, ianhi, ignapk, Jakov Smolic, James T. Lee, Jan
6450 Janssen, Jan Klötzke, Jan Palus, Jay Burger, Jeremy Cline, Jérémy
6451 Rosen, Jian-Hong Pan, Jiri Slaby, Joel Shapiro, Joerg Behrmann, Jörg
6452 Thalheim, Jouke Witteveen, Kai-Heng Feng, Kenny Levinsen, Kevin
6453 Kuehler, Kumar Kartikeya Dwivedi, layderv, laydervus, Lénaïc Huard,
6454 Lennart Poettering, Lidong Zhong, Luca Boccassi, Luca BRUNO, Lucas
6455 Werkmeister, Lukas Klingsbo, Lukáš Nykrýn, Łukasz Stelmach, Maciej
6456 S. Szmigiero, MadMcCrow, Marc-André Lureau, Marcel Holtmann, Marc
6457 Kleine-Budde, Martin Hundebøll, Matthew Leeds, Matt Ranostay, Maxim
6458 Fomin, MaxVerevkin, Michael Biebl, Michael Chapman, Michael Gubbels,
6459 Michael Marley, Michał Bartoszkiewicz, Michal Koutný, Michal Sekletár,
6460 Mike Gilbert, Mike Kazantsev, Mikhail Novosyolov, ml, Motiejus Jakštys,
6461 nabijaczleweli, nerdopolis, Niccolò Maggioni, Niklas Hambüchen, Norbert
6462 Lange, Paul Cercueil, pelzvieh, Peter Hutterer, Piero La Terza, Pieter
6463 Lexis, Piotr Drąg, Rafael Fontenelle, Richard Petri, Ronan Pigott, Ross
6464 Lagerwall, Rubens Figueiredo, satmandu, Sean-StarLabs, Sebastian
6465 Jennen, sterlinghughes, Surhud More, Susant Sahani, szb512, Thomas
6466 Haller, Tobias Hunger, Tom, Tomáš Pospíšek, Tomer Shechner, Tom Hughes,
6467 Topi Miettinen, Tudor Roman, Uwe Kleine-König, Valery0xff, Vito Caputo,
6468 Vladimir Panteleev, Vladyslav Tronko, Wen Yang, Yegor Vialov, Yigal
6469 Korman, Yi Gao, YmrDtnJu, Yuri Chornoivan, Yu Watanabe, Zbigniew
6470 Jędrzejewski-Szmek, Zhu Li, Дамјан Георгиевски, наб
6472 – Warsaw, 2020-07-30
6476 * A new tool "systemd-repart" has been added, that operates as an
6477 idempotent declarative repartitioner for GPT partition tables.
6478 Specifically, a set of partitions that must or may exist can be
6479 configured via drop-in files, and during every boot the partition
6480 table on disk is compared with these files, creating missing
6481 partitions or growing existing ones based on configurable relative
6482 and absolute size constraints. The tool is strictly incremental,
6483 i.e. does not delete, shrink or move partitions, but only adds and
6484 grows them. The primary use-case is OS images that ship in minimized
6485 form, that on first boot are grown to the size of the underlying
6486 block device or augmented with additional partitions. For example,
6487 the root partition could be extended to cover the whole disk, or a
6488 swap or /home partitions could be added on first boot. It can also be
6489 used for systems that use an A/B update scheme but ship images with
6490 just the A partition, with B added on first boot. The tool is
6491 primarily intended to be run in the initrd, shortly before
6492 transitioning into the host OS, but can also be run after the
6493 transition took place. It automatically discovers the disk backing
6494 the root file system, and should hence not require any additional
6495 configuration besides the partition definition drop-ins. If no
6496 configuration drop-ins are present, no action is taken.
6498 * A new component "userdb" has been added, along with a small daemon
6499 "systemd-userdbd.service" and a client tool "userdbctl". The framework
6500 allows defining rich user and group records in a JSON format,
6501 extending on the classic "struct passwd" and "struct group"
6502 structures. Various components in systemd have been updated to
6503 process records in this format, including systemd-logind and
6504 pam-systemd. The user records are intended to be extensible, and
6505 allow setting various resource management, security and runtime
6506 parameters that shall be applied to processes and sessions of the
6507 user as they log in. This facility is intended to allow associating
6508 such metadata directly with user/group records so that they can be
6509 produced, extended and consumed in unified form. We hope that
6510 eventually frameworks such as sssd will generate records this way, so
6511 that for the first time resource management and various other
6512 per-user settings can be configured in LDAP directories and then
6513 provided to systemd (specifically to systemd-logind and pam-system)
6514 to apply on login. For further details see:
6516 https://systemd.io/USER_RECORD
6517 https://systemd.io/GROUP_RECORD
6518 https://systemd.io/USER_GROUP_API
6520 * A small new service systemd-homed.service has been added, that may be
6521 used to securely manage home directories with built-in encryption.
6522 The complete user record data is unified with the home directory,
6523 thus making home directories naturally migratable. Its primary
6524 back-end is based on LUKS volumes, but fscrypt, plain directories,
6525 and other storage schemes are also supported. This solves a couple of
6526 problems we saw with traditional ways to manage home directories, in
6527 particular when it comes to encryption. For further discussion of
6528 this, see the video of Lennart's talk at AllSystemsGo! 2019:
6530 https://media.ccc.de/v/ASG2019-164-reinventing-home-directories
6532 For further details about the format and expectations on home
6533 directories this new daemon makes, see:
6535 https://systemd.io/HOME_DIRECTORY
6537 * systemd-journald is now multi-instantiable. In addition to the main
6538 instance systemd-journald.service there's now a template unit
6539 systemd-journald@.service, with each instance defining a new named
6540 log 'namespace' (whose name is specified via the instance part of the
6541 unit name). A new unit file setting LogNamespace= has been added,
6542 taking such a namespace name, that assigns services to the specified
6543 log namespaces. As each log namespace is serviced by its own
6544 independent journal daemon, this functionality may be used to improve
6545 performance and increase isolation of applications, at the price of
6546 losing global message ordering. Each instance of journald has a
6547 separate set of configuration files, with possibly different disk
6548 usage limitations and other settings.
6550 journalctl now takes a new option --namespace= to show logs from a
6551 specific log namespace. The sd-journal.h API gained
6552 sd_journal_open_namespace() for opening the log stream of a specific
6553 log namespace. systemd-journald also gained the ability to exit on
6554 idle, which is useful in the context of log namespaces, as this means
6555 log daemons for log namespaces can be activated automatically on
6556 demand and will stop automatically when no longer used, minimizing
6559 * When systemd-tmpfiles copies a file tree using the 'C' line type it
6560 will now label every copied file according to the SELinux database.
6562 * When systemd/PID 1 detects it is used in the initrd it will now boot
6563 into initrd.target rather than default.target by default. This should
6564 make it simpler to build initrds with systemd as for many cases the
6565 only difference between a host OS image and an initrd image now is
6566 the presence of the /etc/initrd-release file.
6568 * A new kernel command line option systemd.cpu_affinity= is now
6569 understood. It's equivalent to the CPUAffinity= option in
6570 /etc/systemd/system.conf and allows setting the CPU mask for PID 1
6571 itself and the default for all other processes.
6573 * When systemd/PID 1 is reloaded (with systemctl daemon-reload or
6574 equivalent), the SELinux database is now reloaded, ensuring that
6575 sockets and other file system objects are generated taking the new
6576 database into account.
6578 * systemd/PID 1 accepts a new "systemd.show-status=error" setting, and
6579 "quiet" has been changed to imply that instead of
6580 "systemd.show-status=auto". In this mode, only messages about errors
6581 and significant delays in boot are shown on the console.
6583 * The sd-event.h API gained native support for the new Linux "pidfd"
6584 concept. This permits watching processes using file descriptors
6585 instead of PID numbers, which fixes a number of races and makes
6586 process supervision more robust and efficient. All of systemd's
6587 components will now use pidfds if the kernel supports it for process
6588 watching, with the exception of PID 1 itself, unfortunately. We hope
6589 to move PID 1 to exclusively using pidfds too eventually, but this
6590 requires some more kernel work first. (Background: PID 1 watches
6591 processes using waitid() with the P_ALL flag, and that does not play
6592 together nicely with pidfds yet.)
6594 * Closely related to this, the sd-event.h API gained two new calls
6595 sd_event_source_send_child_signal() (for sending a signal to a
6596 watched process) and sd_event_source_get_child_process_own() (for
6597 marking a process so that it is killed automatically whenever the
6598 event source watching it is freed).
6600 * systemd-networkd gained support for configuring Token Bucket Filter
6601 (TBF) parameters in its qdisc configuration support. Similarly,
6602 support for Stochastic Fairness Queuing (SFQ), Controlled-Delay
6603 Active Queue Management (CoDel), and Fair Queue (FQ) has been added.
6605 * systemd-networkd gained support for Intermediate Functional Block
6606 (IFB) network devices.
6608 * systemd-networkd gained support for configuring multi-path IP routes,
6609 using the new MultiPathRoute= setting in the [Route] section.
6611 * systemd-networkd's DHCPv4 client has been updated to support a new
6612 SendDecline= option. If enabled, duplicate address detection is done
6613 after a DHCP offer is received from the server. If a conflict is
6614 detected, the address is declined. The DHCPv4 client also gained
6615 support for a new RouteMTUBytes= setting that allows to configure the
6616 MTU size to be used for routes generated from DHCPv4 leases.
6618 * The PrefixRoute= setting in systemd-networkd's [Address] section of
6619 .network files has been deprecated, and replaced by AddPrefixRoute=,
6620 with its sense inverted.
6622 * The Gateway= setting of [Route] sections of .network files gained
6623 support for a special new value "_dhcp". If set, the configured
6624 static route uses the gateway host configured via DHCP.
6626 * New User= and SuppressPrefixLength= settings have been implemented
6627 for the [RoutingPolicyRule] section of .network files to configure
6628 source routing based on UID ranges and prefix length, respectively.
6630 * The Type= match property of .link files has been generalized to
6631 always match the device type shown by 'networkctl status', even for
6632 devices where udev does not set DEVTYPE=. This allows e.g. Type=ether
6635 * sd-bus gained a new API call sd_bus_message_sensitive() that marks a
6636 D-Bus message object as "sensitive". Those objects are erased from
6637 memory when they are freed. This concept is intended to be used for
6638 messages that contain security sensitive data. A new flag
6639 SD_BUS_VTABLE_SENSITIVE has been introduced as well to mark methods
6640 in sd-bus vtables, causing any incoming and outgoing messages of
6641 those methods to be implicitly marked as "sensitive".
6643 * sd-bus gained a new API call sd_bus_message_dump() for dumping the
6644 contents of a message (or parts thereof) to standard output for
6647 * systemd-sysusers gained support for creating users with the primary
6648 group named differently than the user.
6650 * systemd-growfs (i.e. the x-systemd.growfs mount option in /etc/fstab)
6651 gained support for growing XFS partitions. Previously it supported
6652 only ext4 and btrfs partitions.
6654 * The support for /etc/crypttab gained a new x-initrd.attach option. If
6655 set, the specified encrypted volume is unlocked already in the
6656 initrd. This concept corresponds to the x-initrd.mount option in
6659 * systemd-cryptsetup gained native support for unlocking encrypted
6660 volumes utilizing PKCS#11 smartcards, i.e. for example to bind
6661 encryption of volumes to YubiKeys. This is exposed in the new
6662 pkcs11-uri= option in /etc/crypttab.
6664 * The /etc/fstab support in systemd now supports two new mount options
6665 x-systemd.{required,wanted}-by=, for explicitly configuring the units
6666 that the specified mount shall be pulled in by, in place of
6667 the usual local-fs.target/remote-fs.target.
6669 * The https://systemd.io/ web site has been relaunched, directly
6670 populated with most of the documentation included in the systemd
6671 repository. systemd also acquired a new logo, thanks to Tobias
6674 * systemd-udevd gained support for managing "alternative" network
6675 interface names, as supported by new Linux kernels. For the first
6676 time this permits assigning multiple (and longer!) names to a network
6677 interface. systemd-udevd will now by default assign the names
6678 generated via all supported naming schemes to each interface. This
6679 may be further tweaked with .link files and the AlternativeName= and
6680 AlternativeNamesPolicy= settings. Other components of systemd have
6681 been updated to support the new alternative names wherever
6682 appropriate. For example, systemd-nspawn will now generate
6683 alternative interface names for the host-facing side of container
6684 veth links based on the full container name without truncation.
6686 * systemd-nspawn interface naming logic has been updated in another way
6687 too: if the main interface name (i.e. as opposed to new-style
6688 "alternative" names) based on the container name is truncated, a
6689 simple hashing scheme is used to give different interface names to
6690 multiple containers whose names all begin with the same prefix. Since
6691 this changes the primary interface names pointing to containers if
6692 truncation happens, the old scheme may still be requested by
6693 selecting an older naming scheme, via the net.naming-scheme= kernel
6694 command line option.
6696 * PrivateUsers= in service files now works in services run by the
6697 systemd --user per-user instance of the service manager.
6699 * A new per-service sandboxing option ProtectClock= has been added that
6700 locks down write access to the system clock. It takes away device
6701 node access to /dev/rtc as well as the system calls that set the
6702 system clock and the CAP_SYS_TIME and CAP_WAKE_ALARM capabilities.
6703 Note that this option does not affect access to auxiliary services
6704 that allow changing the clock, for example access to
6707 * The systemd-id128 tool gained a new "show" verb for listing or
6708 resolving a number of well-known UUIDs/128-bit IDs, currently mostly
6709 GPT partition table types.
6711 * The Discoverable Partitions Specification has been updated to support
6712 /var and /var/tmp partition discovery. Support for this has been
6713 added to systemd-gpt-auto-generator. For details see:
6715 https://systemd.io/DISCOVERABLE_PARTITIONS
6717 * "systemctl list-unit-files" has been updated to show a new column
6718 with the suggested enablement state based on the vendor preset files
6719 for the respective units.
6721 * "systemctl" gained a new option "--with-dependencies". If specified
6722 commands such as "systemctl status" or "systemctl cat" will now show
6723 all specified units along with all units they depend on.
6725 * networkctl gained support for showing per-interface logs in its
6728 * systemd-networkd-wait-online gained support for specifying the maximum
6729 operational state to wait for, and to wait for interfaces to
6732 * The [Match] section of .link and .network files now supports a new
6733 option PermanentMACAddress= which may be used to check against the
6734 permanent MAC address of a network device even if a randomized MAC
6737 * The [TrafficControlQueueingDiscipline] section in .network files has
6738 been renamed to [NetworkEmulator] with the "NetworkEmulator" prefix
6739 dropped from the individual setting names.
6741 * Any .link and .network files that have an empty [Match] section (this
6742 also includes empty and commented-out files) will now be
6743 rejected. systemd-udev and systemd-networkd started warning about
6744 such files in version 243.
6746 * systemd-logind will now validate access to the operation of changing
6747 the virtual terminal via a polkit action. By default, only users
6748 with at least one session on a local VT are granted permission.
6750 * When systemd sets up PAM sessions that invoked service processes
6751 shall run in, the pam_setcred() API is now invoked, thus permitting
6752 PAM modules to set additional credentials for the processes.
6754 * portablectl attach/detach verbs now accept --now and --enable options
6755 to combine attachment with enablement and invocation, or detachment
6756 with stopping and disablement.
6758 * UPGRADE ISSUE: a bug where some jobs were trimmed as redundant was
6759 fixed, which in turn exposed bugs in unit configuration of services
6760 which have Type=oneshot and should only run once, but do not have
6761 RemainAfterExit=yes set. Without RemainAfterExit=yes, a one-shot
6762 service may be started again after exiting successfully, for example
6763 as a dependency in another transaction. Affected services included
6764 some internal systemd services (most notably
6765 systemd-vconsole-setup.service, which was updated to have
6766 RemainAfterExit=yes), and plymouth-start.service. Please ensure that
6767 plymouth has been suitably updated or patched before upgrading to
6768 this systemd release. See
6769 https://bugzilla.redhat.com/show_bug.cgi?id=1807771 for some
6770 additional discussion.
6772 Contributions from: AJ Bagwell, Alin Popa, Andreas Rammhold, Anita
6773 Zhang, Ansgar Burchardt, Antonio Russo, Arian van Putten, Ashley Davis,
6774 Balint Reczey, Bart Willems, Bastien Nocera, Benjamin Dahlhoff, Charles
6775 (Chas) Williams, cheese1, Chris Down, Chris Murphy, Christian Ehrhardt,
6776 Christian Göttsche, cvoinf, Daan De Meyer, Daniele Medri, Daniel Rusek,
6777 Daniel Shahaf, Dann Frazier, Dan Streetman, Dariusz Gadomski, David
6778 Michael, Dimitri John Ledkov, Emmanuel Bourg, Evgeny Vereshchagin,
6779 ezst036, Felipe Sateler, Filipe Brandenburger, Florian Klink, Franck
6780 Bui, Fran Dieguez, Frantisek Sumsal, Greg "GothAck" Miell, Guilhem
6781 Lettron, Guillaume Douézan-Grard, Hans de Goede, HATAYAMA Daisuke, Iain
6782 Lane, James Buren, Jan Alexander Steffens (heftig), Jérémy Rosen, Jin
6783 Park, Jun'ichi Nomura, Kai Krakow, Kevin Kuehler, Kevin P. Fleming,
6784 Lennart Poettering, Leonid Bloch, Leonid Evdokimov, lothrond, Luca
6785 Boccassi, Lukas K, Lynn Kirby, Mario Limonciello, Mark Deneen, Matthew
6786 Leeds, Michael Biebl, Michal Koutný, Michal Sekletár, Mike Auty, Mike
6787 Gilbert, mtron, nabijaczleweli, Naïm Favier, Nate Jones, Norbert Lange,
6788 Oliver Giles, Paul Davey, Paul Menzel, Peter Hutterer, Piotr Drąg, Rafa
6789 Couto, Raphael, rhn, Robert Scheck, Rocka, Romain Naour, Ryan Attard,
6790 Sascha Dewald, Shengjing Zhu, Slava Kardakov, Spencer Michaels, Sylvain
6791 Plantefeve, Stanislav Angelovič, Susant Sahani, Thomas Haller, Thomas
6792 Schmitt, Timo Schlüßler, Timo Wilken, Tobias Bernard, Tobias Klauser,
6793 Tobias Stoeckmann, Topi Miettinen, tsia, WataruMatsuoka, Wieland
6794 Hoffmann, Wilhelm Schuster, Will Fleming, xduugu, Yong Cong Sin, Yuri
6795 Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zeyu
6798 – Warsaw, 2020-03-06
6802 * Support for the cpuset cgroups v2 controller has been added.
6803 Processes may be restricted to specific CPUs using the new
6804 AllowedCPUs= setting, and to specific memory NUMA nodes using the new
6805 AllowedMemoryNodes= setting.
6807 * The signal used in restart jobs (as opposed to e.g. stop jobs) may
6808 now be configured using a new RestartKillSignal= setting. This
6809 allows units which signals to request termination to implement
6810 different behaviour when stopping in preparation for a restart.
6812 * "systemctl clean" may now be used also for socket, mount, and swap
6815 * systemd will also read configuration options from the EFI variable
6816 SystemdOptions. This may be used to configure systemd behaviour when
6817 modifying the kernel command line is inconvenient, but configuration
6818 on disk is read too late, for example for the options related to
6819 cgroup hierarchy setup. 'bootctl systemd-efi-options' may be used to
6820 set the EFI variable.
6822 * systemd will now disable printk ratelimits in early boot. This should
6823 allow us to capture more logs from the early boot phase where normal
6824 storage is not available and the kernel ring buffer is used for
6825 logging. Configuration on the kernel command line has higher priority
6826 and overrides the systemd setting.
6828 systemd programs which log to /dev/kmsg directly use internal
6829 ratelimits to prevent runaway logging. (Normally this is only used
6830 during early boot, so in practice this change has very little
6833 * Unit files now support top level dropin directories of the form
6834 <unit_type>.d/ (e.g. service.d/) that may be used to add configuration
6835 that affects all corresponding unit files.
6837 * systemctl gained support for 'stop --job-mode=triggering' which will
6838 stop the specified unit and any units which could trigger it.
6840 * Unit status display now includes units triggering and triggered by
6841 the unit being shown.
6843 * The RuntimeMaxSec= setting is now supported by scopes, not just
6844 .service units. This is particularly useful for PAM sessions which
6845 create a scope unit for the user login. systemd.runtime_max_sec=
6846 setting may used with the pam_systemd module to limit the duration
6847 of the PAM session, for example for time-limited logins.
6849 * A new @pkey system call group is now defined to make it easier to
6850 allow-list memory protection syscalls for containers and services
6851 which need to use them.
6853 * systemd-udevd: removed the 30s timeout for killing stale workers on
6854 exit. systemd-udevd now waits for workers to finish. The hard-coded
6855 exit timeout of 30s was too short for some large installations, where
6856 driver initialization could be prematurely interrupted during initrd
6857 processing if the root file system had been mounted and init was
6858 preparing to switch root. If udevd is run without systemd and workers
6859 are hanging while udevd receives an exit signal, udevd will now exit
6860 when udev.event_timeout is reached for the last hanging worker. With
6861 systemd, the exit timeout can additionally be configured using
6862 TimeoutStopSec= in systemd-udevd.service.
6864 * udev now provides a program (fido_id) that identifies FIDO CTAP1
6865 ("U2F")/CTAP2 security tokens based on the usage declared in their
6866 report and descriptor and outputs suitable environment variables.
6867 This replaces the externally maintained allow lists of all known
6868 security tokens that were used previously.
6870 * Automatically generated autosuspend udev rules for allow-listed
6871 devices have been imported from the Chromium OS project. This should
6872 improve power saving with many more devices.
6874 * udev gained a new "CONST{key}=value" setting that allows matching
6875 against system-wide constants without forking a helper binary.
6876 Currently "arch" and "virt" keys are supported.
6878 * udev now opens CDROMs in non-exclusive mode when querying their
6879 capabilities. This should fix issues where other programs trying to
6880 use the CDROM cannot gain access to it, but carries a risk of
6881 interfering with programs writing to the disk, if they did not open
6882 the device in exclusive mode as they should.
6884 * systemd-networkd does not create a default route for IPv4 link local
6885 addressing anymore. The creation of the route was unexpected and was
6886 breaking routing in various cases, but people who rely on it being
6887 created implicitly will need to adjust. Such a route may be requested
6888 with DefaultRouteOnDevice=yes.
6890 Similarly, systemd-networkd will not assign a link-local IPv6 address
6891 when IPv6 link-local routing is not enabled.
6893 * Receive and transmit buffers may now be configured on links with
6894 the new RxBufferSize= and TxBufferSize= settings.
6896 * systemd-networkd may now advertise additional IPv6 routes. A new
6897 [IPv6RoutePrefix] section with Route= and LifetimeSec= options is
6900 * systemd-networkd may now configure "next hop" routes using the
6901 [NextHop] section and Gateway= and Id= settings.
6903 * systemd-networkd will now retain DHCP config on restarts by default
6904 (but this may be overridden using the KeepConfiguration= setting).
6905 The default for SendRelease= has been changed to true.
6907 * The DHCPv4 client now uses the OPTION_INFORMATION_REFRESH_TIME option
6908 received from the server.
6910 The client will use the received SIP server list if UseSIP=yes is
6913 The client may be configured to request specific options from the
6914 server using a new RequestOptions= setting.
6916 The client may be configured to send arbitrary options to the server
6917 using a new SendOption= setting.
6919 A new IPServiceType= setting has been added to configure the "IP
6920 service type" value used by the client.
6922 * The DHCPv6 client learnt a new PrefixDelegationHint= option to
6923 request prefix hints in the DHCPv6 solicitation.
6925 * The DHCPv4 server may be configured to send arbitrary options using
6926 a new SendOption= setting.
6928 * The DHCPv4 server may now be configured to emit SIP server list using
6929 the new EmitSIP= and SIP= settings.
6931 * systemd-networkd and networkctl may now renew DHCP leases on demand.
6932 networkctl has a new 'networkctl renew' verb.
6934 * systemd-networkd may now reconfigure links on demand. networkctl
6935 gained two new verbs: "reload" will reload the configuration, and
6936 "reconfigure DEVICE…" will reconfigure one or more devices.
6938 * .network files may now match on SSID and BSSID of a wireless network,
6939 i.e. the access point name and hardware address using the new SSID=
6940 and BSSID= options. networkctl will display the current SSID and
6941 BSSID for wireless links.
6943 .network files may also match on the wireless network type using the
6944 new WLANInterfaceType= option.
6946 * systemd-networkd now includes default configuration that enables
6947 link-local addressing when connected to an ad-hoc wireless network.
6949 * systemd-networkd may configure the Traffic Control queueing
6950 disciplines in the kernel using the new
6951 [TrafficControlQueueingDiscipline] section and Parent=,
6952 NetworkEmulatorDelaySec=, NetworkEmulatorDelayJitterSec=,
6953 NetworkEmulatorPacketLimit=, NetworkEmulatorLossRate=,
6954 NetworkEmulatorDuplicateRate= settings.
6956 * systemd-tmpfiles gained a new w+ setting to append to files.
6958 * systemd-analyze dump will now report when the memory configuration in
6959 the kernel does not match what systemd has configured (usually,
6960 because some external program has modified the kernel configuration
6963 * systemd-analyze gained a new --base-time= switch instructs the
6964 'calendar' verb to resolve times relative to that timestamp instead
6965 of the present time.
6967 * journalctl --update-catalog now produces deterministic output (making
6968 reproducible image builds easier).
6970 * A new devicetree-overlay setting is now documented in the Boot Loader
6973 * The default value of the WatchdogSec= setting used in systemd
6974 services (the ones bundled with the project itself) may be set at
6975 configuration time using the -Dservice-watchdog= setting. If set to
6976 empty, the watchdogs will be disabled.
6978 * systemd-resolved validates IP addresses in certificates now when GnuTLS
6981 * libcryptsetup >= 2.0.1 is now required.
6983 * A configuration option -Duser-path= may be used to override the $PATH
6984 used by the user service manager. The default is again to use the same
6985 path as the system manager.
6987 * The systemd-id128 tool gained a new switch "-u" (or "--uuid") for
6988 outputting the 128-bit IDs in UUID format (i.e. in the "canonical
6991 * Service units gained a new sandboxing option ProtectKernelLogs= which
6992 makes sure the program cannot get direct access to the kernel log
6993 buffer anymore, i.e. the syslog() system call (not to be confused
6994 with the API of the same name in libc, which is not affected), the
6995 /proc/kmsg and /dev/kmsg nodes and the CAP_SYSLOG capability are made
6996 inaccessible to the service. It's recommended to enable this setting
6997 for all services that should not be able to read from or write to the
6998 kernel log buffer, which are probably almost all.
7000 Contributions from: Aaron Plattner, Alcaro, Anita Zhang, Balint Reczey,
7001 Bastien Nocera, Baybal Ni, Benjamin Bouvier, Benjamin Gilbert, Carlo
7002 Teubner, cbzxt, Chen Qi, Chris Down, Christian Rebischke, Claudio
7003 Zumbo, ClydeByrdIII, crashfistfight, Cyprien Laplace, Daniel Edgecumbe,
7004 Daniel Gorbea, Daniel Rusek, Daniel Stuart, Dan Streetman, David
7005 Pedersen, David Tardon, Dimitri John Ledkov, Dominique Martinet, Donald
7006 A. Cupp Jr, Evgeny Vereshchagin, Fabian Henneke, Filipe Brandenburger,
7007 Franck Bui, Frantisek Sumsal, Georg Müller, Hans de Goede, Haochen
7008 Tong, HATAYAMA Daisuke, Iwan Timmer, Jan Janssen, Jan Kundrát, Jan
7009 Synacek, Jan Tojnar, Jay Strict, Jérémy Rosen, Jóhann B. Guðmundsson,
7010 Jonas Jelten, Jonas Thelemann, Justin Trudell, J. Xing, Kai-Heng Feng,
7011 Kenneth D'souza, Kevin Becker, Kevin Kuehler, Lennart Poettering,
7012 Léonard Gérard, Lorenz Bauer, Luca Boccassi, Maciej Stanczew, Mario
7013 Limonciello, Marko Myllynen, Mark Stosberg, Martin Wilck, matthiasroos,
7014 Michael Biebl, Michael Olbrich, Michael Tretter, Michal Sekletar,
7015 Michal Sekletár, Michal Suchanek, Mike Gilbert, Mike Kazantsev, Nicolas
7016 Douma, nikolas, Norbert Lange, pan93412, Pascal de Bruijn, Paul Menzel,
7017 Pavel Hrdina, Peter Wu, Philip Withnall, Piotr Drąg, Rafael Fontenelle,
7018 Renaud Métrich, Riccardo Schirone, RoadrunnerWMC, Ronan Pigott, Ryan
7019 Attard, Sebastian Wick, Serge, Siddharth Chandrasekara, Steve Ramage,
7020 Steve Traylen, Susant Sahani, Thibault Nélis, Tim Teichmann, Tom
7021 Fitzhenry, Tommy J, Torsten Hilbrich, Vito Caputo, ypf791, Yu Watanabe,
7022 Zach Smith, Zbigniew Jędrzejewski-Szmek
7024 – Warsaw, 2019-11-29
7028 * This release enables unprivileged programs (i.e. requiring neither
7029 setuid nor file capabilities) to send ICMP Echo (i.e. ping) requests
7030 by turning on the "net.ipv4.ping_group_range" sysctl of the Linux
7031 kernel for the whole UNIX group range, i.e. all processes. This
7032 change should be reasonably safe, as the kernel support for it was
7033 specifically implemented to allow safe access to ICMP Echo for
7034 processes lacking any privileges. If this is not desirable, it can be
7035 disabled again by setting the parameter to "1 0".
7037 * Previously, filters defined with SystemCallFilter= would have the
7038 effect that any calling of an offending system call would terminate
7039 the calling thread. This behaviour never made much sense, since
7040 killing individual threads of unsuspecting processes is likely to
7041 create more problems than it solves. With this release the default
7042 action changed from killing the thread to killing the whole
7043 process. For this to work correctly both a kernel version (>= 4.14)
7044 and a libseccomp version (>= 2.4.0) supporting this new seccomp
7045 action is required. If an older kernel or libseccomp is used the old
7046 behaviour continues to be used. This change does not affect any
7047 services that have no system call filters defined, or that use
7048 SystemCallErrorNumber= (and thus see EPERM or another error instead
7049 of being killed when calling an offending system call). Note that
7050 systemd documentation always claimed that the whole process is
7051 killed. With this change behaviour is thus adjusted to match the
7054 * On 64 bit systems, the "kernel.pid_max" sysctl is now bumped to
7055 4194304 by default, i.e. the full 22bit range the kernel allows, up
7056 from the old 16-bit range. This should improve security and
7057 robustness, as PID collisions are made less likely (though certainly
7058 still possible). There are rumours this might create compatibility
7059 problems, though at this moment no practical ones are known to
7060 us. Downstream distributions are hence advised to undo this change in
7061 their builds if they are concerned about maximum compatibility, but
7062 for everybody else we recommend leaving the value bumped. Besides
7063 improving security and robustness this should also simplify things as
7064 the maximum number of allowed concurrent tasks was previously bounded
7065 by both "kernel.pid_max" and "kernel.threads-max" and now effectively
7066 only a single knob is left ("kernel.threads-max"). There have been
7067 concerns that usability is affected by this change because larger PID
7068 numbers are harder to type, but we believe the change from 5 digits
7069 to 7 digits doesn't hamper usability.
7071 * MemoryLow= and MemoryMin= gained hierarchy-aware counterparts,
7072 DefaultMemoryLow= and DefaultMemoryMin=, which can be used to
7073 hierarchically set default memory protection values for a particular
7074 subtree of the unit hierarchy.
7076 * Memory protection directives can now take a value of zero, allowing
7077 explicit opting out of a default value propagated by an ancestor.
7079 * systemd now defaults to the "unified" cgroup hierarchy setup during
7080 build-time, i.e. -Ddefault-hierarchy=unified is now the build-time
7081 default. Previously, -Ddefault-hierarchy=hybrid was the default. This
7082 change reflects the fact that cgroupsv2 support has matured
7083 substantially in both systemd and in the kernel, and is clearly the
7084 way forward. Downstream production distributions might want to
7085 continue to use -Ddefault-hierarchy=hybrid (or even =legacy) for
7086 their builds as unfortunately the popular container managers have not
7087 caught up with the kernel API changes.
7089 * Man pages are not built by default anymore (html pages were already
7090 disabled by default), to make development builds quicker. When
7091 building systemd for a full installation with documentation, meson
7092 should be called with -Dman=true and/or -Dhtml=true as appropriate.
7093 The default was changed based on the assumption that quick one-off or
7094 repeated development builds are much more common than full optimized
7095 builds for installation, and people need to pass various other
7096 options to when doing "proper" builds anyway, so the gain from making
7097 development builds quicker is bigger than the one time disruption for
7100 Two scripts are created in the *build* directory to generate and
7101 preview man and html pages on demand, e.g.:
7103 build/man/man systemctl
7104 build/man/html systemd.index
7106 * libidn2 is used by default if both libidn2 and libidn are installed.
7107 Please use -Dlibidn=true if libidn is preferred.
7109 * The D-Bus "wire format" of the CPUAffinity= attribute is changed on
7110 big-endian machines. Before, bytes were written and read in native
7111 machine order as exposed by the native libc __cpu_mask interface.
7112 Now, little-endian order is always used (CPUs 0–7 are described by
7113 bits 0–7 in byte 0, CPUs 8–15 are described by byte 1, and so on).
7114 This change fixes D-Bus calls that cross endianness boundary.
7116 The presentation format used for CPUAffinity= by "systemctl show" and
7117 "systemd-analyze dump" is changed to present CPU indices instead of
7118 the raw __cpu_mask bitmask. For example, CPUAffinity=0-1 would be
7119 shown as CPUAffinity=03000000000000000000000000000… (on
7120 little-endian) or CPUAffinity=00000000000000300000000000000… (on
7121 64-bit big-endian), and is now shown as CPUAffinity=0-1, matching the
7122 input format. The maximum integer that will be printed in the new
7123 format is 8191 (four digits), while the old format always used a very
7124 long number (with the length varying by architecture), so they can be
7125 unambiguously distinguished.
7127 * /usr/sbin/halt.local is no longer supported. Implementation in
7128 distributions was inconsistent and it seems this functionality was
7131 To replace this functionality, users should:
7132 - either define a new unit and make it a dependency of final.target
7133 (systemctl add-wants final.target my-halt-local.service)
7134 - or move the shutdown script to /usr/lib/systemd/system-shutdown/
7135 and ensure that it accepts "halt", "poweroff", "reboot", and
7136 "kexec" as an argument, see the description in systemd-shutdown(8).
7138 * When a [Match] section in .link or .network file is empty (contains
7139 no match patterns), a warning will be emitted. Please add any "match
7140 all" pattern instead, e.g. OriginalName=* or Name=* in case all
7141 interfaces should really be matched.
7143 * A new setting NUMAPolicy= may be used to set process memory
7144 allocation policy. This setting can be specified in
7145 /etc/systemd/system.conf and hence will set the default policy for
7146 PID1. The default policy can be overridden on a per-service
7147 basis. The related setting NUMAMask= is used to specify NUMA node
7148 mask that should be associated with the selected policy.
7150 * PID 1 will now listen to Out-Of-Memory (OOM) events the kernel
7151 generates when processes it manages are reaching their memory limits,
7152 and will place their units in a special state, and optionally kill or
7153 stop the whole unit.
7155 * The service manager will now expose bus properties for the IO
7156 resources used by units. This information is also shown in "systemctl
7157 status" now (for services that have IOAccounting=yes set). Moreover,
7158 the IO accounting data is included in the resource log message
7159 generated whenever a unit stops.
7161 * Units may now configure an explicit timeout to wait for when killed
7162 with SIGABRT, for example when a service watchdog is hit. Previously,
7163 the regular TimeoutStopSec= timeout was applied in this case too —
7164 now a separate timeout may be set using TimeoutAbortSec=.
7166 * Services may now send a special WATCHDOG=trigger message with
7167 sd_notify() to trigger an immediate "watchdog missed" event, and thus
7168 trigger service termination. This is useful both for testing watchdog
7169 handling, but also for defining error paths in services, that shall
7170 be handled the same way as watchdog events.
7172 * There are two new per-unit settings IPIngressFilterPath= and
7173 IPEgressFilterPath= which allow configuration of a BPF program
7174 (usually by specifying a path to a program uploaded to /sys/fs/bpf/)
7175 to apply to the IP packet ingress/egress path of all processes of a
7176 unit. This is useful to allow running systemd services with BPF
7177 programs set up externally.
7179 * systemctl gained a new "clean" verb for removing the state, cache,
7180 runtime or logs directories of a service while it is terminated. The
7181 new verb may also be used to remove the state maintained on disk for
7182 timer units that have Persistent= configured.
7184 * During the last phase of shutdown systemd will now automatically
7185 increase the log level configured in the "kernel.printk" sysctl so
7186 that any relevant loggable events happening during late shutdown are
7187 made visible. Previously, loggable events happening so late during
7188 shutdown were generally lost if the "kernel.printk" sysctl was set to
7189 high thresholds, as regular logging daemons are terminated at that
7190 time and thus nothing is written to disk.
7192 * If processes terminated during the last phase of shutdown do not exit
7193 quickly systemd will now show their names after a short time, to make
7194 debugging easier. After a longer timeout they are forcibly killed,
7197 * journalctl (and the other tools that display logs) will now highlight
7198 warnings in yellow (previously, both LOG_NOTICE and LOG_WARNING where
7199 shown in bright bold, now only LOG_NOTICE is). Moreover, audit logs
7200 are now shown in blue color, to separate them visually from regular
7201 logs. References to configuration files are now turned into clickable
7202 links on terminals that support that.
7204 * systemd-journald will now stop logging to /var/log/journal during
7205 shutdown when /var/ is on a separate mount, so that it can be
7206 unmounted safely during shutdown.
7208 * systemd-resolved gained support for a new 'strict' DNS-over-TLS mode.
7210 * systemd-resolved "Cache=" configuration option in resolved.conf has
7211 been extended to also accept the 'no-negative' value. Previously,
7212 only a boolean option was allowed (yes/no), having yes as the
7213 default. If this option is set to 'no-negative', negative answers are
7214 not cached while the old cache heuristics are used positive answers.
7215 The default remains unchanged.
7217 * The predictable naming scheme for network devices now supports
7218 generating predictable names for "netdevsim" devices.
7220 Moreover, the "en" prefix was dropped from the ID_NET_NAME_ONBOARD
7223 Those two changes form a new net.naming-policy-scheme= entry.
7224 Distributions which want to preserve naming stability may want to set
7225 the -Ddefault-net-naming-scheme= configuration option.
7227 * systemd-networkd now supports MACsec, nlmon, IPVTAP and Xfrm
7228 interfaces natively.
7230 * systemd-networkd's bridge FDB support now allows configuration of a
7231 destination address for each entry (Destination=), as well as the
7232 VXLAN VNI (VNI=), as well as an option to declare what an entry is
7233 associated with (AssociatedWith=).
7235 * systemd-networkd's DHCPv4 support now understands a new MaxAttempts=
7236 option for configuring the maximum number of DHCP lease requests. It
7237 also learnt a new BlackList= option for deny-listing DHCP servers (a
7238 similar setting has also been added to the IPv6 RA client), as well
7239 as a SendRelease= option for configuring whether to send a DHCP
7240 RELEASE message when terminating.
7242 * systemd-networkd's DHCPv4 and DHCPv6 stacks can now be configured
7243 separately in the [DHCPv4] and [DHCPv6] sections.
7245 * systemd-networkd's DHCP support will now optionally create an
7246 implicit host route to the DNS server specified in the DHCP lease, in
7247 addition to the routes listed explicitly in the lease. This should
7248 ensure that in multi-homed systems DNS traffic leaves the systems on
7249 the interface that acquired the DNS server information even if other
7250 routes such as default routes exist. This behaviour may be turned on
7251 with the new RoutesToDNS= option.
7253 * systemd-networkd's VXLAN support gained a new option
7254 GenericProtocolExtension= for enabling VXLAN Generic Protocol
7255 Extension support, as well as IPDoNotFragment= for setting the IP
7256 "Don't fragment" bit on outgoing packets. A similar option has been
7257 added to the GENEVE support.
7259 * In systemd-networkd's [Route] section you may now configure
7260 FastOpenNoCookie= for configuring per-route TCP fast-open support, as
7261 well as TTLPropagate= for configuring Label Switched Path (LSP) TTL
7262 propagation. The Type= setting now supports local, broadcast,
7263 anycast, multicast, any, xresolve routes, too.
7265 * systemd-networkd's [Network] section learnt a new option
7266 DefaultRouteOnDevice= for automatically configuring a default route
7267 onto the network device.
7269 * systemd-networkd's bridging support gained two new options ProxyARP=
7270 and ProxyARPWifi= for configuring proxy ARP behaviour as well as
7271 MulticastRouter= for configuring multicast routing behaviour. A new
7272 option MulticastIGMPVersion= may be used to change bridge's multicast
7273 Internet Group Management Protocol (IGMP) version.
7275 * systemd-networkd's FooOverUDP support gained the ability to configure
7276 local and peer IP addresses via Local= and Peer=. A new option
7277 PeerPort= may be used to configure the peer's IP port.
7279 * systemd-networkd's TUN support gained a new setting VnetHeader= for
7280 tweaking Generic Segment Offload support.
7282 * The address family for policy rules may be specified using the new
7283 Family= option in the [RoutingPolicyRule] section.
7285 * networkctl gained a new "delete" command for removing virtual network
7286 devices, as well as a new "--stats" switch for showing device
7289 * networkd.conf gained a new setting SpeedMeter= and
7290 SpeedMeterIntervalSec=, to measure bitrate of network interfaces. The
7291 measured speed may be shown by 'networkctl status'.
7293 * "networkctl status" now displays MTU and queue lengths, and more
7294 detailed information about VXLAN and bridge devices.
7296 * systemd-networkd's .network and .link files gained a new Property=
7297 setting in the [Match] section, to match against devices with
7298 specific udev properties.
7300 * systemd-networkd's tunnel support gained a new option
7301 AssignToLoopback= for selecting whether to use the loopback device
7302 "lo" as underlying device.
7304 * systemd-networkd's MACAddress= setting in the [Neighbor] section has
7305 been renamed to LinkLayerAddress=, and it now allows configuration of
7308 * systemd-networkd's handling of the kernel's disable_ipv6 sysctl is
7309 simplified: systemd-networkd will disable the sysctl (enable IPv6) if
7310 IPv6 configuration (static or DHCPv6) was found for a given
7311 interface. It will not touch the sysctl otherwise.
7313 * The order of entries is $PATH used by the user manager instance was
7314 changed to put bin/ entries before the corresponding sbin/ entries.
7315 It is recommended to not rely on this order, and only ever have one
7316 binary with a given name in the system paths under /usr.
7318 * A new tool systemd-network-generator has been added that may generate
7319 .network, .netdev and .link files from IP configuration specified on
7320 the kernel command line in the format used by Dracut.
7322 * The CriticalConnection= setting in .network files is now deprecated,
7323 and replaced by a new KeepConfiguration= setting which allows more
7324 detailed configuration of the IP configuration to keep in place.
7326 * systemd-analyze gained a few new verbs:
7328 - "systemd-analyze timestamp" parses and converts timestamps. This is
7329 similar to the existing "systemd-analyze calendar" command which
7330 does the same for recurring calendar events.
7332 - "systemd-analyze timespan" parses and converts timespans (i.e.
7333 durations as opposed to points in time).
7335 - "systemd-analyze condition" will parse and test ConditionXYZ=
7338 - "systemd-analyze exit-status" will parse and convert exit status
7339 codes to their names and back.
7341 - "systemd-analyze unit-files" will print a list of all unit
7342 file paths and unit aliases.
7344 * SuccessExitStatus=, RestartPreventExitStatus=, and
7345 RestartForceExitStatus= now accept exit status names (e.g. "DATAERR"
7346 is equivalent to "65"). Those exit status name mappings may be
7347 displayed with the systemd-analyze exit-status verb describe above.
7349 * systemd-logind now exposes a per-session SetBrightness() bus call,
7350 which may be used to securely change the brightness of a kernel
7351 brightness device, if it belongs to the session's seat. By using this
7352 call unprivileged clients can make changes to "backlight" and "leds"
7353 devices securely with strict requirements on session membership.
7354 Desktop environments may use this to generically make brightness
7355 changes to such devices without shipping private SUID binaries or
7356 udev rules for that purpose.
7358 * "udevadm info" gained a --wait-for-initialization switch to wait for
7359 a device to be initialized.
7361 * systemd-hibernate-resume-generator will now look for resumeflags= on
7362 the kernel command line, which is similar to rootflags= and may be
7363 used to configure device timeout for the hibernation device.
7365 * sd-event learnt a new API call sd_event_source_disable_unref() for
7366 disabling and unref'ing an event source in a single function. A
7367 related call sd_event_source_disable_unrefp() has been added for use
7368 with gcc's cleanup extension.
7370 * The sd-id128.h public API gained a new definition
7371 SD_ID128_UUID_FORMAT_STR for formatting a 128-bit ID in UUID format
7374 * "busctl introspect" gained a new switch --xml-interface for dumping
7375 XML introspection data unmodified.
7377 * PID 1 may now show the unit name instead of the unit description
7378 string in its status output during boot. This may be configured in
7379 the StatusUnitFormat= setting in /etc/systemd/system.conf or the
7380 kernel command line option systemd.status_unit_format=.
7382 * PID 1 now understands a new option KExecWatchdogSec= in
7383 /etc/systemd/system.conf to set a watchdog timeout for kexec reboots.
7384 Previously watchdog functionality was only available for regular
7385 reboots. The new setting defaults to off, because we don't know in
7386 the general case if the watchdog will be reset after kexec (some
7387 drivers do reset it, but not all), and the new userspace might not be
7388 configured to handle the watchdog.
7390 Moreover, the old ShutdownWatchdogSec= setting has been renamed to
7391 RebootWatchdogSec= to more clearly communicate what it is about. The
7392 old name is still accepted for compatibility.
7394 * The systemd.debug_shell kernel command line option now optionally
7395 takes a tty name to spawn the debug shell on, which allows a
7396 different tty to be selected than the built-in default.
7398 * Service units gained a new ExecCondition= setting which will run
7399 before ExecStartPre= and either continue execution of the unit (for
7400 clean exit codes), stop execution without marking the unit failed
7401 (for exit codes 1 through 254), or stop execution and fail the unit
7402 (for exit code 255 or abnormal termination).
7404 * A new service systemd-pstore.service has been added that pulls data
7405 from /sys/fs/pstore/ and saves it to /var/lib/pstore for later
7408 * timedatectl gained new verbs for configuring per-interface NTP
7409 service configuration for systemd-timesyncd.
7411 * "localectl list-locales" won't list non-UTF-8 locales anymore. It's
7412 2019. (You can set non-UTF-8 locales though, if you know their name.)
7414 * If variable assignments in sysctl.d/ files are prefixed with "-" any
7415 failures to apply them are now ignored.
7417 * systemd-random-seed.service now optionally credits entropy when
7418 applying the seed to the system. Set $SYSTEMD_RANDOM_SEED_CREDIT to
7419 true for the service to enable this behaviour, but please consult the
7420 documentation first, since this comes with a couple of caveats.
7422 * systemd-random-seed.service is now a synchronization point for full
7423 initialization of the kernel's entropy pool. Services that require
7424 /dev/urandom to be correctly initialized should be ordered after this
7427 * The systemd-boot boot loader has been updated to optionally maintain
7428 a random seed file in the EFI System Partition (ESP). During the boot
7429 phase, this random seed is read and updated with a new seed
7430 cryptographically derived from it. Another derived seed is passed to
7431 the OS. The latter seed is then credited to the kernel's entropy pool
7432 very early during userspace initialization (from PID 1). This allows
7433 systems to boot up with a fully initialized kernel entropy pool from
7434 earliest boot on, and thus entirely removes all entropy pool
7435 initialization delays from systems using systemd-boot. Special care
7436 is taken to ensure different seeds are derived on system images
7437 replicated to multiple systems. "bootctl status" will show whether
7438 a seed was received from the boot loader.
7440 * bootctl gained two new verbs:
7442 - "bootctl random-seed" will generate the file in ESP and an EFI
7443 variable to allow a random seed to be passed to the OS as described
7446 - "bootctl is-installed" checks whether systemd-boot is currently
7449 * bootctl will warn if it detects that boot entries are misconfigured
7450 (for example if the kernel image was removed without purging the
7453 * A new document has been added describing systemd's use and support
7454 for the kernel's entropy pool subsystem:
7456 https://systemd.io/RANDOM_SEEDS
7458 * When the system is hibernated the swap device to write the
7459 hibernation image to is now automatically picked from all available
7460 swap devices, preferring the swap device with the highest configured
7461 priority over all others, and picking the device with the most free
7462 space if there are multiple devices with the highest priority.
7464 * /etc/crypttab support has learnt a new keyfile-timeout= per-device
7465 option that permits selecting the timeout how long to wait for a
7466 device with an encryption key before asking for the password.
7468 * IOWeight= has learnt to properly set the IO weight when using the
7469 BFQ scheduler officially found in kernels 5.0+.
7471 * A new mailing list has been created for reporting of security issues:
7472 systemd-security@redhat.com. For mode details, see
7473 https://systemd.io/CONTRIBUTING#security-vulnerability-reports.
