1 systemd System and Service Manager
3 CHANGES WITH 254 in spe:
5 Security relevant changes:
7 * pam_systemd will now by default pass the CAP_WAKE_ALARM ambient
8 process capability to invoked session processes of regular users on
9 local seats (as well as to systemd --user), unless configured
10 otherwise via data from JSON user records, or via the PAM module's
11 parameter list. This is useful in order allow desktop tools such as
12 GNOME's Alarm Clock application to set a timer for
13 CLOCK_REALTIME_ALARM that wakes up the system when it elapses. A
14 per-user service unit file may thus use AmbientCapability= to pass
15 the capability to invoked processes. Note that this capability is
16 relatively narrow in focus (in particular compared to other process
17 capabilities such as CAP_SYS_ADMIN) and we already — by default —
18 permit more impactful operations such as system suspend to local
23 Announcements of Future Feature Removals and Incompatible Changes:
25 * We intend to remove cgroup v1 support from systemd release after the
26 end of 2023. If you run services that make explicit use of cgroup v1
27 features (i.e. the "legacy hierarchy" with separate hierarchies for
28 each controller), please implement compatibility with cgroup v2 (i.e.
29 the "unified hierarchy") sooner rather than later. Most of Linux
30 userspace has been ported over already.
32 * We intend to remove support for split-usr (/usr mounted separately
33 during boot) and unmerged-usr (parallel directories /bin and
34 /usr/bin, /lib and /usr/lib, etc). This will happen in the second
35 half of 2023, in the first release that falls into that time window.
36 For more details, see:
37 https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html
39 * We intend to change behaviour w.r.t. units of the per-user service
40 manager and sandboxing options, so that they work without having to
41 manually enable PrivateUsers= as well, which is not required for
42 system units. To make this work, we will implicitly enable user
43 namespaces (PrivateUsers=yes) when a sandboxing option is enabled in a
44 user unit. The drawback is that system users will no longer be visible
45 (and appear as 'nobody') to the user unit when a sandboxing option is
46 enabled. By definition a sandboxed user unit should run with reduced
47 privileges, so impact should be small. This will remove a great source
48 of confusion that has been reported by users over the years, due to
49 how these options require an extra setting to be manually enabled when
50 used in the per-user service manager, as opposed as to the system
51 service manager. We plan to enable this change in the next release
52 later this year. For more details, see:
53 https://lists.freedesktop.org/archives/systemd-devel/2022-December/048682.html
55 Deprecations and incompatible changes:
57 * systemctl will now warn when invoked without /proc/ mounted
58 (e.g. when invoked after chroot() into an directory tree without the
59 API mount points like /proc/ being set up.) Operation in such an
60 environment is not fully supported.
62 * The return value of 'systemctl is-active|is-enabled|is-failed' for
63 unknown units is changed: previously 1 or 3 were returned, but now 4
64 (EXIT_PROGRAM_OR_SERVICES_STATUS_UNKNOWN) is used as documented.
66 * 'udevadm hwdb' subcommand is deprecated and will emit a warning.
67 systemd-hwdb (added in 2014) should be used instead.
69 * 'bootctl --json' now outputs a single JSON array, instead of a stream
70 of newline-separated JSON objects.
72 * Udev rules in 60-evdev.rules have been changed to load hwdb
73 properties for all modalias patterns. Previously only the first
74 matching pattern was used. This could change what properties are
75 assigned if the user has more and less specific patterns that could
76 match the same device, but it is expected that the change will have
77 no effect for most users.
79 * systemd-networkd-wait-online exits successfully when all interfaces
80 are ready or unmanaged. Previously, if neither '--any' nor
81 '--interface=' options were used, at least one interface had to be in
82 configured state. This change allows the case where systemd-networkd
83 is enabled, but no interfaces are configured, to be handled
84 gracefully. It may occur in particular when a different network
85 manager is also enabled and used.
87 * Some compatibility helpers were dropped: EmergencyAction= in the user
88 manager, as well as measuring kernel command line into PCR 8 in
89 systemd-stub, along with the -Defi-tpm-pcr-compat compile-time
92 * The '-Dupdate-helper-user-timeout=' build-time option has been
93 renamed to '-Dupdate-helper-user-timeout-sec=', and now takes an
94 integer as parameter instead of a string.
96 * The DDI image dissection logic (which backs RootImage= in service
97 unit files, the --image= switch in various tools such as
98 systemd-nspawn, as well as systemd-dissect) will now only mount file
99 systems of types btrfs, ext4, xfs, erofs, squashfs, vfat. This list
100 can be overridden via the $SYSTEMD_DISSECT_FILE_SYSTEMS environment
101 variable. These file systems are fairly well supported and maintained
102 in current kernels, while others are usually more niche, exotic or
103 legacy and thus typically do not receive the same level of security
106 * The default per-link multicast DNS mode is changed to "yes"
107 (that was previously "no"). As the default global multicast DNS mode
108 has been "yes" (but can be changed by the build option), now the
109 multicast DNS is enabled on all links by default. You can disable the
110 multicast DNS on all links by setting MulticastDNS= in resolved.conf,
111 or on an interface by calling "resolvectl mdns INTERFACE no".
115 * A tool 'ukify' tool to build, measure, and sign Unified Kernel Images
116 (UKIs) has been added. This replaces functionality provided by
117 'dracut --uefi' and extends it with automatic calculation of PE file
118 offsets, insertion of signed PCR policies generated by
119 systemd-measure, support for initrd concatenation, signing of the
120 embedded Linux image and the combined image with sbsign, and
121 heuristics to autodetect the kernel uname and verify the splash
124 Changes in systemd and units:
126 * A new service type Type=notify-reload is defined. When such a unit is
127 reloaded a UNIX process signal (typically SIGHUP) is sent to the main
128 service process. The manager will then wait until it receives a
129 "RELOADING=1" followed by a "READY=1" notification from the unit as
130 response (via sd_notify()). Otherwise, this type is the same as
131 Type=notify. A new setting ReloadSignal= may be used to change the
132 signal to send from the default of SIGHUP.
134 user@.service, systemd-networkd.service, systemd-udevd.service, and
135 systemd-logind have been updated to this type.
137 * Initrd environments which are not on a pure memory file system (e.g.
138 overlayfs combination as opposed to tmpfs) are now supported. With
139 this change, during the initrd → host transition ("switch root")
140 systemd will erase all files of the initrd only when the initrd is
141 backed by a memory file system such as tmpfs.
143 * New per-unit MemoryZSwapMax= option has been added to configure
144 memory.zswap.max cgroup properties (the maximum amount of zswap
147 * A new LogFilterPatterns= option has been added for units. It may be
148 used to specify accept/deny regular expressions for log messages
149 generated by the unit, that shall be enforced by systemd-journald.
150 Rejected messages are neither stored in the journal nor forwarded.
151 This option may be used to suppress noisy or uninteresting messages
154 * The manager has a new
155 org.freedesktop.systemd1.Manager.GetUnitByPIDFD() D-Bus method to
156 query process ownership via a PIDFD, which is more resilient against
157 PID recycling issues.
159 * Scope units now support OOMPolicy=. Login session scopes default to
160 OOMPolicy=continue, allowing login scopes to survive the OOM killer
161 terminating some processes in the scope.
163 * systemd-fstab-generator now supports x-systemd.makefs option for
164 /sysroot/ (in the initrd).
166 * The maximum rate at which daemon reloads are executed can now be
167 limited with the new ReloadLimitIntervalSec=/ReloadLimitBurst=
168 options. (Or the equivalent on the kernel command line:
169 systemd.reload_limit_interval_sec=/systemd.reload_limit_burst=). In
170 addition, systemd now logs the originating unit and PID when a reload
171 request is received over D-Bus.
173 * When enabling a swap device systemd will now reinitialize the device
174 when the page size of the swap space does not match the page size of
175 the running kernel. Note that this requires the 'swapon' utility to
176 provide the '--fixpgsz' option, as implemented by util-linux, and it
177 is not supported by busybox at the time of writing.
179 * systemd now executes generator programs in a mount namespace
180 "sandbox" with most of the file system read-only and write access
181 restricted to the output directories, and with a temporary /tmp/
182 mount provided. This provides a safeguard against programming errors
183 in the generators, but also fixes here-docs in shells, which
184 previously didn't work in early boot when /tmp/ wasn't available
185 yet. (This feature has no security implications, because the code is
186 still privileged and can trivially exit the sandbox.)
188 * The system manager will now parse a new "vmm.notify_socket"
189 system credential, which may be supplied to a VM via SMBIOS. If
190 found, the manager will send a "READY=1" notification on the
191 specified socket after boot is complete. This allows readiness
192 notification to be sent from a VM guest to the VM host over a VSOCK
195 * The sample PAM configuration file for systemd-user@.service now
196 includes a call to pam_namespace. This puts children of user@.service
197 in the expected namespace. (Many distributions replace their file
198 with something custom, so this change has limited effect.)
200 * A new environment variable $SYSTEMD_DEFAULT_MOUNT_RATE_LIMIT_BURST
201 can be used to override the mount units burst late limit for
202 parsing '/proc/self/mountinfo', which was introduced in v249.
205 * Drop-ins for init.scope changing control group resource limits are
206 now applied, while they were previously ignored.
208 * New build-time configuration options '-Ddefault-timeout-sec=' and
209 '-Ddefault-user-timeout-sec=' have been added, to let distributions
210 choose the default timeout for starting/stopping/aborting system and
211 user units respectively.
213 * Service units gained a new setting OpenFile= which may be used to
214 open arbitrary files in the file system (or connect to arbitrary
215 AF_UNIX sockets in the file system), and pass the open file
216 descriptor to the invoked process via the usual file descriptor
217 passing protocol. This is useful to give unprivileged services access
218 to select files which have restrictive access modes that would
219 normally not allow this. It's also useful in case RootDirectory= or
220 RootImage= is used to allow access to files from the host environment
221 (which is after all not visible from the service if these two options
226 * The new net naming scheme "v253" has been introduced. In the new
227 scheme, ID_NET_NAME_PATH is also set for USB devices not connected via
228 a PCI bus. This extends the coverage of predictable interface names
229 in some embedded systems.
231 The "amba" bus path is now included in ID_NET_NAME_PATH, resulting in
232 a more informative path on some embedded systems.
234 * Partition block devices will now also get symlinks in
235 /dev/disk/by-diskseq/<seq>-part<n>, which may be used to reference
236 block device nodes via the kernel's "diskseq" value. Previously those
237 symlinks were only created for the main block device.
239 * A new operator '-=' is supported for SYMLINK variables. This allows
240 symlinks to be unconfigured even if an earlier rule added them.
242 * 'udevadm --trigger --settle' now also works for network devices
243 that are being renamed.
245 Changes in sd-boot, bootctl, and the Boot Loader Specification:
247 * systemd-boot now passes its random seed directly to the kernel's RNG
248 via the LINUX_EFI_RANDOM_SEED_TABLE_GUID configuration table, which
249 means the RNG gets seeded very early in boot before userspace has
252 * systemd-boot will pass a disk-backed random seed – even when secure
253 boot is enabled – if it can additionally get a random seed from EFI
254 itself (via EFI's RNG protocol), or a prior seed in
255 LINUX_EFI_RANDOM_SEED_TABLE_GUID from a preceding bootloader.
257 * systemd-boot-system-token.service was renamed to
258 systemd-boot-random-seed.service and extended to always save a random
259 seed to ESP on every boot when a compatible boot loader is used. This
260 allows a refreshed random seed to be used in the boot loader.
262 * systemd-boot handles various seed inputs using a domain- and
263 field-separated hashing scheme.
265 * systemd-boot's 'random-seed-mode' option has been removed. A system
266 token is now always required to be present for random seeds to be
269 * systemd-boot now supports being loaded from other locations than the
270 ESP, for example for direct kernel boot under QEMU or when embedded
273 * systemd-boot now parses SMBIOS information to detect
274 virtualization. This information is used to skip some warnings which
275 are not useful in a VM and to conditionalize other aspects of
278 * systemd-boot now supports a new 'if-safe' mode that will perform UEFI
279 Secure Boot automated certificate enrollment from the ESP only if it
280 is considered 'safe' to do so. At the moment 'safe' means running in
283 * systemd-stub now processes random seeds in the same way as
284 systemd-boot already does, in case a unified kernel image is being
285 used from a different bootloader than systemd-boot, or without any
288 * bootctl will now generate a system token on all EFI systems, even
289 virtualized ones, and is activated in the case that the system token
290 is missing from either sd-boot and sd-stub booted systems.
292 * bootctl now implements two new verbs: 'kernel-identify' prints the
293 type of a kernel image file, and 'kernel-inspect' provides
294 information about the embedded command line and kernel version of
297 * bootctl now honours $KERNEL_INSTALL_CONF_ROOT with the same meaning
298 as for kernel-install.
300 * The JSON output of "bootctl list" will now contain two more fields:
301 isDefault and isSelected are boolean fields set to true on the
302 default and currently booted boot menu entries.
304 * bootctl gained a new verb "unlink" for removing a boot loader entry
305 type #1 file from disk in a safe and robust way.
307 * bootctl also gained a new verb "cleanup" that automatically removes
308 all files from the ESP's and XBOOTLDR's "entry-token" directory, that
309 is not referenced anymore by any installed Type #1 boot loader
310 specification entry. This is particularly useful in environments where
311 a large number of entries reference the same or partly the same
312 resources (for example, for snapshot-based setups).
314 Changes in kernel-install:
316 * A new "installation layout" can be configured as layout=uki. With
317 this setting, a Boot Loader Specification Type#1 entry will not be
318 created. Instead, a new kernel-install plugin 90-uki-copy.install
319 will copy any .efi files from the staging area into the boot
320 partition. A plugin to generate the UKI .efi file must be provided
323 Changes in systemctl:
325 * 'systemctl reboot' has dropped support for accepting a positional
326 argument as the argument to the reboot(2) syscall. Please use the
327 --reboot-argument= option instead.
329 * 'systemctl disable' will now warn when called on units without
330 install information. A new --no-warn option has been added that
331 silences this warning.
333 * New option '--drop-in=' can be used to tell 'systemctl edit' the name
334 of the drop-in to edit. (Previously, 'override.conf' was always
337 * 'systemctl list-dependencies' now respects --type= and --state=.
339 * 'systemctl kexec' now supports XEN VMM environments.
341 * 'systemctl edit' will now tell the invoked editor to jump into the
342 first line with actual unit file data, skipping over synthesized
345 Changes in systemd-networkd and related tools:
347 * The [DHCPv4] section in .network file gained new SocketPriority=
348 setting that assigns the Linux socket priority used by the DHCPv4 raw
349 socket. This may be used in conjunction with the
350 EgressQOSMaps=setting in [VLAN] section of .netdev file to send the
351 desired ethernet 802.1Q frame priority for DHCPv4 initial
352 packets. This cannot be achieved with netfilter mangle tables because
353 of the raw socket bypass.
355 * The [DHCPv4] and [IPv6AcceptRA] sections in .network file gained a
356 new QuickAck= boolean setting that enables the TCP quick ACK mode for
357 the routes configured by the acquired DHCPv4 lease or received router
358 advertisements (RAs).
360 * The RouteMetric= option (for DHCPv4, DHCPv6, and IPv6 advertised
361 routes) now accepts three values, for high, medium, and low preference
362 of the router (which can be set with the RouterPreference=) setting.
364 * systemd-networkd-wait-online now supports matching via alternative
367 * The [DHCPv6] section in .network file gained new SendRelease=
368 setting which enables the DHCPv6 client to send release when
369 it stops. This is the analog of the [DHCPv4] SendRelease= setting.
370 It is enabled by default.
372 * If the Address= setting in [Network] or [Address] sections in .network
373 specified without its prefix length, then now systemd-networkd assumes
374 /32 for IPv4 or /128 for IPv6 addresses.
376 * networkctl shows network and link file dropins in status output.
378 Changes in systemd-dissect:
380 * systemd-dissect gained a new option --list, to print the paths of
381 all files and directories in a DDI.
383 * systemd-dissect gained a new option --mtree, to generate a file
384 manifest compatible with BSD mtree(5) of a DDI
386 * systemd-dissect gained a new option --with, to execute a command with
387 the specified DDI temporarily mounted and used as working
388 directory. This is for example useful to convert a DDI to "tar"
389 simply by running it within a "systemd-dissect --with" invocation.
391 * systemd-dissect gained a new option --discover, to search for
392 Discoverable Disk Images (DDIs) in well-known directories of the
393 system. This will list machine, portable service and system extension
396 * systemd-dissect now understands 2nd stage initrd images stored as a
397 Discoverable Disk Image (DDI).
399 * systemd-dissect will now display the main UUID of GPT DDIs (i.e. the
400 disk UUID stored in the GPT header) among the other data it can show.
402 * systemd-dissect gained a new --in-memory switch to operate on an
403 in-memory copy of the specified DDI file. This is useful to access a
404 DDI with write access without persisting any changes. It's also
405 useful for accessing a DDI without keeping the originating file
408 * The DDI dissection logic will now automatically detect the intended
409 sector size of disk images stored in files, based on the GPT
410 partition table arrangement. Loopback block devices for such DDIs
411 will then be configured automatically for the right sector size. This
412 is useful to make dealing with modern 4K sector size DDIs fully
413 automatic. The systemd-dissect tool will now show the detected sector
414 size among the other DDI information in its output.
416 Changes in systemd-repart:
418 * systemd-repart gained new options --include-partitions= and
419 --exclude-partitions= to filter operation on partitions by type UUID.
420 This allows systemd-repart to be used to build images in which the
421 type of one partition is set based on the contents of another
422 partition (for example when the boot partition shall include a verity
423 hash of the root partition).
425 * systemd-repart also gained a --defer-partitions= option that is
426 similar to --exclude-partitions=, but the size of the partition is
427 still taken into account when sizing partitions, but without
430 * systemd-repart gained a new --sector-size= option to specify what
431 sector size should be used when an image is created.
433 * systemd-repart now supports generating erofs file systems via
434 CopyFiles= (a read-only file system similar to squashfs).
436 * The Minimize= option was extended to accept "best" (which means the
437 most minimal image possible, but may require multiple attempts) and
438 "guess" (which means a reasonably small image).
440 * The systemd-growfs binary now comes with a regular unit file template
441 systemd-growfs@.service which can be instantiated directly for any
442 desired file system. (Previously, the unit was generated dynamically
443 by various generators, but no regular unit file template was
446 Changes in journal tools:
448 * Various systemd tools will append extra fields to log messages when
449 in debug mode, or when SYSTEMD_ENABLE_LOG_CONTEXT=1 is set. Currently
450 this includes information about D-Bus messages when sd-bus is used,
451 e.g. DBUS_SENDER=, DBUS_DESTINATION=, and DBUS_PATH=, and information
452 about devices when sd-device is used, e.g. DEVNAME= and DRIVER=.
453 Details of what is logged and when are subject to change.
455 * The systemd-journald-audit.socket can now be disabled via the usual
456 "systemctl disable" mechanism to stop collection of audit
457 messages. Please note that it is not enabled statically anymore and
458 must be handled by the preset/enablement logic in package
459 installation scripts.
461 * New options MaxUse=, KeepFree=, MaxFileSize=, and MaxFiles= can
462 be used to curtail disk use by systemd-journal-remote. This is
463 similar to the options supported by systemd-journald.
465 Changes in systemd-cryptenroll, systemd-cryptsetup, and related
468 * When enrolling new keys systemd-cryptenroll now supports unlocking
469 via FIDO2 tokens (option --unlock-fido2-device=). Previously, a
470 password was strictly required to be specified.
472 * systemd-cryptsetup now supports pre-flight requests for FIDO2 tokens
473 (except for tokens with user verification, UV) to identify tokens
474 before authentication. Multiple FIDO2 tokens can now be enrolled at
475 the same time, and systemd-cryptsetup will automatically select one
476 that corresponds to one of the available LUKS key slots.
478 * systemd-cryptsetup now supports new options tpm2-measure-bank= and
479 tpm2-measure-pcr= in crypttab(5). These allow specifying the TPM2 PCR
480 bank and number into which the volume key should be measured. This is
481 automatically enabled for the encrypted root volume discovered and
482 activated by systemd-gpt-auto-generator.
484 * systemd-gpt-auto-generator mounts the ESP and XBOOTLDR partitions with
485 "noexec,nosuid,nodev".
487 * systemd-gpt-auto-generator will now honour the rootfstype= and
488 rootflags= kernel command line switches for root file systems it
489 discovers, to match behaviour in case an explicit root fs is
492 * systemd-pcrphase gained new options --machine-id and --file-system=
493 to measure the machine-id and mount point information into PCR 15. New
494 service unit files systemd-pcrmachine.service and
495 systemd-pcrfs@.service have been added that invoke the tool with
496 these switches during early boot.
498 * systemd-pcrphase gained a --graceful switch will make it exit cleanly
499 with a success exit code even if no TPM device is detected.
501 * systemd-cryptenroll now stores the user-supplied PIN with a salt,
502 making it harder to brute-force.
504 Changes in other tools:
506 * systemd-homed gained support for luksPbkdfForceIterations (the
507 intended number of iterations for the PBKDF operation on LUKS).
509 * Environment variables $SYSTEMD_HOME_MKFS_OPTIONS_BTRFS,
510 $SYSTEMD_HOME_MKFS_OPTIONS_EXT4, and $SYSTEMD_HOME_MKFS_OPTIONS_XFS
511 may now be used to specify additional arguments for mkfs when
512 systemd-homed formats a file system.
514 * systemd-hostnamed now exports the contents of
515 /sys/class/dmi/id/bios_vendor and /sys/class/dmi/id/bios_date via two
516 new D-Bus properties: FirmwareVendor and FirmwareDate. This allows
517 unprivileged code to access those values.
519 systemd-hostnamed also exports the SUPPORT_END= field from
520 os-release(5) as OperatingSystemSupportEnd. hostnamectl make uses of
521 this to show the status of the installed system.
523 * systemd-measure gained an --append= option to sign multiple phase
524 paths with different signing keys. This allows secrets to be
525 accessible only in certain parts of the boot sequence. Note that
526 'ukify' provides similar functionality in a more accessible form.
528 * systemd-timesyncd will now write a structured log message with
529 MESSAGE_ID set to SD_MESSAGE_TIME_BUMP when it bumps the clock based
530 on a on-disk timestamp, similarly to what it did when reaching
531 synchronization via NTP.
533 * systemd-timesyncd will now update the on-disk timestamp file on each
534 boot at least once, making it more likely that the system time
535 increases in subsequent boots.
537 * systemd-vconsole-setup gained support for system/service credentials:
538 vconsole.keymap/vconsole.keymap_toggle and
539 vconsole.font/vconsole.font_map/vconsole.font_unimap are analogous
540 the similarly-named options in vconsole.conf.
542 * systemd-localed will now save the XKB keyboard configuration to
543 /etc/vconsole.conf, and also read it from there with a higher
544 preference than the /etc/X11/xorg.conf.d/00-keyboard.conf config
545 file. Previously, this information was stored in the former file in
546 converted form, and only in latter file in the original form. Tools
547 which want to access keyboard configuration can now do so from a
550 * systemd-resolved gained support for configuring the nameservers and
551 search domains via kernel command line (nameserver=, domain=) and
552 credentials (network.dns, network.search_domains).
554 * systemd-resolved will now synthesize host names for the DNS stub
555 addresses it supports. Specifically when "_localdnsstub" is resolved,
556 127.0.0.53 is returned, and if "_localdnsproxy" is resolved
557 127.0.0.54 is returned.
559 * systemd-notify will now send a "RELOADING=1" notification when called
560 with --reloading, and "STOPPING=1" when called with --stopping. This
561 can be used to implement notifications from units where it's easier
562 to call a program than to use the sd-daemon library.
564 * systemd-analyze's 'plot' command can now output its information in
565 JSON, controlled via the --json= switch. Also, new --table, and
566 --no-legend options have been added.
568 * 'machinectl enable' will now automatically enable machines.target
569 unit in addition to adding the machine unit to the target.
571 Similarly, 'machinectl start|stop' gained a --now option to enable or
572 disable the machine unit when starting or stopping it.
574 * systemd-sysusers will now create /etc/ if it is missing.
576 * systemd-sleep 'HibernateDelaySec=' setting is changed back to
577 pre-v252's behaviour, and a new 'SuspendEstimationSec=' setting is
578 added to provide the new initial value for the new automated battery
579 estimation functionality. If 'HibernateDelaySec=' is set to any value,
580 the automated estimate (and thus the automated hibernation on low
581 battery to avoid data loss) functionality will be disabled.
583 * Default tmpfiles.d/ configuration will now automatically create
584 credentials storage directory '/etc/credstore/' with the appropriate,
585 secure permissions. If '/run/credstore/' exists, its permissions will
586 be fixed too in case they are not correct.
588 Changes in libsystemd and shared code:
590 * sd-bus gained new convenience functions sd_bus_emit_signal_to(),
591 sd_bus_emit_signal_tov(), and sd_bus_message_new_signal_to().
593 * sd-id128 functions now return -EUCLEAN (instead of -EIO) when the
594 128bit ID in files such as /etc/machine-id has an invalid
595 format. They also accept NULL as output parameter in more places,
596 which is useful when the caller only wants to validate the inputs and
597 does not need the output value.
599 * sd-login gained new functions sd_pidfd_get_session(),
600 sd_pidfd_get_owner_uid(), sd_pidfd_get_unit(),
601 sd_pidfd_get_user_unit(), sd_pidfd_get_slice(),
602 sd_pidfd_get_user_slice(), sd_pidfd_get_machine_name(), and
603 sd_pidfd_get_cgroup(), that are analogous to sd_pid_get_*(),
604 but accept a PIDFD instead of a PID.
606 * sd-path (and systemd-path) now export four new paths:
607 SD_PATH_SYSTEMD_SYSTEM_ENVIRONMENT_GENERATOR,
608 SD_PATH_SYSTEMD_USER_ENVIRONMENT_GENERATOR,
609 SD_PATH_SYSTEMD_SEARCH_SYSTEM_ENVIRONMENT_GENERATOR, and
610 SD_PATH_SYSTEMD_SEARCH_USER_ENVIRONMENT_GENERATOR,
612 * sd_notify() now supports AF_VSOCK as transport for notification
613 messages (in addition to the existing AF_UNIX support). This is
614 enabled if $NOTIFY_SOCKET is set in a "vsock:CID:port" format.
616 * Detection of chroot() environments now works if /proc/ is not
617 mounted. This affects systemd-detect-virt --chroot, but also means
618 that systemd tools will silently skip various operations in such an
621 * "Lockheed Martin Hardened Security for Intel Processors" (HS SRE)
622 virtualization is now detected.
624 Changes in the build system:
626 * Standalone variants of systemd-repart and systemd-shutdown may now be
627 built (if -Dstandalone=true).
629 * systemd-ac-power has been moved from /usr/lib/ to /usr/bin/, to, for
630 example, allow scripts to conditionalize execution on AC power
633 * The libp11kit library is now loaded through dlopen(3).
635 Changes in the documentation:
637 * Specifications that are not closely tied to systemd have moved to
638 https://uapi-group.org/specifications/: the Boot Loader Specification
639 and the Discoverable Partitions Specification.
641 Contributions from: 김인수, 13r0ck, Aidan Dang, Alberto Planas,
642 Alvin Šipraga, Andika Triwidada, AndyChi, angus-p, Anita Zhang,
643 Antonio Alvarez Feijoo, Arsen Arsenović, asavah, Benjamin Fogle,
644 Benjamin Tissoires, berenddeschouwer, BerndAdameit,
645 Bernd Steinhauser, blutch112, cake03, Callum Farmer, Carlo Teubner,
646 Charles Hardin, chris, Christian Brauner, Christian Göttsche,
647 Cristian Rodríguez, Daan De Meyer, Dan Streetman, DaPigGuy,
648 Darrell Kavanagh, David Tardon, dependabot[bot], Dirk Su,
649 Dmitry V. Levin, drosdeck, Edson Juliano Drosdeck, edupont,
650 Eric DeVolder, Erik Moqvist, Evgeny Vereshchagin, Fabian Gurtner,
651 Felix Riemann, Franck Bui, Frantisek Sumsal, Geert Lorang,
652 Gerd Hoffmann, Gio, Hannoskaj, Hans de Goede, Hugo Carvalho,
653 igo95862, Ilya Leoshkevich, Ivan Shapovalov, Jacek Migacz,
654 Jade Lovelace, Jan Engelhardt, Jan Janssen, Jan Macku, January,
655 Jason A. Donenfeld, jcg, Jean-Tiare Le Bigot, Jelle van der Waa,
656 Jeremy Linton, Jian Zhang, Jiayi Chen, Jia Zhang, Joerg Behrmann,
657 Jörg Thalheim, Joshua Goins, joshuazivkovic, Joshua Zivkovic,
658 Kai-Chuan Hsieh, Khem Raj, Koba Ko, Lennart Poettering, lichao,
659 Li kunyu, Luca Boccassi, Luca BRUNO, Ludwig Nussel,
660 Łukasz Stelmach, Lycowolf, marcel151, Marcus Schäfer, Marek Vasut,
661 Mark Laws, Michael Biebl, Michał Kotyla, Michal Koutný,
662 Michal Sekletár, Mike Gilbert, Mike Yuan, MkfsSion, ml,
663 msizanoen1, mvzlb, MVZ Ludwigsburg, Neil Moore, Nick Rosbrook,
664 noodlejetski, Pasha Vorobyev, Peter Cai, p-fpv, Phaedrus Leeds,
665 Philipp Jungkamp, Quentin Deslandes, Raul Tambre, Ray Strode,
666 reuben olinsky, Richard E. van der Luit, Richard Phibel,
667 Ricky Tigg, Robin Humble, rogg, Rudi Heitbaum, Sam James,
668 Samuel Cabrero, Samuel Thibault, Siddhesh Poyarekar, Simon Brand,
669 Space Meyer, Spindle Security, Steve Ramage, Takashi Sakamoto,
670 Thomas Haller, Tonći Galić, Topi Miettinen, Torsten Hilbrich,
671 Tuetuopay, uerdogan, Ulrich Ölmann, Valentin David,
672 Vitaly Kuznetsov, Vito Caputo, Waltibaba, Will Fancher,
673 William Roberts, wouter bolsterlee, Youfu Zhang, Yu Watanabe,
674 Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски,
681 Announcements of Future Feature Removals:
683 * We intend to remove cgroup v1 support from systemd release after the
684 end of 2023. If you run services that make explicit use of cgroup v1
685 features (i.e. the "legacy hierarchy" with separate hierarchies for
686 each controller), please implement compatibility with cgroup v2 (i.e.
687 the "unified hierarchy") sooner rather than later. Most of Linux
688 userspace has been ported over already.
690 * We intend to remove support for split-usr (/usr mounted separately
691 during boot) and unmerged-usr (parallel directories /bin and
692 /usr/bin, /lib and /usr/lib, etc). This will happen in the second
693 half of 2023, in the first release that falls into that time window.
694 For more details, see:
695 https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html
697 Compatibility Breaks:
699 * ConditionKernelVersion= checks that use the '=' or '!=' operators
700 will now do simple string comparisons (instead of version comparisons
701 à la stverscmp()). Version comparisons are still done for the
702 ordering operators '<', '>', '<=', '>='. Moreover, if no operator is
703 specified, a shell-style glob match is now done. This creates a minor
704 incompatibility compared to older systemd versions when the '*', '?',
705 '[', ']' characters are used, as these will now match as shell globs
706 instead of literally. Given that kernel version strings typically do
707 not include these characters we expect little breakage through this
710 * The service manager will now read the SELinux label used for SELinux
711 access checks from the unit file at the time it loads the file.
712 Previously, the label would be read at the moment of the access
713 check, which was problematic since at that time the unit file might
714 already have been updated or removed.
718 * systemd-measure is a new tool for calculating and signing expected
719 TPM2 PCR values for a given unified kernel image (UKI) booted via
720 sd-stub. The public key used for the signature and the signed
721 expected PCR information can be embedded inside the UKI. This
722 information can be extracted from the UKI by external tools and code
723 in the image itself and is made available to userspace in the booted
726 systemd-cryptsetup, systemd-cryptenroll, and systemd-creds have been
727 updated to make use of this information if available in the booted
728 kernel: when locking an encrypted volume/credential to the TPM
729 systemd-cryptenroll/systemd-creds will use the public key to bind the
730 volume/credential to any kernel that carries PCR information signed
731 by the same key pair. When unlocking such volumes/credentials
732 systemd-cryptsetup/systemd-creds will use the signature embedded in
733 the booted UKI to gain access.
735 Binding TPM-based disk encryption to public keys/signatures of PCR
736 values — instead of literal PCR values — addresses the inherent
737 "brittleness" of traditional PCR-bound TPM disk encryption schemes:
738 disks remain accessible even if the UKI is updated, without any TPM
739 specific preparation during the OS update — as long as each UKI
740 carries the necessary PCR signature information.
742 Net effect: if you boot a properly prepared kernel, TPM-bound disk
743 encryption now defaults to be locked to kernels which carry PCR
744 signatures from the same key pair. Example: if a hypothetical distro
745 FooOS prepares its UKIs like this, TPM-based disk encryption is now –
746 by default – bound to only FooOS kernels, and encrypted volumes bound
747 to the TPM cannot be unlocked on kernels from other sources. (But do
748 note this behaviour requires preparation/enabling in the UKI, and of
749 course users can always enroll non-TPM ways to unlock the volume.)
751 * systemd-pcrphase is a new tool that is invoked at six places during
752 system runtime, and measures additional words into TPM2 PCR 11, to
753 mark milestones of the boot process. This allows binding access to
754 specific TPM2-encrypted secrets to specific phases of the boot
755 process. (Example: LUKS2 disk encryption key only accessible in the
756 initrd, but not later.)
758 Changes in systemd itself, i.e. the manager and units
760 * The cpu controller is delegated to user manager units by default, and
761 CPUWeight= settings are applied to the top-level user slice units
762 (app.slice, background.slice, session.slice). This provides a degree
763 of resource isolation between different user services competing for
766 * Systemd can optionally do a full preset in the "first boot" condition
767 (instead of just enable-only). This behaviour is controlled by the
768 compile-time option -Dfirst-boot-full-preset. Right now it defaults
769 to 'false', but the plan is to switch it to 'true' for the subsequent
772 * Drop-ins are now allowed for transient units too.
774 * Systemd will set the taint flag 'support-ended' if it detects that
775 the OS image is past its end-of-support date. This date is declared
776 in a new /etc/os-release field SUPPORT_END= described below.
778 * Two new settings ConditionCredential= and AssertCredential= can be
779 used to skip or fail units if a certain system credential is not
782 * ConditionMemory= accepts size suffixes (K, M, G, T, …).
784 * DefaultSmackProcessLabel= can be used in system.conf and user.conf to
785 specify the SMACK security label to use when not specified in a unit
788 * DefaultDeviceTimeoutSec= can be used in system.conf and user.conf to
789 specify the default timeout when waiting for device units to
792 * C.UTF-8 is used as the default locale if nothing else has been
795 * [Condition|Assert]Firmware= have been extended to support certain
796 SMBIOS fields. For example
798 ConditionFirmware=smbios-field(board_name = "Custom Board")
800 conditionalizes the unit to run only when
801 /sys/class/dmi/id/board_name contains "Custom Board" (without the
804 * ConditionFirstBoot= now correctly evaluates as true only during the
805 boot phase of the first boot. A unit executed later, after booting
806 has completed, will no longer evaluate this condition as true.
808 * Socket units will now create sockets in the SELinuxContext= of the
809 associated service unit, if any.
811 * Boot phase transitions (start initrd → exit initrd → boot complete →
812 shutdown) will be measured into TPM2 PCR 11, so that secrets can be
813 bound to a specific runtime phase. E.g.: a LUKS encryption key can be
814 unsealed only in the initrd.
816 * Service credentials (i.e. SetCredential=/LoadCredential=/…) will now
817 also be provided to ExecStartPre= processes.
819 * Various units are now correctly ordered against
820 initrd-switch-root.target where previously a conflict without
821 ordering was configured. A stop job for those units would be queued,
822 but without the ordering it could be executed only after
823 initrd-switch-root.service, leading to units not being restarted in
824 the host system as expected.
826 * In order to fully support the IPMI watchdog driver, which has not yet
827 been ported to the new common watchdog device interface,
828 /dev/watchdog0 will be tried first and systemd will silently fallback
829 to /dev/watchdog if it is not found.
831 * New watchdog-related D-Bus properties are now published by systemd:
832 WatchdogDevice, WatchdogLastPingTimestamp,
833 WatchdogLastPingTimestampMonotonic.
835 * At shutdown, API virtual files systems (proc, sys, etc.) will be
838 * At shutdown, systemd will now log about processes blocking unmounting
841 * A new meson build option 'clock-valid-range-usec-max' was added to
842 allow disabling system time correction if RTC returns a timestamp far
845 * Propagated restart jobs will no longer be discarded while a unit is
848 * PID 1 will now import system credentials from SMBIOS Type 11 fields
849 ("OEM vendor strings"), in addition to qemu_fwcfg. This provides a
850 simple, fast and generic path for supplying credentials to a VM,
851 without involving external tools such as cloud-init/ignition.
853 * The CPUWeight= setting of unit files now accepts a new special value
854 "idle", which configures "idle" level scheduling for the unit.
856 * Service processes that are activated due to a .timer or .path unit
857 triggering will now receive information about this via environment
858 variables. Note that this is information is lossy, as activation
859 might be coalesced and only one of the activating triggers will be
860 reported. This is hence more suited for debugging or tracing rather
861 than for behaviour decisions.
863 * The riscv_flush_icache(2) system call has been added to the list of
864 system calls allowed by default when SystemCallFilter= is used.
866 * The selinux context derived from the target executable, instead of
867 'init_t' used for the manager itself, is now used when creating
868 listening sockets for units that specify SELinuxContextFromNet=yes.
870 Changes in sd-boot, bootctl, and the Boot Loader Specification:
872 * The Boot Loader Specification has been cleaned up and clarified.
873 Various corner cases in version string comparisons have been fixed
874 (e.g. comparisons for empty strings). Boot counting is now part of
875 the main specification.
877 * New PCRs measurements are performed during boot: PCR 11 for the
878 kernel+initrd combo, PCR 13 for any sysext images. If a measurement
879 took place this is now reported to userspace via the new
880 StubPcrKernelImage and StubPcrInitRDSysExts EFI variables.
882 * As before, systemd-stub will measure kernel parameters and system
883 credentials into PCR 12. It will now report this fact via the
884 StubPcrKernelParameters EFI variable to userspace.
886 * The UEFI monotonic boot counter is now included in the updated random
887 seed file maintained by sd-boot, providing some additional entropy.
889 * sd-stub will use LoadImage/StartImage to execute the kernel, instead
890 of arranging the image manually and jumping to the kernel entry
891 point. sd-stub also installs a temporary UEFI SecurityOverride to
892 allow the (unsigned) nested image to be booted. This is safe because
893 the outer (signed) stub+kernel binary must have been verified before
894 the stub was executed.
896 * Booting in EFI mixed mode (a 64-bit kernel over 32-bit UEFI firmware)
897 is now supported by sd-boot.
899 * bootctl gained a bunch of new options: --all-architectures to install
900 binaries for all supported EFI architectures, --root= and --image=
901 options to operate on a directory or disk image, and
902 --install-source= to specify the source for binaries to install,
903 --efi-boot-option-description= to control the name of the boot entry.
905 * The sd-boot stub exports a StubFeatures flag, which is used by
906 bootctl to show features supported by the stub that was used to boot.
908 * The PE section offsets that are used by tools that assemble unified
909 kernel images have historically been hard-coded. This may lead to
910 overlapping PE sections which may break on boot. The UKI will now try
911 to detect and warn about this.
913 Any tools that assemble UKIs must update to calculate these offsets
914 dynamically. Future sd-stub versions may use offsets that will not
915 work with the currently used set of hard-coded offsets!
917 * sd-stub now accepts (and passes to the initrd and then to the full
918 OS) new PE sections '.pcrsig' and '.pcrkey' that can be used to embed
919 signatures of expected PCR values, to allow sealing secrets via the
920 TPM2 against pre-calculated PCR measurements.
922 Changes in the hardware database:
924 * 'systemd-hwdb query' now supports the --root= option.
926 Changes in systemctl:
928 * systemctl now supports --state= and --type= options for the 'show'
931 * systemctl gained a new verb 'list-automounts' to list automount
934 * systemctl gained support for a new --image= switch to be able to
935 operate on the specified disk image (similar to the existing --root=
936 which operates relative to some directory).
938 Changes in systemd-networkd:
940 * networkd can set Linux NetLabel labels for integration with the
941 network control in security modules via a new NetLabel= option.
943 * The RapidCommit= is (re-)introduced to enable faster configuration
944 via DHCPv6 (RFC 3315).
946 * networkd gained a new option TCPCongestionControlAlgorithm= that
947 allows setting a per-route TCP algorithm.
949 * networkd gained a new option KeepFileDescriptor= to allow keeping a
950 reference (file descriptor) open on TUN/TAP interfaces, which is
951 useful to avoid link flaps while the underlying service providing the
952 interface is being serviced.
954 * RouteTable= now also accepts route table names.
956 Changes in systemd-nspawn:
958 * The --bind= and --overlay= options now support relative paths.
960 * The --bind= option now supports a 'rootidmap' value, which will
961 use id-mapped mounts to map the root user inside the container to the
962 owner of the mounted directory on the host.
964 Changes in systemd-resolved:
966 * systemd-resolved now persists DNSOverTLS in its state file too. This
967 fixes a problem when used in combination with NetworkManager, which
968 sends the setting only once, causing it to be lost if resolved was
969 restarted at any point.
971 * systemd-resolved now exposes a varlink socket at
972 /run/systemd/resolve/io.systemd.Resolve.Monitor, accessible only for
973 root. Processed DNS requests in a JSON format will be published to
974 any clients connected to this socket.
976 resolvectl gained a 'monitor' verb to make use of this.
978 * systemd-resolved now treats unsupported DNSSEC algorithms as INSECURE
979 instead of returning SERVFAIL, as per RFC:
980 https://datatracker.ietf.org/doc/html/rfc6840#section-5.2
982 * OpenSSL is the default crypto backend for systemd-resolved. (gnutls
985 Changes in libsystemd and other libraries:
987 * libsystemd now exports sd_bus_error_setfv() (a convenience function
988 for setting bus errors), sd_id128_string_equal (a convenience
989 function for 128bit ID string comparisons), and
990 sd_bus_message_read_strv_extend() (a function to incrementally read
993 * libsystemd now exports sd_device_get_child_first()/_next() as a
994 high-level interface for enumerating child devices. It also supports
995 sd_device_new_child() for opening a child device given a device
998 * libsystemd now exports sd_device_monitor_set()/get_description()
999 which allow setting a custom description that will be used in log
1000 messages by sd_device_monitor*.
1002 * Private shared libraries (libsystemd-shared-nnn.so,
1003 libsystemd-core-nnn.so) are now installed into arch-specific
1004 directories to allow multi-arch installs.
1006 * A new sd-gpt.h header is now published, listing GUIDs from the
1007 Discoverable Partitions specification. For more details see:
1008 https://systemd.io/DISCOVERABLE_PARTITIONS/
1010 * A new function sd_hwdb_new_from_path() has been added to open a hwdb
1011 database given an explicit path to the file.
1013 * The signal number argument to sd_event_add_signal() now can now be
1014 ORed with the SD_EVENT_SIGNAL_PROCMASK flag, causing sigprocmask() to
1015 be automatically invoked to block the specified signal. This is
1016 useful to simplify invocations as the caller doesn't have to do this
1019 * A new convenience call sd_event_set_signal_exit() has been added to
1020 sd-event to set up signal handling so that the event loop
1021 automatically terminates cleanly on SIGTERM/SIGINT.
1023 Changes in other components:
1025 * systemd-sysusers, systemd-tmpfiles, and systemd-sysctl configuration
1026 can now be provided via the credential mechanism.
1028 * systemd-analyze gained a new verb 'compare-versions' that implements
1029 comparisons for versions strings (similarly to 'rpmdev-vercmp' and
1030 'dpkg --compare-versions').
1032 * 'systemd-analyze dump' is extended to accept glob patterns for unit
1033 names to limit the output to matching units.
1035 * tmpfiles.d/ lines can read file contents to write from a credential.
1036 The new modifier char '^' is used to specify that the argument is a
1037 credential name. This mechanism is used to automatically populate
1038 /etc/motd, /etc/issue, and /etc/hosts from credentials.
1040 * tmpfiles.d/ may now be configured to avoid changing uid/gid/mode of
1041 an inode if the specification is prefixed with ':' and the inode
1044 * Default tmpfiles.d/ configuration now carries a line to automatically
1045 use an 'ssh.authorized_keys.root' credential if provided to set up
1046 the SSH authorized_keys file for the root user.
1048 * systemd-tmpfiles will now gracefully handle absent source of "C" copy
1051 * tmpfiles.d/ F/w lines now optionally permit encoding of the payload
1052 in base64. This is useful to write arbitrary binary data into files.
1054 * The pkgconfig and rpm macros files now export the directory for user
1055 units as 'user_tmpfiles_dir' and '%_user_tmpfilesdir'.
1057 * Detection of Apple Virtualization and detection of Parallels and
1058 KubeVirt virtualization on non-x86 archs have been added.
1060 * os-release gained a new field SUPPORT_END=YYYY-MM-DD to inform the
1061 user when their system will become unsupported.
1063 * When performing suspend-then-hibernate, the system will estimate the
1064 discharge rate and use that to set the delay until hibernation and
1065 hibernate immediately instead of suspending when running from a
1066 battery and the capacity is below 5%.
1068 * systemd-sysctl gained a --strict option to fail when a sysctl
1069 setting is unknown to the kernel.
1071 * machinectl supports --force for the 'copy-to' and 'copy-from'
1074 * coredumpctl gained the --root and --image options to look for journal
1075 files under the specified root directory, image, or block device.
1077 * 'journalctl -o' and similar commands now implement a new output mode
1078 "short-delta". It is similar to "short-monotonic", but also shows the
1079 time delta between subsequent messages.
1081 * journalctl now respects the --quiet flag when verifying consistency
1084 * Journal log messages gained a new implicit field _RUNTIME_SCOPE= that
1085 will indicate whether a message was logged in the 'initrd' phase or
1086 in the 'system' phase of the boot process.
1088 * Journal files gained a new compatibility flag
1089 'HEADER_INCOMPATIBLE_COMPACT'. Files with this flag implement changes
1090 to the storage format that allow reducing size on disk. As with other
1091 compatibility flags, older journalctl versions will not be able to
1092 read journal files using this new format. The environment variable
1093 'SYSTEMD_JOURNAL_COMPACT=0' can be passed to systemd-journald to
1094 disable this functionality. It is enabled by default.
1096 * systemd-run's --working-directory= switch now works when used in
1097 combination with --scope.
1099 * portablectl gained a --force flag to skip certain sanity checks. This
1100 is implemented using new flags accepted by systemd-portabled for the
1101 *WithExtensions() D-Bus methods: SD_SYSTEMD_PORTABLE_FORCE_ATTACH
1102 flag now means that the attach/detach checks whether the units are
1103 already present and running will be skipped. Similarly,
1104 SD_SYSTEMD_PORTABLE_FORCE_SYSEXT flag means that the check whether
1105 image name matches the name declared inside of the image will be
1106 skipped. Callers must be sure to do those checks themselves if
1109 * systemd-portabled will now use the original filename to check
1110 extension-release.NAME for correctness, in case it is passed a
1113 * systemd-portabled now uses PrivateTmp=yes in the 'trusted' profile
1116 * sysext's extension-release files now support '_any' as a special
1117 value for the ID= field, to allow distribution-independent extensions
1118 (e.g.: fully statically compiled binaries, scripts). It also gained
1119 support for a new ARCHITECTURE= field that may be used to explicitly
1120 restrict an image to hosts of a specific architecture.
1122 * systemd-repart now supports creating squashfs partitions. This
1123 requires mksquashfs from squashfs-tools.
1125 * systemd-repart gained a --split flag to also generate split
1126 artifacts, i.e. a separate file for each partition. This is useful in
1127 conjunction with systemd-sysupdate or other tools, or to generate
1128 split dm-verity artifacts.
1130 * systemd-repart is now able to generate dm-verity partitions, including
1133 * systemd-repart can now set a partition UUID to zero, allowing it to
1134 be filled in later, such as when using verity partitions.
1136 * systemd-repart now supports drop-ins for its configuration files.
1138 * Package metadata logged by systemd-coredump in the system journal is
1141 * xdg-autostart-service now expands 'tilde' characters in Exec lines.
1143 * systemd-oomd now automatically links against libatomic, if available.
1145 * systemd-oomd now sends out a 'Killed' D-Bus signal when a cgroup is
1148 * scope units now also provide oom-kill status.
1150 * systemd-pstore will now try to load only the efi_pstore kernel module
1151 before running, ensuring that pstore can be used.
1153 * systemd-logind gained a new StopIdleSessionSec= option to stop an idle
1154 session after a preconfigure timeout.
1156 * systemd-homed will now wait up to 30 seconds for workers to terminate,
1157 rather than indefinitely.
1159 * homectl gained a new '--luks-sector-size=' flag that allows users to
1160 select the preferred LUKS sector size. Must be a power of 2 between 512
1161 and 4096. systemd-userdbd records gained a corresponding field.
1163 * systemd-sysusers will now respect the 'SOURCE_DATE_EPOCH' environment
1164 variable when generating the 'sp_lstchg' field, to ensure an image
1165 build can be reproducible.
1167 * 'udevadm wait' will now listen to kernel uevents too when called with
1170 * When naming network devices udev will now consult the Devicetree
1171 "alias" fields for the device.
1173 * systemd-udev will now create infiniband/by-path and
1174 infiniband/by-ibdev links for Infiniband verbs devices.
1176 * systemd-udev-trigger.service will now also prioritize input devices.
1178 * ConditionACPower= and systemd-ac-power will now assume the system is
1179 running on AC power if no battery can be found.
1181 * All features and tools using the TPM2 will now communicate with it
1182 using a bind key. Beforehand, the tpm2 support used encrypted sessions
1183 by creating a primary key that was used to encrypt traffic. This
1184 creates a problem as the key created for encrypting the traffic could
1185 be faked by an active interposer on the bus. In cases when a pin is
1186 used, a bind key will be used. The pin is used as the auth value for
1187 the seal key, aka the disk encryption key, and that auth value will be
1188 used in the session establishment. An attacker would need the pin
1189 value to create the secure session and thus an active interposer
1190 without the pin cannot interpose on TPM2 traffic.
1192 * systemd-growfs no longer requires udev to run.
1194 * systemd-backlight now will better support systems with multiple
1197 * systemd-cryptsetup's keyfile-timeout= option now also works when a
1198 device is used as a keyfile.
1200 * systemd-cryptenroll gained a new --unlock-key-file= option to get the
1201 unlocking key from a key file (instead of prompting the user). Note
1202 that this is the key for unlocking the volume in order to be able to
1203 enroll a new key, but it is not the key that is enrolled.
1205 * systemd-dissect gained a new --umount switch that will safely and
1206 synchronously unmount all partitions of an image previously mounted
1207 with 'systemd-dissect --mount'.
1209 * When using gcrypt, all systemd tools and services will now configure
1210 it to prefer the OS random number generator if present.
1212 * All example code shipped with documentation has been relicensed from CC0
1215 * Unit tests will no longer fail when running on a system without
1218 Experimental features:
1220 * BPF programs can now be compiled with bpf-gcc (requires libbpf >= 1.0
1221 and bpftool >= 7.0).
1223 * sd-boot can automatically enroll SecureBoot keys from files found on
1224 the ESP. This enrollment can be either automatic ('force' mode) or
1225 controlled by the user ('manual' mode). It is sufficient to place the
1226 SecureBoot keys in the right place in the ESP and they will be picked
1227 up by sd-boot and shown in the boot menu.
1229 * The mkosi config in systemd gained support for automatically
1230 compiling a kernel with the configuration appropriate for testing
1231 systemd. This may be useful when developing or testing systemd in
1232 tandem with the kernel.
1234 Contributions from: 김인수, Adam Williamson, adrian5, Aidan Dang,
1235 Akihiko Odaki, Alban Bedel, Albert Mikaelyan, Aleksey Vasenev,
1236 Alexander Graf, Alexander Shopov, Alexander Wilson,
1237 Alper Nebi Yasak, anarcat, Anders Jonsson, Andre Kalb,
1238 Andrew Stone, Andrey Albershteyn, Anita Zhang, Ansgar Burchardt,
1239 Antonio Alvarez Feijoo, Arnaud Ferraris, Aryan singh, asavah,
1240 Avamander, Avram Lubkin, Balázs Meskó, Bastien Nocera,
1241 Benjamin Franzke, BerndAdameit, bin456789, Celeste Liu,
1242 Chih-Hsuan Yen, Christian Brauner, Christian Göttsche,
1243 Christian Hesse, Clyde Byrd III, codefiles, Colin Walters,
1244 Cristian Rodríguez, Daan De Meyer, Daniel Braunwarth,
1245 Daniel Rusek, Dan Streetman, Darsey Litzenberger, David Edmundson,
1246 David Jaša, David Rheinsberg, David Seifert, David Tardon,
1247 dependabot[bot], Devendra Tewari, Dominique Martinet, drosdeck,
1248 Edson Juliano Drosdeck, Eduard Tolosa, eggfly, Einsler Lee,
1249 Elias Probst, Eli Schwartz, Evgeny Vereshchagin, exploide, Fei Li,
1250 Foster Snowhill, Franck Bui, Frank Dana, Frantisek Sumsal,
1251 Gerd Hoffmann, Gio, Goffredo Baroncelli, gtwang01,
1252 Guillaume W. Bres, H A, Hans de Goede, Heinrich Schuchardt,
1253 Hugo Carvalho, i-do-cpp, igo95862, j00512545, Jacek Migacz,
1254 Jade Bilkey, James Hilliard, Jan B, Janis Goldschmidt,
1255 Jan Janssen, Jan Kuparinen, Jan Luebbe, Jan Macku,
1256 Jason A. Donenfeld, Javkhlanbayar Khongorzul, Jeremy Soller,
1257 JeroenHD, jiangchuangang, João Loureiro,
1258 Joaquín Ignacio Aramendía, Jochen Sprickerhof,
1259 Johannes Schauer Marin Rodrigues, Jonas Kümmerlin,
1260 Jonas Witschel, Jonathan Kang, Jonathan Lebon, Joost Heitbrink,
1261 Jörg Thalheim, josh-gordon-fb, Joyce, Kai Lueke, lastkrick,
1262 Lennart Poettering, Leon M. George, licunlong, Li kunyu,
1263 LockBlock-dev, Loïc Collignon, Lubomir Rintel, Luca Boccassi,
1264 Luca BRUNO, Ludwig Nussel, Łukasz Stelmach, Maccraft123,
1265 Marc Kleine-Budde, Marius Vollmer, Martin Wilck, matoro,
1266 Matthias Lisin, Max Gautier, Maxim Mikityanskiy, Michael Biebl,
1267 Michal Koutný, Michal Sekletár, Michal Stanke, Mike Gilbert,
1268 Mitchell Freiderich, msizanoen1, Nick Rosbrook, nl6720, Oğuz Ersen,
1269 Oleg Solovyov, Olga Smirnova, Pablo Ceballos, Pavel Zhukov,
1270 Phaedrus Leeds, Philipp Gortan, Piotr Drąg, Pyfisch,
1271 Quentin Deslandes, Rahil Bhimjiani, Rene Hollander, Richard Huang,
1272 Richard Phibel, Rudi Heitbaum, Sam James, Sarah Brofeldt,
1273 Sean Anderson, Sebastian Scheibner, Shreenidhi Shedi,
1274 Sonali Srivastava, Steve Ramage, Suraj Krishnan, Swapnil Devesh,
1275 Takashi Sakamoto, Ted X. Toth, Temuri Doghonadze, Thomas Blume,
1276 Thomas Haller, Thomas Hebb, Tomáš Hnyk, Tomasz Paweł Gajc,
1277 Topi Miettinen, Ulrich Ölmann, undef, Uriel Corfa,
1278 Victor Westerhuis, Vincent Dagonneau, Vishal Chillara Srinivas,
1279 Vito Caputo, Weblate, Wenchao Hao, William Roberts, williamsumendap,
1280 wineway, xiaoyang, Yuri Chornoivan, Yu Watanabe,
1281 Zbigniew Jędrzejewski-Szmek, Zhaofeng Li, наб
1283 – The Great Beyond, 2022-10-31 👻
1287 Backwards-incompatible changes:
1289 * The minimum kernel version required has been bumped from 3.13 to 4.15,
1290 and CLOCK_BOOTTIME is now assumed to always exist.
1292 * C11 with GNU extensions (aka "gnu11") is now used to build our
1293 components. Public API headers are still restricted to ISO C89.
1295 * In v250, a systemd-networkd feature that automatically configures
1296 routes to addresses specified in AllowedIPs= was added and enabled by
1297 default. However, this causes network connectivity issues in many
1298 existing setups. Hence, it has been disabled by default since
1299 systemd-stable 250.3. The feature can still be used by explicitly
1300 configuring RouteTable= setting in .netdev files.
1302 * Jobs started via StartUnitWithFlags() will no longer return 'skipped'
1303 when a Condition*= check does not succeed, restoring the JobRemoved
1304 signal to the behaviour it had before v250.
1306 * The org.freedesktop.portable1 methods GetMetadataWithExtensions() and
1307 GetImageMetadataWithExtensions() have been fixed to provide an extra
1308 return parameter, containing the actual extension release metadata.
1309 The current implementation was judged to be broken and unusable, and
1310 thus the usual procedure of adding a new set of methods was skipped,
1311 and backward compatibility broken instead on the assumption that
1312 nobody can be affected given the current state of this interface.
1314 * All kernels supported by systemd mix bytes returned by RDRAND (or
1315 similar) into the entropy pool at early boot. This means that on
1316 those systems, even if /dev/urandom is not yet initialized, it still
1317 returns bytes that are of at least RDRAND quality. For that reason,
1318 we no longer have reason to invoke RDRAND from systemd itself, which
1319 has historically been a source of bugs. Furthermore, kernels ≥5.6
1320 provide the getrandom(GRND_INSECURE) interface for returning random
1321 bytes before the entropy pool is initialized without warning into
1322 kmsg, which is what we attempt to use if available. systemd's direct
1323 usage of RDRAND has been removed. x86 systems ≥Broadwell that are
1324 running an older kernel may experience kmsg warnings that were not
1325 seen with 250. For newer kernels, non-x86 systems, or older x86
1326 systems, there should be no visible changes.
1328 * sd-boot will now measure the kernel command line into TPM PCR 12
1329 rather than PCR 8. This improves usefulness of the measurements on
1330 systems where sd-boot is chainloaded from Grub. Grub measures all
1331 commands its executes into PCR 8, which makes it very hard to use
1332 reasonably, hence separate ourselves from that and use PCR 12
1333 instead, which is what certain Ubuntu editions already do. To retain
1334 compatibility with systems running older systemd systems a new meson
1335 option 'efi-tpm-pcr-compat' has been added (which defaults to false).
1336 If enabled, the measurement is done twice: into the new-style PCR 12
1337 *and* the old-style PCR 8. It's strongly advised to migrate all users
1338 to PCR 12 for this purpose in the long run, as we intend to remove
1339 this compatibility feature in two years' time.
1341 * busctl capture now writes output in the newer pcapng format instead
1344 * A udev rule that imported hwdb matches for USB devices with lowercase
1345 hexadecimal vendor/product ID digits was added in systemd 250. This
1346 has been reverted, since uppercase hexadecimal digits are supposed to
1347 be used, and we already had a rule with the appropriate match.
1349 Users might need to adjust their local hwdb entries.
1351 * arch_prctl(2) has been moved to the @default set in the syscall filters
1352 (as exposed via the SystemCallFilter= setting in service unit files).
1353 It is apparently used by the linker now.
1355 * The tmpfiles entries that create the /run/systemd/netif directory and
1356 its subdirectories were moved from tmpfiles.d/systemd.conf to
1357 tmpfiles.d/systemd-network.conf.
1359 Users might need to adjust their files that override tmpfiles.d/systemd.conf
1360 to account for this change.
1362 * The requirement for Portable Services images to contain a well-formed
1363 os-release file (i.e.: contain at least an ID field) is now enforced.
1364 This applies to base images and extensions, and also to systemd-sysext.
1366 Changes in the Boot Loader Specification, kernel-install and sd-boot:
1368 * kernel-install's and bootctl's Boot Loader Specification Type #1
1369 entry generation logic has been reworked. The user may now pick
1370 explicitly by which "token" string to name the installation's boot
1371 entries, via the new /etc/kernel/entry-token file or the new
1372 --entry-token= switch to bootctl. By default — as before — the
1373 entries are named after the local machine ID. However, in "golden
1374 image" environments, where the machine ID shall be initialized on
1375 first boot (as opposed to at installation time before first boot) the
1376 machine ID will not be available at build time. In this case the
1377 --entry-token= switch to bootctl (or the /etc/kernel/entry-token
1378 file) may be used to override the "token" for the entries, for
1379 example the IMAGE_ID= or ID= fields from /etc/os-release. This will
1380 make the OS images independent of any machine ID, and ensure that the
1381 images will not carry any identifiable information before first boot,
1382 but on the other hand means that multiple parallel installations of
1383 the very same image on the same disk cannot be supported.
1385 Summary: if you are building golden images that shall acquire
1386 identity information exclusively on first boot, make sure to both
1387 remove /etc/machine-id *and* to write /etc/kernel/entry-token to the
1388 value of the IMAGE_ID= or ID= field of /etc/os-release or another
1389 suitable identifier before deploying the image.
1391 * The Boot Loader Specification has been extended with
1392 /loader/entries.srel file located in the EFI System Partition (ESP)
1393 that disambiguates the format of the entries in the /loader/entries/
1394 directory (in order to discern them from incompatible uses of this
1395 directory by other projects). For entries that follow the
1396 Specification, the string "type1" is stored in this file.
1398 bootctl will now write this file automatically when installing the
1399 systemd-boot boot loader.
1401 * kernel-install supports a new initrd_generator= setting in
1402 /etc/kernel/install.conf, that is exported as
1403 $KERNEL_INSTALL_INITRD_GENERATOR to kernel-install plugins. This
1404 allows choosing different initrd generators.
1406 * kernel-install will now create a "staging area" (an initially-empty
1407 directory to gather files for a Boot Loader Specification Type #1
1408 entry). The path to this directory is exported as
1409 $KERNEL_INSTALL_STAGING_AREA to kernel-install plugins, which should
1410 drop files there instead of writing them directly to the final
1411 location. kernel-install will move them when all files have been
1412 prepared successfully.
1414 * New option sort-key= has been added to the Boot Loader Specification
1415 to override the sorting order of the entries in the boot menu. It is
1416 read by sd-boot and bootctl, and will be written by kernel-install,
1417 with the default value of IMAGE_ID= or ID= fields from
1418 os-release. Together, this means that on multiboot installations,
1419 entries should be grouped and sorted in a predictable way.
1421 * The sort order of boot entries has been updated: entries which have
1422 the new field sort-key= are sorted by it first, and all entries
1423 without it are ordered later. After that, entries are sorted by
1424 version so that newest entries are towards the beginning of the list.
1426 * The kernel-install tool gained a new 'inspect' verb which shows the
1427 paths and other settings used.
1429 * sd-boot can now optionally beep when the menu is shown and menu
1430 entries are selected, which can be useful on machines without a
1431 working display. (Controllable via a loader.conf setting.)
1433 * The --make-machine-id-directory= switch to bootctl has been replaced
1434 by --make-entry-directory=, given that the entry directory is not
1435 necessarily named after the machine ID, but after some other suitable
1436 ID as selected via --entry-token= described above. The old name of
1437 the option is still understood to maximize compatibility.
1439 * 'bootctl list' gained support for a new --json= switch to output boot
1440 menu entries in JSON format.
1442 * 'bootctl is-installed' now supports the --graceful, and various verbs
1443 omit output with the new option --quiet.
1445 Changes in systemd-homed:
1447 * Starting with v250 systemd-homed uses UID/GID mapping on the mounts
1448 of activated home directories it manages (if the kernel and selected
1449 file systems support it). So far it mapped three UID ranges: the
1450 range from 0…60000, the user's own UID, and the range 60514…65534,
1451 leaving everything else unmapped (in other words, the 16bit UID range
1452 is mapped almost fully, with the exception of the UID subrange used
1453 for systemd-homed users, with one exception: the user's own UID).
1454 Unmapped UIDs may not be used for file ownership in the home
1455 directory — any chown() attempts with them will fail. With this
1456 release a fourth range is added to these mappings:
1457 524288…1879048191. This range is the UID range intended for container
1460 https://systemd.io/UIDS-GIDS
1462 This range may be used for container managers that place container OS
1463 trees in the home directory (which is a questionable approach, for
1464 quota, permission, SUID handling and network file system
1465 compatibility reasons, but nonetheless apparently commonplace). Note
1466 that this mapping is mapped 1:1 in a pass-through fashion, i.e. the
1467 UID assignments from the range are not managed or mapped by
1468 `systemd-homed`, and must be managed with other mechanisms, in the
1469 context of the local system.
1471 Typically, a better approach to user namespacing in relevant
1472 container managers would be to leave container OS trees on disk at
1473 UID offset 0, but then map them to a dynamically allocated runtime
1474 UID range via another UID mount map at container invocation
1475 time. That way user namespace UID ranges become strictly a runtime
1476 concept, and do not leak into persistent file systems, persistent
1477 user databases or persistent configuration, thus greatly simplifying
1478 handling, and improving compatibility with home directories intended
1479 to be portable like the ones managed by systemd-homed.
1481 Changes in shared libraries:
1483 * A new libsystemd-core-<version>.so private shared library is
1484 installed under /usr/lib/systemd/system, mirroring the existing
1485 libsystemd-shared-<version>.so library. This allows the total
1486 installation size to be reduced by binary code reuse.
1488 * The <version> tag used in the name of libsystemd-shared.so and
1489 libsystemd-core.so can be configured via the meson option
1490 'shared-lib-tag'. Distributions may build subsequent versions of the
1491 systemd package with unique tags (e.g. the full package version),
1492 thus allowing multiple installations of those shared libraries to be
1493 available at the same time. This is intended to fix an issue where
1494 programs that link to those libraries would fail to execute because
1495 they were installed earlier or later than the appropriate version of
1498 * The sd-id128 API gained a new call sd_id128_to_uuid_string() that is
1499 similar to sd_id128_to_string() but formats the ID in RFC 4122 UUID
1500 format instead of as a simple series of hex characters.
1502 * The sd-device API gained two new calls sd_device_new_from_devname()
1503 and sd_device_new_from_path() which permit allocating an sd_device
1504 object from a device node name or file system path.
1506 * sd-device also gained a new call sd_device_open() which will open the
1507 device node associated with a device for which an sd_device object
1508 has been allocated. The call is supposed to address races around
1509 device nodes being removed/recycled due to hotplug events, or media
1510 change events: the call checks internally whether the major/minor of
1511 the device node and the "diskseq" (in case of block devices) match
1512 with the metadata loaded in the sd_device object, thus ensuring that
1513 the device once opened really matches the provided sd_device object.
1515 Changes in PID1, systemctl, and systemd-oomd:
1517 * A new set of service monitor environment variables will be passed to
1518 OnFailure=/OnSuccess= handlers, but only if exactly one unit lists the
1519 handler unit as OnFailure=/OnSuccess=. The variables are:
1520 $MONITOR_SERVICE_RESULT, $MONITOR_EXIT_CODE, $MONITOR_EXIT_STATUS,
1521 $MONITOR_INVOCATION_ID and $MONITOR_UNIT. For cases when a single
1522 handler needs to watch multiple units, use a templated handler.
1524 * A new ExtensionDirectories= setting in service unit files allows
1525 system extensions to be loaded from a directory. (It is similar to
1526 ExtensionImages=, but takes paths to directories, instead of
1529 'portablectl attach --extension=' now also accepts directory paths.
1531 * The user.delegate and user.invocation_id extended attributes on
1532 cgroups are used in addition to trusted.delegate and
1533 trusted.invocation_id. The latter pair requires privileges to set,
1534 but the former doesn't and can be also set by the unprivileged user
1537 (Only supported on kernels ≥5.6.)
1539 * Units that were killed by systemd-oomd will now have a service result
1540 of 'oom-kill'. The number of times a service was killed is tallied
1541 in the 'user.oomd_ooms' extended attribute.
1543 The OOMPolicy= unit file setting is now also honoured by
1546 * In unit files the new %y/%Y specifiers can be used to refer to
1547 normalized unit file path, which is particularly useful for symlinked
1550 The new %q specifier resolves to the pretty hostname
1551 (i.e. PRETTY_HOSTNAME= from /etc/machine-info).
1553 The new %d specifier resolves to the credentials directory of a
1554 service (same as $CREDENTIALS_DIRECTORY).
1556 * The RootDirectory=, MountAPIVFS=, ExtensionDirectories=,
1557 *Capabilities*=, ProtectHome=, *Directory=, TemporaryFileSystem=,
1558 PrivateTmp=, PrivateDevices=, PrivateNetwork=, NetworkNamespacePath=,
1559 PrivateIPC=, IPCNamespacePath=, PrivateUsers=, ProtectClock=,
1560 ProtectKernelTunables=, ProtectKernelModules=, ProtectKernelLogs=,
1561 MountFlags= service settings now also work in unprivileged user
1562 services, i.e. those run by the user's --user service manager, as long
1563 as user namespaces are enabled on the system.
1565 * Services with Restart=always and a failing ExecCondition= will no
1566 longer be restarted, to bring ExecCondition= behaviour in line with
1567 Condition*= settings.
1569 * LoadCredential= now accepts a directory as the argument; all files
1570 from the directory will be loaded as credentials.
1572 * A new D-Bus property ControlGroupId is now exposed on service units,
1573 that encapsulates the service's numeric cgroup ID that newer kernels
1574 assign to each cgroup.
1576 * PID 1 gained support for configuring the "pre-timeout" of watchdog
1577 devices and the associated governor, via the new
1578 RuntimeWatchdogPreSec= and RuntimeWatchdogPreGovernor= configuration
1579 options in /etc/systemd/system.conf.
1581 * systemctl's --timestamp= option gained a new choice "unix", to show
1582 timestamp as unix times, i.e. seconds since 1970, Jan 1st.
1584 * A new "taint" flag named "old-kernel" is introduced which is set when
1585 the kernel systemd runs on is older then the current baseline version
1586 (see above). The flag is shown in "systemctl status" output.
1588 * Two additional taint flags "short-uid-range" and "short-gid-range"
1589 have been added as well, which are set when systemd notices it is run
1590 within a userns namespace that does not define the full 0…65535 UID
1593 * A new "unmerged-usr" taint flag has been added that is set whenever
1594 running on systems where /bin/ + /sbin/ are *not* symlinks to their
1595 counterparts in /usr/, i.e. on systems where the /usr/-merge has not
1598 * Generators invoked by PID 1 will now have a couple of useful
1599 environment variables set describing the execution context a
1600 bit. $SYSTEMD_SCOPE encodes whether the generator is called from the
1601 system service manager, or from the per-user service
1602 manager. $SYSTEMD_IN_INITRD encodes whether the generator is invoked
1603 in initrd context or on the host. $SYSTEMD_FIRST_BOOT encodes whether
1604 systemd considers the current boot to be a "first"
1605 boot. $SYSTEMD_VIRTUALIZATION encode whether virtualization is
1606 detected and which type of hypervisor/container
1607 manager. $SYSTEMD_ARCHITECTURE indicates which architecture the
1608 kernel is built for.
1610 * PID 1 will now automatically pick up system credentials from qemu's
1611 fw_cfg interface, thus allowing passing arbitrary data into VM
1612 systems similar to how this is already supported for passing them
1613 into `systemd-nspawn` containers. Credentials may now also be passed
1614 in via the new kernel command line option `systemd.set_credential=`
1615 (note that kernel command line options are world-readable during
1616 runtime, and only useful for credentials that require no
1617 confidentiality). The credentials that can be passed to unified
1618 kernels that use the `systemd-stub` UEFI stub are now similarly
1619 picked up automatically. Automatic importing of system credentials
1620 this way can be turned off via the new
1621 `systemd.import_credentials=no` kernel command line option.
1623 * LoadCredential= will now automatically look for credentials in the
1624 /etc/credstore/, /run/credstore/, /usr/lib/credstore/ directories if
1625 the argument is not an absolute path. Similarly,
1626 LoadCredentialEncrypted= will check the same directories plus
1627 /etc/credstore.encrypted/, /run/credstore.encrypted/ and
1628 /usr/lib/credstore.encrypted/. The idea is to use those directories
1629 as the system-wide location for credentials that services should pick
1632 * System and service credentials are described in great detail in a new
1635 https://systemd.io/CREDENTIALS
1637 Changes in systemd-journald:
1639 * The journal JSON export format has been added to listed of stable
1640 interfaces (https://systemd.io/PORTABILITY_AND_STABILITY/).
1642 * journalctl --list-boots now supports JSON output and the --reverse option.
1644 * Under docs/: JOURNAL_EXPORT_FORMATS was imported from the wiki and
1645 updated, BUILDING_IMAGES is new:
1647 https://systemd.io/JOURNAL_EXPORT_FORMATS
1648 https://systemd.io/BUILDING_IMAGES
1652 * Two new hwdb files have been added. One lists "handhelds" (PDAs,
1653 calculators, etc.), the other AV production devices (DJ tables,
1654 keypads, etc.) that should accessible to the seat owner user by
1657 * udevadm trigger gained a new --prioritized-subsystem= option to
1658 process certain subsystems (and all their parent devices) earlier.
1660 systemd-udev-trigger.service now uses this new option to trigger
1661 block and TPM devices first, hopefully making the boot a bit faster.
1663 * udevadm trigger now implements --type=all, --initialized-match,
1664 --initialized-nomatch to trigger both subsystems and devices, only
1665 already-initialized devices, and only devices which haven't been
1666 initialized yet, respectively.
1668 * udevadm gained a new "wait" command for safely waiting for a specific
1669 device to show up in the udev device database. This is useful in
1670 scripts that asynchronously allocate a block device (e.g. through
1671 repartitioning, or allocating a loopback device or similar) and need
1672 to synchronize on the creation to complete.
1674 * udevadm gained a new "lock" command for locking one or more block
1675 devices while formatting it or writing a partition table to it. It is
1676 an implementation of https://systemd.io/BLOCK_DEVICE_LOCKING and
1677 usable in scripts dealing with block devices.
1679 * udevadm info will show a couple of additional device fields in its
1680 output, and will not apply a limited set of coloring to line types.
1682 * udevadm info --tree will now show a tree of objects (i.e. devices and
1683 suchlike) in the /sys/ hierarchy.
1685 * Block devices will now get a new set of device symlinks in
1686 /dev/disk/by-diskseq/<nr>, which may be used to reference block
1687 device nodes via the kernel's "diskseq" value. Note that this does
1688 not guarantee that opening a device by a symlink like this will
1689 guarantee that the opened device actually matches the specified
1690 diskseq value. To be safe against races, the actual diskseq value of
1691 the opened device (BLKGETDISKSEQ ioctl()) must still be compred with
1692 the one in the symlink path.
1694 * .link files gained support for setting MDI/MID-X on a link.
1696 * .link files gained support for [Match] Firmware= setting to match on
1697 the device firmware description string. By mistake, it was previously
1698 only supported in .network files.
1700 * .link files gained support for [Link] SR-IOVVirtualFunctions= setting
1701 and [SR-IOV] section to configure SR-IOV virtual functions.
1703 Changes in systemd-networkd:
1705 * The default scope for unicast routes configured through [Route]
1706 section is changed to "link", to make the behavior consistent with
1707 "ip route" command. The manual configuration of [Route] Scope= is
1710 * A new unit systemd-networkd-wait-online@<interface>.service has been
1711 added that can be used to wait for a specific network interface to be
1714 * systemd-networkd gained a new [Bridge] Isolated=true|false setting
1715 that configures the eponymous kernel attribute on the bridge.
1717 * .netdev files now can be used to create virtual WLAN devices, and
1718 configure various settings on them, via the [WLAN] section.
1720 * .link/.network files gained support for [Match] Kind= setting to match
1721 on device kind ("bond", "bridge", "gre", "tun", "veth", etc.)
1723 This value is also shown by 'networkctl status'.
1725 * The Local= setting in .netdev files for various virtual network
1726 devices gained support for specifying, in addition to the network
1727 address, the name of a local interface which must have the specified
1730 * systemd-networkd gained a new [Tunnel] External= setting in .netdev
1731 files, to configure tunnels in external mode (a.k.a. collect metadata
1734 * [Network] L2TP= setting was removed. Please use interface specifier in
1735 Local= setting in .netdev files of corresponding L2TP interface.
1737 * New [DHCPServer] BootServerName=, BootServerAddress=, and
1738 BootFilename= settings can be used to configure the server address,
1739 server name, and file name sent in the DHCP packet (e.g. to configure
1742 Changes in systemd-resolved:
1744 * systemd-resolved is started earlier (in sysinit.target), so it
1745 available earlier and will also be started in the initrd if installed
1748 Changes in disk encryption:
1750 * systemd-cryptenroll can now control whether to require the user to
1751 enter a PIN when using TPM-based unlocking of a volume via the new
1752 --tpm2-with-pin= option.
1754 Option tpm2-pin= can be used in /etc/crypttab.
1756 * When unlocking devices via TPM, TPM2 parameter encryption is now
1757 used, to ensure that communication between CPU and discrete TPM chips
1758 cannot be eavesdropped to acquire disk encryption keys.
1760 * A new switch --fido2-credential-algorithm= has been added to
1761 systemd-cryptenroll allowing selection of the credential algorithm to
1762 use when binding encryption to FIDO2 tokens.
1764 Changes in systemd-hostnamed:
1766 * HARDWARE_VENDOR= and HARDWARE_MODEL= can be set in /etc/machine-info
1767 to override the values gleaned from the hwdb.
1769 * A ID_CHASSIS property can be set in the hwdb (for the DMI device
1770 /sys/class/dmi/id) to override the chassis that is reported by
1773 * hostnamed's D-Bus interface gained a new method GetHardwareSerial()
1774 for reading the hardware serial number, as reportd by DMI. It also
1775 exposes a new method D-Bus property FirmwareVersion that encode the
1776 firmware version of the system.
1778 Changes in other components:
1780 * /etc/locale.conf is now populated through tmpfiles.d factory /etc/
1781 handling with the values that were configured during systemd build
1782 (if /etc/locale.conf has not been created through some other
1783 mechanism). This means that /etc/locale.conf should always have
1784 reasonable contents and we avoid a potential mismatch in defaults.
1786 * The userdbctl tool will now show UID range information as part of the
1787 list of known users.
1789 * A new build-time configuration setting default-user-shell= can be
1790 used to set the default shell for user records and nspawn shell
1791 invocations (instead of the default /bin/bash).
1793 * systemd-timesyncd now provides a D-Bus API for receiving NTP server
1794 information dynamically at runtime via IPC.
1796 * The systemd-creds tool gained a new "has-tpm2" verb, which reports
1797 whether a functioning TPM2 infrastructure is available, i.e. if
1798 firmware, kernel driver and systemd all have TPM2 support enabled and
1801 * The systemd-creds tool gained support for generating encrypted
1802 credentials that are using an empty encryption key. While this
1803 provides no integrity nor confidentiality it's useful to implement
1804 codeflows that work the same on TPM-ful and TPM2-less systems. The
1805 service manager will only accept credentials "encrypted" that way if
1806 a TPM2 device cannot be detected, to ensure that credentials
1807 "encrypted" like that cannot be used to trick TPM2 systems.
1809 * When deciding whether to colorize output, all systemd programs now
1810 also check $COLORTERM (in addition to $NO_COLOR, $SYSTEMD_COLORS, and
1813 * Meson's new install_tag feature is now in use for several components,
1814 allowing to build and install select binaries only: pam, nss, devel
1815 (pkg-config files), systemd-boot, libsystemd, libudev. Example:
1816 $ meson build systemd-boot
1817 $ meson install --tags systemd-boot --no-rebuild
1818 https://mesonbuild.com/Installing.html#installation-tags
1820 * A new build configuration option has been added, to allow selecting the
1821 default compression algorithm used by systemd-journald and systemd-coredump.