7475 Contributions from: Aaron Barany, Adrian Bunk, Alan Jenkins, Albrecht
7476 Lohofener, Andrej Valek, Anita Zhang, Arian van Putten, Balint Reczey,
7477 Bastien Nocera, Ben Boeckel, Benjamin Robin, camoz, Chen Qi, Chris
7478 Chiu, Chris Down, Christian Göttsche, Christian Kellner, Clinton Roy,
7479 Connor Reeder, Daniel Black, Daniel Lublin, Daniele Medri, Dan
7480 Streetman, Dave Reisner, Dave Ross, David Art, David Tardon, Debarshi
7481 Ray, Dimitri John Ledkov, Dominick Grift, Donald Buczek, Douglas
7482 Christman, Eric DeVolder, EtherGraf, Evgeny Vereshchagin, Feldwor,
7483 Felix Riemann, Florian Dollinger, Francesco Pennica, Franck Bui,
7484 Frantisek Sumsal, Franz Pletz, frederik, Hans de Goede, Iago López
7485 Galeiras, Insun Pyo, Ivan Shapovalov, Iwan Timmer, Jack, Jakob
7486 Unterwurzacher, Jan Chren, Jan Klötzke, Jan Losinski, Jan Pokorný, Jan
7487 Synacek, Jan-Michael Brummer, Jeka Pats, Jeremy Soller, Jérémy Rosen,
7488 Jiri Pirko, Joe Lin, Joerg Behrmann, Joe Richey, Jóhann B. Guðmundsson,
7489 Johannes Christ, Johannes Schmitz, Jonathan Rouleau, Jorge Niedbalski,
7490 Jörg Thalheim, Kai Krakow, Kai Lüke, Karel Zak, Kashyap Chamarthy,
7491 Krayushkin Konstantin, Lennart Poettering, Lubomir Rintel, Luca
7492 Boccassi, Luís Ferreira, Marc-André Lureau, Markus Felten, Martin Pitt,
7493 Matthew Leeds, Mattias Jernberg, Michael Biebl, Michael Olbrich,
7494 Michael Prokop, Michael Stapelberg, Michael Zhivich, Michal Koutný,
7495 Michal Sekletar, Mike Gilbert, Milan Broz, Miroslav Lichvar, mpe85,
7496 Mr-Foo, Network Silence, Oliver Harley, pan93412, Paul Menzel, pEJipE,
7497 Peter A. Bigot, Philip Withnall, Piotr Drąg, Rafael Fontenelle, Robert
7498 Scheck, Roberto Santalla, Ronan Pigott, root, RussianNeuroMancer,
7499 Sebastian Jennen, shinygold, Shreyas Behera, Simon Schricker, Susant
7500 Sahani, Thadeu Lima de Souza Cascardo, Theo Ouzhinski, Thiebaud
7501 Weksteen, Thomas Haller, Thomas Weißschuh, Tomas Mraz, Tommi Rantala,
7502 Topi Miettinen, VD-Lycos, ven, Vladimir Yerilov, Wieland Hoffmann,
7503 William A. Kennington III, William Wold, Xi Ruoyao, Yuri Chornoivan,
7504 Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zhang Xianwei
7506 – Camerino, 2019-09-03
7510 * In .link files, MACAddressPolicy=persistent (the default) is changed
7511 to cover more devices. For devices like bridges, tun, tap, bond, and
7512 similar interfaces that do not have other identifying information,
7513 the interface name is used as the basis for persistent seed for MAC
7514 and IPv4LL addresses. The way that devices that were handled
7515 previously is not changed, and this change is about covering more
7516 devices then previously by the "persistent" policy.
7518 MACAddressPolicy=random may be used to force randomized MACs and
7519 IPv4LL addresses for a device if desired.
7521 Hint: the log output from udev (at debug level) was enhanced to
7522 clarify what policy is followed and which attributes are used.
7523 `SYSTEMD_LOG_LEVEL=debug udevadm test-builtin net_setup_link /sys/class/net/<name>`
7524 may be used to view this.
7526 Hint: if a bridge interface is created without any slaves, and gains
7527 a slave later, then now the bridge does not inherit slave's MAC.
7528 To inherit slave's MAC, for example, create the following file:
7530 # /etc/systemd/network/98-bridge-inherit-mac.link
7535 MACAddressPolicy=none
7538 * The .device units generated by systemd-fstab-generator and other
7539 generators do not automatically pull in the corresponding .mount unit
7540 as a Wants= dependency. This means that simply plugging in the device
7541 will not cause the mount unit to be started automatically. But please
7542 note that the mount unit may be started for other reasons, in
7543 particular if it is part of local-fs.target, and any unit which
7544 (transitively) depends on local-fs.target is started.
7546 * networkctl list/status/lldp now accept globbing wildcards for network
7547 interface names to match against all existing interfaces.
7549 * The $PIDFILE environment variable is set to point the absolute path
7550 configured with PIDFile= for processes of that service.
7552 * The fallback DNS server list was augmented with Cloudflare public DNS
7553 servers. Use `-Ddns-servers=` to set a different fallback.
7555 * A new special target usb-gadget.target will be started automatically
7556 when a USB Device Controller is detected (which means that the system
7557 is a USB peripheral).
7559 * A new unit setting CPUQuotaPeriodSec= assigns the time period
7560 relatively to which the CPU time quota specified by CPUQuota= is
7563 * A new unit setting ProtectHostname= may be used to prevent services
7564 from modifying hostname information (even if they otherwise would
7565 have privileges to do so).
7567 * A new unit setting NetworkNamespacePath= may be used to specify a
7568 namespace for service or socket units through a path referring to a
7569 Linux network namespace pseudo-file.
7571 * The PrivateNetwork= setting and JoinsNamespaceOf= dependencies now
7572 have an effect on .socket units: when used the listening socket is
7573 created within the configured network namespace instead of the host
7576 * ExecStart= command lines in unit files may now be prefixed with ':'
7577 in which case environment variable substitution is
7578 disabled. (Supported for the other ExecXYZ= settings, too.)
7580 * .timer units gained two new boolean settings OnClockChange= and
7581 OnTimezoneChange= which may be used to also trigger a unit when the
7582 system clock is changed or the local timezone is
7583 modified. systemd-run has been updated to make these options easily
7584 accessible from the command line for transient timers.
7586 * Two new conditions for units have been added: ConditionMemory= may be
7587 used to conditionalize a unit based on installed system
7588 RAM. ConditionCPUs= may be used to conditionalize a unit based on
7589 installed CPU cores.
7591 * The @default system call filter group understood by SystemCallFilter=
7592 has been updated to include the new rseq() system call introduced in
7595 * A new time-set.target has been added that indicates that the system
7596 time has been set from a local source (possibly imprecise). The
7597 existing time-sync.target is stronger and indicates that the time has
7598 been synchronized with a precise external source. Services where
7599 approximate time is sufficient should use the new target.
7601 * "systemctl start" (and related commands) learnt a new
7602 --show-transaction option. If specified brief information about all
7603 jobs queued because of the requested operation is shown.
7605 * systemd-networkd recognizes a new operation state 'enslaved', used
7606 (instead of 'degraded' or 'carrier') for interfaces which form a
7607 bridge, bond, or similar, and an new 'degraded-carrier' operational
7608 state used for the bond or bridge master interface when one of the
7609 enslaved devices is not operational.
7611 * .network files learnt the new IgnoreCarrierLoss= option for leaving
7612 networks configured even if the carrier is lost.
7614 * The RequiredForOnline= setting in .network files may now specify a
7615 minimum operational state required for the interface to be considered
7616 "online" by systemd-networkd-wait-online. Related to this
7617 systemd-networkd-wait-online gained a new option --operational-state=
7618 to configure the same, and its --interface= option was updated to
7619 optionally also take an operational state specific for an interface.
7621 * systemd-networkd-wait-online gained a new setting --any for waiting
7622 for only one of the requested interfaces instead of all of them.
7624 * systemd-networkd now implements L2TP tunnels.
7626 * Two new .network settings UseAutonomousPrefix= and UseOnLinkPrefix=
7627 may be used to cause autonomous and onlink prefixes received in IPv6
7628 Router Advertisements to be ignored.
7630 * New MulticastFlood=, NeighborSuppression=, and Learning= .network
7631 file settings may be used to tweak bridge behaviour.
7633 * The new TripleSampling= option in .network files may be used to
7634 configure CAN triple sampling.
7636 * A new .netdev settings PrivateKeyFile= and PresharedKeyFile= may be
7637 used to point to private or preshared key for a WireGuard interface.
7639 * /etc/crypttab now supports the same-cpu-crypt and
7640 submit-from-crypt-cpus options to tweak encryption work scheduling
7643 * systemd-tmpfiles will now take a BSD file lock before operating on a
7644 contents of directory. This may be used to temporarily exclude
7645 directories from aging by taking the same lock (useful for example
7646 when extracting a tarball into /tmp or /var/tmp as a privileged user,
7647 which might create files with really old timestamps, which
7648 nevertheless should not be deleted). For further details, see:
7650 https://systemd.io/TEMPORARY_DIRECTORIES
7652 * systemd-tmpfiles' h line type gained support for the
7653 FS_PROJINHERIT_FL ('P') file attribute (introduced in kernel 4.5),
7654 controlling project quota inheritance.
7656 * sd-boot and bootctl now implement support for an Extended Boot Loader
7657 (XBOOTLDR) partition, that is intended to be mounted to /boot, in
7658 addition to the ESP partition mounted to /efi or /boot/efi.
7659 Configuration file fragments, kernels, initrds and other EFI images
7660 to boot will be loaded from both the ESP and XBOOTLDR partitions.
7661 The XBOOTLDR partition was previously described by the Boot Loader
7662 Specification, but implementation was missing in sd-boot. Support for
7663 this concept allows using the sd-boot boot loader in more
7664 conservative scenarios where the boot loader itself is placed in the
7665 ESP but the kernels to boot (and their metadata) in a separate
7668 * A system may now be booted with systemd.volatile=overlay on the
7669 kernel command line, which causes the root file system to be set up
7670 an overlayfs mount combining the root-only root directory with a
7671 writable tmpfs. In this setup, the underlying root device is not
7672 modified, and any changes are lost at reboot.
7674 * Similar, systemd-nspawn can now boot containers with a volatile
7675 overlayfs root with the new --volatile=overlay switch.
7677 * systemd-nspawn can now consume OCI runtime bundles using a new
7678 --oci-bundle= option. This implementation is fully usable, with most
7679 features in the specification implemented, but since this a lot of
7680 new code and functionality, this feature should most likely not
7681 be used in production yet.
7683 * systemd-nspawn now supports various options described by the OCI
7684 runtime specification on the command-line and in .nspawn files:
7685 --inaccessible=/Inaccessible= may be used to mask parts of the file
7686 system tree, --console=/--pipe may be used to configure how standard
7687 input, output, and error are set up.
7689 * busctl learned the `emit` verb to generate D-Bus signals.
7691 * systemd-analyze cat-config may be used to gather and display
7692 configuration spread over multiple files, for example system and user
7693 presets, tmpfiles.d, sysusers.d, udev rules, etc.
7695 * systemd-analyze calendar now takes an optional new parameter
7696 --iterations= which may be used to show a maximum number of iterations
7697 the specified expression will elapse next.
7699 * The sd-bus C API gained support for naming method parameters in the
7702 * systemd-logind gained D-Bus APIs to specify the "reboot parameter"
7703 the reboot() system call expects.
7705 * journalctl learnt a new --cursor-file= option that points to a file
7706 from which a cursor should be loaded in the beginning and to which
7707 the updated cursor should be stored at the end.
7709 * ACRN hypervisor and Windows Subsystem for Linux (WSL) are now
7710 detected by systemd-detect-virt (and may also be used in
7711 ConditionVirtualization=).
7713 * The behaviour of systemd-logind may now be modified with environment
7714 variables $SYSTEMD_REBOOT_TO_FIRMWARE_SETUP,
7715 $SYSTEMD_REBOOT_TO_BOOT_LOADER_MENU, and
7716 $SYSTEMD_REBOOT_TO_BOOT_LOADER_ENTRY. They cause logind to either
7717 skip the relevant operation completely (when set to false), or to
7718 create a flag file in /run/systemd (when set to true), instead of
7719 actually commencing the real operation when requested. The presence
7720 of /run/systemd/reboot-to-firmware-setup,
7721 /run/systemd/reboot-to-boot-loader-menu, and
7722 /run/systemd/reboot-to-boot-loader-entry, may be used by alternative
7723 boot loader implementations to replace some steps logind performs
7724 during reboot with their own operations.
7726 * systemctl can be used to request a reboot into the boot loader menu
7727 or a specific boot loader entry with the new --boot-load-menu= and
7728 --boot-loader-entry= options to a reboot command. (This requires a
7729 boot loader that supports this, for example sd-boot.)
7731 * kernel-install will no longer unconditionally create the output
7732 directory (e.g. /efi/<machine-id>/<kernel-version>) for boot loader
7733 snippets, but will do only if the machine-specific parent directory
7734 (i.e. /efi/<machine-id>/) already exists. bootctl has been modified
7735 to create this parent directory during sd-boot installation.
7737 This makes it easier to use kernel-install with plugins which support
7738 a different layout of the bootloader partitions (for example grub2).
7740 * During package installation (with `ninja install`), we would create
7741 symlinks for getty@tty1.service, systemd-networkd.service,
7742 systemd-networkd.socket, systemd-resolved.service,
7743 remote-cryptsetup.target, remote-fs.target,
7744 systemd-networkd-wait-online.service, and systemd-timesyncd.service
7745 in /etc, as if `systemctl enable` was called for those units, to make
7746 the system usable immediately after installation. Now this is not
7747 done anymore, and instead calling `systemctl preset-all` is
7748 recommended after the first installation of systemd.
7750 * A new boolean sandboxing option RestrictSUIDSGID= has been added that
7751 is built on seccomp. When turned on creation of SUID/SGID files is
7754 * The NoNewPrivileges= and the new RestrictSUIDSGID= options are now
7755 implied if DynamicUser= is turned on for a service. This hardens
7756 these services, so that they neither can benefit from nor create
7757 SUID/SGID executables. This is a minor compatibility breakage, given
7758 that when DynamicUser= was first introduced SUID/SGID behaviour was
7759 unaffected. However, the security benefit of these two options is
7760 substantial, and the setting is still relatively new, hence we opted
7761 to make it mandatory for services with dynamic users.
7763 Contributions from: Adam Jackson, Alexander Tsoy, Andrey Yashkin,
7764 Andrzej Pietrasiewicz, Anita Zhang, Balint Reczey, Beniamino Galvani,
7765 Ben Iofel, Benjamin Berg, Benjamin Dahlhoff, Chris, Chris Morin,
7766 Christopher Wong, Claudius Ellsel, Clemens Gruber, dana, Daniel Black,
7767 Davide Cavalca, David Michael, David Rheinsberg, emersion, Evgeny
7768 Vereshchagin, Filipe Brandenburger, Franck Bui, Frantisek Sumsal,
7769 Giacinto Cifelli, Hans de Goede, Hugo Kindel, Ignat Korchagin, Insun
7770 Pyo, Jan Engelhardt, Jonas Dorel, Jonathan Lebon, Jonathon Kowalski,
7771 Jörg Sommer, Jörg Thalheim, Jussi Pakkanen, Kai-Heng Feng, Lennart
7772 Poettering, Lubomir Rintel, Luís Ferreira, Martin Pitt, Matthias
7773 Klumpp, Michael Biebl, Michael Niewöhner, Michael Olbrich, Michal
7774 Sekletar, Mike Lothian, Paul Menzel, Piotr Drąg, Riccardo Schirone,
7775 Robin Elvedi, Roman Kulikov, Ronald Tschalär, Ross Burton, Ryan
7776 Gonzalez, Sebastian Krzyszkowiak, Stephane Chazelas, StKob, Susant
7777 Sahani, Sylvain Plantefève, Szabolcs Fruhwald, Taro Yamada, Theo
7778 Ouzhinski, Thomas Haller, Tobias Jungel, Tom Yan, Tony Asleson, Topi
7779 Miettinen, unixsysadmin, Van Laser, Vesa Jääskeläinen, Yu, Li-Yu,
7780 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
7782 — Warsaw, 2019-04-11
7786 * The default locale can now be configured at compile time. Otherwise,
7787 a suitable default will be selected automatically (one of C.UTF-8,
7788 en_US.UTF-8, and C).
7790 * The version string shown by systemd and other tools now includes the
7791 git commit hash when built from git. An override may be specified
7792 during compilation, which is intended to be used by distributions to
7793 include the package release information.
7795 * systemd-cat can now filter standard input and standard error streams
7796 for different syslog priorities using the new --stderr-priority=
7799 * systemd-journald and systemd-journal-remote reject entries which
7800 contain too many fields (CVE-2018-16865) and set limits on the
7801 process' command line length (CVE-2018-16864).
7803 * $DBUS_SESSION_BUS_ADDRESS environment variable is set by pam_systemd
7806 * A new network device NamePolicy "keep" is implemented for link files,
7807 and used by default in 99-default.link (the fallback configuration
7808 provided by systemd). With this policy, if the network device name
7809 was already set by userspace, the device will not be renamed again.
7810 This matches the naming scheme that was implemented before
7811 systemd-240. If naming-scheme < 240 is specified, the "keep" policy
7812 is also enabled by default, even if not specified. Effectively, this
7813 means that if naming-scheme >= 240 is specified, network devices will
7814 be renamed according to the configuration, even if they have been
7815 renamed already, if "keep" is not specified as the naming policy in
7816 the .link file. The 99-default.link file provided by systemd includes
7817 "keep" for backwards compatibility, but it is recommended for user
7818 installed .link files to *not* include it.
7820 The "kernel" policy, which keeps kernel names declared to be
7821 "persistent", now works again as documented.
7823 * kernel-install script now optionally takes the paths to one or more
7824 initrd files, and passes them to all plugins.
7826 * The mincore() system call has been dropped from the @system-service
7827 system call filter group, as it is pretty exotic and may potentially
7828 used for side-channel attacks.
7830 * -fPIE is dropped from compiler and linker options. Please specify
7831 -Db_pie=true option to meson to build position-independent
7832 executables. Note that the meson option is supported since meson-0.49.
7834 * The fs.protected_regular and fs.protected_fifos sysctls, which were
7835 added in Linux 4.19 to make some data spoofing attacks harder, are
7836 now enabled by default. While this will hopefully improve the
7837 security of most installations, it is technically a backwards
7838 incompatible change; to disable these sysctls again, place the
7839 following lines in /etc/sysctl.d/60-protected.conf or a similar file:
7841 fs.protected_regular = 0
7842 fs.protected_fifos = 0
7844 Note that the similar hardlink and symlink protection has been
7845 enabled since v199, and may be disabled likewise.
7847 * The files read from the EnvironmentFile= setting in unit files now
7848 parse backslashes inside quotes literally, matching the behaviour of
7851 * udevadm trigger, udevadm control, udevadm settle and udevadm monitor
7852 now automatically become NOPs when run in a chroot() environment.
7854 * The tmpfiles.d/ "C" line type will now copy directory trees not only
7855 when the destination is so far missing, but also if it already exists
7856 as a directory and is empty. This is useful to cater for systems
7857 where directory trees are put together from multiple separate mount
7858 points but otherwise empty.
7860 * A new function sd_bus_close_unref() (and the associated
7861 sd_bus_close_unrefp()) has been added to libsystemd, that combines
7862 sd_bus_close() and sd_bus_unref() in one.
7864 * udevadm control learnt a new option for --ping for testing whether a
7865 systemd-udevd instance is running and reacting.
7867 * udevadm trigger learnt a new option for --wait-daemon for waiting
7868 systemd-udevd daemon to be initialized.
7870 Contributions from: Aaron Plattner, Alberts Muktupāvels, Alex Mayer,
7871 Ayman Bagabas, Beniamino Galvani, Burt P, Chris Down, Chris Lamb, Chris
7872 Morin, Christian Hesse, Claudius Ellsel, dana, Daniel Axtens, Daniele
7873 Medri, Dave Reisner, David Santamaría Rogado, Diego Canuhe, Dimitri
7874 John Ledkov, Evgeny Vereshchagin, Fabrice Fontaine, Filipe
7875 Brandenburger, Franck Bui, Frantisek Sumsal, govwin, Hans de Goede,
7876 James Hilliard, Jan Engelhardt, Jani Uusitalo, Jan Janssen, Jan
7877 Synacek, Jonathan McDowell, Jonathan Roemer, Jonathon Kowalski, Joost
7878 Heitbrink, Jörg Thalheim, Lance, Lennart Poettering, Louis Taylor,
7879 Lucas Werkmeister, Mantas Mikulėnas, Marc-Antoine Perennou,
7880 marvelousblack, Michael Biebl, Michael Sloan, Michal Sekletar, Mike
7881 Auty, Mike Gilbert, Mikhail Kasimov, Neil Brown, Niklas Hambüchen,
7882 Patrick Williams, Paul Seyfert, Peter Hutterer, Philip Withnall, Roger
7883 James, Ronnie P. Thomas, Ryan Gonzalez, Sam Morris, Stephan Edel,
7884 Stephan Gerhold, Susant Sahani, Taro Yamada, Thomas Haller, Topi
7885 Miettinen, YiFei Zhu, YmrDtnJu, YunQiang Su, Yu Watanabe, Zbigniew
7886 Jędrzejewski-Szmek, zsergeant77, Дамјан Георгиевски
7888 — Berlin, 2019-02-14
7892 * NoNewPrivileges=yes has been set for all long-running services
7893 implemented by systemd. Previously, this was problematic due to
7894 SELinux (as this would also prohibit the transition from PID1's label
7895 to the service's label). This restriction has since been lifted, but
7896 an SELinux policy update is required.
7897 (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.)
7899 * DynamicUser=yes is dropped from systemd-networkd.service,
7900 systemd-resolved.service and systemd-timesyncd.service, which was
7901 enabled in v239 for systemd-networkd.service and systemd-resolved.service,
7902 and since v236 for systemd-timesyncd.service. The users and groups
7903 systemd-network, systemd-resolve and systemd-timesync are created
7904 by systemd-sysusers again. Distributors or system administrators
7905 may need to create these users and groups if they not exist (or need
7906 to re-enable DynamicUser= for those units) while upgrading systemd.
7907 Also, the clock file for systemd-timesyncd may need to move from
7908 /var/lib/private/systemd/timesync/clock to /var/lib/systemd/timesync/clock.
7910 * When unit files are loaded from disk, previously systemd would
7911 sometimes (depending on the unit loading order) load units from the
7912 target path of symlinks in .wants/ or .requires/ directories of other
7913 units. This meant that unit could be loaded from different paths
7914 depending on whether the unit was requested explicitly or as a
7915 dependency of another unit, not honouring the priority of directories
7916 in search path. It also meant that it was possible to successfully
7917 load and start units which are not found in the unit search path, as
7918 long as they were requested as a dependency and linked to from
7919 .wants/ or .requires/. The target paths of those symlinks are not
7920 used for loading units anymore and the unit file must be found in
7923 * A new service type has been added: Type=exec. It's very similar to
7924 Type=simple but ensures the service manager will wait for both fork()
7925 and execve() of the main service binary to complete before proceeding
7926 with follow-up units. This is primarily useful so that the manager
7927 propagates any errors in the preparation phase of service execution
7928 back to the job that requested the unit to be started. For example,
7929 consider a service that has ExecStart= set to a file system binary
7930 that doesn't exist. With Type=simple starting the unit would be
7931 considered instantly successful, as only fork() has to complete
7932 successfully and the manager does not wait for execve(), and hence
7933 its failure is seen "too late". With the new Type=exec service type
7934 starting the unit will fail, as the manager will wait for the
7935 execve() and notice its failure, which is then propagated back to the
7938 NOTE: with the next release 241 of systemd we intend to change the
7939 systemd-run tool to default to Type=exec for transient services
7940 started by it. This should be mostly safe, but in specific corner
7941 cases might result in problems, as the systemd-run tool will then
7942 block on NSS calls (such as user name look-ups due to User=) done
7943 between the fork() and execve(), which under specific circumstances
7944 might cause problems. It is recommended to specify "-p Type=simple"
7945 explicitly in the few cases where this applies. For regular,
7946 non-transient services (i.e. those defined with unit files on disk)
7947 we will continue to default to Type=simple.
7949 * The Linux kernel's current default RLIMIT_NOFILE resource limit for
7950 userspace processes is set to 1024 (soft) and 4096
7951 (hard). Previously, systemd passed this on unmodified to all
7952 processes it forked off. With this systemd release the hard limit
7953 systemd passes on is increased to 512K, overriding the kernel's
7954 defaults and substantially increasing the number of simultaneous file
7955 descriptors unprivileged userspace processes can allocate. Note that
7956 the soft limit remains at 1024 for compatibility reasons: the
7957 traditional UNIX select() call cannot deal with file descriptors >=
7958 1024 and increasing the soft limit globally might thus result in
7959 programs unexpectedly allocating a high file descriptor and thus
7960 failing abnormally when attempting to use it with select() (of
7961 course, programs shouldn't use select() anymore, and prefer
7962 poll()/epoll, but the call unfortunately remains undeservedly popular
7963 at this time). This change reflects the fact that file descriptor
7964 handling in the Linux kernel has been optimized in more recent
7965 kernels and allocating large numbers of them should be much cheaper
7966 both in memory and in performance than it used to be. Programs that
7967 want to take benefit of the increased limit have to "opt-in" into
7968 high file descriptors explicitly by raising their soft limit. Of
7969 course, when they do that they must acknowledge that they cannot use
7970 select() anymore (and neither can any shared library they use — or
7971 any shared library used by any shared library they use and so on).
7972 Which default hard limit is most appropriate is of course hard to
7973 decide. However, given reports that ~300K file descriptors are used
7974 in real-life applications we believe 512K is sufficiently high as new
7975 default for now. Note that there are also reports that using very
7976 high hard limits (e.g. 1G) is problematic: some software allocates
7977 large arrays with one element for each potential file descriptor
7978 (Java, …) — a high hard limit thus triggers excessively large memory
7979 allocations in these applications. Hopefully, the new default of 512K
7980 is a good middle ground: higher than what real-life applications
7981 currently need, and low enough for avoid triggering excessively large
7982 allocations in problematic software. (And yes, somebody should fix
7985 * The fs.nr_open and fs.file-max sysctls are now automatically bumped
7986 to the highest possible values, as separate accounting of file
7987 descriptors is no longer necessary, as memcg tracks them correctly as
7988 part of the memory accounting anyway. Thus, from the four limits on
7989 file descriptors currently enforced (fs.file-max, fs.nr_open,
7990 RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two,
7991 and keep only the latter two. A set of build-time options
7992 (-Dbump-proc-sys-fs-file-max=false and -Dbump-proc-sys-fs-nr-open=false)
7993 has been added to revert this change in behaviour, which might be
7994 an option for systems that turn off memcg in the kernel.
7996 * When no /etc/locale.conf file exists (and hence no locale settings
7997 are in place), systemd will now use the "C.UTF-8" locale by default,
7998 and set LANG= to it. This locale is supported by various
7999 distributions including Fedora, with clear indications that upstream
8000 glibc is going to make it available too. This locale enables UTF-8
8001 mode by default, which appears appropriate for 2018.
8003 * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by
8004 default. This effectively switches the RFC3704 Reverse Path filtering
8005 from Strict mode to Loose mode. This is more appropriate for hosts
8006 that have multiple links with routes to the same networks (e.g.
8007 a client with a Wi-Fi and Ethernet both connected to the internet).
8009 Consult the kernel documentation for details on this sysctl:
8010 https://docs.kernel.org/networking/ip-sysctl.html
8012 * The v239 change to turn on "net.ipv4.tcp_ecn" by default has been
8015 * CPUAccounting=yes no longer enables the CPU controller when using
8016 kernel 4.15+ and the unified cgroup hierarchy, as required accounting
8017 statistics are now provided independently from the CPU controller.
8019 * Support for disabling a particular cgroup controller within a sub-tree
8020 has been added through the DisableControllers= directive.
8022 * cgroup_no_v1=all on the kernel command line now also implies
8023 using the unified cgroup hierarchy, unless one explicitly passes
8024 systemd.unified_cgroup_hierarchy=0 on the kernel command line.
8026 * The new "MemoryMin=" unit file property may now be used to set the
8027 memory usage protection limit of processes invoked by the unit. This
8028 controls the cgroup v2 memory.min attribute. Similarly, the new
8029 "IODeviceLatencyTargetSec=" property has been added, wrapping the new
8030 cgroup v2 io.latency cgroup property for configuring per-service I/O
8033 * systemd now supports the cgroup v2 devices BPF logic, as counterpart
8034 to the cgroup v1 "devices" cgroup controller.
8036 * systemd-escape now is able to combine --unescape with --template. It
8037 also learnt a new option --instance for extracting and unescaping the
8038 instance part of a unit name.
8040 * sd-bus now provides the sd_bus_message_readv() which is similar to
8041 sd_bus_message_read() but takes a va_list object. The pair
8042 sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout()
8043 has been added for configuring the default method call timeout to
8044 use. sd_bus_error_move() may be used to efficiently move the contents
8045 from one sd_bus_error structure to another, invalidating the
8046 source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may
8047 be used to control whether a bus connection object is automatically
8048 flushed when an sd-event loop is exited.
8050 * When processing classic BSD syslog log messages, journald will now
8051 save the original time-stamp string supplied in the new
8052 SYSLOG_TIMESTAMP= journal field. This permits consumers to
8053 reconstruct the original BSD syslog message more correctly.
8055 * StandardOutput=/StandardError= in service files gained support for
8056 new "append:…" parameters, for connecting STDOUT/STDERR of a service
8057 to a file, and appending to it.
8059 * The signal to use as last step of killing of unit processes is now
8060 configurable. Previously it was hard-coded to SIGKILL, which may now
8061 be overridden with the new KillSignal= setting. Note that this is the
8062 signal used when regular termination (i.e. SIGTERM) does not suffice.
8063 Similarly, the signal used when aborting a program in case of a
8064 watchdog timeout may now be configured too (WatchdogSignal=).
8066 * The XDG_SESSION_DESKTOP environment variable may now be configured in
8067 the pam_systemd argument line, using the new desktop= switch. This is
8068 useful to initialize it properly from a display manager without
8069 having to touch C code.
8071 * Most configuration options that previously accepted percentage values
8072 now also accept permille values with the '‰' suffix (instead of '%').
8074 * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for
8077 * systemd-resolved's configuration file resolved.conf gained a new
8078 option ReadEtcHosts= which may be used to turn off processing and
8079 honoring /etc/hosts entries.
8081 * The "--wait" switch may now be passed to "systemctl
8082 is-system-running", in which case the tool will synchronously wait
8083 until the system finished start-up.
8085 * hostnamed gained a new bus call to determine the DMI product UUID.
8087 * On x86-64 systemd will now prefer using the RDRAND processor
8088 instruction over /dev/urandom whenever it requires randomness that
8089 neither has to be crypto-grade nor should be reproducible. This
8090 should substantially reduce the amount of entropy systemd requests
8091 from the kernel during initialization on such systems, though not
8092 reduce it to zero. (Why not zero? systemd still needs to allocate
8093 UUIDs and such uniquely, which require high-quality randomness.)
8095 * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP
8096 tunnels. It also gained a new option ForceDHCPv6PDOtherInformation=
8097 for forcing the "Other Information" bit in IPv6 RA messages. The
8098 bonding logic gained four new options AdActorSystemPriority=,
8099 AdUserPortKey=, AdActorSystem= for configuring various 802.3ad
8100 aspects, and DynamicTransmitLoadBalancing= for enabling dynamic
8101 shuffling of flows. The tunnel logic gained a new
8102 IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid
8103 Deployment. The policy rule logic gained four new options IPProtocol=,
8104 SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained
8105 support for the MulticastToUnicast= option. networkd also gained
8106 support for configuring static IPv4 ARP or IPv6 neighbor entries.
8108 * .preset files (as read by 'systemctl preset') may now be used to
8109 instantiate services.
8111 * /etc/crypttab now understands the sector-size= option to configure
8112 the sector size for an encrypted partition.
8114 * Key material for encrypted disks may now be placed on a formatted
8115 medium, and referenced from /etc/crypttab by the UUID of the file
8116 system, followed by "=" suffixed by the path to the key file.
8118 * The "collect" udev component has been removed without replacement, as
8119 it is neither used nor maintained.
8121 * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=,
8122 LogsDirectory=, ConfigurationDirectory= settings are used in a
8123 service the executed processes will now receive a set of environment
8124 variables containing the full paths of these directories.
8125 Specifically, RUNTIME_DIRECTORY=, STATE_DIRECTORY, CACHE_DIRECTORY,
8126 LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options
8127 are used. Note that these options may be used multiple times per
8128 service in which case the resulting paths will be concatenated and
8129 separated by colons.
8131 * Predictable interface naming has been extended to cover InfiniBand
8132 NICs. They will be exposed with an "ib" prefix.
8134 * tmpfiles.d/ line types may now be suffixed with a '-' character, in
8135 which case the respective line failing is ignored.
8137 * .link files may now be used to configure the equivalent to the
8138 "ethtool advertise" commands.
8140 * The sd-device.h and sd-hwdb.h APIs are now exported, as an
8141 alternative to libudev.h. Previously, the latter was just an internal
8142 wrapper around the former, but now these two APIs are exposed
8145 * sd-id128.h gained a new function sd_id128_get_boot_app_specific()
8146 which calculates an app-specific boot ID similar to how
8147 sd_id128_get_machine_app_specific() generates an app-specific machine
8150 * A new tool systemd-id128 has been added that can be used to determine
8151 and generate various 128-bit IDs.
8153 * /etc/os-release gained two new standardized fields DOCUMENTATION_URL=
8156 * systemd-hibernate-resume-generator will now honor the "noresume"
8157 kernel command line option, in which case it will bypass resuming
8158 from any hibernated image.
8160 * The systemd-sleep.conf configuration file gained new options
8161 AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=,
8162 AllowHybridSleep= for prohibiting specific sleep modes even if the
8163 kernel exports them.
8165 * portablectl is now officially supported and has thus moved to
8168 * bootctl learnt the two new commands "set-default" and "set-oneshot"
8169 for setting the default boot loader item to boot to (either
8170 persistently or only for the next boot). This is currently only
8171 compatible with sd-boot, but may be implemented on other boot loaders
8172 too, that follow the boot loader interface. The updated interface is
8173 now documented here:
8175 https://systemd.io/BOOT_LOADER_INTERFACE
8177 * A new kernel command line option systemd.early_core_pattern= is now
8178 understood which may be used to influence the core_pattern PID 1
8179 installs during early boot.
8181 * busctl learnt two new options -j and --json= for outputting method
8182 call replies, properties and monitoring output in JSON.
8184 * journalctl's JSON output now supports simple ANSI coloring as well as
8185 a new "json-seq" mode for generating RFC7464 output.
8187 * Unit files now support the %g/%G specifiers that resolve to the UNIX
8188 group/GID of the service manager runs as, similar to the existing
8189 %u/%U specifiers that resolve to the UNIX user/UID.
8191 * systemd-logind learnt a new global configuration option
8192 UserStopDelaySec= that may be set in logind.conf. It specifies how
8193 long the systemd --user instance shall remain started after a user
8194 logs out. This is useful to speed up repetitive re-connections of the
8195 same user, as it means the user's service manager doesn't have to be
8196 stopped/restarted on each iteration, but can be reused between
8197 subsequent options. This setting defaults to 10s. systemd-logind also
8198 exports two new properties on its Manager D-Bus objects indicating
8199 whether the system's lid is currently closed, and whether the system
8202 * systemd gained support for a generic boot counting logic, which
8203 generically permits automatic reverting to older boot loader entries
8204 if newer updated ones don't work. The boot loader side is implemented
8205 in sd-boot, but is kept open for other boot loaders too. For details
8208 https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT
8210 * The SuccessAction=/FailureAction= unit file settings now learnt two
8211 new parameters: "exit" and "exit-force", which result in immediate
8212 exiting of the service manager, and are only useful in systemd --user
8213 and container environments.
8215 * Unit files gained support for a pair of options
8216 FailureActionExitStatus=/SuccessActionExitStatus= for configuring the
8217 exit status to use as service manager exit status when
8218 SuccessAction=/FailureAction= is set to exit or exit-force.
8220 * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service
8221 options may now be used to configure the log rate limiting applied by
8222 journald per-service.
8224 * systemd-analyze gained a new verb "timespan" for parsing and
8225 normalizing time span values (i.e. strings like "5min 7s 8us").
8227 * systemd-analyze also gained a new verb "security" for analyzing the
8228 security and sand-boxing settings of services in order to determine an
8229 "exposure level" for them, indicating whether a service would benefit
8230 from more sand-boxing options turned on for them.
8232 * "systemd-analyze syscall-filter" will now also show system calls
8233 supported by the local kernel but not included in any of the defined
8236 * .nspawn files now understand the Ephemeral= setting, matching the
8237 --ephemeral command line switch.
8239 * sd-event gained the new APIs sd_event_source_get_floating() and
8240 sd_event_source_set_floating() for controlling whether a specific
8241 event source is "floating", i.e. destroyed along with the even loop
8244 * Unit objects on D-Bus gained a new "Refs" property that lists all
8245 clients that currently have a reference on the unit (to ensure it is
8248 * The JoinControllers= option in system.conf is no longer supported, as
8249 it didn't work correctly, is hard to support properly, is legacy (as
8250 the concept only exists on cgroup v1) and apparently wasn't used.
8252 * Journal messages that are generated whenever a unit enters the failed
8253 state are now tagged with a unique MESSAGE_ID. Similarly, messages
8254 generated whenever a service process exits are now made recognizable,
8255 too. A tagged message is also emitted whenever a unit enters the
8256 "dead" state on success.
8258 * systemd-run gained a new switch --working-directory= for configuring
8259 the working directory of the service to start. A shortcut -d is
8260 equivalent, setting the working directory of the service to the
8261 current working directory of the invoking program. The new --shell
8262 (or just -S) option has been added for invoking the $SHELL of the
8263 caller as a service, and implies --pty --same-dir --wait --collect
8264 --service-type=exec. Or in other words, "systemd-run -S" is now the
8265 quickest way to quickly get an interactive in a fully clean and
8266 well-defined system service context.
8268 * machinectl gained a new verb "import-fs" for importing an OS tree
8269 from a directory. Moreover, when a directory or tarball is imported
8270 and single top-level directory found with the OS itself below the OS
8271 tree is automatically mangled and moved one level up.
8273 * systemd-importd will no longer set up an implicit btrfs loop-back
8274 file system on /var/lib/machines. If one is already set up, it will
8275 continue to be used.
8277 * A new generator "systemd-run-generator" has been added. It will
8278 synthesize a unit from one or more program command lines included in
8279 the kernel command line. This is very useful in container managers
8282 # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"'
8284 This will run "systemd-nspawn" on an image, invoke the specified
8285 command line and immediately shut down the container again, returning
8286 the command line's exit code.
8288 * The block device locking logic is now documented:
8290 https://systemd.io/BLOCK_DEVICE_LOCKING
8292 * loginctl and machinectl now optionally output the various tables in
8293 JSON using the --output= switch. It is our intention to add similar
8294 support to systemctl and all other commands.
8296 * udevadm's query and trigger verb now optionally take a .device unit
8299 * systemd-udevd's network naming logic now understands a new
8300 net.naming-scheme= kernel command line switch, which may be used to
8301 pick a specific version of the naming scheme. This helps stabilizing
8302 interface names even as systemd/udev are updated and the naming logic
8305 * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and
8306 SD_ID128_ALLF to test if a 128-bit ID is set to all 0xFF bytes, and to
8307 initialize one to all 0xFF.
8309 * After loading the SELinux policy systemd will now recursively relabel
8310 all files and directories listed in
8311 /run/systemd/relabel-extra.d/*.relabel (which should be simple
8312 newline separated lists of paths) in addition to the ones it already
8313 implicitly relabels in /run, /dev and /sys. After the relabelling is
8314 completed the *.relabel files (and /run/systemd/relabel-extra.d/) are
8315 removed. This is useful to permit initrds (i.e. code running before
8316 the SELinux policy is in effect) to generate files in the host
8317 filesystem safely and ensure that the correct label is applied during
8318 the transition to the host OS.
8320 * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding
8321 mknod() handling in user namespaces. Previously mknod() would always
8322 fail with EPERM in user namespaces. Since 4.18 mknod() will succeed
8323 but device nodes generated that way cannot be opened, and attempts to
8324 open them result in EPERM. This breaks the "graceful fallback" logic
8325 in systemd's PrivateDevices= sand-boxing option. This option is
8326 implemented defensively, so that when systemd detects it runs in a
8327 restricted environment (such as a user namespace, or an environment
8328 where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD)
8329 where device nodes cannot be created the effect of PrivateDevices= is
8330 bypassed (following the logic that 2nd-level sand-boxing is not
8331 essential if the system systemd runs in is itself already sand-boxed
8332 as a whole). This logic breaks with 4.18 in container managers where
8333 user namespacing is used: suddenly PrivateDevices= succeeds setting
8334 up a private /dev/ file system containing devices nodes — but when
8335 these are opened they don't work.
8337 At this point it is recommended that container managers utilizing
8338 user namespaces that intend to run systemd in the payload explicitly
8339 block mknod() with seccomp or similar, so that the graceful fallback
8342 We are very sorry for the breakage and the requirement to change
8343 container configurations for newer kernels. It's purely caused by an
8344 incompatible kernel change. The relevant kernel developers have been
8345 notified about this userspace breakage quickly, but they chose to
8348 * PermissionsStartOnly= setting is deprecated (but is still supported
8349 for backwards compatibility). The same functionality is provided by
8350 the more flexible "+", "!", and "!!" prefixes to ExecStart= and other
8353 * $DBUS_SESSION_BUS_ADDRESS environment variable is not set by
8354 pam_systemd anymore.
8356 * The naming scheme for network devices was changed to always rename
8357 devices, even if they were already renamed by userspace. The "kernel"
8358 policy was changed to only apply as a fallback, if no other naming
8361 * The requirements to build systemd is bumped to meson-0.46 and
8364 Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander
8365 Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson,
8366 Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov,
8367 asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt
8368 Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen
8369 Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius
8370 Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn
8371 Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner,
8372 David Anderson, Davide Cavalca, David Leeds, David Malcolm, David
8373 Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus,
8374 Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters,
8375 Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad,
8376 Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank
8377 Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe
8378 Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit
8379 Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan
8380 Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld,
8381 javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi
8382 Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens,
8383 Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi,
8384 Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry
8385 Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz
8386 Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier,
8387 Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin
8388 Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko
8389 Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck,
8390 Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich,
8391 Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal
8392 Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal
8393 Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby,
8394 Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł
8395 Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller,
8396 Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez,
8397 Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam
8398 Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher,
8399 Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee
8400 (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen
8401 Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim,
8402 Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas
8403 Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias
8404 Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore
8405 Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech
8406 Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward,
8407 Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe,
8408 Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein
8410 — Warsaw, 2018-12-21
8414 * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
8415 builtin will name network interfaces differently than in previous
8416 versions for virtual network interfaces created with SR-IOV and NPAR
8417 and for devices where the PCI network controller device does not have
8418 a slot number associated.
8420 SR-IOV virtual devices are now named based on the name of the parent
8421 interface, with a suffix of "v<N>", where <N> is the virtual device
8422 number. Previously those virtual devices were named as if completely
8425 The ninth and later NPAR virtual devices will be named following the
8426 scheme used for the first eight NPAR partitions. Previously those
8427 devices were not renamed and the kernel default (eth<n>) was used.
8429 "net_id" will also generate names for PCI devices where the PCI
8430 network controller device does not have an associated slot number
8431 itself, but one of its parents does. Previously those devices were
8432 not renamed and the kernel default (eth<n>) was used.
8434 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
8435 systemd-logind.service. Since v235, IPAddressDeny=any has been set to
8436 the unit. So, it is expected that the default behavior of
8437 systemd-logind is not changed. However, if distribution packagers or
8438 administrators disabled or modified IPAddressDeny= setting by a
8439 drop-in config file, then it may be necessary to update the file to
8440 re-enable AF_INET and AF_INET6 to support network user name services,
8443 * When the RestrictNamespaces= unit property is specified multiple
8444 times, then the specified types are merged now. Previously, only the
8445 last assignment was used. So, if distribution packagers or
8446 administrators modified the setting by a drop-in config file, then it
8447 may be necessary to update the file.
8449 * When OnFailure= is used in combination with Restart= on a service
8450 unit, then the specified units will no longer be triggered on
8451 failures that result in restarting. Previously, the specified units
8452 would be activated each time the unit failed, even when the unit was
8453 going to be restarted automatically. This behaviour contradicted the
8454 documentation. With this release the code is adjusted to match the
8457 * systemd-tmpfiles will now print a notice whenever it encounters
8458 tmpfiles.d/ lines referencing the /var/run/ directory. It will
8459 recommend reworking them to use the /run/ directory instead (for
8460 which /var/run/ is simply a symlinked compatibility alias). This way
8461 systemd-tmpfiles can properly detect line conflicts and merge lines
8462 referencing the same file by two paths, without having to access
8465 * systemctl disable/unmask/preset/preset-all cannot be used with
8466 --runtime. Previously this was allowed, but resulted in unintuitive
8467 behaviour that wasn't useful. systemctl disable/unmask will now undo
8468 both runtime and persistent enablement/masking, i.e. it will remove
8469 any relevant symlinks both in /run and /etc.
8471 * Note that all long-running system services shipped with systemd will
8472 now default to a system call allow list (rather than a deny list, as
8473 before). In particular, systemd-udevd will now enforce one too. For
8474 most cases this should be safe, however downstream distributions
8475 which disabled sandboxing of systemd-udevd (specifically the
8476 MountFlags= setting), might want to disable this security feature
8477 too, as the default allow-listing will prohibit all mount, swap,
8478 reboot and clock changing operations from udev rules.
8480 * sd-boot acquired new loader configuration settings to optionally turn
8481 off Windows and MacOS boot partition discovery as well as
8482 reboot-into-firmware menu items. It is also able to pick a better
8483 screen resolution for HiDPI systems, and now provides loader
8484 configuration settings to change the resolution explicitly.
8486 * systemd-resolved now supports DNS-over-TLS. It's still
8487 turned off by default, use DNSOverTLS=opportunistic to turn it on in
8488 resolved.conf. We intend to make this the default as soon as couple
8489 of additional techniques for optimizing the initial latency caused by
8490 establishing a TLS/TCP connection are implemented.
8492 * systemd-resolved.service and systemd-networkd.service now set
8493 DynamicUser=yes. The users systemd-resolve and systemd-network are
8494 not created by systemd-sysusers anymore.
8496 NOTE: This has a chance of breaking nss-ldap and similar NSS modules
8497 that embed a network facing module into any process using getpwuid()
8498 or related call: the dynamic allocation of the user ID for
8499 systemd-resolved.service means the service manager has to check NSS
8500 if the user name is already taken when forking off the service. Since
8501 the user in the common case won't be defined in /etc/passwd the
8502 lookup is likely to trigger nss-ldap which in turn might use NSS to
8503 ask systemd-resolved for hostname lookups. This will hence result in
8504 a deadlock: a user name lookup in order to start
8505 systemd-resolved.service will result in a hostname lookup for which
8506 systemd-resolved.service needs to be started already. There are
8507 multiple ways to work around this problem: pre-allocate the
8508 "systemd-resolve" user on such systems, so that nss-ldap won't be
8509 triggered; or use a different NSS package that doesn't do networking
8510 in-process but provides a local asynchronous name cache; or configure
8511 the NSS package to avoid lookups for UIDs in the range `pkg-config
8512 systemd --variable=dynamicuidmin` … `pkg-config systemd
8513 --variable=dynamicuidmax`, so that it does not consider itself
8514 authoritative for the same UID range systemd allocates dynamic users
8517 * The systemd-resolve tool has been renamed to resolvectl (it also
8518 remains available under the old name, for compatibility), and its
8519 interface is now verb-based, similar in style to the other <xyz>ctl
8520 tools, such as systemctl or loginctl.
8522 * The resolvectl/systemd-resolve tool also provides 'resolvconf'
8523 compatibility. It may be symlinked under the 'resolvconf' name, in
8524 which case it will take arguments and input compatible with the
8525 Debian and FreeBSD resolvconf tool.
8527 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
8528 where the system initially suspends, and after a timeout resumes and
8531 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
8532 set the client will only send a DUID as client identifier. (EDIT: the
8533 option was broken, and was dropped in v255.)
8535 * The nss-systemd glibc NSS module will now enumerate dynamic users and
8536 groups in effect. Previously, it could resolve UIDs/GIDs to user
8537 names/groups and vice versa, but did not support enumeration.
8539 * journald's Compress= configuration setting now optionally accepts a
8540 byte threshold value. All journal objects larger than this threshold
8541 will be compressed, smaller ones will not. Previously this threshold
8542 was not configurable and set to 512.
8544 * A new system.conf setting NoNewPrivileges= is now available which may
8545 be used to turn off acquisition of new privileges system-wide
8546 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
8547 for all its children). Note that turning this option on means setuid
8548 binaries and file system capabilities lose their special powers.
8549 While turning on this option is a big step towards a more secure
8550 system, doing so is likely to break numerous pre-existing UNIX tools,
8551 in particular su and sudo.
8553 * A new service systemd-time-sync-wait.service has been added. If
8554 enabled it will delay the time-sync.target unit at boot until time
8555 synchronization has been received from the network. This
8556 functionality is useful on systems lacking a local RTC or where it is
8557 acceptable that the boot process shall be delayed by external network
8560 * When hibernating, systemd will now inform the kernel of the image
8561 write offset, on kernels new enough to support this. This means swap
8562 files should work for hibernation now.
8564 * When loading unit files, systemd will now look for drop-in unit files
8565 extensions in additional places. Previously, for a unit file name
8566 "foo-bar-baz.service" it would look for dropin files in
8567 "foo-bar-baz.service.d/*.conf". Now, it will also look in
8568 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
8569 service name truncated after all inner dashes. This scheme allows
8570 writing drop-ins easily that apply to a whole set of unit files at
8571 once. It's particularly useful for mount and slice units (as their
8572 naming is prefix based), but is also useful for service and other
8573 units, for packages that install multiple unit files at once,
8574 following a strict naming regime of beginning the unit file name with
8575 the package's name. Two new specifiers are now supported in unit
8576 files to match this: %j and %J are replaced by the part of the unit
8577 name following the last dash.
8579 * Unit files and other configuration files that support specifier
8580 expansion now understand another three new specifiers: %T and %V will
8581 resolve to /tmp and /var/tmp respectively, or whatever temporary
8582 directory has been set for the calling user. %E will expand to either
8583 /etc (for system units) or $XDG_CONFIG_HOME (for user units).
8585 * The ExecStart= lines of unit files are no longer required to
8586 reference absolute paths. If non-absolute paths are specified the
8587 specified binary name is searched within the service manager's
8588 built-in $PATH, which may be queried with 'systemd-path
8589 search-binaries-default'. It's generally recommended to continue to
8590 use absolute paths for all binaries specified in unit files.
8592 * Units gained a new load state "bad-setting", which is used when a
8593 unit file was loaded, but contained fatal errors which prevent it
8594 from being started (for example, a service unit has been defined
8595 lacking both ExecStart= and ExecStop= lines).
8597 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
8598 support alternative debuggers, for example lldb. The old name
8599 continues to be available however, for compatibility reasons. Use the
8600 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
8601 to pick an alternative debugger instead of the default gdb.
8603 * systemctl and the other tools will now output escape sequences that
8604 generate proper clickable hyperlinks in various terminal emulators
8605 where useful (for example, in the "systemctl status" output you can
8606 now click on the unit file name to quickly open it in the
8607 editor/viewer of your choice). Note that not all terminal emulators
8608 support this functionality yet, but many do. Unfortunately, the
8609 "less" pager doesn't support this yet, hence this functionality is
8610 currently automatically turned off when a pager is started (which
8611 happens quite often due to auto-paging). We hope to remove this
8612 limitation as soon as "less" learns these escape sequences. This new
8613 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
8614 environment variable. For details on these escape sequences see:
8615 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
8617 * networkd's .network files now support a new IPv6MTUBytes= option for
8618 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
8619 option in the [Route] section to configure the MTU to use for
8620 specific routes. It also gained support for configuration of the DHCP
8621 "UserClass" option through the new UserClass= setting. It gained
8622 three new options in the new [CAN] section for configuring CAN
8623 networks. The MULTICAST and ALLMULTI interface flags may now be
8624 controlled explicitly with the new Multicast= and AllMulticast=
8627 * networkd will now automatically make use of the kernel's route
8628 expiration feature, if it is available.
8630 * udevd's .link files now support setting the number of receive and
8631 transmit channels, using the RxChannels=, TxChannels=,
8632 OtherChannels=, CombinedChannels= settings.
8634 * Support for UDPSegmentationOffload= has been removed, given its
8635 limited support in hardware, and waning software support.
8637 * networkd's .netdev files now support creating "netdevsim" interfaces.
8639 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
8640 to query the unit belonging to a specific kernel control group.
8642 * systemd-analyze gained a new verb "cat-config", which may be used to
8643 dump the contents of any configuration file, with all its matching
8644 drop-in files added in, and honouring the usual search and masking
8645 logic applied to systemd configuration files. For example use
8646 "systemd-analyze cat-config systemd/system.conf" to get the complete
8647 system configuration file of systemd how it would be loaded by PID 1
8648 itself. Similar to this, various tools such as systemd-tmpfiles or
8649 systemd-sysusers, gained a new option "--cat-config", which does the
8650 corresponding operation for their own configuration settings. For
8651 example, "systemd-tmpfiles --cat-config" will now output the full
8652 list of tmpfiles.d/ lines in place.
8654 * timedatectl gained three new verbs: "show" shows bus properties of
8655 systemd-timedated, "timesync-status" shows the current NTP
8656 synchronization state of systemd-timesyncd, and "show-timesync"
8657 shows bus properties of systemd-timesyncd.
8659 * systemd-timesyncd gained a bus interface on which it exposes details
8662 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
8663 understood by systemd-timedated. It takes a colon-separated list of
8664 unit names of NTP client services. The list is used by
8665 "timedatectl set-ntp".
8667 * systemd-nspawn gained a new --rlimit= switch for setting initial
8668 resource limits for the container payload. There's a new switch
8669 --hostname= to explicitly override the container's hostname. A new
8670 --no-new-privileges= switch may be used to control the
8671 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
8672 --oom-score-adjust= switch controls the OOM scoring adjustment value
8673 for the payload. The new --cpu-affinity= switch controls the CPU
8674 affinity of the container payload. The new --resolv-conf= switch
8675 allows more detailed control of /etc/resolv.conf handling of the
8676 container. Similarly, the new --timezone= switch allows more detailed
8677 control of /etc/localtime handling of the container.
8679 * systemd-detect-virt gained a new --list switch, which will print a
8680 list of all currently known VM and container environments.
8682 * Support for "Portable Services" has been added, see
8683 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
8684 experimental, but this is expected to change soon. Reflecting this
8685 experimental state, the "portablectl" binary is not installed into
8686 /usr/bin yet. The binary has to be called with the full path
8687 /usr/lib/systemd/portablectl instead.
8689 * journalctl's and systemctl's -o switch now knows a new log output
8690 mode "with-unit". The output it generates is very similar to the
8691 regular "short" mode, but displays the unit name instead of the
8692 syslog tag for each log line. Also, the date is shown with timezone
8693 information. This mode is probably more useful than the classic
8694 "short" output mode for most purposes, except where pixel-perfect
8695 compatibility with classic /var/log/messages formatting is required.
8697 * A new --dump-bus-properties switch has been added to the systemd
8698 binary, which may be used to dump all supported D-Bus properties.
8699 (Options which are still supported, but are deprecated, are *not*
8702 * sd-bus gained a set of new calls:
8703 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
8704 enable/disable the "floating" state of a bus slot object,
8705 i.e. whether the slot object pins the bus it is allocated for into
8706 memory or if the bus slot object gets disconnected when the bus goes
8707 away. sd_bus_open_with_description(),
8708 sd_bus_open_user_with_description(),
8709 sd_bus_open_system_with_description() may be used to allocate bus
8710 objects and set their description string already during allocation.
8712 * sd-event gained support for watching inotify events from the event
8713 loop, in an efficient way, sharing inotify handles between multiple
8714 users. For this a new function sd_event_add_inotify() has been added.
8716 * sd-event and sd-bus gained support for calling special user-supplied
8717 destructor functions for userdata pointers associated with
8718 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
8719 functions sd_bus_slot_set_destroy_callback,
8720 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
8721 sd_bus_track_get_destroy_callback,
8722 sd_event_source_set_destroy_callback,
8723 sd_event_source_get_destroy_callback have been added.
8725 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
8727 * PID 1 will now automatically reschedule .timer units whenever the
8728 local timezone changes. (They previously got rescheduled
8729 automatically when the system clock changed.)
8731 * New documentation has been added to document cgroups delegation,
8732 portable services and the various code quality tools we have set up:
8734 https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md
8735 https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md
8736 https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md
8738 * The Boot Loader Specification has been added to the source tree.
8740 https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md
8742 While moving it into our source tree we have updated it and further
8743 changes are now accepted through the usual github PR workflow.
8745 * pam_systemd will now look for PAM userdata fields systemd.memory_max,
8746 systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
8747 earlier PAM modules. The data in these fields is used to initialize
8748 the session scope's resource properties. Thus external PAM modules
8749 may now configure per-session limits, for example sourced from
8750 external user databases.
8752 * socket units with Accept=yes will now maintain a "refused" counter in
8753 addition to the existing "accepted" counter, counting connections
8754 refused due to the enforced limits.
8756 * The "systemd-path search-binaries-default" command may now be use to
8757 query the default, built-in $PATH PID 1 will pass to the services it
8760 * A new unit file setting PrivateMounts= has been added. It's a boolean
8761 option. If enabled the unit's processes are invoked in their own file
8762 system namespace. Note that this behaviour is also implied if any
8763 other file system namespacing options (such as PrivateTmp=,
8764 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
8765 primarily useful for services that do not use any of the other file
8766 system namespacing options. One such service is systemd-udevd.service
8767 where this is now used by default.
8769 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
8770 when the system is booted in UEFI "secure mode".
8772 * A new unit "system-update-pre.target" is added, which defines an
8773 optional synchronization point for offline system updates, as
8774 implemented by the pre-existing "system-update.target" unit. It
8775 allows ordering services before the service that executes the actual
8776 update process in a generic way.
8778 * Systemd now emits warnings whenever .include syntax is used.
8780 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
8781 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
8782 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
8783 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
8784 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
8785 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
8786 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
8787 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
8788 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
8789 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
8790 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
8791 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
8792 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
8793 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
8794 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
8795 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
8796 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
8797 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
8798 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
8799 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
8800 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
8801 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
8802 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
8803 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
8804 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
8805 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
8806 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
8807 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
8808 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
8810 — Berlin, 2018-06-22
8814 * The MemoryAccounting= unit property now defaults to on. After
8815 discussions with the upstream control group maintainers we learnt
8816 that the negative impact of cgroup memory accounting on current
8817 kernels is finally relatively minimal, so that it should be safe to
8818 enable this by default without affecting system performance. Besides
8819 memory accounting only task accounting is turned on by default, all
8820 other forms of resource accounting (CPU, IO, IP) remain off for now,
8821 because it's not clear yet that their impact is small enough to move
8822 from opt-in to opt-out. We recommend downstreams to leave memory
8823 accounting on by default if kernel 4.14 or higher is primarily
8824 used. On very resource constrained systems or when support for old
8825 kernels is a necessity, -Dmemory-accounting-default=false can be used
8826 to revert this change.
8828 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
8829 %udev_rules_update) and the journal catalog (%journal_catalog_update)
8830 from the upgrade scriptlets of individual packages now do nothing.
8831 Transfiletriggers have been added which will perform those updates
8832 once at the end of the transaction.
8834 Similar transfiletriggers have been added to execute any sysctl.d
8835 and binfmt.d rules. Thus, it should be unnecessary to provide any
8836 scriptlets to execute this configuration from package installation
8839 * systemd-sysusers gained a mode where the configuration to execute is
8840 specified on the command line, but this configuration is not executed
8841 directly, but instead it is merged with the configuration on disk,
8842 and the result is executed. This is useful for package installation
8843 scripts which want to create the user before installing any files on
8844 disk (in case some of those files are owned by that user), while
8845 still allowing local admin overrides.
8847 This functionality is exposed to rpm scriptlets through a new
8848 %sysusers_create_package macro. Old %sysusers_create and
8849 %sysusers_create_inline macros are deprecated.
8851 A transfiletrigger for sysusers.d configuration is now installed,
8852 which means that it should be unnecessary to call systemd-sysusers from
8853 package installation scripts, unless the package installs any files
8854 owned by those newly-created users, in which case
8855 %sysusers_create_package should be used.
8857 * Analogous change has been done for systemd-tmpfiles: it gained a mode
8858 where the command-line configuration is merged with the configuration
8859 on disk. This is exposed as the new %tmpfiles_create_package macro,
8860 and %tmpfiles_create is deprecated. A transfiletrigger is installed
8861 for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
8862 from package installation scripts.
8864 * sysusers.d configuration for a user may now also specify the group
8865 number, in addition to the user number ("u username 123:456"), or
8866 without the user number ("u username -:456").
8868 * Configution items for systemd-sysusers can now be specified as
8869 positional arguments when the new --inline switch is used.
8871 * The login shell of users created through sysusers.d may now be
8872 specified (previously, it was always /bin/sh for root and
8873 /sbin/nologin for other users).
8875 * systemd-analyze gained a new --global switch to look at global user
8876 configuration. It also gained a unit-paths verb to list the unit load
8877 paths that are compiled into systemd (which can be used with
8878 --systemd, --user, or --global).
8880 * udevadm trigger gained a new --settle/-w option to wait for any
8881 triggered events to finish (but just those, and not any other events
8882 which are triggered meanwhile).
8884 * The action that systemd-logind takes when the lid is closed and the
8885 machine is connected to external power can now be configured using
8886 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
8887 was determined by HandleLidSwitch=, and, for backwards compatibility,
8888 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
8890 * journalctl will periodically call sd_journal_process() to make it
8891 resilient against inotify queue overruns when journal files are
8892 rotated very quickly.
8894 * Two new functions in libsystemd — sd_bus_get_n_queued_read and
8895 sd_bus_get_n_queued_write — may be used to check the number of
8896 pending bus messages.
8898 * systemd gained a new
8899 org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
8900 which can be used to migrate foreign processes to scope and service
8901 units. The primary user for this new API is systemd itself: the
8902 systemd --user instance uses this call of the systemd --system
8903 instance to migrate processes if it itself gets the request to
8904 migrate processes and the kernel refuses this due to access
8905 restrictions. Thanks to this "systemd-run --scope --user …" works
8906 again in pure cgroup v2 environments when invoked from the user
8909 * A new TemporaryFileSystem= setting can be used to mask out part of
8910 the real file system tree with tmpfs mounts. This may be combined
8911 with BindPaths= and BindReadOnlyPaths= to hide files or directories
8912 not relevant to the unit, while still allowing some paths lower in
8913 the tree to be accessed.
8915 ProtectHome=tmpfs may now be used to hide user home and runtime
8916 directories from units, in a way that is mostly equivalent to
8917 "TemporaryFileSystem=/home /run/user /root".
8919 * Non-service units are now started with KeyringMode=shared by default.
8920 This means that mount and swapon and other mount tools have access
8921 to keys in the main keyring.
8923 * /sys/fs/bpf is now mounted automatically.
8925 * QNX virtualization is now detected by systemd-detect-virt and may
8926 be used in ConditionVirtualization=.
8928 * IPAccounting= may now be enabled also for slice units.
8930 * A new -Dsplit-bin= build configuration switch may be used to specify
8931 whether bin and sbin directories are merged, or if they should be
8932 included separately in $PATH and various listings of executable
8933 directories. The build configuration scripts will try to autodetect
8934 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
8935 system, but distributions are encouraged to configure this
8938 * A new -Dok-color= build configuration switch may be used to change
8939 the colour of "OK" status messages.
8941 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
8942 PrivateNetwork=yes was buggy in previous versions of systemd. This
8943 means that after the upgrade and daemon-reexec, any such units must
8946 * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
8947 will not exclude read-only files owned by root from cleanup.
8949 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
8950 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
8951 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
8952 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
8953 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
8954 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
8955 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
8956 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
8957 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
8958 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
8959 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
8960 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
8961 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
8962 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
8963 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
8964 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
8966 — Warsaw, 2018-03-05
8970 * Some keyboards come with a zoom see-saw or rocker which until now got
8971 mapped to the Linux "zoomin/out" keys in hwdb. However, these
8972 keycodes are not recognized by any major desktop. They now produce
8973 Up/Down key events so that they can be used for scrolling.
8975 * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
8976 slightly: previously, if an argument was specified for lines of this
8977 type (i.e. the right-most column was set) this string was appended to
8978 existing files each time systemd-tmpfiles was run. This behaviour was
8979 different from what the documentation said, and not particularly
8980 useful, as repeated systemd-tmpfiles invocations would not be
8981 idempotent and grow such files without bounds. With this release
8982 behaviour has been altered to match what the documentation says:
8983 lines of this type only have an effect if the indicated files don't
8984 exist yet, and only then the argument string is written to the file.
8986 * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
8987 systemd-tmpfiles behaviour: previously, read-only files owned by root
8988 were always excluded from the file "aging" algorithm (i.e. the
8989 automatic clean-up of directories like /tmp based on
8990 atime/mtime/ctime). We intend to drop this restriction, and age files
8991 by default even when owned by root and read-only. This behaviour was
8992 inherited from older tools, but there have been requests to remove
8993 it, and it's not obvious why this restriction was made in the first
8994 place. Please speak up now, if you are aware of software that requires
8995 this behaviour, otherwise we'll remove the restriction in v238.
8997 * A new environment variable $SYSTEMD_OFFLINE is now understood by
8998 systemctl. It takes a boolean argument. If on, systemctl assumes it
8999 operates on an "offline" OS tree, and will not attempt to talk to the
9000 service manager. Previously, this mode was implicitly enabled if a
9001 chroot() environment was detected, and this new environment variable
9002 now provides explicit control.
9004 * .path and .socket units may now be created transiently, too.
9005 Previously only service, mount, automount and timer units were
9006 supported as transient units. The systemd-run tool has been updated
9007 to expose this new functionality, you may hence use it now to bind
9008 arbitrary commands to path or socket activation on-the-fly from the
9009 command line. Moreover, almost all properties are now exposed for the
9010 unit types that already supported transient operation.
9012 * The systemd-mount command gained support for a new --owner= parameter
9013 which takes a user name, which is then resolved and included in uid=
9014 and gid= mount options string of the file system to mount.
9016 * A new unit condition ConditionControlGroupController= has been added
9017 that checks whether a specific cgroup controller is available.
9019 * Unit files, udev's .link files, and systemd-networkd's .netdev and
9020 .network files all gained support for a new condition
9021 ConditionKernelVersion= for checking against specific kernel
9024 * In systemd-networkd, the [IPVLAN] section in .netdev files gained
9025 support for configuring device flags in the Flags= setting. In the
9026 same files, the [Tunnel] section gained support for configuring
9027 AllowLocalRemote=. The [Route] section in .network files gained
9028 support for configuring InitialCongestionWindow=,
9029 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
9030 understands RapidCommit=.
9032 * systemd-networkd's DHCPv6 support gained support for Prefix
9035 * sd-bus gained support for a new "watch-bind" feature. When this
9036 feature is enabled, an sd_bus connection may be set up to connect to
9037 an AF_UNIX socket in the file system as soon as it is created. This
9038 functionality is useful for writing early-boot services that
9039 automatically connect to the system bus as soon as it is started,
9040 without ugly time-based polling. systemd-networkd and
9041 systemd-resolved have been updated to make use of this
9042 functionality. busctl exposes this functionality in a new
9043 --watch-bind= command line switch.
9045 * sd-bus will now optionally synthesize a local "Connected" signal as
9046 soon as a D-Bus connection is set up fully. This message mirrors the
9047 already existing "Disconnected" signal which is synthesized when the
9048 connection is terminated. This signal is generally useful but
9049 particularly handy in combination with the "watch-bind" feature
9050 described above. Synthesizing of this message has to be requested
9051 explicitly through the new API call sd_bus_set_connected_signal(). In
9052 addition a new call sd_bus_is_ready() has been added that checks
9053 whether a connection is fully set up (i.e. between the "Connected" and
9054 "Disconnected" signals).
9056 * sd-bus gained two new calls sd_bus_request_name_async() and
9057 sd_bus_release_name_async() for asynchronously registering bus
9058 names. Similar, there is now sd_bus_add_match_async() for installing
9059 a signal match asynchronously. All of systemd's own services have
9060 been updated to make use of these calls. Doing these operations
9061 asynchronously has two benefits: it reduces the risk of deadlocks in
9062 case of cyclic dependencies between bus services, and it speeds up
9063 service initialization since synchronization points for bus
9064 round-trips are removed.
9066 * sd-bus gained two new calls sd_bus_match_signal() and
9067 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
9068 and sd_bus_add_match_async() but instead of taking a D-Bus match
9069 string take match fields as normal function parameters.
9071 * sd-bus gained two new calls sd_bus_set_sender() and
9072 sd_bus_message_set_sender() for setting the sender name of outgoing
9073 messages (either for all outgoing messages or for just one specific
9074 one). These calls are only useful in direct connections as on
9075 brokered connections the broker fills in the sender anyway,
9076 overwriting whatever the client filled in.
9078 * sd-event gained a new pseudo-handle that may be specified on all API
9079 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
9080 used this refers to the default event loop object of the calling
9081 thread. Note however that this does not implicitly allocate one —
9082 which has to be done prior by using sd_event_default(). Similarly
9083 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
9084 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
9085 to the default bus of the specified type of the calling thread. Here
9086 too this does not implicitly allocate bus connection objects, this
9087 has to be done prior with sd_bus_default() and friends.
9089 * sd-event gained a new call pair
9090 sd_event_source_{get|set}_io_fd_own(). This may be used to request
9091 automatic closure of the file descriptor an IO event source watches
9092 when the event source is destroyed.
9094 * systemd-networkd gained support for natively configuring WireGuard
9097 * In previous versions systemd synthesized user records both for the
9098 "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
9099 internally. In order to simplify distribution-wide renames of the
9100 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
9101 new transitional flag file has been added: if
9102 /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
9103 user and group record within the systemd codebase is disabled.
9105 * systemd-notify gained a new --uid= option for selecting the source
9106 user/UID to use for notification messages sent to the service
9109 * journalctl gained a new --grep= option to list only entries in which
9110 the message matches a certain pattern. By default matching is case
9111 insensitive if the pattern is lowercase, and case sensitive
9112 otherwise. Option --case-sensitive=yes|no can be used to override
9113 this an specify case sensitivity or case insensitivity.
9115 * There's now a "systemd-analyze service-watchdogs" command for printing
9116 the current state of the service runtime watchdog, and optionally
9117 enabling or disabling the per-service watchdogs system-wide if given a
9118 boolean argument (i.e. the concept you configure in WatchdogSec=), for
9119 debugging purposes. There's also a kernel command line option
9120 systemd.service_watchdogs= for controlling the same.
9122 * Two new "log-level" and "log-target" options for systemd-analyze were
9123 added that merge the now deprecated get-log-level, set-log-level and
9124 get-log-target, set-log-target pairs. The deprecated options are still
9125 understood for backwards compatibility. The two new options print the
9126 current value when no arguments are given, and set them when a
9127 level/target is given as an argument.
9129 * sysusers.d's "u" lines now optionally accept both a UID and a GID
9130 specification, separated by a ":" character, in order to create users
9131 where UID and GID do not match.
9133 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
9134 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
9135 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
9136 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
9137 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
9138 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
9139 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
9140 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
9141 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
9142 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
9143 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
9144 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
9145 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
9146 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
9147 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
9148 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
9149 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
9150 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
9151 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
9152 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
9159 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
9160 in v235 has been extended to also set the dummy.ko module option
9161 numdummies=0, preventing the kernel from automatically creating
9162 dummy0. All dummy interfaces must now be explicitly created.
9164 * Unknown '%' specifiers in configuration files are now rejected. This
9165 applies to units and tmpfiles.d configuration. Any percent characters
9166 that are followed by a letter or digit that are not supposed to be
9167 interpreted as the beginning of a specifier should be escaped by
9168 doubling ("%%"). (So "size=5%" is still accepted, as well as
9169 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
9170 valid specifiers today.)
9172 * systemd-resolved now maintains a new dynamic
9173 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
9174 recommended to make /etc/resolv.conf a symlink to it. This file
9175 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
9176 includes dynamically acquired search domains, achieving more correct
9177 DNS resolution by software that bypasses local DNS APIs such as NSS.
9179 * The "uaccess" udev tag has been dropped from /dev/kvm and
9180 /dev/dri/renderD*. These devices now have the 0666 permissions by
9181 default (but this may be changed at build-time). /dev/dri/renderD*
9182 will now be owned by the "render" group along with /dev/kfd.
9184 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
9185 systemd-journal-gatewayd.service and
9186 systemd-journal-upload.service. This means "nss-systemd" must be
9187 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
9188 services are resolved properly.
9190 * In /etc/fstab two new mount options are now understood:
9191 x-systemd.makefs and x-systemd.growfs. The former has the effect that
9192 the configured file system is formatted before it is mounted, the
9193 latter that the file system is resized to the full block device size
9194 after it is mounted (i.e. if the file system is smaller than the
9195 partition it resides on, it's grown). This is similar to the fsck
9196 logic in /etc/fstab, and pulls in systemd-makefs@.service and
9197 systemd-growfs@.service as necessary, similar to
9198 systemd-fsck@.service. Resizing is currently only supported on ext4
9201 * In systemd-networkd, the IPv6 RA logic now optionally may announce
9202 DNS server and domain information.
9204 * Support for the LUKS2 on-disk format for encrypted partitions has
9205 been added. This requires libcryptsetup2 during compilation and
9208 * The systemd --user instance will now signal "readiness" when its
9209 basic.target unit has been reached, instead of when the run queue ran
9210 empty for the first time.
9212 * Tmpfiles.d with user configuration are now also supported.
9213 systemd-tmpfiles gained a new --user switch, and snippets placed in
9214 ~/.config/user-tmpfiles.d/ and corresponding directories will be
9215 executed by systemd-tmpfiles --user running in the new
9216 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
9217 running in the user session.
9219 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
9220 %S resolves to the top-level state directory (/var/lib for the system
9221 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
9222 top-level cache directory (/var/cache for the system instance,
9223 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
9224 logs directory (/var/log for the system instance,
9225 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
9226 existing %t specifier, that resolves to the top-level runtime
9227 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
9230 * journalctl learnt a new parameter --output-fields= for limiting the
9231 set of journal fields to output in verbose and JSON output modes.
9233 * systemd-timesyncd's configuration file gained a new option
9234 RootDistanceMaxSec= for setting the maximum root distance of servers
9235 it'll use, as well as the new options PollIntervalMinSec= and
9236 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
9238 * bootctl gained a new command "list" for listing all available boot
9239 menu items on systems that follow the boot loader specification.
9241 * systemctl gained a new --dry-run switch that shows what would be done
9242 instead of doing it, and is currently supported by the shutdown and
9245 * ConditionSecurity= can now detect the TOMOYO security module.
9247 * Unit file [Install] sections are now also respected in unit drop-in
9248 files. This is intended to be used by drop-ins under /usr/lib/.
9250 * systemd-firstboot may now also set the initial keyboard mapping.
9252 * Udev "changed" events for devices which are exposed as systemd
9253 .device units are now propagated to units specified in
9254 ReloadPropagatedFrom= as reload requests.
9256 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
9257 unit template name (i.e. a name in the form of 'foobar@.service',
9258 without the instance component between the '@' and - the '.'), then
9259 the escaped sysfs path of the device is automatically used as the
9262 * SystemCallFilter= in unit files has been extended so that an "errno"
9263 can be specified individually for each system call. Example:
9264 SystemCallFilter=~uname:EILSEQ.
9266 * The cgroup delegation logic has been substantially updated. Delegate=
9267 now optionally takes a list of controllers (instead of a boolean, as
9268 before), which lists the controllers to delegate at least.
9270 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
9272 * A new LogLevelMax= setting configures the maximum log level any
9273 process of the service may log at (i.e. anything with a lesser
9274 priority than what is specified is automatically dropped). A new
9275 LogExtraFields= setting allows configuration of additional journal
9276 fields to attach to all log records generated by any of the unit's
9279 * New StandardInputData= and StandardInputText= settings along with the
9280 new option StandardInput=data may be used to configure textual or
9281 binary data that shall be passed to the executed service process via
9282 standard input, encoded in-line in the unit file.
9284 * StandardInput=, StandardOutput= and StandardError= may now be used to
9285 connect stdin/stdout/stderr of executed processes directly with a
9286 file or AF_UNIX socket in the file system, using the new "file:" option.
9288 * A new unit file option CollectMode= has been added, that allows
9289 tweaking the garbage collection logic for units. It may be used to
9290 tell systemd to garbage collect units that have failed automatically
9291 (normally it only GCs units that exited successfully). systemd-run
9292 and systemd-mount expose this new functionality with a new -G option.
9294 * "machinectl bind" may now be used to bind mount non-directories
9295 (i.e. regularfiles, devices, fifos, sockets).
9297 * systemd-analyze gained a new verb "calendar" for validating and
9298 testing calendar time specifications to use for OnCalendar= in timer
9299 units. Besides validating the expression it will calculate the next
9300 time the specified expression would elapse.
9302 * In addition to the pre-existing FailureAction= unit file setting
9303 there's now SuccessAction=, for configuring a shutdown action to
9304 execute when a unit completes successfully. This is useful in
9305 particular inside containers that shall terminate after some workload
9306 has been completed. Also, both options are now supported for all unit
9307 types, not just services.
9309 * networkds's IP rule support gained two new options
9310 IncomingInterface= and OutgoingInterface= for configuring the incoming
9311 and outgoing interfaces of configured rules. systemd-networkd also
9312 gained support for "vxcan" network devices.
9314 * networkd gained a new setting RequiredForOnline=, taking a
9315 boolean. If set, systemd-wait-online will take it into consideration
9316 when determining that the system is up, otherwise it will ignore the
9317 interface for this purpose.
9319 * The sd_notify() protocol gained support for a new operation: with
9320 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
9321 store again, ahead of POLLHUP or POLLERR when they are removed
9324 * A new document doc/UIDS-GIDS.md has been added to the source tree,
9325 that documents the UID/GID range and assignment assumptions and
9326 requirements of systemd.
9328 * The watchdog device PID 1 will ping may now be configured through the
9329 WatchdogDevice= configuration file setting, or by setting the
9330 systemd.watchdog_service= kernel command line option.
9332 * systemd-resolved's gained support for registering DNS-SD services on
9333 the local network using MulticastDNS. Services may either be
9334 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
9335 the same dir below /run, /usr/lib), or through its D-Bus API.
9337 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
9338 extend the effective start, runtime, and stop time. The service must
9339 continue to send EXTEND_TIMEOUT_USEC within the period specified to
9340 prevent the service manager from making the service as timedout.
9342 * systemd-resolved's DNSSEC support gained support for RFC 8080
9343 (Ed25519 keys and signatures).
9345 * The systemd-resolve command line tool gained a new set of options
9346 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
9347 --set-nta= and --revert to configure per-interface DNS configuration
9348 dynamically during runtime. It's useful for pushing DNS information
9349 into systemd-resolved from DNS hook scripts that various interface
9350 managing software supports (such as pppd).
9352 * systemd-nspawn gained a new --network-namespace-path= command line
9353 option, which may be used to make a container join an existing
9354 network namespace, by specifying a path to a "netns" file.
9356 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
9357 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
9358 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
9359 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
9360 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
9361 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
9362 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
9363 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
9364 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
9365 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
9366 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
9367 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
9368 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
9369 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
9370 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
9371 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
9372 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
9373 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
9374 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
9375 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
9376 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
9377 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
9378 Jędrzejewski-Szmek, Zeal Jagannatha
9380 — Berlin, 2017-12-14
9384 * INCOMPATIBILITY: systemd-logind.service and other long-running
9385 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
9386 communication with the outside. This generally improves security of
9387 the system, and is in almost all cases a safe and good choice, as
9388 these services do not and should not provide any network-facing
9389 functionality. However, systemd-logind uses the glibc NSS API to
9390 query the user database. This creates problems on systems where NSS
9391 is set up to directly consult network services for user database
9392 lookups. In particular, this creates incompatibilities with the
9393 "nss-nis" module, which attempts to directly contact the NIS/YP
9394 network servers it is configured for, and will now consistently
9395 fail. In such cases, it is possible to turn off IP sandboxing for
9396 systemd-logind.service (set IPAddressDeny= in its [Service] section
9397 to the empty string, via a .d/ unit file drop-in). Downstream
9398 distributions might want to update their nss-nis packaging to include
9399 such a drop-in snippet, accordingly, to hide this incompatibility
9400 from the user. Another option is to make use of glibc's nscd service
9401 to proxy such network requests through a privilege-separated, minimal
9402 local caching daemon, or to switch to more modern technologies such
9403 sssd, whose NSS hook-ups generally do not involve direct network
9404 access. In general, we think it's definitely time to question the
9405 implementation choices of nss-nis, i.e. whether it's a good idea
9406 today to embed a network-facing loadable module into all local
9407 processes that need to query the user database, including the most
9408 trivial and benign ones, such as "ls". For more details about
9409 IPAddressDeny= see below.
9411 * A new modprobe.d drop-in is now shipped by default that sets the
9412 bonding module option max_bonds=0. This overrides the kernel default,
9413 to avoid conflicts and ambiguity as to whether or not bond0 should be
9414 managed by systemd-networkd or not. This resolves multiple issues
9415 with bond0 properties not being applied, when bond0 is configured
9416 with systemd-networkd. Distributors may choose to not package this,
9417 however in that case users will be prevented from correctly managing
9418 bond0 interface using systemd-networkd.
9420 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
9421 which print the logging level and target of the system manager. They
9422 complement the existing "set-log-level" and "set-log-target" verbs
9423 used to change those values.
9425 * journald.conf gained a new boolean setting ReadKMsg= which defaults
9426 to on. If turned off kernel log messages will not be read by
9427 systemd-journald or included in the logs. It also gained a new
9428 setting LineMax= for configuring the maximum line length in
9429 STDOUT/STDERR log streams. The new default for this value is 48K, up
9430 from the previous hardcoded 2048.
9432 * A new unit setting RuntimeDirectoryPreserve= has been added, which
9433 allows more detailed control of what to do with a runtime directory
9434 configured with RuntimeDirectory= (i.e. a directory below /run or
9435 $XDG_RUNTIME_DIR) after a unit is stopped.
9437 * The RuntimeDirectory= setting for units gained support for creating
9438 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
9439 one top-level directory.
9441 * Units gained new options StateDirectory=, CacheDirectory=,
9442 LogsDirectory= and ConfigurationDirectory= which are closely related
9443 to RuntimeDirectory= but manage per-service directories below
9444 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
9445 possible to write unit files which when activated automatically gain
9446 properly owned service specific directories in these locations, thus
9447 making unit files self-contained and increasing compatibility with
9448 stateless systems and factory reset where /etc or /var are
9449 unpopulated at boot. Matching these new settings there's also
9450 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
9451 ConfigurationDirectoryMode= for configuring the access mode of these
9452 directories. These settings are particularly useful in combination
9453 with DynamicUser=yes as they provide secure, properly-owned,
9454 writable, and stateful locations for storage, excluded from the
9455 sandbox that such services live in otherwise.
9457 * Automake support has been removed from this release. systemd is now
9460 * systemd-journald will now aggressively cache client metadata during
9461 runtime, speeding up log write performance under pressure. This comes
9462 at a small price though: as much of the metadata is read
9463 asynchronously from /proc/ (and isn't implicitly attached to log
9464 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
9465 metadata stored alongside a log entry might be slightly
9466 out-of-date. Previously it could only be slightly newer than the log
9467 message. The time window is small however, and given that the kernel
9468 is unlikely to be improved anytime soon in this regard, this appears
9471 * nss-myhostname/systemd-resolved will now by default synthesize an
9472 A/AAAA resource record for the "_gateway" hostname, pointing to the
9473 current default IP gateway. Previously it did that for the "gateway"
9474 name, hampering adoption, as some distributions wanted to leave that
9475 hostname open for local use. The old behaviour may still be
9476 requested at build time.
9478 * systemd-networkd's [Address] section in .network files gained a new
9479 Scope= setting for configuring the IP address scope. The [Network]
9480 section gained a new boolean setting ConfigureWithoutCarrier= that
9481 tells systemd-networkd to ignore link sensing when configuring the
9482 device. The [DHCP] section gained a new Anonymize= boolean option for
9483 turning on a number of options suggested in RFC 7844. A new
9484 [RoutingPolicyRule] section has been added for configuring the IP
9485 routing policy. The [Route] section has gained support for a new
9486 Type= setting which permits configuring
9487 blackhole/unreachable/prohibit routes.
9489 * The [VRF] section in .netdev files gained a new Table= setting for
9490 configuring the routing table to use. The [Tunnel] section gained a
9491 new Independent= boolean field for configuring tunnels independent of
9492 an underlying network interface. The [Bridge] section gained a new
9493 GroupForwardMask= option for configuration of propagation of link
9494 local frames between bridge ports.
9496 * The WakeOnLan= setting in .link files gained support for a number of
9497 new modes. A new TCP6SegmentationOffload= setting has been added for
9498 configuring TCP/IPv6 hardware segmentation offload.
9500 * The IPv6 RA sender implementation may now optionally send out RDNSS
9501 and RDNSSL records to supply DNS configuration to peers.
9503 * systemd-nspawn gained support for a new --system-call-filter= command
9504 line option for adding and removing entries in the default system
9505 call filter it applies. Moreover systemd-nspawn has been changed to
9506 implement a system call allow list instead of a deny list.
9508 * systemd-run gained support for a new --pipe command line option. If
9509 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
9510 are directly passed on to the activated transient service
9511 executable. This allows invoking arbitrary processes as systemd
9512 services (for example to take benefit of dependency management,
9513 accounting management, resource management or log management that is
9514 done automatically for services) — while still allowing them to be
9515 integrated in a classic UNIX shell pipeline.
9517 * When a service sends RELOAD=1 via sd_notify() and reload propagation
9518 using ReloadPropagationTo= is configured, a reload is now propagated
9519 to configured units. (Previously this was only done on explicitly
9520 requested reloads, using "systemctl reload" or an equivalent
9523 * For each service unit a restart counter is now kept: it is increased
9524 each time the service is restarted due to Restart=, and may be
9525 queried using "systemctl show -p NRestarts …".
9527 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
9528 @signal and @timer have been added, for usage with SystemCallFilter=
9529 in unit files and the new --system-call-filter= command line option
9530 of systemd-nspawn (see above).
9532 * ExecStart= lines in unit files gained two new modifiers: when a
9533 command line is prefixed with "!" the command will be executed as
9534 configured, except for the credentials applied by
9535 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
9536 "+", but does still apply namespacing options unlike "+". There's
9537 also "!!" now, which is mostly identical, but becomes a NOP on
9538 systems that support ambient capabilities. This is useful to write
9539 unit files that work with ambient capabilities where possible but
9540 automatically fall back to traditional privilege dropping mechanisms
9541 on systems where this is not supported.
9543 * ListenNetlink= settings in socket units now support RDMA netlink
9546 * A new unit file setting LockPersonality= has been added which permits
9547 locking down the chosen execution domain ("personality") of a service
9550 * A new special target "getty-pre.target" has been added, which is
9551 ordered before all text logins, and may be used to order services
9552 before textual logins acquire access to the console.
9554 * systemd will now attempt to load the virtio-rng.ko kernel module very
9555 early on if a VM environment supporting this is detected. This should
9556 improve entropy during early boot in virtualized environments.
9558 * A _netdev option is now supported in /etc/crypttab that operates in a
9559 similar way as the same option in /etc/fstab: it permits configuring
9560 encrypted devices that need to be ordered after the network is up.
9561 Following this logic, two new special targets
9562 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
9563 added that are to cryptsetup.target what remote-fs.target and
9564 remote-fs-pre.target are to local-fs.target.
9566 * Service units gained a new UnsetEnvironment= setting which permits
9567 unsetting specific environment variables for services that are
9568 normally passed to it (for example in order to mask out locale
9569 settings for specific services that can't deal with it).
9571 * Units acquired a new boolean option IPAccounting=. When turned on, IP
9572 traffic accounting (packet count as well as byte count) is done for
9573 the service, and shown as part of "systemctl status" or "systemd-run
9576 * Service units acquired two new options IPAddressAllow= and
9577 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
9578 for configuring a simple IP access control list for all sockets of
9579 the unit. These options are available also on .slice and .socket
9580 units, permitting flexible access list configuration for individual
9581 services as well as groups of services (as defined by a slice unit),
9582 including system-wide. Note that IP ACLs configured this way are
9583 enforced on every single IPv4 and IPv6 socket created by any process
9584 of the service unit, and apply to ingress as well as egress traffic.
9586 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
9587 structured log message is generated each time the unit is stopped,
9588 containing information about the consumed resources of this
9591 * A new setting KeyringMode= has been added to unit files, which may be
9592 used to control how the kernel keyring is set up for executed
9595 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
9596 "systemctl kexec" and "systemctl exit" are now always asynchronous in
9597 behaviour (that is: these commands return immediately after the
9598 operation was enqueued instead of waiting for the operation to
9599 complete). Previously, "systemctl poweroff" and "systemctl reboot"
9600 were asynchronous on systems using systemd-logind (i.e. almost
9601 always, and like they were on sysvinit), and the other three commands
9602 were unconditionally synchronous. With this release this is cleaned
9603 up, and callers will see the same asynchronous behaviour on all
9604 systems for all five operations.
9606 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
9609 * .timer units now accept calendar specifications in other timezones
9610 than UTC or the local timezone.
9612 * The tmpfiles snippet var.conf has been changed to create
9613 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
9614 the "utmp" group already, and it appears to be generally understood
9615 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
9616 databases. Previously this was implemented correctly for all these
9617 databases excepts btmp, which has been opened up like this now
9618 too. Note that while the other databases are world-readable
9619 (i.e. 0644), btmp is not and remains more restrictive.
9621 * The systemd-resolve tool gained a new --reset-server-features
9622 switch. When invoked like this systemd-resolved will forget
9623 everything it learnt about the features supported by the configured
9624 upstream DNS servers, and restarts the feature probing logic on the
9625 next resolver look-up for them at the highest feature level
9628 * The status dump systemd-resolved sends to the logs upon receiving
9629 SIGUSR1 now also includes information about all DNS servers it is
9630 configured to use, and the features levels it probed for them.
9632 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
9633 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
9634 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
9635 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
9636 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
9637 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
9638 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
9639 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
9640 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
9641 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
9642 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
9643 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
9644 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
9645 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
9646 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
9647 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
9648 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
9649 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
9650 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
9651 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
9653 — Berlin, 2017-10-06
9657 * Meson is now supported as build system in addition to Automake. It is
9658 our plan to remove Automake in one of our next releases, so that
9659 Meson becomes our exclusive build system. Hence, please start using
9660 the Meson build system in your downstream packaging. There's plenty
9661 of documentation around how to use Meson, the extremely brief
9664 ./autogen.sh && ./configure && make && sudo make install
9668 meson build && ninja -C build && sudo ninja -C build install
9670 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
9671 which permits configuring a timeout on the time a job is
9672 running. This is particularly useful for setting timeouts on jobs for
9675 * Unit files gained two new options ConditionUser= and ConditionGroup=
9676 for conditionalizing units based on the identity of the user/group
9677 running a systemd user instance.
9679 * systemd-networkd now understands a new FlowLabel= setting in the
9680 [VXLAN] section of .network files, as well as a Priority= in
9681 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
9682 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
9683 gained support for configuration of GENEVE links, and IPv6 address
9684 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
9686 * .link files now understand a new Port= setting.
9688 * systemd-networkd's DHCP support gained support for DHCP option 119
9689 (domain search list).
9691 * systemd-networkd gained support for serving IPv6 address ranges using
9692 the Router Advertisement protocol. The new .network configuration
9693 section [IPv6Prefix] may be used to configure the ranges to
9694 serve. This is implemented based on a new, minimal, native server
9695 implementation of RA.
9697 * journalctl's --output= switch gained support for a new parameter
9698 "short-iso-precise" for a mode where timestamps are shown as precise
9701 * systemd-udevd's "net_id" builtin may now generate stable network
9702 interface names from IBM PowerVM VIO devices as well as ACPI platform
9705 * MulticastDNS support in systemd-resolved may now be explicitly
9706 enabled/disabled using the new MulticastDNS= configuration file
9709 * systemd-resolved may now optionally use libidn2 instead of the libidn
9710 for processing internationalized domain names. Support for libidn2
9711 should be considered experimental and should not be enabled by
9714 * "machinectl pull-tar" and related call may now do verification of
9715 downloaded images using SUSE-style .sha256 checksum files in addition
9716 to the already existing support for validating using Ubuntu-style
9719 * sd-bus gained support for a new sd_bus_message_appendv() call which
9720 is va_list equivalent of sd_bus_message_append().
9722 * sd-boot gained support for validating images using SHIM/MOK.
9724 * The SMACK code learnt support for "onlycap".
9726 * systemd-mount --umount is now much smarter in figuring out how to
9727 properly unmount a device given its mount or device path.
9729 * The code to call libnss_dns as a fallback from libnss_resolve when
9730 the communication with systemd-resolved fails was removed. This
9731 fallback was redundant and interfered with the [!UNAVAIL=return]
9732 suffix. See nss-resolve(8) for the recommended configuration.
9734 * systemd-logind may now be restarted without losing state. It stores
9735 the file descriptors for devices it manages in the system manager
9736 using the FDSTORE= mechanism. Please note that further changes in
9737 other components may be required to make use of this (for example
9738 Xorg has code to listen for stops of systemd-logind and terminate
9739 itself when logind is stopped or restarted, in order to avoid using
9740 stale file descriptors for graphical devices, which is now
9741 counterproductive and must be reverted in order for restarts of
9742 systemd-logind to be safe. See
9743 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
9745 * All kernel-install plugins are called with the environment variable
9746 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
9747 /etc/machine-id. If the machine ID could not be determined,
9748 $KERNEL_INSTALL_MACHINE_ID will be empty. Plugins should not put
9749 anything in the entry directory (passed as the second argument) if
9750 $KERNEL_INSTALL_MACHINE_ID is empty. For backwards compatibility, a
9751 temporary directory is passed as the entry directory and removed
9752 after all the plugins exit.
9754 * If KERNEL_INSTALL_MACHINE_ID is set in /etc/machine-info, kernel-install
9755 will now use its value as the machine ID instead of the machine ID
9756 from /etc/machine-id. If KERNEL_INSTALL_MACHINE_ID isn't set in
9757 /etc/machine-info and no machine ID is set in /etc/machine-id,
9758 kernel-install will try to store the current machine ID there as
9759 KERNEL_INSTALL_MACHINE_ID. If there is no machine ID, kernel-install
9760 will generate a new UUID, store it in /etc/machine-info as
9761 KERNEL_INSTALL_MACHINE_ID and use it as the machine ID.
9763 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
9764 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
9765 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
9766 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
9767 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
9768 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
9769 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
9770 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
9771 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
9772 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
9773 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
9774 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
9775 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
9776 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
9777 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
9778 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
9779 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
9780 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
9781 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
9782 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
9783 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
9784 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
9785 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
9786 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
9787 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
9788 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
9789 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
9790 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
9791 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
9794 — Berlin, 2017-07-12
9798 * The "hybrid" control group mode has been modified to improve
9799 compatibility with "legacy" cgroups-v1 setups. Specifically, the
9800 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
9801 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
9802 cgroups-v1 hierarchy), the only externally visible change being that
9803 the cgroups-v2 hierarchy is also mounted, to
9804 /sys/fs/cgroup/unified. This should provide a large degree of
9805 compatibility with "legacy" cgroups-v1, while taking benefit of the
9806 better management capabilities of cgroups-v2.
9808 * The default control group setup mode may be selected both a boot-time
9809 via a set of kernel command line parameters (specifically:
9810 systemd.unified_cgroup_hierarchy= and
9811 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
9812 default selected on the configure command line
9813 (--with-default-hierarchy=). The upstream default is "hybrid"
9814 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
9815 this will change in a future systemd version to be "unified" (pure
9816 cgroups-v2 mode). The third option for the compile time option is
9817 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
9818 distributions to default to "hybrid" mode for release distributions,
9819 starting with v233. We recommend "unified" for development
9820 distributions (specifically: distributions such as Fedora's rawhide)
9821 as that's where things are headed in the long run. Use "legacy" for
9822 greatest stability and compatibility only.
9824 * Note one current limitation of "unified" and "hybrid" control group
9825 setup modes: the kernel currently does not permit the systemd --user
9826 instance (i.e. unprivileged code) to migrate processes between two
9827 disconnected cgroup subtrees, even if both are managed and owned by
9828 the user. This effectively means "systemd-run --user --scope" doesn't
9829 work when invoked from outside of any "systemd --user" service or
9830 scope. Specifically, it is not supported from session scopes. We are
9831 working on fixing this in a future systemd version. (See #3388 for
9832 further details about this.)
9834 * DBus policy files are now installed into /usr rather than /etc. Make
9835 sure your system has dbus >= 1.9.18 running before upgrading to this
9836 version, or override the install path with --with-dbuspolicydir= .
9838 * All python scripts shipped with systemd (specifically: the various
9839 tests written in Python) now require Python 3.
9841 * systemd unit tests can now run standalone (without the source or
9842 build directories), and can be installed into /usr/lib/systemd/tests/
9843 with 'make install-tests'.
9845 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
9846 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
9849 * Support for the %c, %r, %R specifiers in unit files has been
9850 removed. Specifiers are not supposed to be dependent on configuration
9851 in the unit file itself (so that they resolve the same regardless
9852 where used in the unit files), but these specifiers were influenced
9853 by the Slice= option.
9855 * The shell invoked by debug-shell.service now defaults to /bin/sh in
9856 all cases. If distributions want to use a different shell for this
9857 purpose (for example Fedora's /sbin/sushell) they need to specify
9858 this explicitly at configure time using --with-debug-shell=.
9860 * The confirmation spawn prompt has been reworked to offer the
9863 (c)ontinue, proceed without asking anymore
9864 (D)ump, show the state of the unit
9865 (f)ail, don't execute the command and pretend it failed
9867 (i)nfo, show a short summary of the unit
9868 (j)obs, show jobs that are in progress
9869 (s)kip, don't execute the command and pretend it succeeded
9870 (y)es, execute the command
9872 The 'n' choice for the confirmation spawn prompt has been removed,
9873 because its meaning was confusing.
9875 The prompt may now also be redirected to an alternative console by
9876 specifying the console as parameter to systemd.confirm_spawn=.
9878 * Services of Type=notify require a READY=1 notification to be sent
9879 during startup. If no such message is sent, the service now fails,
9880 even if the main process exited with a successful exit code.
9882 * Services that fail to start up correctly now always have their
9883 ExecStopPost= commands executed. Previously, they'd enter "failed"
9884 state directly, without executing these commands.
9886 * The option MulticastDNS= of network configuration files has acquired
9887 an actual implementation. With MulticastDNS=yes a host can resolve
9888 names of remote hosts and reply to mDNS A and AAAA requests.
9890 * When units are about to be started an additional check is now done to
9891 ensure that all dependencies of type BindsTo= (when used in
9892 combination with After=) have been started.
9894 * systemd-analyze gained a new verb "syscall-filter" which shows which
9895 system call groups are defined for the SystemCallFilter= unit file
9896 setting, and which system calls they contain.
9898 * A new system call filter group "@filesystem" has been added,
9899 consisting of various file system related system calls. Group
9900 "@reboot" has been added, covering reboot, kexec and shutdown related
9901 calls. Finally, group "@swap" has been added covering swap
9902 configuration related calls.
9904 * A new unit file option RestrictNamespaces= has been added that may be
9905 used to restrict access to the various process namespace types the
9906 Linux kernel provides. Specifically, it may be used to take away the
9907 right for a service unit to create additional file system, network,
9908 user, and other namespaces. This sandboxing option is particularly
9909 relevant due to the high amount of recently discovered namespacing
9910 related vulnerabilities in the kernel.
9912 * systemd-udev's .link files gained support for a new AutoNegotiation=
9913 setting for configuring Ethernet auto-negotiation.
9915 * systemd-networkd's .network files gained support for a new
9916 ListenPort= setting in the [DHCP] section to explicitly configure the
9917 UDP client port the DHCP client shall listen on.
9919 * .network files gained a new Unmanaged= boolean setting for explicitly
9920 excluding one or more interfaces from management by systemd-networkd.
9922 * The systemd-networkd ProxyARP= option has been renamed to
9923 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
9924 renamed to ReduceARPProxy=. The old names continue to be available
9927 * systemd-networkd gained support for configuring IPv6 Proxy NDP
9928 addresses via the new IPv6ProxyNDPAddress= .network file setting.
9930 * systemd-networkd's bonding device support gained support for two new
9931 configuration options ActiveSlave= and PrimarySlave=.
9933 * The various options in the [Match] section of .network files gained
9934 support for negative matching.
9936 * New systemd-specific mount options are now understood in /etc/fstab:
9938 x-systemd.mount-timeout= may be used to configure the maximum
9939 permitted runtime of the mount command.
9941 x-systemd.device-bound may be set to bind a mount point to its
9942 backing device unit, in order to automatically remove a mount point
9943 if its backing device is unplugged. This option may also be
9944 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
9945 on the block device, which is now automatically set for all CDROM
9946 drives, so that mounted CDs are automatically unmounted when they are
9947 removed from the drive.