1822 This allows to build-in support for decompressing all supported formats,
1823 but choose a specific one for compression. E.g.:
1824 $ meson -Ddefault-compression=xz
1826 Experimental features:
1828 * sd-boot gained a new *experimental* setting "reboot-for-bitlocker" in
1829 loader.conf that implements booting Microsoft Windows from the
1830 sd-boot in a way that first reboots the system, to reset the TPM
1831 PCRs. This improves compatibility with BitLocker's TPM use, as the
1832 PCRs will only record the Windows boot process, and not sd-boot
1833 itself, thus retaining the PCR measurements not involving sd-boot.
1834 Note that this feature is experimental for now, and is likely going
1835 to be generalized and renamed in a future release, without retaining
1836 compatibility with the current implementation.
1838 * A new systemd-sysupdate component has been added that automatically
1839 discovers, downloads, and installs A/B-style updates for the host
1840 installation itself, or container images, portable service images,
1841 and other assets. See the new systemd-sysupdate man page for updates.
1843 Contributions from: 4piu, Adam Williamson, adrian5, Albert Brox,
1844 AlexCatze, Alex Henrie, Alfonso Sánchez-Beato, Alice S,
1845 Alvin Šipraga, amarjargal, Amarjargal, Andrea Pappacoda,
1846 Andreas Rammhold, Andy Chi, Anita Zhang, Antonio Alvarez Feijoo,
1847 Arfrever Frehtes Taifersar Arahesis, ash, Bastien Nocera, Be,
1848 bearhoney, Ben Efros, Benjamin Berg, Benjamin Franzke,
1849 Brett Holman, Christian Brauner, Clyde Byrd III, Curtis Klein,
1850 Daan De Meyer, Daniele Medri, Daniel Mack, Danilo Krummrich,
1851 David, David Bond, Davide Cavalca, David Tardon, davijosw,
1852 dependabot[bot], Donald Chan, Dorian Clay, Eduard Tolosa,
1853 Elias Probst, Eli Schwartz, Erik Sjölund, Evgeny Vereshchagin,
1854 Federico Ceratto, Franck Bui, Frantisek Sumsal, Gaël PORTAY,
1855 Georges Basile Stavracas Neto, Gibeom Gwon, Goffredo Baroncelli,
1856 Grigori Goronzy, Hans de Goede, Heiko Becker, Hugo Carvalho,
1857 Jakob Lell, James Hilliard, Jan Janssen, Jason A. Donenfeld,
1858 Joan Bruguera, Joerie de Gram, Josh Triplett, Julia Kartseva,
1859 Kazuo Moriwaka, Khem Raj, ksa678491784, Lance, Lan Tian,
1860 Laura Barcziova, Lennart Poettering, Leviticoh, licunlong,
1861 Lidong Zhong, lincoln auster, Lubomir Rintel, Luca Boccassi,
1862 Luca BRUNO, lucagoc, Ludwig Nussel, Marcel Hellwig, march1993,
1863 Marco Scardovi, Mario Limonciello, Mariusz Tkaczyk,
1864 Markus Weippert, Martin, Martin Liska, Martin Wilck, Matija Skala,
1865 Matthew Blythe, Matthias Lisin, Matthijs van Duin, Matt Walton,
1866 Max Gautier, Michael Biebl, Michael Olbrich, Michal Koutný,
1867 Michal Sekletár, Mike Gilbert, MkfsSion, Morten Linderud,
1868 Nick Rosbrook, Nikolai Grigoriev, Nikolai Kostrigin,
1869 Nishal Kulkarni, Noel Kuntze, Pablo Ceballos, Peter Hutterer,
1870 Peter Morrow, Pigmy-penguin, Piotr Drąg, prumian, Richard Neill,
1871 Rike-Benjamin Schuppner, rodin-ia, Romain Naour, Ruben Kerkhof,
1872 Ryan Hendrickson, Santa Wiryaman, Sebastian Pucilowski, Seth Falco,
1873 Simon Ellmann, Sonali Srivastava, Stefan Seering,
1874 Stephen Hemminger, tawefogo, techtino, Temuri Doghonadze,
1875 Thomas Batten, Thomas Haller, Thomas Weißschuh, Tobias Stoeckmann,
1876 Tomasz Pala, Tyson Whitehead, Vishal Chillara Srinivas,
1877 Vivien Didelot, w30023233, wangyuhang, Weblate, Xiaotian Wu,
1878 yangmingtai, YmrDtnJu, Yonathan Randolph, Yutsuten, Yu Watanabe,
1879 Zbigniew Jędrzejewski-Szmek, наб
1881 — Edinburgh, 2022-05-21
1885 * Support for encrypted and authenticated credentials has been added.
1886 This extends the credential logic introduced with v247 to support
1887 non-interactive symmetric encryption and authentication, based on a
1888 key that is stored on the /var/ file system or in the TPM2 chip (if
1889 available), or the combination of both (by default if a TPM2 chip
1890 exists the combination is used, otherwise the /var/ key only). The
1891 credentials are automatically decrypted at the moment a service is
1892 started, and are made accessible to the service itself in unencrypted
1893 form. A new tool 'systemd-creds' encrypts credentials for this
1894 purpose, and two new service file settings LoadCredentialEncrypted=
1895 and SetCredentialEncrypted= configure such credentials.
1897 This feature is useful to store sensitive material such as SSL
1898 certificates, passwords and similar securely at rest and only decrypt
1899 them when needed, and in a way that is tied to the local OS
1900 installation or hardware.
1902 * systemd-gpt-auto-generator can now automatically set up discoverable
1903 LUKS2 encrypted swap partitions.
1905 * The GPT Discoverable Partitions Specification has been substantially
1906 extended with support for root and /usr/ partitions for the majority
1907 of architectures systemd supports. This includes platforms that do
1908 not natively support UEFI, because even though GPT is specified under
1909 UEFI umbrella, it is useful on other systems too. Specifically,
1910 systemd-nspawn, systemd-sysext, systemd-gpt-auto-generator and
1911 Portable Services use the concept without requiring UEFI.
1913 * The GPT Discoverable Partitions Specifications has been extended with
1914 a new set of partitions that may carry PKCS#7 signatures for Verity
1915 partitions, encoded in a simple JSON format. This implements a simple
1916 mechanism for building disk images that are fully authenticated and
1917 can be tested against a set of cryptographic certificates. This is
1918 now implemented for the various systemd tools that can operate with
1919 disk images, such as systemd-nspawn, systemd-sysext, systemd-dissect,
1920 Portable services/RootImage=, systemd-tmpfiles, and systemd-sysusers.
1921 The PKCS#7 signatures are passed to the kernel (where they are
1922 checked against certificates from the kernel keyring), or can be
1923 verified against certificates provided in userspace (via a simple
1924 drop-in file mechanism).
1926 * systemd-dissect's inspection logic will now report for which uses a
1927 disk image is intended. Specifically, it will display whether an
1928 image is suitable for booting on UEFI or in a container (using
1929 systemd-nspawn's --image= switch), whether it can be used as portable
1930 service, or attached as system extension.
1932 * The system-extension.d/ drop-in files now support a new field
1933 SYSEXT_SCOPE= that may encode which purpose a system extension image
1934 is for: one of "initrd", "system" or "portable". This is useful to
1935 make images more self-descriptive, and to ensure system extensions
1936 cannot be attached in the wrong contexts.
1938 * The os-release file learnt a new PORTABLE_PREFIXES= field which may
1939 be used in portable service images to indicate which unit prefixes
1942 * The GPT image dissection logic in systemd-nspawn/systemd-dissect/…
1943 now is able to decode images for non-native architectures as well.
1944 This allows systemd-nspawn to boot images of non-native architectures
1945 if the corresponding user mode emulator is installed and
1946 systemd-binfmtd is running.
1948 * systemd-logind gained new settings HandlePowerKeyLongPress=,
1949 HandleRebootKeyLongPress=, HandleSuspendKeyLongPress= and
1950 HandleHibernateKeyLongPress= which may be used to configure actions
1951 when the relevant keys are pressed for more than 5s. This is useful
1952 on devices that only have hardware for a subset of these keys. By
1953 default, if the reboot key is pressed long the poweroff operation is
1954 now triggered, and when the suspend key is pressed long the hibernate
1955 operation is triggered. Long pressing the other two keys currently
1956 does not trigger any operation by default.
1958 * When showing unit status updates on the console during boot and
1959 shutdown, and a service is slow to start so that the cylon animation
1960 is shown, the most recent sd_notify() STATUS= text is now shown as
1961 well. Services may use this to make the boot/shutdown output easier
1962 to understand, and to indicate what precisely a service that is slow
1963 to start or stop is waiting for. In particular, the per-user service
1964 manager instance now reports what it is doing and which service it is
1965 waiting for this way to the system service manager.
1967 * The service manager will now re-execute on reception of the
1968 SIGRTMIN+25 signal. It previously already did that on SIGTERM — but
1969 only when running as PID 1. There was no signal to request this when
1970 running as per-user service manager, i.e. as any other PID than 1.
1971 SIGRTMIN+25 works for both system and user managers.
1973 * The hardware watchdog logic in PID 1 gained support for operating
1974 with the default timeout configured in the hardware, instead of
1975 insisting on re-configuring it. Set RuntimeWatchdogSec=default to
1976 request this behavior.
1978 * A new kernel command line option systemd.watchdog_sec= is now
1979 understood which may be used to override the hardware watchdog
1980 time-out for the boot.
1982 * A new setting DefaultOOMScoreAdjust= is now supported in
1983 /etc/systemd/system.conf and /etc/systemd/user.conf. It may be used
1984 to set the default process OOM score adjustment value for processes
1985 started by the service manager. For per-user service managers this
1986 now defaults to 100, but for per-system service managers is left as
1987 is. This means that by default now services forked off the user
1988 service manager are more likely to be killed by the OOM killer than
1989 system services or the managers themselves.
1991 * A new per-service setting RestrictFileSystems= as been added that
1992 restricts the file systems a service has access to by their type.
1993 This is based on the new BPF LSM of the Linux kernel. It provides an
1994 effective way to make certain API file systems unavailable to
1995 services (and thus minimizing attack surface). A new command
1996 "systemd-analyze filesystems" has been added that lists all known
1997 file system types (and how they are grouped together under useful
2000 * Services now support a new setting RestrictNetworkInterfaces= for
2001 restricting access to specific network interfaces.
2003 * Service unit files gained new settings StartupAllowedCPUs= and
2004 StartupAllowedMemoryNodes=. These are similar to their counterparts
2005 without the "Startup" prefix and apply during the boot process
2006 only. This is useful to improve boot-time behavior of the system and
2007 assign resources differently during boot than during regular
2008 runtime. This is similar to the preexisting StartupCPUWeight=
2011 * Related to this: the various StartupXYZ= settings
2012 (i.e. StartupCPUWeight=, StartupAllowedCPUs=, …) are now also applied
2013 during shutdown. The settings not prefixed with "Startup" hence apply
2014 during regular runtime, and those that are prefixed like that apply
2015 during boot and shutdown.
2017 * A new per-unit set of conditions/asserts
2018 [Condition|Assert][Memory|CPU|IO]Pressure= have been added to make a
2019 unit skip/fail activation if the system's (or a slice's) memory/cpu/io
2020 pressure is above the configured threshold, using the kernel PSI
2021 feature. For more details see systemd.unit(5) and
2022 https://docs.kernel.org/accounting/psi.html
2024 * The combination of ProcSubset=pid and ProtectKernelTunables=yes and/or
2025 ProtectKernelLogs=yes can now be used.
2027 * The default maximum numbers of inodes have been raised from 64k to 1M
2028 for /dev/, and from 400k to 1M for /tmp/.
2030 * The per-user service manager learnt support for communicating with
2031 systemd-oomd to acquire OOM kill information.
2033 * A new service setting ExecSearchPath= has been added that allows
2034 changing the search path for executables for services. It affects
2035 where we look for the binaries specified in ExecStart= and similar,
2036 and the specified directories are also added the $PATH environment
2037 variable passed to invoked processes.
2039 * A new setting RuntimeRandomizedExtraSec= has been added for service
2040 and scope units that allows extending the runtime time-out as
2041 configured by RuntimeMaxSec= with a randomized amount.
2043 * The syntax of the service unit settings RuntimeDirectory=,
2044 StateDirectory=, CacheDirectory=, LogsDirectory= has been extended:
2045 if the specified value is now suffixed with a colon, followed by
2046 another filename, the latter will be created as symbolic link to the
2047 specified directory. This allows creating these service directories
2048 together with alias symlinks to make them available under multiple
2051 * Service unit files gained two new settings TTYRows=/TTYColumns= for
2052 configuring rows/columns of the TTY device passed to
2053 stdin/stdout/stderr of the service. This is useful to propagate TTY
2054 dimensions to a virtual machine.
2056 * A new service unit file setting ExitType= has been added that
2057 specifies when to assume a service has exited. By default systemd
2058 only watches the main process of a service. By setting
2059 ExitType=cgroup it can be told to wait for the last process in a
2062 * Automount unit files gained a new setting ExtraOptions= that can be
2063 used to configure additional mount options to pass to the kernel when
2064 mounting the autofs instance.
2066 * "Urlification" (generation of ESC sequences that generate clickable
2067 hyperlinks in modern terminals) may now be turned off altogether
2070 * Path units gained new TriggerLimitBurst= and TriggerLimitIntervalSec=
2071 settings that default to 200 and 2 s respectively. The ratelimit
2072 ensures that a path unit cannot cause PID1 to busy-loop when it is
2073 trying to trigger a service that is skipped because of a Condition*=
2074 not being satisfied. This matches the configuration and behaviour of
2077 * The TPM2/FIDO2/PKCS11 support in systemd-cryptsetup is now also built
2078 as a plug-in for cryptsetup. This means the plain cryptsetup command
2079 may now be used to unlock volumes set up this way.
2081 * The TPM2 logic in cryptsetup will now automatically detect systems
2082 where the TPM2 chip advertises SHA256 PCR banks but the firmware only
2083 updates the SHA1 banks. In such a case PCR policies will be
2084 automatically bound to the latter, not the former. This makes the PCR
2085 policies reliable, but of course do not provide the same level of
2086 trust as SHA256 banks.
2088 * The TPM2 logic in systemd-cryptsetup/systemd-cryptsetup now supports
2089 RSA primary keys in addition to ECC, improving compatibility with
2090 TPM2 chips that do not support ECC. RSA keys are much slower to use
2091 than ECC, and hence are only used if ECC is not available.
2093 * /etc/crypttab gained support for a new token-timeout= setting for
2094 encrypted volumes that allows configuration of the maximum time to
2095 wait for PKCS#11/FIDO2 tokens to be plugged in. If the time elapses
2096 the logic will query the user for a regular passphrase/recovery key
2099 * Support for activating dm-integrity volumes at boot via a new file
2100 /etc/integritytab and the tool systemd-integritysetup have been
2101 added. This is similar to /etc/crypttab and /etc/veritytab, but deals
2102 with dm-integrity instead of dm-crypt/dm-verity.
2104 * The systemd-veritysetup-generator now understands a new usrhash=
2105 kernel command line option for specifying the Verity root hash for
2106 the partition backing the /usr/ file system. A matching set of
2107 systemd.verity_usr_* kernel command line options has been added as
2108 well. These all work similar to the corresponding options for the
2111 * The sd-device API gained a new API call sd_device_get_diskseq() to
2112 return the DISKSEQ property of a device structure. The "disk
2113 sequence" concept is a new feature recently introduced to the Linux
2114 kernel that allows detecting reuse cycles of block devices, i.e. can
2115 be used to recognize when loopback block devices are reused for a
2116 different purpose or CD-ROM drives get their media changed.
2118 * A new unit systemd-boot-update.service has been added. If enabled
2119 (the default) and the sd-boot loader is detected to be installed, it
2120 is automatically updated to the newest version when out of date. This
2121 is useful to ensure the boot loader remains up-to-date, and updates
2122 automatically propagate from the OS tree in /usr/.
2124 * sd-boot will now build with SBAT by default in order to facilitate
2125 working with recent versions of Shim that require it to be present.
2127 * sd-boot can now parse Microsoft Windows' Boot Configuration Data.
2128 This is used to robustly generate boot entry titles for Windows.
2130 * A new generic target unit factory-reset.target has been added. It is
2131 hooked into systemd-logind similar in fashion to
2132 reboot/poweroff/suspend/hibernate, and is supposed to be used to
2133 initiate a factory reset operation. What precisely this operation
2134 entails is up for the implementer to decide, the primary goal of the
2135 new unit is provide a framework where to plug in the implementation
2136 and how to trigger it.
2138 * A new meson build-time option 'clock-valid-range-usec-max' has been
2139 added which takes a time in µs and defaults to 15 years. If the RTC
2140 time is noticed to be more than the specified time ahead of the
2141 built-in epoch of systemd (which by default is the release timestamp
2142 of systemd) it is assumed that the RTC is not working correctly, and
2143 the RTC is reset to the epoch. (It already is reset to the epoch when
2144 noticed to be before it.) This should increase the chance that time
2145 doesn't accidentally jump too far ahead due to faulty hardware or
2148 * A new setting SaveIntervalSec= has been added to systemd-timesyncd,
2149 which may be used to automatically save the current system time to
2150 disk in regular intervals. This is useful to maintain a roughly
2151 monotonic clock even without RTC hardware and with some robustness
2152 against abnormal system shutdown.
2154 * systemd-analyze verify gained support for a pair of new --image= +
2155 --root= switches for verifying units below a specific root
2156 directory/image instead of on the host.
2158 * systemd-analyze verify gained support for verifying unit files under
2159 an explicitly specified unit name, independently of what the filename
2162 * systemd-analyze verify gained a new switch --recursive-errors= which
2163 controls whether to only fail on errors found in the specified units
2164 or recursively any dependent units.
2166 * systemd-analyze security now supports a new --offline mode for
2167 analyzing unit files stored on disk instead of loaded units. It may
2168 be combined with --root=/--image to analyze unit files under a root
2169 directory or disk image. It also learnt a new --threshold= parameter
2170 for specifying an exposure level threshold: if the exposure level
2171 exceeds the specified value the call will fail. It also gained a new
2172 --security-policy= switch for configuring security policies to
2173 enforce on the units. A policy is a JSON file that lists which tests
2174 shall be weighted how much to determine the overall exposure
2175 level. Altogether these new features are useful for fully automatic
2176 analysis and enforcement of security policies on unit files.
2178 * systemd-analyze security gain a new --json= switch for JSON output.
2180 * systemd-analyze learnt a new --quiet switch for reducing
2181 non-essential output. It's honored by the "dot", "syscall-filter",
2182 "filesystems" commands.
2184 * systemd-analyze security gained a --profile= option that can be used
2185 to take into account a portable profile when analyzing portable
2186 services, since a lot of the security-related settings are enabled
2189 * systemd-analyze learnt a new inspect-elf verb that parses ELF core
2190 files, binaries and executables and prints metadata information,
2191 including the build-id and other info described on:
2192 https://systemd.io/COREDUMP_PACKAGE_METADATA/
2194 * .network files gained a new UplinkInterface= in the [IPv6SendRA]
2195 section, for automatically propagating DNS settings from other
2198 * The static lease DHCP server logic in systemd-networkd may now serve
2199 IP addresses outside of the configured IP pool range for the server.
2201 * CAN support in systemd-networkd gained four new settings Loopback=,
2202 OneShot=, PresumeAck=, ClassicDataLengthCode= for tweaking CAN
2203 control modes. It gained a number of further settings for tweaking
2206 * The [CAN] section in .network file gained new TimeQuantaNSec=,
2207 PropagationSegment=, PhaseBufferSegment1=, PhaseBufferSegment2=,
2208 SyncJumpWidth=, DataTimeQuantaNSec=, DataPropagationSegment=,
2209 DataPhaseBufferSegment1=, DataPhaseBufferSegment2=, and
2210 DataSyncJumpWidth= settings to control bit-timing processed by the
2213 * DHCPv4 client support in systemd-networkd learnt a new Label= option
2214 for configuring the address label to apply to configure IPv4
2217 * The [IPv6AcceptRA] section of .network files gained support for a new
2218 UseMTU= setting that may be used to control whether to apply the
2219 announced MTU settings to the local interface.
2221 * The [DHCPv4] section in .network file gained a new Use6RD= boolean
2222 setting to control whether the DHCPv4 client request and process the
2225 * The [DHCPv6PrefixDelegation] section in .network file is renamed to
2226 [DHCPPrefixDelegation], as now the prefix delegation is also supported
2227 with DHCPv4 protocol by enabling the Use6RD= setting.
2229 * The [DHCPPrefixDelegation] section in .network file gained a new
2230 setting UplinkInterface= to specify the upstream interface.
2232 * The [DHCPv6] section in .network file gained a new setting
2233 UseDelegatedPrefix= to control whether the delegated prefixes will be
2234 propagated to the downstream interfaces.
2236 * The [IPv6AcceptRA] section of .network files now understands two new
2237 settings UseGateway=/UseRoutePrefix= for explicitly configuring
2238 whether to use the relevant fields from the IPv6 Router Advertisement
2241 * The ForceDHCPv6PDOtherInformation= setting in the [DHCPv6] section
2242 has been removed. Please use the WithoutRA= and UseDelegatedPrefix=
2243 settings in the [DHCPv6] section and the DHCPv6Client= setting in the
2244 [IPv6AcceptRA] section to control when the DHCPv6 client is started
2245 and how the delegated prefixes are handled by the DHCPv6 client.
2247 * The IPv6Token= section in the [Network] section is deprecated, and
2248 the [IPv6AcceptRA] section gained the Token= setting for its
2249 replacement. The [IPv6Prefix] section also gained the Token= setting.
2250 The Token= setting gained 'eui64' mode to explicitly configure an
2251 address with the EUI64 algorithm based on the interface MAC address.
2252 The 'prefixstable' mode can now optionally take a secret key. The
2253 Token= setting in the [DHCPPrefixDelegation] section now supports all
2254 algorithms supported by the same settings in the other sections.
2256 * The [RoutingPolicyRule] section of .network file gained a new
2257 SuppressInterfaceGroup= setting.
2259 * The IgnoreCarrierLoss= setting in the [Network] section of .network
2260 files now allows a duration to be specified, controlling how long to
2261 wait before reacting to carrier loss.
2263 * The [DHCPServer] section of .network file gained a new Router=
2264 setting to specify the router address.
2266 * The [CAKE] section of .network files gained various new settings
2267 AutoRateIngress=, CompensationMode=, FlowIsolationMode=, NAT=,
2268 MPUBytes=, PriorityQueueingPreset=, FirewallMark=, Wash=, SplitGSO=,
2269 and UseRawPacketSize= for configuring CAKE.
2271 * systemd-networkd now ships with new default .network files:
2272 80-container-vb.network which matches host-side network bridge device
2273 created by systemd-nspawn's --network-bridge or --network-zone
2274 switch, and 80-6rd-tunnel.network which matches automatically created
2275 sit tunnel with 6rd prefix when the DHCP 6RD option is received.
2277 * systemd-networkd's handling of Endpoint= resolution for WireGuard
2278 interfaces has been improved.
2280 * systemd-networkd will now automatically configure routes to addresses
2281 specified in AllowedIPs=. This feature can be controlled via
2282 RouteTable= and RouteMetric= settings in [WireGuard] or
2283 [WireGuardPeer] sections.
2285 * systemd-networkd will now once again automatically generate persistent
2286 MAC addresses for batadv and bridge interfaces. Users can disable this
2287 by using MACAddress=none in .netdev files.
2289 * systemd-networkd and systemd-udevd now support IP over InfiniBand
2290 interfaces. The Kind= setting in .netdev file accepts "ipoib". And
2291 systemd.netdev files gained the [IPoIB] section.
2293 * systemd-networkd and systemd-udevd now support net.ifname-policy=
2294 option on the kernel command-line. This is implemented through the
2295 systemd-network-generator service that automatically generates
2296 appropriate .link, .network, and .netdev files.
2298 * The various systemd-udevd "ethtool" buffer settings now understand
2299 the special value "max" to configure the buffers to the maximum the
2302 * systemd-udevd's .link files may now configure a large variety of
2303 NIC coalescing settings, plus more hardware offload settings.
2305 * .link files gained a new WakeOnLanPassword= setting in the [Link]
2306 section that allows to specify a WoL "SecureOn" password on hardware
2309 * systemd-nspawn's --setenv= switch now supports an additional syntax:
2310 if only a variable name is specified (i.e. without being suffixed by
2311 a '=' character and a value) the current value of the environment
2312 variable is propagated to the container. e.g. --setenv=FOO will
2313 lookup the current value of $FOO in the environment, and pass it down
2314 to the container. Similar behavior has been added to homectl's,
2315 machinectl's and systemd-run's --setenv= switch.
2317 * systemd-nspawn gained a new switch --suppress-sync= which may be used
2318 to optionally suppress the effect of the sync()/fsync()/fdatasync()
2319 system calls for the container payload. This is useful for build
2320 system environments where safety against abnormal system shutdown is
2321 not essential as all build artifacts can be regenerated any time, but
2322 the performance win is beneficial.
2324 * systemd-nspawn will now raise the RLIMIT_NOFILE hard limit to the
2325 same value that PID 1 uses for most forked off processes.
2327 * systemd-nspawn's --bind=/--bind-ro= switches now optionally take
2328 uidmap/nouidmap options as last parameter. If "uidmap" is used the
2329 bind mounts are created with UID mapping taking place that ensures
2330 the host's file ownerships are mapped 1:1 to container file
2331 ownerships, even if user namespacing is used. This way
2332 files/directories bound into containers will no longer show up as
2333 owned by the nobody user as they typically did if no special care was
2334 taken to shift them manually.
2336 * When discovering Windows installations sd-boot will now attempt to
2337 show the Windows version.
2339 * The color scheme to use in sd-boot may now be configured at
2342 * sd-boot gained the ability to change screen resolution during
2343 boot-time, by hitting the "r" key. This will cycle through available
2344 resolutions and save the last selection.
2346 * sd-boot learnt a new hotkey "f". When pressed the system will enter
2347 firmware setup. This is useful in environments where it is difficult
2348 to hit the right keys early enough to enter the firmware, and works
2349 on any firmware regardless which key it natively uses.
2351 * sd-boot gained support for automatically booting into the menu item
2352 selected on the last boot (using the "@saved" identifier for menu
2355 * sd-boot gained support for automatically loading all EFI drivers
2356 placed in the /EFI/systemd/drivers/ subdirectory of the EFI System
2357 Partition (ESP). These drivers are loaded before the menu entries are
2358 loaded. This is useful e.g. to load additional file system drivers
2359 for the XBOOTLDR partition.
2361 * systemd-boot will now paint the input cursor on its own instead of
2362 relying on the firmware to do so, increasing compatibility with broken
2363 firmware that doesn't make the cursor reasonably visible.
2365 * sd-boot now embeds a .osrel PE section like we expect from Boot
2366 Loader Specification Type #2 Unified Kernels. This means sd-boot
2367 itself may be used in place of a Type #2 Unified Kernel. This is
2368 useful for debugging purposes as it allows chain-loading one a
2369 (development) sd-boot instance from another.
2371 * sd-boot now supports a new "devicetree" field in Boot Loader
2372 Specification Type #1 entries: if configured the specified device
2373 tree file is installed before the kernel is invoked. This is useful
2374 for installing/applying new devicetree files without updating the
2377 * Similarly, sd-stub now can read devicetree data from a PE section
2378 ".dtb" and apply it before invoking the kernel.
2380 * sd-stub (the EFI stub that can be glued in front of a Linux kernel)
2381 gained the ability to pick up credentials and sysext files, wrap them
2382 in a cpio archive, and pass as an additional initrd to the invoked
2383 Linux kernel, in effect placing those files in the /.extra/ directory
2384 of the initrd environment. This is useful to implement trusted initrd
2385 environments which are fully authenticated but still can be extended
2386 (via sysexts) and parameterized (via encrypted/authenticated
2387 credentials, see above).
2389 Credentials can be located next to the kernel image file (credentials
2390 specific to a single boot entry), or in one of the shared directories
2391 (credentials applicable to multiple boot entries).
2393 * sd-stub now comes with a full man page, that explains its feature set
2394 and how to combine a kernel image, an initrd and the stub to build a
2395 complete EFI unified kernel image, implementing Boot Loader
2396 Specification Type #2.
2398 * sd-stub may now provide the initrd to the executed kernel via the
2399 LINUX_EFI_INITRD_MEDIA_GUID EFI protocol, adding compatibility for
2400 non-x86 architectures.
2402 * bootctl learnt new set-timeout and set-timeout-oneshot commands that
2403 may be used to set the boot menu time-out of the boot loader (for all
2404 or just the subsequent boot).
2406 * bootctl and kernel-install will now read variables
2407 KERNEL_INSTALL_LAYOUT= from /etc/machine-info and layout= from
2408 /etc/kernel/install.conf. When set, it specifies the layout to use
2409 for installation directories on the boot partition, so that tools
2410 don't need to guess it based on the already-existing directories. The
2411 only value that is defined natively is "bls", corresponding to the
2413 https://systemd.io/BOOT_LOADER_SPECIFICATION/. Plugins for
2414 kernel-install that implement a different layout can declare other
2415 values for this variable.
2417 'bootctl install' will now write KERNEL_INSTALL_LAYOUT=bls, on the
2418 assumption that if the user installed sd-boot to the ESP, they intend
2419 to use the entry layout understood by sd-boot. It'll also write
2420 KERNEL_INSTALL_MACHINE_ID= if it creates any directories using the ID
2421 (and it wasn't specified in the config file yet). Similarly,
2422 kernel-install will now write KERNEL_INSTALL_MACHINE_ID= (if it
2423 wasn't specified in the config file yet). Effectively, those changes
2424 mean that the machine-id used for boot loader entry installation is
2425 "frozen" upon first use and becomes independent of the actual
2428 Configuring KERNEL_INSTALL_MACHINE_ID fixes the following problem:
2429 images created for distribution ("golden images") are built with no
2430 machine-id, so that a unique machine-id can be created on the first
2431 boot. But those images may contain boot loader entries with the
2432 machine-id used during build included in paths. Using a "frozen"
2433 value allows unambiguously identifying entries that match the
2434 specific installation, while still permitting parallel installations
2437 Configuring KERNEL_INSTALL_LAYOUT obviates the need for
2438 kernel-install to guess the installation layout. This fixes the
2439 problem where a (possibly empty) directory in the boot partition is
2440 created from a different layout causing kernel-install plugins to
2441 assume the wrong layout. A particular example of how this may happen
2442 is the grub2 package in Fedora which includes directories under /boot
2443 directly in its file list. Various other packages pull in grub2 as a
2444 dependency, so it may be installed even if unused, breaking
2445 installations that use the bls layout.
2447 * bootctl and systemd-bless-boot can now be linked statically.
2449 * systemd-sysext now optionally doesn't insist on extension-release.d/
2450 files being placed in the image under the image's file name. If the
2451 file system xattr user.extension-release.strict is set on the
2452 extension release file, it is accepted regardless of its name. This
2453 relaxes security restrictions a bit, as system extension may be
2454 attached under a wrong name this way.
2456 * udevadm's test-builtin command learnt a new --action= switch for
2457 testing the built-in with the specified action (in place of the
2460 * udevadm info gained new switches --property=/--value for showing only
2461 specific udev properties/values instead of all.
2463 * A new hwdb database has been added that contains matches for various
2464 types of signal analyzers (protocol analyzers, logic analyzers,
2465 oscilloscopes, multimeters, bench power supplies, etc.) that should
2466 be accessible to regular users.
2468 * A new hwdb database entry has been added that carries information
2469 about types of cameras (regular or infrared), and in which direction
2470 they point (front or back).
2472 * A new rule to allow console users access to rfkill by default has been
2475 * Device nodes for the Software Guard eXtension enclaves (sgx_vepc) are
2476 now also owned by the system group "sgx".
2478 * A new build-time meson option "extra-net-naming-schemes=" has been
2479 added to define additional naming schemes for udev's network
2480 interface naming logic. This is useful for enterprise distributions
2481 and similar which want to pin the schemes of certain distribution
2482 releases under a specific name and previously had to patch the
2483 sources to introduce new named schemes.
2485 * The predictable naming logic for network interfaces has been extended
2486 to generate stable names from Xen netfront device information.
2488 * hostnamed's chassis property can now be sourced from chassis-type
2489 field encoded in devicetree (in addition to the existing DMI
2492 * systemd-cgls now optionally displays cgroup IDs and extended
2493 attributes for each cgroup. (Controllable via the new --xattr= +
2494 --cgroup-id= switches.)
2496 * coredumpctl gained a new --all switch for operating on all
2497 Journal files instead of just the local ones.
2499 * systemd-coredump will now use libdw/libelf via dlopen() rather than
2500 directly linking, allowing users to easily opt-out of backtrace/metadata
2501 analysis of core files, and reduce image sizes when this is not needed.
2503 * systemd-coredump will now analyze core files with libdw/libelf in a
2504 forked, sandboxed process.
2506 * systemd-homed will now try to unmount an activate home area in
2507 regular intervals once the user logged out fully. Previously this was
2508 attempted exactly once but if the home directory was busy for some
2509 reason it was not tried again.
2511 * systemd-homed's LUKS2 home area backend will now create a BSD file
2512 system lock on the image file while the home area is active
2513 (i.e. mounted). If a home area is found to be locked, logins are
2514 politely refused. This should improve behavior when using home areas
2515 images that are accessible via the network from multiple clients, and
2516 reduce the chance of accidental file system corruption in that case.
2518 * Optionally, systemd-homed will now drop the kernel buffer cache once
2519 a user has fully logged out, configurable via the new --drop-caches=
2522 * systemd-homed now makes use of UID mapped mounts for the home areas.
2523 If the kernel and used file system support it, files are now
2524 internally owned by the "nobody" user (i.e. the user typically used
2525 for indicating "this ownership is not mapped"), and dynamically
2526 mapped to the UID used locally on the system via the UID mapping
2527 mount logic of recent kernels. This makes migrating home areas
2528 between different systems cheaper because recursively chown()ing file
2529 system trees is no longer necessary.
2531 * systemd-homed's CIFS backend now optionally supports CIFS service
2532 names with a directory suffix, in order to place home directories in
2533 a subdirectory of a CIFS share, instead of the top-level directory.
2535 * systemd-homed's CIFS backend gained support for specifying additional
2536 mount options in the JSON user record (cifsExtraMountOptions field,
2537 and --cifs-extra-mount-options= homectl switch). This is for example
2538 useful for configuring mount options such as "noserverino" that some
2539 SMB3 services require (use that to run a homed home directory from a
2540 FritzBox SMB3 share this way).
2542 * systemd-homed will now default to btrfs' zstd compression for home
2543 areas. This is inspired by Fedora's recent decision to switch to zstd
2546 * Additional mount options to use when mounting the file system of
2547 LUKS2 volumes in systemd-homed has been added. Via the
2548 $SYSTEMD_HOME_MOUNT_OPTIONS_BTRFS, $SYSTEMD_HOME_MOUNT_OPTIONS_EXT4,
2549 $SYSTEMD_HOME_MOUNT_OPTIONS_XFS environment variables to
2550 systemd-homed or via the luksExtraMountOptions user record JSON
2551 property. (Exposed via homectl --luks-extra-mount-options)
2553 * homectl's resize command now takes the special size specifications
2554 "min" and "max" to shrink/grow the home area to the minimum/maximum
2555 size possible, taking disk usage/space constraints and file system
2556 limitations into account. Resizing is now generally graceful: the
2557 logic will try to get as close to the specified size as possible, but
2558 not consider it a failure if the request couldn't be fulfilled
2561 * systemd-homed gained the ability to automatically shrink home areas
2562 on logout to their minimal size and grow them again on next
2563 login. This ensures that while inactive, a home area only takes up
2564 the minimal space necessary, but once activated, it provides
2565 sufficient space for the user's needs. This behavior is only
2566 supported if btrfs is used as file system inside the home area
2567 (because only for btrfs online growing/shrinking is implemented in
2568 the kernel). This behavior is now enabled by default, but may be
2569 controlled via the new --auto-resize-mode= setting of homectl.
2571 * systemd-homed gained support for automatically re-balancing free disk
2572 space among active home areas, in case the LUKS2 backends are used,
2573 and no explicit disk size was requested. This way disk space is
2574 automatically managed and home areas resized in regular intervals and
2575 manual resizing when disk space becomes scarce should not be
2576 necessary anymore. This behavior is only supported if btrfs is used
2577 within the home areas (as only then online shrinking and growing is
2578 supported), and may be configured via the new rebalanceWeight JSON
2579 user record field (as exposed via the new --rebalance-weight= homectl
2580 setting). Re-balancing is mostly automatic, but can also be requested
2581 explicitly via "homectl rebalance", which is synchronous, and thus
2582 may be used to wait until the rebalance run is complete.
2584 * userdbctl gained a --json= switch for configured the JSON formatting
2585 to use when outputting user or group records.
2587 * userdbctl gained a new --multiplexer= switch for explicitly
2588 configuring whether to use the systemd-userdbd server side user
2589 record resolution logic.
2591 * userdbctl's ssh-authorized-keys command learnt a new --chain switch,
2592 for chaining up another command to execute after completing the
2593 look-up. Since the OpenSSH's AuthorizedKeysCommand only allows
2594 configuration of a single command to invoke, this maybe used to
2595 invoke multiple: first userdbctl's own implementation, and then any
2596 other also configured in the command line.
2598 * The sd-event API gained a new function sd_event_add_inotify_fd() that
2599 is similar to sd_event_add_inotify() but accepts a file descriptor
2600 instead of a path in the file system for referencing the inode to
2603 * The sd-event API gained a new function
2604 sd_event_source_set_ratelimit_expire_callback() that may be used to
2605 define a callback function that is called whenever an event source
2606 leaves the rate limiting phase.
2608 * New documentation has been added explaining which steps are necessary
2609 to port systemd to a new architecture:
2611 https://systemd.io/PORTING_TO_NEW_ARCHITECTURES
2613 * The x-systemd.makefs option in /etc/fstab now explicitly supports
2614 ext2, ext3, and f2fs file systems.
2616 * Mount units and units generated from /etc/fstab entries with 'noauto'
2617 are now ordered the same as other units. Effectively, they will be
2618 started earlier (if something actually pulled them in) and stopped
2619 later, similarly to normal mount units that are part of
2620 fs-local.target. This change should be invisible to users, but
2621 should prevent those units from being stopped too early during
2624 * The systemd-getty-generator now honors a new kernel command line
2625 argument systemd.getty_auto= and a new environment variable
2626 $SYSTEMD_GETTY_AUTO that allows turning it off at boot. This is for
2627 example useful to turn off gettys inside of containers or similar
2630 * systemd-resolved now listens on a second DNS stub address: 127.0.0.54
2631 (in addition to 127.0.0.53, as before). If DNS requests are sent to
2632 this address they are propagated in "bypass" mode only, i.e. are
2633 almost not processed locally, but mostly forwarded as-is to the
2634 current upstream DNS servers. This provides a stable DNS server
2635 address that proxies all requests dynamically to the right upstream
2636 DNS servers even if these dynamically change. This stub does not do
2637 mDNS/LLMNR resolution. However, it will translate look-ups to
2638 DNS-over-TLS if necessary. This new stub is particularly useful in
2639 container/VM environments, or for tethering setups: use DNAT to
2640 redirect traffic to any IP address to this stub.
2642 * systemd-importd now honors new environment variables
2643 $SYSTEMD_IMPORT_BTRFS_SUBVOL, $SYSTEMD_IMPORT_BTRFS_QUOTA,
2644 $SYSTEMD_IMPORT_SYNC, which may be used disable btrfs subvolume
2645 generation, btrfs quota setup and disk synchronization.
2647 * systemd-importd and systemd-resolved can now be optionally built with
2648 OpenSSL instead of libgcrypt.
2650 * systemd-repart no longer requires OpenSSL.
2652 * systemd-sysusers will no longer create the redundant 'nobody' group
2653 by default, as the 'nobody' user is already created with an
2654 appropriate primary group.
2656 * If a unit uses RuntimeMaxSec, systemctl show will now display it.
2658 * systemctl show-environment gained support for --output=json.
2660 * pam_systemd will now first try to use the X11 abstract socket, and
2661 fallback to the socket file in /tmp/.X11-unix/ only if that does not
2664 * systemd-journald will no longer go back to volatile storage
2665 regardless of configuration when its unit is restarted.
2667 * Initial support for the LoongArch architecture has been added (system
2668 call lists, GPT partition table UUIDs, etc).
2670 * systemd-journald's own logging messages are now also logged to the
2671 journal itself when systemd-journald logs to /dev/kmsg.
2673 * systemd-journald now re-enables COW for archived journal files on
2674 filesystems that support COW. One benefit of this change is that
2675 archived journal files will now get compressed on btrfs filesystems
2676 that have compression enabled.
2678 * systemd-journald now deduplicates fields in a single log message
2679 before adding it to the journal. In archived journal files, it will
2680 also punch holes for unused parts and truncate the file as
2681 appropriate, leading to reductions in disk usage.
2683 * journalctl --verify was extended with more informative error
2686 * More of sd-journal's functions are now resistant against journal file
2689 * The shutdown command learnt a new option --show, to display the
2692 * A LICENSES/ directory is now included in the git tree. It contains a
2693 README.md file that explains the licenses used by source files in
2694 this repository. It also contains the text of all applicable
2695 licenses as they appear on spdx.org.
2697 Contributions from: Aakash Singh, acsfer, Adolfo Jayme Barrientos,
2698 Adrian Vovk, Albert Brox, Alberto Mardegan, Alexander Kanavin,
2699 alexlzhu, Alfonso Sánchez-Beato, Alvin Šipraga, Alyssa Ross,
2700 Amir Omidi, Anatol Pomozov, Andika Triwidada, Andreas Rammhold,
2701 Andreas Valder, Andrej Lajovic, Andrew Soutar, Andrew Stone, Andy Chi,
2702 Anita Zhang, Anssi Hannula, Antonio Alvarez Feijoo,
2703 Antony Deepak Thomas, Arnaud Ferraris, Arvid E. Picciani,
2704 Bastien Nocera, Benjamin Berg, Benjamin Herrenschmidt, Ben Stockett,
2705 Bogdan Seniuc, Boqun Feng, Carl Lei, chlorophyll-zz, Chris Packham,
2706 Christian Brauner, Christian Göttsche, Christian Wehrli,
2707 Christoph Anton Mitterer, Cristian Rodríguez, Daan De Meyer,
2708 Daniel Maixner, Dann Frazier, Dan Streetman, Davide Cavalca,
2709 David Seifert, David Tardon, dependabot[bot], Dimitri John Ledkov,
2710 Dimitri Papadopoulos, Dimitry Ishenko, Dmitry Khlebnikov,
2711 Dominique Martinet, duament, Egor, Egor Ignatov, Emil Renner Berthing,
2712 Emily Gonyer, Ettore Atalan, Evgeny Vereshchagin, Florian Klink,
2713 Franck Bui, Frantisek Sumsal, Geass-LL, Gibeom Gwon, GnunuX,
2714 Gogo Gogsi, gregzuro, Greg Zuro, Gustavo Costa, Hans de Goede,
2715 Hela Basa, Henri Chain, hikigaya58, Hugo Carvalho,
2716 Hugo Osvaldo Barrera, Iago Lopez Galeiras, Iago López Galeiras,
2717 I-dont-need-name, igo95862, Jack Dähn, James Hilliard, Jan Janssen,
2718 Jan Kuparinen, Jan Macku, Jan Palus, Jarkko Sakkinen, Jayce Fayne,
2719 jiangchuangang, jlempen, John Lindgren, Jonas Dreßler, Jonas Jelten,
2720 Jonas Witschel, Joris Hartog, José Expósito, Julia Kartseva,
2721 Kai-Heng Feng, Kai Wohlfahrt, Kay Siver Bø, KennthStailey,
2722 Kevin Kuehler, Kevin Orr, Khem Raj, Kristian Klausen, Kyle Laker,
2723 lainahai, LaserEyess, Lennart Poettering, Lia Lenckowski, longpanda,
2724 Luca Boccassi, Luca BRUNO, Ludwig Nussel, Lukas Senionis,
2725 Maanya Goenka, Maciek Borzecki, Marcel Menzel, Marco Scardovi,
2726 Marcus Harrison, Mark Boudreau, Matthijs van Duin, Mauricio Vásquez,
2727 Maxime de Roucy, Max Resch, MertsA, Michael Biebl, Michael Catanzaro,
2728 Michal Koutný, Michal Sekletár, Miika Karanki, Mike Gilbert,
2729 Milo Turner, ml, monosans, Nacho Barrientos, nassir90, Nishal Kulkarni,
2730 nl6720, Ondrej Kozina, Paulo Neves, Pavel Březina, pedro martelletto,
2731 Peter Hutterer, Peter Morrow, Piotr Drąg, Rasmus Villemoes, ratijas,
2732 Raul Tambre, rene, Riccardo Schirone, Robert-L-Turner, Robert Scheck,
2733 Ross Jennings, saikat0511, Scott Lamb, Scott Worley,
2734 Sergei Trofimovich, Sho Iizuka, Slava Bacherikov, Slimane Selyan Amiri,
2735 StefanBruens, Steven Siloti, svonohr, Taiki Sugawara, Takashi Sakamoto,
2736 Takuro Onoue, Thomas Blume, Thomas Haller, Thomas Mühlbacher,
2737 Tianlu Shao, Toke Høiland-Jørgensen, Tom Yan, Tony Asleson,
2738 Topi Miettinen, Ulrich Ölmann, Urs Ritzmann, Vincent Bernat,
2739 Vito Caputo, Vladimir Panteleev, WANG Xuerui, Wind/owZ, Wu Xiaotian,
2740 xdavidwu, Xiaotian Wu, xujing, yangmingtai, Yao Wei, Yao Wei (魏銘廷),
2741 Yegor Alexeyev, Yu Watanabe, Zbigniew Jędrzejewski-Szmek,
2742 Дамјан Георгиевски, наб
2744 — Warsaw, 2021-12-23
2748 * When operating on disk images via the --image= switch of various
2749 tools (such as systemd-nspawn or systemd-dissect), or when udev finds
2750 no 'root=' parameter on the kernel command line, and multiple
2751 suitable root or /usr/ partitions exist in the image, then a simple
2752 comparison inspired by strverscmp() is done on the GPT partition
2753 label, and the newest partition is picked. This permits a simple and
2754 generic whole-file-system A/B update logic where new operating system
2755 versions are dropped into partitions whose label is then updated with
2756 a matching version identifier.
2758 * systemd-sysusers now supports querying the passwords to set for the
2759 users it creates via the "credentials" logic introduced in v247: the
2760 passwd.hashed-password.<user> and passwd.plaintext-password.<user>
2761 credentials are consulted for the password to use (either in UNIX
2762 hashed form, or literally). By default these credentials are inherited
2763 down from PID1 (which in turn imports it from a container manager if
2764 there is one). This permits easy configuration of user passwords
2765 during first boot. Example:
2767 # systemd-nspawn -i foo.raw --volatile=yes --set-credential=passwd.plaintext-password.root:foo
2769 Note that systemd-sysusers operates in purely additive mode: it
2770 executes no operation if the declared users already exist, and hence
2771 doesn't set any passwords as effect of the command line above if the
2772 specified root user exists already in the image. (Note that
2773 --volatile=yes ensures it doesn't, though.)
2775 * systemd-firstboot now also supports querying various system
2776 parameters via the credential subsystems. Thus, as above this may be
2777 used to initialize important system parameters on first boot of
2778 previously unprovisioned images (i.e. images with a mostly empty
2781 * PID 1 may now show both the unit name and the unit description
2782 strings in its status output during boot. This may be configured with
2783 StatusUnitFormat=combined in system.conf or
2784 systemd.status-unit-format=combined on the kernel command line.
2786 * The systemd-machine-id-setup tool now supports a --image= switch for
2787 provisioning a machine ID file into an OS disk image, similar to how
2788 --root= operates on an OS file tree. This matches the existing switch
2789 of the same name for systemd-tmpfiles, systemd-firstboot, and
2790 systemd-sysusers tools.
2792 * Similarly, systemd-repart gained support for the --image= switch too.
2793 In combination with the existing --size= option, this makes the tool
2794 particularly useful for easily growing disk images in a single
2795 invocation, following the declarative rules included in the image
2798 * systemd-repart's partition configuration files gained support for a
2799 new switch MakeDirectories= which may be used to create arbitrary
2800 directories inside file systems that are created, before registering
2801 them in the partition table. This is useful in particular for root
2802 partitions to create mount point directories for other partitions
2803 included in the image. For example, a disk image that contains a
2804 root, /home/, and /var/ partitions, may set MakeDirectories=yes to
2805 create /home/ and /var/ as empty directories in the root file system
2806 on its creation, so that the resulting image can be mounted
2807 immediately, even in read-only mode.
2809 * systemd-repart's CopyBlocks= setting gained support for the special
2810 value "auto". If used, a suitable matching partition on the booted OS
2811 is found as source to copy blocks from. This is useful when
2812 implementing replicating installers, that are booted from one medium
2813 and then stream their own root partition onto the target medium.
2815 * systemd-repart's partition configuration files gained support for a
2816 Flags=, a ReadOnly= and a NoAuto= setting, allowing control of these
2817 GPT partition flags for the created partitions: this is useful for
2818 marking newly created partitions as read-only, or as not being
2819 subject for automatic mounting from creation on.
2821 * The /etc/os-release file has been extended with two new (optional)
2822 variables IMAGE_VERSION= and IMAGE_ID=, carrying identity and version
2823 information for OS images that are updated comprehensively and
2824 atomically as one image. Two new specifiers %M, %A now resolve to
2825 these two fields in the various configuration options that resolve
2828 * portablectl gained a new switch --extension= for enabling portable
2829 service images with extensions that follow the extension image
2830 concept introduced with v248, and thus allows layering multiple
2831 images when setting up the root filesystem of the service.
2833 * systemd-coredump will now extract ELF build-id information from
2834 processes dumping core and include it in the coredump report.
2835 Moreover, it will look for ELF .note.package sections with
2836 distribution packaging meta-information about the crashing process.
2837 This is useful to directly embed the rpm or deb (or any other)
2838 package name and version in ELF files, making it easy to match
2839 coredump reports with the specific package for which the software was
2840 compiled. This is particularly useful on environments with ELF files
2841 from multiple vendors, different distributions and versions, as is
2842 common today in our containerized and sand-boxed world. For further
2845 https://systemd.io/COREDUMP_PACKAGE_METADATA
2847 * A new udev hardware database has been added for FireWire devices
2850 * The "net_id" built-in of udev has been updated with three
2851 backwards-incompatible changes:
2853 - PCI hotplug slot names on s390 systems are now parsed as
2854 hexadecimal numbers. They were incorrectly parsed as decimal
2855 previously, or ignored if the name was not a valid decimal
2858 - PCI onboard indices up to 65535 are allowed. Previously, numbers
2859 above 16383 were rejected. This primarily impacts s390 systems,
2860 where values up to 65535 are used.
2862 - Invalid characters in interface names are replaced with "_".
2864 The new version of the net naming scheme is "v249". The previous
2865 scheme can be selected via the "net.naming-scheme=v247" kernel
2866 command line parameter.
2868 * sd-bus' sd_bus_is_ready() and sd_bus_is_open() calls now accept a
2869 NULL bus object, for which they will return false. Or in other words,
2870 an unallocated bus connection is neither ready nor open.
2872 * The sd-device API acquired a new API function
2873 sd_device_get_usec_initialized() that returns the monotonic time when
2874 the udev device first appeared in the database.
2876 * sd-device gained a new APIs sd_device_trigger_with_uuid() and
2877 sd_device_get_trigger_uuid(). The former is similar to
2878 sd_device_trigger() but returns a randomly generated UUID that is
2879 associated with the synthetic uevent generated by the call. This UUID
2880 may be read from the sd_device object a monitor eventually receives,
2881 via the sd_device_get_trigger_uuid(). This interface requires kernel
2882 4.13 or above to work, and allows tracking a synthetic uevent through
2883 the entire device management stack. The "udevadm trigger --settle"
2884 logic has been updated to make use of this concept if available to
2885 wait precisely for the uevents it generates. "udevadm trigger" also
2886 gained a new parameter --uuid that prints the UUID for each generated
2889 * sd-device also gained new APIs sd_device_new_from_ifname() and
2890 sd_device_new_from_ifindex() for allocating an sd-device object for
2891 the specified network interface. The former accepts an interface name
2892 (either a primary or an alternative name), the latter an interface
2895 * The native Journal protocol has been documented. Clients may talk
2896 this as alternative to the classic BSD syslog protocol for locally
2897 delivering log records to the Journal. The protocol has been stable
2898 for a long time and in fact been implemented already in a variety
2899 of alternative client libraries. This documentation makes the support
2902 https://systemd.io/JOURNAL_NATIVE_PROTOCOL
2904 * A new BPFProgram= setting has been added to service files. It may be
2905 set to a path to a loaded kernel BPF program, i.e. a path to a bpffs
2906 file, or a bind mount or symlink to one. This may be used to upload
2907 and manage BPF programs externally and then hook arbitrary systemd
2910 * The "home.arpa" domain that has been officially declared as the
2911 choice for domain for local home networks per RFC 8375 has been added
2912 to the default NTA list of resolved, since DNSSEC is generally not
2913 available on private domains.
2915 * The CPUAffinity= setting of unit files now resolves "%" specifiers.
2917 * A new ManageForeignRoutingPolicyRules= setting has been added to
2918 .network files which may be used to exclude foreign-created routing
2919 policy rules from systemd-networkd management.
2921 * systemd-network-wait-online gained two new switches -4 and -6 that
2922 may be used to tweak whether to wait for only IPv4 or only IPv6
2925 * .network files gained a new RequiredFamilyForOnline= setting to
2926 fine-tune whether to require an IPv4 or IPv6 address in order to
2927 consider an interface "online".
2929 * networkctl will now show an over-all "online" state in the per-link
2932 * In .network files a new OutgoingInterface= setting has been added to
2933 specify the output interface in bridge FDB setups.
2935 * In .network files the Multipath group ID may now be configured for
2936 [NextHop] entries, via the new Group= setting.
2938 * The DHCP server logic configured in .network files gained a new
2939 setting RelayTarget= that turns the server into a DHCP server relay.
2940 The RelayAgentCircuitId= and RelayAgentRemoteId= settings may be used
2941 to further tweak the DHCP relay behaviour.
2943 * The DHCP server logic also gained a new ServerAddress= setting in
2944 .network files that explicitly specifies the server IP address to
2945 use. If not specified, the address is determined automatically, as
2948 * The DHCP server logic in systemd-networkd gained support for static
2949 DHCP leases, configurable via the [DHCPServerStaticLease]
2950 section. This allows explicitly mapping specific MAC addresses to
2951 fixed IP addresses and vice versa.
2953 * The RestrictAddressFamilies= setting in service files now supports a
2954 new special value "none". If specified sockets of all address
2955 families will be made unavailable to services configured that way.
2957 * systemd-fstab-generator and systemd-repart have been updated to
2958 support booting from disks that carry only a /usr/ partition but no
2959 root partition yet, and where systemd-repart can add it in on the
2960 first boot. This is useful for implementing systems that ship with a
2961 single /usr/ file system, and whose root file system shall be set up
2962 and formatted on a LUKS-encrypted volume whose key is generated
2963 locally (and possibly enrolled in the TPM) during the first boot.
2965 * The [Address] section of .network files now accepts a new
2966 RouteMetric= setting that configures the routing metric to use for
2967 the prefix route created as effect of the address configuration.
2968 Similarly, the [DHCPv6PrefixDelegation] and [IPv6Prefix] sections
2969 gained matching settings for their prefix routes. (The option of the
2970 same name in the [DHCPv6] section is moved to [IPv6AcceptRA], since
2971 it conceptually belongs there; the old option is still understood for
2974 * The DHCPv6 IAID and DUID are now explicitly configurable in .network
2977 * A new udev property ID_NET_DHCP_BROADCAST on network interface
2978 devices is now honoured by systemd-networkd, controlling whether to
2979 issue DHCP offers via broadcasting. This is used to ensure that s390
2980 layer 3 network interfaces work out-of-the-box with systemd-networkd.
2982 * nss-myhostname and systemd-resolved will now synthesize address
2983 records for a new special hostname "_outbound". The name will always
2984 resolve to the local IP addresses most likely used for outbound
2985 connections towards the default routes. On multi-homed hosts this is
2986 useful to have a stable handle referring to "the" local IP address
2987 that matters most, to the point where this is defined.
2989 * The Discoverable Partition Specification has been updated with a new
2990 GPT partition flag "grow-file-system" defined for its partition
2991 types. Whenever partitions with this flag set are automatically
2992 mounted (i.e. via systemd-gpt-auto-generator or the --image= switch
2993 of systemd-nspawn or other tools; and as opposed to explicit mounting
2994 via /etc/fstab), the file system within the partition is
2995 automatically grown to the full size of the partition. If the file
2996 system size already matches the partition size this flag has no
2997 effect. Previously, this functionality has been available via the
2998 explicit x-systemd.growfs mount option, and this new flag extends
2999 this to automatically discovered mounts. A new GrowFileSystem=
3000 setting has been added to systemd-repart drop-in files that allows
3001 configuring this partition flag. This new flag defaults to on for
3002 partitions automatically created by systemd-repart, except if they
3003 are marked read-only. See the specification for further details:
3005 https://systemd.io/DISCOVERABLE_PARTITIONS
3007 * .network files gained a new setting RoutesToNTP= in the [DHCPv4]
3008 section. If enabled (which is the default), and an NTP server address
3009 is acquired through a DHCP lease on this interface an explicit route
3010 to this address is created on this interface to ensure that NTP
3011 traffic to the NTP server acquired on an interface is also routed
3012 through that interface. The pre-existing RoutesToDNS= setting that
3013 implements the same for DNS servers is now enabled by default.
3015 * A pair of service settings SocketBindAllow= + SocketBindDeny= have
3016 been added that may be used to restrict the network interfaces
3017 sockets created by the service may be bound to. This is implemented
3020 * A new ConditionFirmware= setting has been added to unit files to
3021 conditionalize on certain firmware features. At the moment it may
3022 check whether running on an UEFI system, a device.tree system, or if
3023 the system is compatible with some specified device-tree feature.
3025 * A new ConditionOSRelease= setting has been added to unit files to
3026 check os-release(5) fields. The "=", "!=", "<", "<=", ">=", ">"
3027 operators may be used to check if some field has some specific value
3028 or do an alphanumerical comparison. Equality comparisons are useful
3029 for fields like ID, but relative comparisons for fields like
3030 VERSION_ID or IMAGE_VERSION.
3032 * hostnamed gained a new Describe() D-Bus method that returns a JSON
3033 serialization of the host data it exposes. This is exposed via
3034 "hostnamectl --json=" to acquire a host identity description in JSON.
3035 It's our intention to add a similar features to most services and
3036 objects systemd manages, in order to simplify integration with
3037 program code that can consume JSON.
3039 * Similarly, networkd gained a Describe() method on its Manager and
3040 Link bus objects. This is exposed via "networkctl --json=".
3042 * hostnamectl's various "get-xyz"/"set-xyz" verb pairs
3043 (e.g. "hostnamectl get-hostname", "hostnamectl "set-hostname") have
3044 been replaced by a single "xyz" verb (e.g. "hostnamectl hostname")
3045 that is used both to get the value (when no argument is given), and
3046 to set the value (when an argument is specified). The old names
3047 continue to be supported for compatibility.
3049 * systemd-detect-virt and ConditionVirtualization= are now able to
3050 correctly identify Amazon EC2 environments.
3052 * The LogLevelMax= setting of unit files now applies not only to log
3053 messages generated *by* the service, but also to log messages
3054 generated *about* the service by PID 1. To suppress logs concerning a
3055 specific service comprehensively, set this option to a high log
3058 * bootctl gained support for a new --make-machine-id-directory= switch
3059 that allows precise control on whether to create the top-level
3060 per-machine directory in the boot partition that typically contains
3061 Type 1 boot loader entries.
3063 * During build SBAT data to include in the systemd-boot EFI PE binaries
3064 may be specified now.
3066 * /etc/crypttab learnt a new option "headless". If specified any
3067 requests to query the user interactively for passwords or PINs will
3068 be skipped. This is useful on systems that are headless, i.e. where
3069 an interactive user is generally not present.
3071 * /etc/crypttab also learnt a new option "password-echo=" that allows
3072 configuring whether the encryption password prompt shall echo the
3073 typed password and if so, do so literally or via asterisks. (The
3074 default is the same behaviour as before: provide echo feedback via
3077 * FIDO2 support in systemd-cryptenroll/systemd-cryptsetup and
3078 systemd-homed has been updated to allow explicit configuration of the
3079 "user presence" and "user verification" checks, as well as whether a
3080 PIN is required for authentication, via the new switches
3081 --fido2-with-user-presence=, --fido2-with-user-verification=,
3082 --fido2-with-client-pin= to systemd-cryptenroll and homectl. Which
3083 features are available, and may be enabled or disabled depends on the
3086 * systemd-nspawn's --private-user= switch now accepts the special value
3087 "identity" which configures a user namespacing environment with an
3088 identity mapping of 65535 UIDs. This means the container UID 0 is
3089 mapped to the host UID 0, and the UID 1 to host UID 1. On first look
3090 this doesn't appear to be useful, however it does reduce the attack
3091 surface a bit, since the resulting container will possess process
3092 capabilities only within its namespace and not on the host.
3094 * systemd-nspawn's --private-user-chown switch has been replaced by a
3095 more generic --private-user-ownership= switch that accepts one of
3096 three values: "chown" is equivalent to the old --private-user-chown,
3097 and "off" is equivalent to the absence of the old switch. The value
3098 "map" uses the new UID mapping mounts of Linux 5.12 to map ownership
3099 of files and directories of the underlying image to the chosen UID
3100 range for the container. "auto" is equivalent to "map" if UID mapping
3101 mount are supported, otherwise it is equivalent to "chown". The short
3102 -U switch systemd-nspawn now implies --private-user-ownership=auto
3103 instead of the old --private-user-chown. Effectively this means: if
3104 the backing file system supports UID mapping mounts the feature is
3105 now used by default if -U is used. Generally, it's a good idea to use
3106 UID mapping mounts instead of recursive chown()ing, since it allows
3107 running containers off immutable images (since no modifications of
3108 the images need to take place), and share images between multiple
3109 instances. Moreover, the recursive chown()ing operation is slow and
3110 can be avoided. Conceptually it's also a good thing if transient UID
3111 range uses do not leak into persistent file ownership anymore. TLDR:
3112 finally, the last major drawback of user namespacing has been
3113 removed, and -U should always be used (unless you use btrfs, where
3114 UID mapped mounts do not exist; or your container actually needs
3115 privileges on the host).
3117 * nss-systemd now synthesizes user and group shadow records in addition
3118 to the main user and group records. Thus, hashed passwords managed by
3119 systemd-homed are now accessible via the shadow database.
3121 * The userdb logic (and thus nss-systemd, and so on) now read
3122 additional user/group definitions in JSON format from the drop-in
3123 directories /etc/userdb/, /run/userdb/, /run/host/userdb/ and
3124 /usr/lib/userdb/. This is a simple and powerful mechanism for making
3125 additional users available to the system, with full integration into
3126 NSS including the shadow databases. Since the full JSON user/group
3127 record format is supported this may also be used to define users with
3128 resource management settings and other runtime settings that
3129 pam_systemd and systemd-logind enforce at login.
3131 * The userdbctl tool gained two new switches --with-dropin= and
3132 --with-varlink= which can be used to fine-tune the sources used for
3133 user database lookups.
3135 * systemd-nspawn gained a new switch --bind-user= for binding a host
3136 user account into the container. This does three things: the user's
3137 home directory is bind mounted from the host into the container,
3138 below the /run/userdb/home/ hierarchy. A free UID is picked in the
3139 container, and a user namespacing UID mapping to the host user's UID
3140 installed. And finally, a minimal JSON user and group record (along
3141 with its hashed password) is dropped into /run/host/userdb/. These
3142 records are picked up automatically by the userdb drop-in logic
3143 describe above, and allow the user to login with the same password as
3144 on the host. Effectively this means: if host and container run new
3145 enough systemd versions making a host user available to the container
3146 is trivially simple.
3148 * systemd-journal-gatewayd now supports the switches --user, --system,
3149 --merge, --file= that are equivalent to the same switches of
3150 journalctl, and permit exposing only the specified subset of the
3153 * The OnFailure= dependency between units is now augmented with a
3154 implicit reverse dependency OnFailureOf= (this new dependency cannot
3155 be configured directly it's only created as effect of an OnFailure=
3156 dependency in the reverse order — it's visible in "systemctl show"
3157 however). Similar, Slice= now has an reverse dependency SliceOf=,
3158 that is also not configurable directly, but useful to determine all
3159 units that are members of a slice.
3161 * A pair of new dependency types between units PropagatesStopTo= +
3162 StopPropagatedFrom= has been added, that allows propagation of unit
3163 stop events between two units. It operates similar to the existing
3164 PropagatesReloadTo= + ReloadPropagatedFrom= dependencies.
3166 * A new dependency type OnSuccess= has been added (plus the reverse
3167 dependency OnSuccessOf=, which cannot be configured directly, but
3168 exists only as effect of the reverse OnSuccess=). It is similar to
3169 OnFailure=, but triggers in the opposite case: when a service exits
3170 cleanly. This allows "chaining up" of services where one or more
3171 services are started once another service has successfully completed.
3173 * A new dependency type Upholds= has been added (plus the reverse
3174 dependency UpheldBy=, which cannot be configured directly, but exists
3175 only as effect of Upholds=). This dependency type is a stronger form
3176 of Wants=: if a unit has an UpHolds= dependency on some other unit
3177 and the former is active then the latter is started whenever it is
3178 found inactive (and no job is queued for it). This is an alternative
3179 to Restart= inside service units, but less configurable, and the
3180 request to uphold a unit is not encoded in the unit itself but in
3181 another unit that intends to uphold it.
3183 * The systemd-ask-password tool now also supports reading passwords
3184 from the credentials subsystem, via the new --credential= switch.
3186 * The systemd-ask-password tool learnt a new switch --emoji= which may
3187 be used to explicit control whether the lock and key emoji (🔐) is
3188 shown in the password prompt on suitable TTYs.
3190 * The --echo switch of systemd-ask-password now optionally takes a
3191 parameter that controls character echo. It may either show asterisks
3192 (default, as before), turn echo off entirely, or echo the typed
3193 characters literally.
3195 * The systemd-ask-password tool also gained a new -n switch for
3196 suppressing output of a trailing newline character when writing the
3197 acquired password to standard output, similar to /bin/echo's -n
3200 * New documentation has been added that describes the organization of
3201 the systemd source code tree:
3203 https://systemd.io/ARCHITECTURE
3205 * Units using ConditionNeedsUpdate= will no longer be activated in
3208 * It is now possible to list a template unit in the WantedBy= or
3209 RequiredBy= settings of the [Install] section of another template
3210 unit, which will be instantiated using the same instance name.
3212 * A new MemoryAvailable property is available for units. If the unit,
3213 or the slices it is part of, have a memory limit set via MemoryMax=/
3214 MemoryHigh=, MemoryAvailable will indicate how much more memory the
3215 unit can claim before hitting the limits.
3217 * systemd-coredump will now try to stay below the cgroup memory limit
3218 placed on itself or one of the slices it runs under, if the storage
3219 area for core files (/var/lib/systemd/coredump/) is placed on a tmpfs,
3220 since files written on such filesystems count toward the cgroup memory
3221 limit. If there is not enough available memory in such cases to store
3222 the core file uncompressed, systemd-coredump will skip to compressed
3223 storage directly (if enabled) and it will avoid analyzing the core file
3224 to print backtrace and metadata in the journal.
3226 * tmpfiles.d/ drop-ins gained a new '=' modifier to check if the type
3227 of a path matches the configured expectations, and remove it if not.
3229 * tmpfiles.d/'s 'Age' now accepts an 'age-by' argument, which allows to
3230 specify which of the several available filesystem timestamps (access
3231 time, birth time, change time, modification time) to look at when
3232 deciding whether a path has aged enough to be cleaned.
3234 * A new IPv6StableSecretAddress= setting has been added to .network
3235 files, which takes an IPv6 address to use as secret for IPv6 address
3238 * The [DHCPServer] logic in .network files gained support for a new
3239 UplinkInterface= setting that permits configuration of the uplink
3240 interface name to propagate DHCP lease information from.
3242 * The WakeOnLan= setting in .link files now accepts a list of flags
3243 instead of a single one, to configure multiple wake-on-LAN policies.
3245 * User-space defined tracepoints (USDT) have been added to udev at
3246 strategic locations. This is useful for tracing udev behaviour and
3247 performance with bpftrace and similar tools.
3249 * systemd-journald-upload gained a new NetworkTimeoutSec= option for
3250 setting a network timeout time.
3252 * If a system service is running in a new mount namespace (RootDirectory=
3253 and friends), all file systems will be mounted with MS_NOSUID by
3254 default, unless the system is running with SELinux enabled.
3256 * When enumerating time zones the timedatectl tool will now consult the
3257 'tzdata.zi' file shipped by the IANA time zone database package, in
3258 addition to 'zone1970.tab', as before. This makes sure time zone
3259 aliases are now correctly supported. Some distributions so far did
3260 not install this additional file, most do however. If you
3261 distribution does not install it yet, it might make sense to change
3264 * Intel HID rfkill event is no longer masked, since it's the only
3265 source of rfkill event on newer HP laptops. To have both backward and
3266 forward compatibility, userspace daemon needs to debounce duplicated
3267 events in a short time window.
3269 Contributions from: Aakash Singh, adrian5, Albert Brox,
3270 Alexander Sverdlin, Alexander Tsoy, Alexey Rubtsov, alexlzhu,
3271 Allen Webb, Alvin Šipraga, Alyssa Ross, Anders Wenhaug,
3272 Andrea Pappacoda, Anita Zhang, asavah, Balint Reczey, Bertrand Jacquin,
3273 borna-blazevic, caoxia2008cxx, Carlo Teubner, Christian Göttsche,
3274 Christian Hesse, Daniel Schaefer, Dan Streetman,
3275 David Santamaría Rogado, David Tardon, Deepak Rawat, dgcampea,
3276 Dimitri John Ledkov, ei-ke, Emilio Herrera, Emil Renner Berthing,
3277 Eric Cook, Flos Lonicerae, Franck Bui, Francois Gervais,
3278 Frantisek Sumsal, Gibeom Gwon, gitm0, Hamish Moffatt, Hans de Goede,
3279 Harsh Barsaiyan, Henri Chain, Hristo Venev, Icenowy Zheng, Igor Zhbanov,
3280 imayoda, Jakub Warczarek, James Buren, Jan Janssen, Jan Macku,
3281 Jan Synacek, Jason Francis, Jayanth Ananthapadmanaban, Jeremy Szu,
3282 Jérôme Carretero, Jesse Stricker, jiangchuangang, Joerg Behrmann,
3283 Jóhann B. Guðmundsson, Jörg Deckert, Jörg Thalheim, Juergen Hoetzel,
3284 Julia Kartseva, Kai-Heng Feng, Khem Raj, KoyamaSohei, laineantti,
3285 Lennart Poettering, LetzteInstanz, Luca Adrian L, Luca Boccassi,
3286 Lucas Magasweran, Mantas Mikulėnas, Marco Antonio Mauro, Mark Wielaard,
3287 Masahiro Matsuya, Matt Johnston, Michael Catanzaro, Michal Koutný,
3288 Michal Sekletár, Mike Crowe, Mike Kazantsev, Milan, milaq,
3289 Miroslav Suchý, Morten Linderud, nerdopolis, nl6720, Noah Meyerhans,
3290 Oleg Popov, Olle Lundberg, Ondrej Kozina, Paweł Marciniak, Perry.Yuan,
3291 Peter Hutterer, Peter Kjellerstedt, Peter Morrow, Phaedrus Leeds,
3292 plattrap, qhill, Raul Tambre, Roman Beranek, Roshan Shariff,
3293 Ryan Hendrickson, Samuel BF, scootergrisen, Sebastian Blunt,
3294 Seong-ho Cho, Sergey Bugaev, Sevan Janiyan, Sibo Dong, simmon,
3295 Simon Watts, Srinidhi Kaushik, Štěpán Němec, Steve Bonds, Susant Sahani,
3296 sverdlin, syyhao1994, Takashi Sakamoto, Topi Miettinen, tramsay,
3297 Trent Piepho, Uwe Kleine-König, Viktor Mihajlovski, Vincent Dechenaux,
3298 Vito Caputo, William A. Kennington III, Yangyang Shen, Yegor Alexeyev,
3299 Yi Gao, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, zsien, наб
3301 — Edinburgh, 2021-07-07
3305 * A concept of system extension images is introduced. Such images may
3306 be used to extend the /usr/ and /opt/ directory hierarchies at
3307 runtime with additional files (even if the file system is read-only).
3308 When a system extension image is activated, its /usr/ and /opt/
3309 hierarchies and os-release information are combined via overlayfs
3310 with the file system hierarchy of the host OS.
3312 A new systemd-sysext tool can be used to merge, unmerge, list, and
3313 refresh system extension hierarchies. See
3314 https://www.freedesktop.org/software/systemd/man/systemd-sysext.html.
3316 The systemd-sysext.service automatically merges installed system
3317 extensions during boot (before basic.target, but not in very early
3318 boot, since various file systems have to be mounted first).
3320 The SYSEXT_LEVEL= field in os-release(5) may be used to specify the
3321 supported system extension level.
3323 * A new ExtensionImages= unit setting can be used to apply the same
3324 system extension image concept from systemd-sysext to the namespaced
3325 file hierarchy of specific services, following the same rules and
3328 * Support for a new special "root=tmpfs" kernel command-line option has
3329 been added. When specified, a tmpfs is mounted on /, and mount.usr=
3330 should be used to point to the operating system implementation.
3332 * A new configuration file /etc/veritytab may be used to configure
3333 dm-verity integrity protection for block devices. Each line is in the
3334 format "volume-name data-device hash-device roothash options",
3335 similar to /etc/crypttab.
3337 * A new kernel command-line option systemd.verity.root_options= may be
3338 used to configure dm-verity behaviour for the root device.
3340 * The key file specified in /etc/crypttab (the third field) may now
3341 refer to an AF_UNIX/SOCK_STREAM socket in the file system. The key is
3342 acquired by connecting to that socket and reading from it. This
3343 allows the implementation of a service to provide key information
3344 dynamically, at the moment when it is needed.
3346 * When the hostname is set explicitly to "localhost", systemd-hostnamed
3347 will respect this. Previously such a setting would be mostly silently
3348 ignored. The goal is to honour configuration as specified by the
3351 * The fallback hostname that will be used by the system manager and
3352 systemd-hostnamed can now be configured in two new ways: by setting
3353 DEFAULT_HOSTNAME= in os-release(5), or by setting
3354 $SYSTEMD_DEFAULT_HOSTNAME in the environment block. As before, it can
3355 also be configured during compilation. The environment variable is
3356 intended for testing and local overrides, the os-release(5) field is
3357 intended to allow customization by different variants of a
3358 distribution that share the same compiled packages.
3360 * The environment block of the manager itself may be configured through
3361 a new ManagerEnvironment= setting in system.conf or user.conf. This
3362 complements existing ways to set the environment block (the kernel
3363 command line for the system manager, the inherited environment and
3364 user@.service unit file settings for the user manager).
3366 * systemd-hostnamed now exports the default hostname and the source of
3367 the configured hostname ("static", "transient", or "default") as
3370 * systemd-hostnamed now exports the "HardwareVendor" and
3371 "HardwareModel" D-Bus properties, which are supposed to contain a
3372 pair of cleaned up, human readable strings describing the system's
3373 vendor and model. It's typically sourced from the firmware's DMI
3374 tables, but may be augmented from a new hwdb database. hostnamectl
3375 shows this in the status output.
3377 * Support has been added to systemd-cryptsetup for extracting the
3378 PKCS#11 token URI and encrypted key from the LUKS2 JSON embedded
3379 metadata header. This allows the information how to open the
3380 encrypted device to be embedded directly in the device and obviates
3381 the need for configuration in an external file.
3383 * systemd-cryptsetup gained support for unlocking LUKS2 volumes using
3384 TPM2 hardware, as well as FIDO2 security tokens (in addition to the
3385 pre-existing support for PKCS#11 security tokens).
3387 * systemd-repart may enroll encrypted partitions using TPM2
3388 hardware. This may be useful for example to create an encrypted /var
3389 partition bound to the machine on first boot.
3391 * A new systemd-cryptenroll tool has been added to enroll TPM2, FIDO2
3392 and PKCS#11 security tokens to LUKS volumes, list and destroy
3395 http://0pointer.net/blog/unlocking-luks2-volumes-with-tpm2-fido2-pkcs11-security-hardware-on-systemd-248.html
3397 It also supports enrolling "recovery keys" and regular passphrases.
3399 * The libfido2 dependency is now based on dlopen(), so that the library
3400 is used at runtime when installed, but is not a hard runtime
3403 * systemd-cryptsetup gained support for two new options in
3404 /etc/crypttab: "no-write-workqueue" and "no-read-workqueue" which
3405 request synchronous processing of encryption/decryption IO.
3407 * The manager may be configured at compile time to use the fexecve()
3408 instead of the execve() system call when spawning processes. Using
3409 fexecve() closes a window between checking the security context of an
3410 executable and spawning it, but unfortunately the kernel displays
3411 stale information in the process' "comm" field, which impacts ps
3414 * The configuration option -Dcompat-gateway-hostname has been dropped.
3415 "_gateway" is now the only supported name.
3417 * The ConditionSecurity=tpm2 unit file setting may be used to check if
3418 the system has at least one TPM2 (tpmrm class) device.
3420 * A new ConditionCPUFeature= has been added that may be used to
3421 conditionalize units based on CPU features. For example,
3422 ConditionCPUFeature=rdrand will condition a unit so that it is only
3423 run when the system CPU supports the RDRAND opcode.
3425 * The existing ConditionControlGroupController= setting has been
3426 extended with two new values "v1" and "v2". "v2" means that the
3427 unified v2 cgroup hierarchy is used, and "v1" means that legacy v1
3428 hierarchy or the hybrid hierarchy are used.
3430 * A new PrivateIPC= setting on a unit file allows executed processes to
3431 be moved into a private IPC namespace, with separate System V IPC
3432 identifiers and POSIX message queues.
3434 A new IPCNamespacePath= allows the unit to be joined to an existing
3437 * The tables of system calls in seccomp filters are now automatically
3438 generated from kernel lists exported on
3439 https://fedora.juszkiewicz.com.pl/syscalls.html.
3441 The following architectures should now have complete lists:
3442 alpha, arc, arm64, arm, i386, ia64, m68k, mips64n32, mips64, mipso32,
3443 powerpc, powerpc64, s390, s390x, tilegx, sparc, x86_64, x32.
3445 * The MountAPIVFS= service file setting now additionally mounts a tmpfs
3446 on /run/ if it is not already a mount point. A writable /run/ has
3447 always been a requirement for a functioning system, but this was not
3448 guaranteed when using a read-only image.
3450 Users can always specify BindPaths= or InaccessiblePaths= as
3451 overrides, and they will take precedence. If the host's root mount
3452 point is used, there is no change in behaviour.
3454 * New bind mounts and file system image mounts may be injected into the
3455 mount namespace of a service (without restarting it). This is exposed
3456 respectively as 'systemctl bind <unit> <path>…' and
3457 'systemctl mount-image <unit> <image>…'.
3459 * The StandardOutput= and StandardError= settings can now specify files
3460 to be truncated for output (as "truncate:<path>").
3462 * The ExecPaths= and NoExecPaths= settings may be used to specify
3463 noexec for parts of the file system.
3465 * sd-bus has a new function sd_bus_open_user_machine() to open a
3466 connection to the session bus of a specific user in a local container
3467 or on the local host. This is exposed in the existing -M switch to
3468 systemctl and similar tools:
3470 systemctl --user -M lennart@foobar start foo
3472 This will connect to the user bus of a user "lennart" in container
3473 "foobar". If no container name is specified, the specified user on
3474 the host itself is connected to
3476 systemctl --user -M lennart@ start quux
3478 * sd-bus also gained a convenience function sd_bus_message_send() to
3479 simplify invocations of sd_bus_send(), taking only a single
3480 parameter: the message to send.
3482 * sd-event allows rate limits to be set on event sources, for dealing
3483 with high-priority event sources that might starve out others. See
3484 the new man page sd_event_source_set_ratelimit(3) for details.
3486 * systemd.link files gained a [Link] Promiscuous= switch, which allows
3487 the device to be raised in promiscuous mode.
3489 New [Link] TransmitQueues= and ReceiveQueues= settings allow the
3490 number of TX and RX queues to be configured.
3492 New [Link] TransmitQueueLength= setting allows the size of the TX
3493 queue to be configured.
3495 New [Link] GenericSegmentOffloadMaxBytes= and
3496 GenericSegmentOffloadMaxSegments= allow capping the packet size and
3497 the number of segments accepted in Generic Segment Offload.
3499 * systemd-networkd gained support for the "B.A.T.M.A.N. advanced"
3500 wireless routing protocol that operates on ISO/OSI Layer 2 only and
3501 uses ethernet frames to route/bridge packets. This encompasses a new
3502 "batadv" netdev Type=, a new [BatmanAdvanced] section with a bunch of
3503 new settings in .netdev files, and a new BatmanAdvanced= setting in
3506 * systemd.network files gained a [Network] RouteTable= configuration
3507 switch to select the routing policy table.
3509 systemd.network files gained a [RoutingPolicyRule] Type=
3510 configuration switch (one of "blackhole, "unreachable", "prohibit").
3512 systemd.network files gained a [IPv6AcceptRA] RouteDenyList= and
3513 RouteAllowList= settings to ignore/accept route advertisements from
3514 routers matching specified prefixes. The DenyList= setting has been
3515 renamed to PrefixDenyList= and a new PrefixAllowList= option has been
3518 systemd.network files gained a [DHCPv6] UseAddress= setting to
3519 optionally ignore the address provided in the lease.
3521 systemd.network files gained a [DHCPv6PrefixDelegation]
3522 ManageTemporaryAddress= switch.
3524 systemd.network files gained a new ActivationPolicy= setting which
3525 allows configuring how the UP state of an interface shall be managed,
3526 i.e. whether the interface is always upped, always downed, or may be
3527 upped/downed by the user using "ip link set dev".
3529 * The default for the Broadcast= setting in .network files has slightly
3530 changed: the broadcast address will not be configured for wireguard
3533 * systemd.netdev files gained a [VLAN] Protocol=, IngressQOSMaps=,
3534 EgressQOSMaps=, and [MACVLAN] BroadcastMulticastQueueLength=
3535 configuration options for VLAN packet handling.
3537 * udev rules may now set log_level= option. This allows debug logs to
3538 be enabled for select events, e.g. just for a specific subsystem or
3539 even a single device.
3541 * udev now exports the VOLUME_ID, LOGICAL_VOLUME_ID, VOLUME_SET_ID, and
3542 DATA_PREPARED_ID properties for block devices with ISO9660 file
3545 * udev now exports decoded DMI information about installed memory slots
3546 as device properties under the /sys/class/dmi/id/ pseudo device.
3548 * /dev/ is not mounted noexec anymore. This didn't provide any
3549 significant security benefits and would conflict with the executable
3550 mappings used with /dev/sgx device nodes. The previous behaviour can
3551 be restored for individual services with NoExecPaths=/dev (or by allow-
3552 listing and excluding /dev from ExecPaths=).
3554 * Permissions for /dev/vsock are now set to 0o666, and /dev/vhost-vsock
3555 and /dev/vhost-net are owned by the kvm group.
3557 * The hardware database has been extended with a list of fingerprint
3558 readers that correctly support USB auto-suspend using data from
3561 * systemd-resolved can now answer DNSSEC questions through the stub
3562 resolver interface in a way that allows local clients to do DNSSEC
3563 validation themselves. For a question with DO+CD set, it'll proxy the
3564 DNS query and respond with a mostly unmodified packet received from
3565 the upstream server.
3567 * systemd-resolved learnt a new boolean option CacheFromLocalhost= in
3568 resolved.conf. If true the service will provide caching even for DNS
3569 lookups made to an upstream DNS server on the 127.0.0.1/::1
3570 addresses. By default (and when the option is false) systemd-resolved
3571 will not cache such lookups, in order to avoid duplicate local
3572 caching, under the assumption the local upstream server caches
3575 * systemd-resolved now implements RFC5001 NSID in its local DNS
3576 stub. This may be used by local clients to determine whether they are
3577 talking to the DNS resolver stub or a different DNS server.
3579 * When resolving host names and other records resolvectl will now
3580 report where the data was acquired from (i.e. the local cache, the
3581 network, locally synthesized, …) and whether the network traffic it
3582 effected was encrypted or not. Moreover the tool acquired a number of
3583 new options --cache=, --synthesize=, --network=, --zone=,
3584 --trust-anchor=, --validate= that take booleans and may be used to
3585 tweak a lookup, i.e. whether it may be answered from cached
3586 information, locally synthesized information, information acquired
3587 through the network, the local mDNS/LLMNR zone, the DNSSEC trust
3588 anchor, and whether DNSSEC validation shall be executed for the
3591 * systemd-nspawn gained a new --ambient-capability= setting
3592 (AmbientCapability= in .nspawn files) to configure ambient
3593 capabilities passed to the container payload.
3595 * systemd-nspawn gained the ability to configure the firewall using the
3596 nftables subsystem (in addition to the existing iptables
3597 support). Similarly, systemd-networkd's IPMasquerade= option now
3598 supports nftables as back-end, too. In both cases NAT on IPv6 is now
3599 supported too, in addition to IPv4 (the iptables back-end still is
3602 "IPMasquerade=yes", which was the same as "IPMasquerade=ipv4" before,
3603 retains its meaning, but has been deprecated. Please switch to either
3604 "ivp4" or "both" (if covering IPv6 is desired).
3606 * systemd-importd will now download .verity and .roothash.p7s files
3607 along with the machine image (as exposed via machinectl pull-raw).
3609 * systemd-oomd now gained a new DefaultMemoryPressureDurationSec=
3610 setting to configure the time a unit's cgroup needs to exceed memory
3611 pressure limits before action will be taken, and a new
3612 ManagedOOMPreference=none|avoid|omit setting to avoid killing certain
3615 systemd-oomd is now considered fully supported (the usual
3616 backwards-compatibility promises apply). Swap is not required for
3617 operation, but it is still recommended.
3619 * systemd-timesyncd gained a new ConnectionRetrySec= setting which
3620 configures the retry delay when trying to contact servers.
3622 * systemd-stdio-bridge gained --system/--user options to connect to the
3623 system bus (previous default) or the user session bus.
3625 * systemd-localed may now call locale-gen to generate missing locales
3626 on-demand (UTF-8-only). This improves integration with Debian-based
3627 distributions (Debian/Ubuntu/PureOS/Tanglu/...) and Arch Linux.
3629 * systemctl --check-inhibitors=true may now be used to obey inhibitors
3630 even when invoked non-interactively. The old --ignore-inhibitors
3631 switch is now deprecated and replaced by --check-inhibitors=false.
3633 * systemctl import-environment will now emit a warning when called
3634 without any arguments (i.e. to import the full environment block of
3635 the called program). This command will usually be invoked from a
3636 shell, which means that it'll inherit a bunch of variables which are
3637 specific to that shell, and usually to the TTY the shell is connected
3638 to, and don't have any meaning in the global context of the system or
3639 user service manager. Instead, only specific variables should be
3640 imported into the manager environment block.
3642 Similarly, programs which update the manager environment block by
3643 directly calling the D-Bus API of the manager, should also push
3644 specific variables, and not the full inherited environment.
3646 * systemctl's status output now shows unit state with a more careful
3647 choice of Unicode characters: units in maintenance show a "○" symbol
3648 instead of the usual "●", failed units show "×", and services being
3651 * coredumpctl gained a --debugger-arguments= switch to pass arguments
3652 to the debugger. It also gained support for showing coredump info in
3653 a simple JSON format.
3655 * systemctl/loginctl/machinectl's --signal= option now accept a special
3656 value "list", which may be used to show a brief table with known
3657 process signals and their numbers.
3659 * networkctl now shows the link activation policy in status.
3661 * Various tools gained --pager/--no-pager/--json= switches to
3662 enable/disable the pager and provide JSON output.
3664 * Various tools now accept two new values for the SYSTEMD_COLORS
3665 environment variable: "16" and "256", to configure how many terminal
3666 colors are used in output.
3668 * less 568 or newer is now required for the auto-paging logic of the
3669 various tools. Hyperlink ANSI sequences in terminal output are now
3670 used even if a pager is used, and older versions of less are not able
3671 to display these sequences correctly. SYSTEMD_URLIFY=0 may be used to
3672 disable this output again.
3674 * Builds with support for separate / and /usr/ hierarchies ("split-usr"
3675 builds, non-merged-usr builds) are now officially deprecated. A
3676 warning is emitted during build. Support is slated to be removed in
3677 about a year (when the Debian Bookworm release development starts).
3679 * Systems with the legacy cgroup v1 hierarchy are now marked as
3680 "tainted", to make it clearer that using the legacy hierarchy is not
3683 * systemd-localed will now refuse to configure a keymap which is not
3684 installed in the file system. This is intended as a bug fix, but
3685 could break cases where systemd-localed was used to configure the
3686 keymap in advanced of it being installed. It is necessary to install
3687 the keymap file first.
3689 * The main git development branch has been renamed to 'main'.
3691 * mmcblk[0-9]boot[0-9] devices will no longer be probed automatically
3692 for partitions, as in the vast majority of cases they contain none
3693 and are used internally by the bootloader (eg: uboot).
3695 * systemd will now set the $SYSTEMD_EXEC_PID environment variable for
3696 spawned processes to the PID of the process itself. This may be used
3697 by programs for detecting whether they were forked off by the service
3698 manager itself or are a process forked off further down the tree.
3700 * The sd-device API gained four new calls: sd_device_get_action() to
3701 determine the uevent add/remove/change/… action the device object has
3702 been seen for, sd_device_get_seqno() to determine the uevent sequence
3703 number, sd_device_new_from_stat_rdev() to allocate a new sd_device
3704 object from stat(2) data of a device node, and sd_device_trigger() to
3705 write to the 'uevent' attribute of a device.
3707 * For most tools the --no-legend= switch has been replaced by
3708 --legend=no and --legend=yes, to force whether tables are shown with
3711 * Units acquired a new property "Markers" that takes a list of zero,
3712 one or two of the following strings: "needs-reload" and
3713 "needs-restart". These markers may be set via "systemctl
3714 set-property". Once a marker is set, "systemctl reload-or-restart
3715 --marked" may be invoked to execute the operation the units are
3716 marked for. This is useful for package managers that want to mark
3717 units for restart/reload while updating, but effect the actual
3718 operations at a later step at once.
3720 * The sd_bus_message_read_strv() API call of sd-bus may now also be
3721 used to parse arrays of D-Bus signatures and D-Bus paths, in addition
3724 * bootctl will now report whether the UEFI firmware used a TPM2 device
3725 and measured the boot process into it.
3727 * systemd-tmpfiles learnt support for a new environment variable
3728 $SYSTEMD_TMPFILES_FORCE_SUBVOL which takes a boolean value. If true
3729 the v/q/Q lines in tmpfiles.d/ snippets will create btrfs subvolumes
3730 even if the root fs of the system is not itself a btrfs volume.
3732 * systemd-detect-virt/ConditionVirtualization= will now explicitly
3733 detect Docker/Podman environments where possible. Moreover, they
3734 should be able to generically detect any container manager as long as
3735 it assigns the container a cgroup.
3737 * portablectl gained a new "reattach" verb for detaching/reattaching a
3738 portable service image, useful for updating images on-the-fly.
3740 * Intel SGX enclave device nodes (which expose a security feature of
3741 newer Intel CPUs) will now be owned by a new system group "sgx".
3743 Contributions from: Adam Nielsen, Adrian Vovk, AJ Jordan, Alan Perry,
3744 Alastair Pharo, Alexander Batischev, Ali Abdallah, Andrew Balmos,
3745 Anita Zhang, Annika Wickert, Ansgar Burchardt, Antonio Terceiro,
3746 Antonius Frie, Ardy, Arian van Putten, Ariel Fermani, Arnaud T,
3747 A S Alam, Bastien Nocera, Benjamin Berg, Benjamin Robin, Björn Daase,
3748 caoxia, Carlo Wood, Charles Lee, ChopperRob, chri2, Christian Ehrhardt,
3749 Christian Hesse, Christopher Obbard, clayton craft, corvusnix, cprn,
3750 Daan De Meyer, Daniele Medri, Daniel Rusek, Dan Sanders, Dan Streetman,
3751 Darren Ng, David Edmundson, David Tardon, Deepak Rawat, Devon Pringle,
3752 Dmitry Borodaenko, dropsignal, Einsler Lee, Endre Szabo,
3753 Evgeny Vereshchagin, Fabian Affolter, Fangrui Song, Felipe Borges,
3754 feliperodriguesfr, Felix Stupp, Florian Hülsmann, Florian Klink,
3755 Florian Westphal, Franck Bui, Frantisek Sumsal, Gablegritule,
3756 Gaël PORTAY, Gaurav, Giedrius Statkevičius, Greg Depoire-Ferrer,
3757 Gustavo Costa, Hans de Goede, Hela Basa, heretoenhance, hide,
3758 Iago López Galeiras, igo95862, Ilya Dmitrichenko, Jameer Pathan,
3759 Jan Tojnar, Jiehong, Jinyuan Si, Joerg Behrmann, John Slade,
3760 Jonathan G. Underwood, Jonathan McDowell, Josh Triplett, Joshua Watt,
3761 Julia Cartwright, Julien Humbert, Kairui Song, Karel Zak,
3762 Kevin Backhouse, Kevin P. Fleming, Khem Raj, Konomi, krissgjeng,
3763 l4gfcm, Lajos Veres, Lennart Poettering, Lincoln Ramsay, Luca Boccassi,
3764 Luca BRUNO, Lucas Werkmeister, Luka Kudra, Luna Jernberg,
3765 Marc-André Lureau, Martin Wilck, Matthias Klumpp, Matt Turner,
3766 Michael Gisbers, Michael Marley, Michael Trapp, Michal Fabik,
3767 Michał Kopeć, Michal Koutný, Michal Sekletár, Michele Guerini Rocco,
3768 Mike Gilbert, milovlad, moson-mo, Nick, nihilix-melix, Oğuz Ersen,
3769 Ondrej Mosnacek, pali, Pavel Hrdina, Pavel Sapezhko, Perry Yuan,
3770 Peter Hutterer, Pierre Dubouilh, Piotr Drąg, Pjotr Vertaalt,
3771 Richard Laager, RussianNeuroMancer, Sam Lunt, Sebastiaan van Stijn,
3772 Sergey Bugaev, shenyangyang4, simmon, Simonas Kazlauskas,
3773 Slimane Selyan Amiri, Stefan Agner, Steve Ramage, Susant Sahani,
3774 Sven Mueller, Tad Fisher, Takashi Iwai, Thomas Haller, Tom Shield,
3775 Topi Miettinen, Torsten Hilbrich, tpgxyz, Tyler Hicks, ulf-f,
3776 Ulrich Ölmann, Vincent Pelletier, Vinnie Magro, Vito Caputo, Vlad,
3777 walbit-de, Whired Planck, wouter bolsterlee, Xℹ Ruoyao, Yangyang Shen,
3778 Yuri Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek,
3779 Zmicer Turok, Дамјан Георгиевски
3781 — Berlin, 2021-03-30
3785 * KERNEL API INCOMPATIBILITY: Linux 4.14 introduced two new uevents
3786 "bind" and "unbind" to the Linux device model. When this kernel
3787 change was made, systemd-udevd was only minimally updated to handle
3788 and propagate these new event types. The introduction of these new
3789 uevents (which are typically generated for USB devices and devices
3790 needing a firmware upload before being functional) resulted in a
3791 number of issues which we so far didn't address. We hoped the kernel
3792 maintainers would themselves address these issues in some form, but
3793 that did not happen. To handle them properly, many (if not most) udev
3794 rules files shipped in various packages need updating, and so do many
3795 programs that monitor or enumerate devices with libudev or sd-device,
3796 or otherwise process uevents. Please note that this incompatibility
3797 is not fault of systemd or udev, but caused by an incompatible kernel
3798 change that happened back in Linux 4.14, but is becoming more and
3799 more visible as the new uevents are generated by more kernel drivers.
3801 To minimize issues resulting from this kernel change (but not avoid
3802 them entirely) starting with systemd-udevd 247 the udev "tags"
3803 concept (which is a concept for marking and filtering devices during
3804 enumeration and monitoring) has been reworked: udev tags are now
3805 "sticky", meaning that once a tag is assigned to a device it will not
3806 be removed from the device again until the device itself is removed
3807 (i.e. unplugged). This makes sure that any application monitoring
3808 devices that match a specific tag is guaranteed to both see uevents
3809 where the device starts being relevant, and those where it stops
3810 being relevant (the latter now regularly happening due to the new
3811 "unbind" uevent type). The udev tags concept is hence now a concept
3812 tied to a *device* instead of a device *event* — unlike for example
3813 udev properties whose lifecycle (as before) is generally tied to a
3814 device event, meaning that the previously determined properties are
3815 forgotten whenever a new uevent is processed.
3817 With the newly redefined udev tags concept, sometimes it's necessary
3818 to determine which tags are the ones applied by the most recent
3819 uevent/database update, in order to discern them from those
3820 originating from earlier uevents/database updates of the same
3821 device. To accommodate for this a new automatic property CURRENT_TAGS
3822 has been added that works similar to the existing TAGS property but
3823 only lists tags set by the most recent uevent/database
3824 update. Similarly, the libudev/sd-device API has been updated with
3825 new functions to enumerate these 'current' tags, in addition to the
3826 existing APIs that now enumerate the 'sticky' ones.
3828 To properly handle "bind"/"unbind" on Linux 4.14 and newer it is
3829 essential that all udev rules files and applications are updated to
3830 handle the new events. Specifically:
3832 • All rule files that currently use a header guard similar to
3833 ACTION!="add|change",GOTO="xyz_end" should be updated to use
3834 ACTION=="remove",GOTO="xyz_end" instead, so that the
3835 properties/tags they add are also applied whenever "bind" (or
3836 "unbind") is seen. (This is most important for all physical device
3837 types — those for which "bind" and "unbind" are currently
3838 generated, for all other device types this change is still
3839 recommended but not as important — but certainly prepares for
3840 future kernel uevent type additions).
3842 • Similarly, all code monitoring devices that contains an 'if' branch
3843 discerning the "add" + "change" uevent actions from all other
3844 uevents actions (i.e. considering devices only relevant after "add"
3845 or "change", and irrelevant on all other events) should be reworked
3846 to instead negatively check for "remove" only (i.e. considering
3847 devices relevant after all event types, except for "remove", which
3848 invalidates the device). Note that this also means that devices
3849 should be considered relevant on "unbind", even though conceptually
3850 this — in some form — invalidates the device. Since the precise
3851 effect of "unbind" is not generically defined, devices should be
3852 considered relevant even after "unbind", however I/O errors
3853 accessing the device should then be handled gracefully.
3855 • Any code that uses device tags for deciding whether a device is
3856 relevant or not most likely needs to be updated to use the new
3857 udev_device_has_current_tag() API (or sd_device_has_current_tag()
3858 in case sd-device is used), to check whether the tag is set at the
3859 moment an uevent is seen (as opposed to the existing
3860 udev_device_has_tag() API which checks if the tag ever existed on
3861 the device, following the API concept redefinition explained
3864 We are very sorry for this breakage and the requirement to update
3865 packages using these interfaces. We'd again like to underline that
3866 this is not caused by systemd/udev changes, but result of a kernel
3869 * UPCOMING INCOMPATIBILITY: So far most downstream distribution
3870 packages have not retriggered devices once the udev package (or any
3871 auxiliary package installing additional udev rules) is updated. We
3872 intend to work with major distributions to change this, so that
3873 "udevadm trigger -a change" is issued on such upgrades, ensuring that
3874 the updated ruleset is applied to the devices already discovered, so
3875 that (asynchronously) after the upgrade completed the udev database
3876 is consistent with the updated rule set. This means udev rules must
3877 be ready to be retriggered with a "change" action any time, and
3878 result in correct and complete udev database entries. While the
3879 majority of udev rule files known to us currently get this right,
3880 some don't. Specifically, there are udev rules files included in
3881 various packages that only set udev properties on the "add" action,
3882 but do not handle the "change" action. If a device matching those
3883 rules is retriggered with the "change" action (as is intended here)
3884 it would suddenly lose the relevant properties. This always has been
3885 problematic, but as soon as all udev devices are triggered on relevant
3886 package upgrades this will become particularly so. It is strongly
3887 recommended to fix offending rules so that they can handle a "change"
3888 action at any time, and acquire all necessary udev properties even
3889 then. Or in other words: the header guard mentioned above
3890 (ACTION=="remove",GOTO="xyz_end") is the correct approach to handle
3891 this, as it makes sure rules are rerun on "change" correctly, and
3892 accumulate the correct and complete set of udev properties. udev rule
3893 definitions that cannot handle "change" events being triggered at
3894 arbitrary times should be considered buggy.
3896 * The MountAPIVFS= service file setting now defaults to on if
3897 RootImage= and RootDirectory= are used, which means that with those
3898 two settings /proc/, /sys/ and /dev/ are automatically properly set
3899 up for services. Previous behaviour may be restored by explicitly
3900 setting MountAPIVFS=off.
3902 * Since PAM 1.2.0 (2015) configuration snippets may be placed in
3903 /usr/lib/pam.d/ in addition to /etc/pam.d/. If a file exists in the
3904 latter it takes precedence over the former, similar to how most of
3905 systemd's own configuration is handled. Given that PAM stack
3906 definitions are primarily put together by OS vendors/distributions
3907 (though possibly overridden by users), this systemd release moves its
3908 own PAM stack configuration for the "systemd-user" PAM service (i.e.
3909 for the PAM session invoked by the per-user user@.service instance)
3910 from /etc/pam.d/ to /usr/lib/pam.d/. We recommend moving all
3911 packages' vendor versions of their PAM stack definitions from
3912 /etc/pam.d/ to /usr/lib/pam.d/, but if such OS-wide migration is not
3913 desired the location to which systemd installs its PAM stack
3914 configuration may be changed via the -Dpamconfdir Meson option.
3916 * The runtime dependencies on libqrencode, libpcre2, libidn/libidn2,
3917 libpwquality and libcryptsetup have been changed to be based on
3918 dlopen(): instead of regular dynamic library dependencies declared in
3919 the binary ELF headers, these libraries are now loaded on demand
3920 only, if they are available. If the libraries cannot be found the
3921 relevant operations will fail gracefully, or a suitable fallback
3922 logic is chosen. This is supposed to be useful for general purpose
3923 distributions, as it allows minimizing the list of dependencies the
3924 systemd packages pull in, permitting building of more minimal OS
3925 images, while still making use of these "weak" dependencies should
3926 they be installed. Since many package managers automatically
3927 synthesize package dependencies from ELF shared library dependencies,
3928 some additional manual packaging work has to be done now to replace
3929 those (slightly downgraded from "required" to "recommended" or
3930 whatever is conceptually suitable for the package manager). Note that
3931 this change does not alter build-time behaviour: as before the
3932 build-time dependencies have to be installed during build, even if
3933 they now are optional during runtime.
3935 * sd-event.h gained a new call sd_event_add_time_relative() for
3936 installing timers relative to the current time. This is mostly a
3937 convenience wrapper around the pre-existing sd_event_add_time() call
3938 which installs absolute timers.
3940 * sd-event event sources may now be placed in a new "exit-on-failure"
3941 mode, which may be controlled via the new
3942 sd_event_source_get_exit_on_failure() and
3943 sd_event_source_set_exit_on_failure() functions. If enabled, any
3944 failure returned by the event source handler functions will result in
3945 exiting the event loop (unlike the default behaviour of just
3946 disabling the event source but continuing with the event loop). This
3947 feature is useful to set for all event sources that define "primary"
3948 program behaviour (where failure should be fatal) in contrast to
3949 "auxiliary" behaviour (where failure should remain local).
3951 * Most event source types sd-event supports now accept a NULL handler
3952 function, in which case the event loop is exited once the event
3953 source is to be dispatched, using the userdata pointer — converted to
3954 a signed integer — as exit code of the event loop. Previously this
3955 was supported for IO and signal event sources already. Exit event
3956 sources still do not support this (simply because it makes little
3957 sense there, as the event loop is already exiting when they are
3960 * A new per-unit setting RootImageOptions= has been added which allows
3961 tweaking the mount options for any file system mounted as effect of
3962 the RootImage= setting.
3964 * Another new per-unit setting MountImages= has been added, that allows
3965 mounting additional disk images into the file system tree accessible
3968 * Timer units gained a new FixedRandomDelay= boolean setting. If
3969 enabled, the random delay configured with RandomizedDelaySec= is
3970 selected in a way that is stable on a given system (though still
3971 different for different units).
3973 * Socket units gained a new setting Timestamping= that takes "us", "ns"
3974 or "off". This controls the SO_TIMESTAMP/SO_TIMESTAMPNS socket
3977 * systemd-repart now generates JSON output when requested with the new
3980 * systemd-machined's OpenMachineShell() bus call will now pass
3981 additional policy metadata data fields to the PolicyKit
3982 authentication request.
3984 * systemd-tmpfiles gained a new -E switch, which is equivalent to
3985 --exclude-prefix=/dev --exclude-prefix=/proc --exclude=/run
3986 --exclude=/sys. It's particularly useful in combination with --root=,
3987 when operating on OS trees that do not have any of these four runtime
3988 directories mounted, as this means no files below these subtrees are
3989 created or modified, since those mount points should probably remain
3992 * systemd-tmpfiles gained a new --image= switch which is like --root=,
3993 but takes a disk image instead of a directory as argument. The
3994 specified disk image is mounted inside a temporary mount namespace
3995 and the tmpfiles.d/ drop-ins stored in the image are executed and
3996 applied to the image. systemd-sysusers similarly gained a new
3997 --image= switch, that allows the sysusers.d/ drop-ins stored in the
3998 image to be applied onto the image.
4000 * Similarly, the journalctl command also gained an --image= switch,
4001 which is a quick one-step solution to look at the log data included
4004 * journalctl's --output=cat option (which outputs the log content
4005 without any metadata, just the pure text messages) will now make use
4006 of terminal colors when run on a suitable terminal, similarly to the
4009 * JSON group records now support a "description" string that may be
4010 used to add a human-readable textual description to such groups. This
4011 is supposed to match the user's GECOS field which traditionally
4012 didn't have a counterpart for group records.
4014 * The "systemd-dissect" tool that may be used to inspect OS disk images
4015 and that was previously installed to /usr/lib/systemd/ has now been
4016 moved to /usr/bin/, reflecting its updated status of an officially
4017 supported tool with a stable interface. It gained support for a new
4018 --mkdir switch which when combined with --mount has the effect of
4019 creating the directory to mount the image to if it is missing
4020 first. It also gained two new commands --copy-from and --copy-to for
4021 copying files and directories in and out of an OS image without the
4022 need to manually mount it. It also acquired support for a new option
4023 --json= to generate JSON output when inspecting an OS image.
4025 * The cgroup2 file system is now mounted with the
4026 "memory_recursiveprot" mount option, supported since kernel 5.7. This
4027 means that the MemoryLow= and MemoryMin= unit file settings now apply
4028 recursively to whole subtrees.
4030 * systemd-homed now defaults to using the btrfs file system — if
4031 available — when creating home directories in LUKS volumes. This may
4032 be changed with the DefaultFileSystemType= setting in homed.conf.
4033 It's now the default file system in various major distributions and
4034 has the major benefit for homed that it can be grown and shrunk while
4035 mounted, unlike the other contenders ext4 and xfs, which can both be
4036 grown online, but not shrunk (in fact xfs is the technically most
4037 limited option here, as it cannot be shrunk at all).
4039 * JSON user records managed by systemd-homed gained support for
4040 "recovery keys". These are basically secondary passphrases that can
4041 unlock user accounts/home directories. They are computer-generated
4042 rather than user-chosen, and typically have greater entropy.
4043 homectl's --recovery-key= option may be used to add a recovery key to
4044 a user account. The generated recovery key is displayed as a QR code,
4045 so that it can be scanned to be kept in a safe place. This feature is
4046 particularly useful in combination with systemd-homed's support for
4047 FIDO2 or PKCS#11 authentication, as a secure fallback in case the
4048 security tokens are lost. Recovery keys may be entered wherever the
4049 system asks for a password.
4051 * systemd-homed now maintains a "dirty" flag for each LUKS encrypted
4052 home directory which indicates that a home directory has not been
4053 deactivated cleanly when offline. This flag is useful to identify
4054 home directories for which the offline discard logic did not run when
4055 offlining, and where it would be a good idea to log in again to catch
4058 * systemctl gained a new parameter --timestamp= which may be used to
4059 change the style in which timestamps are output, i.e. whether to show
4060 them in local timezone or UTC, or whether to show µs granularity.
4062 * Alibaba's "pouch" container manager is now detected by
4063 systemd-detect-virt, ConditionVirtualization= and similar
4064 constructs. Similar, they now also recognize IBM PowerVM machine
4067 * systemd-nspawn has been reworked to use the /run/host/incoming/ as
4068 place to use for propagating external mounts into the
4069 container. Similarly /run/host/notify is now used as the socket path
4070 for container payloads to communicate with the container manager
4071 using sd_notify(). The container manager now uses the
4072 /run/host/inaccessible/ directory to place "inaccessible" file nodes
4073 of all relevant types which may be used by the container payload as
4074 bind mount source to over-mount inodes to make them inaccessible.
4075 /run/host/container-manager will now be initialized with the same
4076 string as the $container environment variable passed to the
4077 container's PID 1. /run/host/container-uuid will be initialized with
4078 the same string as $container_uuid. This means the /run/host/
4079 hierarchy is now the primary way to make host resources available to
4080 the container. The Container Interface documents these new files and
4083 https://systemd.io/CONTAINER_INTERFACE
4085 * Support for the "ConditionNull=" unit file condition has been
4086 deprecated and undocumented for 6 years. systemd started to warn
4087 about its use 1.5 years ago. It has now been removed entirely.
4089 * sd-bus.h gained a new API call sd_bus_error_has_names(), which takes
4090 a sd_bus_error struct and a list of error names, and checks if the
4091 error matches one of these names. It's a convenience wrapper that is
4092 useful in cases where multiple errors shall be handled the same way.
4094 * A new system call filter list "@known" has been added, that contains
4095 all system calls known at the time systemd was built.
4097 * Behaviour of system call filter allow lists has changed slightly:
4098 system calls that are contained in @known will result in EPERM by
4099 default, while those not contained in it result in ENOSYS. This
4100 should improve compatibility because known system calls will thus be
4101 communicated as prohibited, while unknown (and thus newer ones) will
4102 be communicated as not implemented, which hopefully has the greatest
4103 chance of triggering the right fallback code paths in client
4106 * "systemd-analyze syscall-filter" will now show two separate sections
4107 at the bottom of the output: system calls known during systemd build
4108 time but not included in any of the filter groups shown above, and
4109 system calls defined on the local kernel but known during systemd
4112 * If the $SYSTEMD_LOG_SECCOMP=1 environment variable is set for
4113 systemd-nspawn all system call filter violations will be logged by
4114 the kernel (audit). This is useful for tracking down system calls
4115 invoked by container payloads that are prohibited by the container's
4116 system call filter policy.
4118 * If the $SYSTEMD_SECCOMP=0 environment variable is set for
4119 systemd-nspawn (and other programs that use seccomp) all seccomp
4120 filtering is turned off.
4122 * Two new unit file settings ProtectProc= and ProcSubset= have been
4123 added that expose the hidepid= and subset= mount options of procfs.
4124 All processes of the unit will only see processes in /proc that are
4125 are owned by the unit's user. This is an important new sandboxing
4126 option that is recommended to be set on all system services. All
4127 long-running system services that are included in systemd itself set
4128 this option now. This option is only supported on kernel 5.8 and
4129 above, since the hidepid= option supported on older kernels was not a
4130 per-mount option but actually applied to the whole PID namespace.
4132 * Socket units gained a new boolean setting FlushPending=. If enabled
4133 all pending socket data/connections are flushed whenever the socket
4134 unit enters the "listening" state, i.e. after the associated service
4137 * The unit file setting NUMAMask= gained a new "all" value: when used,
4138 all existing NUMA nodes are added to the NUMA mask.
4140 * A new "credentials" logic has been added to system services. This is
4141 a simple mechanism to pass privileged data to services in a safe and
4142 secure way. It's supposed to be used to pass per-service secret data
4143 such as passwords or cryptographic keys but also associated less
4144 private information such as user names, certificates, and similar to
4145 system services. Each credential is identified by a short user-chosen
4146 name and may contain arbitrary binary data. Two new unit file
4147 settings have been added: SetCredential= and LoadCredential=. The
4148 former allows setting a credential to a literal string, the latter
4149 sets a credential to the contents of a file (or data read from a
4150 user-chosen AF_UNIX stream socket). Credentials are passed to the
4151 service via a special credentials directory, one file for each
4152 credential. The path to the credentials directory is passed in a new
4153 $CREDENTIALS_DIRECTORY environment variable. Since the credentials
4154 are passed in the file system they may be easily referenced in
4155 ExecStart= command lines too, thus no explicit support for the
4156 credentials logic in daemons is required (though ideally daemons
4157 would look for the bits they need in $CREDENTIALS_DIRECTORY
4158 themselves automatically, if set). The $CREDENTIALS_DIRECTORY is
4159 backed by unswappable memory if privileges allow it, immutable if
4160 privileges allow it, is accessible only to the service's UID, and is
4161 automatically destroyed when the service stops.
4163 * systemd-nspawn supports the same credentials logic. It can both
4164 consume credentials passed to it via the aforementioned
4165 $CREDENTIALS_DIRECTORY protocol as well as pass these credentials on
4166 to its payload. The service manager/PID 1 has been updated to match
4167 this: it can also accept credentials from the container manager that
4168 invokes it (in fact: any process that invokes it), and passes them on
4169 to its services. Thus, credentials can be propagated recursively down
4170 the tree: from a system's service manager to a systemd-nspawn
4171 service, to the service manager that runs as container payload and to
4172 the service it runs below. Credentials may also be added on the
4173 systemd-nspawn command line, using new --set-credential= and
4174 --load-credential= command line switches that match the
4175 aforementioned service settings.
4177 * systemd-repart gained new settings Format=, Encrypt=, CopyFiles= in
4178 the partition drop-ins which may be used to format/LUKS
4179 encrypt/populate any created partitions. The partitions are
4180 encrypted/formatted/populated before they are registered in the
4181 partition table, so that they appear atomically: either the
4182 partitions do not exist yet or they exist fully encrypted, formatted,
4183 and populated — there is no time window where they are
4184 "half-initialized". Thus the system is robust to abrupt shutdown: if
4185 the tool is terminated half-way during its operations on next boot it
4186 will start from the beginning.
4188 * systemd-repart's --size= operation gained a new "auto" value. If
4189 specified, and operating on a loopback file it is automatically sized
4190 to the minimal size the size constraints permit. This is useful to
4191 use "systemd-repart" as an image builder for minimally sized images.
4193 * systemd-resolved now gained a third IPC interface for requesting name
4194 resolution: besides D-Bus and local DNS to 127.0.0.53 a Varlink
4195 interface is now supported. The nss-resolve NSS module has been
4196 modified to use this new interface instead of D-Bus. Using Varlink
4197 has a major benefit over D-Bus: it works without a broker service,
4198 and thus already during earliest boot, before the dbus daemon has
4199 been started. This means name resolution via systemd-resolved now
4200 works at the same time systemd-networkd operates: from earliest boot
4201 on, including in the initrd.
4203 * systemd-resolved gained support for a new DNSStubListenerExtra=
4204 configuration file setting which may be used to specify additional IP
4205 addresses the built-in DNS stub shall listen on, in addition to the
4206 main one on 127.0.0.53:53.
4208 * Name lookups issued via systemd-resolved's D-Bus and Varlink
4209 interfaces (and thus also via glibc NSS if nss-resolve is used) will
4210 now honour a trailing dot in the hostname: if specified the search
4211 path logic is turned off. Thus "resolvectl query foo." is now
4212 equivalent to "resolvectl query --search=off foo.".
4214 * systemd-resolved gained a new D-Bus property "ResolvConfMode" that
4215 exposes how /etc/resolv.conf is currently managed: by resolved (and
4216 in which mode if so) or another subsystem. "resolvctl" will display
4217 this property in its status output.
4219 * The resolv.conf snippets systemd-resolved provides will now set "."
4220 as the search domain if no other search domain is known. This turns
4221 off the derivation of an implicit search domain by nss-dns for the
4222 hostname, when the hostname is set to an FQDN. This change is done to
4223 make nss-dns using resolv.conf provided by systemd-resolved behave
4224 more similarly to nss-resolve.
4226 * systemd-tmpfiles' file "aging" logic (i.e. the automatic clean-up of
4227 /tmp/ and /var/tmp/ based on file timestamps) now looks at the
4228 "birth" time (btime) of a file in addition to the atime, mtime, and
4231 * systemd-analyze gained a new verb "capability" that lists all known
4232 capabilities by the systemd build and by the kernel.
4234 * If a file /usr/lib/clock-epoch exists, PID 1 will read its mtime and
4235 advance the system clock to it at boot if it is noticed to be before
4236 that time. Previously, PID 1 would only advance the time to an epoch
4237 time that is set during build-time. With this new file OS builders
4238 can change this epoch timestamp on individual OS images without
4239 having to rebuild systemd.
4241 * systemd-logind will now listen to the KEY_RESTART key from the Linux
4242 input layer and reboot the system if it is pressed, similarly to how
4243 it already handles KEY_POWER, KEY_SUSPEND or KEY_SLEEP. KEY_RESTART
4244 was originally defined in the Multimedia context (to restart playback
4245 of a song or film), but is now primarily used in various embedded
4246 devices for "Reboot" buttons. Accordingly, systemd-logind will now
4247 honour it as such. This may configured in more detail via the new
4248 HandleRebootKey= and RebootKeyIgnoreInhibited=.
4250 * systemd-nspawn/systemd-machined will now reconstruct hardlinks when
4251 copying OS trees, for example in "systemd-nspawn --ephemeral",
4252 "systemd-nspawn --template=", "machinectl clone" and similar. This is
4253 useful when operating with OSTree images, which use hardlinks heavily
4254 throughout, and where such copies previously resulting in "exploding"
4257 * systemd-nspawn's --console= setting gained support for a new
4258 "autopipe" value, which is identical to "interactive" when invoked on
4259 a TTY, and "pipe" otherwise.
4261 * systemd-networkd's .network files gained support for explicitly
4262 configuring the multicast membership entries of bridge devices in the
4263 [BridgeMDB] section. It also gained support for the PIE queuing
4264 discipline in the [FlowQueuePIE] sections.
4266 * systemd-networkd's .netdev files may now be used to create "BareUDP"
4267 tunnels, configured in the new [BareUDP] setting.
4269 * systemd-networkd's Gateway= setting in .network files now accepts the
4270 special values "_dhcp4" and "_ipv6ra" to configure additional,
4271 locally defined, explicit routes to the gateway acquired via DHCP or
4272 IPv6 Router Advertisements. The old setting "_dhcp" is deprecated,
4273 but still accepted for backwards compatibility.
4275 * systemd-networkd's [IPv6PrefixDelegation] section and
4276 IPv6PrefixDelegation= options have been renamed as [IPv6SendRA] and
4277 IPv6SendRA= (the old names are still accepted for backwards
4280 * systemd-networkd's .network files gained the DHCPv6PrefixDelegation=
4281 boolean setting in [Network] section. If enabled, the delegated prefix
4282 gained by another link will be configured, and an address within the
4283 prefix will be assigned.
4285 * systemd-networkd's .network files gained the Announce= boolean setting
4286 in [DHCPv6PrefixDelegation] section. When enabled, the delegated
4287 prefix will be announced through IPv6 router advertisement (IPv6 RA).
4288 The setting is enabled by default.
4290 * VXLAN tunnels may now be marked as independent of any underlying
4291 network interface via the new Independent= boolean setting.
4293 * systemctl gained support for two new verbs: "service-log-level" and
4294 "service-log-target" may be used on services that implement the
4295 generic org.freedesktop.LogControl1 D-Bus interface to dynamically
4296 adjust the log level and target. All of systemd's long-running
4297 services support this now, but ideally all system services would
4298 implement this interface to make the system more uniformly
4301 * The SystemCallErrorNumber= unit file setting now accepts the new
4302 "kill" and "log" actions, in addition to arbitrary error number
4303 specifications as before. If "kill" the processes are killed on the
4304 event, if "log" the offending system call is audit logged.
4306 * A new SystemCallLog= unit file setting has been added that accepts a
4307 list of system calls that shall be logged about (audit).
4309 * The OS image dissection logic (as used by RootImage= in unit files or
4310 systemd-nspawn's --image= switch) has gained support for identifying
4311 and mounting explicit /usr/ partitions, which are now defined in the
4312 discoverable partition specification. This should be useful for
4313 environments where the root file system is
4314 generated/formatted/populated dynamically on first boot and combined
4315 with an immutable /usr/ tree that is supplied by the vendor.
4317 * In the final phase of shutdown, within the systemd-shutdown binary
4318 we'll now try to detach MD devices (i.e software RAID) in addition to
4319 loopback block devices and DM devices as before. This is supposed to
4320 be a safety net only, in order to increase robustness if things go
4321 wrong. Storage subsystems are expected to properly detach their
4322 storage volumes during regular shutdown already (or in case of
4323 storage backing the root file system: in the initrd hook we return to
4326 * If the SYSTEMD_LOG_TID environment variable is set all systemd tools
4327 will now log the thread ID in their log output. This is useful when
4328 working with heavily threaded programs.
4330 * If the SYSTEMD_RDRAND environment variable is set to "0", systemd will
4331 not use the RDRAND CPU instruction. This is useful in environments
4332 such as replay debuggers where non-deterministic behaviour is not
4335 * The autopaging logic in systemd's various tools (such as systemctl)
4336 has been updated to turn on "secure" mode in "less"
4337 (i.e. $LESSECURE=1) if execution in a "sudo" environment is
4338 detected. This disables invoking external programs from the pager,
4339 via the pipe logic. This behaviour may be overridden via the new
4340 $SYSTEMD_PAGERSECURE environment variable.
4342 * Units which have resource limits (.service, .mount, .swap, .slice,
4343 .socket, and .slice) gained new configuration settings
4344 ManagedOOMSwap=, ManagedOOMMemoryPressure=, and
4345 ManagedOOMMemoryPressureLimitPercent= that specify resource pressure
4346 limits and optional action taken by systemd-oomd.
4348 * A new service systemd-oomd has been added. It monitors resource
4349 contention for selected parts of the unit hierarchy using the PSI
4350 information reported by the kernel, and kills processes when memory
4351 or swap pressure is above configured limits. This service is only
4352 enabled by default in developer mode (see below) and should be
4353 considered a preview in this release. Behaviour details and option
4354 names are subject to change without the usual backwards-compatibility
4357 * A new helper oomctl has been added to introspect systemd-oomd state.
4358 It is only enabled by default in developer mode and should be
4359 considered a preview without the usual backwards-compatibility
4362 * New meson option -Dcompat-mutable-uid-boundaries= has been added. If
4363 enabled, systemd reads the system UID boundaries from /etc/login.defs
4364 at runtime, instead of using the built-in values selected during
4365 build. This is an option to improve compatibility for upgrades from
4366 old systems. It's strongly recommended not to make use of this
4367 functionality on new systems (or even enable it during build), as it
4368 makes something runtime-configurable that is mostly an implementation
4369 detail of the OS, and permits avoidable differences in deployments
4370 that create all kinds of problems in the long run.
4372 * New meson option '-Dmode=developer|release' has been added. When
4373 'developer', additional checks and features are enabled that are
4374 relevant during upstream development, e.g. verification that
4375 semi-automatically-generated documentation has been properly updated
4376 following API changes. Those checks are considered hints for
4377 developers and are not actionable in downstream builds. In addition,
4378 extra features that are not ready for general consumption may be
4379 enabled in developer mode. It is thus recommended to set
4380 '-Dmode=release' in end-user and distro builds.
4382 * systemd-cryptsetup gained support for processing detached LUKS
4383 headers specified on the kernel command line via the header=
4384 parameter of the luks.options= kernel command line option. The same
4385 device/path syntax as for key files is supported for header files
4388 * The "net_id" built-in of udev has been updated to ignore ACPI _SUN
4389 slot index data for devices that are connected through a PCI bridge
4390 where the _SUN index is associated with the bridge instead of the
4391 network device itself. Previously this would create ambiguous device
4392 naming if multiple network interfaces were connected to the same PCI
4393 bridge. Since this is a naming scheme incompatibility on systems that
4394 possess hardware like this it has been introduced as new naming
4395 scheme "v247". The previous scheme can be selected via the
4396 "net.naming-scheme=v245" kernel command line parameter.
4398 * ConditionFirstBoot= semantics have been modified to be safe towards
4399 abnormal system power-off during first boot. Specifically, the
4400 "systemd-machine-id-commit.service" service now acts as boot
4401 milestone indicating when the first boot process is sufficiently
4402 complete in order to not consider the next following boot also a
4403 first boot. If the system is reset before this unit is reached the
4404 first time, the next boot will still be considered a first boot; once
4405 it has been reached, no further boots will be considered a first
4406 boot. The "first-boot-complete.target" unit now acts as official hook
4407 point to order against this. If a service shall be run on every boot
4408 until the first boot fully succeeds it may thus be ordered before
4409 this target unit (and pull it in) and carry ConditionFirstBoot=
4412 * bootctl's set-default and set-oneshot commands now accept the three
4413 special strings "@default", "@oneshot", "@current" in place of a boot
4414 entry id. These strings are resolved to the current default and
4415 oneshot boot loader entry, as well as the currently booted one. Thus
4416 a command "bootctl set-default @current" may be used to make the
4417 currently boot menu item the new default for all subsequent boots.
4419 * "systemctl edit" has been updated to show the original effective unit
4420 contents in commented form in the text editor.
4422 * Units in user mode are now segregated into three new slices:
4423 session.slice (units that form the core of graphical session),
4424 app.slice ("normal" user applications), and background.slice
4425 (low-priority tasks). Unless otherwise configured, user units are
4426 placed in app.slice. The plan is to add resource limits and
4427 protections for the different slices in the future.
4429 * New GPT partition types for RISCV32/64 for the root and /usr
4430 partitions, and their associated Verity partitions have been defined,
4431 and are now understood by systemd-gpt-auto-generator, and the OS
4432 image dissection logic.
4434 Contributions from: Adolfo Jayme Barrientos, afg, Alec Moskvin, Alyssa
4435 Ross, Amitanand Chikorde, Andrew Hangsleben, Anita Zhang, Ansgar
4436 Burchardt, Arian van Putten, Aurelien Jarno, Axel Rasmussen, bauen1,
4437 Beniamino Galvani, Benjamin Berg, Bjørn Mork, brainrom, Chandradeep
4438 Dey, Charles Lee, Chris Down, Christian Göttsche, Christof Efkemann,
4439 Christoph Ruegge, Clemens Gruber, Daan De Meyer, Daniele Medri, Daniel
4440 Mack, Daniel Rusek, Dan Streetman, David Tardon, Dimitri John Ledkov,
4441 Dmitry Borodaenko, Elias Probst, Elisei Roca, ErrantSpore, Etienne
4442 Doms, Fabrice Fontaine, fangxiuning, Felix Riemann, Florian Klink,
4443 Franck Bui, Frantisek Sumsal, fwSmit, George Rawlinson, germanztz,
4444 Gibeom Gwon, Glen Whitney, Gogo Gogsi, Göran Uddeborg, Grant Mathews,
4445 Hans de Goede, Hans Ulrich Niedermann, Haochen Tong, Harald Seiler,
4446 huangyong, Hubert Kario, igo95862, Ikey Doherty, Insun Pyo, Jan Chren,
4447 Jan Schlüter, Jérémy Nouhaud, Jian-Hong Pan, Joerg Behrmann, Jonathan
4448 Lebon, Jörg Thalheim, Josh Brobst, Juergen Hoetzel, Julien Humbert,
4449 Kai-Chuan Hsieh, Kairui Song, Kamil Dudka, Kir Kolyshkin, Kristijan
4450 Gjoshev, Kyle Huey, Kyle Russell, Lee Whalen, Lennart Poettering,
4451 lichangze, Luca Boccassi, Lucas Werkmeister, Luca Weiss, Marc
4452 Kleine-Budde, Marco Wang, Martin Wilck, Marti Raudsepp, masmullin2000,
4453 Máté Pozsgay, Matt Fenwick, Michael Biebl, Michael Scherer, Michal
4454 Koutný, Michal Sekletár, Michal Suchanek, Mikael Szreder, Milo
4455 Casagrande, mirabilos, Mitsuha_QuQ, mog422, Muhammet Kara, Nazar
4456 Vinnichuk, Nicholas Narsing, Nicolas Fella, Njibhu, nl6720, Oğuz Ersen,
4457 Olivier Le Moal, Ondrej Kozina, onlybugreports, Pass Automated Testing
4458 Suite, Pat Coulthard, Pavel Sapezhko, Pedro Ruiz, perry_yuan, Peter
4459 Hutterer, Phaedrus Leeds, PhoenixDiscord, Piotr Drąg, Plan C,
4460 Purushottam choudhary, Rasmus Villemoes, Renaud Métrich, Robert Marko,
4461 Roman Beranek, Ronan Pigott, Roy Chen (陳彥廷), RussianNeuroMancer,
4462 Samanta Navarro, Samuel BF, scootergrisen, Sorin Ionescu, Steve Dodd,
4463 Susant Sahani, Timo Rothenpieler, Tobias Hunger, Tobias Kaufmann, Topi
4464 Miettinen, vanou, Vito Caputo, Weblate, Wen Yang, Whired Planck,
4465 williamvds, Yu, Li-Yu, Yuri Chornoivan, Yu Watanabe, Zbigniew
4466 Jędrzejewski-Szmek, Zmicer Turok, Дамјан Георгиевски
4468 – Warsaw, 2020-11-26
4472 * The service manager gained basic support for cgroup v2 freezer. Units
4473 can now be suspended or resumed either using new systemctl verbs,
4474 freeze and thaw respectively, or via D-Bus.
4476 * PID 1 may now automatically load pre-compiled AppArmor policies from
4477 /etc/apparmor/earlypolicy during early boot.
4479 * The CPUAffinity= setting in service unit files now supports a new
4480 special value "numa" that causes the CPU affinity masked to be set
4481 based on the NUMA mask.
4483 * systemd will now log about all left-over processes remaining in a
4484 unit when the unit is stopped. It will now warn about services using
4485 KillMode=none, as this is generally an unsafe thing to make use of.
4487 * Two new unit file settings
4488 ConditionPathIsEncrypted=/AssertPathIsEncrypted= have been
4489 added. They may be used to check whether a specific file system path
4490 resides on a block device that is encrypted on the block level
4491 (i.e. using dm-crypt/LUKS).
4493 * Another pair of new settings ConditionEnvironment=/AssertEnvironment=
4494 has been added that may be used for simple environment checks. This
4495 is particularly useful when passing in environment variables from a
4496 container manager (or from PAM in case of the systemd --user
4499 * .service unit files now accept a new setting CoredumpFilter= which
4500 allows configuration of the memory sections coredumps of the
4501 service's processes shall include.
4503 * .mount units gained a new ReadWriteOnly= boolean option. If set
4504 it will not be attempted to mount a file system read-only if mounting
4505 in read-write mode doesn't succeed. An option x-systemd.rw-only is
4506 available in /etc/fstab to control the same.
4508 * .socket units gained a new boolean setting PassPacketInfo=. If
4509 enabled, the kernel will attach additional per-packet metadata to all
4510 packets read from the socket, as an ancillary message. This controls
4511 the IP_PKTINFO, IPV6_RECVPKTINFO, NETLINK_PKTINFO socket options,
4512 depending on socket type.
4514 * .service units gained a new setting RootHash= which may be used to
4515 specify the root hash for verity enabled disk images which are
4516 specified in RootImage=. RootVerity= may be used to specify a path to
4517 the Verity data matching a RootImage= file system. (The latter is
4518 only useful for images that do not contain the Verity data embedded
4519 into the same image that carries a GPT partition table following the
4520 Discoverable Partition Specification). Similarly, systemd-nspawn
4521 gained a new switch --verity-data= that takes a path to a file with
4522 the verity data of the disk image supplied in --image=, if the image
4523 doesn't contain the verity data itself.
4525 * .service units gained a new setting RootHashSignature= which takes
4526 either a base64 encoded PKCS#7 signature of the root hash specified
4527 with RootHash=, or a path to a file to read the signature from. This
4528 allows validation of the root hash against public keys available in
4529 the kernel keyring, and is only supported on recent kernels
4530 (>= 5.4)/libcryptsetup (>= 2.30). A similar switch has been added to
4531 systemd-nspawn and systemd-dissect (--root-hash-sig=). Support for
4532 this mechanism has also been added to systemd-veritysetup.
4534 * .service unit files gained two new options
4535 TimeoutStartFailureMode=/TimeoutStopFailureMode= that may be used to
4536 tune behaviour if a start or stop timeout is hit, i.e. whether to
4537 terminate the service with SIGTERM, SIGABRT or SIGKILL.
4539 * Most options in systemd that accept hexadecimal values prefixed with
4540 0x in additional to the usual decimal notation now also support octal
4541 notation when the 0o prefix is used and binary notation if the 0b
4544 * Various command line parameters and configuration file settings that
4545 configure key or certificate files now optionally take paths to
4546 AF_UNIX sockets in the file system. If configured that way a stream
4547 connection is made to the socket and the required data read from
4548 it. This is a simple and natural extension to the existing regular
4549 file logic, and permits other software to provide keys or
4550 certificates via simple IPC services, for example when unencrypted
4551 storage on disk is not desired. Specifically, systemd-networkd's
4552 Wireguard and MACSEC key file settings as well as
4553 systemd-journal-gatewayd's and systemd-journal-remote's PEM
4554 key/certificate parameters support this now.
4556 * Unit files, tmpfiles.d/ snippets, sysusers.d/ snippets and other
4557 configuration files that support specifier expansion learnt six new
4558 specifiers: %a resolves to the current architecture, %o/%w/%B/%W
4559 resolve to the various ID fields from /etc/os-release, %l resolves to
4560 the "short" hostname of the system, i.e. the hostname configured in
4561 the kernel truncated at the first dot.
4563 * Support for the .include syntax in unit files has been removed. The
4564 concept has been obsolete for 6 years and we started warning about
4565 its pending removal 2 years ago (also see NEWS file below). It's
4568 * StandardError= and StandardOutput= in unit files no longer support
4569 the "syslog" and "syslog-console" switches. They were long removed
4570 from the documentation, but will now result in warnings when used,
4571 and be converted to "journal" and "journal+console" automatically.
4573 * If the service setting User= is set to the "nobody" user, a warning
4574 message is now written to the logs (but the value is nonetheless
4575 accepted). Setting User=nobody is unsafe, since the primary purpose
4576 of the "nobody" user is to own all files whose owner cannot be mapped
4577 locally. It's in particular used by the NFS subsystem and in user
4578 namespacing. By running a service under this user's UID it might get
4579 read and even write access to all these otherwise unmappable files,
4580 which is quite likely a major security problem.
4582 * tmpfs mounts automatically created by systemd (/tmp, /run, /dev/shm,
4583 and others) now have a size and inode limits applied (50% of RAM for
4584 /tmp and /dev/shm, 10% of RAM for other mounts, etc.). Please note
4585 that the implicit kernel default is 50% too, so there is no change
4586 in the size limit for /tmp and /dev/shm.
4588 * nss-mymachines lost support for resolution of users and groups, and
4589 now only does resolution of hostnames. This functionality is now
4590 provided by nss-systemd. Thus, the 'mymachines' entry should be
4591 removed from the 'passwd:' and 'group:' lines in /etc/nsswitch.conf
4592 (and 'systemd' added if it is not already there).
4594 * A new kernel command line option systemd.hostname= has been added
4595 that allows controlling the hostname that is initialized early during
4598 * A kernel command line option "udev.blockdev_read_only" has been
4599 added. If specified all hardware block devices that show up are
4600 immediately marked as read-only by udev. This option is useful for
4601 making sure that a specific boot under no circumstances modifies data
4602 on disk. Use "blockdev --setrw" to undo the effect of this, per
4605 * A new boolean kernel command line option systemd.swap= has been
4606 added, which may be used to turn off automatic activation of swap
4607 devices listed in /etc/fstab.
4609 * New kernel command line options systemd.condition-needs-update= and
4610 systemd.condition-first-boot= have been added, which override the
4611 result of the ConditionNeedsUpdate= and ConditionFirstBoot=
4614 * A new kernel command line option systemd.clock-usec= has been added
4615 that allows setting the system clock to the specified time in µs
4616 since Jan 1st, 1970 early during boot. This is in particular useful
4617 in order to make test cases more reliable.
4619 * The fs.suid_dumpable sysctl is set to 2 / "suidsafe". This allows
4620 systemd-coredump to save core files for suid processes. When saving
4621 the core file, systemd-coredump will use the effective uid and gid of
4622 the process that faulted.
4624 * The /sys/module/kernel/parameters/crash_kexec_post_notifiers file is
4625 now automatically set to "Y" at boot, in order to enable pstore
4626 generation for collection with systemd-pstore.
4628 * We provide a set of udev rules to enable auto-suspend on PCI and USB
4629 devices that were tested to correctly support it. Previously, this
4630 was distributed as a set of udev rules, but has now been replaced by
4631 by a set of hwdb entries (and a much shorter udev rule to take action
4632 if the device modalias matches one of the new hwdb entries).
4634 As before, entries are periodically imported from the database
4635 maintained by the ChromiumOS project. If you have a device that
4636 supports auto-suspend correctly and where it should be enabled by
4637 default, please submit a patch that adds it to the database (see
4638 /usr/lib/udev/hwdb.d/60-autosuspend.hwdb).
4640 * systemd-udevd gained the new configuration option timeout_signal= as well
4641 as a corresponding kernel command line option udev.timeout_signal=.
4642 The option can be used to configure the UNIX signal that the main
4643 daemon sends to the worker processes on timeout. Setting the signal
4644 to SIGABRT is useful for debugging.
4646 * .link files managed by systemd-udevd gained options RxFlowControl=,
4647 TxFlowControl=, AutoNegotiationFlowControl= in the [Link] section, in
4648 order to configure various flow control parameters. They also gained
4649 RxMiniBufferSize= and RxJumboBufferSize= in order to configure jumbo
4650 frame ring buffer sizes.
4652 * networkd.conf gained a new boolean setting ManageForeignRoutes=. If
4653 enabled systemd-networkd manages all routes configured by other tools.
4655 * .network files managed by systemd-networkd gained a new section
4656 [SR-IOV], in order to configure SR-IOV capable network devices.
4658 * systemd-networkd's [IPv6Prefix] section in .network files gained a
4659 new boolean setting Assign=. If enabled an address from the prefix is
4660 automatically assigned to the interface.
4662 * systemd-networkd gained a new section [DHCPv6PrefixDelegation] which
4663 controls delegated prefixes assigned by DHCPv6 client. The section
4664 has three settings: SubnetID=, Assign=, and Token=. The setting
4665 SubnetID= allows explicit configuration of the preferred subnet that
4666 systemd-networkd's Prefix Delegation logic assigns to interfaces. If
4667 Assign= is enabled (which is the default) an address from any acquired
4668 delegated prefix is automatically chosen and assigned to the
4669 interface. The setting Token= specifies an optional address generation
4672 * systemd-networkd's [Network] section gained a new setting
4673 IPv4AcceptLocal=. If enabled the interface accepts packets with local
4676 * systemd-networkd gained support for configuring the HTB queuing
4677 discipline in the [HierarchyTokenBucket] and
4678 [HierarchyTokenBucketClass] sections. Similar the "pfifo" qdisc may
4679 be configured in the [PFIFO] section, "GRED" in
4680 [GenericRandomEarlyDetection], "SFB" in [StochasticFairBlue], "cake"
4681 in [CAKE], "PIE" in [PIE], "DRR" in [DeficitRoundRobinScheduler] and
4682 [DeficitRoundRobinSchedulerClass], "BFIFO" in [BFIFO],
4683 "PFIFOHeadDrop" in [PFIFOHeadDrop], "PFIFOFast" in [PFIFOFast], "HHF"
4684 in [HeavyHitterFilter], "ETS" in [EnhancedTransmissionSelection] and
4685 "QFQ" in [QuickFairQueueing] and [QuickFairQueueingClass].
4687 * systemd-networkd gained support for a new Termination= setting in the
4688 [CAN] section for configuring the termination resistor. It also
4689 gained a new ListenOnly= setting for controlling whether to only
4690 listen on CAN interfaces, without interfering with traffic otherwise
4691 (which is useful for debugging/monitoring CAN network
4692 traffic). DataBitRate=, DataSamplePoint=, FDMode=, FDNonISO= have
4693 been added to configure various CAN-FD aspects.
4695 * systemd-networkd's [DHCPv6] section gained a new option WithoutRA=.
4696 When enabled, DHCPv6 will be attempted right-away without requiring an
4697 Router Advertisement packet suggesting it first (i.e. without the 'M'
4698 or 'O' flags set). The [IPv6AcceptRA] section gained a boolean option
4699 DHCPv6Client= that may be used to turn off the DHCPv6 client even if
4700 the RA packets suggest it.
4702 * systemd-networkd's [DHCPv4] section gained a new setting UseGateway=
4703 which may be used to turn off use of the gateway information provided
4704 by the DHCP lease. A new FallbackLeaseLifetimeSec= setting may be
4705 used to configure how to process leases that lack a lifetime option.
4707 * systemd-networkd's [DHCPv4] and [DHCPServer] sections gained a new
4708 setting SendVendorOption= allowing configuration of additional vendor
4709 options to send in the DHCP requests/responses. The [DHCPv6] section
4710 gained a new SendOption= setting for sending arbitrary DHCP
4711 options. RequestOptions= has been added to request arbitrary options
4712 from the server. UserClass= has been added to set the DHCP user class
4715 * systemd-networkd's [DHCPServer] section gained a new set of options
4716 EmitPOP3=/POP3=, EmitSMTP=/SMTP=, EmitLPR=/LPR= for including server
4717 information about these three protocols in the DHCP lease. It also
4718 gained support for including "MUD" URLs ("Manufacturer Usage
4719 Description"). Support for "MUD" URLs was also added to the LLDP
4720 stack, configurable in the [LLDP] section in .network files.
4722 * The Mode= settings in [MACVLAN] and [MACVTAP] now support 'source'
4723 mode. Also, the sections now support a new setting SourceMACAddress=.
4725 * systemd-networkd's .netdev files now support a new setting
4726 VLANProtocol= in the [Bridge] section that allows configuration of
4727 the VLAN protocol to use.
4729 * systemd-networkd supports a new Group= setting in the [Link] section
4730 of the .network files, to control the link group.
4732 * systemd-networkd's [Network] section gained a new
4733 IPv6LinkLocalAddressGenerationMode= setting, which specifies how IPv6
4734 link local address is generated.
4736 * A new default .network file is now shipped that matches TUN/TAP
4737 devices that begin with "vt-" in their name. Such interfaces will
4738 have IP routing onto the host links set up automatically. This is
4739 supposed to be used by VM managers to trivially acquire a network
4740 interface which is fully set up for host communication, simply by
4741 carefully picking an interface name to use.
4743 * systemd-networkd's [DHCPv6] section gained a new setting RouteMetric=
4744 which sets the route priority for routes specified by the DHCP server.
4746 * systemd-networkd's [DHCPv6] section gained a new setting VendorClass=
4747 which configures the vendor class information sent to DHCP server.
4749 * The BlackList= settings in .network files' [DHCPv4] and
4750 [IPv6AcceptRA] sections have been renamed DenyList=. The old names
4751 are still understood to provide compatibility.
4753 * networkctl gained the new "forcerenew" command for forcing all DHCP
4754 server clients to renew their lease. The interface "status" output
4755 will now show numerous additional fields of information about an
4756 interface. There are new "up" and "down" commands to bring specific
4757 interfaces up or down.
4759 * systemd-resolved's DNS= configuration option now optionally accepts a
4760 port number (after ":") and a host name (after "#"). When the host
4761 name is specified, the DNS-over-TLS certificate is validated to match
4762 the specified hostname. Additionally, in case of IPv6 addresses, an
4763 interface may be specified (after "%").
4765 * systemd-resolved may be configured to forward single-label DNS names.
4766 This is not standard-conformant, but may make sense in setups where
4767 public DNS servers are not used.
4769 * systemd-resolved's DNS-over-TLS support gained SNI validation.
4771 * systemd-nspawn's --resolv-conf= switch gained a number of new
4772 supported values. Specifically, options starting with "replace-" are
4773 like those prefixed "copy-" but replace any existing resolv.conf
4774 file. And options ending in "-uplink" and "-stub" can now be used to
4775 propagate other flavours of resolv.conf into the container (as
4776 defined by systemd-resolved).
4778 * The various programs included in systemd can now optionally output
4779 their log messages on stderr prefixed with a timestamp, controlled by
4780 the $SYSTEMD_LOG_TIME environment variable.
4782 * systemctl gained a new "-P" switch that is a shortcut for "--value
4785 * "systemctl list-units" and "systemctl list-machines" no longer hide
4786 their first output column with --no-legend. To hide the first column,
4789 * "systemctl reboot" takes the option "--reboot-argument=".
4790 The optional positional argument to "systemctl reboot" is now
4791 being deprecated in favor of this option.
4793 * systemd-run gained a new switch --slice-inherit. If specified the
4794 unit it generates is placed in the same slice as the systemd-run
4797 * systemd-journald gained support for zstd compression of large fields
4798 in journal files. The hash tables in journal files have been hardened
4799 against hash collisions. This is an incompatible change and means
4800 that journal files created with new systemd versions are not readable
4801 with old versions. If the $SYSTEMD_JOURNAL_KEYED_HASH boolean
4802 environment variable for systemd-journald.service is set to 0 this
4803 new hardening functionality may be turned off, so that generated
4804 journal files remain compatible with older journalctl
4807 * journalctl will now include a clickable link in the default output for
4808 each log message for which an URL with further documentation is
4809 known. This is only supported on terminal emulators that support
4810 clickable hyperlinks, and is turned off if a pager is used (since
4811 "less" still doesn't support hyperlinks,
4812 unfortunately). Documentation URLs may be included in log messages
4813 either by including a DOCUMENTATION= journal field in it, or by
4814 associating a journal message catalog entry with the log message's
4815 MESSAGE_ID, which then carries a "Documentation:" tag.
4817 * journald.conf gained a new boolean setting Audit= that may be used to
4818 control whether systemd-journald will enable audit during
4821 * when systemd-journald's log stream is broken up into multiple lines
4822 because the PID of the sender changed this is indicated in the
4823 generated log records via the _LINE_BREAK=pid-change field.
4825 * journalctl's "-o cat" output mode will now show one or more journal
4826 fields specified with --output-fields= instead of unconditionally
4827 MESSAGE=. This is useful to retrieve a very specific set of fields
4828 without any decoration.
4830 * The sd-journal.h API gained two new functions:
4831 sd_journal_enumerate_available_unique() and
4832 sd_journal_enumerate_available_data() that operate like their
4833 counterparts that lack the _available_ in the name, but skip items
4834 that cannot be read and processed by the local implementation
4835 (i.e. are compressed in an unsupported format or such),
4837 * coredumpctl gained a new --file= switch, matching the same one in
4838 journalctl: a specific journal file may be specified to read the
4841 * coredumps collected by systemd-coredump may now be compressed using
4844 * systemd-binfmt gained a new switch --unregister for unregistering all
4845 registered entries at once. This is now invoked automatically at
4846 shutdown, so that binary formats registered with the "F" flag will
4847 not block clean file system unmounting.
4849 * systemd-notify's --pid= switch gained new values: "parent", "self",
4850 "auto" for controlling which PID to send to the service manager: the
4851 systemd-notify process' PID, or the one of the process invoking it.
4853 * systemd-logind's Session bus object learnt a new method call
4854 SetType() for temporarily updating the session type of an already
4855 allocated session. This is useful for upgrading tty sessions to
4856 graphical ones once a compositor is invoked.
4858 * systemd-socket-proxy gained a new switch --exit-idle-time= for
4859 configuring an exit-on-idle time.
4861 * systemd-repart's --empty= setting gained a new value "create". If
4862 specified a new empty regular disk image file is created under the
4863 specified name. Its size may be specified with the new --size=
4864 option. The latter is also supported without the "create" mode, in
4865 order to grow existing disk image files to the specified size. These
4866 two new options are useful when creating or manipulating disk images
4867 instead of operating on actual block devices.
4869 * systemd-repart drop-ins now support a new UUID= setting to control
4870 the UUID to assign to a newly created partition.
4872 * systemd-repart's SizeMin= per-partition parameter now defaults to 10M
4875 * systemd-repart's Label= setting now support the usual, simple
4876 specifier expansion.
4878 * systemd-homed's LUKS backend gained the ability to discard empty file
4879 system blocks automatically when the user logs out. This is enabled
4880 by default to ensure that home directories take minimal space when
4881 logged out but get full size guarantees when logged in. This may be
4882 controlled with the new --luks-offline-discard= switch to homectl.
4884 * If systemd-homed detects that /home/ is encrypted as a whole it will
4885 now default to the directory or subvolume backends instead of the
4886 LUKS backend, in order to avoid double encryption. The default
4887 storage and file system may now be configured explicitly, too, via
4888 the new /etc/systemd/homed.conf configuration file.
4890 * systemd-homed now supports unlocking home directories with FIDO2
4891 security tokens that support the 'hmac-secret' extension, in addition
4892 to the existing support for PKCS#11 security token unlocking
4893 support. Note that many recent hardware security tokens support both
4894 interfaces. The FIDO2 support is accessible via homectl's
4895 --fido2-device= option.
4897 * homectl's --pkcs11-uri= setting now accepts two special parameters:
4898 if "auto" is specified and only one suitable PKCS#11 security token
4899 is plugged in, its URL is automatically determined and enrolled for
4900 unlocking the home directory. If "list" is specified a brief table of
4901 suitable PKCS#11 security tokens is shown. Similar, the new
4902 --fido2-device= option also supports these two special values, for
4903 automatically selecting and listing suitable FIDO2 devices.
4905 * The /etc/crypttab tmp option now optionally takes an argument
4906 selecting the file system to use. Moreover, the default is now
4907 changed from ext2 to ext4.
4909 * There's a new /etc/crypttab option "keyfile-erase". If specified the
4910 key file listed in the same line is removed after use, regardless if
4911 volume activation was successful or not. This is useful if the key
4912 file is only acquired transiently at runtime and shall be erased
4913 before the system continues to boot.
4915 * There's also a new /etc/crypttab option "try-empty-password". If
4916 specified, before asking the user for a password it is attempted to
4917 unlock the volume with an empty password. This is useful for
4918 installing encrypted images whose password shall be set on first boot
4919 instead of at installation time.
4921 * systemd-cryptsetup will now attempt to load the keys to unlock
4922 volumes with automatically from files in
4923 /etc/cryptsetup-keys.d/<volume>.key and
4924 /run/cryptsetup-keys.d/<volume>.key, if any of these files exist.
4926 * systemd-cryptsetup may now activate Microsoft BitLocker volumes via
4927 /etc/crypttab, during boot.
4929 * logind.conf gained a new RuntimeDirectoryInodesMax= setting to
4930 control the inode limit for the per-user $XDG_RUNTIME_DIR tmpfs
4933 * A new generator systemd-xdg-autostart-generator has been added. It
4934 generates systemd unit files from XDG autostart .desktop files, and
4935 may be used to let the systemd user instance manage services that are
4936 started automatically as part of the desktop session.
4938 * "bootctl" gained a new verb "reboot-to-firmware" that may be used
4939 to query and change the firmware's 'reboot into firmware' setup flag.
4941 * systemd-firstboot gained a new switch --kernel-command-line= that may
4942 be used to initialize the /etc/kernel/cmdline file of the image. It
4943 also gained a new switch --root-password-hashed= which is like
4944 --root-password= but accepts a pre-hashed UNIX password as
4945 argument. The new option --delete-root-password may be used to unset
4946 any password for the root user (dangerous!). The --root-shell= switch
4947 may be used to control the shell to use for the root account. A new
4948 --force option may be used to override any already set settings with
4949 the parameters specified on the command line (by default, the tool
4950 will not override what has already been set before, i.e. is purely
4953 * systemd-firstboot gained support for a new --image= switch, which is
4954 similar to --root= but accepts the path to a disk image file, on
4955 which it then operates.
4957 * A new sd-path.h API has been added to libsystemd. It provides a
4958 simple API for retrieving various search paths and primary
4959 directories for various resources.
4961 * A new call sd_notify_barrier() has been added to the sd-daemon.h
4962 API. The call will block until all previously sent sd_notify()
4963 messages have been processed by the service manager. This is useful
4964 to remove races caused by a process already having disappeared at the
4965 time a notification message is processed by the service manager,
4966 making correct attribution impossible. The systemd-notify tool will
4967 now make use of this call implicitly, but this can be turned off again
4968 via the new --no-block switch.
4970 * When sending a file descriptor (fd) to the service manager to keep
4971 track of, using the sd_notify() mechanism, a new parameter FDPOLL=0
4972 may be specified. If passed the service manager will refrain from
4973 poll()ing on the file descriptor. Traditionally (and when the
4974 parameter is not specified), the service manager will poll it for
4975 POLLHUP or POLLERR events, and immediately close the fds in that
4978 * The service manager (PID1) gained a new D-Bus method call
4979 SetShowStatus() which may be used to control whether it shall show
4980 boot-time status output on the console. This method has a similar
4981 effect to sending SIGRTMIN+20/SIGRTMIN+21 to PID 1.
4983 * The sd-bus API gained a number of convenience functions that take
4984 va_list arguments rather than "...". For example, there's now
4985 sd_bus_call_methodv() to match sd_bus_call_method(). Those calls make
4986 it easier to build wrappers that accept variadic arguments and want
4987 to pass a ready va_list structure to sd-bus.
4989 * sd-bus vtable entries can have a new SD_BUS_VTABLE_ABSOLUTE_OFFSET
4990 flag which alters how the userdata pointer to pass to the callbacks
4991 is determined. When the flag is set, the offset field is converted
4992 as-is into a pointer, without adding it to the object pointer the
4993 vtable is associated with.
4995 * sd-bus now exposes four new functions:
4996 sd_bus_interface_name_is_valid() + sd_bus_service_name_is_valid() +
4997 sd_bus_member_name_is_valid() + sd_bus_object_path_is_valid() will
4998 validate strings to check if they qualify as various D-Bus concepts.
5000 * The sd-bus API gained the SD_BUS_METHOD_WITH_ARGS(),
5001 SD_BUS_METHOD_WITH_ARGS_OFFSET() and SD_BUS_SIGNAL_WITH_ARGS() macros
5002 that simplify adding argument names to D-Bus methods and signals.
5004 * The man pages for the sd-bus and sd-hwdb APIs have been completed.
5006 * Various D-Bus APIs of systemd daemons now have man pages that
5007 document the methods, signals and properties.
5009 * The expectations on user/group name syntax are now documented in
5010 detail; documentation on how classic home directories may be
5011 converted into home directories managed by homed has been added;
5012 documentation regarding integration of homed/userdb functionality in
5013 desktops has been added:
5015 https://systemd.io/USER_NAMES
5016 https://systemd.io/CONVERTING_TO_HOMED
5017 https://systemd.io/USERDB_AND_DESKTOPS
5019 * Documentation for the on-disk Journal file format has been updated
5020 and has now moved to:
5022 https://systemd.io/JOURNAL_FILE_FORMAT
5024 * The interface for containers (https://systemd.io/CONTAINER_INTERFACE)
5025 has been extended by a set of environment variables that expose
5026 select fields from the host's os-release file to the container
5027 payload. Similarly, host's os-release files can be mounted into the
5028 container underneath /run/host. Together, those mechanisms provide a
5029 standardized way to expose information about the host to the
5030 container payload. Both interfaces are implemented in systemd-nspawn.
5032 * All D-Bus services shipped in systemd now implement the generic
5033 LogControl1 D-Bus API which allows clients to change log level +
5034 target of the service during runtime.
5036 * Only relevant for developers: the mkosi.default symlink has been
5037 dropped from version control. Please create a symlink to one of the
5038 distribution-specific defaults in .mkosi/ based on your preference.
5040 Contributions from: 24bisquitz, Adam Nielsen, Alan Perry, Alexander
5041 Malafeev, Amitanand.Chikorde, Alin Popa, Alvin Šipraga, Amos Bird,
5042 Andreas Rammhold, AndreRH, Andrew Doran, Anita Zhang, Ankit Jain,
5043 antznin, Arnaud Ferraris, Arthur Moraes do Lago, Arusekk, Balaji
5044 Punnuru, Balint Reczey, Bastien Nocera, bemarek, Benjamin Berg,
5045 Benjamin Dahlhoff, Benjamin Robin, Chris Down, Chris Kerr, Christian
5046 Göttsche, Christian Hesse, Christian Oder, Ciprian Hacman, Clinton Roy,
5047 codicodi, Corey Hinshaw, Daan De Meyer, Dana Olson, Dan Callaghan,
5048 Daniel Fullmer, Daniel Rusek, Dan Streetman, Dave Reisner, David
5049 Edmundson, David Wood, Denis Pronin, Diego Escalante Urrelo, Dimitri
5050 John Ledkov, dolphrundgren, duguxy, Einsler Lee, Elisei Roca, Emmanuel
5051 Garette, Eric Anderson, Eric DeVolder, Evgeny Vereshchagin,
5052 ExtinctFire, fangxiuning, Ferran Pallarès Roca, Filipe Brandenburger,
5053 Filippo Falezza, Finn, Florian Klink, Florian Mayer, Franck Bui,
5054 Frantisek Sumsal, gaurav, Georg Müller, Gergely Polonkai, Giedrius
5055 Statkevičius, Gigadoc2, gogogogi, Gaurav Singh, gzjsgdsb, Hans de
5056 Goede, Haochen Tong, ianhi, ignapk, Jakov Smolic, James T. Lee, Jan
5057 Janssen, Jan Klötzke, Jan Palus, Jay Burger, Jeremy Cline, Jérémy
5058 Rosen, Jian-Hong Pan, Jiri Slaby, Joel Shapiro, Joerg Behrmann, Jörg
5059 Thalheim, Jouke Witteveen, Kai-Heng Feng, Kenny Levinsen, Kevin
5060 Kuehler, Kumar Kartikeya Dwivedi, layderv, laydervus, Lénaïc Huard,
5061 Lennart Poettering, Lidong Zhong, Luca Boccassi, Luca BRUNO, Lucas
5062 Werkmeister, Lukas Klingsbo, Lukáš Nykrýn, Łukasz Stelmach, Maciej
5063 S. Szmigiero, MadMcCrow, Marc-André Lureau, Marcel Holtmann, Marc
5064 Kleine-Budde, Martin Hundebøll, Matthew Leeds, Matt Ranostay, Maxim
5065 Fomin, MaxVerevkin, Michael Biebl, Michael Chapman, Michael Gubbels,
5066 Michael Marley, Michał Bartoszkiewicz, Michal Koutný, Michal Sekletár,
5067 Mike Gilbert, Mike Kazantsev, Mikhail Novosyolov, ml, Motiejus Jakštys,
5068 nabijaczleweli, nerdopolis, Niccolò Maggioni, Niklas Hambüchen, Norbert
5069 Lange, Paul Cercueil, pelzvieh, Peter Hutterer, Piero La Terza, Pieter
5070 Lexis, Piotr Drąg, Rafael Fontenelle, Richard Petri, Ronan Pigott, Ross
5071 Lagerwall, Rubens Figueiredo, satmandu, Sean-StarLabs, Sebastian
5072 Jennen, sterlinghughes, Surhud More, Susant Sahani, szb512, Thomas
5073 Haller, Tobias Hunger, Tom, Tomáš Pospíšek, Tomer Shechner, Tom Hughes,
5074 Topi Miettinen, Tudor Roman, Uwe Kleine-König, Valery0xff, Vito Caputo,
5075 Vladimir Panteleev, Vladyslav Tronko, Wen Yang, Yegor Vialov, Yigal
5076 Korman, Yi Gao, YmrDtnJu, Yuri Chornoivan, Yu Watanabe, Zbigniew
5077 Jędrzejewski-Szmek, Zhu Li, Дамјан Георгиевски, наб
5079 – Warsaw, 2020-07-30
5083 * A new tool "systemd-repart" has been added, that operates as an
5084 idempotent declarative repartitioner for GPT partition tables.