9949 x-systemd.after= and x-systemd.before= may be used to explicitly
9950 order a mount after or before another unit or mount point.
9952 * Enqueued start jobs for device units are now automatically garbage
9953 collected if there are no jobs waiting for them anymore.
9955 * systemctl list-jobs gained two new switches: with --after, for every
9956 queued job the jobs it's waiting for are shown; with --before the
9957 jobs which it's blocking are shown.
9959 * systemd-nspawn gained support for ephemeral boots from disk images
9960 (or in other words: --ephemeral and --image= may now be
9961 combined). Moreover, ephemeral boots are now supported for normal
9962 directories, even if the backing file system is not btrfs. Of course,
9963 if the file system does not support file system snapshots or
9964 reflinks, the initial copy operation will be relatively expensive, but
9965 this should still be suitable for many use cases.
9967 * Calendar time specifications in .timer units now support
9968 specifications relative to the end of a month by using "~" instead of
9969 "-" as separator between month and day. For example, "*-02~03" means
9970 "the third last day in February". In addition a new syntax for
9971 repeated events has been added using the "/" character. For example,
9972 "9..17/2:00" means "every two hours from 9am to 5pm".
9974 * systemd-socket-proxyd gained a new parameter --connections-max= for
9975 configuring the maximum number of concurrent connections.
9977 * sd-id128 gained a new API for generating unique IDs for the host in a
9978 way that does not leak the machine ID. Specifically,
9979 sd_id128_get_machine_app_specific() derives an ID based on the
9980 machine ID in a well-defined, non-reversible, stable way. This is
9981 useful whenever an identifier for the host is needed but where the
9982 identifier shall not be useful to identify the system beyond the
9983 scope of the application itself. (Internally this uses HMAC-SHA256 as
9984 keyed hash function using the machine ID as input.)
9986 * NotifyAccess= gained a new supported value "exec". When set
9987 notifications are accepted from all processes systemd itself invoked,
9988 including all control processes.
9990 * .nspawn files gained support for defining overlay mounts using the
9991 Overlay= and OverlayReadOnly= options. Previously this functionality
9992 was only available on the systemd-nspawn command line.
9994 * systemd-nspawn's --bind= and --overlay= options gained support for
9995 bind/overlay mounts whose source lies within the container tree by
9996 prefixing the source path with "+".
9998 * systemd-nspawn's --bind= and --overlay= options gained support for
9999 automatically allocating a temporary source directory in /var/tmp
10000 that is removed when the container dies. Specifically, if the source
10001 directory is specified as empty string this mechanism is selected. An
10002 example usage is --overlay=+/var::/var, which creates an overlay
10003 mount based on the original /var contained in the image, overlaid
10004 with a temporary directory in the host's /var/tmp. This way changes
10005 to /var are automatically flushed when the container shuts down.
10007 * systemd-nspawn --image= option does now permit raw file system block
10008 devices (in addition to images containing partition tables, as
10011 * The disk image dissection logic in systemd-nspawn gained support for
10012 automatically setting up LUKS encrypted as well as Verity protected
10013 partitions. When a container is booted from an encrypted image the
10014 passphrase is queried at start-up time. When a container with Verity
10015 data is started, the root hash is search in a ".roothash" file
10016 accompanying the disk image (alternatively, pass the root hash via
10017 the new --root-hash= command line option).
10019 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
10020 be used to dissect disk images the same way as systemd-nspawn does
10021 it, following the Bootable Partition Specification. It may even be
10022 used to mount disk images with complex partition setups (including
10023 LUKS and Verity partitions) to a local host directory, in order to
10024 inspect them. This tool is not considered public API (yet), and is
10025 thus not installed into /usr/bin. Please do not rely on its
10026 existence, since it might go away or be changed in later systemd
10029 * A new generator "systemd-verity-generator" has been added, similar in
10030 style to "systemd-cryptsetup-generator", permitting automatic setup of
10031 Verity root partitions when systemd boots up. In order to make use of
10032 this your partition setup should follow the Discoverable Partitions
10033 Specification, and the GPT partition ID of the root file system
10034 partition should be identical to the upper 128-bit of the Verity root
10035 hash. The GPT partition ID of the Verity partition protecting it
10036 should be the lower 128-bit of the Verity root hash. If the partition
10037 image follows this model it is sufficient to specify a single
10038 "roothash=" kernel command line argument to both configure which root
10039 image and verity partition to use as well as the root hash for
10040 it. Note that systemd-nspawn's Verity support follows the same
10041 semantics, meaning that disk images with proper Verity data in place
10042 may be booted in containers with systemd-nspawn as well as on
10043 physical systems via the verity generator. Also note that the "mkosi"
10044 tool available at https://github.com/systemd/mkosi has been updated
10045 to generate Verity protected disk images following this scheme. In
10046 fact, it has been updated to generate disk images that optionally
10047 implement a complete UEFI SecureBoot trust chain, involving a signed
10048 kernel and initrd image that incorporates such a root hash as well as
10049 a Verity-enabled root partition.
10051 * The hardware database (hwdb) udev supports has been updated to carry
10052 accelerometer quirks.
10054 * All system services are now run with a fresh kernel keyring set up
10055 for them. The invocation ID is stored by default in it, thus
10056 providing a safe, non-overridable way to determine the invocation
10057 ID of each service.
10059 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
10060 options for bind mounting arbitrary paths in a service-specific
10061 way. When these options are used, arbitrary host or service files and
10062 directories may be mounted to arbitrary locations in the service's
10065 * Documentation has been added that lists all of systemd's low-level
10066 environment variables:
10068 https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md
10070 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
10071 whether a specific socket file descriptor matches a specified socket
10074 * systemd-firstboot has been updated to check for the
10075 systemd.firstboot= kernel command line option. It accepts a boolean
10076 and when set to false the first boot questions are skipped.
10078 * systemd-fstab-generator has been updated to check for the
10079 systemd.volatile= kernel command line option, which either takes an
10080 optional boolean parameter or the special value "state". If used the
10081 system may be booted in a "volatile" boot mode. Specifically,
10082 "systemd.volatile" is used, the root directory will be mounted as
10083 tmpfs, and only /usr is mounted from the actual root file system. If
10084 "systemd.volatile=state" is used, the root directory will be mounted
10085 as usual, but /var is mounted as tmpfs. This concept provides similar
10086 functionality as systemd-nspawn's --volatile= option, but provides it
10087 on physical boots. Use this option for implementing stateless
10088 systems, or testing systems with all state and/or configuration reset
10089 to the defaults. (Note though that many distributions are not
10090 prepared to boot up without a populated /etc or /var, though.)
10092 * systemd-gpt-auto-generator gained support for LUKS encrypted root
10093 partitions. Previously it only supported LUKS encrypted partitions
10094 for all other uses, except for the root partition itself.
10096 * Socket units gained support for listening on AF_VSOCK sockets for
10097 communication in virtualized QEMU environments.
10099 * The "configure" script gained a new option --with-fallback-hostname=
10100 for specifying the fallback hostname to use if none is configured in
10101 /etc/hostname. For example, by specifying
10102 --with-fallback-hostname=fedora it is possible to default to a
10103 hostname of "fedora" on pristine installations.
10105 * systemd-cgls gained support for a new --unit= switch for listing only
10106 the control groups of a specific unit. Similar --user-unit= has been
10107 added for listing only the control groups of a specific user unit.
10109 * systemd-mount gained a new --umount switch for unmounting a mount or
10110 automount point (and all mount/automount points below it).
10112 * systemd will now refuse full configuration reloads (via systemctl
10113 daemon-reload and related calls) unless at least 16MiB of free space
10114 are available in /run. This is a safety precaution in order to ensure
10115 that generators can safely operate after the reload completed.
10117 * A new unit file option RootImage= has been added, which has a similar
10118 effect as RootDirectory= but mounts the service's root directory from
10119 a disk image instead of plain directory. This logic reuses the same
10120 image dissection and mount logic that systemd-nspawn already uses,
10121 and hence supports any disk images systemd-nspawn supports, including
10122 those following the Discoverable Partition Specification, as well as
10123 Verity enabled images. This option enables systemd to run system
10124 services directly off disk images acting as resource bundles,
10125 possibly even including full integrity data.
10127 * A new MountAPIVFS= unit file option has been added, taking a boolean
10128 argument. If enabled /proc, /sys and /dev (collectively called the
10129 "API VFS") will be mounted for the service. This is only relevant if
10130 RootDirectory= or RootImage= is used for the service, as these mounts
10131 are of course in place in the host mount namespace anyway.
10133 * systemd-nspawn gained support for a new --pivot-root= switch. If
10134 specified the root directory within the container image is pivoted to
10135 the specified mount point, while the original root disk is moved to a
10136 different place. This option enables booting of ostree images
10137 directly with systemd-nspawn.
10139 * The systemd build scripts will no longer complain if the NTP server
10140 addresses are not changed from the defaults. Google now supports
10141 these NTP servers officially. We still recommend downstreams to
10142 properly register an NTP pool with the NTP pool project though.
10144 * coredumpctl gained a new "--reverse" option for printing the list
10145 of coredumps in reverse order.
10147 * coredumpctl will now show additional information about truncated and
10148 inaccessible coredumps, as well as coredumps that are still being
10149 processed. It also gained a new --quiet switch for suppressing
10150 additional informational message in its output.
10152 * coredumpctl gained support for only showing coredumps newer and/or
10153 older than specific timestamps, using the new --since= and --until=
10154 options, reminiscent of journalctl's options by the same name.
10156 * The systemd-coredump logic has been improved so that it may be reused
10157 to collect backtraces in non-compiled languages, for example in
10158 scripting languages such as Python.
10160 * machinectl will now show the UID shift of local containers, if user
10161 namespacing is enabled for them.
10163 * systemd will now optionally run "environment generator" binaries at
10164 configuration load time. They may be used to add environment
10165 variables to the environment block passed to services invoked. One
10166 user environment generator is shipped by default that sets up
10167 environment variables based on files dropped into /etc/environment.d
10168 and ~/.config/environment.d/.
10170 * systemd-resolved now includes the new, recently published 2017 DNSSEC
10173 * hostnamed has been updated to report a new chassis type of
10174 "convertible" to cover "foldable" laptops that can both act as a
10175 tablet and as a laptop, such as various Lenovo Yoga devices.
10177 Contributions from: Adrián López, Alexander Galanin, Alexander
10178 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
10179 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
10180 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
10181 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
10182 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
10183 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
10184 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
10185 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
10186 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
10187 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
10188 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
10189 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
10190 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
10191 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
10192 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
10193 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
10194 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
10195 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
10196 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
10197 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
10198 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
10199 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
10200 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
10201 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
10202 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
10203 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
10204 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
10207 — Berlin, 2017-03-01
10211 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
10212 RestrictAddressFamilies= enabled. These sandboxing options should
10213 generally be compatible with the various external udev call-out
10214 binaries we are aware of, however there may be exceptions, in
10215 particular when exotic languages for these call-outs are used. In
10216 this case, consider turning off these settings locally.
10218 * The new RemoveIPC= option can be used to remove IPC objects owned by
10219 the user or group of a service when that service exits.
10221 * The new ProtectKernelModules= option can be used to disable explicit
10222 load and unload operations of kernel modules by a service. In
10223 addition access to /usr/lib/modules is removed if this option is set.
10225 * ProtectSystem= option gained a new value "strict", which causes the
10226 whole file system tree with the exception of /dev, /proc, and /sys,
10227 to be remounted read-only for a service.
10229 * The new ProtectKernelTunables= option can be used to disable
10230 modification of configuration files in /sys and /proc by a service.
10231 Various directories and files are remounted read-only, so access is
10232 restricted even if the file permissions would allow it.
10234 * The new ProtectControlGroups= option can be used to disable write
10235 access by a service to /sys/fs/cgroup.
10237 * Various systemd services have been hardened with
10238 ProtectKernelTunables=yes, ProtectControlGroups=yes,
10239 RestrictAddressFamilies=.
10241 * Support for dynamically creating users for the lifetime of a service
10242 has been added. If DynamicUser=yes is specified, user and group IDs
10243 will be allocated from the range 61184…65519 for the lifetime of the
10244 service. They can be resolved using the new nss-systemd.so NSS
10245 module. The module must be enabled in /etc/nsswitch.conf. Services
10246 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
10247 any resources allocated by the service will be cleaned up when the
10248 service exits. They also have ProtectHome=read-only and
10249 ProtectSystem=strict enabled, so they are not able to make any
10250 permanent modifications to the system.
10252 * The nss-systemd module also always resolves root and nobody, making
10253 it possible to have no /etc/passwd or /etc/group files in minimal
10254 container or chroot environments.
10256 * Services may be started with their own user namespace using the new
10257 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
10258 under which the service is running are mapped. All other users are
10261 * Support for the cgroup namespace has been added to systemd-nspawn. If
10262 supported by kernel, the container system started by systemd-nspawn
10263 will have its own view of the cgroup hierarchy. This new behaviour
10264 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
10266 * The new MemorySwapMax= option can be used to limit the maximum swap
10267 usage under the unified cgroup hierarchy.
10269 * Support for the CPU controller in the unified cgroup hierarchy has
10270 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
10271 options. This controller requires out-of-tree patches for the kernel
10272 and the support is provisional.
10274 * Mount and automount units may now be created transiently
10275 (i.e. dynamically at runtime via the bus API, instead of requiring
10276 unit files in the file system).
10278 * systemd-mount is a new tool which may mount file systems – much like
10279 mount(8), optionally pulling in additional dependencies through
10280 transient .mount and .automount units. For example, this tool
10281 automatically runs fsck on a backing block device before mounting,
10282 and allows the automount logic to be used dynamically from the
10283 command line for establishing mount points. This tool is particularly
10284 useful when dealing with removable media, as it will ensure fsck is
10285 run – if necessary – before the first access and that the file system
10286 is quickly unmounted after each access by utilizing the automount
10287 logic. This maximizes the chance that the file system on the
10288 removable media stays in a clean state, and if it isn't in a clean
10289 state is fixed automatically.
10291 * LazyUnmount=yes option for mount units has been added to expose the
10292 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
10295 * /efi will be used as the mount point of the EFI boot partition, if
10296 the directory is present, and the mount point was not configured
10297 through other means (e.g. fstab). If /efi directory does not exist,
10298 /boot will be used as before. This makes it easier to automatically
10299 mount the EFI partition on systems where /boot is used for something
10302 * When operating on GPT disk images for containers, systemd-nspawn will
10303 now mount the ESP to /boot or /efi according to the same rules as PID
10304 1 running on a host. This allows tools like "bootctl" to operate
10305 correctly within such containers, in order to make container images
10306 bootable on physical systems.
10308 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
10310 * Two new user session targets have been added to support running
10311 graphical sessions under the systemd --user instance:
10312 graphical-session.target and graphical-session-pre.target. See
10313 systemd.special(7) for a description of how those targets should be
10316 * The vconsole initialization code has been significantly reworked to
10317 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
10318 support unicode keymaps. Font and keymap configuration will now be
10319 copied to all allocated virtual consoles.
10321 * FreeBSD's bhyve virtualization is now detected.
10323 * Information recorded in the journal for core dumps now includes the
10324 contents of /proc/mountinfo and the command line of the process at
10325 the top of the process hierarchy (which is usually the init process
10328 * systemd-journal-gatewayd learned the --directory= option to serve
10329 files from the specified location.
10331 * journalctl --root=… can be used to peruse the journal in the
10332 /var/log/ directories inside of a container tree. This is similar to
10333 the existing --machine= option, but does not require the container to
10336 * The hardware database has been extended to support
10337 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
10340 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
10341 specify the click rate for mice which include a horizontal wheel with
10342 a click rate that is different than the one for the vertical wheel.
10344 * systemd-run gained a new --wait option that makes service execution
10345 synchronous. (Specifically, the command will not return until the
10346 specified service binary exited.)
10348 * systemctl gained a new --wait option that causes the start command to
10349 wait until the units being started have terminated again.
10351 * A new journal output mode "short-full" has been added which displays
10352 timestamps with abbreviated English day names and adds a timezone
10353 suffix. Those timestamps include more information than the default
10354 "short" output mode, and can be passed directly to journalctl's
10355 --since= and --until= options.
10357 * /etc/resolv.conf will be bind-mounted into containers started by
10358 systemd-nspawn, if possible, so any changes to resolv.conf contents
10359 are automatically propagated to the container.
10361 * The number of instances for socket-activated services originating
10362 from a single IP address can be limited with
10363 MaxConnectionsPerSource=, extending the existing setting of
10366 * systemd-networkd gained support for vcan ("Virtual CAN") interface
10369 * .netdev and .network configuration can now be extended through
10372 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
10373 Segmentation Offload, Generic Receive Offload, Large Receive Offload
10374 can be enabled and disabled using the new UDPSegmentationOffload=,
10375 TCPSegmentationOffload=, GenericSegmentationOffload=,
10376 GenericReceiveOffload=, LargeReceiveOffload= options in the
10377 [Link] section of .link files.
10379 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
10380 Port VLAN ID can be configured for bridge devices using the new STP=,
10381 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
10382 section of .netdev files.
10384 * The route table to which routes received over DHCP or RA should be
10385 added can be configured with the new RouteTable= option in the [DHCP]
10386 and [IPv6AcceptRA] sections of .network files.
10388 * The Address Resolution Protocol can be disabled on links managed by
10389 systemd-networkd using the ARP=no setting in the [Link] section of
10392 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
10393 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
10394 encode information about the result and exit codes of the current
10395 service runtime cycle.
10397 * systemd-sysctl will now configure kernel parameters in the order
10398 they occur in the configuration files. This matches what sysctl
10399 has been traditionally doing.
10401 * kernel-install "plugins" that are executed to perform various
10402 tasks after a new kernel is added and before an old one is removed
10403 can now return a special value to terminate the procedure and
10404 prevent any later plugins from running.
10406 * Journald's SplitMode=login setting has been deprecated. It has been
10407 removed from documentation, and its use is discouraged. In a future
10408 release it will be completely removed, and made equivalent to current
10409 default of SplitMode=uid.
10411 * Storage=both option setting in /etc/systemd/coredump.conf has been
10412 removed. With fast LZ4 compression storing the core dump twice is not
10415 * The --share-system systemd-nspawn option has been replaced with an
10416 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
10417 this functionality is discouraged. In addition the variables
10418 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
10419 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
10420 individual namespaces.
10422 * "machinectl list" now shows the IP address of running containers in
10423 the output, as well as OS release information.
10425 * "loginctl list" now shows the TTY of each session in the output.
10427 * sd-bus gained new API calls sd_bus_track_set_recursive(),
10428 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
10429 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
10430 tracking objects in a "recursive" mode, where a single client can be
10431 counted multiple times, if it takes multiple references.
10433 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
10434 sd_bus_get_exit_on_disconnect(). They may be used to make a
10435 process using sd-bus automatically exit if the bus connection is
10438 * Bus clients of the service manager may now "pin" loaded units into
10439 memory, by taking an explicit reference on them. This is useful to
10440 ensure the client can retrieve runtime data about the service even
10441 after the service completed execution. Taking such a reference is
10442 available only for privileged clients and should be helpful to watch
10443 running services in a race-free manner, and in particular collect
10444 information about exit statuses and results.
10446 * The nss-resolve module has been changed to strictly return UNAVAIL
10447 when communication via D-Bus with resolved failed, and NOTFOUND when
10448 a lookup completed but was negative. This means it is now possible to
10449 neatly configure fallbacks using nsswitch.conf result checking
10450 expressions. Taking benefit of this, the new recommended
10451 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
10453 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
10455 * A new setting CtrlAltDelBurstAction= has been added to
10456 /etc/systemd/system.conf which may be used to configure the precise
10457 behaviour if the user on the console presses Ctrl-Alt-Del more often
10458 than 7 times in 2s. Previously this would unconditionally result in
10459 an expedited, immediate reboot. With this new setting the precise
10460 operation may be configured in more detail, and also turned off
10463 * In .netdev files two new settings RemoteChecksumTx= and
10464 RemoteChecksumRx= are now understood that permit configuring the
10465 remote checksumming logic for VXLAN networks.
10467 * The service manager learnt a new "invocation ID" concept for invoked
10468 services. Each runtime cycle of a service will get a new invocation
10469 ID (a 128-bit random UUID) assigned that identifies the current
10470 run of the service uniquely and globally. A new invocation ID
10471 is generated each time a service starts up. The journal will store
10472 the invocation ID of a service along with any logged messages, thus
10473 making the invocation ID useful for matching the online runtime of a
10474 service with the offline log data it generated in a safe way without
10475 relying on synchronized timestamps. In many ways this new service
10476 invocation ID concept is similar to the kernel's boot ID concept that
10477 uniquely and globally identifies the runtime of each boot. The
10478 invocation ID of a service is passed to the service itself via an
10479 environment variable ($INVOCATION_ID). A new bus call
10480 GetUnitByInvocationID() has been added that is similar to GetUnit()
10481 but instead of retrieving the bus path for a unit by its name
10482 retrieves it by its invocation ID. The returned path is valid only as
10483 long as the passed invocation ID is current.
10485 * systemd-resolved gained a new "DNSStubListener" setting in
10486 resolved.conf. It either takes a boolean value or the special values
10487 "udp" and "tcp", and configures whether to enable the stub DNS
10488 listener on 127.0.0.53:53.
10490 * IP addresses configured via networkd may now carry additional
10491 configuration settings supported by the kernel. New options include:
10492 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
10493 PrefixRoute=, AutoJoin=.
10495 * The PAM configuration fragment file for "user@.service" shipped with
10496 systemd (i.e. the --user instance of systemd) has been stripped to
10497 the minimum necessary to make the system boot. Previously, it
10498 contained Fedora-specific stanzas that did not apply to other
10499 distributions. It is expected that downstream distributions add
10500 additional configuration lines, matching their needs to this file,
10501 using it only as rough template of what systemd itself needs. Note
10502 that this reduced fragment does not even include an invocation of
10503 pam_limits which most distributions probably want to add, even though
10504 systemd itself does not need it. (There's also the new build time
10505 option --with-pamconfdir=no to disable installation of the PAM
10506 fragment entirely.)
10508 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
10509 capability is now also dropped from its set (in addition to
10510 CAP_SYS_MKNOD as before).
10512 * In service unit files it is now possible to connect a specific named
10513 file descriptor with stdin/stdout/stdout of an executed service. The
10514 name may be specified in matching .socket units using the
10515 FileDescriptorName= setting.
10517 * A number of journal settings may now be configured on the kernel
10518 command line. Specifically, the following options are now understood:
10519 systemd.journald.max_level_console=,
10520 systemd.journald.max_level_store=,
10521 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
10522 systemd.journald.max_level_wall=.
10524 * "systemctl is-enabled --full" will now show by which symlinks a unit
10525 file is enabled in the unit dependency tree.
10527 * Support for VeraCrypt encrypted partitions has been added to the
10528 "cryptsetup" logic and /etc/crypttab.
10530 * systemd-detect-virt gained support for a new --private-users switch
10531 that checks whether the invoking processes are running inside a user
10532 namespace. Similar, a new special value "private-users" for the
10533 existing ConditionVirtualization= setting has been added, permitting
10534 skipping of specific units in user namespace environments.
10536 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
10537 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
10538 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
10539 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
10540 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
10541 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
10542 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
10543 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
10544 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
10545 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
10546 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
10547 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
10548 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
10549 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
10550 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
10551 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
10552 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
10553 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
10554 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
10555 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
10556 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
10557 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
10558 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
10559 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
10560 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
10561 Jędrzejewski-Szmek, Zeal Jagannatha
10563 — Santa Fe, 2016-11-03
10567 * In service units the various ExecXYZ= settings have been extended
10568 with an additional special character as first argument of the
10569 assigned value: if the character '+' is used the specified command
10570 line it will be run with full privileges, regardless of User=,
10571 Group=, CapabilityBoundingSet= and similar options. The effect is
10572 similar to the existing PermissionsStartOnly= option, but allows
10573 configuration of this concept for each executed command line
10576 * Services may now alter the service watchdog timeout at runtime by
10577 sending a WATCHDOG_USEC= message via sd_notify().
10579 * MemoryLimit= and related unit settings now optionally take percentage
10580 specifications. The percentage is taken relative to the amount of
10581 physical memory in the system (or in case of containers, the assigned
10582 amount of memory). This allows scaling service resources neatly with
10583 the amount of RAM available on the system. Similarly, systemd-logind's
10584 RuntimeDirectorySize= option now also optionally takes percentage
10587 * In similar fashion TasksMax= takes percentage values now, too. The
10588 value is taken relative to the configured maximum number of processes
10589 on the system. The per-service task maximum has been changed to 15%
10590 using this functionality. (Effectively this is an increase of 512 →
10591 4915 for service units, given the kernel's default pid_max setting.)
10593 * Calendar time specifications in .timer units now understand a ".."
10594 syntax for time ranges. Example: "4..7:10" may now be used for
10595 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
10598 * The InaccessableDirectories=, ReadOnlyDirectories= and
10599 ReadWriteDirectories= unit file settings have been renamed to
10600 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
10601 applied to all kinds of file nodes, and not just directories, with
10602 the exception of symlinks. Specifically these settings may now be
10603 used on block and character device nodes, UNIX sockets and FIFOS as
10604 well as regular files. The old names of these settings remain
10605 available for compatibility.
10607 * systemd will now log about all service processes it kills forcibly
10608 (using SIGKILL) because they remained after the clean shutdown phase
10609 of the service completed. This should help identifying services that
10610 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
10611 systemd-logind's configuration a similar log message is generated for
10612 processes killed at the end of each session due to this setting.
10614 * systemd will now set the $JOURNAL_STREAM environment variable for all
10615 services whose stdout/stderr are connected to the Journal (which
10616 effectively means by default: all services). The variable contains
10617 the device and inode number of the file descriptor used for
10618 stdout/stderr. This may be used by invoked programs to detect whether
10619 their stdout/stderr is connected to the Journal, in which case they
10620 can switch over to direct Journal communication, thus being able to
10621 pass extended, structured metadata along with their log messages. As
10622 one example, this is now used by glib's logging primitives.
10624 * When using systemd's default tmp.mount unit for /tmp, the mount point
10625 will now be established with the "nosuid" and "nodev" options. This
10626 avoids privilege escalation attacks that put traps and exploits into
10627 /tmp. However, this might cause problems if you e.g. put container
10628 images or overlays into /tmp; if you need this, override tmp.mount's
10629 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
10632 * systemd now supports the "memory" cgroup controller also on
10635 * The systemd-cgtop tool now optionally takes a control group path as
10636 command line argument. If specified, the control group list shown is
10637 limited to subgroups of that group.
10639 * The SystemCallFilter= unit file setting gained support for
10640 pre-defined, named system call filter sets. For example
10641 SystemCallFilter=@clock is now an effective way to make all clock
10642 changing-related system calls unavailable to a service. A number of
10643 similar pre-defined groups are defined. Writing system call filters
10644 for system services is simplified substantially with this new
10645 concept. Accordingly, all of systemd's own, long-running services now
10646 enable system call filtering based on this, by default.
10648 * A new service setting MemoryDenyWriteExecute= has been added, taking
10649 a boolean value. If turned on, a service may no longer create memory
10650 mappings that are writable and executable at the same time. This
10651 enhances security for services where this is enabled as it becomes
10652 harder to dynamically write and then execute memory in exploited
10653 service processes. This option has been enabled for all of systemd's
10654 own long-running services.
10656 * A new RestrictRealtime= service setting has been added, taking a
10657 boolean argument. If set the service's processes may no longer
10658 acquire realtime scheduling. This improves security as realtime
10659 scheduling may otherwise be used to easily freeze the system.
10661 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
10662 value. This may be used for requesting that the system manager inside
10663 of the container reports start-up completion to nspawn which then
10664 propagates this notification further to the service manager
10665 supervising nspawn itself. A related option NotifyReady= in .nspawn
10666 files has been added too. This functionality allows ordering of the
10667 start-up of multiple containers using the usual systemd ordering
10670 * machinectl gained a new command "stop" that is an alias for
10673 * systemd-resolved gained support for contacting DNS servers on
10674 link-local IPv6 addresses.
10676 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
10677 its caches. A method call for requesting the same operation has been
10678 added to the bus API too, and is made available via "systemd-resolve
10681 * systemd-resolve gained a new --status switch. If passed a brief
10682 summary of the used DNS configuration with per-interface information
10685 * resolved.conf gained a new Cache= boolean option, defaulting to
10686 on. If turned off local DNS caching is disabled. This comes with a
10687 performance penalty in particular when DNSSEC is enabled. Note that
10688 resolved disables its internal caching implicitly anyway, when the
10689 configured DNS server is on a host-local IP address such as ::1 or
10690 127.0.0.1, thus automatically avoiding double local caching.
10692 * systemd-resolved now listens on the local IP address 127.0.0.53:53
10693 for DNS requests. This improves compatibility with local programs
10694 that do not use the libc NSS or systemd-resolved's bus APIs for name
10695 resolution. This minimal DNS service is only available to local
10696 programs and does not implement the full DNS protocol, but enough to
10697 cover local DNS clients. A new, static resolv.conf file, listing just
10698 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
10699 now recommended to make /etc/resolv.conf a symlink to this file in
10700 order to route all DNS lookups to systemd-resolved, regardless if
10701 done via NSS, the bus API or raw DNS packets. Note that this local
10702 DNS service is not as fully featured as the libc NSS or
10703 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
10704 used to deliver link-local address information (as this implies
10705 sending a local interface index along), LLMNR/mDNS support via this
10706 interface is severely restricted. It is thus strongly recommended for
10707 all applications to use the libc NSS API or native systemd-resolved
10710 * systemd-networkd's bridge support learned a new setting
10711 VLANFiltering= for controlling VLAN filtering. Moreover a new section
10712 in .network files has been added for configuring VLAN bridging in
10713 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
10715 * systemd-networkd's IPv6 Router Advertisement code now makes use of
10716 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
10717 now be acquired without relying on DHCPv6. Two new options
10718 UseDomains= and UseDNS= have been added to configure this behaviour.
10720 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
10721 renamed IPv6AcceptRA=, without altering its behaviour. The old
10722 setting name remains available for compatibility reasons.
10724 * The systemd-networkd VTI/VTI6 tunneling support gained new options
10725 Key=, InputKey= and OutputKey=.
10727 * systemd-networkd gained support for VRF ("Virtual Routing Function")
10728 interface configuration.
10730 * "systemctl edit" may now be used to create new unit files by
10731 specifying the --force switch.
10733 * sd-event gained a new function sd_event_get_iteration() for
10734 requesting the current iteration counter of the event loop. It starts
10735 at zero and is increased by one with each event loop iteration.
10737 * A new rpm macro %systemd_ordering is provided by the macros.systemd
10738 file. It can be used in lieu of %systemd_requires in packages which
10739 don't use any systemd functionality and are intended to be installed
10740 in minimal containers without systemd present. This macro provides
10741 ordering dependencies to ensure that if the package is installed in
10742 the same rpm transaction as systemd, systemd will be installed before
10743 the scriptlets for the package are executed, allowing unit presets
10746 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
10747 been added to simplify packaging of generators.
10749 * The os-release file gained VERSION_CODENAME field for the
10750 distribution nickname (e.g. VERSION_CODENAME=woody).
10752 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
10753 can be set to disable parsing of metadata and the creation
10754 of persistent symlinks for that device.
10756 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
10757 to make them available to logged-in users has been reverted.
10759 * Much of the common code of the various systemd components is now
10760 built into an internal shared library libsystemd-shared-231.so
10761 (incorporating the systemd version number in the name, to be updated
10762 with future releases) that the components link to. This should
10763 decrease systemd footprint both in memory during runtime and on
10764 disk. Note that the shared library is not for public use, and is
10765 neither API nor ABI stable, but is likely to change with every new
10766 released update. Packagers need to make sure that binaries
10767 linking to libsystemd-shared.so are updated in step with the
10770 * Configuration for "mkosi" is now part of the systemd
10771 repository. mkosi is a tool to easily build legacy-free OS images,
10772 and is available on github: https://github.com/systemd/mkosi. If
10773 "mkosi" is invoked in the build tree a new raw OS image is generated
10774 incorporating the systemd sources currently being worked on and a
10775 clean, fresh distribution installation. The generated OS image may be
10776 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
10777 UEFI PC. This functionality is particularly useful to easily test
10778 local changes made to systemd in a pristine, defined environment. See
10779 doc/HACKING for details.
10781 * configure learned the --with-support-url= option to specify the
10782 distribution's bugtracker.
10784 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
10785 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
10786 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
10787 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
10788 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
10789 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
10790 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
10791 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
10792 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
10793 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
10794 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
10795 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
10796 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
10797 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
10798 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
10799 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
10800 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
10801 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
10802 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
10804 — Berlin, 2016-07-25
10808 * DNSSEC is now turned on by default in systemd-resolved (in
10809 "allow-downgrade" mode), but may be turned off during compile time by
10810 passing "--with-default-dnssec=no" to "configure" (and of course,
10811 during runtime with DNSSEC= in resolved.conf). We recommend
10812 downstreams to leave this on at least during development cycles and
10813 report any issues with the DNSSEC logic upstream. We are very
10814 interested in collecting feedback about the DNSSEC validator and its
10815 limitations in the wild. Note however, that DNSSEC support is
10816 probably nothing downstreams should turn on in stable distros just
10817 yet, as it might create incompatibilities with a few DNS servers and
10818 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
10819 automatically whenever we detect such incompatible setups, but there
10820 might be systems we do not cover yet. Hence: please help us testing
10821 the DNSSEC code, leave this on where you can, report back, but then
10822 again don't consider turning this on in your stable, LTS or
10823 production release just yet. (Note that you have to enable
10824 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
10825 and its DNSSEC mode for hostname resolution from local
10828 * systemd-resolve conveniently resolves DANE records with the --tlsa
10829 option and OPENPGPKEY records with the --openpgp option. It also
10830 supports dumping raw DNS record data via the new --raw= switch.
10832 * systemd-logind will now by default terminate user processes that are
10833 part of the user session scope unit (session-XX.scope) when the user
10834 logs out. This behavior is controlled by the KillUserProcesses=
10835 setting in logind.conf, and the previous default of "no" is now
10836 changed to "yes". This means that user sessions will be properly
10837 cleaned up after, but additional steps are necessary to allow
10838 intentionally long-running processes to survive logout.
10840 While the user is logged in at least once, user@.service is running,
10841 and any service that should survive the end of any individual login
10842 session can be started at a user service or scope using systemd-run.
10843 systemd-run(1) man page has been extended with an example which shows
10844 how to run screen in a scope unit underneath user@.service. The same
10845 command works for tmux.
10847 After the user logs out of all sessions, user@.service will be
10848 terminated too, by default, unless the user has "lingering" enabled.
10849 To effectively allow users to run long-term tasks even if they are
10850 logged out, lingering must be enabled for them. See loginctl(1) for
10851 details. The default polkit policy was modified to allow users to
10852 set lingering for themselves without authentication.
10854 Previous defaults can be restored at compile time by the
10855 --without-kill-user-processes option to "configure".
10857 * systemd-logind gained new configuration settings SessionsMax= and
10858 InhibitorsMax=, both with a default of 8192. It will not register new
10859 user sessions or inhibitors above this limit.
10861 * systemd-logind will now reload configuration on SIGHUP.
10863 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
10864 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
10865 enable. Also, support for the "io" cgroup controller in the unified
10866 hierarchy has been added, so that the "memory", "pids" and "io" are
10867 now the controllers that are supported on the unified hierarchy.
10869 WARNING: it is not possible to use previous systemd versions with
10870 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
10871 is necessary to also update systemd in the initramfs if using the
10872 unified hierarchy. An updated SELinux policy is also required.
10874 * LLDP support has been extended, and both passive (receive-only) and
10875 active (sender) modes are supported. Passive mode ("routers-only") is
10876 enabled by default in systemd-networkd. Active LLDP mode is enabled
10877 by default for containers on the internal network. The "networkctl
10878 lldp" command may be used to list information gathered. "networkctl
10879 status" will also show basic LLDP information on connected peers now.
10881 * The IAID and DUID unique identifier sent in DHCP requests may now be
10882 configured for the system and each .network file managed by
10883 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
10885 * systemd-networkd gained support for configuring proxy ARP support for
10886 each interface, via the ProxyArp= setting in .network files. It also
10887 gained support for configuring the multicast querier feature of
10888 bridge devices, via the new MulticastQuerier= setting in .netdev
10889 files. Similarly, snooping on the IGMP traffic can be controlled
10890 via the new setting MulticastSnooping=.
10892 A new setting PreferredLifetime= has been added for addresses
10893 configured in .network file to configure the lifetime intended for an
10896 The systemd-networkd DHCP server gained the option EmitRouter=, which
10897 defaults to yes, to configure whether the DHCP Option 3 (Router)
10900 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
10901 systemd-socket-activate and installed into /usr/bin. It is now fully
10904 * systemd-journald now uses separate threads to flush changes to disk
10905 when closing journal files, thus reducing impact of slow disk I/O on
10906 logging performance.
10908 * The sd-journal API gained two new calls
10909 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
10910 can be used to open journal files using file descriptors instead of
10911 file or directory paths. sd_journal_open_container() has been
10912 deprecated, sd_journal_open_directory_fd() should be used instead
10913 with the flag SD_JOURNAL_OS_ROOT.
10915 * journalctl learned a new output mode "-o short-unix" that outputs log
10916 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
10917 UTC). It also gained support for a new --no-hostname setting to
10918 suppress the hostname column in the family of "short" output modes.
10920 * systemd-ask-password now optionally skips printing of the password to
10921 stdout with --no-output which can be useful in scripts.
10923 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
10924 (devices tagged with ID_MAKER_TOOL) are now tagged with
10925 "uaccess" and are available to logged in users.
10927 * The DeviceAllow= unit setting now supports specifiers (with "%").
10929 * "systemctl show" gained a new --value switch, which allows print a
10930 only the contents of a specific unit property, without also printing
10931 the property's name. Similar support was added to "show*" verbs
10932 of loginctl and machinectl that output "key=value" lists.
10934 * A new unit type "generated" was added for files dynamically generated
10935 by generator tools. Similarly, a new unit type "transient" is used
10936 for unit files created using the runtime API. "systemctl enable" will
10937 refuse to operate on such files.
10939 * A new command "systemctl revert" has been added that may be used to
10940 revert to the vendor version of a unit file, in case local changes
10941 have been made by adding drop-ins or overriding the unit file.
10943 * "machinectl clean" gained a new verb to automatically remove all or
10944 just hidden container images.
10946 * systemd-tmpfiles gained support for a new line type "e" for emptying
10947 directories, if they exist, without creating them if they don't.
10949 * systemd-nspawn gained support for automatically patching the UID/GIDs
10950 of the owners and the ACLs of all files and directories in a
10951 container tree to match the UID/GID user namespacing range selected
10952 for the container invocation. This mode is enabled via the new
10953 --private-users-chown switch. It also gained support for
10954 automatically choosing a free, previously unused UID/GID range when
10955 starting a container, via the new --private-users=pick setting (which
10956 implies --private-users-chown). Together, these options for the first
10957 time make user namespacing for nspawn containers fully automatic and
10958 thus deployable. The systemd-nspawn@.service template unit file has
10959 been changed to use this functionality by default.
10961 * systemd-nspawn gained a new --network-zone= switch, that allows
10962 creating ad-hoc virtual Ethernet links between multiple containers,
10963 that only exist as long as at least one container referencing them is
10964 running. This allows easy connecting of multiple containers with a
10965 common link that implements an Ethernet broadcast domain. Each of
10966 these network "zones" may be named relatively freely by the user, and
10967 may be referenced by any number of containers, but each container may
10968 only reference one of these "zones". On the lower level, this is
10969 implemented by an automatically managed bridge network interface for
10970 each zone, that is created when the first container referencing its
10971 zone is created and removed when the last one referencing its zone
10974 * The default start timeout may now be configured on the kernel command
10975 line via systemd.default_timeout_start_sec=. It was already
10976 configurable via the DefaultTimeoutStartSec= option in
10977 /etc/systemd/system.conf.
10979 * Socket units gained a new TriggerLimitIntervalSec= and
10980 TriggerLimitBurst= setting to configure a limit on the activation
10981 rate of the socket unit.
10983 * The LimitNICE= setting now optionally takes normal UNIX nice values
10984 in addition to the raw integer limit value. If the specified
10985 parameter is prefixed with "+" or "-" and is in the range -20…19 the
10986 value is understood as UNIX nice value. If not prefixed like this it
10987 is understood as raw RLIMIT_NICE limit.
10989 * Note that the effect of the PrivateDevices= unit file setting changed
10990 slightly with this release: the per-device /dev file system will be
10991 mounted read-only from this version on, and will have "noexec"
10992 set. This (minor) change of behavior might cause some (exceptional)
10993 legacy software to break, when PrivateDevices=yes is set for its
10994 service. Please leave PrivateDevices= off if you run into problems
10997 * systemd-bootchart has been split out to a separate repository:
10998 https://github.com/systemd/systemd-bootchart
11000 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
11001 merged into the kernel in its current form.
11003 * The compatibility libraries libsystemd-daemon.so,
11004 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
11005 which have been deprecated since systemd-209 have been removed along
11006 with the corresponding pkg-config files. All symbols provided by
11007 those libraries are provided by libsystemd.so.
11009 * The Capabilities= unit file setting has been removed (it is ignored
11010 for backwards compatibility). AmbientCapabilities= and
11011 CapabilityBoundingSet= should be used instead.
11013 * A new special target has been added, initrd-root-device.target,
11014 which creates a synchronization point for dependencies of the root
11015 device in early userspace. Initramfs builders must ensure that this
11016 target is now included in early userspace.
11018 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
11019 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
11020 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
11021 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
11022 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
11023 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
11024 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
11025 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
11026 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
11027 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
11028 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
11029 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
11030 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
11031 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
11032 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
11033 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
11034 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
11035 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
11036 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
11037 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
11038 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
11039 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
11040 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
11041 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
11042 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
11045 — Fairfax, 2016-05-21
11049 * The systemd-resolved DNS resolver service has gained a substantial
11050 set of new features, most prominently it may now act as a DNSSEC
11051 validating stub resolver. DNSSEC mode is currently turned off by
11052 default, but is expected to be turned on by default in one of the
11053 next releases. For now, we invite everybody to test the DNSSEC logic
11054 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
11055 service also gained a full set of D-Bus interfaces, including calls
11056 to configure DNS and DNSSEC settings per link (for use by external
11057 network management software). systemd-resolved and systemd-networkd
11058 now distinguish between "search" and "routing" domains. The former
11059 are used to qualify single-label names, the latter are used purely
11060 for routing lookups within certain domains to specific links.
11061 resolved now also synthesizes RRs for all entries from /etc/hosts.
11063 * The systemd-resolve tool (which is a client utility for
11064 systemd-resolved) has been improved considerably and is now fully
11065 supported and documented. Hence it has moved from /usr/lib/systemd to
11068 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
11071 * The coredump collection logic has been reworked: when a coredump is
11072 collected it is now written to disk, compressed and processed
11073 (including stacktrace extraction) from a new instantiated service
11074 systemd-coredump@.service, instead of directly from the
11075 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
11076 processing large coredumps can take up a substantial amount of
11077 resources and time, and this previously happened entirely outside of
11078 systemd's service supervision. With the new logic the core_pattern
11079 hook only does minimal metadata collection before passing off control
11080 to the new instantiated service, which is configured with a time
11081 limit, a nice level and other settings to minimize negative impact on
11082 the rest of the system. Also note that the new logic will honour the
11083 RLIMIT_CORE setting of the crashed process, which now allows users
11084 and processes to turn off coredumping for their processes by setting
11087 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
11088 and all forked processes by default. Previously, PID 1 would leave
11089 the setting at "0" for all processes, as set by the kernel. Note that
11090 the resource limit traditionally has no effect on the generated
11091 coredumps on the system if the /proc/sys/kernel/core_pattern hook
11092 logic is used. Since the limit is now honoured (see above) its
11093 default has been changed so that the coredumping logic is enabled by
11094 default for all processes, while allowing specific opt-out.
11096 * When the stacktrace is extracted from processes of system users, this
11097 is now done as "systemd-coredump" user, in order to sandbox this
11098 potentially security sensitive parsing operation. (Note that when
11099 processing coredumps of normal users this is done under the user ID
11100 of process that crashed, as before.) Packagers should take notice
11101 that it is now necessary to create the "systemd-coredump" system user
11102 and group at package installation time.
11104 * The systemd-activate socket activation testing tool gained support
11105 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
11106 and --seqpacket switches. It also has been extended to support both
11107 new-style and inetd-style file descriptor passing. Use the new
11108 --inetd switch to request inetd-style file descriptor passing.
11110 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
11111 variable, which takes a boolean value. If set to false, ANSI color
11112 output is disabled in the tools even when run on a terminal that
11115 * The VXLAN support in networkd now supports two new settings
11116 DestinationPort= and PortRange=.
11118 * A new systemd.machine_id= kernel command line switch has been added,
11119 that may be used to set the machine ID in /etc/machine-id if it is
11120 not initialized yet. This command line option has no effect if the
11121 file is already initialized.
11123 * systemd-nspawn gained a new --as-pid2 switch that invokes any
11124 specified command line as PID 2 rather than PID 1 in the
11125 container. In this mode PID 1 is a minimal stub init process that
11126 implements the special POSIX and Linux semantics of PID 1 regarding
11127 signal and child process management. Note that this stub init process
11128 is implemented in nspawn itself and requires no support from the
11129 container image. This new logic is useful to support running
11130 arbitrary commands in the container, as normal processes are
11131 generally not prepared to run as PID 1.
11133 * systemd-nspawn gained a new --chdir= switch for setting the current
11134 working directory for the process started in the container.
11136 * "journalctl /dev/sda" will now output all kernel log messages for
11137 specified device from the current boot, in addition to all devices
11138 that are parents of it. This should make log output about devices
11139 pretty useful, as long as kernel drivers attach enough metadata to
11140 the log messages. (The usual SATA drivers do.)
11142 * The sd-journal API gained two new calls
11143 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
11144 that report whether log data from /run or /var has been found.
11146 * journalctl gained a new switch "--fields" that prints all journal
11147 record field names currently in use in the journal. This is backed
11148 by two new sd-journal API calls sd_journal_enumerate_fields() and
11149 sd_journal_restart_fields().
11151 * Most configurable timeouts in systemd now expect an argument of
11152 "infinity" to turn them off, instead of "0" as before. The semantics
11153 from now on is that a timeout of "0" means "now", and "infinity"
11154 means "never". To maintain backwards compatibility, "0" continues to
11155 turn off previously existing timeout settings.
11157 * "systemctl reload-or-try-restart" has been renamed to "systemctl
11158 try-reload-or-restart" to clarify what it actually does: the "try"
11159 logic applies to both reloading and restarting, not just restarting.
11160 The old name continues to be accepted for compatibility.
11162 * On boot-up, when PID 1 detects that the system clock is behind the
11163 release date of the systemd version in use, the clock is now set
11164 to the latter. Previously, this was already done in timesyncd, in order
11165 to avoid running with clocks set to the various clock epochs such as
11166 1902, 1938 or 1970. With this change the logic is now done in PID 1
11167 in addition to timesyncd during early boot-up, so that it is enforced
11168 before the first process is spawned by systemd. Note that the logic
11169 in timesyncd remains, as it is more comprehensive and ensures
11170 clock monotonicity by maintaining a persistent timestamp file in
11171 /var. Since /var is generally not available in earliest boot or the
11172 initrd, this part of the logic remains in timesyncd, and is not done
11175 * Support for tweaking details in net_cls.class_id through the
11176 NetClass= configuration directive has been removed, as the kernel
11177 people have decided to deprecate that controller in cgroup v2.
11178 Userspace tools such as nftables are moving over to setting rules
11179 that are specific to the full cgroup path of a task, which obsoletes
11180 these controllers anyway. The NetClass= directive is kept around for
11181 legacy compatibility reasons. For a more in-depth description of the
11182 kernel change, please refer to the respective upstream commit:
11184 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
11186 * A new service setting RuntimeMaxSec= has been added that may be used
11187 to specify a maximum runtime for a service. If the timeout is hit, the
11188 service is terminated and put into a failure state.
11190 * A new service setting AmbientCapabilities= has been added. It allows
11191 configuration of additional Linux process capabilities that are
11192 passed to the activated processes. This is only available on very
11195 * The process resource limit settings in service units may now be used
11196 to configure hard and soft limits individually.
11198 * The various libsystemd APIs such as sd-bus or sd-event now publicly
11199 expose support for gcc's __attribute__((cleanup())) C extension.
11200 Specifically, for many object destructor functions alternative
11201 versions have been added that have names suffixed with "p" and take a
11202 pointer to a pointer to the object to destroy, instead of just a
11203 pointer to the object itself. This is useful because these destructor
11204 functions may be used directly as parameters to the cleanup
11205 construct. Internally, systemd has been a heavy user of this GCC
11206 extension for a long time, and with this change similar support is
11207 now available to consumers of the library outside of systemd. Note
11208 that by using this extension in your sources compatibility with old
11209 and strictly ANSI compatible C compilers is lost. However, all gcc or
11210 LLVM versions of recent years support this extension.
11212 * Timer units gained support for a new setting RandomizedDelaySec= that
11213 allows configuring some additional randomized delay to the configured
11214 time. This is useful to spread out timer events to avoid load peaks in
11215 clusters or larger setups.
11217 * Calendar time specifications now support sub-second accuracy.
11219 * Socket units now support listening on SCTP and UDP-lite protocol
11222 * The sd-event API now comes with a full set of man pages.
11224 * Older versions of systemd contained experimental support for
11225 compressing journal files and coredumps with the LZ4 compressor that
11226 was not compatible with the lz4 binary (due to API limitations of the
11227 lz4 library). This support has been removed; only support for files
11228 compatible with the lz4 binary remains. This LZ4 logic is now
11229 officially supported and no longer considered experimental.
11231 * The dkr image import logic has been removed again from importd. dkr's
11232 micro-services focus doesn't fit into the machine image focus of
11233 importd, and quickly got out of date with the upstream dkr API.
11235 * Creation of the /run/lock/lockdev/ directory was dropped from
11236 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
11237 been available for many years. If you still need this, you need to
11238 create your own tmpfiles.d config file with:
11240 d /run/lock/lockdev 0775 root lock -
11242 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
11243 and RebootArgument= have been moved from the [Service] section of
11244 unit files to [Unit], and they are now supported on all unit types,
11245 not just service units. Of course, systemd will continue to
11246 understand these settings also at the old location, in order to
11247 maintain compatibility.
11249 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
11250 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
11251 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
11252 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
11253 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
11254 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
11255 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
11256 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
11257 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
11258 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
11259 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
11260 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
11261 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
11262 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
11263 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
11264 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
11265 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
11266 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
11267 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
11269 — Berlin, 2016-02-11
11273 * A number of properties previously only settable in unit
11274 files are now also available as properties to set when
11275 creating transient units programmatically via the bus, as it
11276 is exposed with systemd-run's --property=
11277 setting. Specifically, these are: SyslogIdentifier=,
11278 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
11279 EnvironmentFile=, ReadWriteDirectories=,
11280 ReadOnlyDirectories=, InaccessibleDirectories=,
11281 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
11283 * When creating transient services via the bus API it is now
11284 possible to pass in a set of file descriptors to use as
11285 STDIN/STDOUT/STDERR for the invoked process.
11287 * Slice units may now be created transiently via the bus APIs,
11288 similar to the way service and scope units may already be
11289 created transiently.
11291 * Wherever systemd expects a calendar timestamp specification
11292 (like in journalctl's --since= and --until= switches) UTC
11293 timestamps are now supported. Timestamps suffixed with "UTC"
11294 are now considered to be in Universal Time Coordinated
11295 instead of the local timezone. Also, timestamps may now
11296 optionally be specified with sub-second accuracy. Both of
11297 these additions also apply to recurring calendar event
11298 specification, such as OnCalendar= in timer units.
11300 * journalctl gained a new "--sync" switch that asks the
11301 journal daemon to write all so far unwritten log messages to
11302 disk and sync the files, before returning.
11304 * systemd-tmpfiles learned two new line types "q" and "Q" that
11305 operate like "v", but also set up a basic btrfs quota
11306 hierarchy when used on a btrfs file system with quota
11309 * tmpfiles' "v", "q" and "Q" will now create a plain directory
11310 instead of a subvolume (even on a btrfs file system) if the
11311 root directory is a plain directory, and not a
11312 subvolume. This should simplify things with certain chroot()
11313 environments which are not aware of the concept of btrfs
11316 * systemd-detect-virt gained a new --chroot switch to detect
11317 whether execution takes place in a chroot() environment.
11319 * CPUAffinity= now takes CPU index ranges in addition to
11320 individual indexes.
11322 * The various memory-related resource limit settings (such as
11323 LimitAS=) now understand the usual K, M, G, … suffixes to
11324 the base of 1024 (IEC). Similar, the time-related resource
11325 limit settings understand the usual min, h, day, … suffixes
11328 * There's a new system.conf setting DefaultTasksMax= to
11329 control the default TasksMax= setting for services and
11330 scopes running on the system. (TasksMax= is the primary
11331 setting that exposes the "pids" cgroup controller on systemd
11332 and was introduced in the previous systemd release.) The
11333 setting now defaults to 512, which means services that are
11334 not explicitly configured otherwise will only be able to
11335 create 512 processes or threads at maximum, from this
11336 version on. Note that this means that thread- or
11337 process-heavy services might need to be reconfigured to set
11338 TasksMax= to a higher value. It is sufficient to set
11339 TasksMax= in these specific unit files to a higher value, or
11340 even "infinity". Similar, there's now a logind.conf setting
11341 UserTasksMax= that defaults to 4096 and limits the total
11342 number of processes or tasks each user may own
11343 concurrently. nspawn containers also have the TasksMax=
11344 value set by default now, to 8192. Note that all of this
11345 only has an effect if the "pids" cgroup controller is
11346 enabled in the kernel. The general benefit of these changes
11347 should be a more robust and safer system, that provides a
11348 certain amount of per-service fork() bomb protection.
11350 * systemd-nspawn gained the new --network-veth-extra= switch
11351 to define additional and arbitrarily-named virtual Ethernet
11352 links between the host and the container.
11354 * A new service execution setting PassEnvironment= has been
11355 added that allows importing select environment variables
11356 from PID1's environment block into the environment block of
11359 * Timer units gained support for a new RemainAfterElapse=
11360 setting which takes a boolean argument. It defaults to on,
11361 exposing behaviour unchanged to previous releases. If set to
11362 off, timer units are unloaded after they elapsed if they
11363 cannot elapse again. This is particularly useful for
11364 transient timer units, which shall not stay around longer
11365 than until they first elapse.
11367 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
11368 default now (the kernel default is 16). This is beneficial
11369 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
11370 allows substantially larger numbers of queued
11371 datagrams. This should increase the capability of systemd to
11372 parallelize boot-up, as logging and sd_notify() are unlikely
11373 to stall execution anymore. If you need to change the value
11374 from the new defaults, use the usual sysctl.d/ snippets.
11376 * The compression framing format used by the journal or
11377 coredump processing has changed to be in line with what the
11378 official LZ4 tools generate. LZ4 compression support in
11379 systemd was considered unsupported previously, as the format
11380 was not compatible with the normal tools. With this release
11381 this has changed now, and it is hence safe for downstream
11382 distributions to turn it on. While not compressing as well
11383 as the XZ, LZ4 is substantially faster, which makes
11384 it a good default choice for the compression logic in the
11385 journal and in coredump handling.
11387 * Any reference to /etc/mtab has been dropped from
11388 systemd. The file has been obsolete since a while, but
11389 systemd refused to work on systems where it was incorrectly
11390 set up (it should be a symlink or non-existent). Please make
11391 sure to update to util-linux 2.27.1 or newer in conjunction
11392 with this systemd release, which also drops any reference to
11393 /etc/mtab. If you maintain a distribution make sure that no
11394 software you package still references it, as this is a
11395 likely source of bugs. There's also a glibc bug pending,
11396 asking for removal of any reference to this obsolete file:
11398 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
11400 Note that only util-linux versions built with
11401 --enable-libmount-force-mountinfo are supported.
11403 * Support for the ".snapshot" unit type has been removed. This
11404 feature turned out to be little useful and little used, and
11405 has now been removed from the core and from systemctl.
11407 * The dependency types RequiresOverridable= and
11408 RequisiteOverridable= have been removed from systemd. They
11409 have been used only very sparingly to our knowledge and
11410 other options that provide a similar effect (such as
11411 systemctl --mode=ignore-dependencies) are much more useful
11412 and commonly used. Moreover, they were only half-way
11413 implemented as the option to control behaviour regarding
11414 these dependencies was never added to systemctl. By removing
11415 these dependency types the execution engine becomes a bit
11416 simpler. Unit files that use these dependencies should be
11417 changed to use the non-Overridable dependency types
11418 instead. In fact, when parsing unit files with these
11419 options, that's what systemd will automatically convert them
11420 too, but it will also warn, asking users to fix the unit
11421 files accordingly. Removal of these dependency types should
11422 only affect a negligible number of unit files in the wild.
11424 * Behaviour of networkd's IPForward= option changed
11425 (again). It will no longer maintain a per-interface setting,
11426 but propagate one way from interfaces where this is enabled
11427 to the global kernel setting. The global setting will be
11428 enabled when requested by a network that is set up, but
11429 never be disabled again. This change was made to make sure
11430 IPv4 and IPv6 behaviour regarding packet forwarding is
11431 similar (as the Linux IPv6 stack does not support
11432 per-interface control of this setting) and to minimize
11435 * In unit files the behaviour of %u, %U, %h, %s has
11436 changed. These specifiers will now unconditionally resolve
11437 to the various user database fields of the user that the
11438 systemd instance is running as, instead of the user
11439 configured in the specific unit via User=. Note that this
11440 effectively doesn't change much, as resolving of these
11441 specifiers was already turned off in the --system instance
11442 of systemd, as we cannot do NSS lookups from PID 1. In the
11443 --user instance of systemd these specifiers where correctly
11444 resolved, but hardly made any sense, since the user instance
11445 lacks privileges to do user switches anyway, and User= is
11446 hence useless. Moreover, even in the --user instance of
11447 systemd behaviour was awkward as it would only take settings
11448 from User= assignment placed before the specifier into
11449 account. In order to unify and simplify the logic around
11450 this the specifiers will now always resolve to the
11451 credentials of the user invoking the manager (which in case
11452 of PID 1 is the root user).
11454 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
11455 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
11456 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
11457 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
11458 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
11459 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
11460 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
11461 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
11462 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
11463 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
11464 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
11465 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
11466 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
11467 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
11470 — Berlin, 2015-11-18
11474 * systemd now depends on util-linux v2.27. More specifically,
11475 the newly added mount monitor feature in libmount now
11476 replaces systemd's former own implementation.
11478 * libmount mandates /etc/mtab not to be regular file, and
11479 systemd now enforces this condition at early boot.
11480 /etc/mtab has been deprecated and warned about for a very
11481 long time, so systems running systemd should already have
11482 stopped having this file around as anything else than a
11483 symlink to /proc/self/mounts.
11485 * Support for the "pids" cgroup controller has been added. It
11486 allows accounting the number of tasks in a cgroup and
11487 enforcing limits on it. This adds two new setting
11488 TasksAccounting= and TasksMax= to each unit, as well as a
11489 global option DefaultTasksAccounting=.
11491 * Support for the "net_cls" cgroup controller has been added.
11492 It allows assigning a net class ID to each task in the
11493 cgroup, which can then be used in firewall rules and traffic
11494 shaping configurations. Note that the kernel netfilter net
11495 class code does not currently work reliably for ingress
11496 packets on unestablished sockets.
11498 This adds a new config directive called NetClass= to CGroup
11499 enabled units. Allowed values are positive numbers for fixed
11500 assignments and "auto" for picking a free value
11503 * 'systemctl is-system-running' now returns 'offline' if the
11504 system is not booted with systemd. This command can now be
11505 used as a substitute for 'systemd-notify --booted'.
11507 * Watchdog timeouts have been increased to 3 minutes for all
11508 in-tree service files. Apparently, disk IO issues are more
11509 frequent than we hoped, and user reported >1 minute waiting
11512 * 'machine-id-commit' functionality has been merged into
11513 'machine-id-setup --commit'. The separate binary has been
11516 * The WorkingDirectory= directive in unit files may now be set
11517 to the special value '~'. In this case, the working
11518 directory is set to the home directory of the user
11519 configured in User=.
11521 * "machinectl shell" will now open the shell in the home
11522 directory of the selected user by default.
11524 * The CrashChVT= configuration file setting is renamed to
11525 CrashChangeVT=, following our usual logic of not
11526 abbreviating unnecessarily. The old directive is still
11527 supported for compat reasons. Also, this directive now takes
11528 an integer value between 1 and 63, or a boolean value. The
11529 formerly supported '-1' value for disabling stays around for
11532 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
11533 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
11534 RootDirectory= properties can now be set for transient
11537 * The systemd-analyze tool gained a new "set-log-target" verb
11538 to change the logging target the system manager logs to
11539 dynamically during runtime. This is similar to how
11540 "systemd-analyze set-log-level" already changes the log
11543 * In nspawn /sys is now mounted as tmpfs, with only a selected
11544 set of subdirectories mounted in from the real sysfs. This
11545 enhances security slightly, and is useful for ensuring user
11546 namespaces work correctly.
11548 * Support for USB FunctionFS activation has been added. This
11549 allows implementation of USB gadget services that are
11550 activated as soon as they are requested, so that they don't
11551 have to run continuously, similar to classic socket
11554 * The "systemctl exit" command now optionally takes an
11555 additional parameter that sets the exit code to return from
11556 the systemd manager when exiting. This is only relevant when
11557 running the systemd user instance, or when running the
11558 system instance in a container.
11560 * sd-bus gained the new API calls sd_bus_path_encode_many()
11561 and sd_bus_path_decode_many() that allow easy encoding and
11562 decoding of multiple identifier strings inside a D-Bus
11563 object path. Another new call sd_bus_default_flush_close()
11564 has been added to flush and close per-thread default
11567 * systemd-cgtop gained support for a -M/--machine= switch to
11568 show the control groups within a certain container only.
11570 * "systemctl kill" gained support for an optional --fail
11571 switch. If specified the requested operation will fail of no
11572 processes have been killed, because the unit had no
11573 processes attached, or similar.
11575 * A new systemd.crash_reboot=1 kernel command line option has
11576 been added that triggers a reboot after crashing. This can
11577 also be set through CrashReboot= in systemd.conf.
11579 * The RuntimeDirectory= setting now understands unit
11580 specifiers like %i or %f.
11582 * A new (still internal) library API sd-ipv4acd has been added,
11583 that implements address conflict detection for IPv4. It's
11584 based on code from sd-ipv4ll, and will be useful for
11585 detecting DHCP address conflicts.
11587 * File descriptors passed during socket activation may now be
11588 named. A new API sd_listen_fds_with_names() is added to
11589 access the names. The default names may be overridden,
11590 either in the .socket file using the FileDescriptorName=
11591 parameter, or by passing FDNAME= when storing the file
11592 descriptors using sd_notify().
11594 * systemd-networkd gained support for:
11596 - Setting the IPv6 Router Advertisement settings via
11597 IPv6AcceptRouterAdvertisements= in .network files.
11599 - Configuring the HelloTimeSec=, MaxAgeSec= and
11600 ForwardDelaySec= bridge parameters in .netdev files.
11602 - Configuring PreferredSource= for static routes in
11605 * The "ask-password" framework used to query for LUKS harddisk
11606 passwords or SSL passwords during boot gained support for
11607 caching passwords in the kernel keyring, if it is
11608 available. This makes sure that the user only has to type in
11609 a passphrase once if there are multiple objects to unlock
11610 with the same one. Previously, such password caching was
11611 available only when Plymouth was used; this moves the
11612 caching logic into the systemd codebase itself. The
11613 "systemd-ask-password" utility gained a new --keyname=
11614 switch to control which kernel keyring key to use for
11615 caching a password in. This functionality is also useful for
11616 enabling display managers such as gdm to automatically
11617 unlock the user's GNOME keyring if its passphrase, the
11618 user's password and the harddisk password are the same, if
11619 gdm-autologin is used.
11621 * When downloading tar or raw images using "machinectl
11622 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
11623 file is now also downloaded, if it is available and stored
11624 next to the image file.
11626 * Units of type ".socket" gained a new boolean setting
11627 Writable= which is only useful in conjunction with
11628 ListenSpecial=. If true, enables opening the specified
11629 special file in O_RDWR mode rather than O_RDONLY mode.
11631 * systemd-rfkill has been reworked to become a singleton
11632 service that is activated through /dev/rfkill on each rfkill
11633 state change and saves the settings to disk. This way,
11634 systemd-rfkill is now compatible with devices that exist
11635 only intermittendly, and even restores state if the previous
11636 system shutdown was abrupt rather than clean.
11638 * The journal daemon gained support for vacuuming old journal
11639 files controlled by the number of files that shall remain,
11640 in addition to the already existing control by size and by
11641 date. This is useful as journal interleaving performance
11642 degrades with too many separate journal files, and allows
11643 putting an effective limit on them. The new setting defaults
11644 to 100, but this may be changed by setting SystemMaxFiles=
11645 and RuntimeMaxFiles= in journald.conf. Also, the
11646 "journalctl" tool gained the new --vacuum-files= switch to
11647 manually vacuum journal files to leave only the specified
11648 number of files in place.
11650 * udev will now create /dev/disk/by-path links for ATA devices
11651 on kernels where that is supported.
11653 * Galician, Serbian, Turkish and Korean translations were added.
11655 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
11656 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
11657 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
11658 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
11659 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
11660 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
11661 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
11662 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
11663 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
11664 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
11665 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
11666 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
11667 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
11668 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
11669 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
11670 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
11671 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
11672 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
11674 — Berlin, 2015-10-07
11678 * The DHCP implementation of systemd-networkd gained a set of
11681 - The DHCP server now supports emitting DNS and NTP
11682 information. It may be enabled and configured via
11683 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
11684 and NTP information is enabled, but no servers are
11685 configured, the corresponding uplink information (if there
11686 is any) is propagated.
11688 - Server and client now support transmission and reception
11689 of timezone information. It can be configured via the
11690 newly introduced network options UseTimezone=,
11691 EmitTimezone=, and Timezone=. Transmission of timezone
11692 information is enabled between host and containers by
11693 default now: the container will change its local timezone
11694 to what the host has set.
11696 - Lease timeouts can now be configured via
11697 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
11699 - The DHCP server improved on the stability of
11700 leases. Clients are more likely to get the same lease
11701 information back, even if the server loses state.
11703 - The DHCP server supports two new configuration options to
11704 control the lease address pool metrics, PoolOffset= and
11707 * The encapsulation limit of tunnels in systemd-networkd may
11708 now be configured via 'EncapsulationLimit='. It allows
11709 modifying the maximum additional levels of encapsulation
11710 that are permitted to be prepended to a packet.
11712 * systemd now supports the concept of user buses replacing
11713 session buses, if used with dbus-1.10 (and enabled via dbus
11714 --enable-user-session). It previously only supported this on
11715 kdbus-enabled systems, and this release expands this to
11716 'dbus-daemon' systems.
11718 * systemd-networkd now supports predictable interface names
11719 for virtio devices.
11721 * systemd now optionally supports the new Linux kernel
11722 "unified" control group hierarchy. If enabled via the kernel
11723 command-line option 'systemd.unified_cgroup_hierarchy=1',
11724 systemd will try to mount the unified cgroup hierarchy
11725 directly on /sys/fs/cgroup. If not enabled, or not
11726 available, systemd will fall back to the legacy cgroup
11727 hierarchy setup, as before. Host system and containers can
11728 mix and match legacy and unified hierarchies as they
11729 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
11730 environment variable to individually select the hierarchy to
11731 use for executed containers. By default, nspawn will use the
11732 unified hierarchy for the containers if the host uses the
11733 unified hierarchy, and the legacy hierarchy otherwise.
11734 Please note that at this point the unified hierarchy is an
11735 experimental kernel feature and is likely to change in one
11736 of the next kernel releases. Therefore, it should not be
11737 enabled by default in downstream distributions yet. The
11738 minimum required kernel version for the unified hierarchy to
11739 work is 4.2. Note that when the unified hierarchy is used
11740 for the first time delegated access to controllers is
11741 safe. Because of this systemd-nspawn containers will get
11742 access to controllers now, as will systemd user
11743 sessions. This means containers and user sessions may now
11744 manage their own resources, partitioning up what the system
11747 * A new special scope unit "init.scope" has been introduced
11748 that encapsulates PID 1 of the system. It may be used to
11749 determine resource usage and enforce resource limits on PID
11750 1 itself. PID 1 hence moved out of the root of the control
11753 * The cgtop tool gained support for filtering out kernel
11754 threads when counting tasks in a control group. Also, the
11755 count of processes is now recursively summed up by
11756 default. Two options -k and --recursive= have been added to
11757 revert to old behaviour. The tool has also been updated to
11758 work correctly in containers now.
11760 * systemd-nspawn's --bind= and --bind-ro= options have been
11761 extended to allow creation of non-recursive bind mounts.
11763 * libsystemd gained two new calls sd_pid_get_cgroup() and
11764 sd_peer_get_cgroup() which return the control group path of
11765 a process or peer of a connected AF_UNIX socket. This
11766 function call is particularly useful when implementing
11767 delegated subtrees support in the control group hierarchy.
11769 * The "sd-event" event loop API of libsystemd now supports
11770 correct dequeuing of real-time signals, without losing
11773 * When systemd requests a polkit decision when managing units it
11774 will now add additional fields to the request, including unit
11775 name and desired operation. This enables more powerful polkit
11776 policies, that make decisions depending on these parameters.
11778 * nspawn learnt support for .nspawn settings files, that may
11779 accompany the image files or directories of containers, and
11780 may contain additional settings for the container. This is
11781 an alternative to configuring container parameters via the
11782 nspawn command line.
11784 Contributions from: Cristian Rodríguez, Daniel Mack, David
11785 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
11786 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
11787 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
11788 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
11789 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
11790 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
11791 Andersen, Tom Gundersen, Torstein Husebø
11793 — Berlin, 2015-09-08
11797 * machinectl gained a new verb 'shell' which opens a fresh
11798 shell on the target container or the host. It is similar to
11799 the existing 'login' command of machinectl, but spawns the
11800 shell directly without prompting for username or
11801 password. The pseudo machine '.host' now refers to the local
11802 host and is used by default. Hence, 'machinectl shell' can
11803 be used as replacement for 'su -' which spawns a session as
11804 a fresh systemd unit in a way that is fully isolated from
11805 the originating session.
11807 * systemd-networkd learned to cope with private-zone DHCP
11808 options and allows other programs to query the values.
11810 * SELinux access control when enabling/disabling units is no
11811 longer enforced with this release. The previous implementation
11812 was incorrect, and a new corrected implementation is not yet
11813 available. As unit file operations are still protected via
11814 polkit and D-Bus policy this is not a security problem. Yet,
11815 distributions which care about optimal SELinux support should
11816 probably not stabilize on this release.
11818 * sd-bus gained support for matches of type "arg0has=", that
11819 test for membership of strings in string arrays sent in bus
11822 * systemd-resolved now dumps the contents of its DNS and LLMNR
11823 caches to the logs on reception of the SIGUSR1 signal. This
11824 is useful to debug DNS behaviour.
11826 * The coredumpctl tool gained a new --directory= option to
11827 operate on journal files in a specific directory.
11829 * "systemctl reboot" and related commands gained a new
11830 "--message=" option which may be used to set a free-text
11831 wall message when shutting down or rebooting the
11832 system. This message is also logged, which is useful for
11833 figuring out the reason for a reboot or shutdown a
11836 * The "systemd-resolve-host" tool's -i switch now takes
11837 network interface numbers as alternative to interface names.
11839 * A new unit file setting for services has been introduced:
11840 UtmpMode= allows configuration of how precisely systemd
11841 handles utmp and wtmp entries for the service if this is
11842 enabled. This allows writing services that appear similar to
11843 user sessions in the output of the "w", "who", "last" and
11846 * systemd-resolved will now locally synthesize DNS resource
11847 records for the "localhost" and "gateway" domains as well as
11848 the local hostname. This should ensure that clients querying
11849 RRs via resolved will get similar results as those going via
11850 NSS, if nss-myhostname is enabled.
11852 Contributions from: Alastair Hughes, Alex Crawford, Daniel
11853 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
11854 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
11855 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
11856 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
11857 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
11858 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
11859 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
11860 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
11861 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
11862 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
11863 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
11865 — Berlin, 2015-08-27
11869 * The systemd-efi-boot-generator functionality was merged into
11870 systemd-gpt-auto-generator.
11872 * systemd-networkd now supports Group Policy for vxlan
11873 devices. It can be enabled via the new boolean configuration
11874 option called 'GroupPolicyExtension='.
11876 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
11877 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
11878 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
11880 — Berlin, 2015-07-31
11884 * The python-systemd code has been removed from the systemd repository.
11885 A new repository has been created which accommodates the code from
11886 now on, and we kindly ask distributions to create a separate package
11887 for this: https://github.com/systemd/python-systemd
11889 * The systemd daemon will now reload its main configuration
11890 (/etc/systemd/system.conf) on daemon-reload.
11892 * sd-dhcp now exposes vendor specific extensions via
11893 sd_dhcp_lease_get_vendor_specific().
11895 * systemd-networkd gained a number of new configuration options.
11897 - A new boolean configuration option for TAP devices called
11898 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
11899 device, thus allowing to send and receive GSO packets.
11901 - A new tunnel configuration option called 'CopyDSCP='.
11902 If enabled, the DSCP field of ip6 tunnels is copied into the
11903 decapsulated packet.
11905 - A set of boolean bridge configuration options were added.
11906 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
11907 and 'UnicastFlood=' are now parsed by networkd and applied to the
11908 respective bridge link device via the respective IFLA_BRPORT_*
11911 - A new string configuration option to override the hostname sent
11912 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
11913 is true, networkd will use the configured hostname instead of the
11914 system hostname when sending DHCP requests.
11916 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
11917 networkd will configure the IPv6 flow-label of the tunnel device
11918 according to RFC2460.
11920 - The 'macvtap' virtual network devices are now supported, similar to
11921 the already supported 'macvlan' devices.
11923 * systemd-resolved now implements RFC5452 to improve resilience against
11924 cache poisoning. Additionally, source port randomization is enabled
11925 by default to further protect against DNS spoofing attacks.
11927 * nss-mymachines now supports translating UIDs and GIDs of running
11928 containers with user-namespaces enabled. If a container 'foo'
11929 translates a host uid 'UID' to the container uid 'TUID', then
11930 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
11931 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
11932 mapped as 'vg-foo-TGID'.
11934 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
11935 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
11936 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
11937 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
11938 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
11939 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
11940 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
11941 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
11942 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
11943 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
11945 — Berlin, 2015-07-29
11949 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
11950 There are no known issues with current sysfs, and udev does not need
11951 or should be used to work around such bugs.
11953 * udev does no longer enable USB HID power management. Several reports
11954 indicate, that some devices cannot handle that setting.
11956 * The udev accelerometer helper was removed. The functionality
11957 is now fully included in iio-sensor-proxy. But this means,
11958 older iio-sensor-proxy versions will no longer provide
11959 accelerometer/orientation data with this systemd version.
11960 Please upgrade iio-sensor-proxy to version 1.0.
11962 * networkd gained a new configuration option IPv6PrivacyExtensions=
11963 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
11964 for Stateless Address") on selected networks.
11966 * For the sake of fewer build-time dependencies and less code in the
11967 main repository, the python bindings are about to be removed in the
11968 next release. A new repository has been created which accommodates
11969 the code from now on, and we kindly ask distributions to create a
11970 separate package for this. The removal will take place in v223.
11972 https://github.com/systemd/python-systemd
11974 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
11975 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
11976 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
11977 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
11978 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
11979 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
11980 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
11981 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
11982 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
11983 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
11985 — Berlin, 2015-07-07
11989 * The sd-bus.h and sd-event.h APIs have now been declared
11990 stable and have been added to the official interface of
11991 libsystemd.so. sd-bus implements an alternative D-Bus client
11992 library, that is relatively easy to use, very efficient and
11993 supports both classic D-Bus as well as kdbus as transport
11994 backend. sd-event is a generic event loop abstraction that
11995 is built around Linux epoll, but adds features such as event
11996 prioritization or efficient timer handling. Both APIs are good
11997 choices for C programs looking for a bus and/or event loop
11998 implementation that is minimal and does not have to be
11999 portable to other kernels.
12001 * kdbus support is no longer compile-time optional. It is now
12002 always built-in. However, it can still be disabled at
12003 runtime using the kdbus=0 kernel command line setting, and
12004 that setting may be changed to default to off, by specifying
12005 --disable-kdbus at build-time. Note though that the kernel
12006 command line setting has no effect if the kdbus.ko kernel
12007 module is not installed, in which case kdbus is (obviously)
12008 also disabled. We encourage all downstream distributions to
12009 begin testing kdbus by adding it to the kernel images in the
12010 development distributions, and leaving kdbus support in
12013 * The minimal required util-linux version has been bumped to
12016 * Support for chkconfig (--enable-chkconfig) was removed in
12017 favor of calling an abstraction tool
12018 /lib/systemd/systemd-sysv-install. This needs to be
12019 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
12020 in README for details.
12022 * If there's a systemd unit and a SysV init script for the
12023 same service name, and the user executes "systemctl enable"
12024 for it (or a related call), then this will now enable both
12025 (or execute the related operation on both), not just the
12028 * The libudev API documentation has been converted from gtkdoc
12031 * gudev has been removed from the systemd tree, it is now an
12034 * The systemd-cgtop tool learnt a new --raw switch to generate
12035 "raw" (machine parsable) output.
12037 * networkd's IPForwarding= .network file setting learnt the
12038 new setting "kernel", which ensures that networkd does not
12039 change the IP forwarding sysctl from the default kernel
12042 * The systemd-logind bus API now exposes a new boolean
12043 property "Docked" that reports whether logind considers the
12044 system "docked", i.e. connected to a docking station or not.
12046 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
12047 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
12048 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
12049 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
12050 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
12051 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
12052 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
12053 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
12054 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
12055 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
12056 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
12057 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
12058 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
12059 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
12060 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
12061 Fink, Zbigniew Jędrzejewski-Szmek
12063 — Berlin, 2015-06-19
12067 * The gudev library has been extracted into a separate repository
12068 available at: https://git.gnome.org/browse/libgudev/
12069 It is now managed as part of the Gnome project. Distributions
12070 are recommended to pass --disable-gudev to systemd and use
12071 gudev from the Gnome project instead. gudev is still included
12072 in systemd, for now. It will be removed soon, though. Please
12073 also see the announcement-thread on systemd-devel:
12074 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
12076 * systemd now exposes a CPUUsageNSec= property for each
12077 service unit on the bus, that contains the overall consumed
12078 CPU time of a service (the sum of what each process of the
12079 service consumed). This value is only available if
12080 CPUAccounting= is turned on for a service, and is then shown
12081 in the "systemctl status" output.
12083 * Support for configuring alternative mappings of the old SysV
12084 runlevels to systemd targets has been removed. They are now
12085 hardcoded in a way that runlevels 2, 3, 4 all map to
12086 multi-user.target and 5 to graphical.target (which
12087 previously was already the default behaviour).
12089 * The auto-mounter logic gained support for mount point
12090 expiry, using a new TimeoutIdleSec= setting in .automount
12091 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
12093 * The EFI System Partition (ESP) as mounted to /boot by
12094 systemd-efi-boot-generator will now be unmounted
12095 automatically after 2 minutes of not being used. This should
12096 minimize the risk of ESP corruptions.
12098 * New /etc/fstab options x-systemd.requires= and
12099 x-systemd.requires-mounts-for= are now supported to express
12100 additional dependencies for mounts. This is useful for
12101 journaling file systems that support external journal
12102 devices or overlay file systems that require underlying file
12103 systems to be mounted.
12105 * systemd does not support direct live-upgrades (via systemctl
12106 daemon-reexec) from versions older than v44 anymore. As no
12107 distribution we are aware of shipped such old versions in a
12108 stable release this should not be problematic.
12110 * When systemd forks off a new per-connection service instance
12111 it will now set the $REMOTE_ADDR environment variable to the
12112 remote IP address, and $REMOTE_PORT environment variable to
12113 the remote IP port. This behaviour is similar to the
12114 corresponding environment variables defined by CGI.
12116 * systemd-networkd gained support for uplink failure
12117 detection. The BindCarrier= option allows binding interface
12118 configuration dynamically to the link sense of other
12119 interfaces. This is useful to achieve behaviour like in
12122 * systemd-networkd gained support for configuring the DHCP
12123 client identifier to use when requesting leases.
12125 * systemd-networkd now has a per-network UseNTP= option to
12126 configure whether NTP server information acquired via DHCP
12127 is passed on to services like systemd-timesyncd.
12129 * systemd-networkd gained support for vti6 tunnels.
12131 * Note that systemd-networkd manages the sysctl variable
12132 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
12133 it is configured for since v219. The variable controls IP
12134 forwarding, and is a per-interface alternative to the global
12135 /proc/sys/net/ipv[46]/ip_forward. This setting is
12136 configurable in the IPForward= option, which defaults to
12137 "no". This means if networkd is used for an interface it is
12138 no longer sufficient to set the global sysctl option to turn
12139 on IP forwarding! Instead, the .network file option
12140 IPForward= needs to be turned on! Note that the
12141 implementation of this behaviour was broken in v219 and has
12142 been fixed in v220.
12144 * Many bonding and vxlan options are now configurable in
12147 * systemd-nspawn gained a new --property= setting to set unit
12148 properties for the container scope. This is useful for
12149 setting resource parameters (e.g. "CPUShares=500") on
12150 containers started from the command line.
12152 * systemd-nspawn gained a new --private-users= switch to make
12153 use of user namespacing available on recent Linux kernels.
12155 * systemd-nspawn may now be called as part of a shell pipeline
12156 in which case the pipes used for stdin and stdout are passed
12157 directly to the process invoked in the container, without
12158 indirection via a pseudo tty.
12160 * systemd-nspawn gained a new switch to control the UNIX
12161 signal to use when killing the init process of the container
12162 when shutting down.
12164 * systemd-nspawn gained a new --overlay= switch for mounting
12165 overlay file systems into the container using the new kernel
12168 * When a container image is imported via systemd-importd and
12169 the host file system is not btrfs, a loopback block device
12170 file is created in /var/lib/machines.raw with a btrfs file
12171 system inside. It is then mounted to /var/lib/machines to
12172 enable btrfs features for container management. The loopback
12173 file and btrfs file system is grown as needed when container
12174 images are imported via systemd-importd.
12176 * systemd-machined/systemd-importd gained support for btrfs
12177 quota, to enforce container disk space limits on disk. This
12178 is exposed in "machinectl set-limit".
12180 * systemd-importd now can import containers from local .tar,
12181 .raw and .qcow2 images, and export them to .tar and .raw. It
12182 can also import dkr v2 images now from the network (on top
12185 * systemd-importd gained support for verifying downloaded
12186 images with gpg2 (previously only gpg1 was supported).
12188 * systemd-machined, systemd-logind, systemd: most bus calls are
12189 now accessible to unprivileged processes via polkit. Also,
12190 systemd-logind will now allow users to kill their own sessions
12191 without further privileges or authorization.
12193 * systemd-shutdownd has been removed. This service was
12194 previously responsible for implementing scheduled shutdowns
12195 as exposed in /usr/bin/shutdown's time parameter. This
12196 functionality has now been moved into systemd-logind and is
12197 accessible via a bus interface.
12199 * "systemctl reboot" gained a new switch --firmware-setup that
12200 can be used to reboot into the EFI firmware setup, if that
12201 is available. systemd-logind now exposes an API on the bus
12202 to trigger such reboots, in case graphical desktop UIs want
12203 to cover this functionality.
12205 * "systemctl enable", "systemctl disable" and "systemctl mask"
12206 now support a new "--now" switch. If specified the units
12207 that are enabled will also be started, and the ones
12208 disabled/masked also stopped.
12210 * The Gummiboot EFI boot loader tool has been merged into
12211 systemd, and renamed to "systemd-boot". The bootctl tool has been
12212 updated to support systemd-boot.
12214 * An EFI kernel stub has been added that may be used to create
12215 kernel EFI binaries that contain not only the actual kernel,
12216 but also an initrd, boot splash, command line and OS release
12217 information. This combined binary can then be signed as a
12218 single image, so that the firmware can verify it all in one
12219 step. systemd-boot has special support for EFI binaries created
12220 like this and can extract OS release information from them
12221 and show them in the boot menu. This functionality is useful
12222 to implement cryptographically verified boot schemes.
12224 * Optional support has been added to systemd-fsck to pass
12225 fsck's progress report to an AF_UNIX socket in the file
12228 * udev will no longer create device symlinks for all block devices by
12229 default. A deny list for excluding special block devices from this
12230 logic has been turned into an allow list that requires picking block
12231 devices explicitly that require device symlinks.
12233 * A new (currently still internal) API sd-device.h has been
12234 added to libsystemd. This modernized API is supposed to
12235 replace libudev eventually. In fact, already much of libudev
12236 is now just a wrapper around sd-device.h.
12238 * A new hwdb database for storing metadata about pointing
12239 stick devices has been added.
12241 * systemd-tmpfiles gained support for setting file attributes
12242 similar to the "chattr" tool with new 'h' and 'H' lines.
12244 * systemd-journald will no longer unconditionally set the
12245 btrfs NOCOW flag on new journal files. This is instead done
12246 with tmpfiles snippet using the new 'h' line type. This
12247 allows easy disabling of this logic, by masking the
12248 journal-nocow.conf tmpfiles file.
12250 * systemd-journald will now translate audit message types to
12251 human readable identifiers when writing them to the
12252 journal. This should improve readability of audit messages.
12254 * The LUKS logic gained support for the offset= and skip=
12255 options in /etc/crypttab, as previously implemented by
12258 * /usr/lib/os-release gained a new optional field VARIANT= for
12259 distributions that support multiple variants (such as a
12260 desktop edition, a server edition, …)
12262 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
12263 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
12264 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
12265 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
12266 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
12267 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
12268 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
12269 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
12270 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
12271 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
12272 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
12273 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
12274 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
12275 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
12276 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
12277 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
12278 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
12279 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
12280 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
12281 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
12282 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
12283 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
12284 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
12285 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
12286 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
12287 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
12288 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
12290 — Berlin, 2015-05-22
12294 * Introduce a new API "sd-hwdb.h" for querying the hardware
12295 metadata database. With this minimal interface one can query
12296 and enumerate the udev hwdb, decoupled from the old libudev
12297 library. libudev's interface for this is now only a wrapper
12298 around sd-hwdb. A new tool systemd-hwdb has been added to
12299 interface with and update the database.
12301 * When any of systemd's tools copies files (for example due to
12302 tmpfiles' C lines) a btrfs reflink will attempted first,
12303 before bytewise copying is done.
12305 * systemd-nspawn gained a new --ephemeral switch. When
12306 specified a btrfs snapshot is taken of the container's root
12307 directory, and immediately removed when the container
12308 terminates again. Thus, a container can be started whose
12309 changes never alter the container's root directory, and are
12310 lost on container termination. This switch can also be used
12311 for starting a container off the root file system of the
12312 host without affecting the host OS. This switch is only
12313 available on btrfs file systems.
12315 * systemd-nspawn gained a new --template= switch. It takes the
12316 path to a container tree to use as template for the tree
12317 specified via --directory=, should that directory be
12318 missing. This allows instantiating containers dynamically,
12319 on first run. This switch is only available on btrfs file
12322 * When a .mount unit refers to a mount point on which multiple
12323 mounts are stacked, and the .mount unit is stopped all of
12324 the stacked mount points will now be unmounted until no
12325 mount point remains.
12327 * systemd now has an explicit notion of supported and
12328 unsupported unit types. Jobs enqueued for unsupported unit
12329 types will now fail with an "unsupported" error code. More
12330 specifically .swap, .automount and .device units are not
12331 supported in containers, .busname units are not supported on
12332 non-kdbus systems. .swap and .automount are also not
12333 supported if their respective kernel compile time options
12336 * machinectl gained support for two new "copy-from" and
12337 "copy-to" commands for copying files from a running
12338 container to the host or vice versa.
12340 * machinectl gained support for a new "bind" command to bind
12341 mount host directories into local containers. This is
12342 currently only supported for nspawn containers.
12344 * networkd gained support for configuring bridge forwarding
12345 database entries (fdb) from .network files.
12347 * A new tiny daemon "systemd-importd" has been added that can
12348 download container images in tar, raw, qcow2 or dkr formats,
12349 and make them available locally in /var/lib/machines, so
12350 that they can run as nspawn containers. The daemon can GPG
12351 verify the downloads (not supported for dkr, since it has no
12352 provisions for verifying downloads). It will transparently
12353 decompress bz2, xz, gzip compressed downloads if necessary,
12354 and restore sparse files on disk. The daemon uses privilege
12355 separation to ensure the actual download logic runs with
12356 fewer privileges than the daemon itself. machinectl has
12357 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
12358 make the functionality of importd available to the
12359 user. With this in place the Fedora and Ubuntu "Cloud"
12360 images can be downloaded and booted as containers unmodified
12361 (the Fedora images lack the appropriate GPG signature files
12362 currently, so they cannot be verified, but this will change
12363 soon, hopefully). Note that downloading images is currently
12364 only fully supported on btrfs.
12366 * machinectl is now able to list container images found in
12367 /var/lib/machines, along with some metadata about sizes of
12368 disk and similar. If the directory is located on btrfs and
12369 quota is enabled, this includes quota display. A new command
12370 "image-status" has been added that shows additional
12371 information about images.
12373 * machinectl is now able to clone container images
12374 efficiently, if the underlying file system (btrfs) supports
12375 it, with the new "machinectl clone" command. It also
12376 gained commands for renaming and removing images, as well as
12377 marking them read-only or read-write (supported also on
12378 legacy file systems).
12380 * networkd gained support for collecting LLDP network
12381 announcements, from hardware that supports this. This is
12382 shown in networkctl output.
12384 * systemd-run gained support for a new -t (--pty) switch for
12385 invoking a binary on a pty whose input and output is
12386 connected to the invoking terminal. This allows executing
12387 processes as system services while interactively
12388 communicating with them via the terminal. Most interestingly
12389 this is supported across container boundaries. Invoking
12390 "systemd-run -t /bin/bash" is an alternative to running a
12391 full login session, the difference being that the former
12392 will not register a session, nor go through the PAM session
12395 * tmpfiles gained support for a new "v" line type for creating
12396 btrfs subvolumes. If the underlying file system is a legacy
12397 file system, this automatically degrades to creating a
12398 normal directory. Among others /var/lib/machines is now
12399 created like this at boot, should it be missing.
12401 * The directory /var/lib/containers/ has been deprecated and
12402 been replaced by /var/lib/machines. The term "machines" has
12403 been used in the systemd context as generic term for both
12404 VMs and containers, and hence appears more appropriate for
12405 this, as the directory can also contain raw images bootable
12408 * systemd-nspawn when invoked with -M but without --directory=
12409 or --image= is now capable of searching for the container
12410 root directory, subvolume or disk image automatically, in
12411 /var/lib/machines. systemd-nspawn@.service has been updated
12412 to make use of this, thus allowing it to be used for raw
12415 * A new machines.target unit has been introduced that is
12416 supposed to group all containers/VMs invoked as services on
12417 the system. systemd-nspawn@.service has been updated to
12418 integrate with that.
12420 * machinectl gained a new "start" command, for invoking a
12421 container as a service. "machinectl start foo" is mostly
12422 equivalent to "systemctl start systemd-nspawn@foo.service",
12423 but handles escaping in a nicer way.
12425 * systemd-nspawn will now mount most of the cgroupfs tree
12426 read-only into each container, with the exception of the
12427 container's own subtree in the name=systemd hierarchy.
12429 * journald now sets the special FS_NOCOW file flag for its
12430 journal files. This should improve performance on btrfs, by
12431 avoiding heavy fragmentation when journald's write-pattern
12432 is used on COW file systems. It degrades btrfs' data
12433 integrity guarantees for the files to the same levels as for
12434 ext3/ext4 however. This should be OK though as journald does
12435 its own data integrity checks and all its objects are
12436 checksummed on disk. Also, journald should handle btrfs disk
12437 full events a lot more gracefully now, by processing SIGBUS
12438 errors, and not relying on fallocate() anymore.
12440 * When journald detects that journal files it is writing to
12441 have been deleted it will immediately start new journal
12444 * systemd now provides a way to store file descriptors
12445 per-service in PID 1. This is useful for daemons to ensure
12446 that fds they require are not lost during a daemon
12447 restart. The fds are passed to the daemon on the next
12448 invocation in the same way socket activation fds are
12449 passed. This is now used by journald to ensure that the
12450 various sockets connected to all the system's stdout/stderr
12451 are not lost when journald is restarted. File descriptors
12452 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
12453 an extension to sd_notify(). Note that a limit is enforced
12454 on the number of fds a service can store in PID 1, and it
12455 defaults to 0, so that no fds may be stored, unless this is
12456 explicitly turned on.
12458 * The default TERM variable to use for units connected to a
12459 terminal, when no other value is explicitly is set is now
12460 vt220 rather than vt102. This should be fairly safe still,
12461 but allows PgUp/PgDn work.
12463 * The /etc/crypttab option header= as known from Debian is now
12466 * "loginctl user-status" and "loginctl session-status" will
12467 now show the last 10 lines of log messages of the
12468 user/session following the status output. Similar,
12469 "machinectl status" will show the last 10 log lines
12470 associated with a virtual machine or container
12471 service. (Note that this is usually not the log messages
12472 done in the VM/container itself, but simply what the
12473 container manager logs. For nspawn this includes all console
12476 * "loginctl session-status" without further argument will now
12477 show the status of the session of the caller. Similar,
12478 "lock-session", "unlock-session", "activate",
12479 "enable-linger", "disable-linger" may now be called without
12480 session/user parameter in which case they apply to the
12481 caller's session/user.
12483 * An X11 session scriptlet is now shipped that uploads
12484 $DISPLAY and $XAUTHORITY into the environment of the systemd
12485 --user daemon if a session begins. This should improve
12486 compatibility with X11 enabled applications run as systemd
12489 * Generators are now subject to masking via /etc and /run, the
12490 same way as unit files.
12492 * networkd .network files gained support for configuring
12493 per-link IPv4/IPv6 packet forwarding as well as IPv4
12494 masquerading. This is by default turned on for veth links to
12495 containers, as registered by systemd-nspawn. This means that
12496 nspawn containers run with --network-veth will now get
12497 automatic routed access to the host's networks without any
12498 further configuration or setup, as long as networkd runs on
12501 * systemd-nspawn gained the --port= (-p) switch to expose TCP
12502 or UDP posts of a container on the host. With this in place
12503 it is possible to run containers with private veth links
12504 (--network-veth), and have their functionality exposed on
12505 the host as if their services were running directly on the
12508 * systemd-nspawn's --network-veth switch now gained a short
12509 version "-n", since with the changes above it is now truly
12510 useful out-of-the-box. The systemd-nspawn@.service has been
12511 updated to make use of it too by default.
12513 * systemd-nspawn will now maintain a per-image R/W lock, to
12514 ensure that the same image is not started more than once
12515 writable. (It's OK to run an image multiple times
12516 simultaneously in read-only mode.)
12518 * systemd-nspawn's --image= option is now capable of
12519 dissecting and booting MBR and GPT disk images that contain
12520 only a single active Linux partition. Previously it
12521 supported only GPT disk images with proper GPT type
12522 IDs. This allows running cloud images from major
12523 distributions directly with systemd-nspawn, without
12526 * In addition to collecting mouse dpi data in the udev
12527 hardware database, there's now support for collecting angle
12528 information for mouse scroll wheels. The database is
12529 supposed to guarantee similar scrolling behavior on mice
12530 that it knows about. There's also support for collecting
12531 information about Touchpad types.
12533 * udev's input_id built-in will now also collect touch screen
12534 dimension data and attach it to probed devices.
12536 * /etc/os-release gained support for a Distribution Privacy
12539 * networkd gained support for creating "ipvlan", "gretap",
12540 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
12542 * systemd-tmpfiles gained support for "a" lines for setting
12545 * systemd-nspawn will now mount /tmp in the container to
12546 tmpfs, automatically.
12548 * systemd now exposes the memory.usage_in_bytes cgroup
12549 attribute and shows it for each service in the "systemctl
12550 status" output, if available.
12552 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
12553 immediate reboot is triggered. This useful if shutdown is
12554 hung and is unable to complete, to expedite the
12555 operation. Note that this kind of reboot will still unmount
12556 all file systems, and hence should not result in fsck being
12557 run on next reboot.
12559 * A .device unit for an optical block device will now be
12560 considered active only when a medium is in the drive. Also,
12561 mount units are now bound to their backing devices thus
12562 triggering automatic unmounting when devices become
12563 unavailable. With this in place systemd will now
12564 automatically unmount left-over mounts when a CD-ROM is
12565 ejected or a USB stick is yanked from the system.
12567 * networkd-wait-online now has support for waiting for
12568 specific interfaces only (with globbing), and for giving up
12569 after a configurable timeout.
12571 * networkd now exits when idle. It will be automatically
12572 restarted as soon as interfaces show up, are removed or
12573 change state. networkd will stay around as long as there is
12574 at least one DHCP state machine or similar around, that keep
12577 * networkd may now configure IPv6 link-local addressing in
12578 addition to IPv4 link-local addressing.
12580 * The IPv6 "token" for use in SLAAC may now be configured for
12581 each .network interface in networkd.
12583 * Routes configured with networkd may now be assigned a scope
12586 * networkd's [Match] sections now support globbing and lists
12587 of multiple space-separated matches per item.
12589 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
12590 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
12591 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
12592 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
12593 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
12594 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
12595 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
12596 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
12597 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
12598 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
12599 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
12600 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
12601 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
12602 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
12603 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
12604 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
12605 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
12606 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
12607 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
12608 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
12609 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
12610 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
12611 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
12612 Hoffmann, Zbigniew Jędrzejewski-Szmek
12614 — Berlin, 2015-02-16
12618 * When querying unit file enablement status (for example via
12619 "systemctl is-enabled"), a new state "indirect" is now known
12620 which indicates that a unit might not be enabled itself, but
12621 another unit listed in its Also= setting might be.
12623 * Similar to the various existing ConditionXYZ= settings for
12624 units, there are now matching AssertXYZ= settings. While
12625 failing conditions cause a unit to be skipped, but its job
12626 to succeed, failing assertions declared like this will cause
12627 a unit start operation and its job to fail.
12629 * hostnamed now knows a new chassis type "embedded".
12631 * systemctl gained a new "edit" command. When used on a unit
12632 file, this allows extending unit files with .d/ drop-in
12633 configuration snippets or editing the full file (after
12634 copying it from /usr/lib to /etc). This will invoke the
12635 user's editor (as configured with $EDITOR), and reload the
12636 modified configuration after editing.
12638 * "systemctl status" now shows the suggested enablement state
12639 for a unit, as declared in the (usually vendor-supplied)
12640 system preset files.
12642 * nss-myhostname will now resolve the single-label hostname
12643 "gateway" to the locally configured default IP routing
12644 gateways, ordered by their metrics. This assigns a stable
12645 name to the used gateways, regardless which ones are
12646 currently configured. Note that the name will only be
12647 resolved after all other name sources (if nss-myhostname is
12648 configured properly) and should hence not negatively impact
12649 systems that use the single-label hostname "gateway" in
12652 * systemd-inhibit now allows filtering by mode when listing
12655 * Scope and service units gained a new "Delegate" boolean
12656 property, which, when set, allows processes running inside the
12657 unit to further partition resources. This is primarily
12658 useful for systemd user instances as well as container
12661 * journald will now pick up audit messages directly from
12662 the kernel, and log them like any other log message. The
12663 audit fields are split up and fully indexed. This means that
12664 journalctl in many ways is now a (nicer!) alternative to
12665 ausearch, the traditional audit client. Note that this
12666 implements only a minimal audit client. If you want the
12667 special audit modes like reboot-on-log-overflow, please use
12668 the traditional auditd instead, which can be used in
12669 parallel to journald.
12671 * The ConditionSecurity= unit file option now understands the
12672 special string "audit" to check whether auditing is
12675 * journalctl gained two new commands --vacuum-size= and
12676 --vacuum-time= to delete old journal files until the
12677 remaining ones take up no more than the specified size on disk,
12678 or are not older than the specified time.