5085 Specifically, a set of partitions that must or may exist can be
5086 configured via drop-in files, and during every boot the partition
5087 table on disk is compared with these files, creating missing
5088 partitions or growing existing ones based on configurable relative
5089 and absolute size constraints. The tool is strictly incremental,
5090 i.e. does not delete, shrink or move partitions, but only adds and
5091 grows them. The primary use-case is OS images that ship in minimized
5092 form, that on first boot are grown to the size of the underlying
5093 block device or augmented with additional partitions. For example,
5094 the root partition could be extended to cover the whole disk, or a
5095 swap or /home partitions could be added on first boot. It can also be
5096 used for systems that use an A/B update scheme but ship images with
5097 just the A partition, with B added on first boot. The tool is
5098 primarily intended to be run in the initrd, shortly before
5099 transitioning into the host OS, but can also be run after the
5100 transition took place. It automatically discovers the disk backing
5101 the root file system, and should hence not require any additional
5102 configuration besides the partition definition drop-ins. If no
5103 configuration drop-ins are present, no action is taken.
5105 * A new component "userdb" has been added, along with a small daemon
5106 "systemd-userdbd.service" and a client tool "userdbctl". The framework
5107 allows defining rich user and group records in a JSON format,
5108 extending on the classic "struct passwd" and "struct group"
5109 structures. Various components in systemd have been updated to
5110 process records in this format, including systemd-logind and
5111 pam-systemd. The user records are intended to be extensible, and
5112 allow setting various resource management, security and runtime
5113 parameters that shall be applied to processes and sessions of the
5114 user as they log in. This facility is intended to allow associating
5115 such metadata directly with user/group records so that they can be
5116 produced, extended and consumed in unified form. We hope that
5117 eventually frameworks such as sssd will generate records this way, so
5118 that for the first time resource management and various other
5119 per-user settings can be configured in LDAP directories and then
5120 provided to systemd (specifically to systemd-logind and pam-system)
5121 to apply on login. For further details see:
5123 https://systemd.io/USER_RECORD
5124 https://systemd.io/GROUP_RECORD
5125 https://systemd.io/USER_GROUP_API
5127 * A small new service systemd-homed.service has been added, that may be
5128 used to securely manage home directories with built-in encryption.
5129 The complete user record data is unified with the home directory,
5130 thus making home directories naturally migratable. Its primary
5131 back-end is based on LUKS volumes, but fscrypt, plain directories,
5132 and other storage schemes are also supported. This solves a couple of
5133 problems we saw with traditional ways to manage home directories, in
5134 particular when it comes to encryption. For further discussion of
5135 this, see the video of Lennart's talk at AllSystemsGo! 2019:
5137 https://media.ccc.de/v/ASG2019-164-reinventing-home-directories
5139 For further details about the format and expectations on home
5140 directories this new daemon makes, see:
5142 https://systemd.io/HOME_DIRECTORY
5144 * systemd-journald is now multi-instantiable. In addition to the main
5145 instance systemd-journald.service there's now a template unit
5146 systemd-journald@.service, with each instance defining a new named
5147 log 'namespace' (whose name is specified via the instance part of the
5148 unit name). A new unit file setting LogNamespace= has been added,
5149 taking such a namespace name, that assigns services to the specified
5150 log namespaces. As each log namespace is serviced by its own
5151 independent journal daemon, this functionality may be used to improve
5152 performance and increase isolation of applications, at the price of
5153 losing global message ordering. Each instance of journald has a
5154 separate set of configuration files, with possibly different disk
5155 usage limitations and other settings.
5157 journalctl now takes a new option --namespace= to show logs from a
5158 specific log namespace. The sd-journal.h API gained
5159 sd_journal_open_namespace() for opening the log stream of a specific
5160 log namespace. systemd-journald also gained the ability to exit on
5161 idle, which is useful in the context of log namespaces, as this means
5162 log daemons for log namespaces can be activated automatically on
5163 demand and will stop automatically when no longer used, minimizing
5166 * When systemd-tmpfiles copies a file tree using the 'C' line type it
5167 will now label every copied file according to the SELinux database.
5169 * When systemd/PID 1 detects it is used in the initrd it will now boot
5170 into initrd.target rather than default.target by default. This should
5171 make it simpler to build initrds with systemd as for many cases the
5172 only difference between a host OS image and an initrd image now is
5173 the presence of the /etc/initrd-release file.
5175 * A new kernel command line option systemd.cpu_affinity= is now
5176 understood. It's equivalent to the CPUAffinity= option in
5177 /etc/systemd/system.conf and allows setting the CPU mask for PID 1
5178 itself and the default for all other processes.
5180 * When systemd/PID 1 is reloaded (with systemctl daemon-reload or
5181 equivalent), the SELinux database is now reloaded, ensuring that
5182 sockets and other file system objects are generated taking the new
5183 database into account.
5185 * systemd/PID 1 accepts a new "systemd.show-status=error" setting, and
5186 "quiet" has been changed to imply that instead of
5187 "systemd.show-status=auto". In this mode, only messages about errors
5188 and significant delays in boot are shown on the console.
5190 * The sd-event.h API gained native support for the new Linux "pidfd"
5191 concept. This permits watching processes using file descriptors
5192 instead of PID numbers, which fixes a number of races and makes
5193 process supervision more robust and efficient. All of systemd's
5194 components will now use pidfds if the kernel supports it for process
5195 watching, with the exception of PID 1 itself, unfortunately. We hope
5196 to move PID 1 to exclusively using pidfds too eventually, but this
5197 requires some more kernel work first. (Background: PID 1 watches
5198 processes using waitid() with the P_ALL flag, and that does not play
5199 together nicely with pidfds yet.)
5201 * Closely related to this, the sd-event.h API gained two new calls
5202 sd_event_source_send_child_signal() (for sending a signal to a
5203 watched process) and sd_event_source_get_child_process_own() (for
5204 marking a process so that it is killed automatically whenever the
5205 event source watching it is freed).
5207 * systemd-networkd gained support for configuring Token Bucket Filter
5208 (TBF) parameters in its qdisc configuration support. Similarly,
5209 support for Stochastic Fairness Queuing (SFQ), Controlled-Delay
5210 Active Queue Management (CoDel), and Fair Queue (FQ) has been added.
5212 * systemd-networkd gained support for Intermediate Functional Block
5213 (IFB) network devices.
5215 * systemd-networkd gained support for configuring multi-path IP routes,
5216 using the new MultiPathRoute= setting in the [Route] section.
5218 * systemd-networkd's DHCPv4 client has been updated to support a new
5219 SendDecline= option. If enabled, duplicate address detection is done
5220 after a DHCP offer is received from the server. If a conflict is
5221 detected, the address is declined. The DHCPv4 client also gained
5222 support for a new RouteMTUBytes= setting that allows to configure the
5223 MTU size to be used for routes generated from DHCPv4 leases.
5225 * The PrefixRoute= setting in systemd-networkd's [Address] section of
5226 .network files has been deprecated, and replaced by AddPrefixRoute=,
5227 with its sense inverted.
5229 * The Gateway= setting of [Route] sections of .network files gained
5230 support for a special new value "_dhcp". If set, the configured
5231 static route uses the gateway host configured via DHCP.
5233 * New User= and SuppressPrefixLength= settings have been implemented
5234 for the [RoutingPolicyRule] section of .network files to configure
5235 source routing based on UID ranges and prefix length, respectively.
5237 * The Type= match property of .link files has been generalized to
5238 always match the device type shown by 'networkctl status', even for
5239 devices where udev does not set DEVTYPE=. This allows e.g. Type=ether
5242 * sd-bus gained a new API call sd_bus_message_sensitive() that marks a
5243 D-Bus message object as "sensitive". Those objects are erased from
5244 memory when they are freed. This concept is intended to be used for
5245 messages that contain security sensitive data. A new flag
5246 SD_BUS_VTABLE_SENSITIVE has been introduced as well to mark methods
5247 in sd-bus vtables, causing any incoming and outgoing messages of
5248 those methods to be implicitly marked as "sensitive".
5250 * sd-bus gained a new API call sd_bus_message_dump() for dumping the
5251 contents of a message (or parts thereof) to standard output for
5254 * systemd-sysusers gained support for creating users with the primary
5255 group named differently than the user.
5257 * systemd-growfs (i.e. the x-systemd.growfs mount option in /etc/fstab)
5258 gained support for growing XFS partitions. Previously it supported
5259 only ext4 and btrfs partitions.
5261 * The support for /etc/crypttab gained a new x-initrd.attach option. If
5262 set, the specified encrypted volume is unlocked already in the
5263 initrd. This concept corresponds to the x-initrd.mount option in
5266 * systemd-cryptsetup gained native support for unlocking encrypted
5267 volumes utilizing PKCS#11 smartcards, i.e. for example to bind
5268 encryption of volumes to YubiKeys. This is exposed in the new
5269 pkcs11-uri= option in /etc/crypttab.
5271 * The /etc/fstab support in systemd now supports two new mount options
5272 x-systemd.{required,wanted}-by=, for explicitly configuring the units
5273 that the specified mount shall be pulled in by, in place of
5274 the usual local-fs.target/remote-fs.target.
5276 * The https://systemd.io/ web site has been relaunched, directly
5277 populated with most of the documentation included in the systemd
5278 repository. systemd also acquired a new logo, thanks to Tobias
5281 * systemd-udevd gained support for managing "alternative" network
5282 interface names, as supported by new Linux kernels. For the first
5283 time this permits assigning multiple (and longer!) names to a network
5284 interface. systemd-udevd will now by default assign the names
5285 generated via all supported naming schemes to each interface. This
5286 may be further tweaked with .link files and the AlternativeName= and
5287 AlternativeNamesPolicy= settings. Other components of systemd have
5288 been updated to support the new alternative names wherever
5289 appropriate. For example, systemd-nspawn will now generate
5290 alternative interface names for the host-facing side of container
5291 veth links based on the full container name without truncation.
5293 * systemd-nspawn interface naming logic has been updated in another way
5294 too: if the main interface name (i.e. as opposed to new-style
5295 "alternative" names) based on the container name is truncated, a
5296 simple hashing scheme is used to give different interface names to
5297 multiple containers whose names all begin with the same prefix. Since
5298 this changes the primary interface names pointing to containers if
5299 truncation happens, the old scheme may still be requested by
5300 selecting an older naming scheme, via the net.naming-scheme= kernel
5301 command line option.
5303 * PrivateUsers= in service files now works in services run by the
5304 systemd --user per-user instance of the service manager.
5306 * A new per-service sandboxing option ProtectClock= has been added that
5307 locks down write access to the system clock. It takes away device
5308 node access to /dev/rtc as well as the system calls that set the
5309 system clock and the CAP_SYS_TIME and CAP_WAKE_ALARM capabilities.
5310 Note that this option does not affect access to auxiliary services
5311 that allow changing the clock, for example access to
5314 * The systemd-id128 tool gained a new "show" verb for listing or
5315 resolving a number of well-known UUIDs/128bit IDs, currently mostly
5316 GPT partition table types.
5318 * The Discoverable Partitions Specification has been updated to support
5319 /var and /var/tmp partition discovery. Support for this has been
5320 added to systemd-gpt-auto-generator. For details see:
5322 https://systemd.io/DISCOVERABLE_PARTITIONS
5324 * "systemctl list-unit-files" has been updated to show a new column
5325 with the suggested enablement state based on the vendor preset files
5326 for the respective units.
5328 * "systemctl" gained a new option "--with-dependencies". If specified
5329 commands such as "systemctl status" or "systemctl cat" will now show
5330 all specified units along with all units they depend on.
5332 * networkctl gained support for showing per-interface logs in its
5335 * systemd-networkd-wait-online gained support for specifying the maximum
5336 operational state to wait for, and to wait for interfaces to
5339 * The [Match] section of .link and .network files now supports a new
5340 option PermanentMACAddress= which may be used to check against the
5341 permanent MAC address of a network device even if a randomized MAC
5344 * The [TrafficControlQueueingDiscipline] section in .network files has
5345 been renamed to [NetworkEmulator] with the "NetworkEmulator" prefix
5346 dropped from the individual setting names.
5348 * Any .link and .network files that have an empty [Match] section (this
5349 also includes empty and commented-out files) will now be
5350 rejected. systemd-udev and systemd-networkd started warning about
5351 such files in version 243.
5353 * systemd-logind will now validate access to the operation of changing
5354 the virtual terminal via a polkit action. By default, only users
5355 with at least one session on a local VT are granted permission.
5357 * When systemd sets up PAM sessions that invoked service processes
5358 shall run in, the pam_setcred() API is now invoked, thus permitting
5359 PAM modules to set additional credentials for the processes.
5361 * portablectl attach/detach verbs now accept --now and --enable options
5362 to combine attachment with enablement and invocation, or detachment
5363 with stopping and disablement.
5365 * UPGRADE ISSUE: a bug where some jobs were trimmed as redundant was
5366 fixed, which in turn exposed bugs in unit configuration of services
5367 which have Type=oneshot and should only run once, but do not have
5368 RemainAfterExit=yes set. Without RemainAfterExit=yes, a one-shot
5369 service may be started again after exiting successfully, for example
5370 as a dependency in another transaction. Affected services included
5371 some internal systemd services (most notably
5372 systemd-vconsole-setup.service, which was updated to have
5373 RemainAfterExit=yes), and plymouth-start.service. Please ensure that
5374 plymouth has been suitably updated or patched before upgrading to
5375 this systemd release. See
5376 https://bugzilla.redhat.com/show_bug.cgi?id=1807771 for some
5377 additional discussion.
5379 Contributions from: AJ Bagwell, Alin Popa, Andreas Rammhold, Anita
5380 Zhang, Ansgar Burchardt, Antonio Russo, Arian van Putten, Ashley Davis,
5381 Balint Reczey, Bart Willems, Bastien Nocera, Benjamin Dahlhoff, Charles
5382 (Chas) Williams, cheese1, Chris Down, Chris Murphy, Christian Ehrhardt,
5383 Christian Göttsche, cvoinf, Daan De Meyer, Daniele Medri, Daniel Rusek,
5384 Daniel Shahaf, Dann Frazier, Dan Streetman, Dariusz Gadomski, David
5385 Michael, Dimitri John Ledkov, Emmanuel Bourg, Evgeny Vereshchagin,
5386 ezst036, Felipe Sateler, Filipe Brandenburger, Florian Klink, Franck
5387 Bui, Fran Dieguez, Frantisek Sumsal, Greg "GothAck" Miell, Guilhem
5388 Lettron, Guillaume Douézan-Grard, Hans de Goede, HATAYAMA Daisuke, Iain
5389 Lane, James Buren, Jan Alexander Steffens (heftig), Jérémy Rosen, Jin
5390 Park, Jun'ichi Nomura, Kai Krakow, Kevin Kuehler, Kevin P. Fleming,
5391 Lennart Poettering, Leonid Bloch, Leonid Evdokimov, lothrond, Luca
5392 Boccassi, Lukas K, Lynn Kirby, Mario Limonciello, Mark Deneen, Matthew
5393 Leeds, Michael Biebl, Michal Koutný, Michal Sekletár, Mike Auty, Mike
5394 Gilbert, mtron, nabijaczleweli, Naïm Favier, Nate Jones, Norbert Lange,
5395 Oliver Giles, Paul Davey, Paul Menzel, Peter Hutterer, Piotr Drąg, Rafa
5396 Couto, Raphael, rhn, Robert Scheck, Rocka, Romain Naour, Ryan Attard,
5397 Sascha Dewald, Shengjing Zhu, Slava Kardakov, Spencer Michaels, Sylvain
5398 Plantefeve, Stanislav Angelovič, Susant Sahani, Thomas Haller, Thomas
5399 Schmitt, Timo Schlüßler, Timo Wilken, Tobias Bernard, Tobias Klauser,
5400 Tobias Stoeckmann, Topi Miettinen, tsia, WataruMatsuoka, Wieland
5401 Hoffmann, Wilhelm Schuster, Will Fleming, xduugu, Yong Cong Sin, Yuri
5402 Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zeyu
5405 – Warsaw, 2020-03-06
5409 * Support for the cpuset cgroups v2 controller has been added.
5410 Processes may be restricted to specific CPUs using the new
5411 AllowedCPUs= setting, and to specific memory NUMA nodes using the new
5412 AllowedMemoryNodes= setting.
5414 * The signal used in restart jobs (as opposed to e.g. stop jobs) may
5415 now be configured using a new RestartKillSignal= setting. This
5416 allows units which signals to request termination to implement
5417 different behaviour when stopping in preparation for a restart.
5419 * "systemctl clean" may now be used also for socket, mount, and swap
5422 * systemd will also read configuration options from the EFI variable
5423 SystemdOptions. This may be used to configure systemd behaviour when
5424 modifying the kernel command line is inconvenient, but configuration
5425 on disk is read too late, for example for the options related to
5426 cgroup hierarchy setup. 'bootctl systemd-efi-options' may be used to
5427 set the EFI variable.
5429 * systemd will now disable printk ratelimits in early boot. This should
5430 allow us to capture more logs from the early boot phase where normal
5431 storage is not available and the kernel ring buffer is used for
5432 logging. Configuration on the kernel command line has higher priority
5433 and overrides the systemd setting.
5435 systemd programs which log to /dev/kmsg directly use internal
5436 ratelimits to prevent runaway logging. (Normally this is only used
5437 during early boot, so in practice this change has very little
5440 * Unit files now support top level dropin directories of the form
5441 <unit_type>.d/ (e.g. service.d/) that may be used to add configuration
5442 that affects all corresponding unit files.
5444 * systemctl gained support for 'stop --job-mode=triggering' which will
5445 stop the specified unit and any units which could trigger it.
5447 * Unit status display now includes units triggering and triggered by
5448 the unit being shown.
5450 * The RuntimeMaxSec= setting is now supported by scopes, not just
5451 .service units. This is particularly useful for PAM sessions which
5452 create a scope unit for the user login. systemd.runtime_max_sec=
5453 setting may used with the pam_systemd module to limit the duration
5454 of the PAM session, for example for time-limited logins.
5456 * A new @pkey system call group is now defined to make it easier to
5457 allow-list memory protection syscalls for containers and services
5458 which need to use them.
5460 * systemd-udevd: removed the 30s timeout for killing stale workers on
5461 exit. systemd-udevd now waits for workers to finish. The hard-coded
5462 exit timeout of 30s was too short for some large installations, where
5463 driver initialization could be prematurely interrupted during initrd
5464 processing if the root file system had been mounted and init was
5465 preparing to switch root. If udevd is run without systemd and workers
5466 are hanging while udevd receives an exit signal, udevd will now exit
5467 when udev.event_timeout is reached for the last hanging worker. With
5468 systemd, the exit timeout can additionally be configured using
5469 TimeoutStopSec= in systemd-udevd.service.
5471 * udev now provides a program (fido_id) that identifies FIDO CTAP1
5472 ("U2F")/CTAP2 security tokens based on the usage declared in their
5473 report and descriptor and outputs suitable environment variables.
5474 This replaces the externally maintained allow lists of all known
5475 security tokens that were used previously.
5477 * Automatically generated autosuspend udev rules for allow-listed
5478 devices have been imported from the Chromium OS project. This should
5479 improve power saving with many more devices.
5481 * udev gained a new "CONST{key}=value" setting that allows matching
5482 against system-wide constants without forking a helper binary.
5483 Currently "arch" and "virt" keys are supported.
5485 * udev now opens CDROMs in non-exclusive mode when querying their
5486 capabilities. This should fix issues where other programs trying to
5487 use the CDROM cannot gain access to it, but carries a risk of
5488 interfering with programs writing to the disk, if they did not open
5489 the device in exclusive mode as they should.
5491 * systemd-networkd does not create a default route for IPv4 link local
5492 addressing anymore. The creation of the route was unexpected and was
5493 breaking routing in various cases, but people who rely on it being
5494 created implicitly will need to adjust. Such a route may be requested
5495 with DefaultRouteOnDevice=yes.
5497 Similarly, systemd-networkd will not assign a link-local IPv6 address
5498 when IPv6 link-local routing is not enabled.
5500 * Receive and transmit buffers may now be configured on links with
5501 the new RxBufferSize= and TxBufferSize= settings.
5503 * systemd-networkd may now advertise additional IPv6 routes. A new
5504 [IPv6RoutePrefix] section with Route= and LifetimeSec= options is
5507 * systemd-networkd may now configure "next hop" routes using the
5508 [NextHop] section and Gateway= and Id= settings.
5510 * systemd-networkd will now retain DHCP config on restarts by default
5511 (but this may be overridden using the KeepConfiguration= setting).
5512 The default for SendRelease= has been changed to true.
5514 * The DHCPv4 client now uses the OPTION_INFORMATION_REFRESH_TIME option
5515 received from the server.
5517 The client will use the received SIP server list if UseSIP=yes is
5520 The client may be configured to request specific options from the
5521 server using a new RequestOptions= setting.
5523 The client may be configured to send arbitrary options to the server
5524 using a new SendOption= setting.
5526 A new IPServiceType= setting has been added to configure the "IP
5527 service type" value used by the client.
5529 * The DHCPv6 client learnt a new PrefixDelegationHint= option to
5530 request prefix hints in the DHCPv6 solicitation.
5532 * The DHCPv4 server may be configured to send arbitrary options using
5533 a new SendOption= setting.
5535 * The DHCPv4 server may now be configured to emit SIP server list using
5536 the new EmitSIP= and SIP= settings.
5538 * systemd-networkd and networkctl may now renew DHCP leases on demand.
5539 networkctl has a new 'networkctl renew' verb.
5541 * systemd-networkd may now reconfigure links on demand. networkctl
5542 gained two new verbs: "reload" will reload the configuration, and
5543 "reconfigure DEVICE…" will reconfigure one or more devices.
5545 * .network files may now match on SSID and BSSID of a wireless network,
5546 i.e. the access point name and hardware address using the new SSID=
5547 and BSSID= options. networkctl will display the current SSID and
5548 BSSID for wireless links.
5550 .network files may also match on the wireless network type using the
5551 new WLANInterfaceType= option.
5553 * systemd-networkd now includes default configuration that enables
5554 link-local addressing when connected to an ad-hoc wireless network.
5556 * systemd-networkd may configure the Traffic Control queueing
5557 disciplines in the kernel using the new
5558 [TrafficControlQueueingDiscipline] section and Parent=,
5559 NetworkEmulatorDelaySec=, NetworkEmulatorDelayJitterSec=,
5560 NetworkEmulatorPacketLimit=, NetworkEmulatorLossRate=,
5561 NetworkEmulatorDuplicateRate= settings.
5563 * systemd-tmpfiles gained a new w+ setting to append to files.
5565 * systemd-analyze dump will now report when the memory configuration in
5566 the kernel does not match what systemd has configured (usually,
5567 because some external program has modified the kernel configuration
5570 * systemd-analyze gained a new --base-time= switch instructs the
5571 'calendar' verb to resolve times relative to that timestamp instead
5572 of the present time.
5574 * journalctl --update-catalog now produces deterministic output (making
5575 reproducible image builds easier).
5577 * A new devicetree-overlay setting is now documented in the Boot Loader
5580 * The default value of the WatchdogSec= setting used in systemd
5581 services (the ones bundled with the project itself) may be set at
5582 configuration time using the -Dservice-watchdog= setting. If set to
5583 empty, the watchdogs will be disabled.
5585 * systemd-resolved validates IP addresses in certificates now when GnuTLS
5588 * libcryptsetup >= 2.0.1 is now required.
5590 * A configuration option -Duser-path= may be used to override the $PATH
5591 used by the user service manager. The default is again to use the same
5592 path as the system manager.
5594 * The systemd-id128 tool gained a new switch "-u" (or "--uuid") for
5595 outputting the 128bit IDs in UUID format (i.e. in the "canonical
5598 * Service units gained a new sandboxing option ProtectKernelLogs= which
5599 makes sure the program cannot get direct access to the kernel log
5600 buffer anymore, i.e. the syslog() system call (not to be confused
5601 with the API of the same name in libc, which is not affected), the
5602 /proc/kmsg and /dev/kmsg nodes and the CAP_SYSLOG capability are made
5603 inaccessible to the service. It's recommended to enable this setting
5604 for all services that should not be able to read from or write to the
5605 kernel log buffer, which are probably almost all.
5607 Contributions from: Aaron Plattner, Alcaro, Anita Zhang, Balint Reczey,
5608 Bastien Nocera, Baybal Ni, Benjamin Bouvier, Benjamin Gilbert, Carlo
5609 Teubner, cbzxt, Chen Qi, Chris Down, Christian Rebischke, Claudio
5610 Zumbo, ClydeByrdIII, crashfistfight, Cyprien Laplace, Daniel Edgecumbe,
5611 Daniel Gorbea, Daniel Rusek, Daniel Stuart, Dan Streetman, David
5612 Pedersen, David Tardon, Dimitri John Ledkov, Dominique Martinet, Donald
5613 A. Cupp Jr, Evgeny Vereshchagin, Fabian Henneke, Filipe Brandenburger,
5614 Franck Bui, Frantisek Sumsal, Georg Müller, Hans de Goede, Haochen
5615 Tong, HATAYAMA Daisuke, Iwan Timmer, Jan Janssen, Jan Kundrát, Jan
5616 Synacek, Jan Tojnar, Jay Strict, Jérémy Rosen, Jóhann B. Guðmundsson,
5617 Jonas Jelten, Jonas Thelemann, Justin Trudell, J. Xing, Kai-Heng Feng,
5618 Kenneth D'souza, Kevin Becker, Kevin Kuehler, Lennart Poettering,
5619 Léonard Gérard, Lorenz Bauer, Luca Boccassi, Maciej Stanczew, Mario
5620 Limonciello, Marko Myllynen, Mark Stosberg, Martin Wilck, matthiasroos,
5621 Michael Biebl, Michael Olbrich, Michael Tretter, Michal Sekletar,
5622 Michal Sekletár, Michal Suchanek, Mike Gilbert, Mike Kazantsev, Nicolas
5623 Douma, nikolas, Norbert Lange, pan93412, Pascal de Bruijn, Paul Menzel,
5624 Pavel Hrdina, Peter Wu, Philip Withnall, Piotr Drąg, Rafael Fontenelle,
5625 Renaud Métrich, Riccardo Schirone, RoadrunnerWMC, Ronan Pigott, Ryan
5626 Attard, Sebastian Wick, Serge, Siddharth Chandrasekara, Steve Ramage,
5627 Steve Traylen, Susant Sahani, Thibault Nélis, Tim Teichmann, Tom
5628 Fitzhenry, Tommy J, Torsten Hilbrich, Vito Caputo, ypf791, Yu Watanabe,
5629 Zach Smith, Zbigniew Jędrzejewski-Szmek
5631 – Warsaw, 2019-11-29
5635 * This release enables unprivileged programs (i.e. requiring neither
5636 setuid nor file capabilities) to send ICMP Echo (i.e. ping) requests
5637 by turning on the "net.ipv4.ping_group_range" sysctl of the Linux
5638 kernel for the whole UNIX group range, i.e. all processes. This
5639 change should be reasonably safe, as the kernel support for it was
5640 specifically implemented to allow safe access to ICMP Echo for
5641 processes lacking any privileges. If this is not desirable, it can be
5642 disabled again by setting the parameter to "1 0".
5644 * Previously, filters defined with SystemCallFilter= would have the
5645 effect that any calling of an offending system call would terminate
5646 the calling thread. This behaviour never made much sense, since
5647 killing individual threads of unsuspecting processes is likely to
5648 create more problems than it solves. With this release the default
5649 action changed from killing the thread to killing the whole
5650 process. For this to work correctly both a kernel version (>= 4.14)
5651 and a libseccomp version (>= 2.4.0) supporting this new seccomp
5652 action is required. If an older kernel or libseccomp is used the old
5653 behaviour continues to be used. This change does not affect any
5654 services that have no system call filters defined, or that use
5655 SystemCallErrorNumber= (and thus see EPERM or another error instead
5656 of being killed when calling an offending system call). Note that
5657 systemd documentation always claimed that the whole process is
5658 killed. With this change behaviour is thus adjusted to match the
5661 * On 64 bit systems, the "kernel.pid_max" sysctl is now bumped to
5662 4194304 by default, i.e. the full 22bit range the kernel allows, up
5663 from the old 16bit range. This should improve security and
5664 robustness, as PID collisions are made less likely (though certainly
5665 still possible). There are rumours this might create compatibility
5666 problems, though at this moment no practical ones are known to
5667 us. Downstream distributions are hence advised to undo this change in
5668 their builds if they are concerned about maximum compatibility, but
5669 for everybody else we recommend leaving the value bumped. Besides
5670 improving security and robustness this should also simplify things as
5671 the maximum number of allowed concurrent tasks was previously bounded
5672 by both "kernel.pid_max" and "kernel.threads-max" and now effectively
5673 only a single knob is left ("kernel.threads-max"). There have been
5674 concerns that usability is affected by this change because larger PID
5675 numbers are harder to type, but we believe the change from 5 digits
5676 to 7 digits doesn't hamper usability.
5678 * MemoryLow= and MemoryMin= gained hierarchy-aware counterparts,
5679 DefaultMemoryLow= and DefaultMemoryMin=, which can be used to
5680 hierarchically set default memory protection values for a particular
5681 subtree of the unit hierarchy.
5683 * Memory protection directives can now take a value of zero, allowing
5684 explicit opting out of a default value propagated by an ancestor.
5686 * systemd now defaults to the "unified" cgroup hierarchy setup during
5687 build-time, i.e. -Ddefault-hierarchy=unified is now the build-time
5688 default. Previously, -Ddefault-hierarchy=hybrid was the default. This
5689 change reflects the fact that cgroupsv2 support has matured
5690 substantially in both systemd and in the kernel, and is clearly the
5691 way forward. Downstream production distributions might want to
5692 continue to use -Ddefault-hierarchy=hybrid (or even =legacy) for
5693 their builds as unfortunately the popular container managers have not
5694 caught up with the kernel API changes.
5696 * Man pages are not built by default anymore (html pages were already
5697 disabled by default), to make development builds quicker. When
5698 building systemd for a full installation with documentation, meson
5699 should be called with -Dman=true and/or -Dhtml=true as appropriate.
5700 The default was changed based on the assumption that quick one-off or
5701 repeated development builds are much more common than full optimized
5702 builds for installation, and people need to pass various other
5703 options to when doing "proper" builds anyway, so the gain from making
5704 development builds quicker is bigger than the one time disruption for
5707 Two scripts are created in the *build* directory to generate and
5708 preview man and html pages on demand, e.g.:
5710 build/man/man systemctl
5711 build/man/html systemd.index
5713 * libidn2 is used by default if both libidn2 and libidn are installed.
5714 Please use -Dlibidn=true if libidn is preferred.
5716 * The D-Bus "wire format" of the CPUAffinity= attribute is changed on
5717 big-endian machines. Before, bytes were written and read in native
5718 machine order as exposed by the native libc __cpu_mask interface.
5719 Now, little-endian order is always used (CPUs 0–7 are described by
5720 bits 0–7 in byte 0, CPUs 8–15 are described by byte 1, and so on).
5721 This change fixes D-Bus calls that cross endianness boundary.
5723 The presentation format used for CPUAffinity= by "systemctl show" and
5724 "systemd-analyze dump" is changed to present CPU indices instead of
5725 the raw __cpu_mask bitmask. For example, CPUAffinity=0-1 would be
5726 shown as CPUAffinity=03000000000000000000000000000… (on
5727 little-endian) or CPUAffinity=00000000000000300000000000000… (on
5728 64-bit big-endian), and is now shown as CPUAffinity=0-1, matching the
5729 input format. The maximum integer that will be printed in the new
5730 format is 8191 (four digits), while the old format always used a very
5731 long number (with the length varying by architecture), so they can be
5732 unambiguously distinguished.
5734 * /usr/sbin/halt.local is no longer supported. Implementation in
5735 distributions was inconsistent and it seems this functionality was
5738 To replace this functionality, users should:
5739 - either define a new unit and make it a dependency of final.target
5740 (systemctl add-wants final.target my-halt-local.service)
5741 - or move the shutdown script to /usr/lib/systemd/system-shutdown/
5742 and ensure that it accepts "halt", "poweroff", "reboot", and
5743 "kexec" as an argument, see the description in systemd-shutdown(8).
5745 * When a [Match] section in .link or .network file is empty (contains
5746 no match patterns), a warning will be emitted. Please add any "match
5747 all" pattern instead, e.g. OriginalName=* or Name=* in case all
5748 interfaces should really be matched.
5750 * A new setting NUMAPolicy= may be used to set process memory
5751 allocation policy. This setting can be specified in
5752 /etc/systemd/system.conf and hence will set the default policy for
5753 PID1. The default policy can be overridden on a per-service
5754 basis. The related setting NUMAMask= is used to specify NUMA node
5755 mask that should be associated with the selected policy.
5757 * PID 1 will now listen to Out-Of-Memory (OOM) events the kernel
5758 generates when processes it manages are reaching their memory limits,
5759 and will place their units in a special state, and optionally kill or
5760 stop the whole unit.
5762 * The service manager will now expose bus properties for the IO
5763 resources used by units. This information is also shown in "systemctl
5764 status" now (for services that have IOAccounting=yes set). Moreover,
5765 the IO accounting data is included in the resource log message
5766 generated whenever a unit stops.
5768 * Units may now configure an explicit timeout to wait for when killed
5769 with SIGABRT, for example when a service watchdog is hit. Previously,
5770 the regular TimeoutStopSec= timeout was applied in this case too —
5771 now a separate timeout may be set using TimeoutAbortSec=.
5773 * Services may now send a special WATCHDOG=trigger message with
5774 sd_notify() to trigger an immediate "watchdog missed" event, and thus
5775 trigger service termination. This is useful both for testing watchdog
5776 handling, but also for defining error paths in services, that shall
5777 be handled the same way as watchdog events.
5779 * There are two new per-unit settings IPIngressFilterPath= and
5780 IPEgressFilterPath= which allow configuration of a BPF program
5781 (usually by specifying a path to a program uploaded to /sys/fs/bpf/)
5782 to apply to the IP packet ingress/egress path of all processes of a
5783 unit. This is useful to allow running systemd services with BPF
5784 programs set up externally.
5786 * systemctl gained a new "clean" verb for removing the state, cache,
5787 runtime or logs directories of a service while it is terminated. The
5788 new verb may also be used to remove the state maintained on disk for
5789 timer units that have Persistent= configured.
5791 * During the last phase of shutdown systemd will now automatically
5792 increase the log level configured in the "kernel.printk" sysctl so
5793 that any relevant loggable events happening during late shutdown are
5794 made visible. Previously, loggable events happening so late during
5795 shutdown were generally lost if the "kernel.printk" sysctl was set to
5796 high thresholds, as regular logging daemons are terminated at that
5797 time and thus nothing is written to disk.
5799 * If processes terminated during the last phase of shutdown do not exit
5800 quickly systemd will now show their names after a short time, to make
5801 debugging easier. After a longer timeout they are forcibly killed,
5804 * journalctl (and the other tools that display logs) will now highlight
5805 warnings in yellow (previously, both LOG_NOTICE and LOG_WARNING where
5806 shown in bright bold, now only LOG_NOTICE is). Moreover, audit logs
5807 are now shown in blue color, to separate them visually from regular
5808 logs. References to configuration files are now turned into clickable
5809 links on terminals that support that.
5811 * systemd-journald will now stop logging to /var/log/journal during
5812 shutdown when /var/ is on a separate mount, so that it can be
5813 unmounted safely during shutdown.
5815 * systemd-resolved gained support for a new 'strict' DNS-over-TLS mode.
5817 * systemd-resolved "Cache=" configuration option in resolved.conf has
5818 been extended to also accept the 'no-negative' value. Previously,
5819 only a boolean option was allowed (yes/no), having yes as the
5820 default. If this option is set to 'no-negative', negative answers are
5821 not cached while the old cache heuristics are used positive answers.
5822 The default remains unchanged.
5824 * The predictable naming scheme for network devices now supports
5825 generating predictable names for "netdevsim" devices.
5827 Moreover, the "en" prefix was dropped from the ID_NET_NAME_ONBOARD
5830 Those two changes form a new net.naming-policy-scheme= entry.
5831 Distributions which want to preserve naming stability may want to set
5832 the -Ddefault-net-naming-scheme= configuration option.
5834 * systemd-networkd now supports MACsec, nlmon, IPVTAP and Xfrm
5835 interfaces natively.
5837 * systemd-networkd's bridge FDB support now allows configuration of a
5838 destination address for each entry (Destination=), as well as the
5839 VXLAN VNI (VNI=), as well as an option to declare what an entry is
5840 associated with (AssociatedWith=).
5842 * systemd-networkd's DHCPv4 support now understands a new MaxAttempts=
5843 option for configuring the maximum number of DHCP lease requests. It
5844 also learnt a new BlackList= option for deny-listing DHCP servers (a
5845 similar setting has also been added to the IPv6 RA client), as well
5846 as a SendRelease= option for configuring whether to send a DHCP
5847 RELEASE message when terminating.
5849 * systemd-networkd's DHCPv4 and DHCPv6 stacks can now be configured
5850 separately in the [DHCPv4] and [DHCPv6] sections.
5852 * systemd-networkd's DHCP support will now optionally create an
5853 implicit host route to the DNS server specified in the DHCP lease, in
5854 addition to the routes listed explicitly in the lease. This should
5855 ensure that in multi-homed systems DNS traffic leaves the systems on
5856 the interface that acquired the DNS server information even if other
5857 routes such as default routes exist. This behaviour may be turned on
5858 with the new RoutesToDNS= option.
5860 * systemd-networkd's VXLAN support gained a new option
5861 GenericProtocolExtension= for enabling VXLAN Generic Protocol
5862 Extension support, as well as IPDoNotFragment= for setting the IP
5863 "Don't fragment" bit on outgoing packets. A similar option has been
5864 added to the GENEVE support.
5866 * In systemd-networkd's [Route] section you may now configure
5867 FastOpenNoCookie= for configuring per-route TCP fast-open support, as
5868 well as TTLPropagate= for configuring Label Switched Path (LSP) TTL
5869 propagation. The Type= setting now supports local, broadcast,
5870 anycast, multicast, any, xresolve routes, too.
5872 * systemd-networkd's [Network] section learnt a new option
5873 DefaultRouteOnDevice= for automatically configuring a default route
5874 onto the network device.
5876 * systemd-networkd's bridging support gained two new options ProxyARP=
5877 and ProxyARPWifi= for configuring proxy ARP behaviour as well as
5878 MulticastRouter= for configuring multicast routing behaviour. A new
5879 option MulticastIGMPVersion= may be used to change bridge's multicast
5880 Internet Group Management Protocol (IGMP) version.
5882 * systemd-networkd's FooOverUDP support gained the ability to configure
5883 local and peer IP addresses via Local= and Peer=. A new option
5884 PeerPort= may be used to configure the peer's IP port.
5886 * systemd-networkd's TUN support gained a new setting VnetHeader= for
5887 tweaking Generic Segment Offload support.
5889 * The address family for policy rules may be specified using the new
5890 Family= option in the [RoutingPolicyRule] section.
5892 * networkctl gained a new "delete" command for removing virtual network
5893 devices, as well as a new "--stats" switch for showing device
5896 * networkd.conf gained a new setting SpeedMeter= and
5897 SpeedMeterIntervalSec=, to measure bitrate of network interfaces. The
5898 measured speed may be shown by 'networkctl status'.
5900 * "networkctl status" now displays MTU and queue lengths, and more
5901 detailed information about VXLAN and bridge devices.
5903 * systemd-networkd's .network and .link files gained a new Property=
5904 setting in the [Match] section, to match against devices with
5905 specific udev properties.
5907 * systemd-networkd's tunnel support gained a new option
5908 AssignToLoopback= for selecting whether to use the loopback device
5909 "lo" as underlying device.
5911 * systemd-networkd's MACAddress= setting in the [Neighbor] section has
5912 been renamed to LinkLayerAddress=, and it now allows configuration of
5915 * systemd-networkd's handling of the kernel's disable_ipv6 sysctl is
5916 simplified: systemd-networkd will disable the sysctl (enable IPv6) if
5917 IPv6 configuration (static or DHCPv6) was found for a given
5918 interface. It will not touch the sysctl otherwise.
5920 * The order of entries is $PATH used by the user manager instance was
5921 changed to put bin/ entries before the corresponding sbin/ entries.
5922 It is recommended to not rely on this order, and only ever have one
5923 binary with a given name in the system paths under /usr.
5925 * A new tool systemd-network-generator has been added that may generate
5926 .network, .netdev and .link files from IP configuration specified on
5927 the kernel command line in the format used by Dracut.
5929 * The CriticalConnection= setting in .network files is now deprecated,
5930 and replaced by a new KeepConfiguration= setting which allows more
5931 detailed configuration of the IP configuration to keep in place.
5933 * systemd-analyze gained a few new verbs:
5935 - "systemd-analyze timestamp" parses and converts timestamps. This is
5936 similar to the existing "systemd-analyze calendar" command which
5937 does the same for recurring calendar events.
5939 - "systemd-analyze timespan" parses and converts timespans (i.e.
5940 durations as opposed to points in time).
5942 - "systemd-analyze condition" will parse and test ConditionXYZ=
5945 - "systemd-analyze exit-status" will parse and convert exit status
5946 codes to their names and back.
5948 - "systemd-analyze unit-files" will print a list of all unit
5949 file paths and unit aliases.
5951 * SuccessExitStatus=, RestartPreventExitStatus=, and
5952 RestartForceExitStatus= now accept exit status names (e.g. "DATAERR"
5953 is equivalent to "65"). Those exit status name mappings may be
5954 displayed with the systemd-analyze exit-status verb describe above.
5956 * systemd-logind now exposes a per-session SetBrightness() bus call,
5957 which may be used to securely change the brightness of a kernel
5958 brightness device, if it belongs to the session's seat. By using this
5959 call unprivileged clients can make changes to "backlight" and "leds"
5960 devices securely with strict requirements on session membership.
5961 Desktop environments may use this to generically make brightness
5962 changes to such devices without shipping private SUID binaries or
5963 udev rules for that purpose.
5965 * "udevadm info" gained a --wait-for-initialization switch to wait for
5966 a device to be initialized.
5968 * systemd-hibernate-resume-generator will now look for resumeflags= on
5969 the kernel command line, which is similar to rootflags= and may be
5970 used to configure device timeout for the hibernation device.
5972 * sd-event learnt a new API call sd_event_source_disable_unref() for
5973 disabling and unref'ing an event source in a single function. A
5974 related call sd_event_source_disable_unrefp() has been added for use
5975 with gcc's cleanup extension.
5977 * The sd-id128.h public API gained a new definition
5978 SD_ID128_UUID_FORMAT_STR for formatting a 128bit ID in UUID format
5981 * "busctl introspect" gained a new switch --xml-interface for dumping
5982 XML introspection data unmodified.
5984 * PID 1 may now show the unit name instead of the unit description
5985 string in its status output during boot. This may be configured in
5986 the StatusUnitFormat= setting in /etc/systemd/system.conf or the
5987 kernel command line option systemd.status_unit_format=.
5989 * PID 1 now understands a new option KExecWatchdogSec= in
5990 /etc/systemd/system.conf to set a watchdog timeout for kexec reboots.
5991 Previously watchdog functionality was only available for regular
5992 reboots. The new setting defaults to off, because we don't know in
5993 the general case if the watchdog will be reset after kexec (some
5994 drivers do reset it, but not all), and the new userspace might not be
5995 configured to handle the watchdog.
5997 Moreover, the old ShutdownWatchdogSec= setting has been renamed to
5998 RebootWatchdogSec= to more clearly communicate what it is about. The
5999 old name is still accepted for compatibility.
6001 * The systemd.debug_shell kernel command line option now optionally
6002 takes a tty name to spawn the debug shell on, which allows a
6003 different tty to be selected than the built-in default.
6005 * Service units gained a new ExecCondition= setting which will run
6006 before ExecStartPre= and either continue execution of the unit (for
6007 clean exit codes), stop execution without marking the unit failed
6008 (for exit codes 1 through 254), or stop execution and fail the unit
6009 (for exit code 255 or abnormal termination).
6011 * A new service systemd-pstore.service has been added that pulls data
6012 from /sys/fs/pstore/ and saves it to /var/lib/pstore for later
6015 * timedatectl gained new verbs for configuring per-interface NTP
6016 service configuration for systemd-timesyncd.
6018 * "localectl list-locales" won't list non-UTF-8 locales anymore. It's
6019 2019. (You can set non-UTF-8 locales though, if you know their name.)
6021 * If variable assignments in sysctl.d/ files are prefixed with "-" any
6022 failures to apply them are now ignored.
6024 * systemd-random-seed.service now optionally credits entropy when
6025 applying the seed to the system. Set $SYSTEMD_RANDOM_SEED_CREDIT to
6026 true for the service to enable this behaviour, but please consult the
6027 documentation first, since this comes with a couple of caveats.
6029 * systemd-random-seed.service is now a synchronization point for full
6030 initialization of the kernel's entropy pool. Services that require
6031 /dev/urandom to be correctly initialized should be ordered after this
6034 * The systemd-boot boot loader has been updated to optionally maintain
6035 a random seed file in the EFI System Partition (ESP). During the boot
6036 phase, this random seed is read and updated with a new seed
6037 cryptographically derived from it. Another derived seed is passed to
6038 the OS. The latter seed is then credited to the kernel's entropy pool
6039 very early during userspace initialization (from PID 1). This allows
6040 systems to boot up with a fully initialized kernel entropy pool from
6041 earliest boot on, and thus entirely removes all entropy pool
6042 initialization delays from systems using systemd-boot. Special care
6043 is taken to ensure different seeds are derived on system images
6044 replicated to multiple systems. "bootctl status" will show whether
6045 a seed was received from the boot loader.
6047 * bootctl gained two new verbs:
6049 - "bootctl random-seed" will generate the file in ESP and an EFI
6050 variable to allow a random seed to be passed to the OS as described
6053 - "bootctl is-installed" checks whether systemd-boot is currently
6056 * bootctl will warn if it detects that boot entries are misconfigured
6057 (for example if the kernel image was removed without purging the
6060 * A new document has been added describing systemd's use and support
6061 for the kernel's entropy pool subsystem:
6063 https://systemd.io/RANDOM_SEEDS
6065 * When the system is hibernated the swap device to write the
6066 hibernation image to is now automatically picked from all available
6067 swap devices, preferring the swap device with the highest configured
6068 priority over all others, and picking the device with the most free
6069 space if there are multiple devices with the highest priority.
6071 * /etc/crypttab support has learnt a new keyfile-timeout= per-device
6072 option that permits selecting the timeout how long to wait for a
6073 device with an encryption key before asking for the password.
6075 * IOWeight= has learnt to properly set the IO weight when using the
6076 BFQ scheduler officially found in kernels 5.0+.
6078 * A new mailing list has been created for reporting of security issues:
6079 systemd-security@redhat.com. For mode details, see
6080 https://systemd.io/CONTRIBUTING#security-vulnerability-reports.
6082 Contributions from: Aaron Barany, Adrian Bunk, Alan Jenkins, Albrecht
6083 Lohofener, Andrej Valek, Anita Zhang, Arian van Putten, Balint Reczey,
6084 Bastien Nocera, Ben Boeckel, Benjamin Robin, camoz, Chen Qi, Chris
6085 Chiu, Chris Down, Christian Göttsche, Christian Kellner, Clinton Roy,
6086 Connor Reeder, Daniel Black, Daniel Lublin, Daniele Medri, Dan
6087 Streetman, Dave Reisner, Dave Ross, David Art, David Tardon, Debarshi
6088 Ray, Dimitri John Ledkov, Dominick Grift, Donald Buczek, Douglas
6089 Christman, Eric DeVolder, EtherGraf, Evgeny Vereshchagin, Feldwor,
6090 Felix Riemann, Florian Dollinger, Francesco Pennica, Franck Bui,
6091 Frantisek Sumsal, Franz Pletz, frederik, Hans de Goede, Iago López
6092 Galeiras, Insun Pyo, Ivan Shapovalov, Iwan Timmer, Jack, Jakob
6093 Unterwurzacher, Jan Chren, Jan Klötzke, Jan Losinski, Jan Pokorný, Jan
6094 Synacek, Jan-Michael Brummer, Jeka Pats, Jeremy Soller, Jérémy Rosen,
6095 Jiri Pirko, Joe Lin, Joerg Behrmann, Joe Richey, Jóhann B. Guðmundsson,
6096 Johannes Christ, Johannes Schmitz, Jonathan Rouleau, Jorge Niedbalski,
6097 Jörg Thalheim, Kai Krakow, Kai Lüke, Karel Zak, Kashyap Chamarthy,
6098 Krayushkin Konstantin, Lennart Poettering, Lubomir Rintel, Luca
6099 Boccassi, Luís Ferreira, Marc-André Lureau, Markus Felten, Martin Pitt,
6100 Matthew Leeds, Mattias Jernberg, Michael Biebl, Michael Olbrich,
6101 Michael Prokop, Michael Stapelberg, Michael Zhivich, Michal Koutný,
6102 Michal Sekletar, Mike Gilbert, Milan Broz, Miroslav Lichvar, mpe85,
6103 Mr-Foo, Network Silence, Oliver Harley, pan93412, Paul Menzel, pEJipE,
6104 Peter A. Bigot, Philip Withnall, Piotr Drąg, Rafael Fontenelle, Robert
6105 Scheck, Roberto Santalla, Ronan Pigott, root, RussianNeuroMancer,
6106 Sebastian Jennen, shinygold, Shreyas Behera, Simon Schricker, Susant
6107 Sahani, Thadeu Lima de Souza Cascardo, Theo Ouzhinski, Thiebaud
6108 Weksteen, Thomas Haller, Thomas Weißschuh, Tomas Mraz, Tommi Rantala,
6109 Topi Miettinen, VD-Lycos, ven, Vladimir Yerilov, Wieland Hoffmann,
6110 William A. Kennington III, William Wold, Xi Ruoyao, Yuri Chornoivan,
6111 Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zhang Xianwei
6113 – Camerino, 2019-09-03
6117 * In .link files, MACAddressPolicy=persistent (the default) is changed
6118 to cover more devices. For devices like bridges, tun, tap, bond, and
6119 similar interfaces that do not have other identifying information,
6120 the interface name is used as the basis for persistent seed for MAC
6121 and IPv4LL addresses. The way that devices that were handled
6122 previously is not changed, and this change is about covering more
6123 devices then previously by the "persistent" policy.
6125 MACAddressPolicy=random may be used to force randomized MACs and
6126 IPv4LL addresses for a device if desired.
6128 Hint: the log output from udev (at debug level) was enhanced to
6129 clarify what policy is followed and which attributes are used.
6130 `SYSTEMD_LOG_LEVEL=debug udevadm test-builtin net_setup_link /sys/class/net/<name>`
6131 may be used to view this.
6133 Hint: if a bridge interface is created without any slaves, and gains
6134 a slave later, then now the bridge does not inherit slave's MAC.
6135 To inherit slave's MAC, for example, create the following file:
6137 # /etc/systemd/network/98-bridge-inherit-mac.link
6142 MACAddressPolicy=none
6145 * The .device units generated by systemd-fstab-generator and other
6146 generators do not automatically pull in the corresponding .mount unit
6147 as a Wants= dependency. This means that simply plugging in the device
6148 will not cause the mount unit to be started automatically. But please
6149 note that the mount unit may be started for other reasons, in
6150 particular if it is part of local-fs.target, and any unit which
6151 (transitively) depends on local-fs.target is started.
6153 * networkctl list/status/lldp now accept globbing wildcards for network
6154 interface names to match against all existing interfaces.
6156 * The $PIDFILE environment variable is set to point the absolute path
6157 configured with PIDFile= for processes of that service.
6159 * The fallback DNS server list was augmented with Cloudflare public DNS
6160 servers. Use `-Ddns-servers=` to set a different fallback.
6162 * A new special target usb-gadget.target will be started automatically
6163 when a USB Device Controller is detected (which means that the system
6164 is a USB peripheral).
6166 * A new unit setting CPUQuotaPeriodSec= assigns the time period
6167 relatively to which the CPU time quota specified by CPUQuota= is
6170 * A new unit setting ProtectHostname= may be used to prevent services
6171 from modifying hostname information (even if they otherwise would
6172 have privileges to do so).
6174 * A new unit setting NetworkNamespacePath= may be used to specify a
6175 namespace for service or socket units through a path referring to a
6176 Linux network namespace pseudo-file.
6178 * The PrivateNetwork= setting and JoinsNamespaceOf= dependencies now
6179 have an effect on .socket units: when used the listening socket is
6180 created within the configured network namespace instead of the host
6183 * ExecStart= command lines in unit files may now be prefixed with ':'
6184 in which case environment variable substitution is
6185 disabled. (Supported for the other ExecXYZ= settings, too.)
6187 * .timer units gained two new boolean settings OnClockChange= and
6188 OnTimezoneChange= which may be used to also trigger a unit when the
6189 system clock is changed or the local timezone is
6190 modified. systemd-run has been updated to make these options easily
6191 accessible from the command line for transient timers.
6193 * Two new conditions for units have been added: ConditionMemory= may be
6194 used to conditionalize a unit based on installed system
6195 RAM. ConditionCPUs= may be used to conditionalize a unit based on
6196 installed CPU cores.
6198 * The @default system call filter group understood by SystemCallFilter=
6199 has been updated to include the new rseq() system call introduced in
6202 * A new time-set.target has been added that indicates that the system
6203 time has been set from a local source (possibly imprecise). The
6204 existing time-sync.target is stronger and indicates that the time has
6205 been synchronized with a precise external source. Services where
6206 approximate time is sufficient should use the new target.
6208 * "systemctl start" (and related commands) learnt a new
6209 --show-transaction option. If specified brief information about all
6210 jobs queued because of the requested operation is shown.
6212 * systemd-networkd recognizes a new operation state 'enslaved', used
6213 (instead of 'degraded' or 'carrier') for interfaces which form a
6214 bridge, bond, or similar, and an new 'degraded-carrier' operational
6215 state used for the bond or bridge master interface when one of the
6216 enslaved devices is not operational.
6218 * .network files learnt the new IgnoreCarrierLoss= option for leaving
6219 networks configured even if the carrier is lost.
6221 * The RequiredForOnline= setting in .network files may now specify a
6222 minimum operational state required for the interface to be considered
6223 "online" by systemd-networkd-wait-online. Related to this
6224 systemd-networkd-wait-online gained a new option --operational-state=
6225 to configure the same, and its --interface= option was updated to
6226 optionally also take an operational state specific for an interface.
6228 * systemd-networkd-wait-online gained a new setting --any for waiting
6229 for only one of the requested interfaces instead of all of them.
6231 * systemd-networkd now implements L2TP tunnels.
6233 * Two new .network settings UseAutonomousPrefix= and UseOnLinkPrefix=
6234 may be used to cause autonomous and onlink prefixes received in IPv6
6235 Router Advertisements to be ignored.
6237 * New MulticastFlood=, NeighborSuppression=, and Learning= .network
6238 file settings may be used to tweak bridge behaviour.
6240 * The new TripleSampling= option in .network files may be used to
6241 configure CAN triple sampling.
6243 * A new .netdev settings PrivateKeyFile= and PresharedKeyFile= may be
6244 used to point to private or preshared key for a WireGuard interface.
6246 * /etc/crypttab now supports the same-cpu-crypt and
6247 submit-from-crypt-cpus options to tweak encryption work scheduling
6250 * systemd-tmpfiles will now take a BSD file lock before operating on a
6251 contents of directory. This may be used to temporarily exclude
6252 directories from aging by taking the same lock (useful for example
6253 when extracting a tarball into /tmp or /var/tmp as a privileged user,
6254 which might create files with really old timestamps, which
6255 nevertheless should not be deleted). For further details, see:
6257 https://systemd.io/TEMPORARY_DIRECTORIES
6259 * systemd-tmpfiles' h line type gained support for the
6260 FS_PROJINHERIT_FL ('P') file attribute (introduced in kernel 4.5),
6261 controlling project quota inheritance.
6263 * sd-boot and bootctl now implement support for an Extended Boot Loader
6264 (XBOOTLDR) partition, that is intended to be mounted to /boot, in
6265 addition to the ESP partition mounted to /efi or /boot/efi.
6266 Configuration file fragments, kernels, initrds and other EFI images
6267 to boot will be loaded from both the ESP and XBOOTLDR partitions.
6268 The XBOOTLDR partition was previously described by the Boot Loader
6269 Specification, but implementation was missing in sd-boot. Support for
6270 this concept allows using the sd-boot boot loader in more
6271 conservative scenarios where the boot loader itself is placed in the
6272 ESP but the kernels to boot (and their metadata) in a separate
6275 * A system may now be booted with systemd.volatile=overlay on the
6276 kernel command line, which causes the root file system to be set up
6277 an overlayfs mount combining the root-only root directory with a
6278 writable tmpfs. In this setup, the underlying root device is not
6279 modified, and any changes are lost at reboot.
6281 * Similar, systemd-nspawn can now boot containers with a volatile
6282 overlayfs root with the new --volatile=overlay switch.
6284 * systemd-nspawn can now consume OCI runtime bundles using a new
6285 --oci-bundle= option. This implementation is fully usable, with most
6286 features in the specification implemented, but since this a lot of
6287 new code and functionality, this feature should most likely not
6288 be used in production yet.
6290 * systemd-nspawn now supports various options described by the OCI
6291 runtime specification on the command-line and in .nspawn files:
6292 --inaccessible=/Inaccessible= may be used to mask parts of the file
6293 system tree, --console=/--pipe may be used to configure how standard
6294 input, output, and error are set up.
6296 * busctl learned the `emit` verb to generate D-Bus signals.
6298 * systemd-analyze cat-config may be used to gather and display
6299 configuration spread over multiple files, for example system and user
6300 presets, tmpfiles.d, sysusers.d, udev rules, etc.
6302 * systemd-analyze calendar now takes an optional new parameter
6303 --iterations= which may be used to show a maximum number of iterations
6304 the specified expression will elapse next.
6306 * The sd-bus C API gained support for naming method parameters in the
6309 * systemd-logind gained D-Bus APIs to specify the "reboot parameter"
6310 the reboot() system call expects.
6312 * journalctl learnt a new --cursor-file= option that points to a file
6313 from which a cursor should be loaded in the beginning and to which
6314 the updated cursor should be stored at the end.
6316 * ACRN hypervisor and Windows Subsystem for Linux (WSL) are now
6317 detected by systemd-detect-virt (and may also be used in
6318 ConditionVirtualization=).
6320 * The behaviour of systemd-logind may now be modified with environment
6321 variables $SYSTEMD_REBOOT_TO_FIRMWARE_SETUP,
6322 $SYSTEMD_REBOOT_TO_BOOT_LOADER_MENU, and
6323 $SYSTEMD_REBOOT_TO_BOOT_LOADER_ENTRY. They cause logind to either
6324 skip the relevant operation completely (when set to false), or to
6325 create a flag file in /run/systemd (when set to true), instead of
6326 actually commencing the real operation when requested. The presence
6327 of /run/systemd/reboot-to-firmware-setup,
6328 /run/systemd/reboot-to-boot-loader-menu, and
6329 /run/systemd/reboot-to-boot-loader-entry, may be used by alternative
6330 boot loader implementations to replace some steps logind performs
6331 during reboot with their own operations.
6333 * systemctl can be used to request a reboot into the boot loader menu
6334 or a specific boot loader entry with the new --boot-load-menu= and
6335 --boot-loader-entry= options to a reboot command. (This requires a
6336 boot loader that supports this, for example sd-boot.)
6338 * kernel-install will no longer unconditionally create the output
6339 directory (e.g. /efi/<machine-id>/<kernel-version>) for boot loader
6340 snippets, but will do only if the machine-specific parent directory
6341 (i.e. /efi/<machine-id>/) already exists. bootctl has been modified
6342 to create this parent directory during sd-boot installation.
6344 This makes it easier to use kernel-install with plugins which support
6345 a different layout of the bootloader partitions (for example grub2).
6347 * During package installation (with `ninja install`), we would create
6348 symlinks for getty@tty1.service, systemd-networkd.service,
6349 systemd-networkd.socket, systemd-resolved.service,
6350 remote-cryptsetup.target, remote-fs.target,
6351 systemd-networkd-wait-online.service, and systemd-timesyncd.service
6352 in /etc, as if `systemctl enable` was called for those units, to make
6353 the system usable immediately after installation. Now this is not
6354 done anymore, and instead calling `systemctl preset-all` is
6355 recommended after the first installation of systemd.
6357 * A new boolean sandboxing option RestrictSUIDSGID= has been added that
6358 is built on seccomp. When turned on creation of SUID/SGID files is
6361 * The NoNewPrivileges= and the new RestrictSUIDSGID= options are now
6362 implied if DynamicUser= is turned on for a service. This hardens
6363 these services, so that they neither can benefit from nor create
6364 SUID/SGID executables. This is a minor compatibility breakage, given
6365 that when DynamicUser= was first introduced SUID/SGID behaviour was
6366 unaffected. However, the security benefit of these two options is
6367 substantial, and the setting is still relatively new, hence we opted
6368 to make it mandatory for services with dynamic users.
6370 Contributions from: Adam Jackson, Alexander Tsoy, Andrey Yashkin,
6371 Andrzej Pietrasiewicz, Anita Zhang, Balint Reczey, Beniamino Galvani,
6372 Ben Iofel, Benjamin Berg, Benjamin Dahlhoff, Chris, Chris Morin,
6373 Christopher Wong, Claudius Ellsel, Clemens Gruber, dana, Daniel Black,
6374 Davide Cavalca, David Michael, David Rheinsberg, emersion, Evgeny
6375 Vereshchagin, Filipe Brandenburger, Franck Bui, Frantisek Sumsal,
6376 Giacinto Cifelli, Hans de Goede, Hugo Kindel, Ignat Korchagin, Insun
6377 Pyo, Jan Engelhardt, Jonas Dorel, Jonathan Lebon, Jonathon Kowalski,
6378 Jörg Sommer, Jörg Thalheim, Jussi Pakkanen, Kai-Heng Feng, Lennart
6379 Poettering, Lubomir Rintel, Luís Ferreira, Martin Pitt, Matthias
6380 Klumpp, Michael Biebl, Michael Niewöhner, Michael Olbrich, Michal
6381 Sekletar, Mike Lothian, Paul Menzel, Piotr Drąg, Riccardo Schirone,
6382 Robin Elvedi, Roman Kulikov, Ronald Tschalär, Ross Burton, Ryan
6383 Gonzalez, Sebastian Krzyszkowiak, Stephane Chazelas, StKob, Susant
6384 Sahani, Sylvain Plantefève, Szabolcs Fruhwald, Taro Yamada, Theo
6385 Ouzhinski, Thomas Haller, Tobias Jungel, Tom Yan, Tony Asleson, Topi
6386 Miettinen, unixsysadmin, Van Laser, Vesa Jääskeläinen, Yu, Li-Yu,
6387 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
6389 — Warsaw, 2019-04-11
6393 * The default locale can now be configured at compile time. Otherwise,
6394 a suitable default will be selected automatically (one of C.UTF-8,
6395 en_US.UTF-8, and C).
6397 * The version string shown by systemd and other tools now includes the
6398 git commit hash when built from git. An override may be specified
6399 during compilation, which is intended to be used by distributions to
6400 include the package release information.
6402 * systemd-cat can now filter standard input and standard error streams
6403 for different syslog priorities using the new --stderr-priority=
6406 * systemd-journald and systemd-journal-remote reject entries which
6407 contain too many fields (CVE-2018-16865) and set limits on the
6408 process' command line length (CVE-2018-16864).
6410 * $DBUS_SESSION_BUS_ADDRESS environment variable is set by pam_systemd
6413 * A new network device NamePolicy "keep" is implemented for link files,
6414 and used by default in 99-default.link (the fallback configuration
6415 provided by systemd). With this policy, if the network device name
6416 was already set by userspace, the device will not be renamed again.
6417 This matches the naming scheme that was implemented before
6418 systemd-240. If naming-scheme < 240 is specified, the "keep" policy
6419 is also enabled by default, even if not specified. Effectively, this
6420 means that if naming-scheme >= 240 is specified, network devices will
6421 be renamed according to the configuration, even if they have been
6422 renamed already, if "keep" is not specified as the naming policy in
6423 the .link file. The 99-default.link file provided by systemd includes
6424 "keep" for backwards compatibility, but it is recommended for user
6425 installed .link files to *not* include it.
6427 The "kernel" policy, which keeps kernel names declared to be
6428 "persistent", now works again as documented.
6430 * kernel-install script now optionally takes the paths to one or more
6431 initrd files, and passes them to all plugins.
6433 * The mincore() system call has been dropped from the @system-service
6434 system call filter group, as it is pretty exotic and may potentially
6435 used for side-channel attacks.
6437 * -fPIE is dropped from compiler and linker options. Please specify
6438 -Db_pie=true option to meson to build position-independent
6439 executables. Note that the meson option is supported since meson-0.49.
6441 * The fs.protected_regular and fs.protected_fifos sysctls, which were
6442 added in Linux 4.19 to make some data spoofing attacks harder, are
6443 now enabled by default. While this will hopefully improve the
6444 security of most installations, it is technically a backwards
6445 incompatible change; to disable these sysctls again, place the
6446 following lines in /etc/sysctl.d/60-protected.conf or a similar file:
6448 fs.protected_regular = 0
6449 fs.protected_fifos = 0
6451 Note that the similar hardlink and symlink protection has been
6452 enabled since v199, and may be disabled likewise.
6454 * The files read from the EnvironmentFile= setting in unit files now
6455 parse backslashes inside quotes literally, matching the behaviour of
6458 * udevadm trigger, udevadm control, udevadm settle and udevadm monitor
6459 now automatically become NOPs when run in a chroot() environment.
6461 * The tmpfiles.d/ "C" line type will now copy directory trees not only
6462 when the destination is so far missing, but also if it already exists
6463 as a directory and is empty. This is useful to cater for systems
6464 where directory trees are put together from multiple separate mount
6465 points but otherwise empty.
6467 * A new function sd_bus_close_unref() (and the associated
6468 sd_bus_close_unrefp()) has been added to libsystemd, that combines
6469 sd_bus_close() and sd_bus_unref() in one.
6471 * udevadm control learnt a new option for --ping for testing whether a
6472 systemd-udevd instance is running and reacting.
6474 * udevadm trigger learnt a new option for --wait-daemon for waiting
6475 systemd-udevd daemon to be initialized.
6477 Contributions from: Aaron Plattner, Alberts Muktupāvels, Alex Mayer,
6478 Ayman Bagabas, Beniamino Galvani, Burt P, Chris Down, Chris Lamb, Chris
6479 Morin, Christian Hesse, Claudius Ellsel, dana, Daniel Axtens, Daniele
6480 Medri, Dave Reisner, David Santamaría Rogado, Diego Canuhe, Dimitri
6481 John Ledkov, Evgeny Vereshchagin, Fabrice Fontaine, Filipe
6482 Brandenburger, Franck Bui, Frantisek Sumsal, govwin, Hans de Goede,
6483 James Hilliard, Jan Engelhardt, Jani Uusitalo, Jan Janssen, Jan
6484 Synacek, Jonathan McDowell, Jonathan Roemer, Jonathon Kowalski, Joost
6485 Heitbrink, Jörg Thalheim, Lance, Lennart Poettering, Louis Taylor,
6486 Lucas Werkmeister, Mantas Mikulėnas, Marc-Antoine Perennou,
6487 marvelousblack, Michael Biebl, Michael Sloan, Michal Sekletar, Mike
6488 Auty, Mike Gilbert, Mikhail Kasimov, Neil Brown, Niklas Hambüchen,
6489 Patrick Williams, Paul Seyfert, Peter Hutterer, Philip Withnall, Roger
6490 James, Ronnie P. Thomas, Ryan Gonzalez, Sam Morris, Stephan Edel,
6491 Stephan Gerhold, Susant Sahani, Taro Yamada, Thomas Haller, Topi
6492 Miettinen, YiFei Zhu, YmrDtnJu, YunQiang Su, Yu Watanabe, Zbigniew
6493 Jędrzejewski-Szmek, zsergeant77, Дамјан Георгиевски
6495 — Berlin, 2019-02-14
6499 * NoNewPrivileges=yes has been set for all long-running services
6500 implemented by systemd. Previously, this was problematic due to
6501 SELinux (as this would also prohibit the transition from PID1's label
6502 to the service's label). This restriction has since been lifted, but
6503 an SELinux policy update is required.
6504 (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.)
6506 * DynamicUser=yes is dropped from systemd-networkd.service,
6507 systemd-resolved.service and systemd-timesyncd.service, which was
6508 enabled in v239 for systemd-networkd.service and systemd-resolved.service,
6509 and since v236 for systemd-timesyncd.service. The users and groups
6510 systemd-network, systemd-resolve and systemd-timesync are created
6511 by systemd-sysusers again. Distributors or system administrators
6512 may need to create these users and groups if they not exist (or need
6513 to re-enable DynamicUser= for those units) while upgrading systemd.
6514 Also, the clock file for systemd-timesyncd may need to move from
6515 /var/lib/private/systemd/timesync/clock to /var/lib/systemd/timesync/clock.
6517 * When unit files are loaded from disk, previously systemd would
6518 sometimes (depending on the unit loading order) load units from the
6519 target path of symlinks in .wants/ or .requires/ directories of other
6520 units. This meant that unit could be loaded from different paths
6521 depending on whether the unit was requested explicitly or as a
6522 dependency of another unit, not honouring the priority of directories
6523 in search path. It also meant that it was possible to successfully
6524 load and start units which are not found in the unit search path, as
6525 long as they were requested as a dependency and linked to from
6526 .wants/ or .requires/. The target paths of those symlinks are not
6527 used for loading units anymore and the unit file must be found in
6530 * A new service type has been added: Type=exec. It's very similar to
6531 Type=simple but ensures the service manager will wait for both fork()
6532 and execve() of the main service binary to complete before proceeding
6533 with follow-up units. This is primarily useful so that the manager
6534 propagates any errors in the preparation phase of service execution
6535 back to the job that requested the unit to be started. For example,
6536 consider a service that has ExecStart= set to a file system binary
6537 that doesn't exist. With Type=simple starting the unit would be
6538 considered instantly successful, as only fork() has to complete
6539 successfully and the manager does not wait for execve(), and hence
6540 its failure is seen "too late". With the new Type=exec service type
6541 starting the unit will fail, as the manager will wait for the
6542 execve() and notice its failure, which is then propagated back to the
6545 NOTE: with the next release 241 of systemd we intend to change the
6546 systemd-run tool to default to Type=exec for transient services
6547 started by it. This should be mostly safe, but in specific corner
6548 cases might result in problems, as the systemd-run tool will then
6549 block on NSS calls (such as user name look-ups due to User=) done
6550 between the fork() and execve(), which under specific circumstances
6551 might cause problems. It is recommended to specify "-p Type=simple"
6552 explicitly in the few cases where this applies. For regular,
6553 non-transient services (i.e. those defined with unit files on disk)
6554 we will continue to default to Type=simple.
6556 * The Linux kernel's current default RLIMIT_NOFILE resource limit for
6557 userspace processes is set to 1024 (soft) and 4096
6558 (hard). Previously, systemd passed this on unmodified to all
6559 processes it forked off. With this systemd release the hard limit
6560 systemd passes on is increased to 512K, overriding the kernel's
6561 defaults and substantially increasing the number of simultaneous file
6562 descriptors unprivileged userspace processes can allocate. Note that
6563 the soft limit remains at 1024 for compatibility reasons: the
6564 traditional UNIX select() call cannot deal with file descriptors >=
6565 1024 and increasing the soft limit globally might thus result in
6566 programs unexpectedly allocating a high file descriptor and thus
6567 failing abnormally when attempting to use it with select() (of
6568 course, programs shouldn't use select() anymore, and prefer
6569 poll()/epoll, but the call unfortunately remains undeservedly popular
6570 at this time). This change reflects the fact that file descriptor
6571 handling in the Linux kernel has been optimized in more recent
6572 kernels and allocating large numbers of them should be much cheaper
6573 both in memory and in performance than it used to be. Programs that
6574 want to take benefit of the increased limit have to "opt-in" into
6575 high file descriptors explicitly by raising their soft limit. Of
6576 course, when they do that they must acknowledge that they cannot use
6577 select() anymore (and neither can any shared library they use — or
6578 any shared library used by any shared library they use and so on).
6579 Which default hard limit is most appropriate is of course hard to
6580 decide. However, given reports that ~300K file descriptors are used
6581 in real-life applications we believe 512K is sufficiently high as new
6582 default for now. Note that there are also reports that using very
6583 high hard limits (e.g. 1G) is problematic: some software allocates
6584 large arrays with one element for each potential file descriptor
6585 (Java, …) — a high hard limit thus triggers excessively large memory
6586 allocations in these applications. Hopefully, the new default of 512K
6587 is a good middle ground: higher than what real-life applications
6588 currently need, and low enough for avoid triggering excessively large
6589 allocations in problematic software. (And yes, somebody should fix
6592 * The fs.nr_open and fs.file-max sysctls are now automatically bumped
6593 to the highest possible values, as separate accounting of file
6594 descriptors is no longer necessary, as memcg tracks them correctly as
6595 part of the memory accounting anyway. Thus, from the four limits on
6596 file descriptors currently enforced (fs.file-max, fs.nr_open,
6597 RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two,
6598 and keep only the latter two. A set of build-time options
6599 (-Dbump-proc-sys-fs-file-max=false and -Dbump-proc-sys-fs-nr-open=false)
6600 has been added to revert this change in behaviour, which might be
6601 an option for systems that turn off memcg in the kernel.
6603 * When no /etc/locale.conf file exists (and hence no locale settings
6604 are in place), systemd will now use the "C.UTF-8" locale by default,
6605 and set LANG= to it. This locale is supported by various
6606 distributions including Fedora, with clear indications that upstream
6607 glibc is going to make it available too. This locale enables UTF-8
6608 mode by default, which appears appropriate for 2018.
6610 * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by
6611 default. This effectively switches the RFC3704 Reverse Path filtering
6612 from Strict mode to Loose mode. This is more appropriate for hosts
6613 that have multiple links with routes to the same networks (e.g.
6614 a client with a Wi-Fi and Ethernet both connected to the internet).
6616 Consult the kernel documentation for details on this sysctl:
6617 https://docs.kernel.org/networking/ip-sysctl.html
6619 * The v239 change to turn on "net.ipv4.tcp_ecn" by default has been
6622 * CPUAccounting=yes no longer enables the CPU controller when using
6623 kernel 4.15+ and the unified cgroup hierarchy, as required accounting
6624 statistics are now provided independently from the CPU controller.
6626 * Support for disabling a particular cgroup controller within a sub-tree
6627 has been added through the DisableControllers= directive.
6629 * cgroup_no_v1=all on the kernel command line now also implies
6630 using the unified cgroup hierarchy, unless one explicitly passes
6631 systemd.unified_cgroup_hierarchy=0 on the kernel command line.
6633 * The new "MemoryMin=" unit file property may now be used to set the
6634 memory usage protection limit of processes invoked by the unit. This
6635 controls the cgroup v2 memory.min attribute. Similarly, the new
6636 "IODeviceLatencyTargetSec=" property has been added, wrapping the new
6637 cgroup v2 io.latency cgroup property for configuring per-service I/O
6640 * systemd now supports the cgroup v2 devices BPF logic, as counterpart
6641 to the cgroup v1 "devices" cgroup controller.
6643 * systemd-escape now is able to combine --unescape with --template. It
6644 also learnt a new option --instance for extracting and unescaping the
6645 instance part of a unit name.
6647 * sd-bus now provides the sd_bus_message_readv() which is similar to
6648 sd_bus_message_read() but takes a va_list object. The pair
6649 sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout()
6650 has been added for configuring the default method call timeout to
6651 use. sd_bus_error_move() may be used to efficiently move the contents
6652 from one sd_bus_error structure to another, invalidating the
6653 source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may
6654 be used to control whether a bus connection object is automatically
6655 flushed when an sd-event loop is exited.
6657 * When processing classic BSD syslog log messages, journald will now
6658 save the original time-stamp string supplied in the new
6659 SYSLOG_TIMESTAMP= journal field. This permits consumers to
6660 reconstruct the original BSD syslog message more correctly.
6662 * StandardOutput=/StandardError= in service files gained support for
6663 new "append:…" parameters, for connecting STDOUT/STDERR of a service
6664 to a file, and appending to it.
6666 * The signal to use as last step of killing of unit processes is now
6667 configurable. Previously it was hard-coded to SIGKILL, which may now
6668 be overridden with the new KillSignal= setting. Note that this is the
6669 signal used when regular termination (i.e. SIGTERM) does not suffice.
6670 Similarly, the signal used when aborting a program in case of a
6671 watchdog timeout may now be configured too (WatchdogSignal=).
6673 * The XDG_SESSION_DESKTOP environment variable may now be configured in
6674 the pam_systemd argument line, using the new desktop= switch. This is
6675 useful to initialize it properly from a display manager without
6676 having to touch C code.
6678 * Most configuration options that previously accepted percentage values
6679 now also accept permille values with the '‰' suffix (instead of '%').
6681 * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for
6684 * systemd-resolved's configuration file resolved.conf gained a new
6685 option ReadEtcHosts= which may be used to turn off processing and
6686 honoring /etc/hosts entries.
6688 * The "--wait" switch may now be passed to "systemctl
6689 is-system-running", in which case the tool will synchronously wait
6690 until the system finished start-up.
6692 * hostnamed gained a new bus call to determine the DMI product UUID.
6694 * On x86-64 systemd will now prefer using the RDRAND processor
6695 instruction over /dev/urandom whenever it requires randomness that
6696 neither has to be crypto-grade nor should be reproducible. This
6697 should substantially reduce the amount of entropy systemd requests
6698 from the kernel during initialization on such systems, though not
6699 reduce it to zero. (Why not zero? systemd still needs to allocate
6700 UUIDs and such uniquely, which require high-quality randomness.)
6702 * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP
6703 tunnels. It also gained a new option ForceDHCPv6PDOtherInformation=
6704 for forcing the "Other Information" bit in IPv6 RA messages. The
6705 bonding logic gained four new options AdActorSystemPriority=,
6706 AdUserPortKey=, AdActorSystem= for configuring various 802.3ad
6707 aspects, and DynamicTransmitLoadBalancing= for enabling dynamic
6708 shuffling of flows. The tunnel logic gained a new
6709 IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid
6710 Deployment. The policy rule logic gained four new options IPProtocol=,
6711 SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained
6712 support for the MulticastToUnicast= option. networkd also gained
6713 support for configuring static IPv4 ARP or IPv6 neighbor entries.
6715 * .preset files (as read by 'systemctl preset') may now be used to
6716 instantiate services.
6718 * /etc/crypttab now understands the sector-size= option to configure
6719 the sector size for an encrypted partition.
6721 * Key material for encrypted disks may now be placed on a formatted
6722 medium, and referenced from /etc/crypttab by the UUID of the file
6723 system, followed by "=" suffixed by the path to the key file.
6725 * The "collect" udev component has been removed without replacement, as
6726 it is neither used nor maintained.
6728 * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=,
6729 LogsDirectory=, ConfigurationDirectory= settings are used in a
6730 service the executed processes will now receive a set of environment
6731 variables containing the full paths of these directories.