12680 * A new, native PPPoE library has been added to sd-network,
12681 systemd's library of light-weight networking protocols. This
12682 library will be used in a future version of networkd to
12683 enable PPPoE communication without an external pppd daemon.
12685 * The busctl tool now understands a new "capture" verb that
12686 works similar to "monitor", but writes a packet capture
12687 trace to STDOUT that can be redirected to a file which is
12688 compatible with libcap's capture file format. This can then
12689 be loaded in Wireshark and similar tools to inspect bus
12692 * The busctl tool now understands a new "tree" verb that shows
12693 the object trees of a specific service on the bus, or of all
12696 * The busctl tool now understands a new "introspect" verb that
12697 shows all interfaces and members of objects on the bus,
12698 including their signature and values. This is particularly
12699 useful to get more information about bus objects shown by
12700 the new "busctl tree" command.
12702 * The busctl tool now understands new verbs "call",
12703 "set-property" and "get-property" for invoking bus method
12704 calls, setting and getting bus object properties in a
12707 * busctl gained a new --augment-creds= argument that controls
12708 whether the tool shall augment credential information it
12709 gets from the bus with data from /proc, in a possibly
12712 * nspawn's --link-journal= switch gained two new values
12713 "try-guest" and "try-host" that work like "guest" and
12714 "host", but do not fail if the host has no persistent
12715 journaling enabled. -j is now equivalent to
12716 --link-journal=try-guest.
12718 * macvlan network devices created by nspawn will now have
12719 stable MAC addresses.
12721 * A new SmackProcessLabel= unit setting has been added, which
12722 controls the SMACK security label processes forked off by
12723 the respective unit shall use.
12725 * If compiled with --enable-xkbcommon, systemd-localed will
12726 verify x11 keymap settings by compiling the given keymap. It
12727 will spew out warnings if the compilation fails. This
12728 requires libxkbcommon to be installed.
12730 * When a coredump is collected, a larger number of metadata
12731 fields is now collected and included in the journal records
12732 created for it. More specifically, control group membership,
12733 environment variables, memory maps, working directory,
12734 chroot directory, /proc/$PID/status, and a list of open file
12735 descriptors is now stored in the log entry.
12737 * The udev hwdb now contains DPI information for mice. For
12740 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
12742 * All systemd programs that read standalone configuration
12743 files in /etc now also support a corresponding series of
12744 .conf.d configuration directories in /etc/, /run/,
12745 /usr/local/lib/, /usr/lib/, and (if configured with
12746 --enable-split-usr) /lib/. In particular, the following
12747 configuration files now have corresponding configuration
12748 directories: system.conf user.conf, logind.conf,
12749 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
12750 resolved.conf, timesyncd.conf, journal-remote.conf, and
12751 journal-upload.conf. Note that distributions should use the
12752 configuration directories in /usr/lib/; the directories in
12753 /etc/ are reserved for the system administrator.
12755 * systemd-rfkill will no longer take the rfkill device name
12756 into account when storing rfkill state on disk, as the name
12757 might be dynamically assigned and not stable. Instead, the
12758 ID_PATH udev variable combined with the rfkill type (wlan,
12759 bluetooth, …) is used.
12761 * A new service systemd-machine-id-commit.service has been
12762 added. When used on systems where /etc is read-only during
12763 boot, and /etc/machine-id is not initialized (but an empty
12764 file), this service will copy the temporary machine ID
12765 created as replacement into /etc after the system is fully
12766 booted up. This is useful for systems that are freshly
12767 installed with a non-initialized machine ID, but should get
12768 a fixed machine ID for subsequent boots.
12770 * networkd's .netdev files now provide a large set of
12771 configuration parameters for VXLAN devices. Similarly, the
12772 bridge port cost parameter is now configurable in .network
12773 files. There's also new support for configuring IP source
12774 routing. networkd .link files gained support for a new
12775 OriginalName= match that is useful to match against the
12776 original interface name the kernel assigned. .network files
12777 may include MTU= and MACAddress= fields for altering the MTU
12778 and MAC address while being connected to a specific network
12781 * The LUKS logic gained supported for configuring
12782 UUID-specific key files. There's also new support for naming
12783 LUKS device from the kernel command line, using the new
12784 luks.name= argument.
12786 * Timer units may now be transiently created via the bus API
12787 (this was previously already available for scope and service
12788 units). In addition it is now possible to create multiple
12789 transient units at the same time with a single bus call. The
12790 "systemd-run" tool has been updated to make use of this for
12791 running commands on a specified time, in at(1)-style.
12793 * tmpfiles gained support for "t" lines, for assigning
12794 extended attributes to files. Among other uses this may be
12795 used to assign SMACK labels to files.
12797 Contributions from: Alin Rauta, Alison Chaiken, Andrej
12798 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
12799 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
12800 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
12801 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
12802 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
12803 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
12804 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
12805 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
12806 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
12807 Michael Biebl, Michael Chapman, Michael Marineau, Michal
12808 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
12809 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
12810 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
12811 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
12812 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
12813 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
12814 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
12816 — Berlin, 2014-12-10
12820 * journalctl gained the new options -t/--identifier= to match
12821 on the syslog identifier (aka "tag"), as well as --utc to
12822 show log timestamps in the UTC timezone. journalctl now also
12823 accepts -n/--lines=all to disable line capping in a pager.
12825 * journalctl gained a new switch, --flush, that synchronously
12826 flushes logs from /run/log/journal to /var/log/journal if
12827 persistent storage is enabled. systemd-journal-flush.service
12828 now waits until the operation is complete.
12830 * Services can notify the manager before they start a reload
12831 (by sending RELOADING=1) or shutdown (by sending
12832 STOPPING=1). This allows the manager to track and show the
12833 internal state of daemons and closes a race condition when
12834 the process is still running but has closed its D-Bus
12837 * Services with Type=oneshot do not have to have any ExecStart
12840 * User units are now loaded also from
12841 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
12842 /run/systemd/user directory that was already previously
12843 supported, but is under the control of the user.
12845 * Job timeouts (i.e. timeouts on the time a job that is
12846 queued stays in the run queue) can now optionally result in
12847 immediate reboot or power-off actions (JobTimeoutAction= and
12848 JobTimeoutRebootArgument=). This is useful on ".target"
12849 units, to limit the maximum time a target remains
12850 undispatched in the run queue, and to trigger an emergency
12851 operation in such a case. This is now used by default to
12852 turn off the system if boot-up (as defined by everything in
12853 basic.target) hangs and does not complete for at least
12854 15min. Also, if power-off or reboot hang for at least 30min
12855 an immediate power-off/reboot operation is triggered. This
12856 functionality is particularly useful to increase reliability
12857 on embedded devices, but also on laptops which might
12858 accidentally get powered on when carried in a backpack and
12859 whose boot stays stuck in a hard disk encryption passphrase
12862 * systemd-logind can be configured to also handle lid switch
12863 events even when the machine is docked or multiple displays
12864 are attached (HandleLidSwitchDocked= option).
12866 * A helper binary and a service have been added which can be
12867 used to resume from hibernation in the initramfs. A
12868 generator will parse the resume= option on the kernel
12869 command line to trigger resume.
12871 * A user console daemon systemd-consoled has been
12872 added. Currently, it is a preview, and will so far open a
12873 single terminal on each session of the user marked as
12874 Desktop=systemd-console.
12876 * Route metrics can be specified for DHCP routes added by
12879 * The SELinux context of socket-activated services can be set
12880 from the information provided by the networking stack
12881 (SELinuxContextFromNet= option).
12883 * Userspace firmware loading support has been removed and
12884 the minimum supported kernel version is thus bumped to 3.7.
12886 * Timeout for udev workers has been increased from 1 to 3
12887 minutes, but a warning will be printed after 1 minute to
12888 help diagnose kernel modules that take a long time to load.
12890 * Udev rules can now remove tags on devices with TAG-="foobar".
12892 * systemd's readahead implementation has been removed. In many
12893 circumstances it didn't give expected benefits even for
12894 rotational disk drives and was becoming less relevant in the
12895 age of SSDs. As none of the developers has been using
12896 rotating media anymore, and nobody stepped up to actively
12897 maintain this component of systemd it has now been removed.
12899 * Swap units can use Options= to specify discard options.
12900 Discard options specified for swaps in /etc/fstab are now
12903 * Docker containers are now detected as a separate type of
12906 * The Password Agent protocol gained support for queries where
12907 the user input is shown, useful e.g. for user names.
12908 systemd-ask-password gained a new --echo option to turn that
12911 * The default sysctl.d/ snippets will now set:
12913 net.core.default_qdisc = fq_codel
12915 This selects Fair Queuing Controlled Delay as the default
12916 queuing discipline for network interfaces. fq_codel helps
12917 fight the network bufferbloat problem. It is believed to be
12918 a good default with no tuning required for most workloads.
12919 Downstream distributions may override this choice. On 10Gbit
12920 servers that do not do forwarding, "fq" may perform better.
12921 Systems without a good clocksource should use "pfifo_fast".
12923 * If kdbus is enabled during build a new option BusPolicy= is
12924 available for service units, that allows locking all service
12925 processes into a stricter bus policy, in order to limit
12926 access to various bus services, or even hide most of them
12927 from the service's view entirely.
12929 * networkctl will now show the .network and .link file
12930 networkd has applied to a specific interface.
12932 * sd-login gained a new API call sd_session_get_desktop() to
12933 query which desktop environment has been selected for a
12936 * UNIX utmp support is now compile-time optional to support
12937 legacy-free systems.
12939 * systemctl gained two new commands "add-wants" and
12940 "add-requires" for pulling in units from specific targets
12943 * If the word "rescue" is specified on the kernel command line
12944 the system will now boot into rescue mode (aka
12945 rescue.target), which was previously available only by
12946 specifying "1" or "systemd.unit=rescue.target" on the kernel
12947 command line. This new kernel command line option nicely
12948 mirrors the already existing "emergency" kernel command line
12951 * New kernel command line options mount.usr=, mount.usrflags=,
12952 mount.usrfstype= have been added that match root=, rootflags=,
12953 rootfstype= but allow mounting a specific file system to
12956 * The $NOTIFY_SOCKET is now also passed to control processes of
12957 services, not only the main process.
12959 * This version reenables support for fsck's -l switch. This
12960 means at least version v2.25 of util-linux is required for
12961 operation, otherwise dead-locks on device nodes may
12962 occur. Again: you need to update util-linux to at least
12963 v2.25 when updating systemd to v217.
12965 * The "multi-seat-x" tool has been removed from systemd, as
12966 its functionality has been integrated into X servers 1.16,
12967 and the tool is hence redundant. It is recommended to update
12968 display managers invoking this tool to simply invoke X
12969 directly from now on, again.
12971 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
12972 message flag has been added for all of systemd's polkit
12973 authenticated method calls has been added. In particular this
12974 now allows optional interactive authorization via polkit for
12975 many of PID1's privileged operations such as unit file
12976 enabling and disabling.
12978 * "udevadm hwdb --update" learnt a new switch "--usr" for
12979 placing the rebuilt hardware database in /usr instead of
12980 /etc. When used only hardware database entries stored in
12981 /usr will be used, and any user database entries in /etc are
12982 ignored. This functionality is useful for vendors to ship a
12983 pre-built database on systems where local configuration is
12984 unnecessary or unlikely.
12986 * Calendar time specifications in .timer units now also
12987 understand the strings "semi-annually", "quarterly" and
12988 "minutely" as shortcuts (in addition to the preexisting
12989 "annually", "hourly", …).
12991 * systemd-tmpfiles will now correctly create files in /dev
12992 at boot which are marked for creation only at boot. It is
12993 recommended to always create static device nodes with 'c!'
12994 and 'b!', so that they are created only at boot and not
12995 overwritten at runtime.
12997 * When the watchdog logic is used for a service (WatchdogSec=)
12998 and the watchdog timeout is hit the service will now be
12999 terminated with SIGABRT (instead of just SIGTERM), in order
13000 to make sure a proper coredump and backtrace is
13001 generated. This ensures that hanging services will result in
13002 similar coredump/backtrace behaviour as services that hit a
13003 segmentation fault.
13005 Contributions from: Andreas Henriksson, Andrei Borzenkov,
13006 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
13007 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
13008 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
13009 Herrmann, David Sommerseth, David Strauss, Emil Renner
13010 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
13011 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
13012 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
13013 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
13014 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
13015 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
13016 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
13017 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
13018 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
13019 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
13020 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
13021 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
13022 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
13023 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
13024 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
13025 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
13028 — Berlin, 2014-10-28
13032 * timedated no longer reads NTP implementation unit names from
13033 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
13034 implementations should add a
13036 Conflicts=systemd-timesyncd.service
13038 to their unit files to take over and replace systemd's NTP
13039 default functionality.
13041 * systemd-sysusers gained a new line type "r" for configuring
13042 which UID/GID ranges to allocate system users/groups
13043 from. Lines of type "u" may now add an additional column
13044 that specifies the home directory for the system user to be
13045 created. Also, systemd-sysusers may now optionally read user
13046 information from STDIN instead of a file. This is useful for
13047 invoking it from RPM preinst scriptlets that need to create
13048 users before the first RPM file is installed since these
13049 files might need to be owned by them. A new
13050 %sysusers_create_inline RPM macro has been introduced to do
13051 just that. systemd-sysusers now updates the shadow files as
13052 well as the user/group databases, which should enhance
13053 compatibility with certain tools like grpck.
13055 * A number of bus APIs of PID 1 now optionally consult polkit to
13056 permit access for otherwise unprivileged clients under certain
13057 conditions. Note that this currently doesn't support
13058 interactive authentication yet, but this is expected to be
13059 added eventually, too.
13061 * /etc/machine-info now has new fields for configuring the
13062 deployment environment of the machine, as well as the
13063 location of the machine. hostnamectl has been updated with
13064 new command to update these fields.
13066 * systemd-timesyncd has been updated to automatically acquire
13067 NTP server information from systemd-networkd, which might
13068 have been discovered via DHCP.
13070 * systemd-resolved now includes a caching DNS stub resolver
13071 and a complete LLMNR name resolution implementation. A new
13072 NSS module "nss-resolve" has been added which can be used
13073 instead of glibc's own "nss-dns" to resolve hostnames via
13074 systemd-resolved. Hostnames, addresses and arbitrary RRs may
13075 be resolved via systemd-resolved D-Bus APIs. In contrast to
13076 the glibc internal resolver systemd-resolved is aware of
13077 multi-homed system, and keeps DNS server and caches separate
13078 and per-interface. Queries are sent simultaneously on all
13079 interfaces that have DNS servers configured, in order to
13080 properly handle VPNs and local LANs which might resolve
13081 separate sets of domain names. systemd-resolved may acquire
13082 DNS server information from systemd-networkd automatically,
13083 which in turn might have discovered them via DHCP. A tool
13084 "systemd-resolve-host" has been added that may be used to
13085 query the DNS logic in resolved. systemd-resolved implements
13086 IDNA and automatically uses IDNA or UTF-8 encoding depending
13087 on whether classic DNS or LLMNR is used as transport. In the
13088 next releases we intend to add a DNSSEC and mDNS/DNS-SD
13089 implementation to systemd-resolved.
13091 * A new NSS module nss-mymachines has been added, that
13092 automatically resolves the names of all local registered
13093 containers to their respective IP addresses.
13095 * A new client tool "networkctl" for systemd-networkd has been
13096 added. It currently is entirely passive and will query
13097 networking configuration from udev, rtnetlink and networkd,
13098 and present it to the user in a very friendly
13099 way. Eventually, we hope to extend it to become a full
13100 control utility for networkd.
13102 * .socket units gained a new DeferAcceptSec= setting that
13103 controls the kernels' TCP_DEFER_ACCEPT sockopt for
13104 TCP. Similarly, support for controlling TCP keep-alive
13105 settings has been added (KeepAliveTimeSec=,
13106 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
13107 turning off Nagle's algorithm on TCP has been added
13110 * logind learned a new session type "web", for use in projects
13111 like Cockpit which register web clients as PAM sessions.
13113 * timer units with at least one OnCalendar= setting will now
13114 be started only after time-sync.target has been
13115 reached. This way they will not elapse before the system
13116 clock has been corrected by a local NTP client or
13117 similar. This is particular useful on RTC-less embedded
13118 machines, that come up with an invalid system clock.
13120 * systemd-nspawn's --network-veth= switch should now result in
13121 stable MAC addresses for both the outer and the inner side
13124 * systemd-nspawn gained a new --volatile= switch for running
13125 container instances with /etc or /var unpopulated.
13127 * The kdbus client code has been updated to use the new Linux
13128 3.17 memfd subsystem instead of the old kdbus-specific one.
13130 * systemd-networkd's DHCP client and server now support
13131 FORCERENEW. There are also new configuration options to
13132 configure the vendor client identifier and broadcast mode
13135 * systemd will no longer inform the kernel about the current
13136 timezone, as this is necessarily incorrect and racy as the
13137 kernel has no understanding of DST and similar
13138 concepts. This hence means FAT timestamps will be always
13139 considered UTC, similar to what Android is already
13140 doing. Also, when the RTC is configured to the local time
13141 (rather than UTC) systemd will never synchronize back to it,
13142 as this might confuse Windows at a later boot.
13144 * systemd-analyze gained a new command "verify" for offline
13145 validation of unit files.
13147 * systemd-networkd gained support for a couple of additional
13148 settings for bonding networking setups. Also, the metric for
13149 statically configured routes may now be configured. For
13150 network interfaces where this is appropriate the peer IP
13151 address may now be configured.
13153 * systemd-networkd's DHCP client will no longer request
13154 broadcasting by default, as this tripped up some networks.
13155 For hardware where broadcast is required the feature should
13156 be switched back on using RequestBroadcast=yes.
13158 * systemd-networkd will now set up IPv4LL addresses (when
13159 enabled) even if DHCP is configured successfully.
13161 * udev will now default to respect network device names given
13162 by the kernel when the kernel indicates that these are
13163 predictable. This behavior can be tweaked by changing
13164 NamePolicy= in the relevant .link file.
13166 * A new library systemd-terminal has been added that
13167 implements full TTY stream parsing and rendering. This
13168 library is supposed to be used later on for implementing a
13169 full userspace VT subsystem, replacing the current kernel
13172 * A new tool systemd-journal-upload has been added to push
13173 journal data to a remote system running
13174 systemd-journal-remote.
13176 * journald will no longer forward all local data to another
13177 running syslog daemon. This change has been made because
13178 rsyslog (which appears to be the most commonly used syslog
13179 implementation these days) no longer makes use of this, and
13180 instead pulls the data out of the journal on its own. Since
13181 forwarding the messages to a non-existent syslog server is
13182 more expensive than we assumed we have now turned this
13183 off. If you run a syslog server that is not a recent rsyslog
13184 version, you have to turn this option on again
13185 (ForwardToSyslog= in journald.conf).
13187 * journald now optionally supports the LZ4 compressor for
13188 larger journal fields. This compressor should perform much
13189 better than XZ which was the previous default.
13191 * machinectl now shows the IP addresses of local containers,
13192 if it knows them, plus the interface name of the container.
13194 * A new tool "systemd-escape" has been added that makes it
13195 easy to escape strings to build unit names and similar.
13197 * sd_notify() messages may now include a new ERRNO= field
13198 which is parsed and collected by systemd and shown among the
13199 "systemctl status" output for a service.
13201 * A new component "systemd-firstboot" has been added that
13202 queries the most basic systemd information (timezone,
13203 hostname, root password) interactively on first
13204 boot. Alternatively it may also be used to provision these
13205 things offline on OS images installed into directories.
13207 * The default sysctl.d/ snippets will now set
13209 net.ipv4.conf.default.promote_secondaries=1
13211 This has the benefit of no flushing secondary IP addresses
13212 when primary addresses are removed.
13214 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
13215 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
13216 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
13217 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
13218 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
13219 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
13220 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
13221 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
13222 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
13223 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
13224 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
13225 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
13226 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
13227 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
13228 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
13230 — Berlin, 2014-08-19
13234 * A new tool systemd-sysusers has been added. This tool
13235 creates system users and groups in /etc/passwd and
13236 /etc/group, based on static declarative system user/group
13237 definitions in /usr/lib/sysusers.d/. This is useful to
13238 enable factory resets and volatile systems that boot up with
13239 an empty /etc directory, and thus need system users and
13240 groups created during early boot. systemd now also ships
13241 with two default sysusers.d/ files for the most basic
13242 users and groups systemd and the core operating system
13245 * A new tmpfiles snippet has been added that rebuilds the
13246 essential files in /etc on boot, should they be missing.
13248 * A directive for ensuring automatic clean-up of
13249 /var/cache/man/ has been removed from the default
13250 configuration. This line should now be shipped by the man
13251 implementation. The necessary change has been made to the
13252 man-db implementation. Note that you need to update your man
13253 implementation to one that ships this line, otherwise no
13254 automatic clean-up of /var/cache/man will take place.
13256 * A new condition ConditionNeedsUpdate= has been added that
13257 may conditionalize services to only run when /etc or /var
13258 are "older" than the vendor operating system resources in
13259 /usr. This is useful for reconstructing or updating /etc
13260 after an offline update of /usr or a factory reset, on the
13261 next reboot. Services that want to run once after such an
13262 update or reset should use this condition and order
13263 themselves before the new systemd-update-done.service, which
13264 will mark the two directories as fully updated. A number of
13265 service files have been added making use of this, to rebuild
13266 the udev hardware database, the journald message catalog and
13267 dynamic loader cache (ldconfig). The systemd-sysusers tool
13268 described above also makes use of this now. With this in
13269 place it is now possible to start up a minimal operating
13270 system with /etc empty cleanly. For more information on the
13271 concepts involved see this recent blog story:
13273 https://0pointer.de/blog/projects/stateless.html
13275 * A new system group "input" has been introduced, and all
13276 input device nodes get this group assigned. This is useful
13277 for system-level software to get access to input devices. It
13278 complements what is already done for "audio" and "video".
13280 * systemd-networkd learnt minimal DHCPv4 server support in
13281 addition to the existing DHCPv4 client support. It also
13282 learnt DHCPv6 client and IPv6 Router Solicitation client
13283 support. The DHCPv4 client gained support for static routes
13284 passed in from the server. Note that the [DHCPv4] section
13285 known in older systemd-networkd versions has been renamed to
13286 [DHCP] and is now also used by the DHCPv6 client. Existing
13287 .network files using settings of this section should be
13288 updated, though compatibility is maintained. Optionally, the
13289 client hostname may now be sent to the DHCP server.
13291 * networkd gained support for vxlan virtual networks as well
13292 as tun/tap and dummy devices.
13294 * networkd gained support for automatic allocation of address
13295 ranges for interfaces from a system-wide pool of
13296 addresses. This is useful for dynamically managing a large
13297 number of interfaces with a single network configuration
13298 file. In particular this is useful to easily assign
13299 appropriate IP addresses to the veth links of a large number
13300 of nspawn instances.
13302 * RPM macros for processing sysusers, sysctl and binfmt
13303 drop-in snippets at package installation time have been
13306 * The /etc/os-release file should now be placed in
13307 /usr/lib/os-release. The old location is automatically
13308 created as symlink. /usr/lib is the more appropriate
13309 location of this file, since it shall actually describe the
13310 vendor operating system shipped in /usr, and not the
13311 configuration stored in /etc.
13313 * .mount units gained a new boolean SloppyOptions= setting
13314 that maps to mount(8)'s -s option which enables permissive
13315 parsing of unknown mount options.
13317 * tmpfiles learnt a new "L+" directive which creates a symlink
13318 but (unlike "L") deletes a pre-existing file first, should
13319 it already exist and not already be the correct
13320 symlink. Similarly, "b+", "c+" and "p+" directives have been
13321 added as well, which create block and character devices, as
13322 well as fifos in the filesystem, possibly removing any
13323 pre-existing files of different types.
13325 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
13326 'argument' field (which so far specified the source to
13327 symlink/copy the files from) is now optional. If omitted the
13328 same file os copied from /usr/share/factory/ suffixed by the
13329 full destination path. This is useful for populating /etc
13330 with essential files, by copying them from vendor defaults
13331 shipped in /usr/share/factory/etc.
13333 * A new command "systemctl preset-all" has been added that
13334 applies the service preset settings to all installed unit
13335 files. A new switch --preset-mode= has been added that
13336 controls whether only enable or only disable operations
13339 * A new command "systemctl is-system-running" has been added
13340 that allows checking the overall state of the system, for
13341 example whether it is fully up and running.
13343 * When the system boots up with an empty /etc, the equivalent
13344 to "systemctl preset-all" is executed during early boot, to
13345 make sure all default services are enabled after a factory
13348 * systemd now contains a minimal preset file that enables the
13349 most basic services systemd ships by default.
13351 * Unit files' [Install] section gained a new DefaultInstance=
13352 field for defining the default instance to create if a
13353 template unit is enabled with no instance specified.
13355 * A new passive target cryptsetup-pre.target has been added
13356 that may be used by services that need to make they run and
13357 finish before the first LUKS cryptographic device is set up.
13359 * The /dev/loop-control and /dev/btrfs-control device nodes
13360 are now owned by the "disk" group by default, opening up
13361 access to this group.
13363 * systemd-coredump will now automatically generate a
13364 stack trace of all core dumps taking place on the system,
13365 based on elfutils' libdw library. This stack trace is logged
13368 * systemd-coredump may now optionally store coredumps directly
13369 on disk (in /var/lib/systemd/coredump, possibly compressed),
13370 instead of storing them unconditionally in the journal. This
13371 mode is the new default. A new configuration file
13372 /etc/systemd/coredump.conf has been added to configure this
13373 and other parameters of systemd-coredump.
13375 * coredumpctl gained a new "info" verb to show details about a
13376 specific coredump. A new switch "-1" has also been added
13377 that makes sure to only show information about the most
13378 recent entry instead of all entries. Also, as the tool is
13379 generally useful now the "systemd-" prefix of the binary
13380 name has been removed. Distributions that want to maintain
13381 compatibility with the old name should add a symlink from
13382 the old name to the new name.
13384 * journald's SplitMode= now defaults to "uid". This makes sure
13385 that unprivileged users can access their own coredumps with
13386 coredumpctl without restrictions.
13388 * New kernel command line options "systemd.wants=" (for
13389 pulling an additional unit during boot), "systemd.mask="
13390 (for masking a specific unit for the boot), and
13391 "systemd.debug-shell" (for enabling the debug shell on tty9)
13392 have been added. This is implemented in the new generator
13393 "systemd-debug-generator".
13395 * systemd-nspawn will now by default filter a couple of
13396 syscalls for containers, among them those required for
13397 kernel module loading, direct x86 IO port access, swap
13398 management, and kexec. Most importantly though
13399 open_by_handle_at() is now prohibited for containers,
13400 closing a hole similar to a recently discussed vulnerability
13401 in docker regarding access to files on file hierarchies the
13402 container should normally not have access to. Note that, for
13403 nspawn, we generally make no security claims anyway (and
13404 this is explicitly documented in the man page), so this is
13405 just a fix for one of the most obvious problems.
13407 * A new man page file-hierarchy(7) has been added that
13408 contains a minimized, modernized version of the file system
13409 layout systemd expects, similar in style to the FHS
13410 specification or hier(5). A new tool systemd-path(1) has
13411 been added to query many of these paths for the local
13414 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
13415 longer done. Since the directory now has a per-user size
13416 limit, and is cleaned on logout this appears unnecessary,
13417 in particular since this now brings the lifecycle of this
13418 directory closer in line with how IPC objects are handled.
13420 * systemd.pc now exports a number of additional directories,
13421 including $libdir (which is useful to identify the library
13422 path for the primary architecture of the system), and a
13423 couple of drop-in directories.
13425 * udev's predictable network interface names now use the dev_port
13426 sysfs attribute, introduced in linux 3.15 instead of dev_id to
13427 distinguish between ports of the same PCI function. dev_id should
13428 only be used for ports using the same HW address, hence the need
13431 * machined has been updated to export the OS version of a
13432 container (read from /etc/os-release and
13433 /usr/lib/os-release) on the bus. This is now shown in
13434 "machinectl status" for a machine.
13436 * A new service setting RestartForceExitStatus= has been
13437 added. If configured to a set of exit signals or process
13438 return values, the service will be restarted when the main
13439 daemon process exits with any of them, regardless of the
13442 * systemctl's -H switch for connecting to remote systemd
13443 machines has been extended so that it may be used to
13444 directly connect to a specific container on the
13445 host. "systemctl -H root@foobar:waldi" will now connect as
13446 user "root" to host "foobar", and then proceed directly to
13447 the container named "waldi". Note that currently you have to
13448 authenticate as user "root" for this to work, as entering
13449 containers is a privileged operation.
13451 Contributions from: Andreas Henriksson, Benjamin Steinwender,
13452 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
13453 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
13454 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
13455 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
13456 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
13457 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
13458 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
13459 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
13460 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
13461 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
13462 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
13464 — Berlin, 2014-07-03
13468 * As an experimental feature, udev now tries to lock the
13469 disk device node (flock(LOCK_SH|LOCK_NB)) while it
13470 executes events for the disk or any of its partitions.
13471 Applications like partitioning programs can lock the
13472 disk device node (flock(LOCK_EX)) and claim temporary
13473 device ownership that way; udev will entirely skip all event
13474 handling for this disk and its partitions. If the disk
13475 was opened for writing, the close will trigger a partition
13476 table rescan in udev's "watch" facility, and if needed
13477 synthesize "change" events for the disk and all its partitions.
13478 This is now unconditionally enabled, and if it turns out to
13479 cause major problems, we might turn it on only for specific
13480 devices, or might need to disable it entirely. Device Mapper
13481 devices are excluded from this logic.
13483 * We temporarily dropped the "-l" switch for fsck invocations,
13484 since they collide with the flock() logic above. util-linux
13485 upstream has been changed already to avoid this conflict,
13486 and we will re-add "-l" as soon as util-linux with this
13487 change has been released.
13489 * The dependency on libattr has been removed. Since a long
13490 time, the extended attribute calls have moved to glibc, and
13491 libattr is thus unnecessary.
13493 * Virtualization detection works without privileges now. This
13494 means the systemd-detect-virt binary no longer requires
13495 CAP_SYS_PTRACE file capabilities, and our daemons can run
13496 with fewer privileges.
13498 * systemd-networkd now runs under its own "systemd-network"
13499 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
13500 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
13501 loses the ability to write to files owned by root this way.
13503 * Similarly, systemd-resolved now runs under its own
13504 "systemd-resolve" user with no capabilities remaining.
13506 * Similarly, systemd-bus-proxyd now runs under its own
13507 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
13509 * systemd-networkd gained support for setting up "veth"
13510 virtual Ethernet devices for container connectivity, as well
13511 as GRE and VTI tunnels.
13513 * systemd-networkd will no longer automatically attempt to
13514 manually load kernel modules necessary for certain tunnel
13515 transports. Instead, it is assumed the kernel loads them
13516 automatically when required. This only works correctly on
13517 very new kernels. On older kernels, please consider adding
13518 the kernel modules to /etc/modules-load.d/ as a work-around.
13520 * The resolv.conf file systemd-resolved generates has been
13521 moved to /run/systemd/resolve/. If you have a symlink from
13522 /etc/resolv.conf, it might be necessary to correct it.
13524 * Two new service settings, ProtectHome= and ProtectSystem=,
13525 have been added. When enabled, they will make the user data
13526 (such as /home) inaccessible or read-only and the system
13527 (such as /usr) read-only, for specific services. This allows
13528 very light-weight per-service sandboxing to avoid
13529 modifications of user data or system files from
13530 services. These two new switches have been enabled for all
13531 of systemd's long-running services, where appropriate.
13533 * Socket units gained new SocketUser= and SocketGroup=
13534 settings to set the owner user and group of AF_UNIX sockets
13535 and FIFOs in the file system.
13537 * Socket units gained a new RemoveOnStop= setting. If enabled,
13538 all FIFOS and sockets in the file system will be removed
13539 when the specific socket unit is stopped.
13541 * Socket units gained a new Symlinks= setting. It takes a list
13542 of symlinks to create to file system sockets or FIFOs
13543 created by the specific Unix sockets. This is useful to
13544 manage symlinks to socket nodes with the same lifecycle as
13547 * The /dev/log socket and /dev/initctl FIFO have been moved to
13548 /run, and have been replaced by symlinks. This allows
13549 connecting to these facilities even if PrivateDevices=yes is
13550 used for a service (which makes /dev/log itself unavailable,
13551 but /run is left). This also has the benefit of ensuring
13552 that /dev only contains device nodes, directories and
13553 symlinks, and nothing else.
13555 * sd-daemon gained two new calls sd_pid_notify() and
13556 sd_pid_notifyf(). They are similar to sd_notify() and
13557 sd_notifyf(), but allow overriding of the source PID of
13558 notification messages if permissions permit this. This is
13559 useful to send notify messages on behalf of a different
13560 process (for example, the parent process). The
13561 systemd-notify tool has been updated to make use of this
13562 when sending messages (so that notification messages now
13563 originate from the shell script invoking systemd-notify and
13564 not the systemd-notify process itself. This should minimize
13565 a race where systemd fails to associate notification
13566 messages to services when the originating process already
13569 * A new "on-abnormal" setting for Restart= has been added. If
13570 set, it will result in automatic restarts on all "abnormal"
13571 reasons for a process to exit, which includes unclean
13572 signals, core dumps, timeouts and watchdog timeouts, but
13573 does not include clean and unclean exit codes or clean
13574 signals. Restart=on-abnormal is an alternative for
13575 Restart=on-failure for services that shall be able to
13576 terminate and avoid restarts on certain errors, by
13577 indicating so with an unclean exit code. Restart=on-failure
13578 or Restart=on-abnormal is now the recommended setting for
13579 all long-running services.
13581 * If the InaccessibleDirectories= service setting points to a
13582 mount point (or if there are any submounts contained within
13583 it), it is now attempted to completely unmount it, to make
13584 the file systems truly unavailable for the respective
13587 * The ReadOnlyDirectories= service setting and
13588 systemd-nspawn's --read-only parameter are now recursively
13589 applied to all submounts, too.
13591 * Mount units may now be created transiently via the bus APIs.
13593 * The support for SysV and LSB init scripts has been removed
13594 from the systemd daemon itself. Instead, it is now
13595 implemented as a generator that creates native systemd units
13596 from these scripts when needed. This enables us to remove a
13597 substantial amount of legacy code from PID 1, following the
13598 fact that many distributions only ship a very small number
13599 of LSB/SysV init scripts nowadays.
13601 * Privileged Xen (dom0) domains are not considered
13602 virtualization anymore by the virtualization detection
13603 logic. After all, they generally have unrestricted access to
13604 the hardware and usually are used to manage the unprivileged
13607 * systemd-tmpfiles gained a new "C" line type, for copying
13608 files or entire directories.
13610 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
13611 lines. So far, they have been non-globbing versions of the
13612 latter, and have thus been redundant. In future, it is
13613 recommended to only use "z". "m" has hence been removed
13614 from the documentation, even though it stays supported.
13616 * A tmpfiles snippet to recreate the most basic structure in
13617 /var has been added. This is enough to create the /var/run →
13618 /run symlink and create a couple of structural
13619 directories. This allows systems to boot up with an empty or
13620 volatile /var. Of course, while with this change, the core OS
13621 now is capable with dealing with a volatile /var, not all
13622 user services are ready for it. However, we hope that sooner
13623 or later, many service daemons will be changed upstream so
13624 that they are able to automatically create their necessary
13625 directories in /var at boot, should they be missing. This is
13626 the first step to allow state-less systems that only require
13627 the vendor image for /usr to boot.
13629 * systemd-nspawn has gained a new --tmpfs= switch to mount an
13630 empty tmpfs instance to a specific directory. This is
13631 particularly useful for making use of the automatic
13632 reconstruction of /var (see above), by passing --tmpfs=/var.
13634 * Access modes specified in tmpfiles snippets may now be
13635 prefixed with "~", which indicates that they shall be masked
13636 by whether the existing file or directory is currently
13637 writable, readable or executable at all. Also, if specified,
13638 the sgid/suid/sticky bits will be masked for all
13641 * A new passive target unit "network-pre.target" has been
13642 added which is useful for services that shall run before any
13643 network is configured, for example firewall scripts.
13645 * The "floppy" group that previously owned the /dev/fd*
13646 devices is no longer used. The "disk" group is now used
13647 instead. Distributions should probably deprecate usage of
13650 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
13651 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
13652 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
13653 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
13654 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
13655 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
13658 — Berlin, 2014-06-11
13662 * A new "systemd-timesyncd" daemon has been added for
13663 synchronizing the system clock across the network. It
13664 implements an SNTP client. In contrast to NTP
13665 implementations such as chrony or the NTP reference server,
13666 this only implements a client side, and does not bother with
13667 the full NTP complexity, focusing only on querying time from
13668 one remote server and synchronizing the local clock to
13669 it. Unless you intend to serve NTP to networked clients or
13670 want to connect to local hardware clocks, this simple NTP
13671 client should be more than appropriate for most
13672 installations. The daemon runs with minimal privileges, and
13673 has been hooked up with networkd to only operate when
13674 network connectivity is available. The daemon saves the
13675 current clock to disk every time a new NTP sync has been
13676 acquired, and uses this to possibly correct the system clock
13677 early at bootup, in order to accommodate for systems that
13678 lack an RTC such as the Raspberry Pi and embedded devices,
13679 and to make sure that time monotonically progresses on these
13680 systems, even if it is not always correct. To make use of
13681 this daemon, a new system user and group "systemd-timesync"
13682 needs to be created on installation of systemd.
13684 * The queue "seqnum" interface of libudev has been disabled, as
13685 it was generally incompatible with device namespacing as
13686 sequence numbers of devices go "missing" if the devices are
13687 part of a different namespace.
13689 * "systemctl list-timers" and "systemctl list-sockets" gained
13690 a --recursive switch for showing units of these types also
13691 for all local containers, similar in style to the already
13692 supported --recursive switch for "systemctl list-units".
13694 * A new RebootArgument= setting has been added for service
13695 units, which may be used to specify a kernel reboot argument
13696 to use when triggering reboots with StartLimitAction=.
13698 * A new FailureAction= setting has been added for service
13699 units which may be used to specify an operation to trigger
13700 when a service fails. This works similarly to
13701 StartLimitAction=, but unlike it, controls what is done
13702 immediately rather than only after several attempts to
13703 restart the service in question.
13705 * hostnamed got updated to also expose the kernel name,
13706 release, and version on the bus. This is useful for
13707 executing commands like hostnamectl with the -H switch.
13708 systemd-analyze makes use of this to properly display
13709 details when running non-locally.
13711 * The bootchart tool can now show cgroup information in the
13712 graphs it generates.
13714 * The CFS CPU quota cgroup attribute is now exposed for
13715 services. The new CPUQuota= switch has been added for this
13716 which takes a percentage value. Setting this will have the
13717 result that a service may never get more CPU time than the
13718 specified percentage, even if the machine is otherwise idle.
13720 * systemd-networkd learned IPIP and SIT tunnel support.
13722 * LSB init scripts exposing a dependency on $network will now
13723 get a dependency on network-online.target rather than simply
13724 network.target. This should bring LSB handling closer to
13725 what it was on SysV systems.
13727 * A new fsck.repair= kernel option has been added to control
13728 how fsck shall deal with unclean file systems at boot.
13730 * The (.ini) configuration file parser will now silently ignore
13731 sections whose names begin with "X-". This may be used to maintain
13732 application-specific extension sections in unit files.
13734 * machined gained a new API to query the IP addresses of
13735 registered containers. "machinectl status" has been updated
13736 to show these addresses in its output.
13738 * A new call sd_uid_get_display() has been added to the
13739 sd-login APIs for querying the "primary" session of a
13740 user. The "primary" session of the user is elected from the
13741 user's sessions and generally a graphical session is
13742 preferred over a text one.
13744 * A minimal systemd-resolved daemon has been added. It
13745 currently simply acts as a companion to systemd-networkd and
13746 manages resolv.conf based on per-interface DNS
13747 configuration, possibly supplied via DHCP. In the long run
13748 we hope to extend this into a local DNSSEC enabled DNS and
13751 * The systemd-networkd-wait-online tool is now enabled by
13752 default. It will delay network-online.target until a network
13753 connection has been configured. The tool primarily integrates
13754 with networkd, but will also make a best effort to make sense
13755 of network configuration performed in some other way.
13757 * Two new service options StartupCPUShares= and
13758 StartupBlockIOWeight= have been added that work similarly to
13759 CPUShares= and BlockIOWeight= however only apply during
13760 system startup. This is useful to prioritize certain services
13761 differently during bootup than during normal runtime.
13763 * hostnamed has been changed to prefer the statically
13764 configured hostname in /etc/hostname (unless set to
13765 'localhost' or empty) over any dynamic one supplied by
13766 dhcp. With this change, the rules for picking the hostname
13767 match more closely the rules of other configuration settings
13768 where the local administrator's configuration in /etc always
13769 overrides any other settings.
13771 Contributions from: Ali H. Caliskan, Alison Chaiken, Bas van
13772 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
13773 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
13774 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
13775 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
13776 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
13777 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
13778 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
13779 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
13780 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
13781 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
13782 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
13783 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
13784 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
13785 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
13786 Lindskog, WaLyong Cho, Will Woods, Zbigniew
13789 — Beijing, 2014-05-28
13793 * When restoring the screen brightness at boot, stay away from
13794 the darkest setting or from the lowest 5% of the available
13795 range, depending on which is the larger value of both. This
13796 should effectively protect the user from rebooting into a
13797 black screen, should the brightness have been set to minimum
13800 * sd-login gained a new sd_machine_get_class() call to
13801 determine the class ("vm" or "container") of a machine
13802 registered with machined.
13804 * sd-login gained new calls
13805 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
13806 to query the identity of the peer of a local AF_UNIX
13807 connection. They operate similarly to their sd_pid_get_xyz()
13810 * PID 1 will now maintain a system-wide system state engine
13811 with the states "starting", "running", "degraded",
13812 "maintenance", "stopping". These states are bound to system
13813 startup, normal runtime, runtime with at least one failed
13814 service, rescue/emergency mode and system shutdown. This
13815 state is shown in the "systemctl status" output when no unit
13816 name is passed. It is useful to determine system state, in
13817 particularly when doing so for many systems or containers at
13820 * A new command "list-machines" has been added to "systemctl"
13821 that lists all local OS containers and shows their system
13822 state (see above), if systemd runs inside of them.
13824 * systemctl gained a new "-r" switch to recursively enumerate
13825 units on all local containers, when used with the
13826 "list-unit" command (which is the default one that is
13827 executed when no parameters are specified).
13829 * The GPT automatic partition discovery logic will now honour
13830 two GPT partition flags: one may be set on a partition to
13831 cause it to be mounted read-only, and the other may be set
13832 on a partition to ignore it during automatic discovery.
13834 * Two new GPT type UUIDs have been added for automatic root
13835 partition discovery, for 32-bit and 64-bit ARM. This is not
13836 particularly useful for discovering the root directory on
13837 these architectures during bare-metal boots (since UEFI is
13838 not common there), but still very useful to allow booting of
13839 ARM disk images in nspawn with the -i option.
13841 * MAC addresses of interfaces created with nspawn's
13842 --network-interface= switch will now be generated from the
13843 machine name, and thus be stable between multiple invocations
13846 * logind will now automatically remove all IPC objects owned
13847 by a user if she or he fully logs out. This makes sure that
13848 users who are logged out cannot continue to consume IPC
13849 resources. This covers SysV memory, semaphores and message
13850 queues as well as POSIX shared memory and message
13851 queues. Traditionally, SysV and POSIX IPC had no lifecycle
13852 limits. With this functionality, that is corrected. This may
13853 be turned off by using the RemoveIPC= switch of logind.conf.
13855 * The systemd-machine-id-setup and tmpfiles tools gained a
13856 --root= switch to operate on a specific root directory,
13859 * journald can now forward logged messages to the TTYs of all
13860 logged in users ("wall"). This is the default for all
13861 emergency messages now.
13863 * A new tool systemd-journal-remote has been added to stream
13864 journal log messages across the network.
13866 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
13867 controller trees are mounted into it. Note that the
13868 directories mounted beneath it are not read-only. This is a
13869 security measure and is particularly useful because glibc
13870 actually includes a search logic to pick any tmpfs it can
13871 find to implement shm_open() if /dev/shm is not available
13872 (which it might very well be in namespaced setups).
13874 * machinectl gained a new "poweroff" command to cleanly power
13875 down a local OS container.
13877 * The PrivateDevices= unit file setting will now also drop the
13878 CAP_MKNOD capability from the capability bound set, and
13879 imply DevicePolicy=closed.
13881 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
13882 comprehensively on all long-running systemd services where
13883 this is appropriate.
13885 * systemd-udevd will now run in a disassociated mount
13886 namespace. To mount directories from udev rules, make sure to
13887 pull in mount units via SYSTEMD_WANTS properties.
13889 * The kdbus support gained support for uploading policy into
13890 the kernel. sd-bus gained support for creating "monitoring"
13891 connections that can eavesdrop into all bus communication
13892 for debugging purposes.
13894 * Timestamps may now be specified in seconds since the UNIX
13895 epoch Jan 1st, 1970 by specifying "@" followed by the value
13898 * Native tcpwrap support in systemd has been removed. tcpwrap
13899 is old code, not really maintained anymore and has serious
13900 shortcomings, and better options such as firewalls
13901 exist. For setups that require tcpwrap usage, please
13902 consider invoking your socket-activated service via tcpd,
13903 like on traditional inetd.
13905 * A new system.conf configuration option
13906 DefaultTimerAccuracySec= has been added that controls the
13907 default AccuracySec= setting of .timer units.
13909 * Timer units gained a new WakeSystem= switch. If enabled,
13910 timers configured this way will cause the system to resume
13911 from system suspend (if the system supports that, which most
13914 * Timer units gained a new Persistent= switch. If enabled,
13915 timers configured this way will save to disk when they have
13916 been last triggered. This information is then used on next
13917 reboot to possible execute overdue timer events, that
13918 could not take place because the system was powered off.
13919 This enables simple anacron-like behaviour for timer units.
13921 * systemctl's "list-timers" will now also list the time a
13922 timer unit was last triggered in addition to the next time
13923 it will be triggered.
13925 * systemd-networkd will now assign predictable IPv4LL
13926 addresses to its local interfaces.
13928 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
13929 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
13930 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
13931 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
13932 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
13933 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
13934 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
13935 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
13938 — Berlin, 2014-03-25
13942 * A new unit file setting RestrictAddressFamilies= has been
13943 added to restrict which socket address families unit
13944 processes gain access to. This takes address family names
13945 like "AF_INET" or "AF_UNIX", and is useful to minimize the
13946 attack surface of services via exotic protocol stacks. This
13947 is built on seccomp system call filters.
13949 * Two new unit file settings RuntimeDirectory= and
13950 RuntimeDirectoryMode= have been added that may be used to
13951 manage a per-daemon runtime directories below /run. This is
13952 an alternative for setting up directory permissions with
13953 tmpfiles snippets, and has the advantage that the runtime
13954 directory's lifetime is bound to the daemon runtime and that
13955 the daemon starts up with an empty directory each time. This
13956 is particularly useful when writing services that drop
13957 privileges using the User= or Group= setting.
13959 * The DeviceAllow= unit setting now supports globbing for
13960 matching against device group names.
13962 * The systemd configuration file system.conf gained new
13963 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
13964 DefaultMemoryAccounting= to globally turn on/off accounting
13965 for specific resources (cgroups) for all units. These
13966 settings may still be overridden individually in each unit
13969 * systemd-gpt-auto-generator is now able to discover /srv and
13970 root partitions in addition to /home and swap partitions. It
13971 also supports LUKS-encrypted partitions now. With this in
13972 place, automatic discovery of partitions to mount following
13973 the Discoverable Partitions Specification
13974 (https://systemd.io/DISCOVERABLE_PARTITIONS/)
13975 is now a lot more complete. This allows booting without
13976 /etc/fstab and without root= on the kernel command line on
13977 systems prepared appropriately.
13979 * systemd-nspawn gained a new --image= switch which allows
13980 booting up disk images and Linux installations on any block
13981 device that follow the Discoverable Partitions Specification
13982 (see above). This means that installations made with
13983 appropriately updated installers may now be started and
13984 deployed using container managers, completely
13985 unmodified. (We hope that libvirt-lxc will add support for
13986 this feature soon, too.)
13988 * systemd-nspawn gained a new --network-macvlan= setting to
13989 set up a private macvlan interface for the
13990 container. Similarly, systemd-networkd gained a new
13991 Kind=macvlan setting in .netdev files.
13993 * systemd-networkd now supports configuring local addresses
13996 * A new tool systemd-network-wait-online has been added to
13997 synchronously wait for network connectivity using
14000 * The sd-bus.h bus API gained a new sd_bus_track object for
14001 tracking the lifecycle of bus peers. Note that sd-bus.h is
14002 still not a public API though (unless you specify
14003 --enable-kdbus on the configure command line, which however
14004 voids your warranty and you get no API stability guarantee).
14006 * The $XDG_RUNTIME_DIR runtime directories for each user are
14007 now individual tmpfs instances, which has the benefit of
14008 introducing separate pools for each user, with individual
14009 size limits, and thus making sure that unprivileged clients
14010 can no longer negatively impact the system or other users by
14011 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
14012 RuntimeDirectorySize= has been introduced that allows
14013 controlling the default size limit for all users. It
14014 defaults to 10% of the available physical memory. This is no
14015 replacement for quotas on tmpfs though (which the kernel
14016 still does not support), as /dev/shm and /tmp are still
14017 shared resources used by both the system and unprivileged
14020 * logind will now automatically turn off automatic suspending
14021 on laptop lid close when more than one display is
14022 connected. This was previously expected to be implemented
14023 individually in desktop environments (such as GNOME),
14024 however has been added to logind now, in order to fix a
14025 boot-time race where a desktop environment might not have
14026 been started yet and thus not been able to take an inhibitor
14027 lock at the time where logind already suspends the system
14028 due to a closed lid.
14030 * logind will now wait at least 30s after each system
14031 suspend/resume cycle, and 3min after system boot before
14032 suspending the system due to a closed laptop lid. This
14033 should give USB docking stations and similar enough time to
14034 be probed and configured after system resume and boot in
14035 order to then act as suspend blocker.
14037 * systemd-run gained a new --property= setting which allows
14038 initialization of resource control properties (and others)
14039 for the created scope or service unit. Example: "systemd-run
14040 --property=BlockIOWeight=10 updatedb" may be used to run
14041 updatedb at a low block IO scheduling weight.