6732 Specifically, RUNTIME_DIRECTORY=, STATE_DIRECTORY, CACHE_DIRECTORY,
6733 LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options
6734 are used. Note that these options may be used multiple times per
6735 service in which case the resulting paths will be concatenated and
6736 separated by colons.
6738 * Predictable interface naming has been extended to cover InfiniBand
6739 NICs. They will be exposed with an "ib" prefix.
6741 * tmpfiles.d/ line types may now be suffixed with a '-' character, in
6742 which case the respective line failing is ignored.
6744 * .link files may now be used to configure the equivalent to the
6745 "ethtool advertise" commands.
6747 * The sd-device.h and sd-hwdb.h APIs are now exported, as an
6748 alternative to libudev.h. Previously, the latter was just an internal
6749 wrapper around the former, but now these two APIs are exposed
6752 * sd-id128.h gained a new function sd_id128_get_boot_app_specific()
6753 which calculates an app-specific boot ID similar to how
6754 sd_id128_get_machine_app_specific() generates an app-specific machine
6757 * A new tool systemd-id128 has been added that can be used to determine
6758 and generate various 128bit IDs.
6760 * /etc/os-release gained two new standardized fields DOCUMENTATION_URL=
6763 * systemd-hibernate-resume-generator will now honor the "noresume"
6764 kernel command line option, in which case it will bypass resuming
6765 from any hibernated image.
6767 * The systemd-sleep.conf configuration file gained new options
6768 AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=,
6769 AllowHybridSleep= for prohibiting specific sleep modes even if the
6770 kernel exports them.
6772 * portablectl is now officially supported and has thus moved to
6775 * bootctl learnt the two new commands "set-default" and "set-oneshot"
6776 for setting the default boot loader item to boot to (either
6777 persistently or only for the next boot). This is currently only
6778 compatible with sd-boot, but may be implemented on other boot loaders
6779 too, that follow the boot loader interface. The updated interface is
6780 now documented here:
6782 https://systemd.io/BOOT_LOADER_INTERFACE
6784 * A new kernel command line option systemd.early_core_pattern= is now
6785 understood which may be used to influence the core_pattern PID 1
6786 installs during early boot.
6788 * busctl learnt two new options -j and --json= for outputting method
6789 call replies, properties and monitoring output in JSON.
6791 * journalctl's JSON output now supports simple ANSI coloring as well as
6792 a new "json-seq" mode for generating RFC7464 output.
6794 * Unit files now support the %g/%G specifiers that resolve to the UNIX
6795 group/GID of the service manager runs as, similar to the existing
6796 %u/%U specifiers that resolve to the UNIX user/UID.
6798 * systemd-logind learnt a new global configuration option
6799 UserStopDelaySec= that may be set in logind.conf. It specifies how
6800 long the systemd --user instance shall remain started after a user
6801 logs out. This is useful to speed up repetitive re-connections of the
6802 same user, as it means the user's service manager doesn't have to be
6803 stopped/restarted on each iteration, but can be reused between
6804 subsequent options. This setting defaults to 10s. systemd-logind also
6805 exports two new properties on its Manager D-Bus objects indicating
6806 whether the system's lid is currently closed, and whether the system
6809 * systemd gained support for a generic boot counting logic, which
6810 generically permits automatic reverting to older boot loader entries
6811 if newer updated ones don't work. The boot loader side is implemented
6812 in sd-boot, but is kept open for other boot loaders too. For details
6815 https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT
6817 * The SuccessAction=/FailureAction= unit file settings now learnt two
6818 new parameters: "exit" and "exit-force", which result in immediate
6819 exiting of the service manager, and are only useful in systemd --user
6820 and container environments.
6822 * Unit files gained support for a pair of options
6823 FailureActionExitStatus=/SuccessActionExitStatus= for configuring the
6824 exit status to use as service manager exit status when
6825 SuccessAction=/FailureAction= is set to exit or exit-force.
6827 * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service
6828 options may now be used to configure the log rate limiting applied by
6829 journald per-service.
6831 * systemd-analyze gained a new verb "timespan" for parsing and
6832 normalizing time span values (i.e. strings like "5min 7s 8us").
6834 * systemd-analyze also gained a new verb "security" for analyzing the
6835 security and sand-boxing settings of services in order to determine an
6836 "exposure level" for them, indicating whether a service would benefit
6837 from more sand-boxing options turned on for them.
6839 * "systemd-analyze syscall-filter" will now also show system calls
6840 supported by the local kernel but not included in any of the defined
6843 * .nspawn files now understand the Ephemeral= setting, matching the
6844 --ephemeral command line switch.
6846 * sd-event gained the new APIs sd_event_source_get_floating() and
6847 sd_event_source_set_floating() for controlling whether a specific
6848 event source is "floating", i.e. destroyed along with the even loop
6851 * Unit objects on D-Bus gained a new "Refs" property that lists all
6852 clients that currently have a reference on the unit (to ensure it is
6855 * The JoinControllers= option in system.conf is no longer supported, as
6856 it didn't work correctly, is hard to support properly, is legacy (as
6857 the concept only exists on cgroup v1) and apparently wasn't used.
6859 * Journal messages that are generated whenever a unit enters the failed
6860 state are now tagged with a unique MESSAGE_ID. Similarly, messages
6861 generated whenever a service process exits are now made recognizable,
6862 too. A tagged message is also emitted whenever a unit enters the
6863 "dead" state on success.
6865 * systemd-run gained a new switch --working-directory= for configuring
6866 the working directory of the service to start. A shortcut -d is
6867 equivalent, setting the working directory of the service to the
6868 current working directory of the invoking program. The new --shell
6869 (or just -S) option has been added for invoking the $SHELL of the
6870 caller as a service, and implies --pty --same-dir --wait --collect
6871 --service-type=exec. Or in other words, "systemd-run -S" is now the
6872 quickest way to quickly get an interactive in a fully clean and
6873 well-defined system service context.
6875 * machinectl gained a new verb "import-fs" for importing an OS tree
6876 from a directory. Moreover, when a directory or tarball is imported
6877 and single top-level directory found with the OS itself below the OS
6878 tree is automatically mangled and moved one level up.
6880 * systemd-importd will no longer set up an implicit btrfs loop-back
6881 file system on /var/lib/machines. If one is already set up, it will
6882 continue to be used.
6884 * A new generator "systemd-run-generator" has been added. It will
6885 synthesize a unit from one or more program command lines included in
6886 the kernel command line. This is very useful in container managers
6889 # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"'
6891 This will run "systemd-nspawn" on an image, invoke the specified
6892 command line and immediately shut down the container again, returning
6893 the command line's exit code.
6895 * The block device locking logic is now documented:
6897 https://systemd.io/BLOCK_DEVICE_LOCKING
6899 * loginctl and machinectl now optionally output the various tables in
6900 JSON using the --output= switch. It is our intention to add similar
6901 support to systemctl and all other commands.
6903 * udevadm's query and trigger verb now optionally take a .device unit
6906 * systemd-udevd's network naming logic now understands a new
6907 net.naming-scheme= kernel command line switch, which may be used to
6908 pick a specific version of the naming scheme. This helps stabilizing
6909 interface names even as systemd/udev are updated and the naming logic
6912 * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and
6913 SD_ID128_ALLF to test if a 128bit ID is set to all 0xFF bytes, and to
6914 initialize one to all 0xFF.
6916 * After loading the SELinux policy systemd will now recursively relabel
6917 all files and directories listed in
6918 /run/systemd/relabel-extra.d/*.relabel (which should be simple
6919 newline separated lists of paths) in addition to the ones it already
6920 implicitly relabels in /run, /dev and /sys. After the relabelling is
6921 completed the *.relabel files (and /run/systemd/relabel-extra.d/) are
6922 removed. This is useful to permit initrds (i.e. code running before
6923 the SELinux policy is in effect) to generate files in the host
6924 filesystem safely and ensure that the correct label is applied during
6925 the transition to the host OS.
6927 * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding
6928 mknod() handling in user namespaces. Previously mknod() would always
6929 fail with EPERM in user namespaces. Since 4.18 mknod() will succeed
6930 but device nodes generated that way cannot be opened, and attempts to
6931 open them result in EPERM. This breaks the "graceful fallback" logic
6932 in systemd's PrivateDevices= sand-boxing option. This option is
6933 implemented defensively, so that when systemd detects it runs in a
6934 restricted environment (such as a user namespace, or an environment
6935 where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD)
6936 where device nodes cannot be created the effect of PrivateDevices= is
6937 bypassed (following the logic that 2nd-level sand-boxing is not
6938 essential if the system systemd runs in is itself already sand-boxed
6939 as a whole). This logic breaks with 4.18 in container managers where
6940 user namespacing is used: suddenly PrivateDevices= succeeds setting
6941 up a private /dev/ file system containing devices nodes — but when
6942 these are opened they don't work.
6944 At this point it is recommended that container managers utilizing
6945 user namespaces that intend to run systemd in the payload explicitly
6946 block mknod() with seccomp or similar, so that the graceful fallback
6949 We are very sorry for the breakage and the requirement to change
6950 container configurations for newer kernels. It's purely caused by an
6951 incompatible kernel change. The relevant kernel developers have been
6952 notified about this userspace breakage quickly, but they chose to
6955 * PermissionsStartOnly= setting is deprecated (but is still supported
6956 for backwards compatibility). The same functionality is provided by
6957 the more flexible "+", "!", and "!!" prefixes to ExecStart= and other
6960 * $DBUS_SESSION_BUS_ADDRESS environment variable is not set by
6961 pam_systemd anymore.
6963 * The naming scheme for network devices was changed to always rename
6964 devices, even if they were already renamed by userspace. The "kernel"
6965 policy was changed to only apply as a fallback, if no other naming
6968 * The requirements to build systemd is bumped to meson-0.46 and
6971 Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander
6972 Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson,
6973 Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov,
6974 asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt
6975 Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen
6976 Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius
6977 Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn
6978 Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner,
6979 David Anderson, Davide Cavalca, David Leeds, David Malcolm, David
6980 Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus,
6981 Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters,
6982 Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad,
6983 Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank
6984 Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe
6985 Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit
6986 Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan
6987 Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld,
6988 javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi
6989 Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens,
6990 Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi,
6991 Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry
6992 Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz
6993 Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier,
6994 Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin
6995 Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko
6996 Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck,
6997 Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich,
6998 Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal
6999 Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal
7000 Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby,
7001 Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł
7002 Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller,
7003 Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez,
7004 Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam
7005 Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher,
7006 Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee
7007 (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen
7008 Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim,
7009 Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas
7010 Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias
7011 Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore
7012 Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech
7013 Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward,
7014 Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe,
7015 Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein
7017 — Warsaw, 2018-12-21
7021 * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
7022 builtin will name network interfaces differently than in previous
7023 versions for virtual network interfaces created with SR-IOV and NPAR
7024 and for devices where the PCI network controller device does not have
7025 a slot number associated.
7027 SR-IOV virtual devices are now named based on the name of the parent
7028 interface, with a suffix of "v<N>", where <N> is the virtual device
7029 number. Previously those virtual devices were named as if completely
7032 The ninth and later NPAR virtual devices will be named following the
7033 scheme used for the first eight NPAR partitions. Previously those
7034 devices were not renamed and the kernel default (eth<n>) was used.
7036 "net_id" will also generate names for PCI devices where the PCI
7037 network controller device does not have an associated slot number
7038 itself, but one of its parents does. Previously those devices were
7039 not renamed and the kernel default (eth<n>) was used.
7041 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
7042 systemd-logind.service. Since v235, IPAddressDeny=any has been set to
7043 the unit. So, it is expected that the default behavior of
7044 systemd-logind is not changed. However, if distribution packagers or
7045 administrators disabled or modified IPAddressDeny= setting by a
7046 drop-in config file, then it may be necessary to update the file to
7047 re-enable AF_INET and AF_INET6 to support network user name services,
7050 * When the RestrictNamespaces= unit property is specified multiple
7051 times, then the specified types are merged now. Previously, only the
7052 last assignment was used. So, if distribution packagers or
7053 administrators modified the setting by a drop-in config file, then it
7054 may be necessary to update the file.
7056 * When OnFailure= is used in combination with Restart= on a service
7057 unit, then the specified units will no longer be triggered on
7058 failures that result in restarting. Previously, the specified units
7059 would be activated each time the unit failed, even when the unit was
7060 going to be restarted automatically. This behaviour contradicted the
7061 documentation. With this release the code is adjusted to match the
7064 * systemd-tmpfiles will now print a notice whenever it encounters
7065 tmpfiles.d/ lines referencing the /var/run/ directory. It will
7066 recommend reworking them to use the /run/ directory instead (for
7067 which /var/run/ is simply a symlinked compatibility alias). This way
7068 systemd-tmpfiles can properly detect line conflicts and merge lines
7069 referencing the same file by two paths, without having to access
7072 * systemctl disable/unmask/preset/preset-all cannot be used with
7073 --runtime. Previously this was allowed, but resulted in unintuitive
7074 behaviour that wasn't useful. systemctl disable/unmask will now undo
7075 both runtime and persistent enablement/masking, i.e. it will remove
7076 any relevant symlinks both in /run and /etc.
7078 * Note that all long-running system services shipped with systemd will
7079 now default to a system call allow list (rather than a deny list, as
7080 before). In particular, systemd-udevd will now enforce one too. For
7081 most cases this should be safe, however downstream distributions
7082 which disabled sandboxing of systemd-udevd (specifically the
7083 MountFlags= setting), might want to disable this security feature
7084 too, as the default allow-listing will prohibit all mount, swap,
7085 reboot and clock changing operations from udev rules.
7087 * sd-boot acquired new loader configuration settings to optionally turn
7088 off Windows and MacOS boot partition discovery as well as
7089 reboot-into-firmware menu items. It is also able to pick a better
7090 screen resolution for HiDPI systems, and now provides loader
7091 configuration settings to change the resolution explicitly.
7093 * systemd-resolved now supports DNS-over-TLS. It's still
7094 turned off by default, use DNSOverTLS=opportunistic to turn it on in
7095 resolved.conf. We intend to make this the default as soon as couple
7096 of additional techniques for optimizing the initial latency caused by
7097 establishing a TLS/TCP connection are implemented.
7099 * systemd-resolved.service and systemd-networkd.service now set
7100 DynamicUser=yes. The users systemd-resolve and systemd-network are
7101 not created by systemd-sysusers anymore.
7103 NOTE: This has a chance of breaking nss-ldap and similar NSS modules
7104 that embed a network facing module into any process using getpwuid()
7105 or related call: the dynamic allocation of the user ID for
7106 systemd-resolved.service means the service manager has to check NSS
7107 if the user name is already taken when forking off the service. Since
7108 the user in the common case won't be defined in /etc/passwd the
7109 lookup is likely to trigger nss-ldap which in turn might use NSS to
7110 ask systemd-resolved for hostname lookups. This will hence result in
7111 a deadlock: a user name lookup in order to start
7112 systemd-resolved.service will result in a hostname lookup for which
7113 systemd-resolved.service needs to be started already. There are
7114 multiple ways to work around this problem: pre-allocate the
7115 "systemd-resolve" user on such systems, so that nss-ldap won't be
7116 triggered; or use a different NSS package that doesn't do networking
7117 in-process but provides a local asynchronous name cache; or configure
7118 the NSS package to avoid lookups for UIDs in the range `pkg-config
7119 systemd --variable=dynamicuidmin` … `pkg-config systemd
7120 --variable=dynamicuidmax`, so that it does not consider itself
7121 authoritative for the same UID range systemd allocates dynamic users
7124 * The systemd-resolve tool has been renamed to resolvectl (it also
7125 remains available under the old name, for compatibility), and its
7126 interface is now verb-based, similar in style to the other <xyz>ctl
7127 tools, such as systemctl or loginctl.
7129 * The resolvectl/systemd-resolve tool also provides 'resolvconf'
7130 compatibility. It may be symlinked under the 'resolvconf' name, in
7131 which case it will take arguments and input compatible with the
7132 Debian and FreeBSD resolvconf tool.
7134 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
7135 where the system initially suspends, and after a timeout resumes and
7138 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
7139 set the client will only send a DUID as client identifier.
7141 * The nss-systemd glibc NSS module will now enumerate dynamic users and
7142 groups in effect. Previously, it could resolve UIDs/GIDs to user
7143 names/groups and vice versa, but did not support enumeration.
7145 * journald's Compress= configuration setting now optionally accepts a
7146 byte threshold value. All journal objects larger than this threshold
7147 will be compressed, smaller ones will not. Previously this threshold
7148 was not configurable and set to 512.
7150 * A new system.conf setting NoNewPrivileges= is now available which may
7151 be used to turn off acquisition of new privileges system-wide
7152 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
7153 for all its children). Note that turning this option on means setuid
7154 binaries and file system capabilities lose their special powers.
7155 While turning on this option is a big step towards a more secure
7156 system, doing so is likely to break numerous pre-existing UNIX tools,
7157 in particular su and sudo.
7159 * A new service systemd-time-sync-wait.service has been added. If
7160 enabled it will delay the time-sync.target unit at boot until time
7161 synchronization has been received from the network. This
7162 functionality is useful on systems lacking a local RTC or where it is
7163 acceptable that the boot process shall be delayed by external network
7166 * When hibernating, systemd will now inform the kernel of the image
7167 write offset, on kernels new enough to support this. This means swap
7168 files should work for hibernation now.
7170 * When loading unit files, systemd will now look for drop-in unit files
7171 extensions in additional places. Previously, for a unit file name
7172 "foo-bar-baz.service" it would look for dropin files in
7173 "foo-bar-baz.service.d/*.conf". Now, it will also look in
7174 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
7175 service name truncated after all inner dashes. This scheme allows
7176 writing drop-ins easily that apply to a whole set of unit files at
7177 once. It's particularly useful for mount and slice units (as their
7178 naming is prefix based), but is also useful for service and other
7179 units, for packages that install multiple unit files at once,
7180 following a strict naming regime of beginning the unit file name with
7181 the package's name. Two new specifiers are now supported in unit
7182 files to match this: %j and %J are replaced by the part of the unit
7183 name following the last dash.
7185 * Unit files and other configuration files that support specifier
7186 expansion now understand another three new specifiers: %T and %V will
7187 resolve to /tmp and /var/tmp respectively, or whatever temporary
7188 directory has been set for the calling user. %E will expand to either
7189 /etc (for system units) or $XDG_CONFIG_HOME (for user units).
7191 * The ExecStart= lines of unit files are no longer required to
7192 reference absolute paths. If non-absolute paths are specified the
7193 specified binary name is searched within the service manager's
7194 built-in $PATH, which may be queried with 'systemd-path
7195 search-binaries-default'. It's generally recommended to continue to
7196 use absolute paths for all binaries specified in unit files.
7198 * Units gained a new load state "bad-setting", which is used when a
7199 unit file was loaded, but contained fatal errors which prevent it
7200 from being started (for example, a service unit has been defined
7201 lacking both ExecStart= and ExecStop= lines).
7203 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
7204 support alternative debuggers, for example lldb. The old name
7205 continues to be available however, for compatibility reasons. Use the
7206 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
7207 to pick an alternative debugger instead of the default gdb.
7209 * systemctl and the other tools will now output escape sequences that
7210 generate proper clickable hyperlinks in various terminal emulators
7211 where useful (for example, in the "systemctl status" output you can
7212 now click on the unit file name to quickly open it in the
7213 editor/viewer of your choice). Note that not all terminal emulators
7214 support this functionality yet, but many do. Unfortunately, the
7215 "less" pager doesn't support this yet, hence this functionality is
7216 currently automatically turned off when a pager is started (which
7217 happens quite often due to auto-paging). We hope to remove this
7218 limitation as soon as "less" learns these escape sequences. This new
7219 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
7220 environment variable. For details on these escape sequences see:
7221 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
7223 * networkd's .network files now support a new IPv6MTUBytes= option for
7224 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
7225 option in the [Route] section to configure the MTU to use for
7226 specific routes. It also gained support for configuration of the DHCP
7227 "UserClass" option through the new UserClass= setting. It gained
7228 three new options in the new [CAN] section for configuring CAN
7229 networks. The MULTICAST and ALLMULTI interface flags may now be
7230 controlled explicitly with the new Multicast= and AllMulticast=
7233 * networkd will now automatically make use of the kernel's route
7234 expiration feature, if it is available.
7236 * udevd's .link files now support setting the number of receive and
7237 transmit channels, using the RxChannels=, TxChannels=,
7238 OtherChannels=, CombinedChannels= settings.
7240 * Support for UDPSegmentationOffload= has been removed, given its
7241 limited support in hardware, and waning software support.
7243 * networkd's .netdev files now support creating "netdevsim" interfaces.
7245 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
7246 to query the unit belonging to a specific kernel control group.
7248 * systemd-analyze gained a new verb "cat-config", which may be used to
7249 dump the contents of any configuration file, with all its matching
7250 drop-in files added in, and honouring the usual search and masking
7251 logic applied to systemd configuration files. For example use
7252 "systemd-analyze cat-config systemd/system.conf" to get the complete
7253 system configuration file of systemd how it would be loaded by PID 1
7254 itself. Similar to this, various tools such as systemd-tmpfiles or
7255 systemd-sysusers, gained a new option "--cat-config", which does the
7256 corresponding operation for their own configuration settings. For
7257 example, "systemd-tmpfiles --cat-config" will now output the full
7258 list of tmpfiles.d/ lines in place.
7260 * timedatectl gained three new verbs: "show" shows bus properties of
7261 systemd-timedated, "timesync-status" shows the current NTP
7262 synchronization state of systemd-timesyncd, and "show-timesync"
7263 shows bus properties of systemd-timesyncd.
7265 * systemd-timesyncd gained a bus interface on which it exposes details
7268 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
7269 understood by systemd-timedated. It takes a colon-separated list of
7270 unit names of NTP client services. The list is used by
7271 "timedatectl set-ntp".
7273 * systemd-nspawn gained a new --rlimit= switch for setting initial
7274 resource limits for the container payload. There's a new switch
7275 --hostname= to explicitly override the container's hostname. A new
7276 --no-new-privileges= switch may be used to control the
7277 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
7278 --oom-score-adjust= switch controls the OOM scoring adjustment value
7279 for the payload. The new --cpu-affinity= switch controls the CPU
7280 affinity of the container payload. The new --resolv-conf= switch
7281 allows more detailed control of /etc/resolv.conf handling of the
7282 container. Similarly, the new --timezone= switch allows more detailed
7283 control of /etc/localtime handling of the container.
7285 * systemd-detect-virt gained a new --list switch, which will print a
7286 list of all currently known VM and container environments.
7288 * Support for "Portable Services" has been added, see
7289 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
7290 experimental, but this is expected to change soon. Reflecting this
7291 experimental state, the "portablectl" binary is not installed into
7292 /usr/bin yet. The binary has to be called with the full path
7293 /usr/lib/systemd/portablectl instead.
7295 * journalctl's and systemctl's -o switch now knows a new log output
7296 mode "with-unit". The output it generates is very similar to the
7297 regular "short" mode, but displays the unit name instead of the
7298 syslog tag for each log line. Also, the date is shown with timezone
7299 information. This mode is probably more useful than the classic
7300 "short" output mode for most purposes, except where pixel-perfect
7301 compatibility with classic /var/log/messages formatting is required.
7303 * A new --dump-bus-properties switch has been added to the systemd
7304 binary, which may be used to dump all supported D-Bus properties.
7305 (Options which are still supported, but are deprecated, are *not*
7308 * sd-bus gained a set of new calls:
7309 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
7310 enable/disable the "floating" state of a bus slot object,
7311 i.e. whether the slot object pins the bus it is allocated for into
7312 memory or if the bus slot object gets disconnected when the bus goes
7313 away. sd_bus_open_with_description(),
7314 sd_bus_open_user_with_description(),
7315 sd_bus_open_system_with_description() may be used to allocate bus
7316 objects and set their description string already during allocation.
7318 * sd-event gained support for watching inotify events from the event
7319 loop, in an efficient way, sharing inotify handles between multiple
7320 users. For this a new function sd_event_add_inotify() has been added.
7322 * sd-event and sd-bus gained support for calling special user-supplied
7323 destructor functions for userdata pointers associated with
7324 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
7325 functions sd_bus_slot_set_destroy_callback,
7326 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
7327 sd_bus_track_get_destroy_callback,
7328 sd_event_source_set_destroy_callback,
7329 sd_event_source_get_destroy_callback have been added.
7331 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
7333 * PID 1 will now automatically reschedule .timer units whenever the
7334 local timezone changes. (They previously got rescheduled
7335 automatically when the system clock changed.)
7337 * New documentation has been added to document cgroups delegation,
7338 portable services and the various code quality tools we have set up:
7340 https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md
7341 https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md
7342 https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md
7344 * The Boot Loader Specification has been added to the source tree.
7346 https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md
7348 While moving it into our source tree we have updated it and further
7349 changes are now accepted through the usual github PR workflow.
7351 * pam_systemd will now look for PAM userdata fields systemd.memory_max,
7352 systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
7353 earlier PAM modules. The data in these fields is used to initialize
7354 the session scope's resource properties. Thus external PAM modules
7355 may now configure per-session limits, for example sourced from
7356 external user databases.
7358 * socket units with Accept=yes will now maintain a "refused" counter in
7359 addition to the existing "accepted" counter, counting connections
7360 refused due to the enforced limits.
7362 * The "systemd-path search-binaries-default" command may now be use to
7363 query the default, built-in $PATH PID 1 will pass to the services it
7366 * A new unit file setting PrivateMounts= has been added. It's a boolean
7367 option. If enabled the unit's processes are invoked in their own file
7368 system namespace. Note that this behaviour is also implied if any
7369 other file system namespacing options (such as PrivateTmp=,
7370 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
7371 primarily useful for services that do not use any of the other file
7372 system namespacing options. One such service is systemd-udevd.service
7373 where this is now used by default.
7375 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
7376 when the system is booted in UEFI "secure mode".
7378 * A new unit "system-update-pre.target" is added, which defines an
7379 optional synchronization point for offline system updates, as
7380 implemented by the pre-existing "system-update.target" unit. It
7381 allows ordering services before the service that executes the actual
7382 update process in a generic way.
7384 * Systemd now emits warnings whenever .include syntax is used.
7386 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
7387 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
7388 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
7389 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
7390 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
7391 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
7392 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
7393 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
7394 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
7395 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
7396 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
7397 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
7398 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
7399 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
7400 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
7401 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
7402 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
7403 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
7404 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
7405 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
7406 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
7407 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
7408 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
7409 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
7410 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
7411 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
7412 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
7413 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
7414 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
7416 — Berlin, 2018-06-22
7420 * The MemoryAccounting= unit property now defaults to on. After
7421 discussions with the upstream control group maintainers we learnt
7422 that the negative impact of cgroup memory accounting on current
7423 kernels is finally relatively minimal, so that it should be safe to
7424 enable this by default without affecting system performance. Besides
7425 memory accounting only task accounting is turned on by default, all
7426 other forms of resource accounting (CPU, IO, IP) remain off for now,
7427 because it's not clear yet that their impact is small enough to move
7428 from opt-in to opt-out. We recommend downstreams to leave memory
7429 accounting on by default if kernel 4.14 or higher is primarily
7430 used. On very resource constrained systems or when support for old
7431 kernels is a necessity, -Dmemory-accounting-default=false can be used
7432 to revert this change.
7434 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
7435 %udev_rules_update) and the journal catalog (%journal_catalog_update)
7436 from the upgrade scriptlets of individual packages now do nothing.
7437 Transfiletriggers have been added which will perform those updates
7438 once at the end of the transaction.
7440 Similar transfiletriggers have been added to execute any sysctl.d
7441 and binfmt.d rules. Thus, it should be unnecessary to provide any
7442 scriptlets to execute this configuration from package installation
7445 * systemd-sysusers gained a mode where the configuration to execute is
7446 specified on the command line, but this configuration is not executed
7447 directly, but instead it is merged with the configuration on disk,
7448 and the result is executed. This is useful for package installation
7449 scripts which want to create the user before installing any files on
7450 disk (in case some of those files are owned by that user), while
7451 still allowing local admin overrides.
7453 This functionality is exposed to rpm scriptlets through a new
7454 %sysusers_create_package macro. Old %sysusers_create and
7455 %sysusers_create_inline macros are deprecated.
7457 A transfiletrigger for sysusers.d configuration is now installed,
7458 which means that it should be unnecessary to call systemd-sysusers from
7459 package installation scripts, unless the package installs any files
7460 owned by those newly-created users, in which case
7461 %sysusers_create_package should be used.
7463 * Analogous change has been done for systemd-tmpfiles: it gained a mode
7464 where the command-line configuration is merged with the configuration
7465 on disk. This is exposed as the new %tmpfiles_create_package macro,
7466 and %tmpfiles_create is deprecated. A transfiletrigger is installed
7467 for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
7468 from package installation scripts.
7470 * sysusers.d configuration for a user may now also specify the group
7471 number, in addition to the user number ("u username 123:456"), or
7472 without the user number ("u username -:456").
7474 * Configution items for systemd-sysusers can now be specified as
7475 positional arguments when the new --inline switch is used.
7477 * The login shell of users created through sysusers.d may now be
7478 specified (previously, it was always /bin/sh for root and
7479 /sbin/nologin for other users).
7481 * systemd-analyze gained a new --global switch to look at global user
7482 configuration. It also gained a unit-paths verb to list the unit load
7483 paths that are compiled into systemd (which can be used with
7484 --systemd, --user, or --global).
7486 * udevadm trigger gained a new --settle/-w option to wait for any
7487 triggered events to finish (but just those, and not any other events
7488 which are triggered meanwhile).
7490 * The action that systemd-logind takes when the lid is closed and the
7491 machine is connected to external power can now be configured using
7492 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
7493 was determined by HandleLidSwitch=, and, for backwards compatibility,
7494 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
7496 * journalctl will periodically call sd_journal_process() to make it
7497 resilient against inotify queue overruns when journal files are
7498 rotated very quickly.
7500 * Two new functions in libsystemd — sd_bus_get_n_queued_read and
7501 sd_bus_get_n_queued_write — may be used to check the number of
7502 pending bus messages.
7504 * systemd gained a new
7505 org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
7506 which can be used to migrate foreign processes to scope and service
7507 units. The primary user for this new API is systemd itself: the
7508 systemd --user instance uses this call of the systemd --system
7509 instance to migrate processes if it itself gets the request to
7510 migrate processes and the kernel refuses this due to access
7511 restrictions. Thanks to this "systemd-run --scope --user …" works
7512 again in pure cgroup v2 environments when invoked from the user
7515 * A new TemporaryFileSystem= setting can be used to mask out part of
7516 the real file system tree with tmpfs mounts. This may be combined
7517 with BindPaths= and BindReadOnlyPaths= to hide files or directories
7518 not relevant to the unit, while still allowing some paths lower in
7519 the tree to be accessed.
7521 ProtectHome=tmpfs may now be used to hide user home and runtime
7522 directories from units, in a way that is mostly equivalent to
7523 "TemporaryFileSystem=/home /run/user /root".
7525 * Non-service units are now started with KeyringMode=shared by default.
7526 This means that mount and swapon and other mount tools have access
7527 to keys in the main keyring.
7529 * /sys/fs/bpf is now mounted automatically.
7531 * QNX virtualization is now detected by systemd-detect-virt and may
7532 be used in ConditionVirtualization=.
7534 * IPAccounting= may now be enabled also for slice units.
7536 * A new -Dsplit-bin= build configuration switch may be used to specify
7537 whether bin and sbin directories are merged, or if they should be
7538 included separately in $PATH and various listings of executable
7539 directories. The build configuration scripts will try to autodetect
7540 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
7541 system, but distributions are encouraged to configure this
7544 * A new -Dok-color= build configuration switch may be used to change
7545 the colour of "OK" status messages.
7547 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
7548 PrivateNetwork=yes was buggy in previous versions of systemd. This
7549 means that after the upgrade and daemon-reexec, any such units must
7552 * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
7553 will not exclude read-only files owned by root from cleanup.
7555 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
7556 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
7557 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
7558 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
7559 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
7560 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
7561 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
7562 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
7563 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
7564 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
7565 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
7566 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
7567 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
7568 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
7569 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
7570 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
7572 — Warsaw, 2018-03-05
7576 * Some keyboards come with a zoom see-saw or rocker which until now got
7577 mapped to the Linux "zoomin/out" keys in hwdb. However, these
7578 keycodes are not recognized by any major desktop. They now produce
7579 Up/Down key events so that they can be used for scrolling.
7581 * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
7582 slightly: previously, if an argument was specified for lines of this
7583 type (i.e. the right-most column was set) this string was appended to
7584 existing files each time systemd-tmpfiles was run. This behaviour was
7585 different from what the documentation said, and not particularly
7586 useful, as repeated systemd-tmpfiles invocations would not be
7587 idempotent and grow such files without bounds. With this release
7588 behaviour has been altered to match what the documentation says:
7589 lines of this type only have an effect if the indicated files don't
7590 exist yet, and only then the argument string is written to the file.
7592 * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
7593 systemd-tmpfiles behaviour: previously, read-only files owned by root
7594 were always excluded from the file "aging" algorithm (i.e. the
7595 automatic clean-up of directories like /tmp based on
7596 atime/mtime/ctime). We intend to drop this restriction, and age files
7597 by default even when owned by root and read-only. This behaviour was
7598 inherited from older tools, but there have been requests to remove
7599 it, and it's not obvious why this restriction was made in the first
7600 place. Please speak up now, if you are aware of software that requires
7601 this behaviour, otherwise we'll remove the restriction in v238.
7603 * A new environment variable $SYSTEMD_OFFLINE is now understood by
7604 systemctl. It takes a boolean argument. If on, systemctl assumes it
7605 operates on an "offline" OS tree, and will not attempt to talk to the
7606 service manager. Previously, this mode was implicitly enabled if a
7607 chroot() environment was detected, and this new environment variable
7608 now provides explicit control.
7610 * .path and .socket units may now be created transiently, too.
7611 Previously only service, mount, automount and timer units were
7612 supported as transient units. The systemd-run tool has been updated
7613 to expose this new functionality, you may hence use it now to bind
7614 arbitrary commands to path or socket activation on-the-fly from the
7615 command line. Moreover, almost all properties are now exposed for the
7616 unit types that already supported transient operation.
7618 * The systemd-mount command gained support for a new --owner= parameter
7619 which takes a user name, which is then resolved and included in uid=
7620 and gid= mount options string of the file system to mount.
7622 * A new unit condition ConditionControlGroupController= has been added
7623 that checks whether a specific cgroup controller is available.
7625 * Unit files, udev's .link files, and systemd-networkd's .netdev and
7626 .network files all gained support for a new condition
7627 ConditionKernelVersion= for checking against specific kernel
7630 * In systemd-networkd, the [IPVLAN] section in .netdev files gained
7631 support for configuring device flags in the Flags= setting. In the
7632 same files, the [Tunnel] section gained support for configuring
7633 AllowLocalRemote=. The [Route] section in .network files gained
7634 support for configuring InitialCongestionWindow=,
7635 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
7636 understands RapidCommit=.
7638 * systemd-networkd's DHCPv6 support gained support for Prefix
7641 * sd-bus gained support for a new "watch-bind" feature. When this
7642 feature is enabled, an sd_bus connection may be set up to connect to
7643 an AF_UNIX socket in the file system as soon as it is created. This
7644 functionality is useful for writing early-boot services that
7645 automatically connect to the system bus as soon as it is started,
7646 without ugly time-based polling. systemd-networkd and
7647 systemd-resolved have been updated to make use of this
7648 functionality. busctl exposes this functionality in a new
7649 --watch-bind= command line switch.
7651 * sd-bus will now optionally synthesize a local "Connected" signal as
7652 soon as a D-Bus connection is set up fully. This message mirrors the
7653 already existing "Disconnected" signal which is synthesized when the
7654 connection is terminated. This signal is generally useful but
7655 particularly handy in combination with the "watch-bind" feature
7656 described above. Synthesizing of this message has to be requested
7657 explicitly through the new API call sd_bus_set_connected_signal(). In
7658 addition a new call sd_bus_is_ready() has been added that checks
7659 whether a connection is fully set up (i.e. between the "Connected" and
7660 "Disconnected" signals).
7662 * sd-bus gained two new calls sd_bus_request_name_async() and
7663 sd_bus_release_name_async() for asynchronously registering bus
7664 names. Similar, there is now sd_bus_add_match_async() for installing
7665 a signal match asynchronously. All of systemd's own services have
7666 been updated to make use of these calls. Doing these operations
7667 asynchronously has two benefits: it reduces the risk of deadlocks in
7668 case of cyclic dependencies between bus services, and it speeds up
7669 service initialization since synchronization points for bus
7670 round-trips are removed.
7672 * sd-bus gained two new calls sd_bus_match_signal() and
7673 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
7674 and sd_bus_add_match_async() but instead of taking a D-Bus match
7675 string take match fields as normal function parameters.
7677 * sd-bus gained two new calls sd_bus_set_sender() and
7678 sd_bus_message_set_sender() for setting the sender name of outgoing
7679 messages (either for all outgoing messages or for just one specific
7680 one). These calls are only useful in direct connections as on
7681 brokered connections the broker fills in the sender anyway,
7682 overwriting whatever the client filled in.
7684 * sd-event gained a new pseudo-handle that may be specified on all API
7685 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
7686 used this refers to the default event loop object of the calling
7687 thread. Note however that this does not implicitly allocate one —
7688 which has to be done prior by using sd_event_default(). Similarly
7689 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
7690 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
7691 to the default bus of the specified type of the calling thread. Here
7692 too this does not implicitly allocate bus connection objects, this
7693 has to be done prior with sd_bus_default() and friends.
7695 * sd-event gained a new call pair
7696 sd_event_source_{get|set}_io_fd_own(). This may be used to request
7697 automatic closure of the file descriptor an IO event source watches
7698 when the event source is destroyed.
7700 * systemd-networkd gained support for natively configuring WireGuard
7703 * In previous versions systemd synthesized user records both for the
7704 "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
7705 internally. In order to simplify distribution-wide renames of the
7706 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
7707 new transitional flag file has been added: if
7708 /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
7709 user and group record within the systemd codebase is disabled.
7711 * systemd-notify gained a new --uid= option for selecting the source
7712 user/UID to use for notification messages sent to the service
7715 * journalctl gained a new --grep= option to list only entries in which
7716 the message matches a certain pattern. By default matching is case
7717 insensitive if the pattern is lowercase, and case sensitive
7718 otherwise. Option --case-sensitive=yes|no can be used to override
7719 this an specify case sensitivity or case insensitivity.
7721 * There's now a "systemd-analyze service-watchdogs" command for printing
7722 the current state of the service runtime watchdog, and optionally
7723 enabling or disabling the per-service watchdogs system-wide if given a
7724 boolean argument (i.e. the concept you configure in WatchdogSec=), for
7725 debugging purposes. There's also a kernel command line option
7726 systemd.service_watchdogs= for controlling the same.
7728 * Two new "log-level" and "log-target" options for systemd-analyze were
7729 added that merge the now deprecated get-log-level, set-log-level and
7730 get-log-target, set-log-target pairs. The deprecated options are still
7731 understood for backwards compatibility. The two new options print the
7732 current value when no arguments are given, and set them when a
7733 level/target is given as an argument.
7735 * sysusers.d's "u" lines now optionally accept both a UID and a GID
7736 specification, separated by a ":" character, in order to create users
7737 where UID and GID do not match.
7739 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
7740 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
7741 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
7742 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
7743 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
7744 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
7745 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
7746 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
7747 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
7748 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
7749 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
7750 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
7751 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
7752 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
7753 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
7754 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
7755 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
7756 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
7757 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
7758 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
7765 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
7766 in v235 has been extended to also set the dummy.ko module option
7767 numdummies=0, preventing the kernel from automatically creating
7768 dummy0. All dummy interfaces must now be explicitly created.
7770 * Unknown '%' specifiers in configuration files are now rejected. This
7771 applies to units and tmpfiles.d configuration. Any percent characters
7772 that are followed by a letter or digit that are not supposed to be
7773 interpreted as the beginning of a specifier should be escaped by
7774 doubling ("%%"). (So "size=5%" is still accepted, as well as
7775 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
7776 valid specifiers today.)
7778 * systemd-resolved now maintains a new dynamic
7779 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
7780 recommended to make /etc/resolv.conf a symlink to it. This file
7781 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
7782 includes dynamically acquired search domains, achieving more correct
7783 DNS resolution by software that bypasses local DNS APIs such as NSS.
7785 * The "uaccess" udev tag has been dropped from /dev/kvm and
7786 /dev/dri/renderD*. These devices now have the 0666 permissions by
7787 default (but this may be changed at build-time). /dev/dri/renderD*
7788 will now be owned by the "render" group along with /dev/kfd.
7790 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
7791 systemd-journal-gatewayd.service and
7792 systemd-journal-upload.service. This means "nss-systemd" must be
7793 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
7794 services are resolved properly.
7796 * In /etc/fstab two new mount options are now understood:
7797 x-systemd.makefs and x-systemd.growfs. The former has the effect that
7798 the configured file system is formatted before it is mounted, the
7799 latter that the file system is resized to the full block device size
7800 after it is mounted (i.e. if the file system is smaller than the
7801 partition it resides on, it's grown). This is similar to the fsck
7802 logic in /etc/fstab, and pulls in systemd-makefs@.service and
7803 systemd-growfs@.service as necessary, similar to
7804 systemd-fsck@.service. Resizing is currently only supported on ext4
7807 * In systemd-networkd, the IPv6 RA logic now optionally may announce
7808 DNS server and domain information.
7810 * Support for the LUKS2 on-disk format for encrypted partitions has
7811 been added. This requires libcryptsetup2 during compilation and
7814 * The systemd --user instance will now signal "readiness" when its
7815 basic.target unit has been reached, instead of when the run queue ran
7816 empty for the first time.
7818 * Tmpfiles.d with user configuration are now also supported.
7819 systemd-tmpfiles gained a new --user switch, and snippets placed in
7820 ~/.config/user-tmpfiles.d/ and corresponding directories will be
7821 executed by systemd-tmpfiles --user running in the new
7822 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
7823 running in the user session.
7825 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
7826 %S resolves to the top-level state directory (/var/lib for the system
7827 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
7828 top-level cache directory (/var/cache for the system instance,
7829 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
7830 logs directory (/var/log for the system instance,
7831 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
7832 existing %t specifier, that resolves to the top-level runtime
7833 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
7836 * journalctl learnt a new parameter --output-fields= for limiting the
7837 set of journal fields to output in verbose and JSON output modes.
7839 * systemd-timesyncd's configuration file gained a new option
7840 RootDistanceMaxSec= for setting the maximum root distance of servers
7841 it'll use, as well as the new options PollIntervalMinSec= and
7842 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
7844 * bootctl gained a new command "list" for listing all available boot
7845 menu items on systems that follow the boot loader specification.
7847 * systemctl gained a new --dry-run switch that shows what would be done
7848 instead of doing it, and is currently supported by the shutdown and
7851 * ConditionSecurity= can now detect the TOMOYO security module.
7853 * Unit file [Install] sections are now also respected in unit drop-in
7854 files. This is intended to be used by drop-ins under /usr/lib/.
7856 * systemd-firstboot may now also set the initial keyboard mapping.
7858 * Udev "changed" events for devices which are exposed as systemd
7859 .device units are now propagated to units specified in
7860 ReloadPropagatedFrom= as reload requests.
7862 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
7863 unit template name (i.e. a name in the form of 'foobar@.service',
7864 without the instance component between the '@' and - the '.'), then
7865 the escaped sysfs path of the device is automatically used as the
7868 * SystemCallFilter= in unit files has been extended so that an "errno"
7869 can be specified individually for each system call. Example:
7870 SystemCallFilter=~uname:EILSEQ.
7872 * The cgroup delegation logic has been substantially updated. Delegate=
7873 now optionally takes a list of controllers (instead of a boolean, as
7874 before), which lists the controllers to delegate at least.
7876 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
7878 * A new LogLevelMax= setting configures the maximum log level any
7879 process of the service may log at (i.e. anything with a lesser
7880 priority than what is specified is automatically dropped). A new
7881 LogExtraFields= setting allows configuration of additional journal
7882 fields to attach to all log records generated by any of the unit's
7885 * New StandardInputData= and StandardInputText= settings along with the
7886 new option StandardInput=data may be used to configure textual or
7887 binary data that shall be passed to the executed service process via
7888 standard input, encoded in-line in the unit file.
7890 * StandardInput=, StandardOutput= and StandardError= may now be used to
7891 connect stdin/stdout/stderr of executed processes directly with a
7892 file or AF_UNIX socket in the file system, using the new "file:" option.
7894 * A new unit file option CollectMode= has been added, that allows
7895 tweaking the garbage collection logic for units. It may be used to
7896 tell systemd to garbage collect units that have failed automatically
7897 (normally it only GCs units that exited successfully). systemd-run
7898 and systemd-mount expose this new functionality with a new -G option.
7900 * "machinectl bind" may now be used to bind mount non-directories
7901 (i.e. regularfiles, devices, fifos, sockets).
7903 * systemd-analyze gained a new verb "calendar" for validating and
7904 testing calendar time specifications to use for OnCalendar= in timer
7905 units. Besides validating the expression it will calculate the next
7906 time the specified expression would elapse.
7908 * In addition to the pre-existing FailureAction= unit file setting
7909 there's now SuccessAction=, for configuring a shutdown action to
7910 execute when a unit completes successfully. This is useful in
7911 particular inside containers that shall terminate after some workload
7912 has been completed. Also, both options are now supported for all unit
7913 types, not just services.
7915 * networkds's IP rule support gained two new options
7916 IncomingInterface= and OutgoingInterface= for configuring the incoming
7917 and outgoing interfaces of configured rules. systemd-networkd also
7918 gained support for "vxcan" network devices.
7920 * networkd gained a new setting RequiredForOnline=, taking a
7921 boolean. If set, systemd-wait-online will take it into consideration
7922 when determining that the system is up, otherwise it will ignore the
7923 interface for this purpose.
7925 * The sd_notify() protocol gained support for a new operation: with
7926 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
7927 store again, ahead of POLLHUP or POLLERR when they are removed
7930 * A new document doc/UIDS-GIDS.md has been added to the source tree,
7931 that documents the UID/GID range and assignment assumptions and
7932 requirements of systemd.
7934 * The watchdog device PID 1 will ping may now be configured through the
7935 WatchdogDevice= configuration file setting, or by setting the
7936 systemd.watchdog_service= kernel commandline option.
7938 * systemd-resolved's gained support for registering DNS-SD services on
7939 the local network using MulticastDNS. Services may either be
7940 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
7941 the same dir below /run, /usr/lib), or through its D-Bus API.
7943 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
7944 extend the effective start, runtime, and stop time. The service must
7945 continue to send EXTEND_TIMEOUT_USEC within the period specified to
7946 prevent the service manager from making the service as timedout.
7948 * systemd-resolved's DNSSEC support gained support for RFC 8080
7949 (Ed25519 keys and signatures).
7951 * The systemd-resolve command line tool gained a new set of options
7952 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
7953 --set-nta= and --revert to configure per-interface DNS configuration
7954 dynamically during runtime. It's useful for pushing DNS information
7955 into systemd-resolved from DNS hook scripts that various interface
7956 managing software supports (such as pppd).
7958 * systemd-nspawn gained a new --network-namespace-path= command line
7959 option, which may be used to make a container join an existing
7960 network namespace, by specifying a path to a "netns" file.
7962 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
7963 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
7964 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
7965 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
7966 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
7967 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
7968 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
7969 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
7970 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
7971 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
7972 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
7973 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
7974 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
7975 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
7976 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
7977 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
7978 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
7979 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
7980 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
7981 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
7982 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
7983 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
7984 Jędrzejewski-Szmek, Zeal Jagannatha
7986 — Berlin, 2017-12-14
7990 * INCOMPATIBILITY: systemd-logind.service and other long-running
7991 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
7992 communication with the outside. This generally improves security of
7993 the system, and is in almost all cases a safe and good choice, as
7994 these services do not and should not provide any network-facing
7995 functionality. However, systemd-logind uses the glibc NSS API to
7996 query the user database. This creates problems on systems where NSS
7997 is set up to directly consult network services for user database
7998 lookups. In particular, this creates incompatibilities with the
7999 "nss-nis" module, which attempts to directly contact the NIS/YP
8000 network servers it is configured for, and will now consistently
8001 fail. In such cases, it is possible to turn off IP sandboxing for
8002 systemd-logind.service (set IPAddressDeny= in its [Service] section
8003 to the empty string, via a .d/ unit file drop-in). Downstream
8004 distributions might want to update their nss-nis packaging to include
8005 such a drop-in snippet, accordingly, to hide this incompatibility
8006 from the user. Another option is to make use of glibc's nscd service
8007 to proxy such network requests through a privilege-separated, minimal
8008 local caching daemon, or to switch to more modern technologies such
8009 sssd, whose NSS hook-ups generally do not involve direct network
8010 access. In general, we think it's definitely time to question the
8011 implementation choices of nss-nis, i.e. whether it's a good idea
8012 today to embed a network-facing loadable module into all local
8013 processes that need to query the user database, including the most
8014 trivial and benign ones, such as "ls". For more details about
8015 IPAddressDeny= see below.
8017 * A new modprobe.d drop-in is now shipped by default that sets the
8018 bonding module option max_bonds=0. This overrides the kernel default,
8019 to avoid conflicts and ambiguity as to whether or not bond0 should be
8020 managed by systemd-networkd or not. This resolves multiple issues
8021 with bond0 properties not being applied, when bond0 is configured
8022 with systemd-networkd. Distributors may choose to not package this,
8023 however in that case users will be prevented from correctly managing
8024 bond0 interface using systemd-networkd.
8026 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
8027 which print the logging level and target of the system manager. They
8028 complement the existing "set-log-level" and "set-log-target" verbs
8029 used to change those values.
8031 * journald.conf gained a new boolean setting ReadKMsg= which defaults
8032 to on. If turned off kernel log messages will not be read by
8033 systemd-journald or included in the logs. It also gained a new
8034 setting LineMax= for configuring the maximum line length in
8035 STDOUT/STDERR log streams. The new default for this value is 48K, up
8036 from the previous hardcoded 2048.
8038 * A new unit setting RuntimeDirectoryPreserve= has been added, which
8039 allows more detailed control of what to do with a runtime directory
8040 configured with RuntimeDirectory= (i.e. a directory below /run or
8041 $XDG_RUNTIME_DIR) after a unit is stopped.
8043 * The RuntimeDirectory= setting for units gained support for creating
8044 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
8045 one top-level directory.
8047 * Units gained new options StateDirectory=, CacheDirectory=,
8048 LogsDirectory= and ConfigurationDirectory= which are closely related
8049 to RuntimeDirectory= but manage per-service directories below
8050 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
8051 possible to write unit files which when activated automatically gain
8052 properly owned service specific directories in these locations, thus
8053 making unit files self-contained and increasing compatibility with
8054 stateless systems and factory reset where /etc or /var are
8055 unpopulated at boot. Matching these new settings there's also
8056 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
8057 ConfigurationDirectoryMode= for configuring the access mode of these
8058 directories. These settings are particularly useful in combination
8059 with DynamicUser=yes as they provide secure, properly-owned,
8060 writable, and stateful locations for storage, excluded from the
8061 sandbox that such services live in otherwise.
8063 * Automake support has been removed from this release. systemd is now
8066 * systemd-journald will now aggressively cache client metadata during
8067 runtime, speeding up log write performance under pressure. This comes
8068 at a small price though: as much of the metadata is read
8069 asynchronously from /proc/ (and isn't implicitly attached to log
8070 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
8071 metadata stored alongside a log entry might be slightly
8072 out-of-date. Previously it could only be slightly newer than the log
8073 message. The time window is small however, and given that the kernel
8074 is unlikely to be improved anytime soon in this regard, this appears
8077 * nss-myhostname/systemd-resolved will now by default synthesize an
8078 A/AAAA resource record for the "_gateway" hostname, pointing to the
8079 current default IP gateway. Previously it did that for the "gateway"
8080 name, hampering adoption, as some distributions wanted to leave that
8081 hostname open for local use. The old behaviour may still be
8082 requested at build time.
8084 * systemd-networkd's [Address] section in .network files gained a new
8085 Scope= setting for configuring the IP address scope. The [Network]
8086 section gained a new boolean setting ConfigureWithoutCarrier= that
8087 tells systemd-networkd to ignore link sensing when configuring the
8088 device. The [DHCP] section gained a new Anonymize= boolean option for
8089 turning on a number of options suggested in RFC 7844. A new
8090 [RoutingPolicyRule] section has been added for configuring the IP
8091 routing policy. The [Route] section has gained support for a new
8092 Type= setting which permits configuring
8093 blackhole/unreachable/prohibit routes.
8095 * The [VRF] section in .netdev files gained a new Table= setting for
8096 configuring the routing table to use. The [Tunnel] section gained a
8097 new Independent= boolean field for configuring tunnels independent of
8098 an underlying network interface. The [Bridge] section gained a new
8099 GroupForwardMask= option for configuration of propagation of link
8100 local frames between bridge ports.
8102 * The WakeOnLan= setting in .link files gained support for a number of
8103 new modes. A new TCP6SegmentationOffload= setting has been added for
8104 configuring TCP/IPv6 hardware segmentation offload.
8106 * The IPv6 RA sender implementation may now optionally send out RDNSS
8107 and RDNSSL records to supply DNS configuration to peers.
8109 * systemd-nspawn gained support for a new --system-call-filter= command
8110 line option for adding and removing entries in the default system
8111 call filter it applies. Moreover systemd-nspawn has been changed to
8112 implement a system call allow list instead of a deny list.
8114 * systemd-run gained support for a new --pipe command line option. If
8115 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
8116 are directly passed on to the activated transient service
8117 executable. This allows invoking arbitrary processes as systemd
8118 services (for example to take benefit of dependency management,
8119 accounting management, resource management or log management that is
8120 done automatically for services) — while still allowing them to be
8121 integrated in a classic UNIX shell pipeline.
8123 * When a service sends RELOAD=1 via sd_notify() and reload propagation
8124 using ReloadPropagationTo= is configured, a reload is now propagated
8125 to configured units. (Previously this was only done on explicitly
8126 requested reloads, using "systemctl reload" or an equivalent
8129 * For each service unit a restart counter is now kept: it is increased
8130 each time the service is restarted due to Restart=, and may be
8131 queried using "systemctl show -p NRestarts …".
8133 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
8134 @signal and @timer have been added, for usage with SystemCallFilter=
8135 in unit files and the new --system-call-filter= command line option
8136 of systemd-nspawn (see above).
8138 * ExecStart= lines in unit files gained two new modifiers: when a
8139 command line is prefixed with "!" the command will be executed as
8140 configured, except for the credentials applied by
8141 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
8142 "+", but does still apply namespacing options unlike "+". There's
8143 also "!!" now, which is mostly identical, but becomes a NOP on
8144 systems that support ambient capabilities. This is useful to write
8145 unit files that work with ambient capabilities where possible but
8146 automatically fall back to traditional privilege dropping mechanisms
8147 on systems where this is not supported.
8149 * ListenNetlink= settings in socket units now support RDMA netlink
8152 * A new unit file setting LockPersonality= has been added which permits
8153 locking down the chosen execution domain ("personality") of a service
8156 * A new special target "getty-pre.target" has been added, which is
8157 ordered before all text logins, and may be used to order services
8158 before textual logins acquire access to the console.
8160 * systemd will now attempt to load the virtio-rng.ko kernel module very
8161 early on if a VM environment supporting this is detected. This should
8162 improve entropy during early boot in virtualized environments.
8164 * A _netdev option is now supported in /etc/crypttab that operates in a
8165 similar way as the same option in /etc/fstab: it permits configuring
8166 encrypted devices that need to be ordered after the network is up.
8167 Following this logic, two new special targets
8168 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
8169 added that are to cryptsetup.target what remote-fs.target and
8170 remote-fs-pre.target are to local-fs.target.
8172 * Service units gained a new UnsetEnvironment= setting which permits
8173 unsetting specific environment variables for services that are
8174 normally passed to it (for example in order to mask out locale
8175 settings for specific services that can't deal with it).
8177 * Units acquired a new boolean option IPAccounting=. When turned on, IP
8178 traffic accounting (packet count as well as byte count) is done for
8179 the service, and shown as part of "systemctl status" or "systemd-run
8182 * Service units acquired two new options IPAddressAllow= and
8183 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
8184 for configuring a simple IP access control list for all sockets of
8185 the unit. These options are available also on .slice and .socket
8186 units, permitting flexible access list configuration for individual
8187 services as well as groups of services (as defined by a slice unit),
8188 including system-wide. Note that IP ACLs configured this way are
8189 enforced on every single IPv4 and IPv6 socket created by any process
8190 of the service unit, and apply to ingress as well as egress traffic.
8192 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
8193 structured log message is generated each time the unit is stopped,
8194 containing information about the consumed resources of this
8197 * A new setting KeyringMode= has been added to unit files, which may be
8198 used to control how the kernel keyring is set up for executed
8201 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
8202 "systemctl kexec" and "systemctl exit" are now always asynchronous in
8203 behaviour (that is: these commands return immediately after the
8204 operation was enqueued instead of waiting for the operation to
8205 complete). Previously, "systemctl poweroff" and "systemctl reboot"
8206 were asynchronous on systems using systemd-logind (i.e. almost
8207 always, and like they were on sysvinit), and the other three commands
8208 were unconditionally synchronous. With this release this is cleaned
8209 up, and callers will see the same asynchronous behaviour on all
8210 systems for all five operations.
8212 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
8215 * .timer units now accept calendar specifications in other timezones
8216 than UTC or the local timezone.
8218 * The tmpfiles snippet var.conf has been changed to create
8219 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
8220 the "utmp" group already, and it appears to be generally understood
8221 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
8222 databases. Previously this was implemented correctly for all these
8223 databases excepts btmp, which has been opened up like this now
8224 too. Note that while the other databases are world-readable
8225 (i.e. 0644), btmp is not and remains more restrictive.
8227 * The systemd-resolve tool gained a new --reset-server-features
8228 switch. When invoked like this systemd-resolved will forget
8229 everything it learnt about the features supported by the configured
8230 upstream DNS servers, and restarts the feature probing logic on the
8231 next resolver look-up for them at the highest feature level
8234 * The status dump systemd-resolved sends to the logs upon receiving
8235 SIGUSR1 now also includes information about all DNS servers it is
8236 configured to use, and the features levels it probed for them.
8238 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
8239 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
8240 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
8241 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
8242 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
8243 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
8244 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
8245 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
8246 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
8247 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
8248 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
8249 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
8250 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
8251 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
8252 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
8253 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
8254 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
8255 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
8256 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
8257 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
8259 — Berlin, 2017-10-06
8263 * Meson is now supported as build system in addition to Automake. It is
8264 our plan to remove Automake in one of our next releases, so that
8265 Meson becomes our exclusive build system. Hence, please start using
8266 the Meson build system in your downstream packaging. There's plenty
8267 of documentation around how to use Meson, the extremely brief
8270 ./autogen.sh && ./configure && make && sudo make install
8274 meson build && ninja -C build && sudo ninja -C build install
8276 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
8277 which permits configuring a timeout on the time a job is
8278 running. This is particularly useful for setting timeouts on jobs for
8281 * Unit files gained two new options ConditionUser= and ConditionGroup=
8282 for conditionalizing units based on the identity of the user/group
8283 running a systemd user instance.
8285 * systemd-networkd now understands a new FlowLabel= setting in the
8286 [VXLAN] section of .network files, as well as a Priority= in
8287 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
8288 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
8289 gained support for configuration of GENEVE links, and IPv6 address
8290 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
8292 * .link files now understand a new Port= setting.
8294 * systemd-networkd's DHCP support gained support for DHCP option 119
8295 (domain search list).
8297 * systemd-networkd gained support for serving IPv6 address ranges using
8298 the Router Advertisement protocol. The new .network configuration
8299 section [IPv6Prefix] may be used to configure the ranges to
8300 serve. This is implemented based on a new, minimal, native server
8301 implementation of RA.
8303 * journalctl's --output= switch gained support for a new parameter
8304 "short-iso-precise" for a mode where timestamps are shown as precise
8307 * systemd-udevd's "net_id" builtin may now generate stable network
8308 interface names from IBM PowerVM VIO devices as well as ACPI platform
8311 * MulticastDNS support in systemd-resolved may now be explicitly
8312 enabled/disabled using the new MulticastDNS= configuration file
8315 * systemd-resolved may now optionally use libidn2 instead of the libidn
8316 for processing internationalized domain names. Support for libidn2
8317 should be considered experimental and should not be enabled by
8320 * "machinectl pull-tar" and related call may now do verification of
8321 downloaded images using SUSE-style .sha256 checksum files in addition
8322 to the already existing support for validating using Ubuntu-style
8325 * sd-bus gained support for a new sd_bus_message_appendv() call which
8326 is va_list equivalent of sd_bus_message_append().
8328 * sd-boot gained support for validating images using SHIM/MOK.
8330 * The SMACK code learnt support for "onlycap".
8332 * systemd-mount --umount is now much smarter in figuring out how to
8333 properly unmount a device given its mount or device path.
8335 * The code to call libnss_dns as a fallback from libnss_resolve when
8336 the communication with systemd-resolved fails was removed. This
8337 fallback was redundant and interfered with the [!UNAVAIL=return]
8338 suffix. See nss-resolve(8) for the recommended configuration.
8340 * systemd-logind may now be restarted without losing state. It stores
8341 the file descriptors for devices it manages in the system manager
8342 using the FDSTORE= mechanism. Please note that further changes in
8343 other components may be required to make use of this (for example
8344 Xorg has code to listen for stops of systemd-logind and terminate
8345 itself when logind is stopped or restarted, in order to avoid using
8346 stale file descriptors for graphical devices, which is now
8347 counterproductive and must be reverted in order for restarts of
8348 systemd-logind to be safe. See
8349 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
8351 * All kernel-install plugins are called with the environment variable
8352 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
8353 /etc/machine-id. If the machine ID could not be determined,
8354 $KERNEL_INSTALL_MACHINE_ID will be empty. Plugins should not put
8355 anything in the entry directory (passed as the second argument) if
8356 $KERNEL_INSTALL_MACHINE_ID is empty. For backwards compatibility, a
8357 temporary directory is passed as the entry directory and removed
8358 after all the plugins exit.
8360 * If KERNEL_INSTALL_MACHINE_ID is set in /etc/machine-info, kernel-install
8361 will now use its value as the machine ID instead of the machine ID
8362 from /etc/machine-id. If KERNEL_INSTALL_MACHINE_ID isn't set in
8363 /etc/machine-info and no machine ID is set in /etc/machine-id,
8364 kernel-install will try to store the current machine ID there as
8365 KERNEL_INSTALL_MACHINE_ID. If there is no machine ID, kernel-install
8366 will generate a new UUID, store it in /etc/machine-info as
8367 KERNEL_INSTALL_MACHINE_ID and use it as the machine ID.
8369 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
8370 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
8371 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
8372 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
8373 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
8374 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
8375 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
8376 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
8377 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
8378 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
8379 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
8380 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
8381 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
8382 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
8383 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
8384 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
8385 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
8386 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
8387 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
8388 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
8389 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
8390 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
8391 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
8392 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
8393 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
8394 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
8395 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
8396 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
8397 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
8400 — Berlin, 2017-07-12
8404 * The "hybrid" control group mode has been modified to improve
8405 compatibility with "legacy" cgroups-v1 setups. Specifically, the
8406 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
8407 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
8408 cgroups-v1 hierarchy), the only externally visible change being that
8409 the cgroups-v2 hierarchy is also mounted, to
8410 /sys/fs/cgroup/unified. This should provide a large degree of
8411 compatibility with "legacy" cgroups-v1, while taking benefit of the
8412 better management capabilities of cgroups-v2.
8414 * The default control group setup mode may be selected both a boot-time
8415 via a set of kernel command line parameters (specifically:
8416 systemd.unified_cgroup_hierarchy= and
8417 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
8418 default selected on the configure command line
8419 (--with-default-hierarchy=). The upstream default is "hybrid"
8420 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
8421 this will change in a future systemd version to be "unified" (pure
8422 cgroups-v2 mode). The third option for the compile time option is
8423 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
8424 distributions to default to "hybrid" mode for release distributions,
8425 starting with v233. We recommend "unified" for development
8426 distributions (specifically: distributions such as Fedora's rawhide)
8427 as that's where things are headed in the long run. Use "legacy" for
8428 greatest stability and compatibility only.
8430 * Note one current limitation of "unified" and "hybrid" control group
8431 setup modes: the kernel currently does not permit the systemd --user
8432 instance (i.e. unprivileged code) to migrate processes between two
8433 disconnected cgroup subtrees, even if both are managed and owned by
8434 the user. This effectively means "systemd-run --user --scope" doesn't
8435 work when invoked from outside of any "systemd --user" service or
8436 scope. Specifically, it is not supported from session scopes. We are
8437 working on fixing this in a future systemd version. (See #3388 for
8438 further details about this.)
8440 * DBus policy files are now installed into /usr rather than /etc. Make
8441 sure your system has dbus >= 1.9.18 running before upgrading to this
8442 version, or override the install path with --with-dbuspolicydir= .
8444 * All python scripts shipped with systemd (specifically: the various
8445 tests written in Python) now require Python 3.
8447 * systemd unit tests can now run standalone (without the source or
8448 build directories), and can be installed into /usr/lib/systemd/tests/
8449 with 'make install-tests'.
8451 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
8452 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
8455 * Support for the %c, %r, %R specifiers in unit files has been
8456 removed. Specifiers are not supposed to be dependent on configuration
8457 in the unit file itself (so that they resolve the same regardless
8458 where used in the unit files), but these specifiers were influenced
8459 by the Slice= option.
8461 * The shell invoked by debug-shell.service now defaults to /bin/sh in
8462 all cases. If distributions want to use a different shell for this
8463 purpose (for example Fedora's /sbin/sushell) they need to specify
8464 this explicitly at configure time using --with-debug-shell=.
8466 * The confirmation spawn prompt has been reworked to offer the
8469 (c)ontinue, proceed without asking anymore
8470 (D)ump, show the state of the unit
8471 (f)ail, don't execute the command and pretend it failed
8473 (i)nfo, show a short summary of the unit
8474 (j)obs, show jobs that are in progress
8475 (s)kip, don't execute the command and pretend it succeeded
8476 (y)es, execute the command
8478 The 'n' choice for the confirmation spawn prompt has been removed,
8479 because its meaning was confusing.
8481 The prompt may now also be redirected to an alternative console by
8482 specifying the console as parameter to systemd.confirm_spawn=.
8484 * Services of Type=notify require a READY=1 notification to be sent
8485 during startup. If no such message is sent, the service now fails,
8486 even if the main process exited with a successful exit code.
8488 * Services that fail to start up correctly now always have their
8489 ExecStopPost= commands executed. Previously, they'd enter "failed"
8490 state directly, without executing these commands.
8492 * The option MulticastDNS= of network configuration files has acquired
8493 an actual implementation. With MulticastDNS=yes a host can resolve
8494 names of remote hosts and reply to mDNS A and AAAA requests.
8496 * When units are about to be started an additional check is now done to
8497 ensure that all dependencies of type BindsTo= (when used in
8498 combination with After=) have been started.
8500 * systemd-analyze gained a new verb "syscall-filter" which shows which
8501 system call groups are defined for the SystemCallFilter= unit file
8502 setting, and which system calls they contain.
8504 * A new system call filter group "@filesystem" has been added,
8505 consisting of various file system related system calls. Group
8506 "@reboot" has been added, covering reboot, kexec and shutdown related
8507 calls. Finally, group "@swap" has been added covering swap
8508 configuration related calls.
8510 * A new unit file option RestrictNamespaces= has been added that may be
8511 used to restrict access to the various process namespace types the
8512 Linux kernel provides. Specifically, it may be used to take away the
8513 right for a service unit to create additional file system, network,
8514 user, and other namespaces. This sandboxing option is particularly
8515 relevant due to the high amount of recently discovered namespacing
8516 related vulnerabilities in the kernel.
8518 * systemd-udev's .link files gained support for a new AutoNegotiation=
8519 setting for configuring Ethernet auto-negotiation.
8521 * systemd-networkd's .network files gained support for a new
8522 ListenPort= setting in the [DHCP] section to explicitly configure the
8523 UDP client port the DHCP client shall listen on.
8525 * .network files gained a new Unmanaged= boolean setting for explicitly
8526 excluding one or more interfaces from management by systemd-networkd.
8528 * The systemd-networkd ProxyARP= option has been renamed to
8529 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
8530 renamed to ReduceARPProxy=. The old names continue to be available
8533 * systemd-networkd gained support for configuring IPv6 Proxy NDP
8534 addresses via the new IPv6ProxyNDPAddress= .network file setting.
8536 * systemd-networkd's bonding device support gained support for two new
8537 configuration options ActiveSlave= and PrimarySlave=.
8539 * The various options in the [Match] section of .network files gained
8540 support for negative matching.
8542 * New systemd-specific mount options are now understood in /etc/fstab:
8544 x-systemd.mount-timeout= may be used to configure the maximum
8545 permitted runtime of the mount command.
8547 x-systemd.device-bound may be set to bind a mount point to its
8548 backing device unit, in order to automatically remove a mount point
8549 if its backing device is unplugged. This option may also be
8550 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
8551 on the block device, which is now automatically set for all CDROM
8552 drives, so that mounted CDs are automatically unmounted when they are
8553 removed from the drive.
8555 x-systemd.after= and x-systemd.before= may be used to explicitly
8556 order a mount after or before another unit or mount point.
8558 * Enqueued start jobs for device units are now automatically garbage
8559 collected if there are no jobs waiting for them anymore.
8561 * systemctl list-jobs gained two new switches: with --after, for every
8562 queued job the jobs it's waiting for are shown; with --before the
8563 jobs which it's blocking are shown.
8565 * systemd-nspawn gained support for ephemeral boots from disk images
8566 (or in other words: --ephemeral and --image= may now be
8567 combined). Moreover, ephemeral boots are now supported for normal
8568 directories, even if the backing file system is not btrfs. Of course,
8569 if the file system does not support file system snapshots or
8570 reflinks, the initial copy operation will be relatively expensive, but
8571 this should still be suitable for many use cases.
8573 * Calendar time specifications in .timer units now support
8574 specifications relative to the end of a month by using "~" instead of
8575 "-" as separator between month and day. For example, "*-02~03" means
8576 "the third last day in February". In addition a new syntax for
8577 repeated events has been added using the "/" character. For example,
8578 "9..17/2:00" means "every two hours from 9am to 5pm".
8580 * systemd-socket-proxyd gained a new parameter --connections-max= for
8581 configuring the maximum number of concurrent connections.
8583 * sd-id128 gained a new API for generating unique IDs for the host in a
8584 way that does not leak the machine ID. Specifically,
8585 sd_id128_get_machine_app_specific() derives an ID based on the
8586 machine ID in a well-defined, non-reversible, stable way. This is
8587 useful whenever an identifier for the host is needed but where the
8588 identifier shall not be useful to identify the system beyond the
8589 scope of the application itself. (Internally this uses HMAC-SHA256 as
8590 keyed hash function using the machine ID as input.)
8592 * NotifyAccess= gained a new supported value "exec". When set
8593 notifications are accepted from all processes systemd itself invoked,
8594 including all control processes.
8596 * .nspawn files gained support for defining overlay mounts using the
8597 Overlay= and OverlayReadOnly= options. Previously this functionality
8598 was only available on the systemd-nspawn command line.
8600 * systemd-nspawn's --bind= and --overlay= options gained support for
8601 bind/overlay mounts whose source lies within the container tree by
8602 prefixing the source path with "+".
8604 * systemd-nspawn's --bind= and --overlay= options gained support for
8605 automatically allocating a temporary source directory in /var/tmp
8606 that is removed when the container dies. Specifically, if the source
8607 directory is specified as empty string this mechanism is selected. An
8608 example usage is --overlay=+/var::/var, which creates an overlay
8609 mount based on the original /var contained in the image, overlaid
8610 with a temporary directory in the host's /var/tmp. This way changes
8611 to /var are automatically flushed when the container shuts down.
8613 * systemd-nspawn --image= option does now permit raw file system block
8614 devices (in addition to images containing partition tables, as
8617 * The disk image dissection logic in systemd-nspawn gained support for
8618 automatically setting up LUKS encrypted as well as Verity protected
8619 partitions. When a container is booted from an encrypted image the
8620 passphrase is queried at start-up time. When a container with Verity
8621 data is started, the root hash is search in a ".roothash" file
8622 accompanying the disk image (alternatively, pass the root hash via
8623 the new --root-hash= command line option).
8625 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
8626 be used to dissect disk images the same way as systemd-nspawn does
8627 it, following the Bootable Partition Specification. It may even be
8628 used to mount disk images with complex partition setups (including
8629 LUKS and Verity partitions) to a local host directory, in order to
8630 inspect them. This tool is not considered public API (yet), and is
8631 thus not installed into /usr/bin. Please do not rely on its
8632 existence, since it might go away or be changed in later systemd
8635 * A new generator "systemd-verity-generator" has been added, similar in
8636 style to "systemd-cryptsetup-generator", permitting automatic setup of
8637 Verity root partitions when systemd boots up. In order to make use of
8638 this your partition setup should follow the Discoverable Partitions
8639 Specification, and the GPT partition ID of the root file system
8640 partition should be identical to the upper 128bit of the Verity root
8641 hash. The GPT partition ID of the Verity partition protecting it
8642 should be the lower 128bit of the Verity root hash. If the partition
8643 image follows this model it is sufficient to specify a single
8644 "roothash=" kernel command line argument to both configure which root
8645 image and verity partition to use as well as the root hash for
8646 it. Note that systemd-nspawn's Verity support follows the same
8647 semantics, meaning that disk images with proper Verity data in place
8648 may be booted in containers with systemd-nspawn as well as on
8649 physical systems via the verity generator. Also note that the "mkosi"
8650 tool available at https://github.com/systemd/mkosi has been updated
8651 to generate Verity protected disk images following this scheme. In
8652 fact, it has been updated to generate disk images that optionally
8653 implement a complete UEFI SecureBoot trust chain, involving a signed
8654 kernel and initrd image that incorporates such a root hash as well as
8655 a Verity-enabled root partition.
8657 * The hardware database (hwdb) udev supports has been updated to carry
8658 accelerometer quirks.
8660 * All system services are now run with a fresh kernel keyring set up
8661 for them. The invocation ID is stored by default in it, thus
8662 providing a safe, non-overridable way to determine the invocation
8665 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
8666 options for bind mounting arbitrary paths in a service-specific
8667 way. When these options are used, arbitrary host or service files and
8668 directories may be mounted to arbitrary locations in the service's
8671 * Documentation has been added that lists all of systemd's low-level
8672 environment variables:
8674 https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md
8676 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
8677 whether a specific socket file descriptor matches a specified socket
8680 * systemd-firstboot has been updated to check for the
8681 systemd.firstboot= kernel command line option. It accepts a boolean
8682 and when set to false the first boot questions are skipped.
8684 * systemd-fstab-generator has been updated to check for the
8685 systemd.volatile= kernel command line option, which either takes an
8686 optional boolean parameter or the special value "state". If used the
8687 system may be booted in a "volatile" boot mode. Specifically,
8688 "systemd.volatile" is used, the root directory will be mounted as
8689 tmpfs, and only /usr is mounted from the actual root file system. If
8690 "systemd.volatile=state" is used, the root directory will be mounted
8691 as usual, but /var is mounted as tmpfs. This concept provides similar
8692 functionality as systemd-nspawn's --volatile= option, but provides it
8693 on physical boots. Use this option for implementing stateless
8694 systems, or testing systems with all state and/or configuration reset
8695 to the defaults. (Note though that many distributions are not
8696 prepared to boot up without a populated /etc or /var, though.)
8698 * systemd-gpt-auto-generator gained support for LUKS encrypted root
8699 partitions. Previously it only supported LUKS encrypted partitions
8700 for all other uses, except for the root partition itself.
8702 * Socket units gained support for listening on AF_VSOCK sockets for
8703 communication in virtualized QEMU environments.
8705 * The "configure" script gained a new option --with-fallback-hostname=
8706 for specifying the fallback hostname to use if none is configured in
8707 /etc/hostname. For example, by specifying
8708 --with-fallback-hostname=fedora it is possible to default to a
8709 hostname of "fedora" on pristine installations.
8711 * systemd-cgls gained support for a new --unit= switch for listing only
8712 the control groups of a specific unit. Similar --user-unit= has been
8713 added for listing only the control groups of a specific user unit.
8715 * systemd-mount gained a new --umount switch for unmounting a mount or
8716 automount point (and all mount/automount points below it).
8718 * systemd will now refuse full configuration reloads (via systemctl
8719 daemon-reload and related calls) unless at least 16MiB of free space
8720 are available in /run. This is a safety precaution in order to ensure
8721 that generators can safely operate after the reload completed.
8723 * A new unit file option RootImage= has been added, which has a similar
8724 effect as RootDirectory= but mounts the service's root directory from
8725 a disk image instead of plain directory. This logic reuses the same
8726 image dissection and mount logic that systemd-nspawn already uses,
8727 and hence supports any disk images systemd-nspawn supports, including
8728 those following the Discoverable Partition Specification, as well as
8729 Verity enabled images. This option enables systemd to run system
8730 services directly off disk images acting as resource bundles,
8731 possibly even including full integrity data.
8733 * A new MountAPIVFS= unit file option has been added, taking a boolean
8734 argument. If enabled /proc, /sys and /dev (collectively called the
8735 "API VFS") will be mounted for the service. This is only relevant if
8736 RootDirectory= or RootImage= is used for the service, as these mounts
8737 are of course in place in the host mount namespace anyway.
8739 * systemd-nspawn gained support for a new --pivot-root= switch. If
8740 specified the root directory within the container image is pivoted to
8741 the specified mount point, while the original root disk is moved to a
8742 different place. This option enables booting of ostree images
8743 directly with systemd-nspawn.
8745 * The systemd build scripts will no longer complain if the NTP server
8746 addresses are not changed from the defaults. Google now supports
8747 these NTP servers officially. We still recommend downstreams to
8748 properly register an NTP pool with the NTP pool project though.
8750 * coredumpctl gained a new "--reverse" option for printing the list
8751 of coredumps in reverse order.
8753 * coredumpctl will now show additional information about truncated and
8754 inaccessible coredumps, as well as coredumps that are still being
8755 processed. It also gained a new --quiet switch for suppressing
8756 additional informational message in its output.
8758 * coredumpctl gained support for only showing coredumps newer and/or
8759 older than specific timestamps, using the new --since= and --until=
8760 options, reminiscent of journalctl's options by the same name.
8762 * The systemd-coredump logic has been improved so that it may be reused
8763 to collect backtraces in non-compiled languages, for example in
8764 scripting languages such as Python.
8766 * machinectl will now show the UID shift of local containers, if user
8767 namespacing is enabled for them.
8769 * systemd will now optionally run "environment generator" binaries at
8770 configuration load time. They may be used to add environment
8771 variables to the environment block passed to services invoked. One
8772 user environment generator is shipped by default that sets up
8773 environment variables based on files dropped into /etc/environment.d
8774 and ~/.config/environment.d/.
8776 * systemd-resolved now includes the new, recently published 2017 DNSSEC
8779 * hostnamed has been updated to report a new chassis type of
8780 "convertible" to cover "foldable" laptops that can both act as a
8781 tablet and as a laptop, such as various Lenovo Yoga devices.
8783 Contributions from: Adrián López, Alexander Galanin, Alexander
8784 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
8785 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
8786 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
8787 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
8788 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
8789 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
8790 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
8791 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
8792 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
8793 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
8794 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
8795 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
8796 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
8797 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
8798 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
8799 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
8800 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
8801 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
8802 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
8803 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
8804 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
8805 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
8806 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
8807 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
8808 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
8809 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
8810 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
8813 — Berlin, 2017-03-01
8817 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
8818 RestrictAddressFamilies= enabled. These sandboxing options should
8819 generally be compatible with the various external udev call-out
8820 binaries we are aware of, however there may be exceptions, in
8821 particular when exotic languages for these call-outs are used. In
8822 this case, consider turning off these settings locally.
8824 * The new RemoveIPC= option can be used to remove IPC objects owned by
8825 the user or group of a service when that service exits.