14043 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
14044 now also work in --scope mode.
14046 * When systemd is compiled with kdbus support, basic support
14047 for enforced policies is now in place. (Note that enabling
14048 kdbus still voids your warranty and no API compatibility
14049 promises are made.)
14051 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
14052 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
14053 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
14054 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
14055 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
14056 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
14057 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
14058 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
14059 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
14060 Zbigniew Jędrzejewski-Szmek
14062 — Berlin, 2014-03-12
14066 * systemd will now relabel /dev after loading the SMACK policy
14067 according to SMACK rules.
14069 * A new unit file option AppArmorProfile= has been added to
14070 set the AppArmor profile for the processes of a unit.
14072 * A new condition check ConditionArchitecture= has been added
14073 to conditionalize units based on the system architecture, as
14074 reported by uname()'s "machine" field.
14076 * systemd-networkd now supports matching on the system
14077 virtualization, architecture, kernel command line, hostname
14080 * logind is now a lot more aggressive when suspending the
14081 machine due to a closed laptop lid. Instead of acting only
14082 on the lid close action, it will continuously watch the lid
14083 status and act on it. This is useful for laptops where the
14084 power button is on the outside of the chassis so that it can
14085 be reached without opening the lid (such as the Lenovo
14086 Yoga). On those machines, logind will now immediately
14087 re-suspend the machine if the power button has been
14088 accidentally pressed while the laptop was suspended and in a
14089 backpack or similar.
14091 * logind will now watch SW_DOCK switches and inhibit reaction
14092 to the lid switch if it is pressed. This means that logind
14093 will not suspend the machine anymore if the lid is closed
14094 and the system is docked, if the laptop supports SW_DOCK
14095 notifications via the input layer. Note that ACPI docking
14096 stations do not generate this currently. Also note that this
14097 logic is usually not fully sufficient and Desktop
14098 Environments should take a lid switch inhibitor lock when an
14099 external display is connected, as systemd will not watch
14102 * nspawn will now make use of the devices cgroup controller by
14103 default, and only permit creation of and access to the usual
14104 API device nodes like /dev/null or /dev/random, as well as
14105 access to (but not creation of) the pty devices.
14107 * We will now ship a default .network file for
14108 systemd-networkd that automatically configures DHCP for
14109 network interfaces created by nspawn's --network-veth or
14110 --network-bridge= switches.
14112 * systemd will now understand the usual M, K, G, T suffixes
14113 according to SI conventions (i.e. to the base 1000) when
14114 referring to throughput and hardware metrics. It will stay
14115 with IEC conventions (i.e. to the base 1024) for software
14116 metrics, according to what is customary according to
14117 Wikipedia. We explicitly document which base applies for
14118 each configuration option.
14120 * The DeviceAllow= setting in unit files now supports a syntax to
14121 allow-list an entire group of devices node majors at once, based on
14122 the /proc/devices listing. For example, with the string "char-pts",
14123 it is now possible to allow-list all current and future pseudo-TTYs
14126 * sd-event learned a new "post" event source. Event sources of
14127 this type are triggered by the dispatching of any event
14128 source of a type that is not "post". This is useful for
14129 implementing clean-up and check event sources that are
14130 triggered by other work being done in the program.
14132 * systemd-networkd is no longer statically enabled, but uses
14133 the usual [Install] sections so that it can be
14134 enabled/disabled using systemctl. It still is enabled by
14137 * When creating a veth interface pair with systemd-nspawn, the
14138 host side will now be prefixed with "vb-" if
14139 --network-bridge= is used, and with "ve-" if --network-veth
14140 is used. This way, it is easy to distinguish these cases on
14141 the host, for example to apply different configuration to
14142 them with systemd-networkd.
14144 * The compatibility libraries for libsystemd-journal.so,
14145 libsystem-id128.so, libsystemd-login.so and
14146 libsystemd-daemon.so do not make use of IFUNC
14147 anymore. Instead, we now build libsystemd.so multiple times
14148 under these alternative names. This means that the footprint
14149 is drastically increased, but given that these are
14150 transitional compatibility libraries, this should not matter
14151 much. This change has been made necessary to support the ARM
14152 platform for these compatibility libraries, as the ARM
14153 toolchain is not really at the same level as the toolchain
14154 for other architectures like x86 and does not support
14155 IFUNC. Please make sure to use --enable-compat-libs only
14156 during a transitional period!
14158 * The .include syntax has been deprecated and is not documented
14159 anymore. Drop-in files in .d directories should be used instead.
14161 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
14162 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
14163 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
14164 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
14165 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
14166 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
14167 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
14168 Zbigniew Jędrzejewski-Szmek
14170 — Berlin, 2014-02-24
14174 * A new component "systemd-networkd" has been added that can
14175 be used to configure local network interfaces statically or
14176 via DHCP. It is capable of bringing up bridges, VLANs, and
14177 bonding. Currently, no hook-ups for interactive network
14178 configuration are provided. Use this for your initrd,
14179 container, embedded, or server setup if you need a simple,
14180 yet powerful, network configuration solution. This
14181 configuration subsystem is quite nifty, as it allows wildcard
14182 hotplug matching in interfaces. For example, with a single
14183 configuration snippet, you can configure that all Ethernet
14184 interfaces showing up are automatically added to a bridge,
14185 or similar. It supports link-sensing and more.
14187 * A new tool "systemd-socket-proxyd" has been added which can
14188 act as a bidirectional proxy for TCP sockets. This is
14189 useful for adding socket activation support to services that
14190 do not actually support socket activation, including virtual
14191 machines and the like.
14193 * Add a new tool to save/restore rfkill state on
14196 * Save/restore state of keyboard backlights in addition to
14197 display backlights on shutdown/boot.
14199 * udev learned a new SECLABEL{} construct to label device
14200 nodes with a specific security label when they appear. For
14201 now, only SECLABEL{selinux} is supported, but the syntax is
14202 prepared for additional security frameworks.
14204 * udev gained a new scheme to configure link-level attributes
14205 from files in /etc/systemd/network/*.link. These files can
14206 match against MAC address, device path, driver name and type,
14207 and will apply attributes like the naming policy, link speed,
14208 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
14209 address assignment policy (randomized, …).
14211 * The configuration of network interface naming rules for
14212 "permanent interface names" has changed: a new NamePolicy=
14213 setting in the [Link] section of .link files determines the
14214 priority of possible naming schemes (onboard, slot, MAC,
14215 path). The default value of this setting is determined by
14216 /usr/lib/net/links/99-default.link. Old
14217 80-net-name-slot.rules udev configuration file has been
14218 removed, so local configuration overriding this file should
14219 be adapted to override 99-default.link instead.
14221 * When the User= switch is used in a unit file, also
14222 initialize $SHELL= based on the user database entry.
14224 * systemd no longer depends on libdbus. All communication is
14225 now done with sd-bus, systemd's low-level bus library
14228 * kdbus support has been added to PID 1 itself. When kdbus is
14229 enabled, this causes PID 1 to set up the system bus and
14230 enable support for a new ".busname" unit type that
14231 encapsulates bus name activation on kdbus. It works a little
14232 bit like ".socket" units, except for bus names. A new
14233 generator has been added that converts classic dbus1 service
14234 activation files automatically into native systemd .busname
14235 and .service units.
14237 * sd-bus: add a light-weight vtable implementation that allows
14238 defining objects on the bus with a simple static const
14239 vtable array of its methods, signals and properties.
14241 * systemd will not generate or install static dbus
14242 introspection data anymore to /usr/share/dbus-1/interfaces,
14243 as the precise format of these files is unclear, and
14244 nothing makes use of it.
14246 * A proxy daemon is now provided to proxy clients connecting
14247 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
14248 compatibility with classic D-Bus.
14250 * A bus driver implementation has been added that supports the
14251 classic D-Bus bus driver calls on kdbus, also for
14252 compatibility purposes.
14254 * A new API "sd-event.h" has been added that implements a
14255 minimal event loop API built around epoll. It provides a
14256 couple of features that direct epoll usage is lacking:
14257 prioritization of events, scales to large numbers of timer
14258 events, per-event timer slack (accuracy), system-wide
14259 coalescing of timer events, exit handlers, watchdog
14260 supervision support using systemd's sd_notify() API, child
14263 * A new API "sd-rntl.h" has been added that provides an API
14264 around the route netlink interface of the kernel, similar in
14265 style to "sd-bus.h".
14267 * A new API "sd-dhcp-client.h" has been added that provides a
14268 small DHCPv4 client-side implementation. This is used by
14269 "systemd-networkd".
14271 * There is a new kernel command line option
14272 "systemd.restore_state=0|1". When set to "0", none of the
14273 systemd tools will restore saved runtime state to hardware
14274 devices. More specifically, the rfkill and backlight states
14277 * The FsckPassNo= compatibility option in mount/service units
14278 has been removed. The fstab generator will now add the
14279 necessary dependencies automatically, and does not require
14280 PID1's support for that anymore.
14282 * journalctl gained a new switch, --list-boots, that lists
14283 recent boots with their times and boot IDs.
14285 * The various tools like systemctl, loginctl, timedatectl,
14286 busctl, systemd-run, … have gained a new switch "-M" to
14287 connect to a specific, local OS container (as direct
14288 connection, without requiring SSH). This works on any
14289 container that is registered with machined, such as those
14290 created by libvirt-lxc or nspawn.
14292 * systemd-run and systemd-analyze also gained support for "-H"
14293 to connect to remote hosts via SSH. This is particularly
14294 useful for systemd-run because it enables queuing of jobs
14295 onto remote systems.
14297 * machinectl gained a new command "login" to open a getty
14298 login in any local container. This works with any container
14299 that is registered with machined (such as those created by
14300 libvirt-lxc or nspawn), and which runs systemd inside.
14302 * machinectl gained a new "reboot" command that may be used to
14303 trigger a reboot on a specific container that is registered
14304 with machined. This works on any container that runs an init
14305 system of some kind.
14307 * systemctl gained a new "list-timers" command to print a nice
14308 listing of installed timer units with the times they elapse
14311 * Alternative reboot() parameters may now be specified on the
14312 "systemctl reboot" command line and are passed to the
14313 reboot() system call.
14315 * systemctl gained a new --job-mode= switch to configure the
14316 mode to queue a job with. This is a more generic version of
14317 --fail, --irreversible, and --ignore-dependencies, which are
14318 still available but not advertised anymore.
14320 * /etc/systemd/system.conf gained new settings to configure
14321 various default timeouts of units, as well as the default
14322 start limit interval and burst. These may still be overridden
14325 * PID1 will now export on the bus profile data of the security
14326 policy upload process (such as the SELinux policy upload to
14329 * journald: when forwarding logs to the console, include
14330 timestamps (following the setting in
14331 /sys/module/printk/parameters/time).
14333 * OnCalendar= in timer units now understands the special
14334 strings "yearly" and "annually". (Both are equivalent)
14336 * The accuracy of timer units is now configurable with the new
14337 AccuracySec= setting. It defaults to 1min.
14339 * A new dependency type JoinsNamespaceOf= has been added that
14340 allows running two services within the same /tmp and network
14341 namespace, if PrivateNetwork= or PrivateTmp= are used.
14343 * A new command "cat" has been added to systemctl. It outputs
14344 the original unit file of a unit, and concatenates the
14345 contents of additional "drop-in" unit file snippets, so that
14346 the full configuration is shown.
14348 * systemctl now supports globbing on the various "list-xyz"
14349 commands, like "list-units" or "list-sockets", as well as on
14350 those commands which take multiple unit names.
14352 * journalctl's --unit= switch gained support for globbing.
14354 * All systemd daemons now make use of the watchdog logic so
14355 that systemd automatically notices when they hang.
14357 * If the $container_ttys environment variable is set,
14358 getty-generator will automatically spawn a getty for each
14359 listed tty. This is useful for container managers to request
14360 login gettys to be spawned on as many ttys as needed.
14362 * %h, %s, %U specifier support is not available anymore when
14363 used in unit files for PID 1. This is because NSS calls are
14364 not safe from PID 1. They stay available for --user
14365 instances of systemd, and as special case for the root user.
14367 * loginctl gained a new "--no-legend" switch to turn off output
14368 of the legend text.
14370 * The "sd-login.h" API gained three new calls:
14371 sd_session_is_remote(), sd_session_get_remote_user(),
14372 sd_session_get_remote_host() to query information about
14375 * The udev hardware database now also carries vendor/product
14376 information of SDIO devices.
14378 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
14379 determine whether watchdog notifications are requested by
14380 the system manager.
14382 * Socket-activated per-connection services now include a
14383 short description of the connection parameters in the
14386 * tmpfiles gained a new "--boot" option. When this is not used,
14387 only lines where the command character is not suffixed with
14388 "!" are executed. When this option is specified, those
14389 options are executed too. This partitions tmpfiles
14390 directives into those that can be safely executed at any
14391 time, and those which should be run only at boot (for
14392 example, a line that creates /run/nologin).
14394 * A new API "sd-resolve.h" has been added which provides a simple
14395 asynchronous wrapper around glibc NSS hostname resolution
14396 calls, such as getaddrinfo(). In contrast to glibc's
14397 getaddrinfo_a(), it does not use signals. In contrast to most
14398 other asynchronous name resolution libraries, this one does
14399 not reimplement DNS, but reuses NSS, so that alternate
14400 hostname resolution systems continue to work, such as mDNS,
14401 LDAP, etc. This API is based on libasyncns, but it has been
14402 cleaned up for inclusion in systemd.
14404 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
14405 "sd-daemon.h" are no longer found in individual libraries
14406 libsystemd-journal.so, libsystemd-login.so,
14407 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
14408 merged them into a single library, libsystemd.so, which
14409 provides all symbols. The reason for this is cyclic
14410 dependencies, as these libraries tend to use each other's
14411 symbols. So far, we have managed to workaround that by linking
14412 a copy of a good part of our code into each of these
14413 libraries again and again, which, however, makes certain
14414 things hard to do, like sharing static variables. Also, it
14415 substantially increases footprint. With this change, there
14416 is only one library for the basic APIs systemd
14417 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
14418 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
14419 library as well, however are subject to the --enable-kdbus
14420 switch (see below). Note that "sd-dhcp-client.h" is not part
14421 of this library (this is because it only consumes, never
14422 provides, services of/to other APIs). To make the transition
14423 easy from the separate libraries to the unified one, we
14424 provide the --enable-compat-libs compile-time switch which
14425 will generate stub libraries that are compatible with the
14426 old ones but redirect all calls to the new one.
14428 * All of the kdbus logic and the new APIs "sd-bus.h",
14429 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
14430 and "sd-utf8.h" are compile-time optional via the
14431 "--enable-kdbus" switch, and they are not compiled in by
14432 default. To make use of kdbus, you have to explicitly enable
14433 the switch. Note however, that neither the kernel nor the
14434 userspace API for all of this is considered stable yet. We
14435 want to maintain the freedom to still change the APIs for
14436 now. By specifying this build-time switch, you acknowledge
14437 that you are aware of the instability of the current
14440 * Also, note that while kdbus is pretty much complete,
14441 it lacks one thing: proper policy support. This means you
14442 can build a fully working system with all features; however,
14443 it will be highly insecure. Policy support will be added in
14444 one of the next releases, at the same time that we will
14445 declare the APIs stable.
14447 * When the kernel command line argument "kdbus" is specified,
14448 systemd will automatically load the kdbus.ko kernel module. At
14449 this stage of development, it is only useful for testing kdbus
14450 and should not be used in production. Note: if "--enable-kdbus"
14451 is specified, and the kdbus.ko kernel module is available, and
14452 "kdbus" is added to the kernel command line, the entire system
14453 runs with kdbus instead of dbus-daemon, with the above mentioned
14454 problem of missing the system policy enforcement. Also a future
14455 version of kdbus.ko or a newer systemd will not be compatible with
14456 each other, and will unlikely be able to boot the machine if only
14457 one of them is updated.
14459 * systemctl gained a new "import-environment" command which
14460 uploads the caller's environment (or parts thereof) into the
14461 service manager so that it is inherited by services started
14462 by the manager. This is useful to upload variables like
14463 $DISPLAY into the user service manager.
14465 * A new PrivateDevices= switch has been added to service units
14466 which allows running a service with a namespaced /dev
14467 directory that does not contain any device nodes for
14468 physical devices. More specifically, it only includes devices
14469 such as /dev/null, /dev/urandom, and /dev/zero which are API
14472 * logind has been extended to support behaviour like VT
14473 switching on seats that do not support a VT. This makes
14474 multi-session available on seats that are not the first seat
14475 (seat0), and on systems where kernel support for VTs has
14476 been disabled at compile-time.
14478 * If a process holds a delay lock for system sleep or shutdown
14479 and fails to release it in time, we will now log its
14480 identity. This makes it easier to identify processes that
14481 cause slow suspends or power-offs.
14483 * When parsing /etc/crypttab, support for a new key-slot=
14484 option as supported by Debian is added. It allows indicating
14485 which LUKS slot to use on disk, speeding up key loading.
14487 * The sd_journal_sendv() API call has been checked and
14488 officially declared to be async-signal-safe so that it may
14489 be invoked from signal handlers for logging purposes.
14491 * Boot-time status output is now enabled automatically after a
14492 short timeout if boot does not progress, in order to give
14493 the user an indication what she or he is waiting for.
14495 * The boot-time output has been improved to show how much time
14496 remains until jobs expire.
14498 * The KillMode= switch in service units gained a new possible
14499 value "mixed". If set, and the unit is shut down, then the
14500 initial SIGTERM signal is sent only to the main daemon
14501 process, while the following SIGKILL signal is sent to
14502 all remaining processes of the service.
14504 * When a scope unit is registered, a new property "Controller"
14505 may be set. If set to a valid bus name, systemd will send a
14506 RequestStop() signal to this name when it would like to shut
14507 down the scope. This may be used to hook manager logic into
14508 the shutdown logic of scope units. Also, scope units may now
14509 be put in a special "abandoned" state, in which case the
14510 manager process which created them takes no further
14511 responsibilities for it.
14513 * When reading unit files, systemd will now verify
14514 the access mode of these files, and warn about certain
14515 suspicious combinations. This has been added to make it
14516 easier to track down packaging bugs where unit files are
14517 marked executable or world-writable.
14519 * systemd-nspawn gained a new "--setenv=" switch to set
14520 container-wide environment variables. The similar option in
14521 systemd-activate was renamed from "--environment=" to
14522 "--setenv=" for consistency.
14524 * systemd-nspawn has been updated to create a new kdbus domain
14525 for each container that is invoked, thus allowing each
14526 container to have its own set of system and user buses,
14527 independent of the host.
14529 * systemd-nspawn gained a new --drop-capability= switch to run
14530 the container with less capabilities than the default. Both
14531 --drop-capability= and --capability= now take the special
14532 string "all" for dropping or keeping all capabilities.
14534 * systemd-nspawn gained new switches for executing containers
14535 with specific SELinux labels set.
14537 * systemd-nspawn gained a new --quiet switch to not generate
14538 any additional output but the container's own console
14541 * systemd-nspawn gained a new --share-system switch to run a
14542 container without PID namespacing enabled.
14544 * systemd-nspawn gained a new --register= switch to control
14545 whether the container is registered with systemd-machined or
14546 not. This is useful for containers that do not run full
14547 OS images, but only specific apps.
14549 * systemd-nspawn gained a new --keep-unit which may be used
14550 when invoked as the only program from a service unit, and
14551 results in registration of the unit service itself in
14552 systemd-machined, instead of a newly opened scope unit.
14554 * systemd-nspawn gained a new --network-interface= switch for
14555 moving arbitrary interfaces to the container. The new
14556 --network-veth switch creates a virtual Ethernet connection
14557 between host and container. The new --network-bridge=
14558 switch then allows assigning the host side of this virtual
14559 Ethernet connection to a bridge device.
14561 * systemd-nspawn gained a new --personality= switch for
14562 setting the kernel personality for the container. This is
14563 useful when running a 32-bit container on a 64-bit host. A
14564 similar option Personality= is now also available for service
14567 * logind will now also track a "Desktop" identifier for each
14568 session which encodes the desktop environment of it. This is
14569 useful for desktop environments that want to identify
14570 multiple running sessions of itself easily.
14572 * A new SELinuxContext= setting for service units has been
14573 added that allows setting a specific SELinux execution
14574 context for a service.
14576 * Most systemd client tools will now honour $SYSTEMD_LESS for
14577 settings of the "less" pager. By default, these tools will
14578 override $LESS to allow certain operations to work, such as
14579 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
14580 influence this logic.
14582 * systemd's "seccomp" hook-up has been changed to make use of
14583 the libseccomp library instead of using its own
14584 implementation. This has benefits for portability among
14587 * For usage together with SystemCallFilter=, a new
14588 SystemCallErrorNumber= setting has been introduced that
14589 allows configuration of a system error number to be returned
14590 on filtered system calls, instead of immediately killing the
14591 process. Also, SystemCallArchitectures= has been added to
14592 limit access to system calls of a particular architecture
14593 (in order to turn off support for unused secondary
14594 architectures). There is also a global
14595 SystemCallArchitectures= setting in system.conf now to turn
14596 off support for non-native system calls system-wide.
14598 * systemd requires a kernel with a working name_to_handle_at(),
14599 please see the kernel config requirements in the README file.
14601 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
14602 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
14603 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
14604 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
14605 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
14606 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
14607 Elia Pinto, Florian Weimer, George McCollister, Goffredo
14608 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
14609 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
14610 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
14611 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
14612 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
14613 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
14614 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
14615 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
14616 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
14617 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
14618 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
14619 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
14620 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
14621 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
14622 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
14623 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
14624 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
14626 — Berlin, 2014-02-20
14630 * logind has gained support for facilitating privileged input
14631 and drm device access for unprivileged clients. This work is
14632 useful to allow Wayland display servers (and similar
14633 programs, such as kmscon) to run under the user's ID and
14634 access input and drm devices which are normally
14635 protected. When this is used (and the kernel is new enough)
14636 logind will "mute" IO on the file descriptors passed to
14637 Wayland as long as it is in the background and "unmute" it
14638 if it returns into the foreground. This allows secure
14639 session switching without allowing background sessions to
14640 eavesdrop on input and display data. This also introduces
14641 session switching support if VT support is turned off in the
14642 kernel, and on seats that are not seat0.
14644 * A new kernel command line option luks.options= is understood
14645 now which allows specifying LUKS options for usage for LUKS
14646 encrypted partitions specified with luks.uuid=.
14648 * tmpfiles.d(5) snippets may now use specifier expansion in
14649 path names. More specifically %m, %b, %H, %v, are now
14650 replaced by the local machine id, boot id, hostname, and
14651 kernel version number.
14653 * A new tmpfiles.d(5) command "m" has been introduced which
14654 may be used to change the owner/group/access mode of a file
14655 or directory if it exists, but do nothing if it does not.
14657 * This release removes high-level support for the
14658 MemorySoftLimit= cgroup setting. The underlying kernel
14659 cgroup attribute memory.soft_limit= is currently badly
14660 designed and likely to be removed from the kernel API in its
14661 current form, hence we should not expose it for now.
14663 * The memory.use_hierarchy cgroup attribute is now enabled for
14664 all cgroups systemd creates in the memory cgroup
14665 hierarchy. This option is likely to be come the built-in
14666 default in the kernel anyway, and the non-hierarchical mode
14667 never made much sense in the intrinsically hierarchical
14670 * A new field _SYSTEMD_SLICE= is logged along with all journal
14671 messages containing the slice a message was generated
14672 from. This is useful to allow easy per-customer filtering of
14673 logs among other things.
14675 * systemd-journald will no longer adjust the group of journal
14676 files it creates to the "systemd-journal" group. Instead we
14677 rely on the journal directory to be owned by the
14678 "systemd-journal" group, and its setgid bit set, so that the
14679 kernel file system layer will automatically enforce that
14680 journal files inherit this group assignment. The reason for
14681 this change is that we cannot allow NSS look-ups from
14682 journald which would be necessary to resolve
14683 "systemd-journal" to a numeric GID, because this might
14684 create deadlocks if NSS involves synchronous queries to
14685 other daemons (such as nscd, or sssd) which in turn are
14686 logging clients of journald and might block on it, which
14687 would then dead lock. A tmpfiles.d(5) snippet included in
14688 systemd will make sure the setgid bit and group are
14689 properly set on the journal directory if it exists on every
14690 boot. However, we recommend adjusting it manually after
14691 upgrades too (or from RPM scriptlets), so that the change is
14692 not delayed until next reboot.
14694 * Backlight and random seed files in /var/lib/ have moved into
14695 the /var/lib/systemd/ directory, in order to centralize all
14696 systemd generated files in one directory.
14698 * Boot time performance measurements (as displayed by
14699 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
14700 performance information if that's available to determine how
14701 much time BIOS and boot loader initialization required. With
14702 a sufficiently new BIOS you hence no longer need to boot
14703 with Gummiboot to get access to such information.
14705 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
14706 Cristian Rodríguez, Dave Reisner, David Herrmann, David
14707 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
14708 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
14709 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
14710 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
14711 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
14713 — Berlin, 2013-10-02
14717 * The Restart= option for services now understands a new
14718 on-watchdog setting, which will restart the service
14719 automatically if the service stops sending out watchdog keep
14720 alive messages (as configured with WatchdogSec=).
14722 * The getty generator (which is responsible for bringing up a
14723 getty on configured serial consoles) will no longer only
14724 start a getty on the primary kernel console but on all
14725 others, too. This makes the order in which console= is
14726 specified on the kernel command line less important.
14728 * libsystemd-logind gained a new sd_session_get_vt() call to
14729 retrieve the VT number of a session.
14731 * If the option "tries=0" is set for an entry of /etc/crypttab
14732 its passphrase is queried indefinitely instead of any
14733 maximum number of tries.
14735 * If a service with a configure PID file terminates its PID
14736 file will now be removed automatically if it still exists
14737 afterwards. This should put an end to stale PID files.
14739 * systemd-run will now also take relative binary path names
14740 for execution and no longer insists on absolute paths.
14742 * InaccessibleDirectories= and ReadOnlyDirectories= now take
14743 paths that are optionally prefixed with "-" to indicate that
14744 it should not be considered a failure if they do not exist.
14746 * journalctl -o (and similar commands) now understands a new
14747 output mode "short-precise", it is similar to "short" but
14748 shows timestamps with usec accuracy.
14750 * The option "discard" (as known from Debian) is now
14751 synonymous to "allow-discards" in /etc/crypttab. In fact,
14752 "discard" is preferred now (since it is easier to remember
14755 * Some licensing clean-ups were made, so that more code is now
14756 LGPL-2.1 licensed than before.
14758 * A minimal tool to save/restore the display backlight
14759 brightness across reboots has been added. It will store the
14760 backlight setting as late as possible at shutdown, and
14761 restore it as early as possible during reboot.
14763 * A logic to automatically discover and enable home and swap
14764 partitions on GPT disks has been added. With this in place
14765 /etc/fstab becomes optional for many setups as systemd can
14766 discover certain partitions located on the root disk
14767 automatically. Home partitions are recognized under their
14768 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
14769 partitions are recognized under their GPT type ID
14770 0657fd6da4ab43c484e50933c84b4f4f.
14772 * systemd will no longer pass any environment from the kernel
14773 or initrd to system services. If you want to set an
14774 environment for all services, do so via the kernel command
14775 line systemd.setenv= assignment.
14777 * The systemd-sysctl tool no longer natively reads the file
14778 /etc/sysctl.conf. If desired, the file should be symlinked
14779 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
14780 legacy support by a symlink rather than built-in code, it
14781 also makes the otherwise hidden order of application of the
14782 different files visible. (Note that this partly reverts to a
14783 pre-198 application order of sysctl knobs!)
14785 * The "systemctl set-log-level" and "systemctl dump" commands
14786 have been moved to systemd-analyze.
14788 * systemd-run learned the new --remain-after-exit switch,
14789 which causes the scope unit not to be cleaned up
14790 automatically after the process terminated.
14792 * tmpfiles learned a new --exclude-prefix= switch to exclude
14793 certain paths from operation.
14795 * journald will now automatically flush all messages to disk
14796 as soon as a message at the log level CRIT, ALERT or EMERG
14799 Contributions from: Andrew Cook, Brandon Philips, Christian
14800 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
14801 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
14802 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
14803 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
14804 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
14805 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
14806 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
14807 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
14808 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
14809 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
14810 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
14811 William Giokas, Zbigniew Jędrzejewski-Szmek
14813 — Berlin, 2013-09-13
14817 * The documentation has been updated to cover the various new
14818 concepts introduced with 205.
14820 * Unit files now understand the new %v specifier which
14821 resolves to the kernel version string as returned by "uname
14824 * systemctl now supports filtering the unit list output by
14825 load state, active state and sub state, using the new
14826 --state= parameter.
14828 * "systemctl status" will now show the results of the
14829 condition checks (like ConditionPathExists= and similar) of
14830 the last start attempts of the unit. They are also logged to
14833 * "journalctl -b" may now be used to look for boot output of a
14834 specific boot. Try "journalctl -b -1" for the previous boot,
14835 but the syntax is substantially more powerful.
14837 * "journalctl --show-cursor" has been added which prints the
14838 cursor string the last shown log line. This may then be used
14839 with the new "journalctl --after-cursor=" switch to continue
14840 browsing logs from that point on.
14842 * "journalctl --force" may now be used to force regeneration
14845 * Creation of "dead" device nodes has been moved from udev
14846 into kmod and tmpfiles. Previously, udev would read the kmod
14847 databases to pre-generate dead device nodes based on meta
14848 information contained in kernel modules, so that these would
14849 be auto-loaded on access rather then at boot. As this
14850 does not really have much to do with the exposing actual
14851 kernel devices to userspace this has always been slightly
14852 alien in the udev codebase. Following the new scheme kmod
14853 will now generate a runtime snippet for tmpfiles from the
14854 module meta information and it now is tmpfiles' job to the
14855 create the nodes. This also allows overriding access and
14856 other parameters for the nodes using the usual tmpfiles
14857 facilities. As side effect this allows us to remove the
14858 CAP_SYS_MKNOD capability bit from udevd entirely.
14860 * logind's device ACLs may now be applied to these "dead"
14861 devices nodes too, thus finally allowing managed access to
14862 devices such as /dev/snd/sequencer without loading the
14863 backing module right-away.
14865 * A new RPM macro has been added that may be used to apply
14866 tmpfiles configuration during package installation.
14868 * systemd-detect-virt and ConditionVirtualization= now can
14869 detect User-Mode-Linux machines (UML).
14871 * journald will now implicitly log the effective capabilities
14872 set of processes in the message metadata.
14874 * systemd-cryptsetup has gained support for TrueCrypt volumes.
14876 * The initrd interface has been simplified (more specifically,
14877 support for passing performance data via environment
14878 variables and fsck results via files in /run has been
14879 removed). These features were non-essential, and are
14880 nowadays available in a much nicer way by having systemd in
14881 the initrd serialize its state and have the hosts systemd
14882 deserialize it again.
14884 * The udev "keymap" data files and tools to apply keyboard
14885 specific mappings of scan to key codes, and force-release
14886 scan code lists have been entirely replaced by a udev
14887 "keyboard" builtin and a hwdb data file.
14889 * systemd will now honour the kernel's "quiet" command line
14890 argument also during late shutdown, resulting in a
14891 completely silent shutdown when used.
14893 * There's now an option to control the SO_REUSEPORT socket
14894 option in .socket units.
14896 * Instance units will now automatically get a per-template
14897 subslice of system.slice unless something else is explicitly
14898 configured. For example, instances of sshd@.service will now
14899 implicitly be placed in system-sshd.slice rather than
14900 system.slice as before.
14902 * Test coverage support may now be enabled at build time.
14904 Contributions from: Dave Reisner, Frederic Crozat, Harald
14905 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
14906 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
14907 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
14908 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
14909 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
14910 Giokas, Zbigniew Jędrzejewski-Szmek
14912 — Berlin, 2013-07-23
14916 * Two new unit types have been introduced:
14918 Scope units are very similar to service units, however, are
14919 created out of pre-existing processes — instead of PID 1
14920 forking off the processes. By using scope units it is
14921 possible for system services and applications to group their
14922 own child processes (worker processes) in a powerful way
14923 which then maybe used to organize them, or kill them
14924 together, or apply resource limits on them.
14926 Slice units may be used to partition system resources in an
14927 hierarchical fashion and then assign other units to them. By
14928 default there are now three slices: system.slice (for all
14929 system services), user.slice (for all user sessions),
14930 machine.slice (for VMs and containers).
14932 Slices and scopes have been introduced primarily in
14933 context of the work to move cgroup handling to a
14934 single-writer scheme, where only PID 1
14935 creates/removes/manages cgroups.
14937 * There's a new concept of "transient" units. In contrast to
14938 normal units these units are created via an API at runtime,
14939 not from configuration from disk. More specifically this
14940 means it is now possible to run arbitrary programs as
14941 independent services, with all execution parameters passed
14942 in via bus APIs rather than read from disk. Transient units
14943 make systemd substantially more dynamic then it ever was,
14944 and useful as a general batch manager.
14946 * logind has been updated to make use of scope and slice units
14947 for managing user sessions. As a user logs in he will get
14948 his own private slice unit, to which all sessions are added
14949 as scope units. We also added support for automatically
14950 adding an instance of user@.service for the user into the
14951 slice. Effectively logind will no longer create cgroup
14952 hierarchies on its own now, it will defer entirely to PID 1
14953 for this by means of scope, service and slice units. Since
14954 user sessions this way become entities managed by PID 1
14955 the output of "systemctl" is now a lot more comprehensive.
14957 * A new mini-daemon "systemd-machined" has been added which
14958 may be used by virtualization managers to register local
14959 VMs/containers. nspawn has been updated accordingly, and
14960 libvirt will be updated shortly. machined will collect a bit
14961 of meta information about the VMs/containers, and assign
14962 them their own scope unit (see above). The collected
14963 meta-data is then made available via the "machinectl" tool,
14964 and exposed in "ps" and similar tools. machined/machinectl
14965 is compile-time optional.
14967 * As discussed earlier, the low-level cgroup configuration
14968 options ControlGroup=, ControlGroupModify=,
14969 ControlGroupPersistent=, ControlGroupAttribute= have been
14970 removed. Please use high-level attribute settings instead as
14971 well as slice units.
14973 * A new bus call SetUnitProperties() has been added to alter
14974 various runtime parameters of a unit. This is primarily
14975 useful to alter cgroup parameters dynamically in a nice way,
14976 but will be extended later on to make more properties
14977 modifiable at runtime. systemctl gained a new set-properties
14978 command that wraps this call.
14980 * A new tool "systemd-run" has been added which can be used to
14981 run arbitrary command lines as transient services or scopes,
14982 while configuring a number of settings via the command
14983 line. This tool is currently very basic, however already
14984 very useful. We plan to extend this tool to even allow
14985 queuing of execution jobs with time triggers from the
14986 command line, similar in fashion to "at".
14988 * nspawn will now inform the user explicitly that kernels with
14989 audit enabled break containers, and suggest the user to turn
14992 * Support for detecting the IMA and AppArmor security
14993 frameworks with ConditionSecurity= has been added.
14995 * journalctl gained a new "-k" switch for showing only kernel
14996 messages, mimicking dmesg output; in addition to "--user"
14997 and "--system" switches for showing only user's own logs
15000 * systemd-delta can now show information about drop-in
15001 snippets extending unit files.
15003 * libsystemd-bus has been substantially updated but is still
15004 not available as public API.
15006 * systemd will now look for the "debug" argument on the kernel
15007 command line and enable debug logging, similar to what
15008 "systemd.log_level=debug" already did before.
15010 * "systemctl set-default", "systemctl get-default" has been
15011 added to configure the default.target symlink, which
15012 controls what to boot into by default.
15014 * "systemctl set-log-level" has been added as a convenient
15015 way to raise and lower systemd logging threshold.
15017 * "systemd-analyze plot" will now show the time the various
15018 generators needed for execution, as well as information
15019 about the unit file loading.
15021 * libsystemd-journal gained a new sd_journal_open_files() call
15022 for opening specific journal files. journactl also gained a
15023 new switch to expose this new functionality. Previously we
15024 only supported opening all files from a directory, or all
15025 files from the system, as opening individual files only is
15026 racy due to journal file rotation.
15028 * systemd gained the new DefaultEnvironment= setting in
15029 /etc/systemd/system.conf to set environment variables for
15032 * If a privileged process logs a journal message with the
15033 OBJECT_PID= field set, then journald will automatically
15034 augment this with additional OBJECT_UID=, OBJECT_GID=,
15035 OBJECT_COMM=, OBJECT_EXE=, … fields. This is useful if
15036 system services want to log events about specific client
15037 processes. journactl/systemctl has been updated to make use
15038 of this information if all log messages regarding a specific
15041 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
15042 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
15043 Reisner, David Coppa, David King, David Strauss, Eelco
15044 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
15045 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
15046 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
15047 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
15048 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
15049 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
15050 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
15051 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
15052 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
15053 Łukasz Stelmach, 장동준
15057 * The Python bindings gained some minimal support for the APIs
15058 exposed by libsystemd-logind.
15060 * ConditionSecurity= gained support for detecting SMACK. Since
15061 this condition already supports SELinux and AppArmor we only
15062 miss IMA for this. Patches welcome!
15064 Contributions from: Karol Lewandowski, Lennart Poettering,
15065 Zbigniew Jędrzejewski-Szmek
15069 * systemd-nspawn will now create /etc/resolv.conf if
15070 necessary, before bind-mounting the host's file onto it.
15072 * systemd-nspawn will now store meta information about a
15073 container on the container's cgroup as extended attribute
15074 fields, including the root directory.
15076 * The cgroup hierarchy has been reworked in many ways. All
15077 objects any of the components systemd creates in the cgroup
15078 tree are now suffixed. More specifically, user sessions are
15079 now placed in cgroups suffixed with ".session", users in
15080 cgroups suffixed with ".user", and nspawn containers in
15081 cgroups suffixed with ".nspawn". Furthermore, all cgroup
15082 names are now escaped in a simple scheme to avoid collision
15083 of userspace object names with kernel filenames. This work
15084 is preparation for making these objects relocatable in the
15085 cgroup tree, in order to allow easy resource partitioning of
15086 these objects without causing naming conflicts.
15088 * systemctl list-dependencies gained the new switches
15089 --plain, --reverse, --after and --before.
15091 * systemd-inhibit now shows the process name of processes that
15092 have taken an inhibitor lock.
15094 * nss-myhostname will now also resolve "localhost"
15095 implicitly. This makes /etc/hosts an optional file and
15096 nicely handles that on IPv6 ::1 maps to both "localhost" and
15097 the local hostname.
15099 * libsystemd-logind.so gained a new call
15100 sd_get_machine_names() to enumerate running containers and
15101 VMs (currently only supported by very new libvirt and
15102 nspawn). sd_login_monitor can now be used to watch
15103 VMs/containers coming and going.
15105 * .include is not allowed recursively anymore, and only in
15106 unit files. Usually it is better to use drop-in snippets in
15107 .d/*.conf anyway, as introduced with systemd 198.
15109 * systemd-analyze gained a new "critical-chain" command that
15110 determines the slowest chain of units run during system
15111 boot-up. It is very useful for tracking down where
15112 optimizing boot time is the most beneficial.
15114 * systemd will no longer allow manipulating service paths in
15115 the name=systemd:/system cgroup tree using ControlGroup= in
15116 units. (But is still fine with it in all other dirs.)
15118 * There's a new systemd-nspawn@.service service file that may
15119 be used to easily run nspawn containers as system
15120 services. With the container's root directory in
15121 /var/lib/container/foobar it is now sufficient to run
15122 "systemctl start systemd-nspawn@foobar.service" to boot it.
15124 * systemd-cgls gained a new parameter "--machine" to list only
15125 the processes within a certain container.
15127 * ConditionSecurity= now can check for "apparmor". We still
15128 are lacking checks for SMACK and IMA for this condition
15129 check though. Patches welcome!
15131 * A new configuration file /etc/systemd/sleep.conf has been
15132 added that may be used to configure which kernel operation
15133 systemd is supposed to execute when "suspend", "hibernate"
15134 or "hybrid-sleep" is requested. This makes the new kernel
15135 "freeze" state accessible to the user.
15137 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
15138 the passed argument if applicable.
15140 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
15141 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
15142 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
15143 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
15144 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
15145 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
15146 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
15151 * The output of 'systemctl list-jobs' got some polishing. The
15152 '--type=' argument may now be passed more than once. A new
15153 command 'systemctl list-sockets' has been added which shows
15154 a list of kernel sockets systemd is listening on with the
15155 socket units they belong to, plus the units these socket
15158 * The experimental libsystemd-bus library got substantial
15159 updates to work in conjunction with the (also experimental)
15160 kdbus kernel project. It works well enough to exchange
15161 messages with some sophistication. Note that kdbus is not
15162 ready yet, and the library is mostly an elaborate test case
15163 for now, and not installable.
15165 * systemd gained a new unit 'systemd-static-nodes.service'
15166 that generates static device nodes earlier during boot, and
15167 can run in conjunction with udev.
15169 * libsystemd-login gained a new call sd_pid_get_user_unit()
15170 to retrieve the user systemd unit a process is running
15171 in. This is useful for systems where systemd is used as
15174 * systemd-nspawn now places all containers in the new /machine
15175 top-level cgroup directory in the name=systemd
15176 hierarchy. libvirt will soon do the same, so that we get a
15177 uniform separation of /system, /user and /machine for system
15178 services, user processes and containers/virtual
15179 machines. This new cgroup hierarchy is also useful to stick
15180 stable names to specific container instances, which can be
15181 recognized later this way (this name may be controlled
15182 via systemd-nspawn's new -M switch). libsystemd-login also
15183 gained a new call sd_pid_get_machine_name() to retrieve the
15184 name of the container/VM a specific process belongs to.
15186 * bootchart can now store its data in the journal.
15188 * libsystemd-journal gained a new call
15189 sd_journal_add_conjunction() for AND expressions to the
15190 matching logic. This can be used to express more complex
15191 logical expressions.
15193 * journactl can now take multiple --unit= and --user-unit=
15196 * The cryptsetup logic now understands the "luks.key=" kernel
15197 command line switch for specifying a file to read the
15198 decryption key from. Also, if a configured key file is not
15199 found the tool will now automatically fall back to prompting
15202 * Python systemd.journal module was updated to wrap recently
15203 added functions from libsystemd-journal. The interface was
15204 changed to bring the low level interface in s.j._Reader
15205 closer to the C API, and the high level interface in
15206 s.j.Reader was updated to wrap and convert all data about
15209 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
15210 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
15211 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
15212 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
15213 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
15214 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
15218 * journalctl --update-catalog now understands a new --root=
15219 option to operate on catalogs found in a different root
15222 * During shutdown after systemd has terminated all running
15223 services a final killing loop kills all remaining left-over
15224 processes. We will now print the name of these processes
15225 when we send SIGKILL to them, since this usually indicates a
15228 * If /etc/crypttab refers to password files stored on
15229 configured mount points automatic dependencies will now be
15230 generated to ensure the specific mount is established first
15231 before the key file is attempted to be read.
15233 * 'systemctl status' will now show information about the
15234 network sockets a socket unit is listening on.
15236 * 'systemctl status' will also shown information about any
15237 drop-in configuration file for units. (Drop-In configuration
15238 files in this context are files such as
15239 /etc/systemd/system/foobar.service.d/*.conf)
15241 * systemd-cgtop now optionally shows summed up CPU times of
15242 cgroups. Press '%' while running cgtop to switch between
15243 percentage and absolute mode. This is useful to determine
15244 which cgroups use up the most CPU time over the entire
15245 runtime of the system. systemd-cgtop has also been updated
15246 to be 'pipeable' for processing with further shell tools.
15248 * 'hostnamectl set-hostname' will now allow setting of FQDN
15251 * The formatting and parsing of time span values has been
15252 changed. The parser now understands fractional expressions
15253 such as "5.5h". The formatter will now output fractional
15254 expressions for all time spans under 1min, i.e. "5.123456s"
15255 rather than "5s 123ms 456us". For time spans under 1s
15256 millisecond values are shown, for those under 1ms
15257 microsecond values are shown. This should greatly improve
15258 all time-related output of systemd.
15260 * libsystemd-login and libsystemd-journal gained new
15261 functions for querying the poll() events mask and poll()
15262 timeout value for integration into arbitrary event
15265 * localectl gained the ability to list available X11 keymaps
15266 (models, layouts, variants, options).
15268 * 'systemd-analyze dot' gained the ability to filter for
15269 specific units via shell-style globs, to create smaller,
15270 more useful graphs. I.e. it is now possible to create simple
15271 graphs of all the dependencies between only target units, or
15272 of all units that Avahi has dependencies with.
15274 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
15275 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
15276 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
15277 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
15278 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
15279 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
15280 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
15284 * The boot-time readahead implementation for rotating media
15285 will now read the read-ahead data in multiple passes which
15286 consist of all read requests made in equidistant time
15287 intervals. This means instead of strictly reading read-ahead
15288 data in its physical order on disk we now try to find a
15289 middle ground between physical and access time order.
15291 * /etc/os-release files gained a new BUILD_ID= field for usage
15292 on operating systems that provide continuous builds of OS
15295 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
15296 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
15297 William Douglas, Zbigniew Jędrzejewski-Szmek
15301 * systemd-python gained an API exposing libsystemd-daemon.
15303 * The SMACK setup logic gained support for uploading CIPSO
15306 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
15307 ReadOnlyDirectories= and InaccessibleDirectories= has
15308 changed. The private /tmp and /var/tmp directories are now
15309 shared by all processes of a service (which means
15310 ExecStartPre= may now leave data in /tmp that ExecStart= of
15311 the same service can still access). When a service is
15312 stopped its temporary directories are immediately deleted
15313 (normal clean-up with tmpfiles is still done in addition to
15316 * By default, systemd will now set a couple of sysctl
15317 variables in the kernel: the safe sysrq options are turned
15318 on, IP route verification is turned on, and source routing
15319 disabled. The recently added hardlink and softlink
15320 protection of the kernel is turned on. These settings should
15321 be reasonably safe, and good defaults for all new systems.
15323 * The predictable network naming logic may now be turned off
15324 with a new kernel command line switch: net.ifnames=0.
15326 * A new libsystemd-bus module has been added that implements a
15327 pretty complete D-Bus client library. For details see:
15329 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
15331 * journald will now explicitly flush the journal files to disk
15332 at the latest 5min after each write. The file will then also
15333 be marked offline until the next write. This should increase
15334 reliability in case of a crash. The synchronization delay
15335 can be configured via SyncIntervalSec= in journald.conf.
15337 * There's a new remote-fs-setup.target unit that can be used
15338 to pull in specific services when at least one remote file
15339 system is to be mounted.
15341 * There are new targets timers.target and paths.target as
15342 canonical targets to pull user timer and path units in
15343 from. This complements sockets.target with a similar
15344 purpose for socket units.
15346 * libudev gained a new call udev_device_set_attribute_value()
15347 to set sysfs attributes of a device.
15349 * The udev daemon now sets the default number of worker
15350 processes executed in parallel based on the number of available
15351 CPUs instead of the amount of available RAM. This is supposed
15352 to provide a more reliable default and limit a too aggressive
15353 parallelism for setups with 1000s of devices connected.
15355 Contributions from: Auke Kok, Colin Walters, Cristian
15356 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
15357 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
15358 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
15359 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
15360 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
15361 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
15362 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
15363 Zbigniew Jędrzejewski-Szmek
15367 * Configuration of unit files may now be extended via drop-in
15368 files without having to edit/override the unit files
15369 themselves. More specifically, if the administrator wants to
15370 change one value for a service file foobar.service he can
15371 now do so by dropping in a configuration snippet into
15372 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
15373 will load all these snippets and apply them on top of the
15374 main unit configuration file, possibly extending or
15375 overriding its settings. Using these drop-in snippets is
15376 generally nicer than the two earlier options for changing
15377 unit files locally: copying the files from
15378 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
15379 them there; or creating a new file in /etc/systemd/system/
15380 that incorporates the original one via ".include". Drop-in
15381 snippets into these .d/ directories can be placed in any
15382 directory systemd looks for units in, and the usual
15383 overriding semantics between /usr/lib, /etc and /run apply
15386 * Most unit file settings which take lists of items can now be
15387 reset by assigning the empty string to them. For example,
15388 normally, settings such as Environment=FOO=BAR append a new
15389 environment variable assignment to the environment block,
15390 each time they are used. By assigning Environment= the empty
15391 string the environment block can be reset to empty. This is
15392 particularly useful with the .d/*.conf drop-in snippets
15393 mentioned above, since this adds the ability to reset list
15394 settings from vendor unit files via these drop-ins.
15396 * systemctl gained a new "list-dependencies" command for
15397 listing the dependencies of a unit recursively.
15399 * Inhibitors are now honored and listed by "systemctl
15400 suspend", "systemctl poweroff" (and similar) too, not only
15401 GNOME. These commands will also list active sessions by
15404 * Resource limits (as exposed by the various control group
15405 controllers) can now be controlled dynamically at runtime
15406 for all units. More specifically, you can now use a command
15407 like "systemctl set-cgroup-attr foobar.service cpu.shares
15408 2000" to alter the CPU shares a specific service gets. These
15409 settings are stored persistently on disk, and thus allow the
15410 administrator to easily adjust the resource usage of
15411 services with a few simple commands. This dynamic resource
15412 management logic is also available to other programs via the
15413 bus. Almost any kernel cgroup attribute and controller is
15416 * systemd-vconsole-setup will now copy all font settings to
15417 all allocated VTs, where it previously applied them only to
15420 * libsystemd-login gained the new sd_session_get_tty() API
15423 * This release drops support for a few legacy or
15424 distribution-specific LSB facility names when parsing init
15425 scripts: $x-display-manager, $mail-transfer-agent,
15426 $mail-transport-agent, $mail-transfer-agent, $smtp,
15427 $null. Also, the mail-transfer-agent.target unit backing
15428 this has been removed. Distributions which want to retain
15429 compatibility with this should carry the burden for
15430 supporting this themselves and patch support for these back
15431 in, if they really need to. Also, the facilities $syslog and
15432 $local_fs are now ignored, since systemd does not support
15433 early-boot LSB init scripts anymore, and these facilities
15434 are implied anyway for normal services. syslog.target has
15437 * There are new bus calls on PID1's Manager object for
15438 cancelling jobs, and removing snapshot units. Previously,
15439 both calls were only available on the Job and Snapshot
15440 objects themselves.
15442 * systemd-journal-gatewayd gained SSL support.
15444 * The various "environment" files, such as /etc/locale.conf
15445 now support continuation lines with a backslash ("\") as
15446 last character in the line, similarly in style (but different)
15447 to how this is supported in shells.
15449 * For normal user processes the _SYSTEMD_USER_UNIT= field is
15450 now implicitly appended to every log entry logged. systemctl
15451 has been updated to filter by this field when operating on a
15452 user systemd instance.
15454 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
15455 CAP_AUDIT_CONTROL capabilities to the capabilities set for
15456 the container. This makes it easier to boot unmodified
15457 Fedora systems in a container, which however still requires
15458 audit=0 to be passed on the kernel command line. Auditing in
15459 kernel and userspace is unfortunately still too broken in
15460 context of containers, hence we recommend compiling it out
15461 of the kernel or using audit=0. Hopefully this will be fixed
15462 one day for good in the kernel.
15464 * nspawn gained the new --bind= and --bind-ro= parameters to
15465 bind mount specific directories from the host into the
15468 * nspawn will now mount its own devpts file system instance
15469 into the container, in order not to leak pty devices from
15470 the host into the container.
15472 * systemd will now read the firmware boot time performance
15473 information from the EFI variables, if the used boot loader
15474 supports this, and takes it into account for boot performance
15475 analysis via "systemd-analyze". This is currently supported
15476 only in conjunction with Gummiboot, but could be supported
15477 by other boot loaders too. For details see:
15479 https://systemd.io/BOOT_LOADER_INTERFACE
15481 * A new generator has been added that automatically mounts the
15482 EFI System Partition (ESP) to /boot, if that directory
15483 exists, is empty, and no other file system has been
15484 configured to be mounted there.
15486 * logind will now send out PrepareForSleep(false) out
15487 unconditionally, after coming back from suspend. This may be
15488 used by applications as asynchronous notification for
15489 system resume events.
15491 * "systemctl unlock-sessions" has been added, that allows
15492 unlocking the screens of all user sessions at once, similar
15493 to how "systemctl lock-sessions" already locked all users
15494 sessions. This is backed by a new D-Bus call UnlockSessions().
15496 * "loginctl seat-status" will now show the master device of a
15497 seat. (i.e. the device of a seat that needs to be around for
15498 the seat to be considered available, usually the graphics
15501 * tmpfiles gained a new "X" line type, that allows
15502 configuration of files and directories (with wildcards) that
15503 shall be excluded from automatic cleanup ("aging").
15505 * udev default rules set the device node permissions now only
15506 at "add" events, and do not change them any longer with a
15507 later "change" event.
15509 * The log messages for lid events and power/sleep keypresses
15510 now carry a message ID.
15512 * We now have a substantially larger unit test suite, but this
15513 continues to be work in progress.
15515 * udevadm hwdb gained a new --root= parameter to change the
15516 root directory to operate relative to.
15518 * logind will now issue a background sync() request to the kernel
15519 early at shutdown, so that dirty buffers are flushed to disk early
15520 instead of at the last moment, in order to optimize shutdown
15523 * A new bootctl tool has been added that is an interface for
15524 certain boot loader operations. This is currently a preview
15525 and is likely to be extended into a small mechanism daemon
15526 like timedated, localed, hostnamed, and can be used by
15527 graphical UIs to enumerate available boot options, and
15528 request boot into firmware operations.
15530 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
15531 the rest of the package. It also has been updated to work
15532 correctly in initrds.
15534 * polkit previously has been runtime optional, and is now also
15535 compile time optional via a configure switch.
15537 * systemd-analyze has been reimplemented in C. Also "systemctl
15538 dot" has moved into systemd-analyze.
15540 * "systemctl status" with no further parameters will now print
15541 the status of all active or failed units.
15543 * Operations such as "systemctl start" can now be executed
15544 with a new mode "--irreversible" which may be used to queue
15545 operations that cannot accidentally be reversed by a later
15546 job queuing. This is by default used to make shutdown
15547 requests more robust.
15549 * The Python API of systemd now gained a new module for
15550 reading journal files.
15552 * A new tool kernel-install has been added that can install
15553 kernel images according to the Boot Loader Specification:
15555 https://systemd.io/BOOT_LOADER_SPECIFICATION
15557 * Boot time console output has been improved to provide
15558 animated boot time output for hanging jobs.
15560 * A new tool systemd-activate has been added which can be used
15561 to test socket activation with, directly from the command
15562 line. This should make it much easier to test and debug
15563 socket activation in daemons.
15565 * journalctl gained a new "--reverse" (or -r) option to show
15566 journal output in reverse order (i.e. newest line first).
15568 * journalctl gained a new "--pager-end" (or -e) option to jump
15569 to immediately jump to the end of the journal in the
15570 pager. This is only supported in conjunction with "less".
15572 * journalctl gained a new "--user-unit=" option, that works
15573 similarly to "--unit=" but filters for user units rather than
15576 * A number of unit files to ease adoption of systemd in
15577 initrds has been added. This moves some minimal logic from
15578 the various initrd implementations into systemd proper.
15580 * The journal files are now owned by a new group
15581 "systemd-journal", which exists specifically to allow access
15582 to the journal, and nothing else. Previously, we used the
15583 "adm" group for that, which however possibly covers more
15584 than just journal/log file access. This new group is now
15585 already used by systemd-journal-gatewayd to ensure this
15586 daemon gets access to the journal files and as little else
15587 as possible. Note that "make install" will also set FS ACLs
15588 up for /var/log/journal to give "adm" and "wheel" read
15589 access to it, in addition to "systemd-journal" which owns
15590 the journal files. We recommend that packaging scripts also
15591 add read access to "adm" + "wheel" to /var/log/journal, and
15592 all existing/future journal files. To normal users and
15593 administrators little changes, however packagers need to
15594 ensure to create the "systemd-journal" system group at
15595 package installation time.
15597 * The systemd-journal-gatewayd now runs as unprivileged user
15598 systemd-journal-gateway:systemd-journal-gateway. Packaging
15599 scripts need to create these system user/group at
15602 * timedated now exposes a new boolean property CanNTP that
15603 indicates whether a local NTP service is available or not.
15605 * systemd-detect-virt will now also detect xen PVs
15607 * The pstore file system is now mounted by default, if it is
15610 * In addition to the SELinux and IMA policies we will now also
15611 load SMACK policies at early boot.
15613 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
15614 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
15615 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
15616 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
15617 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
15618 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
15619 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
15620 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
15621 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
15622 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
15623 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
15624 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
15625 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
15626 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
15630 * Timer units now support calendar time events in addition to
15631 monotonic time events. That means you can now trigger a unit
15632 based on a calendar time specification such as "Thu,Fri
15633 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
15634 or fifth day of any month of the year 2013, given that it is
15635 a Thursday or a Friday. This brings timer event support
15636 considerably closer to cron's capabilities. For details on
15637 the supported calendar time specification language see
15640 * udev now supports a number of different naming policies for
15641 network interfaces for predictable names, and a combination
15642 of these policies is now the default. Please see this wiki
15643 document for details:
15645 https://www.freedesktop.org/software/systemd/man/systemd.net-naming-scheme.html
15647 * Auke Kok's bootchart implementation has been added to the
15648 systemd tree. It is an optional component that can graph the
15649 boot in quite some detail. It is one of the best bootchart
15650 implementations around and minimal in its code and
15653 * nss-myhostname has been integrated into the systemd source
15654 tree. nss-myhostname guarantees that the local hostname
15655 always stays resolvable via NSS. It has been a weak
15656 requirement of systemd-hostnamed since a long time, and
15657 since its code is actually trivial we decided to just
15658 include it in systemd's source tree. It can be turned off
15659 with a configure switch.
15661 * The read-ahead logic is now capable of properly detecting
15662 whether a btrfs file system is on SSD or rotating media, in
15663 order to optimize the read-ahead scheme. Previously, it was
15664 only capable of detecting this on traditional file systems
15667 * In udev, additional device properties are now read from the
15668 IAB in addition to the OUI database. Also, Bluetooth company
15669 identities are attached to the devices as well.
15671 * In service files %U may be used as specifier that is
15672 replaced by the configured user name of the service.
15674 * nspawn may now be invoked without a controlling TTY. This
15675 makes it suitable for invocation as its own service. This
15676 may be used to set up a simple containerized server system
15677 using only core OS tools.
15679 * systemd and nspawn can now accept socket file descriptors
15680 when they are started for socket activation. This enables
15681 implementation of socket activated nspawn
15682 containers. i.e. think about autospawning an entire OS image
15683 when the first SSH or HTTP connection is received. We expect
15684 that similar functionality will also be added to libvirt-lxc
15687 * journalctl will now suppress ANSI color codes when
15688 presenting log data.
15690 * systemctl will no longer show control group information for
15691 a unit if the control group is empty anyway.
15693 * logind can now automatically suspend/hibernate/shutdown the
15696 * /etc/machine-info and hostnamed now also expose the chassis
15697 type of the system. This can be used to determine whether
15698 the local system is a laptop, desktop, handset or
15699 tablet. This information may either be configured by the
15700 user/vendor or is automatically determined from ACPI and DMI
15701 information if possible.
15703 * A number of polkit actions are now bound together with "imply"
15704 rules. This should simplify creating UIs because many actions
15705 will now authenticate similar ones as well.
15707 * Unit files learnt a new condition ConditionACPower= which
15708 may be used to conditionalize a unit depending on whether an
15709 AC power source is connected or not, of whether the system
15710 is running on battery power.
15712 * systemctl gained a new "is-failed" verb that may be used in
15713 shell scripts and suchlike to check whether a specific unit
15714 is in the "failed" state.
15716 * The EnvironmentFile= setting in unit files now supports file
15717 globbing, and can hence be used to easily read a number of
15718 environment files at once.
15720 * systemd will no longer detect and recognize specific
15721 distributions. All distribution-specific #ifdeffery has been
15722 removed, systemd is now fully generic and
15723 distribution-agnostic. Effectively, not too much is lost as
15724 a lot of the code is still accessible via explicit configure
15725 switches. However, support for some distribution specific
15726 legacy configuration file formats has been dropped. We
15727 recommend distributions to simply adopt the configuration
15728 files everybody else uses now and convert the old
15729 configuration from packaging scripts. Most distributions
15730 already did that. If that's not possible or desirable,
15731 distributions are welcome to forward port the specific
15732 pieces of code locally from the git history.
15734 * When logging a message about a unit systemd will now always
15735 log the unit name in the message meta data.
15737 * localectl will now also discover system locale data that is
15738 not stored in locale archives, but directly unpacked.
15740 * logind will no longer unconditionally use framebuffer
15741 devices as seat masters, i.e. as devices that are required
15742 to be existing before a seat is considered preset. Instead,
15743 it will now look for all devices that are tagged as
15744 "seat-master" in udev. By default, framebuffer devices will
15745 be marked as such, but depending on local systems, other
15746 devices might be marked as well. This may be used to
15747 integrate graphics cards using closed source drivers (such
15748 as NVidia ones) more nicely into logind. Note however, that
15749 we recommend using the open source NVidia drivers instead,
15750 and no udev rules for the closed-source drivers will be
15751 shipped from us upstream.
15753 Contributions from: Adam Williamson, Alessandro Crismani, Auke
15754 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
15755 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
15756 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
15757 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
15758 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
15759 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
15760 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
15761 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
15762 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
15763 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
15764 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
15769 * udev gained support for loading additional device properties
15770 from an indexed database that is keyed by vendor/product IDs
15771 and similar device identifiers. For the beginning this
15772 "hwdb" is populated with data from the well-known PCI and
15773 USB database, but also includes PNP, ACPI and OID data. In
15774 the longer run this indexed database shall grow into
15775 becoming the one central database for non-essential
15776 userspace device metadata. Previously, data from the PCI/USB
15777 database was only attached to select devices, since the
15778 lookup was a relatively expensive operation due to O(n) time
15779 complexity (with n being the number of entries in the
15780 database). Since this is now O(1), we decided to add in this
15781 data for all devices where this is available, by
15782 default. Note that the indexed database needs to be rebuilt
15783 when new data files are installed. To achieve this you need
15784 to update your packaging scripts to invoke "udevadm hwdb
15785 --update" after installation of hwdb data files. For
15786 RPM-based distributions we introduced the new
15787 %udev_hwdb_update macro for this purpose.
15789 * The Journal gained support for the "Message Catalog", an
15790 indexed database to link up additional information with
15791 journal entries. For further details please check:
15793 https://www.freedesktop.org/wiki/Software/systemd/catalog
15795 The indexed message catalog database also needs to be
15796 rebuilt after installation of message catalog files. Use
15797 "journalctl --update-catalog" for this. For RPM-based
15798 distributions we introduced the %journal_catalog_update
15799 macro for this purpose.
15801 * The Python Journal bindings gained support for the standard
15802 Python logging framework.
15804 * The Journal API gained new functions for checking whether
15805 the underlying file system of a journal file is capable of
15806 properly reporting file change notifications, or whether
15807 applications that want to reflect journal changes "live"
15808 need to recheck journal files continuously in appropriate
15811 * It is now possible to set the "age" field for tmpfiles
15812 entries to 0, indicating that files matching this entry
15813 shall always be removed when the directories are cleaned up.
15815 * coredumpctl gained a new "gdb" verb which invokes gdb
15816 right-away on the selected coredump.
15818 * There's now support for "hybrid sleep" on kernels that
15819 support this, in addition to "suspend" and "hibernate". Use
15820 "systemctl hybrid-sleep" to make use of this.
15822 * logind's HandleSuspendKey= setting (and related settings)
15823 now gained support for a new "lock" setting to simply
15824 request the screen lock on all local sessions, instead of
15825 actually executing a suspend or hibernation.
15827 * systemd will now mount the EFI variables file system by
15830 * Socket units now gained support for configuration of the
15831 SMACK security label.
15833 * timedatectl will now output the time of the last and next
15834 daylight saving change.
15836 * We dropped support for various legacy and distro-specific
15837 concepts, such as insserv, early-boot SysV services
15838 (i.e. those for non-standard runlevels such as 'b' or 'S')
15839 or ArchLinux /etc/rc.conf support. We recommend the
15840 distributions who still need support this to either continue
15841 to maintain the necessary patches downstream, or find a
15842 different solution. (Talk to us if you have questions!)
15844 * Various systemd components will now bypass polkit checks for
15845 root and otherwise handle properly if polkit is not found to
15846 be around. This should fix most issues for polkit-less
15847 systems. Quite frankly this should have been this way since
15848 day one. It is absolutely our intention to make systemd work
15849 fine on polkit-less systems, and we consider it a bug if
15850 something does not work as it should if polkit is not around.
15852 * For embedded systems it is now possible to build udev and
15853 systemd without blkid and/or kmod support.
15855 * "systemctl switch-root" is now capable of switching root
15856 more than once. I.e. in addition to transitions from the
15857 initrd to the host OS it is now possible to transition to
15858 further OS images from the host. This is useful to implement
15859 offline updating tools.
15861 * Various other additions have been made to the RPM macros
15862 shipped with systemd. Use %udev_rules_update() after
15863 installing new udev rules files. %_udevhwdbdir,
15864 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
15865 %_sysctldir are now available which resolve to the right
15866 directories for packages to place various data files in.
15868 * journalctl gained the new --full switch (in addition to
15869 --all, to disable ellipsation for long messages.
15871 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
15872 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
15873 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
15874 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
15875 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
15876 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
15877 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
15878 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
15879 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
15883 * journalctl gained new --since= and --until= switches to
15884 filter by time. It also now supports nice filtering for
15885 units via --unit=/-u.
15887 * Type=oneshot services may use ExecReload= and do the
15890 * The journal daemon now supports time-based rotation and
15891 vacuuming, in addition to the usual disk-space based
15894 * The journal will now index the available field values for
15895 each field name. This enables clients to show pretty drop
15896 downs of available match values when filtering. The bash
15897 completion of journalctl has been updated
15898 accordingly. journalctl gained a new switch -F to list all
15899 values a certain field takes in the journal database.
15901 * More service events are now written as structured messages
15902 to the journal, and made recognizable via message IDs.
15904 * The timedated, localed and hostnamed mini-services which
15905 previously only provided support for changing time, locale
15906 and hostname settings from graphical DEs such as GNOME now
15907 also have a minimal (but very useful) text-based client
15908 utility each. This is probably the nicest way to changing
15909 these settings from the command line now, especially since
15910 it lists available options and is fully integrated with bash
15913 * There's now a new tool "systemd-coredumpctl" to list and
15914 extract coredumps from the journal.
15916 * We now install a README each in /var/log/ and
15917 /etc/rc.d/init.d explaining where the system logs and init
15918 scripts went. This hopefully should help folks who go to
15919 that dirs and look into the otherwise now empty void and
15920 scratch their heads.
15922 * When user-services are invoked (by systemd --user) the
15923 $MANAGERPID env var is set to the PID of systemd.
15925 * SIGRTMIN+24 when sent to a --user instance will now result
15926 in immediate termination of systemd.
15928 * gatewayd received numerous feature additions such as a
15929 "follow" mode, for live syncing and filtering.
15931 * browse.html now allows filtering and showing detailed
15932 information on specific entries. Keyboard navigation and
15933 mouse screen support has been added.
15935 * gatewayd/journalctl now supports HTML5/JSON
15936 Server-Sent-Events as output.
15938 * The SysV init script compatibility logic will now
15939 heuristically determine whether a script supports the
15940 "reload" verb, and only then make this available as
15941 "systemctl reload".
15943 * "systemctl status --follow" has been removed, use "journalctl
15946 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
15947 have been removed since they are hardly useful to be
15950 * And I'd like to take the opportunity to specifically mention
15951 Zbigniew for his great contributions. Zbigniew, you rock!
15953 Contributions from: Andrew Eikum, Christian Hesse, Colin
15954 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
15955 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
15956 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
15957 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
15958 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
15959 Jędrzejewski-Szmek, Сковорода Никита Андреевич
15963 * If /etc/vconsole.conf is non-existent or empty we will no
15964 longer load any console font or key map at boot by
15965 default. Instead the kernel defaults will be left
15966 intact. This is definitely the right thing to do, as no
15967 configuration should mean no configuration, and hard-coding
15968 font names that are different on all archs is probably a bad
15969 idea. Also, the kernel default key map and font should be
15970 good enough for most cases anyway, and mostly identical to
15971 the userspace fonts/key maps we previously overloaded them
15972 with. If distributions want to continue to default to a
15973 non-kernel font or key map they should ship a default
15974 /etc/vconsole.conf with the appropriate contents.
15976 Contributions from: Colin Walters, Daniel J Walsh, Dave
15977 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
15978 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
15982 * journalctl gained a new --cursor= switch to show entries
15983 starting from the specified location in the journal.
15985 * We now enforce a size limit on journal entry fields exported
15986 with "-o json" in journalctl. Fields larger than 4K will be
15987 assigned null. This can be turned off with --all.
15989 * An (optional) journal gateway daemon is now available as
15990 "systemd-journal-gatewayd.service". This service provides
15991 access to the journal via HTTP and JSON. This functionality
15992 will be used to implement live log synchronization in both
15993 pull and push modes, but has various other users too, such
15994 as easy log access for debugging of embedded devices. Right
15995 now it is already useful to retrieve the journal via HTTP:
15997 # systemctl start systemd-journal-gatewayd.service
15998 # wget http://localhost:19531/entries
16000 This will download the journal contents in a
16001 /var/log/messages compatible format. The same as JSON:
16003 # curl -H"Accept: application/json" http://localhost:19531/entries
16005 This service is also accessible via a web browser where a
16006 single static HTML5 app is served that uses the JSON logic
16007 to enable the user to do some basic browsing of the
16008 journal. This will be extended later on. Here's an example
16009 screenshot of this app in its current state:
16011 https://0pointer.de/public/journal-gatewayd
16013 Contributions from: Kay Sievers, Lennart Poettering, Robert
16014 Milasan, Tom Gundersen
16018 * The bash completion logic is now available for journalctl
16021 * We do not mount the "cpuset" controller anymore together with
16022 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
16023 started if no parameters are assigned to it. "cpuset" hence
16024 broke code that assumed it could create "cpu" groups and
16027 * journalctl -f will now subscribe to terminal size changes,
16028 and line break accordingly.
16030 Contributions from: Dave Reisner, Kay Sievers, Lennart
16031 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
16035 * nspawn will now create a symlink /etc/localtime in the
16036 container environment, copying the host's timezone
16037 setting. Previously this has been done via a bind mount, but
16038 since symlinks cannot be bind mounted this has now been
16039 changed to create/update the appropriate symlink.
16041 * journalctl -n's line number argument is now optional, and
16042 will default to 10 if omitted.
16044 * journald will now log the maximum size the journal files may
16045 take up on disk. This is particularly useful if the default
16046 built-in logic of determining this parameter from the file
16047 system size is used. Use "systemctl status
16048 systemd-journald.service" to see this information.
16050 * The multi-seat X wrapper tool has been stripped down. As X
16051 is now capable of enumerating graphics devices via udev in a
16052 seat-aware way the wrapper is not strictly necessary
16053 anymore. A stripped down temporary stop-gap is still shipped
16054 until the upstream display managers have been updated to
16055 fully support the new X logic. Expect this wrapper to be
16056 removed entirely in one of the next releases.
16058 * HandleSleepKey= in logind.conf has been split up into
16059 HandleSuspendKey= and HandleHibernateKey=. The old setting
16060 is not available anymore. X11 and the kernel are
16061 distinguishing between these keys and we should too. This
16062 also means the inhibition lock for these keys has been split
16065 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
16066 Poettering, Lukas Nykryn, Václav Pavlín
16070 * Whenever a unit changes state we will now log this to the
16071 journal and show along the unit's own log output in
16072 "systemctl status".
16074 * ConditionPathIsMountPoint= can now properly detect bind
16075 mount points too. (Previously, a bind mount of one file
16076 system to another place in the same file system could not be
16077 detected as mount, since they shared struct stat's st_dev
16080 * We will now mount the cgroup controllers cpu, cpuacct,
16081 cpuset and the controllers net_cls, net_prio together by
16084 * nspawn containers will now have a virtualized boot
16085 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
16086 over with a randomized ID at container initialization). This
16087 has the effect of making "journalctl -b" do the right thing
16090 * The JSON output journal serialization has been updated not
16091 to generate "endless" list objects anymore, but rather one
16092 JSON object per line. This is more in line how most JSON
16093 parsers expect JSON objects. The new output mode
16094 "json-pretty" has been added to provide similar output, but
16095 neatly aligned for readability by humans.
16097 * We dropped all explicit sync() invocations in the shutdown
16098 code. The kernel does this implicitly anyway in the kernel
16099 reboot() syscall. halt(8)'s -n option is now a compatibility
16102 * We now support virtualized reboot() in containers, as
16103 supported by newer kernels. We will fall back to exit() if
16104 CAP_SYS_REBOOT is not available to the container. Also,
16105 nspawn makes use of this now and will actually reboot the
16106 container if the containerized OS asks for that.
16108 * journalctl will only show local log output by default
16109 now. Use --merge (-m) to show remote log output, too.
16111 * libsystemd-journal gained the new sd_journal_get_usage()
16112 call to determine the current disk usage of all journal
16113 files. This is exposed in the new "journalctl --disk-usage"
16116 * journald gained a new configuration setting SplitMode= in
16117 journald.conf which may be used to control how user journals
16118 are split off. See journald.conf(5) for details.
16120 * A new condition type ConditionFileNotEmpty= has been added.
16122 * tmpfiles' "w" lines now support file globbing, to write
16123 multiple files at once.
16125 * We added Python bindings for the journal submission
16126 APIs. More Python APIs for a number of selected APIs will
16127 likely follow. Note that we intend to add native bindings
16128 only for the Python language, as we consider it common
16129 enough to deserve bindings shipped within systemd. There are
16130 various projects outside of systemd that provide bindings
16131 for languages such as PHP or Lua.
16133 * Many conditions will now resolve specifiers such as %i. In
16134 addition, PathChanged= and related directives of .path units
16135 now support specifiers as well.
16137 * There's now a new RPM macro definition for the system preset
16140 * journald will now warn if it ca not forward a message to the
16141 syslog daemon because its socket is full.
16143 * timedated will no longer write or process /etc/timezone,
16144 except on Debian. As we do not support late mounted /usr
16145 anymore /etc/localtime always being a symlink is now safe,
16146 and hence the information in /etc/timezone is not necessary
16149 * logind will now always reserve one VT for a text getty (VT6
16150 by default). Previously if more than 6 X sessions where
16151 started they took up all the VTs with auto-spawned gettys,
16152 so that no text gettys were available anymore.
16154 * udev will now automatically inform the btrfs kernel logic
16155 about btrfs RAID components showing up. This should make
16156 simple hotplug based btrfs RAID assembly work.
16158 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
16159 (but not for its children which will stay at the kernel
16160 default). This should allow setups with a lot more listening
16163 * systemd will now always pass the configured timezone to the
16164 kernel at boot. timedated will do the same when the timezone
16167 * logind's inhibition logic has been updated. By default,
16168 logind will now handle the lid switch, the power and sleep
16169 keys all the time, even in graphical sessions. If DEs want
16170 to handle these events on their own they should take the new
16171 handle-power-key, handle-sleep-key and handle-lid-switch
16172 inhibitors during their runtime. A simple way to achieve
16173 that is to invoke the DE wrapped in an invocation of:
16175 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch …
16177 * Access to unit operations is now checked via SELinux taking
16178 the unit file label and client process label into account.
16180 * systemd will now notify the administrator in the journal
16181 when he over-mounts a non-empty directory.
16183 * There are new specifiers that are resolved in unit files,
16184 for the hostname (%H), the machine ID (%m) and the boot ID
16187 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
16188 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
16189 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
16190 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
16191 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
16192 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
16193 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
16197 * Support for reading structured kernel messages from
16198 /dev/kmsg has now been added and is enabled by default.
16200 * Support for reading kernel messages from /proc/kmsg has now
16201 been removed. If you want kernel messages in the journal
16202 make sure to run a recent kernel (>= 3.5) that supports
16203 reading structured messages from /dev/kmsg (see
16204 above). /proc/kmsg is now exclusive property of classic
16205 syslog daemons again.
16207 * The libudev API gained the new
16208 udev_device_new_from_device_id() call.
16210 * The logic for file system namespace (ReadOnlyDirectory=,
16211 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
16212 require pivot_root() anymore. This means fewer temporary
16213 directories are created below /tmp for this feature.
16215 * nspawn containers will now see and receive all submounts
16216 made on the host OS below the root file system of the
16219 * Forward Secure Sealing is now supported for Journal files,
16220 which provide cryptographical sealing of journal files so
16221 that attackers cannot alter log history anymore without this
16222 being detectable. Lennart will soon post a blog story about
16223 this explaining it in more detail.
16225 * There are two new service settings RestartPreventExitStatus=
16226 and SuccessExitStatus= which allow configuration of exit
16227 status (exit code or signal) which will be excepted from the
16228 restart logic, resp. consider successful.
16230 * journalctl gained the new --verify switch that can be used
16231 to check the integrity of the structure of journal files and
16232 (if Forward Secure Sealing is enabled) the contents of
16235 * nspawn containers will now be run with /dev/stdin, /dev/fd/
16236 and similar symlinks pre-created. This makes running shells
16237 as container init process a lot more fun.
16239 * The fstab support can now handle PARTUUID= and PARTLABEL=
16242 * A new ConditionHost= condition has been added to match
16243 against the hostname (with globs) and machine ID. This is
16244 useful for clusters where a single OS image is used to
16245 provision a large number of hosts which shall run slightly
16246 different sets of services.
16248 * Services which hit the restart limit will now be placed in a
16251 Contributions from: Bertram Poettering, Dave Reisner, Huang
16252 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
16253 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
16257 * When running in --user mode systemd will now become a
16258 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
16259 tree a lot more organized.
16261 * A new PartOf= unit dependency type has been introduced that
16262 may be used to group services in a natural way.
16264 * "systemctl enable" may now be used to enable instances of
16267 * journalctl now prints error log levels in red, and
16268 warning/notice log levels in bright white. It also supports
16269 filtering by log level now.
16271 * cgtop gained a new -n switch (similar to top), to configure
16272 the maximum number of iterations to run for. It also gained
16273 -b, to run in batch mode (accepting no input).
16275 * The suffix ".service" may now be omitted on most systemctl
16276 command lines involving service unit names.
16278 * There's a new bus call in logind to lock all sessions, as
16279 well as a loginctl verb for it "lock-sessions".
16281 * libsystemd-logind.so gained a new call sd_journal_perror()
16282 that works similar to libc perror() but logs to the journal
16283 and encodes structured information about the error number.
16285 * /etc/crypttab entries now understand the new keyfile-size=
16288 * shutdown(8) now can send a (configurable) wall message when
16289 a shutdown is cancelled.
16291 * The mount propagation mode for the root file system will now
16292 default to "shared", which is useful to make containers work
16293 nicely out-of-the-box so that they receive new mounts from
16294 the host. This can be undone locally by running "mount
16295 --make-rprivate /" if needed.
16297 * The prefdm.service file has been removed. Distributions
16298 should maintain this unit downstream if they intend to keep
16299 it around. However, we recommend writing normal unit files
16300 for display managers instead.
16302 * Since systemd is a crucial part of the OS we will now
16303 default to a number of compiler switches that improve
16304 security (hardening) such as read-only relocations, stack
16305 protection, and suchlike.
16307 * The TimeoutSec= setting for services is now split into
16308 TimeoutStartSec= and TimeoutStopSec= to allow configuration
16309 of individual time outs for the start and the stop phase of
16312 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
16313 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
16314 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
16315 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
16316 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
16317 Gundersen, Zbigniew Jędrzejewski-Szmek
16321 * The journal and id128 C APIs are now fully documented as man
16324 * Extra safety checks have been added when transitioning from
16325 the initial RAM disk to the main system to avoid accidental
16328 * /etc/crypttab entries now understand the new keyfile-offset=
16331 * systemctl -t can now be used to filter by unit load state.
16333 * The journal C API gained the new sd_journal_wait() call to
16334 make writing synchronous journal clients easier.
16336 * journalctl gained the new -D switch to show journals from a
16337 specific directory.
16339 * journalctl now displays a special marker between log
16340 messages of two different boots.
16342 * The journal is now explicitly flushed to /var via a service
16343 systemd-journal-flush.service, rather than implicitly simply
16344 by seeing /var/log/journal to be writable.
16346 * journalctl (and the journal C APIs) can now match for much
16347 more complex expressions, with alternatives and
16350 * When transitioning from the initial RAM disk to the main
16351 system we will now kill all processes in a killing spree to
16352 ensure no processes stay around by accident.
16354 * Three new specifiers may be used in unit files: %u, %h, %s
16355 resolve to the user name, user home directory resp. user
16356 shell. This is useful for running systemd user instances.
16358 * We now automatically rotate journal files if their data
16359 object hash table gets a fill level > 75%. We also size the
16360 hash table based on the configured maximum file size. This
16361 together should lower hash collisions drastically and thus
16362 speed things up a bit.
16364 * journalctl gained the new "--header" switch to introspect
16365 header data of journal files.
16367 * A new setting SystemCallFilters= has been added to services which may
16368 be used to apply deny lists or allow lists to system calls. This is
16369 based on SECCOMP Mode 2 of Linux 3.5.
16371 * nspawn gained a new --link-journal= switch (and quicker: -j)
16372 to link the container journal with the host. This makes it
16373 very easy to centralize log viewing on the host for all
16374 guests while still keeping the journal files separated.
16376 * Many bugfixes and optimizations
16378 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
16379 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
16380 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
16385 * Several tools now understand kernel command line arguments,
16386 which are only read when run in an initial RAM disk. They
16387 usually follow closely their normal counterparts, but are
16390 * There's a new tool to analyze the readahead files that are
16391 automatically generated at boot. Use:
16393 /usr/lib/systemd/systemd-readahead analyze /.readahead
16395 * We now provide an early debug shell on tty9 if this enabled. Use:
16397 systemctl enable debug-shell.service
16399 * All plymouth related units have been moved into the Plymouth
16400 package. Please make sure to upgrade your Plymouth version
16403 * systemd-tmpfiles now supports getting passed the basename of
16404 a configuration file only, in which case it will look for it
16405 in all appropriate directories automatically.
16407 * udevadm info now takes a /dev or /sys path as argument, and
16408 does the right thing. Example:
16410 udevadm info /dev/sda
16411 udevadm info /sys/class/block/sda
16413 * systemctl now prints a warning if a unit is stopped but a
16414 unit that might trigger it continues to run. Example: a
16415 service is stopped but the socket that activates it is left
16418 * "systemctl status" will now mention if the log output was
16419 shortened due to rotation since a service has been started.
16421 * The journal API now exposes functions to determine the
16422 "cutoff" times due to rotation.
16424 * journald now understands SIGUSR1 and SIGUSR2 for triggering
16425 immediately flushing of runtime logs to /var if possible,
16426 resp. for triggering immediate rotation of the journal
16429 * It is now considered an error if a service is attempted to
16430 be stopped that is not loaded.
16432 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
16434 * systemd-analyze now supports Python 3
16436 * tmpfiles now supports cleaning up directories via aging
16437 where the first level dirs are always kept around but
16438 directories beneath it automatically aged. This is enabled
16439 by prefixing the age field with '~'.
16441 * Seat objects now expose CanGraphical, CanTTY properties
16442 which is required to deal with very fast bootups where the
16443 display manager might be running before the graphics drivers
16444 completed initialization.
16446 * Seat objects now expose a State property.
16448 * We now include RPM macros for service enabling/disabling
16449 based on the preset logic. We recommend RPM based
16450 distributions to make use of these macros if possible. This
16451 makes it simpler to reuse RPM spec files across
16454 * We now make sure that the collected systemd unit name is
16455 always valid when services log to the journal via
16458 * There's a new man page kernel-command-line(7) detailing all
16459 command line options we understand.
16461 * The fstab generator may now be disabled at boot by passing
16462 fstab=0 on the kernel command line.
16464 * A new kernel command line option modules-load= is now understood
16465 to load a specific kernel module statically, early at boot.
16467 * Unit names specified on the systemctl command line are now
16468 automatically escaped as needed. Also, if file system or
16469 device paths are specified they are automatically turned
16470 into the appropriate mount or device unit names. Example:
16472 systemctl status /home
16473 systemctl status /dev/sda
16475 * The SysVConsole= configuration option has been removed from
16476 system.conf parsing.
16478 * The SysV search path is no longer exported on the D-Bus
16481 * The Names= option has been removed from unit file parsing.
16483 * There's a new man page bootup(7) detailing the boot process.
16485 * Every unit and every generator we ship with systemd now
16486 comes with full documentation. The self-explanatory boot is
16489 * A couple of services gained "systemd-" prefixes in their
16490 name if they wrap systemd code, rather than only external
16491 code. Among them fsck@.service which is now
16492 systemd-fsck@.service.
16494 * The HaveWatchdog property has been removed from the D-Bus
16497 * systemd.confirm_spawn= on the kernel command line should now
16500 * There's a new man page crypttab(5) which details all options
16501 we actually understand.
16503 * systemd-nspawn gained a new --capability= switch to pass
16504 additional capabilities to the container.
16506 * timedated will now read known NTP implementation unit names
16507 from /usr/lib/systemd/ntp-units.d/*.list,
16508 systemd-timedated-ntp.target has been removed.
16510 * journalctl gained a new switch "-b" that lists log data of
16511 the current boot only.
16513 * The notify socket is in the abstract namespace again, in
16514 order to support daemons which chroot() at start-up.
16516 * There is a new Storage= configuration option for journald
16517 which allows configuration of where log data should go. This
16518 also provides a way to disable journal logging entirely, so
16519 that data collected is only forwarded to the console, the
16520 kernel log buffer or another syslog implementation.
16522 * Many bugfixes and optimizations
16524 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
16525 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
16526 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
16527 Shawn Landden, Tom Gundersen
16531 * "systemctl help <unit>" now shows the man page if one is
16534 * Several new man pages have been added.
16536 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
16537 MaxLevelConsole= can now be specified in
16538 journald.conf. These options allow reducing the amount of
16539 data stored on disk or forwarded by the log level.
16541 * TimerSlackNSec= can now be specified in system.conf for
16542 PID1. This allows system-wide power savings.
16544 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
16545 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
16550 * logind is now capable of (optionally) handling power and
16551 sleep keys as well as the lid switch.
16553 * journalctl now understands the syntax "journalctl
16554 /usr/bin/avahi-daemon" to get all log output of a specific
16557 * CapabilityBoundingSet= in system.conf now also influences
16558 the capability bound set of usermode helpers of the kernel.
16560 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
16561 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
16562 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
16563 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
16567 * Note that we skipped 139 releases here in order to set the
16568 new version to something that is greater than both udev's
16569 and systemd's most recent version number.
16571 * udev: all udev sources are merged into the systemd source tree now.
16572 All future udev development will happen in the systemd tree. It
16573 is still fully supported to use the udev daemon and tools without
16574 systemd running, like in initramfs or other init systems. Building
16575 udev though, will require the *build* of the systemd tree, but
16576 udev can be properly *run* without systemd.
16578 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
16579 should be used to create dead device nodes as workarounds for broken
16582 * udev: RUN+="socket:…" and udev_monitor_new_from_socket() is
16583 no longer supported. udev_monitor_new_from_netlink() needs to be
16584 used to subscribe to events.
16586 * udev: when udevd is started by systemd, processes which are left
16587 behind by forking them off of udev rules, are unconditionally cleaned
16588 up and killed now after the event handling has finished. Services or
16589 daemons must be started as systemd services. Services can be
16590 pulled-in by udev to get started, but they can no longer be directly
16591 forked by udev rules.
16593 * udev: the daemon binary is called systemd-udevd now and installed
16594 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
16595 to adapt to that, create symlink, or rename the binary after building
16598 * libudev no longer provides these symbols:
16599 udev_monitor_from_socket()
16600 udev_queue_get_failed_list_entry()
16601 udev_get_{dev,sys,run}_path()
16602 The versions number was bumped and symbol versioning introduced.
16604 * systemd-loginctl and systemd-journalctl have been renamed
16605 to loginctl and journalctl to match systemctl.
16607 * The config files: /etc/systemd/systemd-logind.conf and
16608 /etc/systemd/systemd-journald.conf have been renamed to
16609 logind.conf and journald.conf. Package updates should rename
16610 the files to the new names on upgrade.
16612 * For almost all files the license is now LGPL2.1+, changed
16613 from the previous GPL2.0+. Exceptions are some minor stuff
16614 of udev (which will be changed to LGPL2.1 eventually, too),
16615 and the MIT licensed sd-daemon.[ch] library that is suitable
16616 to be used as drop-in files.
16618 * systemd and logind now handle system sleep states, in
16619 particular suspending and hibernating.
16621 * logind now implements a sleep/shutdown/idle inhibiting logic
16622 suitable for a variety of uses. Soonishly Lennart will blog
16623 about this in more detail.
16625 * var-run.mount and var-lock.mount are no longer provided
16626 (which previously bind mounted these directories to their new
16627 places). Distributions which have not converted these
16628 directories to symlinks should consider stealing these files
16629 from git history and add them downstream.
16631 * We introduced the Documentation= field for units and added
16632 this to all our shipped units. This is useful to make it
16633 easier to explore the boot and the purpose of the various
16636 * All smaller setup units (such as
16637 systemd-vconsole-setup.service) now detect properly if they
16638 are run in a container and are skipped when
16639 appropriate. This guarantees an entirely noise-free boot in
16640 Linux container environments such as systemd-nspawn.
16642 * A framework for implementing offline system updates is now
16643 integrated, for details see:
16644 https://www.freedesktop.org/software/systemd/man/systemd.offline-updates.html
16646 * A new service type Type=idle is available now which helps us
16647 avoiding ugly interleaving of getty output and boot status
16650 * There's now a system-wide CapabilityBoundingSet= option to
16651 globally reduce the set of capabilities for the
16652 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
16653 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
16654 even CAP_NET_ADMIN system-wide for secure systems.
16656 * There are now system-wide DefaultLimitXXX= options to
16657 globally change the defaults of the various resource limits
16658 for all units started by PID 1.
16660 * Harald Hoyer's systemd test suite has been integrated into
16661 systemd which allows easy testing of systemd builds in qemu
16662 and nspawn. (This is really awesome! Ask us for details!)
16664 * The fstab parser is now implemented as generator, not inside
16667 * systemctl will now warn you if .mount units generated from
16668 /etc/fstab are out of date due to changes in fstab that
16669 have not been read by systemd yet.
16671 * systemd is now suitable for usage in initrds. Dracut has
16672 already been updated to make use of this. With this in place
16673 initrds get a slight bit faster but primarily are much
16674 easier to introspect and debug since "systemctl status" in
16675 the host system can be used to introspect initrd services,
16676 and the journal from the initrd is kept around too.
16678 * systemd-delta has been added, a tool to explore differences
16679 between user/admin configuration and vendor defaults.
16681 * PrivateTmp= now affects both /tmp and /var/tmp.
16683 * Boot time status messages are now much prettier and feature
16684 proper english language. Booting up systemd has never been
16687 * Read-ahead pack files now include the inode number of all
16688 files to pre-cache. When the inode changes the pre-caching
16689 is not attempted. This should be nicer to deal with updated
16690 packages which might result in changes of read-ahead
16693 * We now temporaritly lower the kernel's read_ahead_kb variable
16694 when collecting read-ahead data to ensure the kernel's
16695 built-in read-ahead does not add noise to our measurements
16696 of necessary blocks to pre-cache.
16698 * There's now RequiresMountsFor= to add automatic dependencies
16699 for all mounts necessary for a specific file system path.
16701 * MountAuto= and SwapAuto= have been removed from
16702 system.conf. Mounting file systems at boot has to take place
16705 * nspawn now learned a new switch --uuid= to set the machine
16706 ID on the command line.
16708 * nspawn now learned the -b switch to automatically search
16709 for an init system.
16711 * vt102 is now the default TERM for serial TTYs, upgraded from
16714 * systemd-logind now works on VT-less systems.
16716 * The build tree has been reorganized. The individual
16717 components now have directories of their own.
16719 * A new condition type ConditionPathIsReadWrite= is now available.
16721 * nspawn learned the new -C switch to create cgroups for the
16722 container in other hierarchies.
16724 * We now have support for hardware watchdogs, configurable in
16727 * The scheduled shutdown logic now has a public API.
16729 * We now mount /tmp as tmpfs by default, but this can be
16730 masked and /etc/fstab can override it.
16732 * Since udisks does not make use of /media anymore we are not
16733 mounting a tmpfs on it anymore.
16735 * journalctl gained a new --local switch to only interleave
16736 locally generated journal files.
16738 * We can now load the IMA policy at boot automatically.
16740 * The GTK tools have been split off into a systemd-ui.
16742 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
16743 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
16744 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
16745 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
16746 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
16747 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
16748 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
16749 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
16750 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
16755 * This is mostly a bugfix release
16757 * Support optional initialization of the machine ID from the
16758 KVM or container configured UUID.
16760 * Support immediate reboots with "systemctl reboot -ff"
16762 * Show /etc/os-release data in systemd-analyze output
16764 * Many bugfixes for the journal, including endianness fixes and
16765 ensuring that disk space enforcement works
16767 * sd-login.h is C++ compatible again
16769 * Extend the /etc/os-release format on request of the Debian
16772 * We now refuse non-UTF8 strings used in various configuration
16773 and unit files. This is done to ensure we do not pass invalid
16774 data over D-Bus or expose it elsewhere.
16776 * Register Mimo USB Screens as suitable for automatic seat
16779 * Read SELinux client context from journal clients in a race
16782 * Reorder configuration file lookup order. /etc now always
16783 overrides /run in order to allow the administrator to always
16784 and unconditionally override vendor-supplied or
16785 automatically generated data.
16787 * The various user visible bits of the journal now have man
16788 pages. We still lack man pages for the journal API calls
16791 * We now ship all man pages in HTML format again in the
16794 Contributions from: Dave Reisner, Dirk Eibach, Frederic
16795 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
16796 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
16801 * This is mostly a bugfix release
16803 * systems lacking /etc/os-release are no longer supported.
16805 * Various functionality updates to libsystemd-login.so
16807 * Track class of PAM logins to distinguish greeters from
16808 normal user logins.
16810 Contributions from: Kay Sievers, Lennart Poettering, Michael
16815 * This is an important bugfix release for v41.
16817 * Building man pages is now optional which should be useful
16818 for those building systemd from git but unwilling to install
16821 * Watchdog support for supervising services is now usable. In
16822 a future release support for hardware watchdogs
16823 (i.e. /dev/watchdog) will be added building on this.
16825 * Service start rate limiting is now configurable and can be
16826 turned off per service. When a start rate limit is hit a
16827 reboot can automatically be triggered.
16829 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
16831 Contributions from: Benjamin Franzke, Bill Nottingham,
16832 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
16833 Schmidt, Michał Górny, Piotr Drąg
16837 * The systemd binary is installed /usr/lib/systemd/systemd now;
16838 An existing /sbin/init symlink needs to be adapted with the
16841 * The code that loads kernel modules has been ported to invoke
16842 libkmod directly, instead of modprobe. This means we do not
16843 support systems with module-init-tools anymore.
16845 * Watchdog support is now already useful, but still not
16848 * A new kernel command line option systemd.setenv= is
16849 understood to set system wide environment variables
16850 dynamically at boot.
16852 * We now limit the set of capabilities of systemd-journald.
16854 * We now set SIGPIPE to ignore by default, since it only is
16855 useful in shell pipelines, and has little use in general
16856 code. This can be disabled with IgnoreSIPIPE=no in unit
16859 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
16860 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
16865 * This is mostly a bugfix release
16867 * We now expose the reason why a service failed in the
16868 "Result" D-Bus property.
16870 * Rudimentary service watchdog support (will be completed over
16871 the next few releases.)
16873 * When systemd forks off in order execute some service we will
16874 now immediately changes its argv[0] to reflect which process
16875 it will execute. This is useful to minimize the time window
16876 with a generic argv[0], which makes bootcharts more useful
16878 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
16879 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
16880 Mike Kazantsev, Ray Strode
16884 * This is mostly a test release, but incorporates many
16887 * New systemd-cgtop tool to show control groups by their
16890 * Linking against libacl for ACLs is optional again. If
16891 disabled, support tracking device access for active logins
16892 goes becomes unavailable, and so does access to the user
16893 journals by the respective users.
16895 * If a group "adm" exists, journal files are automatically
16896 owned by them, thus allow members of this group full access
16897 to the system journal as well as all user journals.
16899 * The journal now stores the SELinux context of the logging
16900 client for all entries.
16902 * Add C++ inclusion guards to all public headers
16904 * New output mode "cat" in the journal to print only text
16905 messages, without any meta data like date or time.
16907 * Include tiny X server wrapper as a temporary stop-gap to
16908 teach XOrg udev display enumeration. This is used by display
16909 managers such as gdm, and will go away as soon as XOrg
16910 learned native udev hotplugging for display devices.
16912 * Add new systemd-cat tool for executing arbitrary programs
16913 with STDERR/STDOUT connected to the journal. Can also act as
16914 BSD logger replacement, and does so by default.
16916 * Optionally store all locally generated coredumps in the
16917 journal along with meta data.
16919 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
16920 writing short strings to files (for usage for /sys), and for
16921 creating symlinks, character and block device nodes.
16923 * New unit file option ControlGroupPersistent= to make cgroups
16924 persistent, following the mechanisms outlined in
16925 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
16927 * Support multiple local RTCs in a sane way
16929 * No longer monopolize IO when replaying readahead data on
16930 rotating disks, since we might starve non-file-system IO to
16931 death, since fanotify() will not see accesses done by blkid,
16934 * Do not show kernel threads in systemd-cgls anymore, unless
16935 requested with new -k switch.
16937 Contributions from: Dan Horák, Kay Sievers, Lennart
16938 Poettering, Michal Schmidt
16942 * This is mostly a test release, but incorporates many
16945 * The git repository moved to:
16946 git://anongit.freedesktop.org/systemd/systemd
16947 ssh://git.freedesktop.org/git/systemd/systemd
16949 * First release with the journal
16950 https://0pointer.de/blog/projects/the-journal.html
16952 * The journal replaces both systemd-kmsg-syslogd and
16953 systemd-stdout-bridge.
16955 * New sd_pid_get_unit() API call in libsystemd-logind
16957 * Many systemadm clean-ups
16959 * Introduce remote-fs-pre.target which is ordered before all
16960 remote mounts and may be used to start services before all
16963 * Added Mageia support
16965 * Add bash completion for systemd-loginctl
16967 * Actively monitor PID file creation for daemons which exit in
16968 the parent process before having finished writing the PID
16969 file in the daemon process. Daemons which do this need to be
16970 fixed (i.e. PID file creation must have finished before the
16971 parent exits), but we now react a bit more gracefully to them.
16973 * Add colourful boot output, mimicking the well-known output
16974 of existing distributions.
16976 * New option PassCredentials= for socket units, for
16977 compatibility with a recent kernel ABI breakage.
16979 * /etc/rc.local is now hooked in via a generator binary, and
16980 thus will no longer act as synchronization point during
16983 * systemctl list-unit-files now supports --root=.
16985 * systemd-tmpfiles now understands two new commands: z, Z for
16986 relabelling files according to the SELinux database. This is
16987 useful to apply SELinux labels to specific files in /sys,
16988 among other things.
16990 * Output of SysV services is now forwarded to both the console
16991 and the journal by default, not only just the console.
16993 * New man pages for all APIs from libsystemd-login.
16995 * The build tree got reorganized and the build system is a
16996 lot more modular allowing embedded setups to specifically
16997 select the components of systemd they are interested in.
16999 * Support for Linux systems lacking the kernel VT subsystem is
17002 * configure's --with-rootdir= got renamed to
17003 --with-rootprefix= to follow the naming used by udev and
17006 * Unless specified otherwise we will now install to /usr instead
17007 of /usr/local by default.
17009 * Processes with '@' in argv[0][0] are now excluded from the
17010 final shut-down killing spree, following the logic explained
17012 https://systemd.io/ROOT_STORAGE_DAEMONS/
17014 * All processes remaining in a service cgroup when we enter
17015 the START or START_PRE states are now killed with
17016 SIGKILL. That means it is no longer possible to spawn
17017 background processes from ExecStart= lines (which was never
17018 supported anyway, and bad style).
17020 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
17021 reloading of units together.
17023 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
17024 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
17025 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
17026 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
17027 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
projects / thirdparty / systemd.git / blob