8827 * The new ProtectKernelModules= option can be used to disable explicit
8828 load and unload operations of kernel modules by a service. In
8829 addition access to /usr/lib/modules is removed if this option is set.
8831 * ProtectSystem= option gained a new value "strict", which causes the
8832 whole file system tree with the exception of /dev, /proc, and /sys,
8833 to be remounted read-only for a service.
8835 * The new ProtectKernelTunables= option can be used to disable
8836 modification of configuration files in /sys and /proc by a service.
8837 Various directories and files are remounted read-only, so access is
8838 restricted even if the file permissions would allow it.
8840 * The new ProtectControlGroups= option can be used to disable write
8841 access by a service to /sys/fs/cgroup.
8843 * Various systemd services have been hardened with
8844 ProtectKernelTunables=yes, ProtectControlGroups=yes,
8845 RestrictAddressFamilies=.
8847 * Support for dynamically creating users for the lifetime of a service
8848 has been added. If DynamicUser=yes is specified, user and group IDs
8849 will be allocated from the range 61184…65519 for the lifetime of the
8850 service. They can be resolved using the new nss-systemd.so NSS
8851 module. The module must be enabled in /etc/nsswitch.conf. Services
8852 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
8853 any resources allocated by the service will be cleaned up when the
8854 service exits. They also have ProtectHome=read-only and
8855 ProtectSystem=strict enabled, so they are not able to make any
8856 permanent modifications to the system.
8858 * The nss-systemd module also always resolves root and nobody, making
8859 it possible to have no /etc/passwd or /etc/group files in minimal
8860 container or chroot environments.
8862 * Services may be started with their own user namespace using the new
8863 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
8864 under which the service is running are mapped. All other users are
8867 * Support for the cgroup namespace has been added to systemd-nspawn. If
8868 supported by kernel, the container system started by systemd-nspawn
8869 will have its own view of the cgroup hierarchy. This new behaviour
8870 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
8872 * The new MemorySwapMax= option can be used to limit the maximum swap
8873 usage under the unified cgroup hierarchy.
8875 * Support for the CPU controller in the unified cgroup hierarchy has
8876 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
8877 options. This controller requires out-of-tree patches for the kernel
8878 and the support is provisional.
8880 * Mount and automount units may now be created transiently
8881 (i.e. dynamically at runtime via the bus API, instead of requiring
8882 unit files in the file system).
8884 * systemd-mount is a new tool which may mount file systems – much like
8885 mount(8), optionally pulling in additional dependencies through
8886 transient .mount and .automount units. For example, this tool
8887 automatically runs fsck on a backing block device before mounting,
8888 and allows the automount logic to be used dynamically from the
8889 command line for establishing mount points. This tool is particularly
8890 useful when dealing with removable media, as it will ensure fsck is
8891 run – if necessary – before the first access and that the file system
8892 is quickly unmounted after each access by utilizing the automount
8893 logic. This maximizes the chance that the file system on the
8894 removable media stays in a clean state, and if it isn't in a clean
8895 state is fixed automatically.
8897 * LazyUnmount=yes option for mount units has been added to expose the
8898 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
8901 * /efi will be used as the mount point of the EFI boot partition, if
8902 the directory is present, and the mount point was not configured
8903 through other means (e.g. fstab). If /efi directory does not exist,
8904 /boot will be used as before. This makes it easier to automatically
8905 mount the EFI partition on systems where /boot is used for something
8908 * When operating on GPT disk images for containers, systemd-nspawn will
8909 now mount the ESP to /boot or /efi according to the same rules as PID
8910 1 running on a host. This allows tools like "bootctl" to operate
8911 correctly within such containers, in order to make container images
8912 bootable on physical systems.
8914 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
8916 * Two new user session targets have been added to support running
8917 graphical sessions under the systemd --user instance:
8918 graphical-session.target and graphical-session-pre.target. See
8919 systemd.special(7) for a description of how those targets should be
8922 * The vconsole initialization code has been significantly reworked to
8923 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
8924 support unicode keymaps. Font and keymap configuration will now be
8925 copied to all allocated virtual consoles.
8927 * FreeBSD's bhyve virtualization is now detected.
8929 * Information recorded in the journal for core dumps now includes the
8930 contents of /proc/mountinfo and the command line of the process at
8931 the top of the process hierarchy (which is usually the init process
8934 * systemd-journal-gatewayd learned the --directory= option to serve
8935 files from the specified location.
8937 * journalctl --root=… can be used to peruse the journal in the
8938 /var/log/ directories inside of a container tree. This is similar to
8939 the existing --machine= option, but does not require the container to
8942 * The hardware database has been extended to support
8943 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
8946 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
8947 specify the click rate for mice which include a horizontal wheel with
8948 a click rate that is different than the one for the vertical wheel.
8950 * systemd-run gained a new --wait option that makes service execution
8951 synchronous. (Specifically, the command will not return until the
8952 specified service binary exited.)
8954 * systemctl gained a new --wait option that causes the start command to
8955 wait until the units being started have terminated again.
8957 * A new journal output mode "short-full" has been added which displays
8958 timestamps with abbreviated English day names and adds a timezone
8959 suffix. Those timestamps include more information than the default
8960 "short" output mode, and can be passed directly to journalctl's
8961 --since= and --until= options.
8963 * /etc/resolv.conf will be bind-mounted into containers started by
8964 systemd-nspawn, if possible, so any changes to resolv.conf contents
8965 are automatically propagated to the container.
8967 * The number of instances for socket-activated services originating
8968 from a single IP address can be limited with
8969 MaxConnectionsPerSource=, extending the existing setting of
8972 * systemd-networkd gained support for vcan ("Virtual CAN") interface
8975 * .netdev and .network configuration can now be extended through
8978 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
8979 Segmentation Offload, Generic Receive Offload, Large Receive Offload
8980 can be enabled and disabled using the new UDPSegmentationOffload=,
8981 TCPSegmentationOffload=, GenericSegmentationOffload=,
8982 GenericReceiveOffload=, LargeReceiveOffload= options in the
8983 [Link] section of .link files.
8985 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
8986 Port VLAN ID can be configured for bridge devices using the new STP=,
8987 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
8988 section of .netdev files.
8990 * The route table to which routes received over DHCP or RA should be
8991 added can be configured with the new RouteTable= option in the [DHCP]
8992 and [IPv6AcceptRA] sections of .network files.
8994 * The Address Resolution Protocol can be disabled on links managed by
8995 systemd-networkd using the ARP=no setting in the [Link] section of
8998 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
8999 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
9000 encode information about the result and exit codes of the current
9001 service runtime cycle.
9003 * systemd-sysctl will now configure kernel parameters in the order
9004 they occur in the configuration files. This matches what sysctl
9005 has been traditionally doing.
9007 * kernel-install "plugins" that are executed to perform various
9008 tasks after a new kernel is added and before an old one is removed
9009 can now return a special value to terminate the procedure and
9010 prevent any later plugins from running.
9012 * Journald's SplitMode=login setting has been deprecated. It has been
9013 removed from documentation, and its use is discouraged. In a future
9014 release it will be completely removed, and made equivalent to current
9015 default of SplitMode=uid.
9017 * Storage=both option setting in /etc/systemd/coredump.conf has been
9018 removed. With fast LZ4 compression storing the core dump twice is not
9021 * The --share-system systemd-nspawn option has been replaced with an
9022 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
9023 this functionality is discouraged. In addition the variables
9024 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
9025 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
9026 individual namespaces.
9028 * "machinectl list" now shows the IP address of running containers in
9029 the output, as well as OS release information.
9031 * "loginctl list" now shows the TTY of each session in the output.
9033 * sd-bus gained new API calls sd_bus_track_set_recursive(),
9034 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
9035 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
9036 tracking objects in a "recursive" mode, where a single client can be
9037 counted multiple times, if it takes multiple references.
9039 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
9040 sd_bus_get_exit_on_disconnect(). They may be used to make a
9041 process using sd-bus automatically exit if the bus connection is
9044 * Bus clients of the service manager may now "pin" loaded units into
9045 memory, by taking an explicit reference on them. This is useful to
9046 ensure the client can retrieve runtime data about the service even
9047 after the service completed execution. Taking such a reference is
9048 available only for privileged clients and should be helpful to watch
9049 running services in a race-free manner, and in particular collect
9050 information about exit statuses and results.
9052 * The nss-resolve module has been changed to strictly return UNAVAIL
9053 when communication via D-Bus with resolved failed, and NOTFOUND when
9054 a lookup completed but was negative. This means it is now possible to
9055 neatly configure fallbacks using nsswitch.conf result checking
9056 expressions. Taking benefit of this, the new recommended
9057 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
9059 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
9061 * A new setting CtrlAltDelBurstAction= has been added to
9062 /etc/systemd/system.conf which may be used to configure the precise
9063 behaviour if the user on the console presses Ctrl-Alt-Del more often
9064 than 7 times in 2s. Previously this would unconditionally result in
9065 an expedited, immediate reboot. With this new setting the precise
9066 operation may be configured in more detail, and also turned off
9069 * In .netdev files two new settings RemoteChecksumTx= and
9070 RemoteChecksumRx= are now understood that permit configuring the
9071 remote checksumming logic for VXLAN networks.
9073 * The service manager learnt a new "invocation ID" concept for invoked
9074 services. Each runtime cycle of a service will get a new invocation
9075 ID (a 128bit random UUID) assigned that identifies the current
9076 run of the service uniquely and globally. A new invocation ID
9077 is generated each time a service starts up. The journal will store
9078 the invocation ID of a service along with any logged messages, thus
9079 making the invocation ID useful for matching the online runtime of a
9080 service with the offline log data it generated in a safe way without
9081 relying on synchronized timestamps. In many ways this new service
9082 invocation ID concept is similar to the kernel's boot ID concept that
9083 uniquely and globally identifies the runtime of each boot. The
9084 invocation ID of a service is passed to the service itself via an
9085 environment variable ($INVOCATION_ID). A new bus call
9086 GetUnitByInvocationID() has been added that is similar to GetUnit()
9087 but instead of retrieving the bus path for a unit by its name
9088 retrieves it by its invocation ID. The returned path is valid only as
9089 long as the passed invocation ID is current.
9091 * systemd-resolved gained a new "DNSStubListener" setting in
9092 resolved.conf. It either takes a boolean value or the special values
9093 "udp" and "tcp", and configures whether to enable the stub DNS
9094 listener on 127.0.0.53:53.
9096 * IP addresses configured via networkd may now carry additional
9097 configuration settings supported by the kernel. New options include:
9098 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
9099 PrefixRoute=, AutoJoin=.
9101 * The PAM configuration fragment file for "user@.service" shipped with
9102 systemd (i.e. the --user instance of systemd) has been stripped to
9103 the minimum necessary to make the system boot. Previously, it
9104 contained Fedora-specific stanzas that did not apply to other
9105 distributions. It is expected that downstream distributions add
9106 additional configuration lines, matching their needs to this file,
9107 using it only as rough template of what systemd itself needs. Note
9108 that this reduced fragment does not even include an invocation of
9109 pam_limits which most distributions probably want to add, even though
9110 systemd itself does not need it. (There's also the new build time
9111 option --with-pamconfdir=no to disable installation of the PAM
9114 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
9115 capability is now also dropped from its set (in addition to
9116 CAP_SYS_MKNOD as before).
9118 * In service unit files it is now possible to connect a specific named
9119 file descriptor with stdin/stdout/stdout of an executed service. The
9120 name may be specified in matching .socket units using the
9121 FileDescriptorName= setting.
9123 * A number of journal settings may now be configured on the kernel
9124 command line. Specifically, the following options are now understood:
9125 systemd.journald.max_level_console=,
9126 systemd.journald.max_level_store=,
9127 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
9128 systemd.journald.max_level_wall=.
9130 * "systemctl is-enabled --full" will now show by which symlinks a unit
9131 file is enabled in the unit dependency tree.
9133 * Support for VeraCrypt encrypted partitions has been added to the
9134 "cryptsetup" logic and /etc/crypttab.
9136 * systemd-detect-virt gained support for a new --private-users switch
9137 that checks whether the invoking processes are running inside a user
9138 namespace. Similar, a new special value "private-users" for the
9139 existing ConditionVirtualization= setting has been added, permitting
9140 skipping of specific units in user namespace environments.
9142 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
9143 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
9144 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
9145 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
9146 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
9147 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
9148 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
9149 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
9150 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
9151 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
9152 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
9153 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
9154 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
9155 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
9156 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
9157 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
9158 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
9159 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
9160 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
9161 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
9162 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
9163 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
9164 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
9165 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
9166 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
9167 Jędrzejewski-Szmek, Zeal Jagannatha
9169 — Santa Fe, 2016-11-03
9173 * In service units the various ExecXYZ= settings have been extended
9174 with an additional special character as first argument of the
9175 assigned value: if the character '+' is used the specified command
9176 line it will be run with full privileges, regardless of User=,
9177 Group=, CapabilityBoundingSet= and similar options. The effect is
9178 similar to the existing PermissionsStartOnly= option, but allows
9179 configuration of this concept for each executed command line
9182 * Services may now alter the service watchdog timeout at runtime by
9183 sending a WATCHDOG_USEC= message via sd_notify().
9185 * MemoryLimit= and related unit settings now optionally take percentage
9186 specifications. The percentage is taken relative to the amount of
9187 physical memory in the system (or in case of containers, the assigned
9188 amount of memory). This allows scaling service resources neatly with
9189 the amount of RAM available on the system. Similarly, systemd-logind's
9190 RuntimeDirectorySize= option now also optionally takes percentage
9193 * In similar fashion TasksMax= takes percentage values now, too. The
9194 value is taken relative to the configured maximum number of processes
9195 on the system. The per-service task maximum has been changed to 15%
9196 using this functionality. (Effectively this is an increase of 512 →
9197 4915 for service units, given the kernel's default pid_max setting.)
9199 * Calendar time specifications in .timer units now understand a ".."
9200 syntax for time ranges. Example: "4..7:10" may now be used for
9201 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
9204 * The InaccessableDirectories=, ReadOnlyDirectories= and
9205 ReadWriteDirectories= unit file settings have been renamed to
9206 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
9207 applied to all kinds of file nodes, and not just directories, with
9208 the exception of symlinks. Specifically these settings may now be
9209 used on block and character device nodes, UNIX sockets and FIFOS as
9210 well as regular files. The old names of these settings remain
9211 available for compatibility.
9213 * systemd will now log about all service processes it kills forcibly
9214 (using SIGKILL) because they remained after the clean shutdown phase
9215 of the service completed. This should help identifying services that
9216 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
9217 systemd-logind's configuration a similar log message is generated for
9218 processes killed at the end of each session due to this setting.
9220 * systemd will now set the $JOURNAL_STREAM environment variable for all
9221 services whose stdout/stderr are connected to the Journal (which
9222 effectively means by default: all services). The variable contains
9223 the device and inode number of the file descriptor used for
9224 stdout/stderr. This may be used by invoked programs to detect whether
9225 their stdout/stderr is connected to the Journal, in which case they
9226 can switch over to direct Journal communication, thus being able to
9227 pass extended, structured metadata along with their log messages. As
9228 one example, this is now used by glib's logging primitives.
9230 * When using systemd's default tmp.mount unit for /tmp, the mount point
9231 will now be established with the "nosuid" and "nodev" options. This
9232 avoids privilege escalation attacks that put traps and exploits into
9233 /tmp. However, this might cause problems if you e. g. put container
9234 images or overlays into /tmp; if you need this, override tmp.mount's
9235 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
9238 * systemd now supports the "memory" cgroup controller also on
9241 * The systemd-cgtop tool now optionally takes a control group path as
9242 command line argument. If specified, the control group list shown is
9243 limited to subgroups of that group.
9245 * The SystemCallFilter= unit file setting gained support for
9246 pre-defined, named system call filter sets. For example
9247 SystemCallFilter=@clock is now an effective way to make all clock
9248 changing-related system calls unavailable to a service. A number of
9249 similar pre-defined groups are defined. Writing system call filters
9250 for system services is simplified substantially with this new
9251 concept. Accordingly, all of systemd's own, long-running services now
9252 enable system call filtering based on this, by default.
9254 * A new service setting MemoryDenyWriteExecute= has been added, taking
9255 a boolean value. If turned on, a service may no longer create memory
9256 mappings that are writable and executable at the same time. This
9257 enhances security for services where this is enabled as it becomes
9258 harder to dynamically write and then execute memory in exploited
9259 service processes. This option has been enabled for all of systemd's
9260 own long-running services.
9262 * A new RestrictRealtime= service setting has been added, taking a
9263 boolean argument. If set the service's processes may no longer
9264 acquire realtime scheduling. This improves security as realtime
9265 scheduling may otherwise be used to easily freeze the system.
9267 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
9268 value. This may be used for requesting that the system manager inside
9269 of the container reports start-up completion to nspawn which then
9270 propagates this notification further to the service manager
9271 supervising nspawn itself. A related option NotifyReady= in .nspawn
9272 files has been added too. This functionality allows ordering of the
9273 start-up of multiple containers using the usual systemd ordering
9276 * machinectl gained a new command "stop" that is an alias for
9279 * systemd-resolved gained support for contacting DNS servers on
9280 link-local IPv6 addresses.
9282 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
9283 its caches. A method call for requesting the same operation has been
9284 added to the bus API too, and is made available via "systemd-resolve
9287 * systemd-resolve gained a new --status switch. If passed a brief
9288 summary of the used DNS configuration with per-interface information
9291 * resolved.conf gained a new Cache= boolean option, defaulting to
9292 on. If turned off local DNS caching is disabled. This comes with a
9293 performance penalty in particular when DNSSEC is enabled. Note that
9294 resolved disables its internal caching implicitly anyway, when the
9295 configured DNS server is on a host-local IP address such as ::1 or
9296 127.0.0.1, thus automatically avoiding double local caching.
9298 * systemd-resolved now listens on the local IP address 127.0.0.53:53
9299 for DNS requests. This improves compatibility with local programs
9300 that do not use the libc NSS or systemd-resolved's bus APIs for name
9301 resolution. This minimal DNS service is only available to local
9302 programs and does not implement the full DNS protocol, but enough to
9303 cover local DNS clients. A new, static resolv.conf file, listing just
9304 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
9305 now recommended to make /etc/resolv.conf a symlink to this file in
9306 order to route all DNS lookups to systemd-resolved, regardless if
9307 done via NSS, the bus API or raw DNS packets. Note that this local
9308 DNS service is not as fully featured as the libc NSS or
9309 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
9310 used to deliver link-local address information (as this implies
9311 sending a local interface index along), LLMNR/mDNS support via this
9312 interface is severely restricted. It is thus strongly recommended for
9313 all applications to use the libc NSS API or native systemd-resolved
9316 * systemd-networkd's bridge support learned a new setting
9317 VLANFiltering= for controlling VLAN filtering. Moreover a new section
9318 in .network files has been added for configuring VLAN bridging in
9319 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
9321 * systemd-networkd's IPv6 Router Advertisement code now makes use of
9322 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
9323 now be acquired without relying on DHCPv6. Two new options
9324 UseDomains= and UseDNS= have been added to configure this behaviour.
9326 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
9327 renamed IPv6AcceptRA=, without altering its behaviour. The old
9328 setting name remains available for compatibility reasons.
9330 * The systemd-networkd VTI/VTI6 tunneling support gained new options
9331 Key=, InputKey= and OutputKey=.
9333 * systemd-networkd gained support for VRF ("Virtual Routing Function")
9334 interface configuration.
9336 * "systemctl edit" may now be used to create new unit files by
9337 specifying the --force switch.
9339 * sd-event gained a new function sd_event_get_iteration() for
9340 requesting the current iteration counter of the event loop. It starts
9341 at zero and is increased by one with each event loop iteration.
9343 * A new rpm macro %systemd_ordering is provided by the macros.systemd
9344 file. It can be used in lieu of %systemd_requires in packages which
9345 don't use any systemd functionality and are intended to be installed
9346 in minimal containers without systemd present. This macro provides
9347 ordering dependencies to ensure that if the package is installed in
9348 the same rpm transaction as systemd, systemd will be installed before
9349 the scriptlets for the package are executed, allowing unit presets
9352 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
9353 been added to simplify packaging of generators.
9355 * The os-release file gained VERSION_CODENAME field for the
9356 distribution nickname (e.g. VERSION_CODENAME=woody).
9358 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
9359 can be set to disable parsing of metadata and the creation
9360 of persistent symlinks for that device.
9362 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
9363 to make them available to logged-in users has been reverted.
9365 * Much of the common code of the various systemd components is now
9366 built into an internal shared library libsystemd-shared-231.so
9367 (incorporating the systemd version number in the name, to be updated
9368 with future releases) that the components link to. This should
9369 decrease systemd footprint both in memory during runtime and on
9370 disk. Note that the shared library is not for public use, and is
9371 neither API nor ABI stable, but is likely to change with every new
9372 released update. Packagers need to make sure that binaries
9373 linking to libsystemd-shared.so are updated in step with the
9376 * Configuration for "mkosi" is now part of the systemd
9377 repository. mkosi is a tool to easily build legacy-free OS images,
9378 and is available on github: https://github.com/systemd/mkosi. If
9379 "mkosi" is invoked in the build tree a new raw OS image is generated
9380 incorporating the systemd sources currently being worked on and a
9381 clean, fresh distribution installation. The generated OS image may be
9382 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
9383 UEFI PC. This functionality is particularly useful to easily test
9384 local changes made to systemd in a pristine, defined environment. See
9385 doc/HACKING for details.
9387 * configure learned the --with-support-url= option to specify the
9388 distribution's bugtracker.
9390 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
9391 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
9392 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
9393 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
9394 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
9395 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
9396 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
9397 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
9398 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
9399 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
9400 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
9401 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
9402 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
9403 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
9404 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
9405 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
9406 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
9407 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
9408 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
9410 — Berlin, 2016-07-25
9414 * DNSSEC is now turned on by default in systemd-resolved (in
9415 "allow-downgrade" mode), but may be turned off during compile time by
9416 passing "--with-default-dnssec=no" to "configure" (and of course,
9417 during runtime with DNSSEC= in resolved.conf). We recommend
9418 downstreams to leave this on at least during development cycles and
9419 report any issues with the DNSSEC logic upstream. We are very
9420 interested in collecting feedback about the DNSSEC validator and its
9421 limitations in the wild. Note however, that DNSSEC support is
9422 probably nothing downstreams should turn on in stable distros just
9423 yet, as it might create incompatibilities with a few DNS servers and
9424 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
9425 automatically whenever we detect such incompatible setups, but there
9426 might be systems we do not cover yet. Hence: please help us testing
9427 the DNSSEC code, leave this on where you can, report back, but then
9428 again don't consider turning this on in your stable, LTS or
9429 production release just yet. (Note that you have to enable
9430 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
9431 and its DNSSEC mode for hostname resolution from local
9434 * systemd-resolve conveniently resolves DANE records with the --tlsa
9435 option and OPENPGPKEY records with the --openpgp option. It also
9436 supports dumping raw DNS record data via the new --raw= switch.
9438 * systemd-logind will now by default terminate user processes that are
9439 part of the user session scope unit (session-XX.scope) when the user
9440 logs out. This behavior is controlled by the KillUserProcesses=
9441 setting in logind.conf, and the previous default of "no" is now
9442 changed to "yes". This means that user sessions will be properly
9443 cleaned up after, but additional steps are necessary to allow
9444 intentionally long-running processes to survive logout.
9446 While the user is logged in at least once, user@.service is running,
9447 and any service that should survive the end of any individual login
9448 session can be started at a user service or scope using systemd-run.
9449 systemd-run(1) man page has been extended with an example which shows
9450 how to run screen in a scope unit underneath user@.service. The same
9451 command works for tmux.
9453 After the user logs out of all sessions, user@.service will be
9454 terminated too, by default, unless the user has "lingering" enabled.
9455 To effectively allow users to run long-term tasks even if they are
9456 logged out, lingering must be enabled for them. See loginctl(1) for
9457 details. The default polkit policy was modified to allow users to
9458 set lingering for themselves without authentication.
9460 Previous defaults can be restored at compile time by the
9461 --without-kill-user-processes option to "configure".
9463 * systemd-logind gained new configuration settings SessionsMax= and
9464 InhibitorsMax=, both with a default of 8192. It will not register new
9465 user sessions or inhibitors above this limit.
9467 * systemd-logind will now reload configuration on SIGHUP.
9469 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
9470 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
9471 enable. Also, support for the "io" cgroup controller in the unified
9472 hierarchy has been added, so that the "memory", "pids" and "io" are
9473 now the controllers that are supported on the unified hierarchy.
9475 WARNING: it is not possible to use previous systemd versions with
9476 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
9477 is necessary to also update systemd in the initramfs if using the
9478 unified hierarchy. An updated SELinux policy is also required.
9480 * LLDP support has been extended, and both passive (receive-only) and
9481 active (sender) modes are supported. Passive mode ("routers-only") is
9482 enabled by default in systemd-networkd. Active LLDP mode is enabled
9483 by default for containers on the internal network. The "networkctl
9484 lldp" command may be used to list information gathered. "networkctl
9485 status" will also show basic LLDP information on connected peers now.
9487 * The IAID and DUID unique identifier sent in DHCP requests may now be
9488 configured for the system and each .network file managed by
9489 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
9491 * systemd-networkd gained support for configuring proxy ARP support for
9492 each interface, via the ProxyArp= setting in .network files. It also
9493 gained support for configuring the multicast querier feature of
9494 bridge devices, via the new MulticastQuerier= setting in .netdev
9495 files. Similarly, snooping on the IGMP traffic can be controlled
9496 via the new setting MulticastSnooping=.
9498 A new setting PreferredLifetime= has been added for addresses
9499 configured in .network file to configure the lifetime intended for an
9502 The systemd-networkd DHCP server gained the option EmitRouter=, which
9503 defaults to yes, to configure whether the DHCP Option 3 (Router)
9506 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
9507 systemd-socket-activate and installed into /usr/bin. It is now fully
9510 * systemd-journald now uses separate threads to flush changes to disk
9511 when closing journal files, thus reducing impact of slow disk I/O on
9512 logging performance.
9514 * The sd-journal API gained two new calls
9515 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
9516 can be used to open journal files using file descriptors instead of
9517 file or directory paths. sd_journal_open_container() has been
9518 deprecated, sd_journal_open_directory_fd() should be used instead
9519 with the flag SD_JOURNAL_OS_ROOT.
9521 * journalctl learned a new output mode "-o short-unix" that outputs log
9522 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
9523 UTC). It also gained support for a new --no-hostname setting to
9524 suppress the hostname column in the family of "short" output modes.
9526 * systemd-ask-password now optionally skips printing of the password to
9527 stdout with --no-output which can be useful in scripts.
9529 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
9530 (devices tagged with ID_MAKER_TOOL) are now tagged with
9531 "uaccess" and are available to logged in users.
9533 * The DeviceAllow= unit setting now supports specifiers (with "%").
9535 * "systemctl show" gained a new --value switch, which allows print a
9536 only the contents of a specific unit property, without also printing
9537 the property's name. Similar support was added to "show*" verbs
9538 of loginctl and machinectl that output "key=value" lists.
9540 * A new unit type "generated" was added for files dynamically generated
9541 by generator tools. Similarly, a new unit type "transient" is used
9542 for unit files created using the runtime API. "systemctl enable" will
9543 refuse to operate on such files.
9545 * A new command "systemctl revert" has been added that may be used to
9546 revert to the vendor version of a unit file, in case local changes
9547 have been made by adding drop-ins or overriding the unit file.
9549 * "machinectl clean" gained a new verb to automatically remove all or
9550 just hidden container images.
9552 * systemd-tmpfiles gained support for a new line type "e" for emptying
9553 directories, if they exist, without creating them if they don't.
9555 * systemd-nspawn gained support for automatically patching the UID/GIDs
9556 of the owners and the ACLs of all files and directories in a
9557 container tree to match the UID/GID user namespacing range selected
9558 for the container invocation. This mode is enabled via the new
9559 --private-users-chown switch. It also gained support for
9560 automatically choosing a free, previously unused UID/GID range when
9561 starting a container, via the new --private-users=pick setting (which
9562 implies --private-users-chown). Together, these options for the first
9563 time make user namespacing for nspawn containers fully automatic and
9564 thus deployable. The systemd-nspawn@.service template unit file has
9565 been changed to use this functionality by default.
9567 * systemd-nspawn gained a new --network-zone= switch, that allows
9568 creating ad-hoc virtual Ethernet links between multiple containers,
9569 that only exist as long as at least one container referencing them is
9570 running. This allows easy connecting of multiple containers with a
9571 common link that implements an Ethernet broadcast domain. Each of
9572 these network "zones" may be named relatively freely by the user, and
9573 may be referenced by any number of containers, but each container may
9574 only reference one of these "zones". On the lower level, this is
9575 implemented by an automatically managed bridge network interface for
9576 each zone, that is created when the first container referencing its
9577 zone is created and removed when the last one referencing its zone
9580 * The default start timeout may now be configured on the kernel command
9581 line via systemd.default_timeout_start_sec=. It was already
9582 configurable via the DefaultTimeoutStartSec= option in
9583 /etc/systemd/system.conf.
9585 * Socket units gained a new TriggerLimitIntervalSec= and
9586 TriggerLimitBurst= setting to configure a limit on the activation
9587 rate of the socket unit.
9589 * The LimitNICE= setting now optionally takes normal UNIX nice values
9590 in addition to the raw integer limit value. If the specified
9591 parameter is prefixed with "+" or "-" and is in the range -20…19 the
9592 value is understood as UNIX nice value. If not prefixed like this it
9593 is understood as raw RLIMIT_NICE limit.
9595 * Note that the effect of the PrivateDevices= unit file setting changed
9596 slightly with this release: the per-device /dev file system will be
9597 mounted read-only from this version on, and will have "noexec"
9598 set. This (minor) change of behavior might cause some (exceptional)
9599 legacy software to break, when PrivateDevices=yes is set for its
9600 service. Please leave PrivateDevices= off if you run into problems
9603 * systemd-bootchart has been split out to a separate repository:
9604 https://github.com/systemd/systemd-bootchart
9606 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
9607 merged into the kernel in its current form.
9609 * The compatibility libraries libsystemd-daemon.so,
9610 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
9611 which have been deprecated since systemd-209 have been removed along
9612 with the corresponding pkg-config files. All symbols provided by
9613 those libraries are provided by libsystemd.so.
9615 * The Capabilities= unit file setting has been removed (it is ignored
9616 for backwards compatibility). AmbientCapabilities= and
9617 CapabilityBoundingSet= should be used instead.
9619 * A new special target has been added, initrd-root-device.target,
9620 which creates a synchronization point for dependencies of the root
9621 device in early userspace. Initramfs builders must ensure that this
9622 target is now included in early userspace.
9624 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
9625 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
9626 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
9627 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
9628 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
9629 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
9630 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
9631 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
9632 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
9633 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
9634 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
9635 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
9636 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
9637 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
9638 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
9639 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
9640 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
9641 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
9642 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
9643 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
9644 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
9645 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
9646 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
9647 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
9648 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
9651 — Fairfax, 2016-05-21
9655 * The systemd-resolved DNS resolver service has gained a substantial
9656 set of new features, most prominently it may now act as a DNSSEC
9657 validating stub resolver. DNSSEC mode is currently turned off by
9658 default, but is expected to be turned on by default in one of the
9659 next releases. For now, we invite everybody to test the DNSSEC logic
9660 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
9661 service also gained a full set of D-Bus interfaces, including calls
9662 to configure DNS and DNSSEC settings per link (for use by external
9663 network management software). systemd-resolved and systemd-networkd
9664 now distinguish between "search" and "routing" domains. The former
9665 are used to qualify single-label names, the latter are used purely
9666 for routing lookups within certain domains to specific links.
9667 resolved now also synthesizes RRs for all entries from /etc/hosts.
9669 * The systemd-resolve tool (which is a client utility for
9670 systemd-resolved) has been improved considerably and is now fully
9671 supported and documented. Hence it has moved from /usr/lib/systemd to
9674 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
9677 * The coredump collection logic has been reworked: when a coredump is
9678 collected it is now written to disk, compressed and processed
9679 (including stacktrace extraction) from a new instantiated service
9680 systemd-coredump@.service, instead of directly from the
9681 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
9682 processing large coredumps can take up a substantial amount of
9683 resources and time, and this previously happened entirely outside of
9684 systemd's service supervision. With the new logic the core_pattern
9685 hook only does minimal metadata collection before passing off control
9686 to the new instantiated service, which is configured with a time
9687 limit, a nice level and other settings to minimize negative impact on
9688 the rest of the system. Also note that the new logic will honour the
9689 RLIMIT_CORE setting of the crashed process, which now allows users
9690 and processes to turn off coredumping for their processes by setting
9693 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
9694 and all forked processes by default. Previously, PID 1 would leave
9695 the setting at "0" for all processes, as set by the kernel. Note that
9696 the resource limit traditionally has no effect on the generated
9697 coredumps on the system if the /proc/sys/kernel/core_pattern hook
9698 logic is used. Since the limit is now honoured (see above) its
9699 default has been changed so that the coredumping logic is enabled by
9700 default for all processes, while allowing specific opt-out.
9702 * When the stacktrace is extracted from processes of system users, this
9703 is now done as "systemd-coredump" user, in order to sandbox this
9704 potentially security sensitive parsing operation. (Note that when
9705 processing coredumps of normal users this is done under the user ID
9706 of process that crashed, as before.) Packagers should take notice
9707 that it is now necessary to create the "systemd-coredump" system user
9708 and group at package installation time.
9710 * The systemd-activate socket activation testing tool gained support
9711 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
9712 and --seqpacket switches. It also has been extended to support both
9713 new-style and inetd-style file descriptor passing. Use the new
9714 --inetd switch to request inetd-style file descriptor passing.
9716 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
9717 variable, which takes a boolean value. If set to false, ANSI color
9718 output is disabled in the tools even when run on a terminal that
9721 * The VXLAN support in networkd now supports two new settings
9722 DestinationPort= and PortRange=.
9724 * A new systemd.machine_id= kernel command line switch has been added,
9725 that may be used to set the machine ID in /etc/machine-id if it is
9726 not initialized yet. This command line option has no effect if the
9727 file is already initialized.
9729 * systemd-nspawn gained a new --as-pid2 switch that invokes any
9730 specified command line as PID 2 rather than PID 1 in the
9731 container. In this mode PID 1 is a minimal stub init process that
9732 implements the special POSIX and Linux semantics of PID 1 regarding
9733 signal and child process management. Note that this stub init process
9734 is implemented in nspawn itself and requires no support from the
9735 container image. This new logic is useful to support running
9736 arbitrary commands in the container, as normal processes are
9737 generally not prepared to run as PID 1.
9739 * systemd-nspawn gained a new --chdir= switch for setting the current
9740 working directory for the process started in the container.
9742 * "journalctl /dev/sda" will now output all kernel log messages for
9743 specified device from the current boot, in addition to all devices
9744 that are parents of it. This should make log output about devices
9745 pretty useful, as long as kernel drivers attach enough metadata to
9746 the log messages. (The usual SATA drivers do.)
9748 * The sd-journal API gained two new calls
9749 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
9750 that report whether log data from /run or /var has been found.
9752 * journalctl gained a new switch "--fields" that prints all journal
9753 record field names currently in use in the journal. This is backed
9754 by two new sd-journal API calls sd_journal_enumerate_fields() and
9755 sd_journal_restart_fields().
9757 * Most configurable timeouts in systemd now expect an argument of
9758 "infinity" to turn them off, instead of "0" as before. The semantics
9759 from now on is that a timeout of "0" means "now", and "infinity"
9760 means "never". To maintain backwards compatibility, "0" continues to
9761 turn off previously existing timeout settings.
9763 * "systemctl reload-or-try-restart" has been renamed to "systemctl
9764 try-reload-or-restart" to clarify what it actually does: the "try"
9765 logic applies to both reloading and restarting, not just restarting.
9766 The old name continues to be accepted for compatibility.
9768 * On boot-up, when PID 1 detects that the system clock is behind the
9769 release date of the systemd version in use, the clock is now set
9770 to the latter. Previously, this was already done in timesyncd, in order
9771 to avoid running with clocks set to the various clock epochs such as
9772 1902, 1938 or 1970. With this change the logic is now done in PID 1
9773 in addition to timesyncd during early boot-up, so that it is enforced
9774 before the first process is spawned by systemd. Note that the logic
9775 in timesyncd remains, as it is more comprehensive and ensures
9776 clock monotonicity by maintaining a persistent timestamp file in
9777 /var. Since /var is generally not available in earliest boot or the
9778 initrd, this part of the logic remains in timesyncd, and is not done
9781 * Support for tweaking details in net_cls.class_id through the
9782 NetClass= configuration directive has been removed, as the kernel
9783 people have decided to deprecate that controller in cgroup v2.
9784 Userspace tools such as nftables are moving over to setting rules
9785 that are specific to the full cgroup path of a task, which obsoletes
9786 these controllers anyway. The NetClass= directive is kept around for
9787 legacy compatibility reasons. For a more in-depth description of the
9788 kernel change, please refer to the respective upstream commit:
9790 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
9792 * A new service setting RuntimeMaxSec= has been added that may be used
9793 to specify a maximum runtime for a service. If the timeout is hit, the
9794 service is terminated and put into a failure state.
9796 * A new service setting AmbientCapabilities= has been added. It allows
9797 configuration of additional Linux process capabilities that are
9798 passed to the activated processes. This is only available on very
9801 * The process resource limit settings in service units may now be used
9802 to configure hard and soft limits individually.
9804 * The various libsystemd APIs such as sd-bus or sd-event now publicly
9805 expose support for gcc's __attribute__((cleanup())) C extension.
9806 Specifically, for many object destructor functions alternative
9807 versions have been added that have names suffixed with "p" and take a
9808 pointer to a pointer to the object to destroy, instead of just a
9809 pointer to the object itself. This is useful because these destructor
9810 functions may be used directly as parameters to the cleanup
9811 construct. Internally, systemd has been a heavy user of this GCC
9812 extension for a long time, and with this change similar support is
9813 now available to consumers of the library outside of systemd. Note
9814 that by using this extension in your sources compatibility with old
9815 and strictly ANSI compatible C compilers is lost. However, all gcc or
9816 LLVM versions of recent years support this extension.
9818 * Timer units gained support for a new setting RandomizedDelaySec= that
9819 allows configuring some additional randomized delay to the configured
9820 time. This is useful to spread out timer events to avoid load peaks in
9821 clusters or larger setups.
9823 * Calendar time specifications now support sub-second accuracy.
9825 * Socket units now support listening on SCTP and UDP-lite protocol
9828 * The sd-event API now comes with a full set of man pages.
9830 * Older versions of systemd contained experimental support for
9831 compressing journal files and coredumps with the LZ4 compressor that
9832 was not compatible with the lz4 binary (due to API limitations of the
9833 lz4 library). This support has been removed; only support for files
9834 compatible with the lz4 binary remains. This LZ4 logic is now
9835 officially supported and no longer considered experimental.
9837 * The dkr image import logic has been removed again from importd. dkr's
9838 micro-services focus doesn't fit into the machine image focus of
9839 importd, and quickly got out of date with the upstream dkr API.
9841 * Creation of the /run/lock/lockdev/ directory was dropped from
9842 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
9843 been available for many years. If you still need this, you need to
9844 create your own tmpfiles.d config file with:
9846 d /run/lock/lockdev 0775 root lock -
9848 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
9849 and RebootArgument= have been moved from the [Service] section of
9850 unit files to [Unit], and they are now supported on all unit types,
9851 not just service units. Of course, systemd will continue to
9852 understand these settings also at the old location, in order to
9853 maintain compatibility.
9855 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
9856 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
9857 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
9858 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
9859 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
9860 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
9861 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
9862 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
9863 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
9864 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
9865 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
9866 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
9867 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
9868 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
9869 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
9870 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
9871 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
9872 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
9873 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
9875 — Berlin, 2016-02-11
9879 * A number of properties previously only settable in unit
9880 files are now also available as properties to set when
9881 creating transient units programmatically via the bus, as it
9882 is exposed with systemd-run's --property=
9883 setting. Specifically, these are: SyslogIdentifier=,
9884 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
9885 EnvironmentFile=, ReadWriteDirectories=,
9886 ReadOnlyDirectories=, InaccessibleDirectories=,
9887 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
9889 * When creating transient services via the bus API it is now
9890 possible to pass in a set of file descriptors to use as
9891 STDIN/STDOUT/STDERR for the invoked process.
9893 * Slice units may now be created transiently via the bus APIs,
9894 similar to the way service and scope units may already be
9895 created transiently.
9897 * Wherever systemd expects a calendar timestamp specification
9898 (like in journalctl's --since= and --until= switches) UTC
9899 timestamps are now supported. Timestamps suffixed with "UTC"
9900 are now considered to be in Universal Time Coordinated
9901 instead of the local timezone. Also, timestamps may now
9902 optionally be specified with sub-second accuracy. Both of
9903 these additions also apply to recurring calendar event
9904 specification, such as OnCalendar= in timer units.
9906 * journalctl gained a new "--sync" switch that asks the
9907 journal daemon to write all so far unwritten log messages to
9908 disk and sync the files, before returning.
9910 * systemd-tmpfiles learned two new line types "q" and "Q" that
9911 operate like "v", but also set up a basic btrfs quota
9912 hierarchy when used on a btrfs file system with quota
9915 * tmpfiles' "v", "q" and "Q" will now create a plain directory
9916 instead of a subvolume (even on a btrfs file system) if the
9917 root directory is a plain directory, and not a
9918 subvolume. This should simplify things with certain chroot()
9919 environments which are not aware of the concept of btrfs
9922 * systemd-detect-virt gained a new --chroot switch to detect
9923 whether execution takes place in a chroot() environment.
9925 * CPUAffinity= now takes CPU index ranges in addition to
9928 * The various memory-related resource limit settings (such as
9929 LimitAS=) now understand the usual K, M, G, … suffixes to
9930 the base of 1024 (IEC). Similar, the time-related resource
9931 limit settings understand the usual min, h, day, … suffixes
9934 * There's a new system.conf setting DefaultTasksMax= to
9935 control the default TasksMax= setting for services and
9936 scopes running on the system. (TasksMax= is the primary
9937 setting that exposes the "pids" cgroup controller on systemd
9938 and was introduced in the previous systemd release.) The
9939 setting now defaults to 512, which means services that are
9940 not explicitly configured otherwise will only be able to
9941 create 512 processes or threads at maximum, from this
9942 version on. Note that this means that thread- or
9943 process-heavy services might need to be reconfigured to set
9944 TasksMax= to a higher value. It is sufficient to set
9945 TasksMax= in these specific unit files to a higher value, or
9946 even "infinity". Similar, there's now a logind.conf setting
9947 UserTasksMax= that defaults to 4096 and limits the total
9948 number of processes or tasks each user may own
9949 concurrently. nspawn containers also have the TasksMax=
9950 value set by default now, to 8192. Note that all of this
9951 only has an effect if the "pids" cgroup controller is
9952 enabled in the kernel. The general benefit of these changes
9953 should be a more robust and safer system, that provides a
9954 certain amount of per-service fork() bomb protection.
9956 * systemd-nspawn gained the new --network-veth-extra= switch
9957 to define additional and arbitrarily-named virtual Ethernet
9958 links between the host and the container.
9960 * A new service execution setting PassEnvironment= has been
9961 added that allows importing select environment variables
9962 from PID1's environment block into the environment block of
9965 * Timer units gained support for a new RemainAfterElapse=
9966 setting which takes a boolean argument. It defaults to on,
9967 exposing behaviour unchanged to previous releases. If set to
9968 off, timer units are unloaded after they elapsed if they
9969 cannot elapse again. This is particularly useful for
9970 transient timer units, which shall not stay around longer
9971 than until they first elapse.
9973 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
9974 default now (the kernel default is 16). This is beneficial
9975 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
9976 allows substantially larger numbers of queued
9977 datagrams. This should increase the capability of systemd to
9978 parallelize boot-up, as logging and sd_notify() are unlikely
9979 to stall execution anymore. If you need to change the value
9980 from the new defaults, use the usual sysctl.d/ snippets.
9982 * The compression framing format used by the journal or
9983 coredump processing has changed to be in line with what the
9984 official LZ4 tools generate. LZ4 compression support in
9985 systemd was considered unsupported previously, as the format
9986 was not compatible with the normal tools. With this release
9987 this has changed now, and it is hence safe for downstream
9988 distributions to turn it on. While not compressing as well
9989 as the XZ, LZ4 is substantially faster, which makes
9990 it a good default choice for the compression logic in the
9991 journal and in coredump handling.
9993 * Any reference to /etc/mtab has been dropped from
9994 systemd. The file has been obsolete since a while, but
9995 systemd refused to work on systems where it was incorrectly
9996 set up (it should be a symlink or non-existent). Please make
9997 sure to update to util-linux 2.27.1 or newer in conjunction
9998 with this systemd release, which also drops any reference to
9999 /etc/mtab. If you maintain a distribution make sure that no
10000 software you package still references it, as this is a
10001 likely source of bugs. There's also a glibc bug pending,
10002 asking for removal of any reference to this obsolete file:
10004 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
10006 Note that only util-linux versions built with
10007 --enable-libmount-force-mountinfo are supported.
10009 * Support for the ".snapshot" unit type has been removed. This
10010 feature turned out to be little useful and little used, and
10011 has now been removed from the core and from systemctl.
10013 * The dependency types RequiresOverridable= and
10014 RequisiteOverridable= have been removed from systemd. They
10015 have been used only very sparingly to our knowledge and
10016 other options that provide a similar effect (such as
10017 systemctl --mode=ignore-dependencies) are much more useful
10018 and commonly used. Moreover, they were only half-way
10019 implemented as the option to control behaviour regarding
10020 these dependencies was never added to systemctl. By removing
10021 these dependency types the execution engine becomes a bit
10022 simpler. Unit files that use these dependencies should be
10023 changed to use the non-Overridable dependency types
10024 instead. In fact, when parsing unit files with these
10025 options, that's what systemd will automatically convert them
10026 too, but it will also warn, asking users to fix the unit
10027 files accordingly. Removal of these dependency types should
10028 only affect a negligible number of unit files in the wild.
10030 * Behaviour of networkd's IPForward= option changed
10031 (again). It will no longer maintain a per-interface setting,
10032 but propagate one way from interfaces where this is enabled
10033 to the global kernel setting. The global setting will be
10034 enabled when requested by a network that is set up, but
10035 never be disabled again. This change was made to make sure
10036 IPv4 and IPv6 behaviour regarding packet forwarding is
10037 similar (as the Linux IPv6 stack does not support
10038 per-interface control of this setting) and to minimize
10041 * In unit files the behaviour of %u, %U, %h, %s has
10042 changed. These specifiers will now unconditionally resolve
10043 to the various user database fields of the user that the
10044 systemd instance is running as, instead of the user
10045 configured in the specific unit via User=. Note that this
10046 effectively doesn't change much, as resolving of these
10047 specifiers was already turned off in the --system instance
10048 of systemd, as we cannot do NSS lookups from PID 1. In the
10049 --user instance of systemd these specifiers where correctly
10050 resolved, but hardly made any sense, since the user instance
10051 lacks privileges to do user switches anyway, and User= is
10052 hence useless. Moreover, even in the --user instance of
10053 systemd behaviour was awkward as it would only take settings
10054 from User= assignment placed before the specifier into
10055 account. In order to unify and simplify the logic around
10056 this the specifiers will now always resolve to the
10057 credentials of the user invoking the manager (which in case
10058 of PID 1 is the root user).
10060 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
10061 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
10062 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
10063 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
10064 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
10065 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
10066 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
10067 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
10068 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
10069 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
10070 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
10071 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
10072 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
10073 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
10076 — Berlin, 2015-11-18
10080 * systemd now depends on util-linux v2.27. More specifically,
10081 the newly added mount monitor feature in libmount now
10082 replaces systemd's former own implementation.
10084 * libmount mandates /etc/mtab not to be regular file, and
10085 systemd now enforces this condition at early boot.
10086 /etc/mtab has been deprecated and warned about for a very
10087 long time, so systems running systemd should already have
10088 stopped having this file around as anything else than a
10089 symlink to /proc/self/mounts.
10091 * Support for the "pids" cgroup controller has been added. It
10092 allows accounting the number of tasks in a cgroup and
10093 enforcing limits on it. This adds two new setting
10094 TasksAccounting= and TasksMax= to each unit, as well as a
10095 global option DefaultTasksAccounting=.
10097 * Support for the "net_cls" cgroup controller has been added.
10098 It allows assigning a net class ID to each task in the
10099 cgroup, which can then be used in firewall rules and traffic
10100 shaping configurations. Note that the kernel netfilter net
10101 class code does not currently work reliably for ingress
10102 packets on unestablished sockets.
10104 This adds a new config directive called NetClass= to CGroup
10105 enabled units. Allowed values are positive numbers for fixed
10106 assignments and "auto" for picking a free value
10109 * 'systemctl is-system-running' now returns 'offline' if the
10110 system is not booted with systemd. This command can now be
10111 used as a substitute for 'systemd-notify --booted'.
10113 * Watchdog timeouts have been increased to 3 minutes for all
10114 in-tree service files. Apparently, disk IO issues are more
10115 frequent than we hoped, and user reported >1 minute waiting
10118 * 'machine-id-commit' functionality has been merged into
10119 'machine-id-setup --commit'. The separate binary has been
10122 * The WorkingDirectory= directive in unit files may now be set
10123 to the special value '~'. In this case, the working
10124 directory is set to the home directory of the user
10125 configured in User=.
10127 * "machinectl shell" will now open the shell in the home
10128 directory of the selected user by default.
10130 * The CrashChVT= configuration file setting is renamed to
10131 CrashChangeVT=, following our usual logic of not
10132 abbreviating unnecessarily. The old directive is still
10133 supported for compat reasons. Also, this directive now takes
10134 an integer value between 1 and 63, or a boolean value. The
10135 formerly supported '-1' value for disabling stays around for
10138 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
10139 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
10140 RootDirectory= properties can now be set for transient
10143 * The systemd-analyze tool gained a new "set-log-target" verb
10144 to change the logging target the system manager logs to
10145 dynamically during runtime. This is similar to how
10146 "systemd-analyze set-log-level" already changes the log
10149 * In nspawn /sys is now mounted as tmpfs, with only a selected
10150 set of subdirectories mounted in from the real sysfs. This
10151 enhances security slightly, and is useful for ensuring user
10152 namespaces work correctly.
10154 * Support for USB FunctionFS activation has been added. This
10155 allows implementation of USB gadget services that are
10156 activated as soon as they are requested, so that they don't
10157 have to run continuously, similar to classic socket
10160 * The "systemctl exit" command now optionally takes an
10161 additional parameter that sets the exit code to return from
10162 the systemd manager when exiting. This is only relevant when
10163 running the systemd user instance, or when running the
10164 system instance in a container.
10166 * sd-bus gained the new API calls sd_bus_path_encode_many()
10167 and sd_bus_path_decode_many() that allow easy encoding and
10168 decoding of multiple identifier strings inside a D-Bus
10169 object path. Another new call sd_bus_default_flush_close()
10170 has been added to flush and close per-thread default
10173 * systemd-cgtop gained support for a -M/--machine= switch to
10174 show the control groups within a certain container only.
10176 * "systemctl kill" gained support for an optional --fail
10177 switch. If specified the requested operation will fail of no
10178 processes have been killed, because the unit had no
10179 processes attached, or similar.
10181 * A new systemd.crash_reboot=1 kernel command line option has
10182 been added that triggers a reboot after crashing. This can
10183 also be set through CrashReboot= in systemd.conf.
10185 * The RuntimeDirectory= setting now understands unit
10186 specifiers like %i or %f.
10188 * A new (still internal) library API sd-ipv4acd has been added,
10189 that implements address conflict detection for IPv4. It's
10190 based on code from sd-ipv4ll, and will be useful for
10191 detecting DHCP address conflicts.
10193 * File descriptors passed during socket activation may now be
10194 named. A new API sd_listen_fds_with_names() is added to
10195 access the names. The default names may be overridden,
10196 either in the .socket file using the FileDescriptorName=
10197 parameter, or by passing FDNAME= when storing the file
10198 descriptors using sd_notify().
10200 * systemd-networkd gained support for:
10202 - Setting the IPv6 Router Advertisement settings via
10203 IPv6AcceptRouterAdvertisements= in .network files.
10205 - Configuring the HelloTimeSec=, MaxAgeSec= and
10206 ForwardDelaySec= bridge parameters in .netdev files.
10208 - Configuring PreferredSource= for static routes in
10211 * The "ask-password" framework used to query for LUKS harddisk
10212 passwords or SSL passwords during boot gained support for
10213 caching passwords in the kernel keyring, if it is
10214 available. This makes sure that the user only has to type in
10215 a passphrase once if there are multiple objects to unlock
10216 with the same one. Previously, such password caching was
10217 available only when Plymouth was used; this moves the
10218 caching logic into the systemd codebase itself. The
10219 "systemd-ask-password" utility gained a new --keyname=
10220 switch to control which kernel keyring key to use for
10221 caching a password in. This functionality is also useful for
10222 enabling display managers such as gdm to automatically
10223 unlock the user's GNOME keyring if its passphrase, the
10224 user's password and the harddisk password are the same, if
10225 gdm-autologin is used.
10227 * When downloading tar or raw images using "machinectl
10228 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
10229 file is now also downloaded, if it is available and stored
10230 next to the image file.
10232 * Units of type ".socket" gained a new boolean setting
10233 Writable= which is only useful in conjunction with
10234 ListenSpecial=. If true, enables opening the specified
10235 special file in O_RDWR mode rather than O_RDONLY mode.
10237 * systemd-rfkill has been reworked to become a singleton
10238 service that is activated through /dev/rfkill on each rfkill
10239 state change and saves the settings to disk. This way,
10240 systemd-rfkill is now compatible with devices that exist
10241 only intermittendly, and even restores state if the previous
10242 system shutdown was abrupt rather than clean.
10244 * The journal daemon gained support for vacuuming old journal
10245 files controlled by the number of files that shall remain,
10246 in addition to the already existing control by size and by
10247 date. This is useful as journal interleaving performance
10248 degrades with too many separate journal files, and allows
10249 putting an effective limit on them. The new setting defaults
10250 to 100, but this may be changed by setting SystemMaxFiles=
10251 and RuntimeMaxFiles= in journald.conf. Also, the
10252 "journalctl" tool gained the new --vacuum-files= switch to
10253 manually vacuum journal files to leave only the specified
10254 number of files in place.
10256 * udev will now create /dev/disk/by-path links for ATA devices
10257 on kernels where that is supported.
10259 * Galician, Serbian, Turkish and Korean translations were added.
10261 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
10262 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
10263 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
10264 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
10265 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
10266 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
10267 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
10268 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
10269 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
10270 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
10271 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
10272 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
10273 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
10274 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
10275 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
10276 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
10277 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
10278 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
10280 — Berlin, 2015-10-07
10284 * The DHCP implementation of systemd-networkd gained a set of
10287 - The DHCP server now supports emitting DNS and NTP
10288 information. It may be enabled and configured via
10289 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
10290 and NTP information is enabled, but no servers are
10291 configured, the corresponding uplink information (if there
10292 is any) is propagated.
10294 - Server and client now support transmission and reception
10295 of timezone information. It can be configured via the
10296 newly introduced network options UseTimezone=,
10297 EmitTimezone=, and Timezone=. Transmission of timezone
10298 information is enabled between host and containers by
10299 default now: the container will change its local timezone
10300 to what the host has set.
10302 - Lease timeouts can now be configured via
10303 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
10305 - The DHCP server improved on the stability of
10306 leases. Clients are more likely to get the same lease
10307 information back, even if the server loses state.
10309 - The DHCP server supports two new configuration options to
10310 control the lease address pool metrics, PoolOffset= and
10313 * The encapsulation limit of tunnels in systemd-networkd may
10314 now be configured via 'EncapsulationLimit='. It allows
10315 modifying the maximum additional levels of encapsulation
10316 that are permitted to be prepended to a packet.
10318 * systemd now supports the concept of user buses replacing
10319 session buses, if used with dbus-1.10 (and enabled via dbus
10320 --enable-user-session). It previously only supported this on
10321 kdbus-enabled systems, and this release expands this to
10322 'dbus-daemon' systems.
10324 * systemd-networkd now supports predictable interface names
10325 for virtio devices.
10327 * systemd now optionally supports the new Linux kernel
10328 "unified" control group hierarchy. If enabled via the kernel
10329 command-line option 'systemd.unified_cgroup_hierarchy=1',
10330 systemd will try to mount the unified cgroup hierarchy
10331 directly on /sys/fs/cgroup. If not enabled, or not
10332 available, systemd will fall back to the legacy cgroup
10333 hierarchy setup, as before. Host system and containers can
10334 mix and match legacy and unified hierarchies as they
10335 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
10336 environment variable to individually select the hierarchy to
10337 use for executed containers. By default, nspawn will use the
10338 unified hierarchy for the containers if the host uses the
10339 unified hierarchy, and the legacy hierarchy otherwise.
10340 Please note that at this point the unified hierarchy is an
10341 experimental kernel feature and is likely to change in one
10342 of the next kernel releases. Therefore, it should not be
10343 enabled by default in downstream distributions yet. The
10344 minimum required kernel version for the unified hierarchy to
10345 work is 4.2. Note that when the unified hierarchy is used
10346 for the first time delegated access to controllers is
10347 safe. Because of this systemd-nspawn containers will get
10348 access to controllers now, as will systemd user
10349 sessions. This means containers and user sessions may now
10350 manage their own resources, partitioning up what the system
10353 * A new special scope unit "init.scope" has been introduced
10354 that encapsulates PID 1 of the system. It may be used to
10355 determine resource usage and enforce resource limits on PID
10356 1 itself. PID 1 hence moved out of the root of the control
10359 * The cgtop tool gained support for filtering out kernel
10360 threads when counting tasks in a control group. Also, the
10361 count of processes is now recursively summed up by
10362 default. Two options -k and --recursive= have been added to
10363 revert to old behaviour. The tool has also been updated to
10364 work correctly in containers now.
10366 * systemd-nspawn's --bind= and --bind-ro= options have been
10367 extended to allow creation of non-recursive bind mounts.
10369 * libsystemd gained two new calls sd_pid_get_cgroup() and
10370 sd_peer_get_cgroup() which return the control group path of
10371 a process or peer of a connected AF_UNIX socket. This
10372 function call is particularly useful when implementing
10373 delegated subtrees support in the control group hierarchy.
10375 * The "sd-event" event loop API of libsystemd now supports
10376 correct dequeuing of real-time signals, without losing
10379 * When systemd requests a polkit decision when managing units it
10380 will now add additional fields to the request, including unit
10381 name and desired operation. This enables more powerful polkit
10382 policies, that make decisions depending on these parameters.
10384 * nspawn learnt support for .nspawn settings files, that may
10385 accompany the image files or directories of containers, and
10386 may contain additional settings for the container. This is
10387 an alternative to configuring container parameters via the
10388 nspawn command line.
10390 Contributions from: Cristian Rodríguez, Daniel Mack, David
10391 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
10392 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
10393 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
10394 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
10395 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
10396 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
10397 Andersen, Tom Gundersen, Torstein Husebø
10399 — Berlin, 2015-09-08
10403 * machinectl gained a new verb 'shell' which opens a fresh
10404 shell on the target container or the host. It is similar to
10405 the existing 'login' command of machinectl, but spawns the
10406 shell directly without prompting for username or
10407 password. The pseudo machine '.host' now refers to the local
10408 host and is used by default. Hence, 'machinectl shell' can
10409 be used as replacement for 'su -' which spawns a session as
10410 a fresh systemd unit in a way that is fully isolated from
10411 the originating session.
10413 * systemd-networkd learned to cope with private-zone DHCP
10414 options and allows other programs to query the values.
10416 * SELinux access control when enabling/disabling units is no
10417 longer enforced with this release. The previous implementation
10418 was incorrect, and a new corrected implementation is not yet
10419 available. As unit file operations are still protected via
10420 polkit and D-Bus policy this is not a security problem. Yet,
10421 distributions which care about optimal SELinux support should
10422 probably not stabilize on this release.
10424 * sd-bus gained support for matches of type "arg0has=", that
10425 test for membership of strings in string arrays sent in bus
10428 * systemd-resolved now dumps the contents of its DNS and LLMNR
10429 caches to the logs on reception of the SIGUSR1 signal. This
10430 is useful to debug DNS behaviour.
10432 * The coredumpctl tool gained a new --directory= option to
10433 operate on journal files in a specific directory.
10435 * "systemctl reboot" and related commands gained a new
10436 "--message=" option which may be used to set a free-text
10437 wall message when shutting down or rebooting the
10438 system. This message is also logged, which is useful for
10439 figuring out the reason for a reboot or shutdown a
10442 * The "systemd-resolve-host" tool's -i switch now takes
10443 network interface numbers as alternative to interface names.
10445 * A new unit file setting for services has been introduced:
10446 UtmpMode= allows configuration of how precisely systemd
10447 handles utmp and wtmp entries for the service if this is
10448 enabled. This allows writing services that appear similar to
10449 user sessions in the output of the "w", "who", "last" and
10452 * systemd-resolved will now locally synthesize DNS resource
10453 records for the "localhost" and "gateway" domains as well as
10454 the local hostname. This should ensure that clients querying
10455 RRs via resolved will get similar results as those going via
10456 NSS, if nss-myhostname is enabled.
10458 Contributions from: Alastair Hughes, Alex Crawford, Daniel
10459 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
10460 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
10461 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
10462 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
10463 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
10464 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
10465 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
10466 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
10467 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
10468 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
10469 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
10471 — Berlin, 2015-08-27
10475 * The systemd-efi-boot-generator functionality was merged into
10476 systemd-gpt-auto-generator.
10478 * systemd-networkd now supports Group Policy for vxlan
10479 devices. It can be enabled via the new boolean configuration
10480 option called 'GroupPolicyExtension='.
10482 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
10483 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
10484 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
10486 — Berlin, 2015-07-31
10490 * The python-systemd code has been removed from the systemd repository.
10491 A new repository has been created which accommodates the code from
10492 now on, and we kindly ask distributions to create a separate package
10493 for this: https://github.com/systemd/python-systemd
10495 * The systemd daemon will now reload its main configuration
10496 (/etc/systemd/system.conf) on daemon-reload.
10498 * sd-dhcp now exposes vendor specific extensions via
10499 sd_dhcp_lease_get_vendor_specific().
10501 * systemd-networkd gained a number of new configuration options.
10503 - A new boolean configuration option for TAP devices called
10504 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
10505 device, thus allowing to send and receive GSO packets.
10507 - A new tunnel configuration option called 'CopyDSCP='.
10508 If enabled, the DSCP field of ip6 tunnels is copied into the
10509 decapsulated packet.
10511 - A set of boolean bridge configuration options were added.
10512 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
10513 and 'UnicastFlood=' are now parsed by networkd and applied to the
10514 respective bridge link device via the respective IFLA_BRPORT_*
10517 - A new string configuration option to override the hostname sent
10518 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
10519 is true, networkd will use the configured hostname instead of the
10520 system hostname when sending DHCP requests.
10522 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
10523 networkd will configure the IPv6 flow-label of the tunnel device
10524 according to RFC2460.
10526 - The 'macvtap' virtual network devices are now supported, similar to
10527 the already supported 'macvlan' devices.
10529 * systemd-resolved now implements RFC5452 to improve resilience against
10530 cache poisoning. Additionally, source port randomization is enabled
10531 by default to further protect against DNS spoofing attacks.
10533 * nss-mymachines now supports translating UIDs and GIDs of running
10534 containers with user-namespaces enabled. If a container 'foo'
10535 translates a host uid 'UID' to the container uid 'TUID', then
10536 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
10537 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
10538 mapped as 'vg-foo-TGID'.
10540 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
10541 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
10542 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
10543 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
10544 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
10545 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
10546 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
10547 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
10548 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
10549 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
10551 — Berlin, 2015-07-29
10555 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
10556 There are no known issues with current sysfs, and udev does not need
10557 or should be used to work around such bugs.
10559 * udev does no longer enable USB HID power management. Several reports
10560 indicate, that some devices cannot handle that setting.
10562 * The udev accelerometer helper was removed. The functionality
10563 is now fully included in iio-sensor-proxy. But this means,
10564 older iio-sensor-proxy versions will no longer provide
10565 accelerometer/orientation data with this systemd version.
10566 Please upgrade iio-sensor-proxy to version 1.0.
10568 * networkd gained a new configuration option IPv6PrivacyExtensions=
10569 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
10570 for Stateless Address") on selected networks.
10572 * For the sake of fewer build-time dependencies and less code in the
10573 main repository, the python bindings are about to be removed in the
10574 next release. A new repository has been created which accommodates
10575 the code from now on, and we kindly ask distributions to create a
10576 separate package for this. The removal will take place in v223.
10578 https://github.com/systemd/python-systemd
10580 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
10581 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
10582 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
10583 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
10584 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
10585 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
10586 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
10587 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
10588 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
10589 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
10591 — Berlin, 2015-07-07
10595 * The sd-bus.h and sd-event.h APIs have now been declared
10596 stable and have been added to the official interface of
10597 libsystemd.so. sd-bus implements an alternative D-Bus client
10598 library, that is relatively easy to use, very efficient and
10599 supports both classic D-Bus as well as kdbus as transport
10600 backend. sd-event is a generic event loop abstraction that
10601 is built around Linux epoll, but adds features such as event
10602 prioritization or efficient timer handling. Both APIs are good
10603 choices for C programs looking for a bus and/or event loop
10604 implementation that is minimal and does not have to be
10605 portable to other kernels.
10607 * kdbus support is no longer compile-time optional. It is now
10608 always built-in. However, it can still be disabled at
10609 runtime using the kdbus=0 kernel command line setting, and
10610 that setting may be changed to default to off, by specifying
10611 --disable-kdbus at build-time. Note though that the kernel
10612 command line setting has no effect if the kdbus.ko kernel
10613 module is not installed, in which case kdbus is (obviously)
10614 also disabled. We encourage all downstream distributions to
10615 begin testing kdbus by adding it to the kernel images in the
10616 development distributions, and leaving kdbus support in
10619 * The minimal required util-linux version has been bumped to
10622 * Support for chkconfig (--enable-chkconfig) was removed in
10623 favor of calling an abstraction tool
10624 /lib/systemd/systemd-sysv-install. This needs to be
10625 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
10626 in README for details.
10628 * If there's a systemd unit and a SysV init script for the
10629 same service name, and the user executes "systemctl enable"
10630 for it (or a related call), then this will now enable both
10631 (or execute the related operation on both), not just the
10634 * The libudev API documentation has been converted from gtkdoc
10637 * gudev has been removed from the systemd tree, it is now an
10640 * The systemd-cgtop tool learnt a new --raw switch to generate
10641 "raw" (machine parsable) output.
10643 * networkd's IPForwarding= .network file setting learnt the
10644 new setting "kernel", which ensures that networkd does not
10645 change the IP forwarding sysctl from the default kernel
10648 * The systemd-logind bus API now exposes a new boolean
10649 property "Docked" that reports whether logind considers the
10650 system "docked", i.e. connected to a docking station or not.
10652 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
10653 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
10654 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
10655 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
10656 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
10657 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
10658 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
10659 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
10660 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
10661 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
10662 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
10663 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
10664 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
10665 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
10666 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
10667 Fink, Zbigniew Jędrzejewski-Szmek
10669 — Berlin, 2015-06-19
10673 * The gudev library has been extracted into a separate repository
10674 available at: https://git.gnome.org/browse/libgudev/
10675 It is now managed as part of the Gnome project. Distributions
10676 are recommended to pass --disable-gudev to systemd and use
10677 gudev from the Gnome project instead. gudev is still included
10678 in systemd, for now. It will be removed soon, though. Please
10679 also see the announcement-thread on systemd-devel:
10680 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
10682 * systemd now exposes a CPUUsageNSec= property for each
10683 service unit on the bus, that contains the overall consumed
10684 CPU time of a service (the sum of what each process of the
10685 service consumed). This value is only available if
10686 CPUAccounting= is turned on for a service, and is then shown
10687 in the "systemctl status" output.
10689 * Support for configuring alternative mappings of the old SysV
10690 runlevels to systemd targets has been removed. They are now
10691 hardcoded in a way that runlevels 2, 3, 4 all map to
10692 multi-user.target and 5 to graphical.target (which
10693 previously was already the default behaviour).
10695 * The auto-mounter logic gained support for mount point
10696 expiry, using a new TimeoutIdleSec= setting in .automount
10697 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
10699 * The EFI System Partition (ESP) as mounted to /boot by
10700 systemd-efi-boot-generator will now be unmounted
10701 automatically after 2 minutes of not being used. This should
10702 minimize the risk of ESP corruptions.
10704 * New /etc/fstab options x-systemd.requires= and
10705 x-systemd.requires-mounts-for= are now supported to express
10706 additional dependencies for mounts. This is useful for
10707 journaling file systems that support external journal
10708 devices or overlay file systems that require underlying file
10709 systems to be mounted.
10711 * systemd does not support direct live-upgrades (via systemctl
10712 daemon-reexec) from versions older than v44 anymore. As no
10713 distribution we are aware of shipped such old versions in a
10714 stable release this should not be problematic.
10716 * When systemd forks off a new per-connection service instance
10717 it will now set the $REMOTE_ADDR environment variable to the
10718 remote IP address, and $REMOTE_PORT environment variable to
10719 the remote IP port. This behaviour is similar to the
10720 corresponding environment variables defined by CGI.
10722 * systemd-networkd gained support for uplink failure
10723 detection. The BindCarrier= option allows binding interface
10724 configuration dynamically to the link sense of other
10725 interfaces. This is useful to achieve behaviour like in
10728 * systemd-networkd gained support for configuring the DHCP
10729 client identifier to use when requesting leases.
10731 * systemd-networkd now has a per-network UseNTP= option to
10732 configure whether NTP server information acquired via DHCP
10733 is passed on to services like systemd-timesyncd.
10735 * systemd-networkd gained support for vti6 tunnels.
10737 * Note that systemd-networkd manages the sysctl variable
10738 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
10739 it is configured for since v219. The variable controls IP
10740 forwarding, and is a per-interface alternative to the global
10741 /proc/sys/net/ipv[46]/ip_forward. This setting is
10742 configurable in the IPForward= option, which defaults to
10743 "no". This means if networkd is used for an interface it is
10744 no longer sufficient to set the global sysctl option to turn
10745 on IP forwarding! Instead, the .network file option
10746 IPForward= needs to be turned on! Note that the
10747 implementation of this behaviour was broken in v219 and has
10748 been fixed in v220.
10750 * Many bonding and vxlan options are now configurable in
10753 * systemd-nspawn gained a new --property= setting to set unit
10754 properties for the container scope. This is useful for
10755 setting resource parameters (e.g. "CPUShares=500") on
10756 containers started from the command line.
10758 * systemd-nspawn gained a new --private-users= switch to make
10759 use of user namespacing available on recent Linux kernels.
10761 * systemd-nspawn may now be called as part of a shell pipeline
10762 in which case the pipes used for stdin and stdout are passed
10763 directly to the process invoked in the container, without
10764 indirection via a pseudo tty.
10766 * systemd-nspawn gained a new switch to control the UNIX
10767 signal to use when killing the init process of the container
10768 when shutting down.
10770 * systemd-nspawn gained a new --overlay= switch for mounting
10771 overlay file systems into the container using the new kernel
10774 * When a container image is imported via systemd-importd and
10775 the host file system is not btrfs, a loopback block device
10776 file is created in /var/lib/machines.raw with a btrfs file
10777 system inside. It is then mounted to /var/lib/machines to
10778 enable btrfs features for container management. The loopback
10779 file and btrfs file system is grown as needed when container
10780 images are imported via systemd-importd.
10782 * systemd-machined/systemd-importd gained support for btrfs
10783 quota, to enforce container disk space limits on disk. This
10784 is exposed in "machinectl set-limit".
10786 * systemd-importd now can import containers from local .tar,
10787 .raw and .qcow2 images, and export them to .tar and .raw. It
10788 can also import dkr v2 images now from the network (on top
10791 * systemd-importd gained support for verifying downloaded
10792 images with gpg2 (previously only gpg1 was supported).
10794 * systemd-machined, systemd-logind, systemd: most bus calls are
10795 now accessible to unprivileged processes via polkit. Also,
10796 systemd-logind will now allow users to kill their own sessions
10797 without further privileges or authorization.
10799 * systemd-shutdownd has been removed. This service was
10800 previously responsible for implementing scheduled shutdowns
10801 as exposed in /usr/bin/shutdown's time parameter. This
10802 functionality has now been moved into systemd-logind and is
10803 accessible via a bus interface.
10805 * "systemctl reboot" gained a new switch --firmware-setup that
10806 can be used to reboot into the EFI firmware setup, if that
10807 is available. systemd-logind now exposes an API on the bus
10808 to trigger such reboots, in case graphical desktop UIs want
10809 to cover this functionality.
10811 * "systemctl enable", "systemctl disable" and "systemctl mask"
10812 now support a new "--now" switch. If specified the units
10813 that are enabled will also be started, and the ones
10814 disabled/masked also stopped.
10816 * The Gummiboot EFI boot loader tool has been merged into
10817 systemd, and renamed to "systemd-boot". The bootctl tool has been
10818 updated to support systemd-boot.
10820 * An EFI kernel stub has been added that may be used to create
10821 kernel EFI binaries that contain not only the actual kernel,
10822 but also an initrd, boot splash, command line and OS release
10823 information. This combined binary can then be signed as a
10824 single image, so that the firmware can verify it all in one
10825 step. systemd-boot has special support for EFI binaries created
10826 like this and can extract OS release information from them
10827 and show them in the boot menu. This functionality is useful
10828 to implement cryptographically verified boot schemes.
10830 * Optional support has been added to systemd-fsck to pass
10831 fsck's progress report to an AF_UNIX socket in the file
10834 * udev will no longer create device symlinks for all block devices by
10835 default. A deny list for excluding special block devices from this
10836 logic has been turned into an allow list that requires picking block
10837 devices explicitly that require device symlinks.
10839 * A new (currently still internal) API sd-device.h has been
10840 added to libsystemd. This modernized API is supposed to
10841 replace libudev eventually. In fact, already much of libudev
10842 is now just a wrapper around sd-device.h.
10844 * A new hwdb database for storing metadata about pointing
10845 stick devices has been added.
10847 * systemd-tmpfiles gained support for setting file attributes
10848 similar to the "chattr" tool with new 'h' and 'H' lines.
10850 * systemd-journald will no longer unconditionally set the
10851 btrfs NOCOW flag on new journal files. This is instead done
10852 with tmpfiles snippet using the new 'h' line type. This
10853 allows easy disabling of this logic, by masking the
10854 journal-nocow.conf tmpfiles file.
10856 * systemd-journald will now translate audit message types to
10857 human readable identifiers when writing them to the
10858 journal. This should improve readability of audit messages.
10860 * The LUKS logic gained support for the offset= and skip=
10861 options in /etc/crypttab, as previously implemented by
10864 * /usr/lib/os-release gained a new optional field VARIANT= for
10865 distributions that support multiple variants (such as a
10866 desktop edition, a server edition, …)
10868 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
10869 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
10870 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
10871 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
10872 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
10873 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
10874 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
10875 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
10876 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
10877 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
10878 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
10879 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
10880 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
10881 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
10882 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
10883 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
10884 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
10885 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
10886 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
10887 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
10888 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
10889 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
10890 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
10891 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
10892 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
10893 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
10894 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
10896 — Berlin, 2015-05-22
10900 * Introduce a new API "sd-hwdb.h" for querying the hardware
10901 metadata database. With this minimal interface one can query
10902 and enumerate the udev hwdb, decoupled from the old libudev
10903 library. libudev's interface for this is now only a wrapper
10904 around sd-hwdb. A new tool systemd-hwdb has been added to
10905 interface with and update the database.
10907 * When any of systemd's tools copies files (for example due to
10908 tmpfiles' C lines) a btrfs reflink will attempted first,
10909 before bytewise copying is done.
10911 * systemd-nspawn gained a new --ephemeral switch. When
10912 specified a btrfs snapshot is taken of the container's root
10913 directory, and immediately removed when the container
10914 terminates again. Thus, a container can be started whose
10915 changes never alter the container's root directory, and are
10916 lost on container termination. This switch can also be used
10917 for starting a container off the root file system of the
10918 host without affecting the host OS. This switch is only
10919 available on btrfs file systems.
10921 * systemd-nspawn gained a new --template= switch. It takes the
10922 path to a container tree to use as template for the tree
10923 specified via --directory=, should that directory be
10924 missing. This allows instantiating containers dynamically,
10925 on first run. This switch is only available on btrfs file
10928 * When a .mount unit refers to a mount point on which multiple
10929 mounts are stacked, and the .mount unit is stopped all of
10930 the stacked mount points will now be unmounted until no
10931 mount point remains.
10933 * systemd now has an explicit notion of supported and
10934 unsupported unit types. Jobs enqueued for unsupported unit
10935 types will now fail with an "unsupported" error code. More
10936 specifically .swap, .automount and .device units are not
10937 supported in containers, .busname units are not supported on
10938 non-kdbus systems. .swap and .automount are also not
10939 supported if their respective kernel compile time options
10942 * machinectl gained support for two new "copy-from" and
10943 "copy-to" commands for copying files from a running
10944 container to the host or vice versa.
10946 * machinectl gained support for a new "bind" command to bind
10947 mount host directories into local containers. This is
10948 currently only supported for nspawn containers.
10950 * networkd gained support for configuring bridge forwarding
10951 database entries (fdb) from .network files.
10953 * A new tiny daemon "systemd-importd" has been added that can
10954 download container images in tar, raw, qcow2 or dkr formats,
10955 and make them available locally in /var/lib/machines, so
10956 that they can run as nspawn containers. The daemon can GPG
10957 verify the downloads (not supported for dkr, since it has no
10958 provisions for verifying downloads). It will transparently
10959 decompress bz2, xz, gzip compressed downloads if necessary,
10960 and restore sparse files on disk. The daemon uses privilege
10961 separation to ensure the actual download logic runs with
10962 fewer privileges than the daemon itself. machinectl has
10963 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
10964 make the functionality of importd available to the
10965 user. With this in place the Fedora and Ubuntu "Cloud"
10966 images can be downloaded and booted as containers unmodified
10967 (the Fedora images lack the appropriate GPG signature files
10968 currently, so they cannot be verified, but this will change
10969 soon, hopefully). Note that downloading images is currently
10970 only fully supported on btrfs.
10972 * machinectl is now able to list container images found in
10973 /var/lib/machines, along with some metadata about sizes of
10974 disk and similar. If the directory is located on btrfs and
10975 quota is enabled, this includes quota display. A new command
10976 "image-status" has been added that shows additional
10977 information about images.
10979 * machinectl is now able to clone container images
10980 efficiently, if the underlying file system (btrfs) supports
10981 it, with the new "machinectl clone" command. It also
10982 gained commands for renaming and removing images, as well as
10983 marking them read-only or read-write (supported also on
10984 legacy file systems).
10986 * networkd gained support for collecting LLDP network
10987 announcements, from hardware that supports this. This is
10988 shown in networkctl output.
10990 * systemd-run gained support for a new -t (--pty) switch for
10991 invoking a binary on a pty whose input and output is
10992 connected to the invoking terminal. This allows executing
10993 processes as system services while interactively
10994 communicating with them via the terminal. Most interestingly
10995 this is supported across container boundaries. Invoking
10996 "systemd-run -t /bin/bash" is an alternative to running a
10997 full login session, the difference being that the former
10998 will not register a session, nor go through the PAM session
11001 * tmpfiles gained support for a new "v" line type for creating
11002 btrfs subvolumes. If the underlying file system is a legacy
11003 file system, this automatically degrades to creating a
11004 normal directory. Among others /var/lib/machines is now
11005 created like this at boot, should it be missing.
11007 * The directory /var/lib/containers/ has been deprecated and
11008 been replaced by /var/lib/machines. The term "machines" has
11009 been used in the systemd context as generic term for both
11010 VMs and containers, and hence appears more appropriate for
11011 this, as the directory can also contain raw images bootable
11014 * systemd-nspawn when invoked with -M but without --directory=
11015 or --image= is now capable of searching for the container
11016 root directory, subvolume or disk image automatically, in
11017 /var/lib/machines. systemd-nspawn@.service has been updated
11018 to make use of this, thus allowing it to be used for raw
11021 * A new machines.target unit has been introduced that is
11022 supposed to group all containers/VMs invoked as services on
11023 the system. systemd-nspawn@.service has been updated to
11024 integrate with that.
11026 * machinectl gained a new "start" command, for invoking a
11027 container as a service. "machinectl start foo" is mostly
11028 equivalent to "systemctl start systemd-nspawn@foo.service",
11029 but handles escaping in a nicer way.
11031 * systemd-nspawn will now mount most of the cgroupfs tree
11032 read-only into each container, with the exception of the
11033 container's own subtree in the name=systemd hierarchy.
11035 * journald now sets the special FS_NOCOW file flag for its
11036 journal files. This should improve performance on btrfs, by
11037 avoiding heavy fragmentation when journald's write-pattern
11038 is used on COW file systems. It degrades btrfs' data
11039 integrity guarantees for the files to the same levels as for
11040 ext3/ext4 however. This should be OK though as journald does
11041 its own data integrity checks and all its objects are
11042 checksummed on disk. Also, journald should handle btrfs disk
11043 full events a lot more gracefully now, by processing SIGBUS
11044 errors, and not relying on fallocate() anymore.
11046 * When journald detects that journal files it is writing to
11047 have been deleted it will immediately start new journal
11050 * systemd now provides a way to store file descriptors
11051 per-service in PID 1. This is useful for daemons to ensure
11052 that fds they require are not lost during a daemon
11053 restart. The fds are passed to the daemon on the next
11054 invocation in the same way socket activation fds are
11055 passed. This is now used by journald to ensure that the
11056 various sockets connected to all the system's stdout/stderr
11057 are not lost when journald is restarted. File descriptors
11058 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
11059 an extension to sd_notify(). Note that a limit is enforced
11060 on the number of fds a service can store in PID 1, and it
11061 defaults to 0, so that no fds may be stored, unless this is
11062 explicitly turned on.
11064 * The default TERM variable to use for units connected to a
11065 terminal, when no other value is explicitly is set is now
11066 vt220 rather than vt102. This should be fairly safe still,
11067 but allows PgUp/PgDn work.
11069 * The /etc/crypttab option header= as known from Debian is now
11072 * "loginctl user-status" and "loginctl session-status" will
11073 now show the last 10 lines of log messages of the
11074 user/session following the status output. Similar,
11075 "machinectl status" will show the last 10 log lines
11076 associated with a virtual machine or container
11077 service. (Note that this is usually not the log messages
11078 done in the VM/container itself, but simply what the
11079 container manager logs. For nspawn this includes all console
11082 * "loginctl session-status" without further argument will now
11083 show the status of the session of the caller. Similar,
11084 "lock-session", "unlock-session", "activate",
11085 "enable-linger", "disable-linger" may now be called without
11086 session/user parameter in which case they apply to the
11087 caller's session/user.
11089 * An X11 session scriptlet is now shipped that uploads
11090 $DISPLAY and $XAUTHORITY into the environment of the systemd
11091 --user daemon if a session begins. This should improve
11092 compatibility with X11 enabled applications run as systemd
11095 * Generators are now subject to masking via /etc and /run, the
11096 same way as unit files.
11098 * networkd .network files gained support for configuring
11099 per-link IPv4/IPv6 packet forwarding as well as IPv4
11100 masquerading. This is by default turned on for veth links to
11101 containers, as registered by systemd-nspawn. This means that
11102 nspawn containers run with --network-veth will now get
11103 automatic routed access to the host's networks without any
11104 further configuration or setup, as long as networkd runs on
11107 * systemd-nspawn gained the --port= (-p) switch to expose TCP
11108 or UDP posts of a container on the host. With this in place
11109 it is possible to run containers with private veth links
11110 (--network-veth), and have their functionality exposed on
11111 the host as if their services were running directly on the
11114 * systemd-nspawn's --network-veth switch now gained a short
11115 version "-n", since with the changes above it is now truly
11116 useful out-of-the-box. The systemd-nspawn@.service has been
11117 updated to make use of it too by default.
11119 * systemd-nspawn will now maintain a per-image R/W lock, to
11120 ensure that the same image is not started more than once
11121 writable. (It's OK to run an image multiple times
11122 simultaneously in read-only mode.)
11124 * systemd-nspawn's --image= option is now capable of
11125 dissecting and booting MBR and GPT disk images that contain
11126 only a single active Linux partition. Previously it
11127 supported only GPT disk images with proper GPT type
11128 IDs. This allows running cloud images from major
11129 distributions directly with systemd-nspawn, without
11132 * In addition to collecting mouse dpi data in the udev
11133 hardware database, there's now support for collecting angle
11134 information for mouse scroll wheels. The database is
11135 supposed to guarantee similar scrolling behavior on mice
11136 that it knows about. There's also support for collecting
11137 information about Touchpad types.
11139 * udev's input_id built-in will now also collect touch screen
11140 dimension data and attach it to probed devices.
11142 * /etc/os-release gained support for a Distribution Privacy
11145 * networkd gained support for creating "ipvlan", "gretap",
11146 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
11148 * systemd-tmpfiles gained support for "a" lines for setting
11151 * systemd-nspawn will now mount /tmp in the container to
11152 tmpfs, automatically.
11154 * systemd now exposes the memory.usage_in_bytes cgroup
11155 attribute and shows it for each service in the "systemctl
11156 status" output, if available.
11158 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
11159 immediate reboot is triggered. This useful if shutdown is
11160 hung and is unable to complete, to expedite the
11161 operation. Note that this kind of reboot will still unmount
11162 all file systems, and hence should not result in fsck being
11163 run on next reboot.
11165 * A .device unit for an optical block device will now be
11166 considered active only when a medium is in the drive. Also,
11167 mount units are now bound to their backing devices thus
11168 triggering automatic unmounting when devices become
11169 unavailable. With this in place systemd will now
11170 automatically unmount left-over mounts when a CD-ROM is
11171 ejected or an USB stick is yanked from the system.
11173 * networkd-wait-online now has support for waiting for
11174 specific interfaces only (with globbing), and for giving up
11175 after a configurable timeout.
11177 * networkd now exits when idle. It will be automatically
11178 restarted as soon as interfaces show up, are removed or
11179 change state. networkd will stay around as long as there is
11180 at least one DHCP state machine or similar around, that keep
11183 * networkd may now configure IPv6 link-local addressing in
11184 addition to IPv4 link-local addressing.
11186 * The IPv6 "token" for use in SLAAC may now be configured for
11187 each .network interface in networkd.
11189 * Routes configured with networkd may now be assigned a scope
11192 * networkd's [Match] sections now support globbing and lists
11193 of multiple space-separated matches per item.
11195 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
11196 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
11197 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
11198 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
11199 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
11200 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
11201 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
11202 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
11203 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
11204 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
11205 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
11206 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
11207 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
11208 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
11209 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
11210 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
11211 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
11212 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
11213 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
11214 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
11215 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
11216 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
11217 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
11218 Hoffmann, Zbigniew Jędrzejewski-Szmek
11220 — Berlin, 2015-02-16
11224 * When querying unit file enablement status (for example via
11225 "systemctl is-enabled"), a new state "indirect" is now known
11226 which indicates that a unit might not be enabled itself, but
11227 another unit listed in its Also= setting might be.
11229 * Similar to the various existing ConditionXYZ= settings for
11230 units, there are now matching AssertXYZ= settings. While
11231 failing conditions cause a unit to be skipped, but its job
11232 to succeed, failing assertions declared like this will cause
11233 a unit start operation and its job to fail.
11235 * hostnamed now knows a new chassis type "embedded".
11237 * systemctl gained a new "edit" command. When used on a unit
11238 file, this allows extending unit files with .d/ drop-in
11239 configuration snippets or editing the full file (after
11240 copying it from /usr/lib to /etc). This will invoke the
11241 user's editor (as configured with $EDITOR), and reload the
11242 modified configuration after editing.
11244 * "systemctl status" now shows the suggested enablement state
11245 for a unit, as declared in the (usually vendor-supplied)
11246 system preset files.
11248 * nss-myhostname will now resolve the single-label hostname
11249 "gateway" to the locally configured default IP routing
11250 gateways, ordered by their metrics. This assigns a stable
11251 name to the used gateways, regardless which ones are
11252 currently configured. Note that the name will only be
11253 resolved after all other name sources (if nss-myhostname is
11254 configured properly) and should hence not negatively impact
11255 systems that use the single-label hostname "gateway" in
11258 * systemd-inhibit now allows filtering by mode when listing
11261 * Scope and service units gained a new "Delegate" boolean
11262 property, which, when set, allows processes running inside the
11263 unit to further partition resources. This is primarily
11264 useful for systemd user instances as well as container
11267 * journald will now pick up audit messages directly from
11268 the kernel, and log them like any other log message. The
11269 audit fields are split up and fully indexed. This means that
11270 journalctl in many ways is now a (nicer!) alternative to
11271 ausearch, the traditional audit client. Note that this
11272 implements only a minimal audit client. If you want the
11273 special audit modes like reboot-on-log-overflow, please use
11274 the traditional auditd instead, which can be used in
11275 parallel to journald.
11277 * The ConditionSecurity= unit file option now understands the
11278 special string "audit" to check whether auditing is
11281 * journalctl gained two new commands --vacuum-size= and
11282 --vacuum-time= to delete old journal files until the
11283 remaining ones take up no more than the specified size on disk,
11284 or are not older than the specified time.
11286 * A new, native PPPoE library has been added to sd-network,
11287 systemd's library of light-weight networking protocols. This
11288 library will be used in a future version of networkd to
11289 enable PPPoE communication without an external pppd daemon.
11291 * The busctl tool now understands a new "capture" verb that
11292 works similar to "monitor", but writes a packet capture
11293 trace to STDOUT that can be redirected to a file which is
11294 compatible with libcap's capture file format. This can then
11295 be loaded in Wireshark and similar tools to inspect bus
11298 * The busctl tool now understands a new "tree" verb that shows
11299 the object trees of a specific service on the bus, or of all
11302 * The busctl tool now understands a new "introspect" verb that
11303 shows all interfaces and members of objects on the bus,
11304 including their signature and values. This is particularly
11305 useful to get more information about bus objects shown by
11306 the new "busctl tree" command.
11308 * The busctl tool now understands new verbs "call",
11309 "set-property" and "get-property" for invoking bus method
11310 calls, setting and getting bus object properties in a
11313 * busctl gained a new --augment-creds= argument that controls
11314 whether the tool shall augment credential information it
11315 gets from the bus with data from /proc, in a possibly
11318 * nspawn's --link-journal= switch gained two new values
11319 "try-guest" and "try-host" that work like "guest" and
11320 "host", but do not fail if the host has no persistent
11321 journaling enabled. -j is now equivalent to
11322 --link-journal=try-guest.
11324 * macvlan network devices created by nspawn will now have
11325 stable MAC addresses.
11327 * A new SmackProcessLabel= unit setting has been added, which
11328 controls the SMACK security label processes forked off by
11329 the respective unit shall use.
11331 * If compiled with --enable-xkbcommon, systemd-localed will
11332 verify x11 keymap settings by compiling the given keymap. It
11333 will spew out warnings if the compilation fails. This
11334 requires libxkbcommon to be installed.
11336 * When a coredump is collected, a larger number of metadata
11337 fields is now collected and included in the journal records
11338 created for it. More specifically, control group membership,
11339 environment variables, memory maps, working directory,
11340 chroot directory, /proc/$PID/status, and a list of open file
11341 descriptors is now stored in the log entry.
11343 * The udev hwdb now contains DPI information for mice. For
11346 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
11348 * All systemd programs that read standalone configuration
11349 files in /etc now also support a corresponding series of
11350 .conf.d configuration directories in /etc/, /run/,
11351 /usr/local/lib/, /usr/lib/, and (if configured with
11352 --enable-split-usr) /lib/. In particular, the following
11353 configuration files now have corresponding configuration
11354 directories: system.conf user.conf, logind.conf,
11355 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
11356 resolved.conf, timesyncd.conf, journal-remote.conf, and
11357 journal-upload.conf. Note that distributions should use the
11358 configuration directories in /usr/lib/; the directories in
11359 /etc/ are reserved for the system administrator.
11361 * systemd-rfkill will no longer take the rfkill device name
11362 into account when storing rfkill state on disk, as the name
11363 might be dynamically assigned and not stable. Instead, the
11364 ID_PATH udev variable combined with the rfkill type (wlan,
11365 bluetooth, …) is used.
11367 * A new service systemd-machine-id-commit.service has been
11368 added. When used on systems where /etc is read-only during
11369 boot, and /etc/machine-id is not initialized (but an empty
11370 file), this service will copy the temporary machine ID
11371 created as replacement into /etc after the system is fully
11372 booted up. This is useful for systems that are freshly
11373 installed with a non-initialized machine ID, but should get
11374 a fixed machine ID for subsequent boots.
11376 * networkd's .netdev files now provide a large set of
11377 configuration parameters for VXLAN devices. Similarly, the
11378 bridge port cost parameter is now configurable in .network
11379 files. There's also new support for configuring IP source
11380 routing. networkd .link files gained support for a new
11381 OriginalName= match that is useful to match against the
11382 original interface name the kernel assigned. .network files
11383 may include MTU= and MACAddress= fields for altering the MTU
11384 and MAC address while being connected to a specific network
11387 * The LUKS logic gained supported for configuring
11388 UUID-specific key files. There's also new support for naming
11389 LUKS device from the kernel command line, using the new
11390 luks.name= argument.
11392 * Timer units may now be transiently created via the bus API
11393 (this was previously already available for scope and service
11394 units). In addition it is now possible to create multiple
11395 transient units at the same time with a single bus call. The
11396 "systemd-run" tool has been updated to make use of this for
11397 running commands on a specified time, in at(1)-style.
11399 * tmpfiles gained support for "t" lines, for assigning
11400 extended attributes to files. Among other uses this may be
11401 used to assign SMACK labels to files.
11403 Contributions from: Alin Rauta, Alison Chaiken, Andrej
11404 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
11405 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
11406 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
11407 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
11408 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
11409 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
11410 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
11411 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
11412 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
11413 Michael Biebl, Michael Chapman, Michael Marineau, Michal
11414 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
11415 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
11416 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
11417 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
11418 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
11419 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
11420 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
11422 — Berlin, 2014-12-10
11426 * journalctl gained the new options -t/--identifier= to match
11427 on the syslog identifier (aka "tag"), as well as --utc to
11428 show log timestamps in the UTC timezone. journalctl now also
11429 accepts -n/--lines=all to disable line capping in a pager.
11431 * journalctl gained a new switch, --flush, that synchronously
11432 flushes logs from /run/log/journal to /var/log/journal if
11433 persistent storage is enabled. systemd-journal-flush.service
11434 now waits until the operation is complete.
11436 * Services can notify the manager before they start a reload
11437 (by sending RELOADING=1) or shutdown (by sending
11438 STOPPING=1). This allows the manager to track and show the
11439 internal state of daemons and closes a race condition when
11440 the process is still running but has closed its D-Bus
11443 * Services with Type=oneshot do not have to have any ExecStart
11446 * User units are now loaded also from
11447 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
11448 /run/systemd/user directory that was already previously
11449 supported, but is under the control of the user.
11451 * Job timeouts (i.e. timeouts on the time a job that is
11452 queued stays in the run queue) can now optionally result in
11453 immediate reboot or power-off actions (JobTimeoutAction= and
11454 JobTimeoutRebootArgument=). This is useful on ".target"
11455 units, to limit the maximum time a target remains
11456 undispatched in the run queue, and to trigger an emergency
11457 operation in such a case. This is now used by default to
11458 turn off the system if boot-up (as defined by everything in
11459 basic.target) hangs and does not complete for at least
11460 15min. Also, if power-off or reboot hang for at least 30min
11461 an immediate power-off/reboot operation is triggered. This
11462 functionality is particularly useful to increase reliability
11463 on embedded devices, but also on laptops which might
11464 accidentally get powered on when carried in a backpack and
11465 whose boot stays stuck in a hard disk encryption passphrase
11468 * systemd-logind can be configured to also handle lid switch
11469 events even when the machine is docked or multiple displays
11470 are attached (HandleLidSwitchDocked= option).
11472 * A helper binary and a service have been added which can be
11473 used to resume from hibernation in the initramfs. A
11474 generator will parse the resume= option on the kernel
11475 command line to trigger resume.
11477 * A user console daemon systemd-consoled has been
11478 added. Currently, it is a preview, and will so far open a
11479 single terminal on each session of the user marked as
11480 Desktop=systemd-console.
11482 * Route metrics can be specified for DHCP routes added by
11485 * The SELinux context of socket-activated services can be set
11486 from the information provided by the networking stack
11487 (SELinuxContextFromNet= option).
11489 * Userspace firmware loading support has been removed and
11490 the minimum supported kernel version is thus bumped to 3.7.
11492 * Timeout for udev workers has been increased from 1 to 3
11493 minutes, but a warning will be printed after 1 minute to
11494 help diagnose kernel modules that take a long time to load.
11496 * Udev rules can now remove tags on devices with TAG-="foobar".
11498 * systemd's readahead implementation has been removed. In many
11499 circumstances it didn't give expected benefits even for
11500 rotational disk drives and was becoming less relevant in the
11501 age of SSDs. As none of the developers has been using
11502 rotating media anymore, and nobody stepped up to actively
11503 maintain this component of systemd it has now been removed.
11505 * Swap units can use Options= to specify discard options.
11506 Discard options specified for swaps in /etc/fstab are now
11509 * Docker containers are now detected as a separate type of
11512 * The Password Agent protocol gained support for queries where
11513 the user input is shown, useful e.g. for user names.
11514 systemd-ask-password gained a new --echo option to turn that
11517 * The default sysctl.d/ snippets will now set:
11519 net.core.default_qdisc = fq_codel
11521 This selects Fair Queuing Controlled Delay as the default
11522 queuing discipline for network interfaces. fq_codel helps
11523 fight the network bufferbloat problem. It is believed to be
11524 a good default with no tuning required for most workloads.
11525 Downstream distributions may override this choice. On 10Gbit
11526 servers that do not do forwarding, "fq" may perform better.
11527 Systems without a good clocksource should use "pfifo_fast".
11529 * If kdbus is enabled during build a new option BusPolicy= is
11530 available for service units, that allows locking all service
11531 processes into a stricter bus policy, in order to limit
11532 access to various bus services, or even hide most of them
11533 from the service's view entirely.
11535 * networkctl will now show the .network and .link file
11536 networkd has applied to a specific interface.
11538 * sd-login gained a new API call sd_session_get_desktop() to
11539 query which desktop environment has been selected for a
11542 * UNIX utmp support is now compile-time optional to support
11543 legacy-free systems.
11545 * systemctl gained two new commands "add-wants" and
11546 "add-requires" for pulling in units from specific targets
11549 * If the word "rescue" is specified on the kernel command line
11550 the system will now boot into rescue mode (aka
11551 rescue.target), which was previously available only by
11552 specifying "1" or "systemd.unit=rescue.target" on the kernel
11553 command line. This new kernel command line option nicely
11554 mirrors the already existing "emergency" kernel command line
11557 * New kernel command line options mount.usr=, mount.usrflags=,
11558 mount.usrfstype= have been added that match root=, rootflags=,
11559 rootfstype= but allow mounting a specific file system to
11562 * The $NOTIFY_SOCKET is now also passed to control processes of
11563 services, not only the main process.
11565 * This version reenables support for fsck's -l switch. This
11566 means at least version v2.25 of util-linux is required for
11567 operation, otherwise dead-locks on device nodes may
11568 occur. Again: you need to update util-linux to at least
11569 v2.25 when updating systemd to v217.
11571 * The "multi-seat-x" tool has been removed from systemd, as
11572 its functionality has been integrated into X servers 1.16,
11573 and the tool is hence redundant. It is recommended to update
11574 display managers invoking this tool to simply invoke X
11575 directly from now on, again.
11577 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
11578 message flag has been added for all of systemd's polkit
11579 authenticated method calls has been added. In particular this
11580 now allows optional interactive authorization via polkit for
11581 many of PID1's privileged operations such as unit file
11582 enabling and disabling.
11584 * "udevadm hwdb --update" learnt a new switch "--usr" for
11585 placing the rebuilt hardware database in /usr instead of
11586 /etc. When used only hardware database entries stored in
11587 /usr will be used, and any user database entries in /etc are
11588 ignored. This functionality is useful for vendors to ship a
11589 pre-built database on systems where local configuration is
11590 unnecessary or unlikely.
11592 * Calendar time specifications in .timer units now also
11593 understand the strings "semi-annually", "quarterly" and
11594 "minutely" as shortcuts (in addition to the preexisting
11595 "annually", "hourly", …).
11597 * systemd-tmpfiles will now correctly create files in /dev
11598 at boot which are marked for creation only at boot. It is
11599 recommended to always create static device nodes with 'c!'
11600 and 'b!', so that they are created only at boot and not
11601 overwritten at runtime.
11603 * When the watchdog logic is used for a service (WatchdogSec=)
11604 and the watchdog timeout is hit the service will now be
11605 terminated with SIGABRT (instead of just SIGTERM), in order
11606 to make sure a proper coredump and backtrace is
11607 generated. This ensures that hanging services will result in
11608 similar coredump/backtrace behaviour as services that hit a
11609 segmentation fault.
11611 Contributions from: Andreas Henriksson, Andrei Borzenkov,
11612 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
11613 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
11614 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
11615 Herrmann, David Sommerseth, David Strauss, Emil Renner
11616 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
11617 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
11618 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
11619 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
11620 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
11621 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
11622 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
11623 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
11624 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
11625 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
11626 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
11627 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
11628 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
11629 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
11630 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
11631 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
11634 — Berlin, 2014-10-28
11638 * timedated no longer reads NTP implementation unit names from
11639 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
11640 implementations should add a
11642 Conflicts=systemd-timesyncd.service
11644 to their unit files to take over and replace systemd's NTP
11645 default functionality.
11647 * systemd-sysusers gained a new line type "r" for configuring
11648 which UID/GID ranges to allocate system users/groups
11649 from. Lines of type "u" may now add an additional column
11650 that specifies the home directory for the system user to be
11651 created. Also, systemd-sysusers may now optionally read user
11652 information from STDIN instead of a file. This is useful for
11653 invoking it from RPM preinst scriptlets that need to create
11654 users before the first RPM file is installed since these
11655 files might need to be owned by them. A new
11656 %sysusers_create_inline RPM macro has been introduced to do
11657 just that. systemd-sysusers now updates the shadow files as
11658 well as the user/group databases, which should enhance
11659 compatibility with certain tools like grpck.
11661 * A number of bus APIs of PID 1 now optionally consult polkit to
11662 permit access for otherwise unprivileged clients under certain
11663 conditions. Note that this currently doesn't support
11664 interactive authentication yet, but this is expected to be
11665 added eventually, too.
11667 * /etc/machine-info now has new fields for configuring the
11668 deployment environment of the machine, as well as the
11669 location of the machine. hostnamectl has been updated with
11670 new command to update these fields.
11672 * systemd-timesyncd has been updated to automatically acquire
11673 NTP server information from systemd-networkd, which might
11674 have been discovered via DHCP.
11676 * systemd-resolved now includes a caching DNS stub resolver
11677 and a complete LLMNR name resolution implementation. A new
11678 NSS module "nss-resolve" has been added which can be used
11679 instead of glibc's own "nss-dns" to resolve hostnames via
11680 systemd-resolved. Hostnames, addresses and arbitrary RRs may
11681 be resolved via systemd-resolved D-Bus APIs. In contrast to
11682 the glibc internal resolver systemd-resolved is aware of
11683 multi-homed system, and keeps DNS server and caches separate
11684 and per-interface. Queries are sent simultaneously on all
11685 interfaces that have DNS servers configured, in order to
11686 properly handle VPNs and local LANs which might resolve
11687 separate sets of domain names. systemd-resolved may acquire
11688 DNS server information from systemd-networkd automatically,
11689 which in turn might have discovered them via DHCP. A tool
11690 "systemd-resolve-host" has been added that may be used to
11691 query the DNS logic in resolved. systemd-resolved implements
11692 IDNA and automatically uses IDNA or UTF-8 encoding depending
11693 on whether classic DNS or LLMNR is used as transport. In the
11694 next releases we intend to add a DNSSEC and mDNS/DNS-SD
11695 implementation to systemd-resolved.
11697 * A new NSS module nss-mymachines has been added, that
11698 automatically resolves the names of all local registered
11699 containers to their respective IP addresses.
11701 * A new client tool "networkctl" for systemd-networkd has been
11702 added. It currently is entirely passive and will query
11703 networking configuration from udev, rtnetlink and networkd,
11704 and present it to the user in a very friendly
11705 way. Eventually, we hope to extend it to become a full
11706 control utility for networkd.
11708 * .socket units gained a new DeferAcceptSec= setting that
11709 controls the kernels' TCP_DEFER_ACCEPT sockopt for
11710 TCP. Similarly, support for controlling TCP keep-alive
11711 settings has been added (KeepAliveTimeSec=,
11712 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
11713 turning off Nagle's algorithm on TCP has been added
11716 * logind learned a new session type "web", for use in projects
11717 like Cockpit which register web clients as PAM sessions.
11719 * timer units with at least one OnCalendar= setting will now
11720 be started only after time-sync.target has been
11721 reached. This way they will not elapse before the system
11722 clock has been corrected by a local NTP client or
11723 similar. This is particular useful on RTC-less embedded
11724 machines, that come up with an invalid system clock.
11726 * systemd-nspawn's --network-veth= switch should now result in
11727 stable MAC addresses for both the outer and the inner side
11730 * systemd-nspawn gained a new --volatile= switch for running
11731 container instances with /etc or /var unpopulated.
11733 * The kdbus client code has been updated to use the new Linux
11734 3.17 memfd subsystem instead of the old kdbus-specific one.
11736 * systemd-networkd's DHCP client and server now support
11737 FORCERENEW. There are also new configuration options to
11738 configure the vendor client identifier and broadcast mode
11741 * systemd will no longer inform the kernel about the current
11742 timezone, as this is necessarily incorrect and racy as the
11743 kernel has no understanding of DST and similar
11744 concepts. This hence means FAT timestamps will be always
11745 considered UTC, similar to what Android is already
11746 doing. Also, when the RTC is configured to the local time
11747 (rather than UTC) systemd will never synchronize back to it,
11748 as this might confuse Windows at a later boot.
11750 * systemd-analyze gained a new command "verify" for offline
11751 validation of unit files.
11753 * systemd-networkd gained support for a couple of additional
11754 settings for bonding networking setups. Also, the metric for
11755 statically configured routes may now be configured. For
11756 network interfaces where this is appropriate the peer IP
11757 address may now be configured.
11759 * systemd-networkd's DHCP client will no longer request
11760 broadcasting by default, as this tripped up some networks.
11761 For hardware where broadcast is required the feature should
11762 be switched back on using RequestBroadcast=yes.
11764 * systemd-networkd will now set up IPv4LL addresses (when
11765 enabled) even if DHCP is configured successfully.
11767 * udev will now default to respect network device names given
11768 by the kernel when the kernel indicates that these are
11769 predictable. This behavior can be tweaked by changing
11770 NamePolicy= in the relevant .link file.
11772 * A new library systemd-terminal has been added that
11773 implements full TTY stream parsing and rendering. This
11774 library is supposed to be used later on for implementing a
11775 full userspace VT subsystem, replacing the current kernel
11778 * A new tool systemd-journal-upload has been added to push
11779 journal data to a remote system running
11780 systemd-journal-remote.
11782 * journald will no longer forward all local data to another
11783 running syslog daemon. This change has been made because
11784 rsyslog (which appears to be the most commonly used syslog
11785 implementation these days) no longer makes use of this, and
11786 instead pulls the data out of the journal on its own. Since
11787 forwarding the messages to a non-existent syslog server is
11788 more expensive than we assumed we have now turned this
11789 off. If you run a syslog server that is not a recent rsyslog
11790 version, you have to turn this option on again
11791 (ForwardToSyslog= in journald.conf).
11793 * journald now optionally supports the LZ4 compressor for
11794 larger journal fields. This compressor should perform much
11795 better than XZ which was the previous default.
11797 * machinectl now shows the IP addresses of local containers,
11798 if it knows them, plus the interface name of the container.
11800 * A new tool "systemd-escape" has been added that makes it
11801 easy to escape strings to build unit names and similar.
11803 * sd_notify() messages may now include a new ERRNO= field
11804 which is parsed and collected by systemd and shown among the
11805 "systemctl status" output for a service.
11807 * A new component "systemd-firstboot" has been added that
11808 queries the most basic systemd information (timezone,
11809 hostname, root password) interactively on first
11810 boot. Alternatively it may also be used to provision these
11811 things offline on OS images installed into directories.
11813 * The default sysctl.d/ snippets will now set
11815 net.ipv4.conf.default.promote_secondaries=1
11817 This has the benefit of no flushing secondary IP addresses
11818 when primary addresses are removed.
11820 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
11821 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
11822 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
11823 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
11824 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
11825 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
11826 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
11827 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
11828 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
11829 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
11830 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
11831 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
11832 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
11833 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
11834 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
11836 — Berlin, 2014-08-19
11840 * A new tool systemd-sysusers has been added. This tool
11841 creates system users and groups in /etc/passwd and
11842 /etc/group, based on static declarative system user/group
11843 definitions in /usr/lib/sysusers.d/. This is useful to
11844 enable factory resets and volatile systems that boot up with
11845 an empty /etc directory, and thus need system users and
11846 groups created during early boot. systemd now also ships
11847 with two default sysusers.d/ files for the most basic
11848 users and groups systemd and the core operating system
11851 * A new tmpfiles snippet has been added that rebuilds the
11852 essential files in /etc on boot, should they be missing.
11854 * A directive for ensuring automatic clean-up of
11855 /var/cache/man/ has been removed from the default
11856 configuration. This line should now be shipped by the man
11857 implementation. The necessary change has been made to the
11858 man-db implementation. Note that you need to update your man
11859 implementation to one that ships this line, otherwise no
11860 automatic clean-up of /var/cache/man will take place.
11862 * A new condition ConditionNeedsUpdate= has been added that
11863 may conditionalize services to only run when /etc or /var
11864 are "older" than the vendor operating system resources in
11865 /usr. This is useful for reconstructing or updating /etc
11866 after an offline update of /usr or a factory reset, on the
11867 next reboot. Services that want to run once after such an
11868 update or reset should use this condition and order
11869 themselves before the new systemd-update-done.service, which
11870 will mark the two directories as fully updated. A number of
11871 service files have been added making use of this, to rebuild
11872 the udev hardware database, the journald message catalog and
11873 dynamic loader cache (ldconfig). The systemd-sysusers tool
11874 described above also makes use of this now. With this in
11875 place it is now possible to start up a minimal operating
11876 system with /etc empty cleanly. For more information on the
11877 concepts involved see this recent blog story:
11879 http://0pointer.de/blog/projects/stateless.html
11881 * A new system group "input" has been introduced, and all
11882 input device nodes get this group assigned. This is useful
11883 for system-level software to get access to input devices. It
11884 complements what is already done for "audio" and "video".
11886 * systemd-networkd learnt minimal DHCPv4 server support in
11887 addition to the existing DHCPv4 client support. It also
11888 learnt DHCPv6 client and IPv6 Router Solicitation client
11889 support. The DHCPv4 client gained support for static routes
11890 passed in from the server. Note that the [DHCPv4] section
11891 known in older systemd-networkd versions has been renamed to
11892 [DHCP] and is now also used by the DHCPv6 client. Existing
11893 .network files using settings of this section should be
11894 updated, though compatibility is maintained. Optionally, the
11895 client hostname may now be sent to the DHCP server.
11897 * networkd gained support for vxlan virtual networks as well
11898 as tun/tap and dummy devices.
11900 * networkd gained support for automatic allocation of address
11901 ranges for interfaces from a system-wide pool of
11902 addresses. This is useful for dynamically managing a large
11903 number of interfaces with a single network configuration
11904 file. In particular this is useful to easily assign
11905 appropriate IP addresses to the veth links of a large number
11906 of nspawn instances.
11908 * RPM macros for processing sysusers, sysctl and binfmt
11909 drop-in snippets at package installation time have been
11912 * The /etc/os-release file should now be placed in
11913 /usr/lib/os-release. The old location is automatically
11914 created as symlink. /usr/lib is the more appropriate
11915 location of this file, since it shall actually describe the
11916 vendor operating system shipped in /usr, and not the
11917 configuration stored in /etc.
11919 * .mount units gained a new boolean SloppyOptions= setting
11920 that maps to mount(8)'s -s option which enables permissive
11921 parsing of unknown mount options.
11923 * tmpfiles learnt a new "L+" directive which creates a symlink
11924 but (unlike "L") deletes a pre-existing file first, should
11925 it already exist and not already be the correct
11926 symlink. Similarly, "b+", "c+" and "p+" directives have been
11927 added as well, which create block and character devices, as
11928 well as fifos in the filesystem, possibly removing any
11929 pre-existing files of different types.
11931 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
11932 'argument' field (which so far specified the source to
11933 symlink/copy the files from) is now optional. If omitted the
11934 same file os copied from /usr/share/factory/ suffixed by the
11935 full destination path. This is useful for populating /etc
11936 with essential files, by copying them from vendor defaults
11937 shipped in /usr/share/factory/etc.
11939 * A new command "systemctl preset-all" has been added that
11940 applies the service preset settings to all installed unit
11941 files. A new switch --preset-mode= has been added that
11942 controls whether only enable or only disable operations
11945 * A new command "systemctl is-system-running" has been added
11946 that allows checking the overall state of the system, for
11947 example whether it is fully up and running.
11949 * When the system boots up with an empty /etc, the equivalent
11950 to "systemctl preset-all" is executed during early boot, to
11951 make sure all default services are enabled after a factory
11954 * systemd now contains a minimal preset file that enables the
11955 most basic services systemd ships by default.
11957 * Unit files' [Install] section gained a new DefaultInstance=
11958 field for defining the default instance to create if a
11959 template unit is enabled with no instance specified.
11961 * A new passive target cryptsetup-pre.target has been added
11962 that may be used by services that need to make they run and
11963 finish before the first LUKS cryptographic device is set up.
11965 * The /dev/loop-control and /dev/btrfs-control device nodes
11966 are now owned by the "disk" group by default, opening up
11967 access to this group.
11969 * systemd-coredump will now automatically generate a
11970 stack trace of all core dumps taking place on the system,
11971 based on elfutils' libdw library. This stack trace is logged
11974 * systemd-coredump may now optionally store coredumps directly
11975 on disk (in /var/lib/systemd/coredump, possibly compressed),
11976 instead of storing them unconditionally in the journal. This
11977 mode is the new default. A new configuration file
11978 /etc/systemd/coredump.conf has been added to configure this
11979 and other parameters of systemd-coredump.
11981 * coredumpctl gained a new "info" verb to show details about a
11982 specific coredump. A new switch "-1" has also been added
11983 that makes sure to only show information about the most
11984 recent entry instead of all entries. Also, as the tool is
11985 generally useful now the "systemd-" prefix of the binary
11986 name has been removed. Distributions that want to maintain
11987 compatibility with the old name should add a symlink from
11988 the old name to the new name.
11990 * journald's SplitMode= now defaults to "uid". This makes sure
11991 that unprivileged users can access their own coredumps with
11992 coredumpctl without restrictions.
11994 * New kernel command line options "systemd.wants=" (for
11995 pulling an additional unit during boot), "systemd.mask="
11996 (for masking a specific unit for the boot), and
11997 "systemd.debug-shell" (for enabling the debug shell on tty9)
11998 have been added. This is implemented in the new generator
11999 "systemd-debug-generator".
12001 * systemd-nspawn will now by default filter a couple of
12002 syscalls for containers, among them those required for
12003 kernel module loading, direct x86 IO port access, swap
12004 management, and kexec. Most importantly though
12005 open_by_handle_at() is now prohibited for containers,
12006 closing a hole similar to a recently discussed vulnerability
12007 in docker regarding access to files on file hierarchies the
12008 container should normally not have access to. Note that, for
12009 nspawn, we generally make no security claims anyway (and
12010 this is explicitly documented in the man page), so this is
12011 just a fix for one of the most obvious problems.
12013 * A new man page file-hierarchy(7) has been added that
12014 contains a minimized, modernized version of the file system
12015 layout systemd expects, similar in style to the FHS
12016 specification or hier(5). A new tool systemd-path(1) has
12017 been added to query many of these paths for the local
12020 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
12021 longer done. Since the directory now has a per-user size
12022 limit, and is cleaned on logout this appears unnecessary,
12023 in particular since this now brings the lifecycle of this
12024 directory closer in line with how IPC objects are handled.
12026 * systemd.pc now exports a number of additional directories,
12027 including $libdir (which is useful to identify the library
12028 path for the primary architecture of the system), and a
12029 couple of drop-in directories.
12031 * udev's predictable network interface names now use the dev_port
12032 sysfs attribute, introduced in linux 3.15 instead of dev_id to
12033 distinguish between ports of the same PCI function. dev_id should
12034 only be used for ports using the same HW address, hence the need
12037 * machined has been updated to export the OS version of a
12038 container (read from /etc/os-release and
12039 /usr/lib/os-release) on the bus. This is now shown in
12040 "machinectl status" for a machine.
12042 * A new service setting RestartForceExitStatus= has been
12043 added. If configured to a set of exit signals or process
12044 return values, the service will be restarted when the main
12045 daemon process exits with any of them, regardless of the
12048 * systemctl's -H switch for connecting to remote systemd
12049 machines has been extended so that it may be used to
12050 directly connect to a specific container on the
12051 host. "systemctl -H root@foobar:waldi" will now connect as
12052 user "root" to host "foobar", and then proceed directly to
12053 the container named "waldi". Note that currently you have to
12054 authenticate as user "root" for this to work, as entering
12055 containers is a privileged operation.
12057 Contributions from: Andreas Henriksson, Benjamin Steinwender,
12058 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
12059 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
12060 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
12061 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
12062 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
12063 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
12064 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
12065 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
12066 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
12067 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
12068 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
12070 — Berlin, 2014-07-03
12074 * As an experimental feature, udev now tries to lock the
12075 disk device node (flock(LOCK_SH|LOCK_NB)) while it
12076 executes events for the disk or any of its partitions.
12077 Applications like partitioning programs can lock the
12078 disk device node (flock(LOCK_EX)) and claim temporary
12079 device ownership that way; udev will entirely skip all event
12080 handling for this disk and its partitions. If the disk
12081 was opened for writing, the close will trigger a partition
12082 table rescan in udev's "watch" facility, and if needed
12083 synthesize "change" events for the disk and all its partitions.
12084 This is now unconditionally enabled, and if it turns out to
12085 cause major problems, we might turn it on only for specific
12086 devices, or might need to disable it entirely. Device Mapper
12087 devices are excluded from this logic.
12089 * We temporarily dropped the "-l" switch for fsck invocations,
12090 since they collide with the flock() logic above. util-linux
12091 upstream has been changed already to avoid this conflict,
12092 and we will re-add "-l" as soon as util-linux with this
12093 change has been released.
12095 * The dependency on libattr has been removed. Since a long
12096 time, the extended attribute calls have moved to glibc, and
12097 libattr is thus unnecessary.
12099 * Virtualization detection works without privileges now. This
12100 means the systemd-detect-virt binary no longer requires
12101 CAP_SYS_PTRACE file capabilities, and our daemons can run
12102 with fewer privileges.
12104 * systemd-networkd now runs under its own "systemd-network"
12105 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
12106 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
12107 loses the ability to write to files owned by root this way.
12109 * Similarly, systemd-resolved now runs under its own
12110 "systemd-resolve" user with no capabilities remaining.
12112 * Similarly, systemd-bus-proxyd now runs under its own
12113 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
12115 * systemd-networkd gained support for setting up "veth"
12116 virtual Ethernet devices for container connectivity, as well
12117 as GRE and VTI tunnels.
12119 * systemd-networkd will no longer automatically attempt to
12120 manually load kernel modules necessary for certain tunnel
12121 transports. Instead, it is assumed the kernel loads them
12122 automatically when required. This only works correctly on
12123 very new kernels. On older kernels, please consider adding
12124 the kernel modules to /etc/modules-load.d/ as a work-around.
12126 * The resolv.conf file systemd-resolved generates has been
12127 moved to /run/systemd/resolve/. If you have a symlink from
12128 /etc/resolv.conf, it might be necessary to correct it.
12130 * Two new service settings, ProtectHome= and ProtectSystem=,
12131 have been added. When enabled, they will make the user data
12132 (such as /home) inaccessible or read-only and the system
12133 (such as /usr) read-only, for specific services. This allows
12134 very light-weight per-service sandboxing to avoid
12135 modifications of user data or system files from
12136 services. These two new switches have been enabled for all
12137 of systemd's long-running services, where appropriate.
12139 * Socket units gained new SocketUser= and SocketGroup=
12140 settings to set the owner user and group of AF_UNIX sockets
12141 and FIFOs in the file system.
12143 * Socket units gained a new RemoveOnStop= setting. If enabled,
12144 all FIFOS and sockets in the file system will be removed
12145 when the specific socket unit is stopped.
12147 * Socket units gained a new Symlinks= setting. It takes a list
12148 of symlinks to create to file system sockets or FIFOs
12149 created by the specific Unix sockets. This is useful to
12150 manage symlinks to socket nodes with the same lifecycle as
12153 * The /dev/log socket and /dev/initctl FIFO have been moved to
12154 /run, and have been replaced by symlinks. This allows
12155 connecting to these facilities even if PrivateDevices=yes is
12156 used for a service (which makes /dev/log itself unavailable,
12157 but /run is left). This also has the benefit of ensuring
12158 that /dev only contains device nodes, directories and
12159 symlinks, and nothing else.
12161 * sd-daemon gained two new calls sd_pid_notify() and
12162 sd_pid_notifyf(). They are similar to sd_notify() and
12163 sd_notifyf(), but allow overriding of the source PID of
12164 notification messages if permissions permit this. This is
12165 useful to send notify messages on behalf of a different
12166 process (for example, the parent process). The
12167 systemd-notify tool has been updated to make use of this
12168 when sending messages (so that notification messages now
12169 originate from the shell script invoking systemd-notify and
12170 not the systemd-notify process itself. This should minimize
12171 a race where systemd fails to associate notification
12172 messages to services when the originating process already
12175 * A new "on-abnormal" setting for Restart= has been added. If
12176 set, it will result in automatic restarts on all "abnormal"
12177 reasons for a process to exit, which includes unclean
12178 signals, core dumps, timeouts and watchdog timeouts, but
12179 does not include clean and unclean exit codes or clean
12180 signals. Restart=on-abnormal is an alternative for
12181 Restart=on-failure for services that shall be able to
12182 terminate and avoid restarts on certain errors, by
12183 indicating so with an unclean exit code. Restart=on-failure
12184 or Restart=on-abnormal is now the recommended setting for
12185 all long-running services.
12187 * If the InaccessibleDirectories= service setting points to a
12188 mount point (or if there are any submounts contained within
12189 it), it is now attempted to completely unmount it, to make
12190 the file systems truly unavailable for the respective
12193 * The ReadOnlyDirectories= service setting and
12194 systemd-nspawn's --read-only parameter are now recursively
12195 applied to all submounts, too.
12197 * Mount units may now be created transiently via the bus APIs.
12199 * The support for SysV and LSB init scripts has been removed
12200 from the systemd daemon itself. Instead, it is now
12201 implemented as a generator that creates native systemd units
12202 from these scripts when needed. This enables us to remove a
12203 substantial amount of legacy code from PID 1, following the
12204 fact that many distributions only ship a very small number
12205 of LSB/SysV init scripts nowadays.
12207 * Privileged Xen (dom0) domains are not considered
12208 virtualization anymore by the virtualization detection
12209 logic. After all, they generally have unrestricted access to
12210 the hardware and usually are used to manage the unprivileged
12213 * systemd-tmpfiles gained a new "C" line type, for copying
12214 files or entire directories.
12216 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
12217 lines. So far, they have been non-globbing versions of the
12218 latter, and have thus been redundant. In future, it is
12219 recommended to only use "z". "m" has hence been removed
12220 from the documentation, even though it stays supported.
12222 * A tmpfiles snippet to recreate the most basic structure in
12223 /var has been added. This is enough to create the /var/run →
12224 /run symlink and create a couple of structural
12225 directories. This allows systems to boot up with an empty or
12226 volatile /var. Of course, while with this change, the core OS
12227 now is capable with dealing with a volatile /var, not all
12228 user services are ready for it. However, we hope that sooner
12229 or later, many service daemons will be changed upstream so
12230 that they are able to automatically create their necessary
12231 directories in /var at boot, should they be missing. This is
12232 the first step to allow state-less systems that only require
12233 the vendor image for /usr to boot.
12235 * systemd-nspawn has gained a new --tmpfs= switch to mount an
12236 empty tmpfs instance to a specific directory. This is
12237 particularly useful for making use of the automatic
12238 reconstruction of /var (see above), by passing --tmpfs=/var.
12240 * Access modes specified in tmpfiles snippets may now be
12241 prefixed with "~", which indicates that they shall be masked
12242 by whether the existing file or directory is currently
12243 writable, readable or executable at all. Also, if specified,
12244 the sgid/suid/sticky bits will be masked for all
12247 * A new passive target unit "network-pre.target" has been
12248 added which is useful for services that shall run before any
12249 network is configured, for example firewall scripts.
12251 * The "floppy" group that previously owned the /dev/fd*
12252 devices is no longer used. The "disk" group is now used
12253 instead. Distributions should probably deprecate usage of
12256 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
12257 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
12258 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
12259 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
12260 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
12261 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
12264 — Berlin, 2014-06-11
12268 * A new "systemd-timesyncd" daemon has been added for
12269 synchronizing the system clock across the network. It
12270 implements an SNTP client. In contrast to NTP
12271 implementations such as chrony or the NTP reference server,
12272 this only implements a client side, and does not bother with
12273 the full NTP complexity, focusing only on querying time from
12274 one remote server and synchronizing the local clock to
12275 it. Unless you intend to serve NTP to networked clients or
12276 want to connect to local hardware clocks, this simple NTP
12277 client should be more than appropriate for most
12278 installations. The daemon runs with minimal privileges, and
12279 has been hooked up with networkd to only operate when
12280 network connectivity is available. The daemon saves the
12281 current clock to disk every time a new NTP sync has been
12282 acquired, and uses this to possibly correct the system clock
12283 early at bootup, in order to accommodate for systems that
12284 lack an RTC such as the Raspberry Pi and embedded devices,
12285 and to make sure that time monotonically progresses on these
12286 systems, even if it is not always correct. To make use of
12287 this daemon, a new system user and group "systemd-timesync"
12288 needs to be created on installation of systemd.
12290 * The queue "seqnum" interface of libudev has been disabled, as
12291 it was generally incompatible with device namespacing as
12292 sequence numbers of devices go "missing" if the devices are
12293 part of a different namespace.
12295 * "systemctl list-timers" and "systemctl list-sockets" gained
12296 a --recursive switch for showing units of these types also
12297 for all local containers, similar in style to the already
12298 supported --recursive switch for "systemctl list-units".
12300 * A new RebootArgument= setting has been added for service
12301 units, which may be used to specify a kernel reboot argument
12302 to use when triggering reboots with StartLimitAction=.
12304 * A new FailureAction= setting has been added for service
12305 units which may be used to specify an operation to trigger
12306 when a service fails. This works similarly to
12307 StartLimitAction=, but unlike it, controls what is done
12308 immediately rather than only after several attempts to
12309 restart the service in question.
12311 * hostnamed got updated to also expose the kernel name,
12312 release, and version on the bus. This is useful for
12313 executing commands like hostnamectl with the -H switch.
12314 systemd-analyze makes use of this to properly display
12315 details when running non-locally.
12317 * The bootchart tool can now show cgroup information in the
12318 graphs it generates.
12320 * The CFS CPU quota cgroup attribute is now exposed for
12321 services. The new CPUQuota= switch has been added for this
12322 which takes a percentage value. Setting this will have the
12323 result that a service may never get more CPU time than the
12324 specified percentage, even if the machine is otherwise idle.
12326 * systemd-networkd learned IPIP and SIT tunnel support.
12328 * LSB init scripts exposing a dependency on $network will now
12329 get a dependency on network-online.target rather than simply
12330 network.target. This should bring LSB handling closer to
12331 what it was on SysV systems.
12333 * A new fsck.repair= kernel option has been added to control
12334 how fsck shall deal with unclean file systems at boot.
12336 * The (.ini) configuration file parser will now silently ignore
12337 sections whose names begin with "X-". This may be used to maintain
12338 application-specific extension sections in unit files.
12340 * machined gained a new API to query the IP addresses of
12341 registered containers. "machinectl status" has been updated
12342 to show these addresses in its output.
12344 * A new call sd_uid_get_display() has been added to the
12345 sd-login APIs for querying the "primary" session of a
12346 user. The "primary" session of the user is elected from the
12347 user's sessions and generally a graphical session is
12348 preferred over a text one.
12350 * A minimal systemd-resolved daemon has been added. It
12351 currently simply acts as a companion to systemd-networkd and
12352 manages resolv.conf based on per-interface DNS
12353 configuration, possibly supplied via DHCP. In the long run
12354 we hope to extend this into a local DNSSEC enabled DNS and
12357 * The systemd-networkd-wait-online tool is now enabled by
12358 default. It will delay network-online.target until a network
12359 connection has been configured. The tool primarily integrates
12360 with networkd, but will also make a best effort to make sense
12361 of network configuration performed in some other way.
12363 * Two new service options StartupCPUShares= and
12364 StartupBlockIOWeight= have been added that work similarly to
12365 CPUShares= and BlockIOWeight= however only apply during
12366 system startup. This is useful to prioritize certain services
12367 differently during bootup than during normal runtime.
12369 * hostnamed has been changed to prefer the statically
12370 configured hostname in /etc/hostname (unless set to
12371 'localhost' or empty) over any dynamic one supplied by
12372 dhcp. With this change, the rules for picking the hostname
12373 match more closely the rules of other configuration settings
12374 where the local administrator's configuration in /etc always
12375 overrides any other settings.
12377 Contributions from: Ali H. Caliskan, Alison Chaiken, Bas van
12378 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
12379 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
12380 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
12381 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
12382 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
12383 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
12384 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
12385 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
12386 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
12387 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
12388 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
12389 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
12390 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
12391 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
12392 Lindskog, WaLyong Cho, Will Woods, Zbigniew
12395 — Beijing, 2014-05-28
12399 * When restoring the screen brightness at boot, stay away from
12400 the darkest setting or from the lowest 5% of the available
12401 range, depending on which is the larger value of both. This
12402 should effectively protect the user from rebooting into a
12403 black screen, should the brightness have been set to minimum
12406 * sd-login gained a new sd_machine_get_class() call to
12407 determine the class ("vm" or "container") of a machine
12408 registered with machined.
12410 * sd-login gained new calls
12411 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
12412 to query the identity of the peer of a local AF_UNIX
12413 connection. They operate similarly to their sd_pid_get_xyz()
12416 * PID 1 will now maintain a system-wide system state engine
12417 with the states "starting", "running", "degraded",
12418 "maintenance", "stopping". These states are bound to system
12419 startup, normal runtime, runtime with at least one failed
12420 service, rescue/emergency mode and system shutdown. This
12421 state is shown in the "systemctl status" output when no unit
12422 name is passed. It is useful to determine system state, in
12423 particularly when doing so for many systems or containers at
12426 * A new command "list-machines" has been added to "systemctl"
12427 that lists all local OS containers and shows their system
12428 state (see above), if systemd runs inside of them.
12430 * systemctl gained a new "-r" switch to recursively enumerate
12431 units on all local containers, when used with the
12432 "list-unit" command (which is the default one that is
12433 executed when no parameters are specified).
12435 * The GPT automatic partition discovery logic will now honour
12436 two GPT partition flags: one may be set on a partition to
12437 cause it to be mounted read-only, and the other may be set
12438 on a partition to ignore it during automatic discovery.
12440 * Two new GPT type UUIDs have been added for automatic root
12441 partition discovery, for 32-bit and 64-bit ARM. This is not
12442 particularly useful for discovering the root directory on
12443 these architectures during bare-metal boots (since UEFI is
12444 not common there), but still very useful to allow booting of
12445 ARM disk images in nspawn with the -i option.
12447 * MAC addresses of interfaces created with nspawn's
12448 --network-interface= switch will now be generated from the
12449 machine name, and thus be stable between multiple invocations
12452 * logind will now automatically remove all IPC objects owned
12453 by a user if she or he fully logs out. This makes sure that
12454 users who are logged out cannot continue to consume IPC
12455 resources. This covers SysV memory, semaphores and message
12456 queues as well as POSIX shared memory and message
12457 queues. Traditionally, SysV and POSIX IPC had no lifecycle
12458 limits. With this functionality, that is corrected. This may
12459 be turned off by using the RemoveIPC= switch of logind.conf.
12461 * The systemd-machine-id-setup and tmpfiles tools gained a
12462 --root= switch to operate on a specific root directory,
12465 * journald can now forward logged messages to the TTYs of all
12466 logged in users ("wall"). This is the default for all
12467 emergency messages now.
12469 * A new tool systemd-journal-remote has been added to stream
12470 journal log messages across the network.
12472 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
12473 controller trees are mounted into it. Note that the
12474 directories mounted beneath it are not read-only. This is a
12475 security measure and is particularly useful because glibc
12476 actually includes a search logic to pick any tmpfs it can
12477 find to implement shm_open() if /dev/shm is not available
12478 (which it might very well be in namespaced setups).
12480 * machinectl gained a new "poweroff" command to cleanly power
12481 down a local OS container.
12483 * The PrivateDevices= unit file setting will now also drop the
12484 CAP_MKNOD capability from the capability bound set, and
12485 imply DevicePolicy=closed.
12487 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
12488 comprehensively on all long-running systemd services where
12489 this is appropriate.
12491 * systemd-udevd will now run in a disassociated mount
12492 namespace. To mount directories from udev rules, make sure to
12493 pull in mount units via SYSTEMD_WANTS properties.
12495 * The kdbus support gained support for uploading policy into
12496 the kernel. sd-bus gained support for creating "monitoring"
12497 connections that can eavesdrop into all bus communication
12498 for debugging purposes.
12500 * Timestamps may now be specified in seconds since the UNIX
12501 epoch Jan 1st, 1970 by specifying "@" followed by the value
12504 * Native tcpwrap support in systemd has been removed. tcpwrap
12505 is old code, not really maintained anymore and has serious
12506 shortcomings, and better options such as firewalls
12507 exist. For setups that require tcpwrap usage, please
12508 consider invoking your socket-activated service via tcpd,
12509 like on traditional inetd.
12511 * A new system.conf configuration option
12512 DefaultTimerAccuracySec= has been added that controls the
12513 default AccuracySec= setting of .timer units.
12515 * Timer units gained a new WakeSystem= switch. If enabled,
12516 timers configured this way will cause the system to resume
12517 from system suspend (if the system supports that, which most
12520 * Timer units gained a new Persistent= switch. If enabled,
12521 timers configured this way will save to disk when they have
12522 been last triggered. This information is then used on next
12523 reboot to possible execute overdue timer events, that
12524 could not take place because the system was powered off.
12525 This enables simple anacron-like behaviour for timer units.
12527 * systemctl's "list-timers" will now also list the time a
12528 timer unit was last triggered in addition to the next time
12529 it will be triggered.
12531 * systemd-networkd will now assign predictable IPv4LL
12532 addresses to its local interfaces.
12534 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
12535 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
12536 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
12537 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
12538 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
12539 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
12540 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
12541 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
12544 — Berlin, 2014-03-25
12548 * A new unit file setting RestrictAddressFamilies= has been
12549 added to restrict which socket address families unit
12550 processes gain access to. This takes address family names
12551 like "AF_INET" or "AF_UNIX", and is useful to minimize the
12552 attack surface of services via exotic protocol stacks. This
12553 is built on seccomp system call filters.
12555 * Two new unit file settings RuntimeDirectory= and
12556 RuntimeDirectoryMode= have been added that may be used to
12557 manage a per-daemon runtime directories below /run. This is
12558 an alternative for setting up directory permissions with
12559 tmpfiles snippets, and has the advantage that the runtime
12560 directory's lifetime is bound to the daemon runtime and that
12561 the daemon starts up with an empty directory each time. This
12562 is particularly useful when writing services that drop
12563 privileges using the User= or Group= setting.
12565 * The DeviceAllow= unit setting now supports globbing for
12566 matching against device group names.
12568 * The systemd configuration file system.conf gained new
12569 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
12570 DefaultMemoryAccounting= to globally turn on/off accounting
12571 for specific resources (cgroups) for all units. These
12572 settings may still be overridden individually in each unit
12575 * systemd-gpt-auto-generator is now able to discover /srv and
12576 root partitions in addition to /home and swap partitions. It
12577 also supports LUKS-encrypted partitions now. With this in
12578 place, automatic discovery of partitions to mount following
12579 the Discoverable Partitions Specification
12580 (https://systemd.io/DISCOVERABLE_PARTITIONS/)
12581 is now a lot more complete. This allows booting without
12582 /etc/fstab and without root= on the kernel command line on
12583 systems prepared appropriately.
12585 * systemd-nspawn gained a new --image= switch which allows
12586 booting up disk images and Linux installations on any block
12587 device that follow the Discoverable Partitions Specification
12588 (see above). This means that installations made with
12589 appropriately updated installers may now be started and
12590 deployed using container managers, completely
12591 unmodified. (We hope that libvirt-lxc will add support for
12592 this feature soon, too.)
12594 * systemd-nspawn gained a new --network-macvlan= setting to
12595 set up a private macvlan interface for the
12596 container. Similarly, systemd-networkd gained a new
12597 Kind=macvlan setting in .netdev files.
12599 * systemd-networkd now supports configuring local addresses
12602 * A new tool systemd-network-wait-online has been added to
12603 synchronously wait for network connectivity using
12606 * The sd-bus.h bus API gained a new sd_bus_track object for
12607 tracking the lifecycle of bus peers. Note that sd-bus.h is
12608 still not a public API though (unless you specify
12609 --enable-kdbus on the configure command line, which however
12610 voids your warranty and you get no API stability guarantee).
12612 * The $XDG_RUNTIME_DIR runtime directories for each user are
12613 now individual tmpfs instances, which has the benefit of
12614 introducing separate pools for each user, with individual
12615 size limits, and thus making sure that unprivileged clients
12616 can no longer negatively impact the system or other users by
12617 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
12618 RuntimeDirectorySize= has been introduced that allows
12619 controlling the default size limit for all users. It
12620 defaults to 10% of the available physical memory. This is no
12621 replacement for quotas on tmpfs though (which the kernel
12622 still does not support), as /dev/shm and /tmp are still
12623 shared resources used by both the system and unprivileged
12626 * logind will now automatically turn off automatic suspending
12627 on laptop lid close when more than one display is
12628 connected. This was previously expected to be implemented
12629 individually in desktop environments (such as GNOME),
12630 however has been added to logind now, in order to fix a
12631 boot-time race where a desktop environment might not have
12632 been started yet and thus not been able to take an inhibitor
12633 lock at the time where logind already suspends the system
12634 due to a closed lid.
12636 * logind will now wait at least 30s after each system
12637 suspend/resume cycle, and 3min after system boot before
12638 suspending the system due to a closed laptop lid. This
12639 should give USB docking stations and similar enough time to
12640 be probed and configured after system resume and boot in
12641 order to then act as suspend blocker.
12643 * systemd-run gained a new --property= setting which allows
12644 initialization of resource control properties (and others)
12645 for the created scope or service unit. Example: "systemd-run
12646 --property=BlockIOWeight=10 updatedb" may be used to run
12647 updatedb at a low block IO scheduling weight.
12649 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
12650 now also work in --scope mode.
12652 * When systemd is compiled with kdbus support, basic support
12653 for enforced policies is now in place. (Note that enabling
12654 kdbus still voids your warranty and no API compatibility
12655 promises are made.)
12657 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
12658 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
12659 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
12660 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
12661 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
12662 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
12663 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
12664 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
12665 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
12666 Zbigniew Jędrzejewski-Szmek
12668 — Berlin, 2014-03-12
12672 * systemd will now relabel /dev after loading the SMACK policy
12673 according to SMACK rules.
12675 * A new unit file option AppArmorProfile= has been added to
12676 set the AppArmor profile for the processes of a unit.
12678 * A new condition check ConditionArchitecture= has been added
12679 to conditionalize units based on the system architecture, as
12680 reported by uname()'s "machine" field.
12682 * systemd-networkd now supports matching on the system
12683 virtualization, architecture, kernel command line, hostname
12686 * logind is now a lot more aggressive when suspending the
12687 machine due to a closed laptop lid. Instead of acting only
12688 on the lid close action, it will continuously watch the lid
12689 status and act on it. This is useful for laptops where the
12690 power button is on the outside of the chassis so that it can
12691 be reached without opening the lid (such as the Lenovo
12692 Yoga). On those machines, logind will now immediately
12693 re-suspend the machine if the power button has been
12694 accidentally pressed while the laptop was suspended and in a
12695 backpack or similar.
12697 * logind will now watch SW_DOCK switches and inhibit reaction
12698 to the lid switch if it is pressed. This means that logind
12699 will not suspend the machine anymore if the lid is closed
12700 and the system is docked, if the laptop supports SW_DOCK
12701 notifications via the input layer. Note that ACPI docking
12702 stations do not generate this currently. Also note that this
12703 logic is usually not fully sufficient and Desktop
12704 Environments should take a lid switch inhibitor lock when an
12705 external display is connected, as systemd will not watch
12708 * nspawn will now make use of the devices cgroup controller by
12709 default, and only permit creation of and access to the usual
12710 API device nodes like /dev/null or /dev/random, as well as
12711 access to (but not creation of) the pty devices.
12713 * We will now ship a default .network file for
12714 systemd-networkd that automatically configures DHCP for
12715 network interfaces created by nspawn's --network-veth or
12716 --network-bridge= switches.
12718 * systemd will now understand the usual M, K, G, T suffixes
12719 according to SI conventions (i.e. to the base 1000) when
12720 referring to throughput and hardware metrics. It will stay
12721 with IEC conventions (i.e. to the base 1024) for software
12722 metrics, according to what is customary according to
12723 Wikipedia. We explicitly document which base applies for
12724 each configuration option.
12726 * The DeviceAllow= setting in unit files now supports a syntax to
12727 allow-list an entire group of devices node majors at once, based on
12728 the /proc/devices listing. For example, with the string "char-pts",
12729 it is now possible to allow-list all current and future pseudo-TTYs
12732 * sd-event learned a new "post" event source. Event sources of
12733 this type are triggered by the dispatching of any event
12734 source of a type that is not "post". This is useful for
12735 implementing clean-up and check event sources that are
12736 triggered by other work being done in the program.
12738 * systemd-networkd is no longer statically enabled, but uses
12739 the usual [Install] sections so that it can be
12740 enabled/disabled using systemctl. It still is enabled by
12743 * When creating a veth interface pair with systemd-nspawn, the
12744 host side will now be prefixed with "vb-" if
12745 --network-bridge= is used, and with "ve-" if --network-veth
12746 is used. This way, it is easy to distinguish these cases on
12747 the host, for example to apply different configuration to
12748 them with systemd-networkd.
12750 * The compatibility libraries for libsystemd-journal.so,
12751 libsystem-id128.so, libsystemd-login.so and
12752 libsystemd-daemon.so do not make use of IFUNC
12753 anymore. Instead, we now build libsystemd.so multiple times
12754 under these alternative names. This means that the footprint
12755 is drastically increased, but given that these are
12756 transitional compatibility libraries, this should not matter
12757 much. This change has been made necessary to support the ARM
12758 platform for these compatibility libraries, as the ARM
12759 toolchain is not really at the same level as the toolchain
12760 for other architectures like x86 and does not support
12761 IFUNC. Please make sure to use --enable-compat-libs only
12762 during a transitional period!
12764 * The .include syntax has been deprecated and is not documented
12765 anymore. Drop-in files in .d directories should be used instead.
12767 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
12768 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
12769 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
12770 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
12771 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
12772 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
12773 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
12774 Zbigniew Jędrzejewski-Szmek
12776 — Berlin, 2014-02-24
12780 * A new component "systemd-networkd" has been added that can
12781 be used to configure local network interfaces statically or
12782 via DHCP. It is capable of bringing up bridges, VLANs, and
12783 bonding. Currently, no hook-ups for interactive network
12784 configuration are provided. Use this for your initrd,
12785 container, embedded, or server setup if you need a simple,
12786 yet powerful, network configuration solution. This
12787 configuration subsystem is quite nifty, as it allows wildcard
12788 hotplug matching in interfaces. For example, with a single
12789 configuration snippet, you can configure that all Ethernet
12790 interfaces showing up are automatically added to a bridge,
12791 or similar. It supports link-sensing and more.
12793 * A new tool "systemd-socket-proxyd" has been added which can
12794 act as a bidirectional proxy for TCP sockets. This is
12795 useful for adding socket activation support to services that
12796 do not actually support socket activation, including virtual
12797 machines and the like.
12799 * Add a new tool to save/restore rfkill state on
12802 * Save/restore state of keyboard backlights in addition to
12803 display backlights on shutdown/boot.
12805 * udev learned a new SECLABEL{} construct to label device
12806 nodes with a specific security label when they appear. For
12807 now, only SECLABEL{selinux} is supported, but the syntax is
12808 prepared for additional security frameworks.
12810 * udev gained a new scheme to configure link-level attributes
12811 from files in /etc/systemd/network/*.link. These files can
12812 match against MAC address, device path, driver name and type,
12813 and will apply attributes like the naming policy, link speed,
12814 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
12815 address assignment policy (randomized, …).
12817 * The configuration of network interface naming rules for
12818 "permanent interface names" has changed: a new NamePolicy=
12819 setting in the [Link] section of .link files determines the
12820 priority of possible naming schemes (onboard, slot, MAC,
12821 path). The default value of this setting is determined by
12822 /usr/lib/net/links/99-default.link. Old
12823 80-net-name-slot.rules udev configuration file has been
12824 removed, so local configuration overriding this file should
12825 be adapted to override 99-default.link instead.
12827 * When the User= switch is used in a unit file, also
12828 initialize $SHELL= based on the user database entry.
12830 * systemd no longer depends on libdbus. All communication is
12831 now done with sd-bus, systemd's low-level bus library
12834 * kdbus support has been added to PID 1 itself. When kdbus is
12835 enabled, this causes PID 1 to set up the system bus and
12836 enable support for a new ".busname" unit type that
12837 encapsulates bus name activation on kdbus. It works a little
12838 bit like ".socket" units, except for bus names. A new
12839 generator has been added that converts classic dbus1 service
12840 activation files automatically into native systemd .busname
12841 and .service units.
12843 * sd-bus: add a light-weight vtable implementation that allows
12844 defining objects on the bus with a simple static const
12845 vtable array of its methods, signals and properties.
12847 * systemd will not generate or install static dbus
12848 introspection data anymore to /usr/share/dbus-1/interfaces,
12849 as the precise format of these files is unclear, and
12850 nothing makes use of it.
12852 * A proxy daemon is now provided to proxy clients connecting
12853 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
12854 compatibility with classic D-Bus.
12856 * A bus driver implementation has been added that supports the
12857 classic D-Bus bus driver calls on kdbus, also for
12858 compatibility purposes.
12860 * A new API "sd-event.h" has been added that implements a
12861 minimal event loop API built around epoll. It provides a
12862 couple of features that direct epoll usage is lacking:
12863 prioritization of events, scales to large numbers of timer
12864 events, per-event timer slack (accuracy), system-wide
12865 coalescing of timer events, exit handlers, watchdog
12866 supervision support using systemd's sd_notify() API, child
12869 * A new API "sd-rntl.h" has been added that provides an API
12870 around the route netlink interface of the kernel, similar in
12871 style to "sd-bus.h".
12873 * A new API "sd-dhcp-client.h" has been added that provides a
12874 small DHCPv4 client-side implementation. This is used by
12875 "systemd-networkd".
12877 * There is a new kernel command line option
12878 "systemd.restore_state=0|1". When set to "0", none of the
12879 systemd tools will restore saved runtime state to hardware
12880 devices. More specifically, the rfkill and backlight states
12883 * The FsckPassNo= compatibility option in mount/service units
12884 has been removed. The fstab generator will now add the
12885 necessary dependencies automatically, and does not require
12886 PID1's support for that anymore.
12888 * journalctl gained a new switch, --list-boots, that lists
12889 recent boots with their times and boot IDs.
12891 * The various tools like systemctl, loginctl, timedatectl,
12892 busctl, systemd-run, … have gained a new switch "-M" to
12893 connect to a specific, local OS container (as direct
12894 connection, without requiring SSH). This works on any
12895 container that is registered with machined, such as those
12896 created by libvirt-lxc or nspawn.
12898 * systemd-run and systemd-analyze also gained support for "-H"
12899 to connect to remote hosts via SSH. This is particularly
12900 useful for systemd-run because it enables queuing of jobs
12901 onto remote systems.
12903 * machinectl gained a new command "login" to open a getty
12904 login in any local container. This works with any container
12905 that is registered with machined (such as those created by
12906 libvirt-lxc or nspawn), and which runs systemd inside.
12908 * machinectl gained a new "reboot" command that may be used to
12909 trigger a reboot on a specific container that is registered
12910 with machined. This works on any container that runs an init
12911 system of some kind.
12913 * systemctl gained a new "list-timers" command to print a nice
12914 listing of installed timer units with the times they elapse
12917 * Alternative reboot() parameters may now be specified on the
12918 "systemctl reboot" command line and are passed to the
12919 reboot() system call.
12921 * systemctl gained a new --job-mode= switch to configure the
12922 mode to queue a job with. This is a more generic version of
12923 --fail, --irreversible, and --ignore-dependencies, which are
12924 still available but not advertised anymore.
12926 * /etc/systemd/system.conf gained new settings to configure
12927 various default timeouts of units, as well as the default
12928 start limit interval and burst. These may still be overridden
12931 * PID1 will now export on the bus profile data of the security
12932 policy upload process (such as the SELinux policy upload to
12935 * journald: when forwarding logs to the console, include
12936 timestamps (following the setting in
12937 /sys/module/printk/parameters/time).
12939 * OnCalendar= in timer units now understands the special
12940 strings "yearly" and "annually". (Both are equivalent)
12942 * The accuracy of timer units is now configurable with the new
12943 AccuracySec= setting. It defaults to 1min.
12945 * A new dependency type JoinsNamespaceOf= has been added that
12946 allows running two services within the same /tmp and network
12947 namespace, if PrivateNetwork= or PrivateTmp= are used.
12949 * A new command "cat" has been added to systemctl. It outputs
12950 the original unit file of a unit, and concatenates the
12951 contents of additional "drop-in" unit file snippets, so that
12952 the full configuration is shown.
12954 * systemctl now supports globbing on the various "list-xyz"
12955 commands, like "list-units" or "list-sockets", as well as on
12956 those commands which take multiple unit names.
12958 * journalctl's --unit= switch gained support for globbing.
12960 * All systemd daemons now make use of the watchdog logic so
12961 that systemd automatically notices when they hang.
12963 * If the $container_ttys environment variable is set,
12964 getty-generator will automatically spawn a getty for each
12965 listed tty. This is useful for container managers to request
12966 login gettys to be spawned on as many ttys as needed.
12968 * %h, %s, %U specifier support is not available anymore when
12969 used in unit files for PID 1. This is because NSS calls are
12970 not safe from PID 1. They stay available for --user
12971 instances of systemd, and as special case for the root user.
12973 * loginctl gained a new "--no-legend" switch to turn off output
12974 of the legend text.
12976 * The "sd-login.h" API gained three new calls:
12977 sd_session_is_remote(), sd_session_get_remote_user(),
12978 sd_session_get_remote_host() to query information about
12981 * The udev hardware database now also carries vendor/product
12982 information of SDIO devices.
12984 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
12985 determine whether watchdog notifications are requested by
12986 the system manager.
12988 * Socket-activated per-connection services now include a
12989 short description of the connection parameters in the
12992 * tmpfiles gained a new "--boot" option. When this is not used,
12993 only lines where the command character is not suffixed with
12994 "!" are executed. When this option is specified, those
12995 options are executed too. This partitions tmpfiles
12996 directives into those that can be safely executed at any
12997 time, and those which should be run only at boot (for
12998 example, a line that creates /run/nologin).
13000 * A new API "sd-resolve.h" has been added which provides a simple
13001 asynchronous wrapper around glibc NSS hostname resolution
13002 calls, such as getaddrinfo(). In contrast to glibc's
13003 getaddrinfo_a(), it does not use signals. In contrast to most
13004 other asynchronous name resolution libraries, this one does
13005 not reimplement DNS, but reuses NSS, so that alternate
13006 hostname resolution systems continue to work, such as mDNS,
13007 LDAP, etc. This API is based on libasyncns, but it has been
13008 cleaned up for inclusion in systemd.
13010 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
13011 "sd-daemon.h" are no longer found in individual libraries
13012 libsystemd-journal.so, libsystemd-login.so,
13013 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
13014 merged them into a single library, libsystemd.so, which
13015 provides all symbols. The reason for this is cyclic
13016 dependencies, as these libraries tend to use each other's
13017 symbols. So far, we have managed to workaround that by linking
13018 a copy of a good part of our code into each of these
13019 libraries again and again, which, however, makes certain
13020 things hard to do, like sharing static variables. Also, it
13021 substantially increases footprint. With this change, there
13022 is only one library for the basic APIs systemd
13023 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
13024 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
13025 library as well, however are subject to the --enable-kdbus
13026 switch (see below). Note that "sd-dhcp-client.h" is not part
13027 of this library (this is because it only consumes, never
13028 provides, services of/to other APIs). To make the transition
13029 easy from the separate libraries to the unified one, we
13030 provide the --enable-compat-libs compile-time switch which
13031 will generate stub libraries that are compatible with the
13032 old ones but redirect all calls to the new one.
13034 * All of the kdbus logic and the new APIs "sd-bus.h",
13035 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
13036 and "sd-utf8.h" are compile-time optional via the
13037 "--enable-kdbus" switch, and they are not compiled in by
13038 default. To make use of kdbus, you have to explicitly enable
13039 the switch. Note however, that neither the kernel nor the
13040 userspace API for all of this is considered stable yet. We
13041 want to maintain the freedom to still change the APIs for
13042 now. By specifying this build-time switch, you acknowledge
13043 that you are aware of the instability of the current
13046 * Also, note that while kdbus is pretty much complete,
13047 it lacks one thing: proper policy support. This means you
13048 can build a fully working system with all features; however,
13049 it will be highly insecure. Policy support will be added in
13050 one of the next releases, at the same time that we will
13051 declare the APIs stable.
13053 * When the kernel command line argument "kdbus" is specified,
13054 systemd will automatically load the kdbus.ko kernel module. At
13055 this stage of development, it is only useful for testing kdbus
13056 and should not be used in production. Note: if "--enable-kdbus"
13057 is specified, and the kdbus.ko kernel module is available, and
13058 "kdbus" is added to the kernel command line, the entire system
13059 runs with kdbus instead of dbus-daemon, with the above mentioned
13060 problem of missing the system policy enforcement. Also a future
13061 version of kdbus.ko or a newer systemd will not be compatible with
13062 each other, and will unlikely be able to boot the machine if only
13063 one of them is updated.
13065 * systemctl gained a new "import-environment" command which
13066 uploads the caller's environment (or parts thereof) into the
13067 service manager so that it is inherited by services started
13068 by the manager. This is useful to upload variables like
13069 $DISPLAY into the user service manager.
13071 * A new PrivateDevices= switch has been added to service units
13072 which allows running a service with a namespaced /dev
13073 directory that does not contain any device nodes for
13074 physical devices. More specifically, it only includes devices
13075 such as /dev/null, /dev/urandom, and /dev/zero which are API
13078 * logind has been extended to support behaviour like VT
13079 switching on seats that do not support a VT. This makes
13080 multi-session available on seats that are not the first seat
13081 (seat0), and on systems where kernel support for VTs has
13082 been disabled at compile-time.
13084 * If a process holds a delay lock for system sleep or shutdown
13085 and fails to release it in time, we will now log its
13086 identity. This makes it easier to identify processes that
13087 cause slow suspends or power-offs.
13089 * When parsing /etc/crypttab, support for a new key-slot=
13090 option as supported by Debian is added. It allows indicating
13091 which LUKS slot to use on disk, speeding up key loading.
13093 * The sd_journal_sendv() API call has been checked and
13094 officially declared to be async-signal-safe so that it may
13095 be invoked from signal handlers for logging purposes.
13097 * Boot-time status output is now enabled automatically after a
13098 short timeout if boot does not progress, in order to give
13099 the user an indication what she or he is waiting for.
13101 * The boot-time output has been improved to show how much time
13102 remains until jobs expire.
13104 * The KillMode= switch in service units gained a new possible
13105 value "mixed". If set, and the unit is shut down, then the
13106 initial SIGTERM signal is sent only to the main daemon
13107 process, while the following SIGKILL signal is sent to
13108 all remaining processes of the service.
13110 * When a scope unit is registered, a new property "Controller"
13111 may be set. If set to a valid bus name, systemd will send a
13112 RequestStop() signal to this name when it would like to shut
13113 down the scope. This may be used to hook manager logic into
13114 the shutdown logic of scope units. Also, scope units may now
13115 be put in a special "abandoned" state, in which case the
13116 manager process which created them takes no further
13117 responsibilities for it.
13119 * When reading unit files, systemd will now verify
13120 the access mode of these files, and warn about certain
13121 suspicious combinations. This has been added to make it
13122 easier to track down packaging bugs where unit files are
13123 marked executable or world-writable.
13125 * systemd-nspawn gained a new "--setenv=" switch to set
13126 container-wide environment variables. The similar option in
13127 systemd-activate was renamed from "--environment=" to
13128 "--setenv=" for consistency.
13130 * systemd-nspawn has been updated to create a new kdbus domain
13131 for each container that is invoked, thus allowing each
13132 container to have its own set of system and user buses,
13133 independent of the host.
13135 * systemd-nspawn gained a new --drop-capability= switch to run
13136 the container with less capabilities than the default. Both
13137 --drop-capability= and --capability= now take the special
13138 string "all" for dropping or keeping all capabilities.
13140 * systemd-nspawn gained new switches for executing containers
13141 with specific SELinux labels set.
13143 * systemd-nspawn gained a new --quiet switch to not generate
13144 any additional output but the container's own console
13147 * systemd-nspawn gained a new --share-system switch to run a
13148 container without PID namespacing enabled.
13150 * systemd-nspawn gained a new --register= switch to control
13151 whether the container is registered with systemd-machined or
13152 not. This is useful for containers that do not run full
13153 OS images, but only specific apps.
13155 * systemd-nspawn gained a new --keep-unit which may be used
13156 when invoked as the only program from a service unit, and
13157 results in registration of the unit service itself in
13158 systemd-machined, instead of a newly opened scope unit.
13160 * systemd-nspawn gained a new --network-interface= switch for
13161 moving arbitrary interfaces to the container. The new
13162 --network-veth switch creates a virtual Ethernet connection
13163 between host and container. The new --network-bridge=
13164 switch then allows assigning the host side of this virtual
13165 Ethernet connection to a bridge device.
13167 * systemd-nspawn gained a new --personality= switch for
13168 setting the kernel personality for the container. This is
13169 useful when running a 32-bit container on a 64-bit host. A
13170 similar option Personality= is now also available for service
13173 * logind will now also track a "Desktop" identifier for each
13174 session which encodes the desktop environment of it. This is
13175 useful for desktop environments that want to identify
13176 multiple running sessions of itself easily.
13178 * A new SELinuxContext= setting for service units has been
13179 added that allows setting a specific SELinux execution
13180 context for a service.
13182 * Most systemd client tools will now honour $SYSTEMD_LESS for
13183 settings of the "less" pager. By default, these tools will
13184 override $LESS to allow certain operations to work, such as
13185 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
13186 influence this logic.
13188 * systemd's "seccomp" hook-up has been changed to make use of
13189 the libseccomp library instead of using its own
13190 implementation. This has benefits for portability among
13193 * For usage together with SystemCallFilter=, a new
13194 SystemCallErrorNumber= setting has been introduced that
13195 allows configuration of a system error number to be returned
13196 on filtered system calls, instead of immediately killing the
13197 process. Also, SystemCallArchitectures= has been added to
13198 limit access to system calls of a particular architecture
13199 (in order to turn off support for unused secondary
13200 architectures). There is also a global
13201 SystemCallArchitectures= setting in system.conf now to turn
13202 off support for non-native system calls system-wide.
13204 * systemd requires a kernel with a working name_to_handle_at(),
13205 please see the kernel config requirements in the README file.
13207 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
13208 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
13209 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
13210 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
13211 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
13212 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
13213 Elia Pinto, Florian Weimer, George McCollister, Goffredo
13214 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
13215 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
13216 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
13217 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
13218 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
13219 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
13220 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
13221 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
13222 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
13223 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
13224 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
13225 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
13226 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
13227 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
13228 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
13229 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
13230 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
13232 — Berlin, 2014-02-20
13236 * logind has gained support for facilitating privileged input
13237 and drm device access for unprivileged clients. This work is
13238 useful to allow Wayland display servers (and similar
13239 programs, such as kmscon) to run under the user's ID and
13240 access input and drm devices which are normally
13241 protected. When this is used (and the kernel is new enough)
13242 logind will "mute" IO on the file descriptors passed to
13243 Wayland as long as it is in the background and "unmute" it
13244 if it returns into the foreground. This allows secure
13245 session switching without allowing background sessions to
13246 eavesdrop on input and display data. This also introduces
13247 session switching support if VT support is turned off in the
13248 kernel, and on seats that are not seat0.
13250 * A new kernel command line option luks.options= is understood
13251 now which allows specifying LUKS options for usage for LUKS
13252 encrypted partitions specified with luks.uuid=.
13254 * tmpfiles.d(5) snippets may now use specifier expansion in
13255 path names. More specifically %m, %b, %H, %v, are now
13256 replaced by the local machine id, boot id, hostname, and
13257 kernel version number.
13259 * A new tmpfiles.d(5) command "m" has been introduced which
13260 may be used to change the owner/group/access mode of a file
13261 or directory if it exists, but do nothing if it does not.
13263 * This release removes high-level support for the
13264 MemorySoftLimit= cgroup setting. The underlying kernel
13265 cgroup attribute memory.soft_limit= is currently badly
13266 designed and likely to be removed from the kernel API in its
13267 current form, hence we should not expose it for now.
13269 * The memory.use_hierarchy cgroup attribute is now enabled for
13270 all cgroups systemd creates in the memory cgroup
13271 hierarchy. This option is likely to be come the built-in
13272 default in the kernel anyway, and the non-hierarchical mode
13273 never made much sense in the intrinsically hierarchical
13276 * A new field _SYSTEMD_SLICE= is logged along with all journal
13277 messages containing the slice a message was generated
13278 from. This is useful to allow easy per-customer filtering of
13279 logs among other things.
13281 * systemd-journald will no longer adjust the group of journal
13282 files it creates to the "systemd-journal" group. Instead we
13283 rely on the journal directory to be owned by the
13284 "systemd-journal" group, and its setgid bit set, so that the
13285 kernel file system layer will automatically enforce that
13286 journal files inherit this group assignment. The reason for
13287 this change is that we cannot allow NSS look-ups from
13288 journald which would be necessary to resolve
13289 "systemd-journal" to a numeric GID, because this might
13290 create deadlocks if NSS involves synchronous queries to
13291 other daemons (such as nscd, or sssd) which in turn are
13292 logging clients of journald and might block on it, which
13293 would then dead lock. A tmpfiles.d(5) snippet included in
13294 systemd will make sure the setgid bit and group are
13295 properly set on the journal directory if it exists on every
13296 boot. However, we recommend adjusting it manually after
13297 upgrades too (or from RPM scriptlets), so that the change is
13298 not delayed until next reboot.
13300 * Backlight and random seed files in /var/lib/ have moved into
13301 the /var/lib/systemd/ directory, in order to centralize all
13302 systemd generated files in one directory.
13304 * Boot time performance measurements (as displayed by
13305 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
13306 performance information if that's available to determine how
13307 much time BIOS and boot loader initialization required. With
13308 a sufficiently new BIOS you hence no longer need to boot
13309 with Gummiboot to get access to such information.
13311 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
13312 Cristian Rodríguez, Dave Reisner, David Herrmann, David
13313 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
13314 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
13315 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
13316 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
13317 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
13319 — Berlin, 2013-10-02
13323 * The Restart= option for services now understands a new
13324 on-watchdog setting, which will restart the service
13325 automatically if the service stops sending out watchdog keep
13326 alive messages (as configured with WatchdogSec=).
13328 * The getty generator (which is responsible for bringing up a
13329 getty on configured serial consoles) will no longer only
13330 start a getty on the primary kernel console but on all
13331 others, too. This makes the order in which console= is
13332 specified on the kernel command line less important.
13334 * libsystemd-logind gained a new sd_session_get_vt() call to
13335 retrieve the VT number of a session.
13337 * If the option "tries=0" is set for an entry of /etc/crypttab
13338 its passphrase is queried indefinitely instead of any
13339 maximum number of tries.
13341 * If a service with a configure PID file terminates its PID
13342 file will now be removed automatically if it still exists
13343 afterwards. This should put an end to stale PID files.
13345 * systemd-run will now also take relative binary path names
13346 for execution and no longer insists on absolute paths.
13348 * InaccessibleDirectories= and ReadOnlyDirectories= now take
13349 paths that are optionally prefixed with "-" to indicate that
13350 it should not be considered a failure if they do not exist.
13352 * journalctl -o (and similar commands) now understands a new
13353 output mode "short-precise", it is similar to "short" but
13354 shows timestamps with usec accuracy.
13356 * The option "discard" (as known from Debian) is now
13357 synonymous to "allow-discards" in /etc/crypttab. In fact,
13358 "discard" is preferred now (since it is easier to remember
13361 * Some licensing clean-ups were made, so that more code is now
13362 LGPL-2.1 licensed than before.
13364 * A minimal tool to save/restore the display backlight
13365 brightness across reboots has been added. It will store the
13366 backlight setting as late as possible at shutdown, and
13367 restore it as early as possible during reboot.
13369 * A logic to automatically discover and enable home and swap
13370 partitions on GPT disks has been added. With this in place
13371 /etc/fstab becomes optional for many setups as systemd can
13372 discover certain partitions located on the root disk
13373 automatically. Home partitions are recognized under their
13374 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
13375 partitions are recognized under their GPT type ID
13376 0657fd6da4ab43c484e50933c84b4f4f.
13378 * systemd will no longer pass any environment from the kernel
13379 or initrd to system services. If you want to set an
13380 environment for all services, do so via the kernel command
13381 line systemd.setenv= assignment.
13383 * The systemd-sysctl tool no longer natively reads the file
13384 /etc/sysctl.conf. If desired, the file should be symlinked
13385 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
13386 legacy support by a symlink rather than built-in code, it
13387 also makes the otherwise hidden order of application of the
13388 different files visible. (Note that this partly reverts to a
13389 pre-198 application order of sysctl knobs!)
13391 * The "systemctl set-log-level" and "systemctl dump" commands
13392 have been moved to systemd-analyze.
13394 * systemd-run learned the new --remain-after-exit switch,
13395 which causes the scope unit not to be cleaned up
13396 automatically after the process terminated.
13398 * tmpfiles learned a new --exclude-prefix= switch to exclude
13399 certain paths from operation.
13401 * journald will now automatically flush all messages to disk
13402 as soon as a message at the log level CRIT, ALERT or EMERG
13405 Contributions from: Andrew Cook, Brandon Philips, Christian
13406 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
13407 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
13408 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
13409 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
13410 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
13411 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
13412 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
13413 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
13414 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
13415 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
13416 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
13417 William Giokas, Zbigniew Jędrzejewski-Szmek
13419 — Berlin, 2013-09-13
13423 * The documentation has been updated to cover the various new
13424 concepts introduced with 205.
13426 * Unit files now understand the new %v specifier which
13427 resolves to the kernel version string as returned by "uname
13430 * systemctl now supports filtering the unit list output by
13431 load state, active state and sub state, using the new
13432 --state= parameter.
13434 * "systemctl status" will now show the results of the
13435 condition checks (like ConditionPathExists= and similar) of
13436 the last start attempts of the unit. They are also logged to
13439 * "journalctl -b" may now be used to look for boot output of a
13440 specific boot. Try "journalctl -b -1" for the previous boot,
13441 but the syntax is substantially more powerful.
13443 * "journalctl --show-cursor" has been added which prints the
13444 cursor string the last shown log line. This may then be used
13445 with the new "journalctl --after-cursor=" switch to continue
13446 browsing logs from that point on.
13448 * "journalctl --force" may now be used to force regeneration
13451 * Creation of "dead" device nodes has been moved from udev
13452 into kmod and tmpfiles. Previously, udev would read the kmod
13453 databases to pre-generate dead device nodes based on meta
13454 information contained in kernel modules, so that these would
13455 be auto-loaded on access rather then at boot. As this
13456 does not really have much to do with the exposing actual
13457 kernel devices to userspace this has always been slightly
13458 alien in the udev codebase. Following the new scheme kmod
13459 will now generate a runtime snippet for tmpfiles from the
13460 module meta information and it now is tmpfiles' job to the
13461 create the nodes. This also allows overriding access and
13462 other parameters for the nodes using the usual tmpfiles
13463 facilities. As side effect this allows us to remove the
13464 CAP_SYS_MKNOD capability bit from udevd entirely.
13466 * logind's device ACLs may now be applied to these "dead"
13467 devices nodes too, thus finally allowing managed access to
13468 devices such as /dev/snd/sequencer without loading the
13469 backing module right-away.
13471 * A new RPM macro has been added that may be used to apply
13472 tmpfiles configuration during package installation.
13474 * systemd-detect-virt and ConditionVirtualization= now can
13475 detect User-Mode-Linux machines (UML).
13477 * journald will now implicitly log the effective capabilities
13478 set of processes in the message metadata.
13480 * systemd-cryptsetup has gained support for TrueCrypt volumes.
13482 * The initrd interface has been simplified (more specifically,
13483 support for passing performance data via environment
13484 variables and fsck results via files in /run has been
13485 removed). These features were non-essential, and are
13486 nowadays available in a much nicer way by having systemd in
13487 the initrd serialize its state and have the hosts systemd
13488 deserialize it again.
13490 * The udev "keymap" data files and tools to apply keyboard
13491 specific mappings of scan to key codes, and force-release
13492 scan code lists have been entirely replaced by a udev
13493 "keyboard" builtin and a hwdb data file.
13495 * systemd will now honour the kernel's "quiet" command line
13496 argument also during late shutdown, resulting in a
13497 completely silent shutdown when used.
13499 * There's now an option to control the SO_REUSEPORT socket
13500 option in .socket units.
13502 * Instance units will now automatically get a per-template
13503 subslice of system.slice unless something else is explicitly
13504 configured. For example, instances of sshd@.service will now
13505 implicitly be placed in system-sshd.slice rather than
13506 system.slice as before.
13508 * Test coverage support may now be enabled at build time.
13510 Contributions from: Dave Reisner, Frederic Crozat, Harald
13511 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
13512 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
13513 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
13514 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
13515 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
13516 Giokas, Zbigniew Jędrzejewski-Szmek
13518 — Berlin, 2013-07-23
13522 * Two new unit types have been introduced:
13524 Scope units are very similar to service units, however, are
13525 created out of pre-existing processes — instead of PID 1
13526 forking off the processes. By using scope units it is
13527 possible for system services and applications to group their
13528 own child processes (worker processes) in a powerful way
13529 which then maybe used to organize them, or kill them
13530 together, or apply resource limits on them.
13532 Slice units may be used to partition system resources in an
13533 hierarchical fashion and then assign other units to them. By
13534 default there are now three slices: system.slice (for all
13535 system services), user.slice (for all user sessions),
13536 machine.slice (for VMs and containers).
13538 Slices and scopes have been introduced primarily in
13539 context of the work to move cgroup handling to a
13540 single-writer scheme, where only PID 1
13541 creates/removes/manages cgroups.
13543 * There's a new concept of "transient" units. In contrast to
13544 normal units these units are created via an API at runtime,
13545 not from configuration from disk. More specifically this
13546 means it is now possible to run arbitrary programs as
13547 independent services, with all execution parameters passed
13548 in via bus APIs rather than read from disk. Transient units
13549 make systemd substantially more dynamic then it ever was,
13550 and useful as a general batch manager.
13552 * logind has been updated to make use of scope and slice units
13553 for managing user sessions. As a user logs in he will get
13554 his own private slice unit, to which all sessions are added
13555 as scope units. We also added support for automatically
13556 adding an instance of user@.service for the user into the
13557 slice. Effectively logind will no longer create cgroup
13558 hierarchies on its own now, it will defer entirely to PID 1
13559 for this by means of scope, service and slice units. Since
13560 user sessions this way become entities managed by PID 1
13561 the output of "systemctl" is now a lot more comprehensive.
13563 * A new mini-daemon "systemd-machined" has been added which
13564 may be used by virtualization managers to register local
13565 VMs/containers. nspawn has been updated accordingly, and
13566 libvirt will be updated shortly. machined will collect a bit
13567 of meta information about the VMs/containers, and assign
13568 them their own scope unit (see above). The collected
13569 meta-data is then made available via the "machinectl" tool,
13570 and exposed in "ps" and similar tools. machined/machinectl
13571 is compile-time optional.
13573 * As discussed earlier, the low-level cgroup configuration
13574 options ControlGroup=, ControlGroupModify=,
13575 ControlGroupPersistent=, ControlGroupAttribute= have been
13576 removed. Please use high-level attribute settings instead as
13577 well as slice units.
13579 * A new bus call SetUnitProperties() has been added to alter
13580 various runtime parameters of a unit. This is primarily
13581 useful to alter cgroup parameters dynamically in a nice way,
13582 but will be extended later on to make more properties
13583 modifiable at runtime. systemctl gained a new set-properties
13584 command that wraps this call.
13586 * A new tool "systemd-run" has been added which can be used to
13587 run arbitrary command lines as transient services or scopes,
13588 while configuring a number of settings via the command
13589 line. This tool is currently very basic, however already
13590 very useful. We plan to extend this tool to even allow
13591 queuing of execution jobs with time triggers from the
13592 command line, similar in fashion to "at".
13594 * nspawn will now inform the user explicitly that kernels with
13595 audit enabled break containers, and suggest the user to turn
13598 * Support for detecting the IMA and AppArmor security
13599 frameworks with ConditionSecurity= has been added.
13601 * journalctl gained a new "-k" switch for showing only kernel
13602 messages, mimicking dmesg output; in addition to "--user"
13603 and "--system" switches for showing only user's own logs
13606 * systemd-delta can now show information about drop-in
13607 snippets extending unit files.
13609 * libsystemd-bus has been substantially updated but is still
13610 not available as public API.
13612 * systemd will now look for the "debug" argument on the kernel
13613 command line and enable debug logging, similar to what
13614 "systemd.log_level=debug" already did before.
13616 * "systemctl set-default", "systemctl get-default" has been
13617 added to configure the default.target symlink, which
13618 controls what to boot into by default.
13620 * "systemctl set-log-level" has been added as a convenient
13621 way to raise and lower systemd logging threshold.
13623 * "systemd-analyze plot" will now show the time the various
13624 generators needed for execution, as well as information
13625 about the unit file loading.
13627 * libsystemd-journal gained a new sd_journal_open_files() call
13628 for opening specific journal files. journactl also gained a
13629 new switch to expose this new functionality. Previously we
13630 only supported opening all files from a directory, or all
13631 files from the system, as opening individual files only is
13632 racy due to journal file rotation.
13634 * systemd gained the new DefaultEnvironment= setting in
13635 /etc/systemd/system.conf to set environment variables for
13638 * If a privileged process logs a journal message with the
13639 OBJECT_PID= field set, then journald will automatically
13640 augment this with additional OBJECT_UID=, OBJECT_GID=,
13641 OBJECT_COMM=, OBJECT_EXE=, … fields. This is useful if
13642 system services want to log events about specific client
13643 processes. journactl/systemctl has been updated to make use
13644 of this information if all log messages regarding a specific
13647 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
13648 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
13649 Reisner, David Coppa, David King, David Strauss, Eelco
13650 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
13651 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
13652 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
13653 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
13654 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
13655 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
13656 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
13657 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
13658 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
13659 Łukasz Stelmach, 장동준
13663 * The Python bindings gained some minimal support for the APIs
13664 exposed by libsystemd-logind.
13666 * ConditionSecurity= gained support for detecting SMACK. Since
13667 this condition already supports SELinux and AppArmor we only
13668 miss IMA for this. Patches welcome!
13670 Contributions from: Karol Lewandowski, Lennart Poettering,
13671 Zbigniew Jędrzejewski-Szmek
13675 * systemd-nspawn will now create /etc/resolv.conf if
13676 necessary, before bind-mounting the host's file onto it.
13678 * systemd-nspawn will now store meta information about a
13679 container on the container's cgroup as extended attribute
13680 fields, including the root directory.
13682 * The cgroup hierarchy has been reworked in many ways. All
13683 objects any of the components systemd creates in the cgroup
13684 tree are now suffixed. More specifically, user sessions are
13685 now placed in cgroups suffixed with ".session", users in
13686 cgroups suffixed with ".user", and nspawn containers in
13687 cgroups suffixed with ".nspawn". Furthermore, all cgroup
13688 names are now escaped in a simple scheme to avoid collision
13689 of userspace object names with kernel filenames. This work
13690 is preparation for making these objects relocatable in the
13691 cgroup tree, in order to allow easy resource partitioning of
13692 these objects without causing naming conflicts.
13694 * systemctl list-dependencies gained the new switches
13695 --plain, --reverse, --after and --before.
13697 * systemd-inhibit now shows the process name of processes that
13698 have taken an inhibitor lock.
13700 * nss-myhostname will now also resolve "localhost"
13701 implicitly. This makes /etc/hosts an optional file and
13702 nicely handles that on IPv6 ::1 maps to both "localhost" and
13703 the local hostname.
13705 * libsystemd-logind.so gained a new call
13706 sd_get_machine_names() to enumerate running containers and
13707 VMs (currently only supported by very new libvirt and
13708 nspawn). sd_login_monitor can now be used to watch
13709 VMs/containers coming and going.
13711 * .include is not allowed recursively anymore, and only in
13712 unit files. Usually it is better to use drop-in snippets in
13713 .d/*.conf anyway, as introduced with systemd 198.
13715 * systemd-analyze gained a new "critical-chain" command that
13716 determines the slowest chain of units run during system
13717 boot-up. It is very useful for tracking down where
13718 optimizing boot time is the most beneficial.
13720 * systemd will no longer allow manipulating service paths in
13721 the name=systemd:/system cgroup tree using ControlGroup= in
13722 units. (But is still fine with it in all other dirs.)
13724 * There's a new systemd-nspawn@.service service file that may
13725 be used to easily run nspawn containers as system
13726 services. With the container's root directory in
13727 /var/lib/container/foobar it is now sufficient to run
13728 "systemctl start systemd-nspawn@foobar.service" to boot it.
13730 * systemd-cgls gained a new parameter "--machine" to list only
13731 the processes within a certain container.
13733 * ConditionSecurity= now can check for "apparmor". We still
13734 are lacking checks for SMACK and IMA for this condition
13735 check though. Patches welcome!
13737 * A new configuration file /etc/systemd/sleep.conf has been
13738 added that may be used to configure which kernel operation
13739 systemd is supposed to execute when "suspend", "hibernate"
13740 or "hybrid-sleep" is requested. This makes the new kernel
13741 "freeze" state accessible to the user.
13743 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
13744 the passed argument if applicable.
13746 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
13747 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
13748 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
13749 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
13750 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
13751 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
13752 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
13757 * The output of 'systemctl list-jobs' got some polishing. The
13758 '--type=' argument may now be passed more than once. A new
13759 command 'systemctl list-sockets' has been added which shows
13760 a list of kernel sockets systemd is listening on with the
13761 socket units they belong to, plus the units these socket
13764 * The experimental libsystemd-bus library got substantial
13765 updates to work in conjunction with the (also experimental)
13766 kdbus kernel project. It works well enough to exchange
13767 messages with some sophistication. Note that kdbus is not
13768 ready yet, and the library is mostly an elaborate test case
13769 for now, and not installable.
13771 * systemd gained a new unit 'systemd-static-nodes.service'
13772 that generates static device nodes earlier during boot, and
13773 can run in conjunction with udev.
13775 * libsystemd-login gained a new call sd_pid_get_user_unit()
13776 to retrieve the user systemd unit a process is running
13777 in. This is useful for systems where systemd is used as
13780 * systemd-nspawn now places all containers in the new /machine
13781 top-level cgroup directory in the name=systemd
13782 hierarchy. libvirt will soon do the same, so that we get a
13783 uniform separation of /system, /user and /machine for system
13784 services, user processes and containers/virtual
13785 machines. This new cgroup hierarchy is also useful to stick
13786 stable names to specific container instances, which can be
13787 recognized later this way (this name may be controlled
13788 via systemd-nspawn's new -M switch). libsystemd-login also
13789 gained a new call sd_pid_get_machine_name() to retrieve the
13790 name of the container/VM a specific process belongs to.
13792 * bootchart can now store its data in the journal.
13794 * libsystemd-journal gained a new call
13795 sd_journal_add_conjunction() for AND expressions to the
13796 matching logic. This can be used to express more complex
13797 logical expressions.
13799 * journactl can now take multiple --unit= and --user-unit=
13802 * The cryptsetup logic now understands the "luks.key=" kernel
13803 command line switch for specifying a file to read the
13804 decryption key from. Also, if a configured key file is not
13805 found the tool will now automatically fall back to prompting
13808 * Python systemd.journal module was updated to wrap recently
13809 added functions from libsystemd-journal. The interface was
13810 changed to bring the low level interface in s.j._Reader
13811 closer to the C API, and the high level interface in
13812 s.j.Reader was updated to wrap and convert all data about
13815 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
13816 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
13817 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
13818 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
13819 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
13820 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
13824 * journalctl --update-catalog now understands a new --root=
13825 option to operate on catalogs found in a different root
13828 * During shutdown after systemd has terminated all running
13829 services a final killing loop kills all remaining left-over
13830 processes. We will now print the name of these processes
13831 when we send SIGKILL to them, since this usually indicates a
13834 * If /etc/crypttab refers to password files stored on
13835 configured mount points automatic dependencies will now be
13836 generated to ensure the specific mount is established first
13837 before the key file is attempted to be read.
13839 * 'systemctl status' will now show information about the
13840 network sockets a socket unit is listening on.
13842 * 'systemctl status' will also shown information about any
13843 drop-in configuration file for units. (Drop-In configuration
13844 files in this context are files such as
13845 /etc/systemd/system/foobar.service.d/*.conf)
13847 * systemd-cgtop now optionally shows summed up CPU times of
13848 cgroups. Press '%' while running cgtop to switch between
13849 percentage and absolute mode. This is useful to determine
13850 which cgroups use up the most CPU time over the entire
13851 runtime of the system. systemd-cgtop has also been updated
13852 to be 'pipeable' for processing with further shell tools.
13854 * 'hostnamectl set-hostname' will now allow setting of FQDN
13857 * The formatting and parsing of time span values has been
13858 changed. The parser now understands fractional expressions
13859 such as "5.5h". The formatter will now output fractional
13860 expressions for all time spans under 1min, i.e. "5.123456s"
13861 rather than "5s 123ms 456us". For time spans under 1s
13862 millisecond values are shown, for those under 1ms
13863 microsecond values are shown. This should greatly improve
13864 all time-related output of systemd.
13866 * libsystemd-login and libsystemd-journal gained new
13867 functions for querying the poll() events mask and poll()
13868 timeout value for integration into arbitrary event
13871 * localectl gained the ability to list available X11 keymaps
13872 (models, layouts, variants, options).
13874 * 'systemd-analyze dot' gained the ability to filter for
13875 specific units via shell-style globs, to create smaller,
13876 more useful graphs. I.e. it is now possible to create simple
13877 graphs of all the dependencies between only target units, or
13878 of all units that Avahi has dependencies with.
13880 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
13881 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
13882 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
13883 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
13884 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
13885 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
13886 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
13890 * The boot-time readahead implementation for rotating media
13891 will now read the read-ahead data in multiple passes which
13892 consist of all read requests made in equidistant time
13893 intervals. This means instead of strictly reading read-ahead
13894 data in its physical order on disk we now try to find a
13895 middle ground between physical and access time order.
13897 * /etc/os-release files gained a new BUILD_ID= field for usage
13898 on operating systems that provide continuous builds of OS
13901 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
13902 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
13903 William Douglas, Zbigniew Jędrzejewski-Szmek
13907 * systemd-python gained an API exposing libsystemd-daemon.
13909 * The SMACK setup logic gained support for uploading CIPSO
13912 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
13913 ReadOnlyDirectories= and InaccessibleDirectories= has
13914 changed. The private /tmp and /var/tmp directories are now
13915 shared by all processes of a service (which means
13916 ExecStartPre= may now leave data in /tmp that ExecStart= of
13917 the same service can still access). When a service is
13918 stopped its temporary directories are immediately deleted
13919 (normal clean-up with tmpfiles is still done in addition to
13922 * By default, systemd will now set a couple of sysctl
13923 variables in the kernel: the safe sysrq options are turned
13924 on, IP route verification is turned on, and source routing
13925 disabled. The recently added hardlink and softlink
13926 protection of the kernel is turned on. These settings should
13927 be reasonably safe, and good defaults for all new systems.
13929 * The predictable network naming logic may now be turned off
13930 with a new kernel command line switch: net.ifnames=0.
13932 * A new libsystemd-bus module has been added that implements a
13933 pretty complete D-Bus client library. For details see:
13935 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
13937 * journald will now explicitly flush the journal files to disk
13938 at the latest 5min after each write. The file will then also
13939 be marked offline until the next write. This should increase
13940 reliability in case of a crash. The synchronization delay
13941 can be configured via SyncIntervalSec= in journald.conf.
13943 * There's a new remote-fs-setup.target unit that can be used
13944 to pull in specific services when at least one remote file
13945 system is to be mounted.
13947 * There are new targets timers.target and paths.target as
13948 canonical targets to pull user timer and path units in
13949 from. This complements sockets.target with a similar
13950 purpose for socket units.
13952 * libudev gained a new call udev_device_set_attribute_value()
13953 to set sysfs attributes of a device.
13955 * The udev daemon now sets the default number of worker
13956 processes executed in parallel based on the number of available
13957 CPUs instead of the amount of available RAM. This is supposed
13958 to provide a more reliable default and limit a too aggressive
13959 parallelism for setups with 1000s of devices connected.
13961 Contributions from: Auke Kok, Colin Walters, Cristian
13962 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
13963 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
13964 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
13965 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
13966 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
13967 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
13968 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
13969 Zbigniew Jędrzejewski-Szmek
13973 * Configuration of unit files may now be extended via drop-in
13974 files without having to edit/override the unit files
13975 themselves. More specifically, if the administrator wants to
13976 change one value for a service file foobar.service he can
13977 now do so by dropping in a configuration snippet into
13978 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
13979 will load all these snippets and apply them on top of the
13980 main unit configuration file, possibly extending or
13981 overriding its settings. Using these drop-in snippets is
13982 generally nicer than the two earlier options for changing
13983 unit files locally: copying the files from
13984 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
13985 them there; or creating a new file in /etc/systemd/system/
13986 that incorporates the original one via ".include". Drop-in
13987 snippets into these .d/ directories can be placed in any
13988 directory systemd looks for units in, and the usual
13989 overriding semantics between /usr/lib, /etc and /run apply
13992 * Most unit file settings which take lists of items can now be
13993 reset by assigning the empty string to them. For example,
13994 normally, settings such as Environment=FOO=BAR append a new
13995 environment variable assignment to the environment block,
13996 each time they are used. By assigning Environment= the empty
13997 string the environment block can be reset to empty. This is
13998 particularly useful with the .d/*.conf drop-in snippets
13999 mentioned above, since this adds the ability to reset list
14000 settings from vendor unit files via these drop-ins.
14002 * systemctl gained a new "list-dependencies" command for
14003 listing the dependencies of a unit recursively.
14005 * Inhibitors are now honored and listed by "systemctl
14006 suspend", "systemctl poweroff" (and similar) too, not only
14007 GNOME. These commands will also list active sessions by
14010 * Resource limits (as exposed by the various control group
14011 controllers) can now be controlled dynamically at runtime
14012 for all units. More specifically, you can now use a command
14013 like "systemctl set-cgroup-attr foobar.service cpu.shares
14014 2000" to alter the CPU shares a specific service gets. These
14015 settings are stored persistently on disk, and thus allow the
14016 administrator to easily adjust the resource usage of
14017 services with a few simple commands. This dynamic resource
14018 management logic is also available to other programs via the
14019 bus. Almost any kernel cgroup attribute and controller is
14022 * systemd-vconsole-setup will now copy all font settings to
14023 all allocated VTs, where it previously applied them only to
14026 * libsystemd-login gained the new sd_session_get_tty() API
14029 * This release drops support for a few legacy or
14030 distribution-specific LSB facility names when parsing init
14031 scripts: $x-display-manager, $mail-transfer-agent,
14032 $mail-transport-agent, $mail-transfer-agent, $smtp,
14033 $null. Also, the mail-transfer-agent.target unit backing
14034 this has been removed. Distributions which want to retain
14035 compatibility with this should carry the burden for
14036 supporting this themselves and patch support for these back
14037 in, if they really need to. Also, the facilities $syslog and
14038 $local_fs are now ignored, since systemd does not support
14039 early-boot LSB init scripts anymore, and these facilities
14040 are implied anyway for normal services. syslog.target has
14043 * There are new bus calls on PID1's Manager object for
14044 cancelling jobs, and removing snapshot units. Previously,
14045 both calls were only available on the Job and Snapshot
14046 objects themselves.
14048 * systemd-journal-gatewayd gained SSL support.
14050 * The various "environment" files, such as /etc/locale.conf
14051 now support continuation lines with a backslash ("\") as
14052 last character in the line, similarly in style (but different)
14053 to how this is supported in shells.
14055 * For normal user processes the _SYSTEMD_USER_UNIT= field is
14056 now implicitly appended to every log entry logged. systemctl
14057 has been updated to filter by this field when operating on a
14058 user systemd instance.
14060 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
14061 CAP_AUDIT_CONTROL capabilities to the capabilities set for
14062 the container. This makes it easier to boot unmodified
14063 Fedora systems in a container, which however still requires
14064 audit=0 to be passed on the kernel command line. Auditing in
14065 kernel and userspace is unfortunately still too broken in
14066 context of containers, hence we recommend compiling it out
14067 of the kernel or using audit=0. Hopefully this will be fixed
14068 one day for good in the kernel.
14070 * nspawn gained the new --bind= and --bind-ro= parameters to
14071 bind mount specific directories from the host into the
14074 * nspawn will now mount its own devpts file system instance
14075 into the container, in order not to leak pty devices from
14076 the host into the container.
14078 * systemd will now read the firmware boot time performance
14079 information from the EFI variables, if the used boot loader
14080 supports this, and takes it into account for boot performance
14081 analysis via "systemd-analyze". This is currently supported
14082 only in conjunction with Gummiboot, but could be supported
14083 by other boot loaders too. For details see:
14085 https://systemd.io/BOOT_LOADER_INTERFACE
14087 * A new generator has been added that automatically mounts the
14088 EFI System Partition (ESP) to /boot, if that directory
14089 exists, is empty, and no other file system has been
14090 configured to be mounted there.
14092 * logind will now send out PrepareForSleep(false) out
14093 unconditionally, after coming back from suspend. This may be
14094 used by applications as asynchronous notification for
14095 system resume events.
14097 * "systemctl unlock-sessions" has been added, that allows
14098 unlocking the screens of all user sessions at once, similar
14099 to how "systemctl lock-sessions" already locked all users
14100 sessions. This is backed by a new D-Bus call UnlockSessions().
14102 * "loginctl seat-status" will now show the master device of a
14103 seat. (i.e. the device of a seat that needs to be around for
14104 the seat to be considered available, usually the graphics
14107 * tmpfiles gained a new "X" line type, that allows
14108 configuration of files and directories (with wildcards) that
14109 shall be excluded from automatic cleanup ("aging").
14111 * udev default rules set the device node permissions now only
14112 at "add" events, and do not change them any longer with a
14113 later "change" event.
14115 * The log messages for lid events and power/sleep keypresses
14116 now carry a message ID.
14118 * We now have a substantially larger unit test suite, but this
14119 continues to be work in progress.
14121 * udevadm hwdb gained a new --root= parameter to change the
14122 root directory to operate relative to.
14124 * logind will now issue a background sync() request to the kernel
14125 early at shutdown, so that dirty buffers are flushed to disk early
14126 instead of at the last moment, in order to optimize shutdown
14129 * A new bootctl tool has been added that is an interface for
14130 certain boot loader operations. This is currently a preview
14131 and is likely to be extended into a small mechanism daemon
14132 like timedated, localed, hostnamed, and can be used by
14133 graphical UIs to enumerate available boot options, and
14134 request boot into firmware operations.
14136 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
14137 the rest of the package. It also has been updated to work
14138 correctly in initrds.
14140 * polkit previously has been runtime optional, and is now also
14141 compile time optional via a configure switch.
14143 * systemd-analyze has been reimplemented in C. Also "systemctl
14144 dot" has moved into systemd-analyze.
14146 * "systemctl status" with no further parameters will now print
14147 the status of all active or failed units.
14149 * Operations such as "systemctl start" can now be executed
14150 with a new mode "--irreversible" which may be used to queue
14151 operations that cannot accidentally be reversed by a later
14152 job queuing. This is by default used to make shutdown
14153 requests more robust.
14155 * The Python API of systemd now gained a new module for
14156 reading journal files.
14158 * A new tool kernel-install has been added that can install
14159 kernel images according to the Boot Loader Specification:
14161 https://systemd.io/BOOT_LOADER_SPECIFICATION
14163 * Boot time console output has been improved to provide
14164 animated boot time output for hanging jobs.
14166 * A new tool systemd-activate has been added which can be used
14167 to test socket activation with, directly from the command
14168 line. This should make it much easier to test and debug
14169 socket activation in daemons.
14171 * journalctl gained a new "--reverse" (or -r) option to show
14172 journal output in reverse order (i.e. newest line first).
14174 * journalctl gained a new "--pager-end" (or -e) option to jump
14175 to immediately jump to the end of the journal in the
14176 pager. This is only supported in conjunction with "less".
14178 * journalctl gained a new "--user-unit=" option, that works
14179 similarly to "--unit=" but filters for user units rather than
14182 * A number of unit files to ease adoption of systemd in
14183 initrds has been added. This moves some minimal logic from
14184 the various initrd implementations into systemd proper.
14186 * The journal files are now owned by a new group
14187 "systemd-journal", which exists specifically to allow access
14188 to the journal, and nothing else. Previously, we used the
14189 "adm" group for that, which however possibly covers more
14190 than just journal/log file access. This new group is now
14191 already used by systemd-journal-gatewayd to ensure this
14192 daemon gets access to the journal files and as little else
14193 as possible. Note that "make install" will also set FS ACLs
14194 up for /var/log/journal to give "adm" and "wheel" read
14195 access to it, in addition to "systemd-journal" which owns
14196 the journal files. We recommend that packaging scripts also
14197 add read access to "adm" + "wheel" to /var/log/journal, and
14198 all existing/future journal files. To normal users and
14199 administrators little changes, however packagers need to
14200 ensure to create the "systemd-journal" system group at
14201 package installation time.
14203 * The systemd-journal-gatewayd now runs as unprivileged user
14204 systemd-journal-gateway:systemd-journal-gateway. Packaging
14205 scripts need to create these system user/group at
14208 * timedated now exposes a new boolean property CanNTP that
14209 indicates whether a local NTP service is available or not.
14211 * systemd-detect-virt will now also detect xen PVs
14213 * The pstore file system is now mounted by default, if it is
14216 * In addition to the SELinux and IMA policies we will now also
14217 load SMACK policies at early boot.
14219 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
14220 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
14221 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
14222 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
14223 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
14224 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
14225 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
14226 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
14227 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
14228 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
14229 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
14230 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
14231 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
14232 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
14236 * Timer units now support calendar time events in addition to
14237 monotonic time events. That means you can now trigger a unit
14238 based on a calendar time specification such as "Thu,Fri
14239 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
14240 or fifth day of any month of the year 2013, given that it is
14241 a Thursday or a Friday. This brings timer event support
14242 considerably closer to cron's capabilities. For details on
14243 the supported calendar time specification language see
14246 * udev now supports a number of different naming policies for
14247 network interfaces for predictable names, and a combination
14248 of these policies is now the default. Please see this wiki
14249 document for details:
14251 https://www.freedesktop.org/software/systemd/man/systemd.net-naming-scheme.html
14253 * Auke Kok's bootchart implementation has been added to the
14254 systemd tree. It is an optional component that can graph the
14255 boot in quite some detail. It is one of the best bootchart
14256 implementations around and minimal in its code and
14259 * nss-myhostname has been integrated into the systemd source
14260 tree. nss-myhostname guarantees that the local hostname
14261 always stays resolvable via NSS. It has been a weak
14262 requirement of systemd-hostnamed since a long time, and
14263 since its code is actually trivial we decided to just
14264 include it in systemd's source tree. It can be turned off
14265 with a configure switch.
14267 * The read-ahead logic is now capable of properly detecting
14268 whether a btrfs file system is on SSD or rotating media, in
14269 order to optimize the read-ahead scheme. Previously, it was
14270 only capable of detecting this on traditional file systems
14273 * In udev, additional device properties are now read from the
14274 IAB in addition to the OUI database. Also, Bluetooth company
14275 identities are attached to the devices as well.
14277 * In service files %U may be used as specifier that is
14278 replaced by the configured user name of the service.
14280 * nspawn may now be invoked without a controlling TTY. This
14281 makes it suitable for invocation as its own service. This
14282 may be used to set up a simple containerized server system
14283 using only core OS tools.
14285 * systemd and nspawn can now accept socket file descriptors
14286 when they are started for socket activation. This enables
14287 implementation of socket activated nspawn
14288 containers. i.e. think about autospawning an entire OS image
14289 when the first SSH or HTTP connection is received. We expect
14290 that similar functionality will also be added to libvirt-lxc
14293 * journalctl will now suppress ANSI color codes when
14294 presenting log data.
14296 * systemctl will no longer show control group information for
14297 a unit if the control group is empty anyway.
14299 * logind can now automatically suspend/hibernate/shutdown the
14302 * /etc/machine-info and hostnamed now also expose the chassis
14303 type of the system. This can be used to determine whether
14304 the local system is a laptop, desktop, handset or
14305 tablet. This information may either be configured by the
14306 user/vendor or is automatically determined from ACPI and DMI
14307 information if possible.
14309 * A number of polkit actions are now bound together with "imply"
14310 rules. This should simplify creating UIs because many actions
14311 will now authenticate similar ones as well.
14313 * Unit files learnt a new condition ConditionACPower= which
14314 may be used to conditionalize a unit depending on whether an
14315 AC power source is connected or not, of whether the system
14316 is running on battery power.
14318 * systemctl gained a new "is-failed" verb that may be used in
14319 shell scripts and suchlike to check whether a specific unit
14320 is in the "failed" state.
14322 * The EnvironmentFile= setting in unit files now supports file
14323 globbing, and can hence be used to easily read a number of
14324 environment files at once.
14326 * systemd will no longer detect and recognize specific
14327 distributions. All distribution-specific #ifdeffery has been
14328 removed, systemd is now fully generic and
14329 distribution-agnostic. Effectively, not too much is lost as
14330 a lot of the code is still accessible via explicit configure
14331 switches. However, support for some distribution specific
14332 legacy configuration file formats has been dropped. We
14333 recommend distributions to simply adopt the configuration
14334 files everybody else uses now and convert the old
14335 configuration from packaging scripts. Most distributions
14336 already did that. If that's not possible or desirable,
14337 distributions are welcome to forward port the specific
14338 pieces of code locally from the git history.
14340 * When logging a message about a unit systemd will now always
14341 log the unit name in the message meta data.
14343 * localectl will now also discover system locale data that is
14344 not stored in locale archives, but directly unpacked.
14346 * logind will no longer unconditionally use framebuffer
14347 devices as seat masters, i.e. as devices that are required
14348 to be existing before a seat is considered preset. Instead,
14349 it will now look for all devices that are tagged as
14350 "seat-master" in udev. By default, framebuffer devices will
14351 be marked as such, but depending on local systems, other
14352 devices might be marked as well. This may be used to
14353 integrate graphics cards using closed source drivers (such
14354 as NVidia ones) more nicely into logind. Note however, that
14355 we recommend using the open source NVidia drivers instead,
14356 and no udev rules for the closed-source drivers will be
14357 shipped from us upstream.
14359 Contributions from: Adam Williamson, Alessandro Crismani, Auke
14360 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
14361 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
14362 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
14363 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
14364 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
14365 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
14366 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
14367 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
14368 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
14369 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
14370 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
14375 * udev gained support for loading additional device properties
14376 from an indexed database that is keyed by vendor/product IDs
14377 and similar device identifiers. For the beginning this
14378 "hwdb" is populated with data from the well-known PCI and
14379 USB database, but also includes PNP, ACPI and OID data. In
14380 the longer run this indexed database shall grow into
14381 becoming the one central database for non-essential
14382 userspace device metadata. Previously, data from the PCI/USB
14383 database was only attached to select devices, since the
14384 lookup was a relatively expensive operation due to O(n) time
14385 complexity (with n being the number of entries in the
14386 database). Since this is now O(1), we decided to add in this
14387 data for all devices where this is available, by
14388 default. Note that the indexed database needs to be rebuilt
14389 when new data files are installed. To achieve this you need
14390 to update your packaging scripts to invoke "udevadm hwdb
14391 --update" after installation of hwdb data files. For
14392 RPM-based distributions we introduced the new
14393 %udev_hwdb_update macro for this purpose.
14395 * The Journal gained support for the "Message Catalog", an
14396 indexed database to link up additional information with
14397 journal entries. For further details please check:
14399 https://www.freedesktop.org/wiki/Software/systemd/catalog
14401 The indexed message catalog database also needs to be
14402 rebuilt after installation of message catalog files. Use
14403 "journalctl --update-catalog" for this. For RPM-based
14404 distributions we introduced the %journal_catalog_update
14405 macro for this purpose.
14407 * The Python Journal bindings gained support for the standard
14408 Python logging framework.
14410 * The Journal API gained new functions for checking whether
14411 the underlying file system of a journal file is capable of
14412 properly reporting file change notifications, or whether
14413 applications that want to reflect journal changes "live"
14414 need to recheck journal files continuously in appropriate
14417 * It is now possible to set the "age" field for tmpfiles
14418 entries to 0, indicating that files matching this entry
14419 shall always be removed when the directories are cleaned up.
14421 * coredumpctl gained a new "gdb" verb which invokes gdb
14422 right-away on the selected coredump.
14424 * There's now support for "hybrid sleep" on kernels that
14425 support this, in addition to "suspend" and "hibernate". Use
14426 "systemctl hybrid-sleep" to make use of this.
14428 * logind's HandleSuspendKey= setting (and related settings)
14429 now gained support for a new "lock" setting to simply
14430 request the screen lock on all local sessions, instead of
14431 actually executing a suspend or hibernation.
14433 * systemd will now mount the EFI variables file system by
14436 * Socket units now gained support for configuration of the
14437 SMACK security label.
14439 * timedatectl will now output the time of the last and next
14440 daylight saving change.
14442 * We dropped support for various legacy and distro-specific
14443 concepts, such as insserv, early-boot SysV services
14444 (i.e. those for non-standard runlevels such as 'b' or 'S')
14445 or ArchLinux /etc/rc.conf support. We recommend the
14446 distributions who still need support this to either continue
14447 to maintain the necessary patches downstream, or find a
14448 different solution. (Talk to us if you have questions!)
14450 * Various systemd components will now bypass polkit checks for
14451 root and otherwise handle properly if polkit is not found to
14452 be around. This should fix most issues for polkit-less
14453 systems. Quite frankly this should have been this way since
14454 day one. It is absolutely our intention to make systemd work
14455 fine on polkit-less systems, and we consider it a bug if
14456 something does not work as it should if polkit is not around.
14458 * For embedded systems it is now possible to build udev and
14459 systemd without blkid and/or kmod support.
14461 * "systemctl switch-root" is now capable of switching root
14462 more than once. I.e. in addition to transitions from the
14463 initrd to the host OS it is now possible to transition to
14464 further OS images from the host. This is useful to implement
14465 offline updating tools.
14467 * Various other additions have been made to the RPM macros
14468 shipped with systemd. Use %udev_rules_update() after
14469 installing new udev rules files. %_udevhwdbdir,
14470 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
14471 %_sysctldir are now available which resolve to the right
14472 directories for packages to place various data files in.
14474 * journalctl gained the new --full switch (in addition to
14475 --all, to disable ellipsation for long messages.
14477 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
14478 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
14479 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
14480 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
14481 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
14482 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
14483 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
14484 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
14485 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
14489 * journalctl gained new --since= and --until= switches to
14490 filter by time. It also now supports nice filtering for
14491 units via --unit=/-u.
14493 * Type=oneshot services may use ExecReload= and do the
14496 * The journal daemon now supports time-based rotation and
14497 vacuuming, in addition to the usual disk-space based
14500 * The journal will now index the available field values for
14501 each field name. This enables clients to show pretty drop
14502 downs of available match values when filtering. The bash
14503 completion of journalctl has been updated
14504 accordingly. journalctl gained a new switch -F to list all
14505 values a certain field takes in the journal database.
14507 * More service events are now written as structured messages
14508 to the journal, and made recognizable via message IDs.
14510 * The timedated, localed and hostnamed mini-services which
14511 previously only provided support for changing time, locale
14512 and hostname settings from graphical DEs such as GNOME now
14513 also have a minimal (but very useful) text-based client
14514 utility each. This is probably the nicest way to changing
14515 these settings from the command line now, especially since
14516 it lists available options and is fully integrated with bash
14519 * There's now a new tool "systemd-coredumpctl" to list and
14520 extract coredumps from the journal.
14522 * We now install a README each in /var/log/ and
14523 /etc/rc.d/init.d explaining where the system logs and init
14524 scripts went. This hopefully should help folks who go to
14525 that dirs and look into the otherwise now empty void and
14526 scratch their heads.
14528 * When user-services are invoked (by systemd --user) the
14529 $MANAGERPID env var is set to the PID of systemd.
14531 * SIGRTMIN+24 when sent to a --user instance will now result
14532 in immediate termination of systemd.
14534 * gatewayd received numerous feature additions such as a
14535 "follow" mode, for live syncing and filtering.
14537 * browse.html now allows filtering and showing detailed
14538 information on specific entries. Keyboard navigation and
14539 mouse screen support has been added.
14541 * gatewayd/journalctl now supports HTML5/JSON
14542 Server-Sent-Events as output.
14544 * The SysV init script compatibility logic will now
14545 heuristically determine whether a script supports the
14546 "reload" verb, and only then make this available as
14547 "systemctl reload".
14549 * "systemctl status --follow" has been removed, use "journalctl
14552 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
14553 have been removed since they are hardly useful to be
14556 * And I'd like to take the opportunity to specifically mention
14557 Zbigniew for his great contributions. Zbigniew, you rock!
14559 Contributions from: Andrew Eikum, Christian Hesse, Colin
14560 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
14561 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
14562 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
14563 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
14564 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
14565 Jędrzejewski-Szmek, Сковорода Никита Андреевич
14569 * If /etc/vconsole.conf is non-existent or empty we will no
14570 longer load any console font or key map at boot by
14571 default. Instead the kernel defaults will be left
14572 intact. This is definitely the right thing to do, as no
14573 configuration should mean no configuration, and hard-coding
14574 font names that are different on all archs is probably a bad
14575 idea. Also, the kernel default key map and font should be
14576 good enough for most cases anyway, and mostly identical to
14577 the userspace fonts/key maps we previously overloaded them
14578 with. If distributions want to continue to default to a
14579 non-kernel font or key map they should ship a default
14580 /etc/vconsole.conf with the appropriate contents.
14582 Contributions from: Colin Walters, Daniel J Walsh, Dave
14583 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
14584 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
14588 * journalctl gained a new --cursor= switch to show entries
14589 starting from the specified location in the journal.
14591 * We now enforce a size limit on journal entry fields exported
14592 with "-o json" in journalctl. Fields larger than 4K will be
14593 assigned null. This can be turned off with --all.
14595 * An (optional) journal gateway daemon is now available as
14596 "systemd-journal-gatewayd.service". This service provides
14597 access to the journal via HTTP and JSON. This functionality
14598 will be used to implement live log synchronization in both
14599 pull and push modes, but has various other users too, such
14600 as easy log access for debugging of embedded devices. Right
14601 now it is already useful to retrieve the journal via HTTP:
14603 # systemctl start systemd-journal-gatewayd.service
14604 # wget http://localhost:19531/entries
14606 This will download the journal contents in a
14607 /var/log/messages compatible format. The same as JSON:
14609 # curl -H"Accept: application/json" http://localhost:19531/entries
14611 This service is also accessible via a web browser where a
14612 single static HTML5 app is served that uses the JSON logic
14613 to enable the user to do some basic browsing of the
14614 journal. This will be extended later on. Here's an example
14615 screenshot of this app in its current state:
14617 http://0pointer.de/public/journal-gatewayd
14619 Contributions from: Kay Sievers, Lennart Poettering, Robert
14620 Milasan, Tom Gundersen
14624 * The bash completion logic is now available for journalctl
14627 * We do not mount the "cpuset" controller anymore together with
14628 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
14629 started if no parameters are assigned to it. "cpuset" hence
14630 broke code that assumed it could create "cpu" groups and
14633 * journalctl -f will now subscribe to terminal size changes,
14634 and line break accordingly.
14636 Contributions from: Dave Reisner, Kay Sievers, Lennart
14637 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
14641 * nspawn will now create a symlink /etc/localtime in the
14642 container environment, copying the host's timezone
14643 setting. Previously this has been done via a bind mount, but
14644 since symlinks cannot be bind mounted this has now been
14645 changed to create/update the appropriate symlink.
14647 * journalctl -n's line number argument is now optional, and
14648 will default to 10 if omitted.
14650 * journald will now log the maximum size the journal files may
14651 take up on disk. This is particularly useful if the default
14652 built-in logic of determining this parameter from the file
14653 system size is used. Use "systemctl status
14654 systemd-journald.service" to see this information.
14656 * The multi-seat X wrapper tool has been stripped down. As X
14657 is now capable of enumerating graphics devices via udev in a
14658 seat-aware way the wrapper is not strictly necessary
14659 anymore. A stripped down temporary stop-gap is still shipped
14660 until the upstream display managers have been updated to
14661 fully support the new X logic. Expect this wrapper to be
14662 removed entirely in one of the next releases.
14664 * HandleSleepKey= in logind.conf has been split up into
14665 HandleSuspendKey= and HandleHibernateKey=. The old setting
14666 is not available anymore. X11 and the kernel are
14667 distinguishing between these keys and we should too. This
14668 also means the inhibition lock for these keys has been split
14671 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
14672 Poettering, Lukas Nykryn, Václav Pavlín
14676 * Whenever a unit changes state we will now log this to the
14677 journal and show along the unit's own log output in
14678 "systemctl status".
14680 * ConditionPathIsMountPoint= can now properly detect bind
14681 mount points too. (Previously, a bind mount of one file
14682 system to another place in the same file system could not be
14683 detected as mount, since they shared struct stat's st_dev
14686 * We will now mount the cgroup controllers cpu, cpuacct,
14687 cpuset and the controllers net_cls, net_prio together by
14690 * nspawn containers will now have a virtualized boot
14691 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
14692 over with a randomized ID at container initialization). This
14693 has the effect of making "journalctl -b" do the right thing
14696 * The JSON output journal serialization has been updated not
14697 to generate "endless" list objects anymore, but rather one
14698 JSON object per line. This is more in line how most JSON
14699 parsers expect JSON objects. The new output mode
14700 "json-pretty" has been added to provide similar output, but
14701 neatly aligned for readability by humans.
14703 * We dropped all explicit sync() invocations in the shutdown
14704 code. The kernel does this implicitly anyway in the kernel
14705 reboot() syscall. halt(8)'s -n option is now a compatibility
14708 * We now support virtualized reboot() in containers, as
14709 supported by newer kernels. We will fall back to exit() if
14710 CAP_SYS_REBOOT is not available to the container. Also,
14711 nspawn makes use of this now and will actually reboot the
14712 container if the containerized OS asks for that.
14714 * journalctl will only show local log output by default
14715 now. Use --merge (-m) to show remote log output, too.
14717 * libsystemd-journal gained the new sd_journal_get_usage()
14718 call to determine the current disk usage of all journal
14719 files. This is exposed in the new "journalctl --disk-usage"
14722 * journald gained a new configuration setting SplitMode= in
14723 journald.conf which may be used to control how user journals
14724 are split off. See journald.conf(5) for details.
14726 * A new condition type ConditionFileNotEmpty= has been added.
14728 * tmpfiles' "w" lines now support file globbing, to write
14729 multiple files at once.
14731 * We added Python bindings for the journal submission
14732 APIs. More Python APIs for a number of selected APIs will
14733 likely follow. Note that we intend to add native bindings
14734 only for the Python language, as we consider it common
14735 enough to deserve bindings shipped within systemd. There are
14736 various projects outside of systemd that provide bindings
14737 for languages such as PHP or Lua.
14739 * Many conditions will now resolve specifiers such as %i. In
14740 addition, PathChanged= and related directives of .path units
14741 now support specifiers as well.
14743 * There's now a new RPM macro definition for the system preset
14746 * journald will now warn if it ca not forward a message to the
14747 syslog daemon because its socket is full.
14749 * timedated will no longer write or process /etc/timezone,
14750 except on Debian. As we do not support late mounted /usr
14751 anymore /etc/localtime always being a symlink is now safe,
14752 and hence the information in /etc/timezone is not necessary
14755 * logind will now always reserve one VT for a text getty (VT6
14756 by default). Previously if more than 6 X sessions where
14757 started they took up all the VTs with auto-spawned gettys,
14758 so that no text gettys were available anymore.
14760 * udev will now automatically inform the btrfs kernel logic
14761 about btrfs RAID components showing up. This should make
14762 simple hotplug based btrfs RAID assembly work.
14764 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
14765 (but not for its children which will stay at the kernel
14766 default). This should allow setups with a lot more listening
14769 * systemd will now always pass the configured timezone to the
14770 kernel at boot. timedated will do the same when the timezone
14773 * logind's inhibition logic has been updated. By default,
14774 logind will now handle the lid switch, the power and sleep
14775 keys all the time, even in graphical sessions. If DEs want
14776 to handle these events on their own they should take the new
14777 handle-power-key, handle-sleep-key and handle-lid-switch
14778 inhibitors during their runtime. A simple way to achieve
14779 that is to invoke the DE wrapped in an invocation of:
14781 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch …
14783 * Access to unit operations is now checked via SELinux taking
14784 the unit file label and client process label into account.
14786 * systemd will now notify the administrator in the journal
14787 when he over-mounts a non-empty directory.
14789 * There are new specifiers that are resolved in unit files,
14790 for the hostname (%H), the machine ID (%m) and the boot ID
14793 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
14794 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
14795 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
14796 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
14797 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
14798 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
14799 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
14803 * Support for reading structured kernel messages from
14804 /dev/kmsg has now been added and is enabled by default.
14806 * Support for reading kernel messages from /proc/kmsg has now
14807 been removed. If you want kernel messages in the journal
14808 make sure to run a recent kernel (>= 3.5) that supports
14809 reading structured messages from /dev/kmsg (see
14810 above). /proc/kmsg is now exclusive property of classic
14811 syslog daemons again.
14813 * The libudev API gained the new
14814 udev_device_new_from_device_id() call.
14816 * The logic for file system namespace (ReadOnlyDirectory=,
14817 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
14818 require pivot_root() anymore. This means fewer temporary
14819 directories are created below /tmp for this feature.
14821 * nspawn containers will now see and receive all submounts
14822 made on the host OS below the root file system of the
14825 * Forward Secure Sealing is now supported for Journal files,
14826 which provide cryptographical sealing of journal files so
14827 that attackers cannot alter log history anymore without this
14828 being detectable. Lennart will soon post a blog story about
14829 this explaining it in more detail.
14831 * There are two new service settings RestartPreventExitStatus=
14832 and SuccessExitStatus= which allow configuration of exit
14833 status (exit code or signal) which will be excepted from the
14834 restart logic, resp. consider successful.
14836 * journalctl gained the new --verify switch that can be used
14837 to check the integrity of the structure of journal files and
14838 (if Forward Secure Sealing is enabled) the contents of
14841 * nspawn containers will now be run with /dev/stdin, /dev/fd/
14842 and similar symlinks pre-created. This makes running shells
14843 as container init process a lot more fun.
14845 * The fstab support can now handle PARTUUID= and PARTLABEL=
14848 * A new ConditionHost= condition has been added to match
14849 against the hostname (with globs) and machine ID. This is
14850 useful for clusters where a single OS image is used to
14851 provision a large number of hosts which shall run slightly
14852 different sets of services.
14854 * Services which hit the restart limit will now be placed in a
14857 Contributions from: Bertram Poettering, Dave Reisner, Huang
14858 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
14859 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
14863 * When running in --user mode systemd will now become a
14864 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
14865 tree a lot more organized.
14867 * A new PartOf= unit dependency type has been introduced that
14868 may be used to group services in a natural way.
14870 * "systemctl enable" may now be used to enable instances of
14873 * journalctl now prints error log levels in red, and
14874 warning/notice log levels in bright white. It also supports
14875 filtering by log level now.
14877 * cgtop gained a new -n switch (similar to top), to configure
14878 the maximum number of iterations to run for. It also gained
14879 -b, to run in batch mode (accepting no input).
14881 * The suffix ".service" may now be omitted on most systemctl
14882 command lines involving service unit names.
14884 * There's a new bus call in logind to lock all sessions, as
14885 well as a loginctl verb for it "lock-sessions".
14887 * libsystemd-logind.so gained a new call sd_journal_perror()
14888 that works similar to libc perror() but logs to the journal
14889 and encodes structured information about the error number.
14891 * /etc/crypttab entries now understand the new keyfile-size=
14894 * shutdown(8) now can send a (configurable) wall message when
14895 a shutdown is cancelled.
14897 * The mount propagation mode for the root file system will now
14898 default to "shared", which is useful to make containers work
14899 nicely out-of-the-box so that they receive new mounts from
14900 the host. This can be undone locally by running "mount
14901 --make-rprivate /" if needed.
14903 * The prefdm.service file has been removed. Distributions
14904 should maintain this unit downstream if they intend to keep
14905 it around. However, we recommend writing normal unit files
14906 for display managers instead.
14908 * Since systemd is a crucial part of the OS we will now
14909 default to a number of compiler switches that improve
14910 security (hardening) such as read-only relocations, stack
14911 protection, and suchlike.
14913 * The TimeoutSec= setting for services is now split into
14914 TimeoutStartSec= and TimeoutStopSec= to allow configuration
14915 of individual time outs for the start and the stop phase of
14918 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
14919 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
14920 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
14921 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
14922 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
14923 Gundersen, Zbigniew Jędrzejewski-Szmek
14927 * The journal and id128 C APIs are now fully documented as man
14930 * Extra safety checks have been added when transitioning from
14931 the initial RAM disk to the main system to avoid accidental
14934 * /etc/crypttab entries now understand the new keyfile-offset=
14937 * systemctl -t can now be used to filter by unit load state.
14939 * The journal C API gained the new sd_journal_wait() call to
14940 make writing synchronous journal clients easier.
14942 * journalctl gained the new -D switch to show journals from a
14943 specific directory.
14945 * journalctl now displays a special marker between log
14946 messages of two different boots.
14948 * The journal is now explicitly flushed to /var via a service
14949 systemd-journal-flush.service, rather than implicitly simply
14950 by seeing /var/log/journal to be writable.
14952 * journalctl (and the journal C APIs) can now match for much
14953 more complex expressions, with alternatives and
14956 * When transitioning from the initial RAM disk to the main
14957 system we will now kill all processes in a killing spree to
14958 ensure no processes stay around by accident.
14960 * Three new specifiers may be used in unit files: %u, %h, %s
14961 resolve to the user name, user home directory resp. user
14962 shell. This is useful for running systemd user instances.
14964 * We now automatically rotate journal files if their data
14965 object hash table gets a fill level > 75%. We also size the
14966 hash table based on the configured maximum file size. This
14967 together should lower hash collisions drastically and thus
14968 speed things up a bit.
14970 * journalctl gained the new "--header" switch to introspect
14971 header data of journal files.
14973 * A new setting SystemCallFilters= has been added to services which may
14974 be used to apply deny lists or allow lists to system calls. This is
14975 based on SECCOMP Mode 2 of Linux 3.5.
14977 * nspawn gained a new --link-journal= switch (and quicker: -j)
14978 to link the container journal with the host. This makes it
14979 very easy to centralize log viewing on the host for all
14980 guests while still keeping the journal files separated.
14982 * Many bugfixes and optimizations
14984 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
14985 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
14986 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
14991 * Several tools now understand kernel command line arguments,
14992 which are only read when run in an initial RAM disk. They
14993 usually follow closely their normal counterparts, but are
14996 * There's a new tool to analyze the readahead files that are
14997 automatically generated at boot. Use:
14999 /usr/lib/systemd/systemd-readahead analyze /.readahead
15001 * We now provide an early debug shell on tty9 if this enabled. Use:
15003 systemctl enable debug-shell.service
15005 * All plymouth related units have been moved into the Plymouth
15006 package. Please make sure to upgrade your Plymouth version
15009 * systemd-tmpfiles now supports getting passed the basename of
15010 a configuration file only, in which case it will look for it
15011 in all appropriate directories automatically.
15013 * udevadm info now takes a /dev or /sys path as argument, and
15014 does the right thing. Example:
15016 udevadm info /dev/sda
15017 udevadm info /sys/class/block/sda
15019 * systemctl now prints a warning if a unit is stopped but a
15020 unit that might trigger it continues to run. Example: a
15021 service is stopped but the socket that activates it is left
15024 * "systemctl status" will now mention if the log output was
15025 shortened due to rotation since a service has been started.
15027 * The journal API now exposes functions to determine the
15028 "cutoff" times due to rotation.
15030 * journald now understands SIGUSR1 and SIGUSR2 for triggering
15031 immediately flushing of runtime logs to /var if possible,
15032 resp. for triggering immediate rotation of the journal
15035 * It is now considered an error if a service is attempted to
15036 be stopped that is not loaded.
15038 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
15040 * systemd-analyze now supports Python 3
15042 * tmpfiles now supports cleaning up directories via aging
15043 where the first level dirs are always kept around but
15044 directories beneath it automatically aged. This is enabled
15045 by prefixing the age field with '~'.
15047 * Seat objects now expose CanGraphical, CanTTY properties
15048 which is required to deal with very fast bootups where the
15049 display manager might be running before the graphics drivers
15050 completed initialization.
15052 * Seat objects now expose a State property.
15054 * We now include RPM macros for service enabling/disabling
15055 based on the preset logic. We recommend RPM based
15056 distributions to make use of these macros if possible. This
15057 makes it simpler to reuse RPM spec files across
15060 * We now make sure that the collected systemd unit name is
15061 always valid when services log to the journal via
15064 * There's a new man page kernel-command-line(7) detailing all
15065 command line options we understand.
15067 * The fstab generator may now be disabled at boot by passing
15068 fstab=0 on the kernel command line.
15070 * A new kernel command line option modules-load= is now understood
15071 to load a specific kernel module statically, early at boot.
15073 * Unit names specified on the systemctl command line are now
15074 automatically escaped as needed. Also, if file system or
15075 device paths are specified they are automatically turned
15076 into the appropriate mount or device unit names. Example:
15078 systemctl status /home
15079 systemctl status /dev/sda
15081 * The SysVConsole= configuration option has been removed from
15082 system.conf parsing.
15084 * The SysV search path is no longer exported on the D-Bus
15087 * The Names= option has been removed from unit file parsing.
15089 * There's a new man page bootup(7) detailing the boot process.
15091 * Every unit and every generator we ship with systemd now
15092 comes with full documentation. The self-explanatory boot is
15095 * A couple of services gained "systemd-" prefixes in their
15096 name if they wrap systemd code, rather than only external
15097 code. Among them fsck@.service which is now
15098 systemd-fsck@.service.
15100 * The HaveWatchdog property has been removed from the D-Bus
15103 * systemd.confirm_spawn= on the kernel command line should now
15106 * There's a new man page crypttab(5) which details all options
15107 we actually understand.
15109 * systemd-nspawn gained a new --capability= switch to pass
15110 additional capabilities to the container.
15112 * timedated will now read known NTP implementation unit names
15113 from /usr/lib/systemd/ntp-units.d/*.list,
15114 systemd-timedated-ntp.target has been removed.
15116 * journalctl gained a new switch "-b" that lists log data of
15117 the current boot only.
15119 * The notify socket is in the abstract namespace again, in
15120 order to support daemons which chroot() at start-up.
15122 * There is a new Storage= configuration option for journald
15123 which allows configuration of where log data should go. This
15124 also provides a way to disable journal logging entirely, so
15125 that data collected is only forwarded to the console, the
15126 kernel log buffer or another syslog implementation.
15128 * Many bugfixes and optimizations
15130 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
15131 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
15132 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
15133 Shawn Landden, Tom Gundersen
15137 * "systemctl help <unit>" now shows the man page if one is
15140 * Several new man pages have been added.
15142 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
15143 MaxLevelConsole= can now be specified in
15144 journald.conf. These options allow reducing the amount of
15145 data stored on disk or forwarded by the log level.
15147 * TimerSlackNSec= can now be specified in system.conf for
15148 PID1. This allows system-wide power savings.
15150 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
15151 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
15156 * logind is now capable of (optionally) handling power and
15157 sleep keys as well as the lid switch.
15159 * journalctl now understands the syntax "journalctl
15160 /usr/bin/avahi-daemon" to get all log output of a specific
15163 * CapabilityBoundingSet= in system.conf now also influences
15164 the capability bound set of usermode helpers of the kernel.
15166 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
15167 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
15168 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
15169 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
15173 * Note that we skipped 139 releases here in order to set the
15174 new version to something that is greater than both udev's
15175 and systemd's most recent version number.
15177 * udev: all udev sources are merged into the systemd source tree now.
15178 All future udev development will happen in the systemd tree. It
15179 is still fully supported to use the udev daemon and tools without
15180 systemd running, like in initramfs or other init systems. Building
15181 udev though, will require the *build* of the systemd tree, but
15182 udev can be properly *run* without systemd.
15184 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
15185 should be used to create dead device nodes as workarounds for broken
15188 * udev: RUN+="socket:…" and udev_monitor_new_from_socket() is
15189 no longer supported. udev_monitor_new_from_netlink() needs to be
15190 used to subscribe to events.
15192 * udev: when udevd is started by systemd, processes which are left
15193 behind by forking them off of udev rules, are unconditionally cleaned
15194 up and killed now after the event handling has finished. Services or
15195 daemons must be started as systemd services. Services can be
15196 pulled-in by udev to get started, but they can no longer be directly
15197 forked by udev rules.
15199 * udev: the daemon binary is called systemd-udevd now and installed
15200 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
15201 to adapt to that, create symlink, or rename the binary after building
15204 * libudev no longer provides these symbols:
15205 udev_monitor_from_socket()
15206 udev_queue_get_failed_list_entry()
15207 udev_get_{dev,sys,run}_path()
15208 The versions number was bumped and symbol versioning introduced.
15210 * systemd-loginctl and systemd-journalctl have been renamed
15211 to loginctl and journalctl to match systemctl.
15213 * The config files: /etc/systemd/systemd-logind.conf and
15214 /etc/systemd/systemd-journald.conf have been renamed to
15215 logind.conf and journald.conf. Package updates should rename
15216 the files to the new names on upgrade.
15218 * For almost all files the license is now LGPL2.1+, changed
15219 from the previous GPL2.0+. Exceptions are some minor stuff
15220 of udev (which will be changed to LGPL2.1 eventually, too),
15221 and the MIT licensed sd-daemon.[ch] library that is suitable
15222 to be used as drop-in files.
15224 * systemd and logind now handle system sleep states, in
15225 particular suspending and hibernating.
15227 * logind now implements a sleep/shutdown/idle inhibiting logic
15228 suitable for a variety of uses. Soonishly Lennart will blog
15229 about this in more detail.
15231 * var-run.mount and var-lock.mount are no longer provided
15232 (which previously bind mounted these directories to their new
15233 places). Distributions which have not converted these
15234 directories to symlinks should consider stealing these files
15235 from git history and add them downstream.
15237 * We introduced the Documentation= field for units and added
15238 this to all our shipped units. This is useful to make it
15239 easier to explore the boot and the purpose of the various
15242 * All smaller setup units (such as
15243 systemd-vconsole-setup.service) now detect properly if they
15244 are run in a container and are skipped when
15245 appropriate. This guarantees an entirely noise-free boot in
15246 Linux container environments such as systemd-nspawn.
15248 * A framework for implementing offline system updates is now
15249 integrated, for details see:
15250 https://www.freedesktop.org/software/systemd/man/systemd.offline-updates.html
15252 * A new service type Type=idle is available now which helps us
15253 avoiding ugly interleaving of getty output and boot status
15256 * There's now a system-wide CapabilityBoundingSet= option to
15257 globally reduce the set of capabilities for the
15258 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
15259 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
15260 even CAP_NET_ADMIN system-wide for secure systems.
15262 * There are now system-wide DefaultLimitXXX= options to
15263 globally change the defaults of the various resource limits
15264 for all units started by PID 1.
15266 * Harald Hoyer's systemd test suite has been integrated into
15267 systemd which allows easy testing of systemd builds in qemu
15268 and nspawn. (This is really awesome! Ask us for details!)
15270 * The fstab parser is now implemented as generator, not inside
15273 * systemctl will now warn you if .mount units generated from
15274 /etc/fstab are out of date due to changes in fstab that
15275 have not been read by systemd yet.
15277 * systemd is now suitable for usage in initrds. Dracut has
15278 already been updated to make use of this. With this in place
15279 initrds get a slight bit faster but primarily are much
15280 easier to introspect and debug since "systemctl status" in
15281 the host system can be used to introspect initrd services,
15282 and the journal from the initrd is kept around too.
15284 * systemd-delta has been added, a tool to explore differences
15285 between user/admin configuration and vendor defaults.
15287 * PrivateTmp= now affects both /tmp and /var/tmp.
15289 * Boot time status messages are now much prettier and feature
15290 proper english language. Booting up systemd has never been
15293 * Read-ahead pack files now include the inode number of all
15294 files to pre-cache. When the inode changes the pre-caching
15295 is not attempted. This should be nicer to deal with updated
15296 packages which might result in changes of read-ahead
15299 * We now temporaritly lower the kernel's read_ahead_kb variable
15300 when collecting read-ahead data to ensure the kernel's
15301 built-in read-ahead does not add noise to our measurements
15302 of necessary blocks to pre-cache.
15304 * There's now RequiresMountsFor= to add automatic dependencies
15305 for all mounts necessary for a specific file system path.
15307 * MountAuto= and SwapAuto= have been removed from
15308 system.conf. Mounting file systems at boot has to take place
15311 * nspawn now learned a new switch --uuid= to set the machine
15312 ID on the command line.
15314 * nspawn now learned the -b switch to automatically search
15315 for an init system.
15317 * vt102 is now the default TERM for serial TTYs, upgraded from
15320 * systemd-logind now works on VT-less systems.
15322 * The build tree has been reorganized. The individual
15323 components now have directories of their own.
15325 * A new condition type ConditionPathIsReadWrite= is now available.
15327 * nspawn learned the new -C switch to create cgroups for the
15328 container in other hierarchies.
15330 * We now have support for hardware watchdogs, configurable in
15333 * The scheduled shutdown logic now has a public API.
15335 * We now mount /tmp as tmpfs by default, but this can be
15336 masked and /etc/fstab can override it.
15338 * Since udisks does not make use of /media anymore we are not
15339 mounting a tmpfs on it anymore.
15341 * journalctl gained a new --local switch to only interleave
15342 locally generated journal files.
15344 * We can now load the IMA policy at boot automatically.
15346 * The GTK tools have been split off into a systemd-ui.
15348 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
15349 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
15350 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
15351 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
15352 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
15353 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
15354 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
15355 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
15356 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
15361 * This is mostly a bugfix release
15363 * Support optional initialization of the machine ID from the
15364 KVM or container configured UUID.
15366 * Support immediate reboots with "systemctl reboot -ff"
15368 * Show /etc/os-release data in systemd-analyze output
15370 * Many bugfixes for the journal, including endianness fixes and
15371 ensuring that disk space enforcement works
15373 * sd-login.h is C++ compatible again
15375 * Extend the /etc/os-release format on request of the Debian
15378 * We now refuse non-UTF8 strings used in various configuration
15379 and unit files. This is done to ensure we do not pass invalid
15380 data over D-Bus or expose it elsewhere.
15382 * Register Mimo USB Screens as suitable for automatic seat
15385 * Read SELinux client context from journal clients in a race
15388 * Reorder configuration file lookup order. /etc now always
15389 overrides /run in order to allow the administrator to always
15390 and unconditionally override vendor-supplied or
15391 automatically generated data.
15393 * The various user visible bits of the journal now have man
15394 pages. We still lack man pages for the journal API calls
15397 * We now ship all man pages in HTML format again in the
15400 Contributions from: Dave Reisner, Dirk Eibach, Frederic
15401 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
15402 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
15407 * This is mostly a bugfix release
15409 * systems lacking /etc/os-release are no longer supported.
15411 * Various functionality updates to libsystemd-login.so
15413 * Track class of PAM logins to distinguish greeters from
15414 normal user logins.
15416 Contributions from: Kay Sievers, Lennart Poettering, Michael
15421 * This is an important bugfix release for v41.
15423 * Building man pages is now optional which should be useful
15424 for those building systemd from git but unwilling to install
15427 * Watchdog support for supervising services is now usable. In
15428 a future release support for hardware watchdogs
15429 (i.e. /dev/watchdog) will be added building on this.
15431 * Service start rate limiting is now configurable and can be
15432 turned off per service. When a start rate limit is hit a
15433 reboot can automatically be triggered.
15435 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
15437 Contributions from: Benjamin Franzke, Bill Nottingham,
15438 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
15439 Schmidt, Michał Górny, Piotr Drąg
15443 * The systemd binary is installed /usr/lib/systemd/systemd now;
15444 An existing /sbin/init symlink needs to be adapted with the
15447 * The code that loads kernel modules has been ported to invoke
15448 libkmod directly, instead of modprobe. This means we do not
15449 support systems with module-init-tools anymore.
15451 * Watchdog support is now already useful, but still not
15454 * A new kernel command line option systemd.setenv= is
15455 understood to set system wide environment variables
15456 dynamically at boot.
15458 * We now limit the set of capabilities of systemd-journald.
15460 * We now set SIGPIPE to ignore by default, since it only is
15461 useful in shell pipelines, and has little use in general
15462 code. This can be disabled with IgnoreSIPIPE=no in unit
15465 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
15466 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
15471 * This is mostly a bugfix release
15473 * We now expose the reason why a service failed in the
15474 "Result" D-Bus property.
15476 * Rudimentary service watchdog support (will be completed over
15477 the next few releases.)
15479 * When systemd forks off in order execute some service we will
15480 now immediately changes its argv[0] to reflect which process
15481 it will execute. This is useful to minimize the time window
15482 with a generic argv[0], which makes bootcharts more useful
15484 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
15485 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
15486 Mike Kazantsev, Ray Strode
15490 * This is mostly a test release, but incorporates many
15493 * New systemd-cgtop tool to show control groups by their
15496 * Linking against libacl for ACLs is optional again. If
15497 disabled, support tracking device access for active logins
15498 goes becomes unavailable, and so does access to the user
15499 journals by the respective users.
15501 * If a group "adm" exists, journal files are automatically
15502 owned by them, thus allow members of this group full access
15503 to the system journal as well as all user journals.
15505 * The journal now stores the SELinux context of the logging
15506 client for all entries.
15508 * Add C++ inclusion guards to all public headers
15510 * New output mode "cat" in the journal to print only text
15511 messages, without any meta data like date or time.
15513 * Include tiny X server wrapper as a temporary stop-gap to
15514 teach XOrg udev display enumeration. This is used by display
15515 managers such as gdm, and will go away as soon as XOrg
15516 learned native udev hotplugging for display devices.
15518 * Add new systemd-cat tool for executing arbitrary programs
15519 with STDERR/STDOUT connected to the journal. Can also act as
15520 BSD logger replacement, and does so by default.
15522 * Optionally store all locally generated coredumps in the
15523 journal along with meta data.
15525 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
15526 writing short strings to files (for usage for /sys), and for
15527 creating symlinks, character and block device nodes.
15529 * New unit file option ControlGroupPersistent= to make cgroups
15530 persistent, following the mechanisms outlined in
15531 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
15533 * Support multiple local RTCs in a sane way
15535 * No longer monopolize IO when replaying readahead data on
15536 rotating disks, since we might starve non-file-system IO to
15537 death, since fanotify() will not see accesses done by blkid,
15540 * Do not show kernel threads in systemd-cgls anymore, unless
15541 requested with new -k switch.
15543 Contributions from: Dan Horák, Kay Sievers, Lennart
15544 Poettering, Michal Schmidt
15548 * This is mostly a test release, but incorporates many
15551 * The git repository moved to:
15552 git://anongit.freedesktop.org/systemd/systemd
15553 ssh://git.freedesktop.org/git/systemd/systemd
15555 * First release with the journal
15556 http://0pointer.de/blog/projects/the-journal.html
15558 * The journal replaces both systemd-kmsg-syslogd and
15559 systemd-stdout-bridge.
15561 * New sd_pid_get_unit() API call in libsystemd-logind
15563 * Many systemadm clean-ups
15565 * Introduce remote-fs-pre.target which is ordered before all
15566 remote mounts and may be used to start services before all
15569 * Added Mageia support
15571 * Add bash completion for systemd-loginctl
15573 * Actively monitor PID file creation for daemons which exit in
15574 the parent process before having finished writing the PID
15575 file in the daemon process. Daemons which do this need to be
15576 fixed (i.e. PID file creation must have finished before the
15577 parent exits), but we now react a bit more gracefully to them.
15579 * Add colourful boot output, mimicking the well-known output
15580 of existing distributions.
15582 * New option PassCredentials= for socket units, for
15583 compatibility with a recent kernel ABI breakage.
15585 * /etc/rc.local is now hooked in via a generator binary, and
15586 thus will no longer act as synchronization point during
15589 * systemctl list-unit-files now supports --root=.
15591 * systemd-tmpfiles now understands two new commands: z, Z for
15592 relabelling files according to the SELinux database. This is
15593 useful to apply SELinux labels to specific files in /sys,
15594 among other things.
15596 * Output of SysV services is now forwarded to both the console
15597 and the journal by default, not only just the console.
15599 * New man pages for all APIs from libsystemd-login.
15601 * The build tree got reorganized and the build system is a
15602 lot more modular allowing embedded setups to specifically
15603 select the components of systemd they are interested in.
15605 * Support for Linux systems lacking the kernel VT subsystem is
15608 * configure's --with-rootdir= got renamed to
15609 --with-rootprefix= to follow the naming used by udev and
15612 * Unless specified otherwise we will now install to /usr instead
15613 of /usr/local by default.
15615 * Processes with '@' in argv[0][0] are now excluded from the
15616 final shut-down killing spree, following the logic explained
15618 https://systemd.io/ROOT_STORAGE_DAEMONS/
15620 * All processes remaining in a service cgroup when we enter
15621 the START or START_PRE states are now killed with
15622 SIGKILL. That means it is no longer possible to spawn
15623 background processes from ExecStart= lines (which was never
15624 supported anyway, and bad style).
15626 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
15627 reloading of units together.
15629 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
15630 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
15631 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
15632 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
15633 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek