1 systemd System and Service Manager
3 CHANGES WITH 247 in spe:
5 * KERNEL API INCOMPATIBILITY: Linux 4.12 introduced two new uevents
6 "bind" and "unbind" to the Linux device model. When this kernel
7 change was made, systemd-udevd was only minimally updated to handle
8 and propagate these new event types. The introduction of these new
9 uevents (which are typically generated for USB devices and devices
10 needing a firmware upload before being functional) resulted in a
11 number of issues which we so far didn't address. We hoped the kernel
12 maintainers would themselves address these issues in some form, but
13 that did not happen. To handle them properly, many (if not most) udev
14 rules files shipped in various packages need updating, and so do many
15 programs that monitor or enumerate devices with libudev or sd-device,
16 or otherwise process uevents. Please note that this incompatibility
17 is not fault of systemd or udev, but caused by an incompatible kernel
18 change that happened back in Linux 4.12, but is becoming more and
19 more visible as the new uevents are generated by more kernel drivers.
21 To minimize issues resulting from this kernel change (but not avoid
22 them entirely) starting with systemd-udevd 247 the udev "tags"
23 concept (which is a concept for marking and filtering devices during
24 enumeration and monitoring) has been reworked: udev tags are now
25 "sticky", meaning that once a tag is assigned to a device it will not
26 be removed from the device again until the device itself is removed
27 (i.e. unplugged). This makes sure that any application monitoring
28 devices that match a specific tag is guaranteed to both see uevents
29 where the device starts being relevant, and those where it stops
30 being relevant (the latter now regularly happening due to the new
31 "unbind" uevent type). The udev tags concept is hence now a concept
32 tied to a *device* instead of a device *event* — unlike for example
33 udev properties whose lifecycle (as before) is generally tied to a
34 device event, meaning that the previously determined properties are
35 forgotten whenever a new uevent is processed.
37 With the newly redefined udev tags concept, sometimes it's necessary
38 to determine which tags are the ones applied by the most recent
39 uevent/database update, in order to discern them from those
40 originating from earlier uevents/database updates of the same
41 device. To accommodate for this a new automatic property CURRENT_TAGS
42 has been added that works similar to the existing TAGS property but
43 only lists tags set by the most recent uevent/database
44 update. Similarly, the libudev/sd-device API has been updated with
45 new functions to enumerate these 'current' tags, in addition to the
46 existing APIs that now enumerate the 'sticky' ones.
48 To properly handle "bind"/"unbind" on Linux 4.12 and newer it is
49 essential that all udev rules files and applications are updated to
50 handle the new events. Specifically:
52 • All rule files that currently use a header guard similar to
53 ACTION!="add|change",GOTO="xyz_end" should be updated to use
54 ACTION=="remove",GOTO="xyz_end" instead, so that the
55 properties/tags they add are also applied whenever "bind" (or
56 "unbind") is seen. (This is most important for all physical device
57 types — those for which "bind" and "unbind" are currently
58 generated, for all other device types this change is still
59 recommended but not as important — but certainly prepares for
60 future kernel uevent type additions).
62 • Similarly, all code monitoring devices that contains an 'if' branch
63 discerning the "add" + "change" uevent actions from all other
64 uevents actions (i.e. considering devices only relevant after "add"
65 or "change", and irrelevant on all other events) should be reworked
66 to instead negatively check for "remove" only (i.e. considering
67 devices relevant after all event types, except for "remove", which
68 invalidates the device). Note that this also means that devices
69 should be considered relevant on "unbind", even though conceptually
70 this — in some form — invalidates the device. Since the precise
71 effect of "unbind" is not generically defined, devices should be
72 considered relevant even after "unbind", however I/O errors
73 accessing the device should then be handled gracefully.
75 • Any code that uses device tags for deciding whether a device is
76 relevant or not most likely needs to be updated to use the new
77 udev_device_has_current_tag() API (or sd_device_has_current_tag()
78 in case sd-device is used), to check whether the tag is set at the
79 moment an uevent is seen (as opposed to the existing
80 udev_device_has_tag() API which checks if the tag ever existed on
81 the device, following the API concept redefinition explained
84 We are very sorry for this breakage and the requirement to update
85 packages using these interfaces. We'd again like to underline that
86 this is not caused by systemd/udev changes, but result of a kernel
89 * The MountAPIVFS= service file setting now defaults to on if
90 RootImage= and RootDirectory= are used, which means that with those
91 two settings /proc/, /sys/ and /dev/ are automatically properly set
92 up for services. Previous behaviour may be restored by explicitly
93 setting MountAPIVFS=off.
95 * Since PAM 1.2.0 (2015) configuration snippets may be placed in
96 /usr/lib/pam.d/ in addition to /etc/pam.d/. If a file exists in the
97 latter it takes precedence over the former, similar to how most of
98 systemd's own configuration is handled. Given that PAM stack
99 definitions are primarily put together by OS vendors/distributions
100 (though possibly overridden by users), this systemd release moves its
101 own PAM stack configuration for the "systemd-user" PAM service (i.e.
102 for the PAM session invoked by the per-user user@.service instance)
103 from /etc/pam.d/ to /usr/lib/pam.d/. We recommend moving all
104 packages' vendor versions of their PAM stack definitions from
105 /etc/pam.d/ to /usr/lib/pam.d/, but if such OS-wide migration is not
106 desired the location to which systemd installs its PAM stack
107 configuration may be changed via the -Dpamconfdir Meson option.
109 * The runtime dependencies on libqrencode, libpcre2, libidn/libidn2,
110 libpwquality and libcryptsetup have been changed to be based on
111 dlopen(): instead of regular dynamic library dependencies declared in
112 the binary ELF headers, these libraries are now loaded on demand
113 only, if they are available. If the libraries cannot be found the
114 relevant operations will fail gracefully, or a suitable fallback
115 logic is chosen. This is supposed to be useful for general purpose
116 distributions, as it allows minimizing the list of dependencies the
117 systemd packages pull in, permitting building of more minimal OS
118 images, while still making use of these "weak" dependencies should
119 they be installed. Since many package managers automatically
120 synthesize package dependencies from ELF shared library dependencies,
121 some additional manual packaging work has to be done now to replace
122 those (slightly downgraded from "required" to "recommended" or
123 whatever is conceptually suitable for the package manager). Note that
124 this change does not alter build-time behaviour: as before the
125 build-time dependencies have to be installed during build, even if
126 they now are optional during runtime.
128 * sd-event.h gained a new call sd_event_add_time_relative() for
129 installing timers relative to the current time. This is mostly a
130 convenience wrapper around the pre-existing sd_event_add_time() call
131 which installs absolute timers.
133 * sd-event event sources may now be placed in a new "exit-on-failure"
134 mode, which may be controlled via the new
135 sd_event_source_get_exit_on_failure() and
136 sd_event_source_set_exit_on_failure() functions. If enabled, any
137 failure returned by the event source handler functions will result in
138 exiting the event loop (unlike the default behaviour of just
139 disabling the event source but continuing with the event loop). This
140 feature is useful to set for all event sources that define "primary"
141 program behaviour (where failure should be fatal) in contrast to
142 "auxiliary" behaviour (where failure should remain local).
144 * Most event source types sd-event supports now accept a NULL handler
145 function, in which case the event loop is exited once the event
146 source is to be dispatched, using the userdata pointer — converted to
147 a signed integer — as exit code of the event loop. Previously this
148 was supported for IO and signal event sources already. Exit event
149 sources still do not support this (simply because it makes little
150 sense there, as the event loop is already exiting when they are
153 * A new per-unit setting RootImageOptions= has been added which allows
154 tweaking the mount options for any file system mounted as effect of
155 the RootImage= setting.
157 * Another new per-unit setting MountImages= has been added, that allows
158 mounting additional disk images into the file system tree accessible
161 * systemd-repart now generates JSON output when requested with the new
164 * systemd-machined's OpenMachineShell() bus call will now pass
165 additional policy metadata data fields to the PolicyKit
166 authentication request.
168 * systemd-tmpfiles gained a new -E switch, which is equivalent to
169 --exclude-prefix=/dev --exclude-prefix=/proc --exclude=/run
170 --exclude=/sys. It's particularly useful in combination with --root=,
171 when operating on OS trees that do not have any of these four runtime
172 directories mounted, as this means no files below these subtrees are
173 created or modified, since those mount points should probably remain
176 * systemd-tmpfiles gained a new --image= switch which is like --root=,
177 but takes a disk image instead of a directory as argument. The
178 specified disk image is mounted inside a temporary mount namespace
179 and the tmpfiles.d/ drop-ins stored in the image are executed and
180 applied to the image. systemd-sysusers similarly gained a new
181 --image= switch, that allows the sysusers.d/ drop-ins stored in the
182 image to be applied onto the image.
184 * Similarly, the journalctl command also gained an --image= switch,
185 which is a quick one-step solution to look at the log data included
188 * journalctl's --output=cat option (which outputs the log content
189 without any metadata, just the pure text messages) will now make use
190 of terminal colors when run on a suitable terminal, similarly to the
193 * JSON group records now support a "description" string that may be
194 used to add a human-readable textual description to such groups. This
195 is supposed to match the user's GECOS field which traditionally
196 didn't have a counterpart for group records.
198 * The "systemd-dissect" tool that may be used to inspect OS disk images
199 and that was previously installed to /usr/lib/systemd/ has now been
200 moved to /usr/bin/, reflecting its updated status of an officially
201 supported tool with a stable interface. It gained support for a new
202 --mkdir switch which when combined with --mount has the effect of
203 creating the directory to mount the image to if it is missing
204 first. It also gained two new commands --copy-from and --copy-to for
205 copying files and directories in and out of an OS image without the
206 need to manually mount it. It also acquired support for a new option
207 --json= to generate JSON output when inspecting an OS image.
209 * The cgroup2 file system is now mounted with the
210 "memory_recursiveprot" mount option, supported since kernel 5.7. This
211 means that the MemoryLow= and MemoryMin= unit file settings now apply
212 recursively to whole subtrees.
214 * systemd-homed now defaults to using the btrfs file system — if
215 available — when creating home directories in LUKS volumes. This may
216 be changed with the DefaultFileSystemType= setting in homed.conf.
217 It's now the default file system in various major distributions and
218 has the major benefit for homed that it can be grown and shrunk while
219 mounted, unlike the other contenders ext4 and xfs, which can both be
220 grown online, but not shrunk (in fact xfs is the technically most
221 limited option here, as it cannot be shrunk at all).
223 * JSON user records managed by systemd-homed gained support for
224 "recovery keys". These are basically secondary passphrases that can
225 unlock user accounts/home directories. They are computer-generated
226 rather than user-chosen, and typically have greater entropy.
227 homectl's --recovery-key= option may be used to add a recovery key to
228 a user account. The generated recovery key is displayed as a QR code,
229 so that it can be scanned to be kept in a safe place. This feature is
230 particularly useful in combination with systemd-homed's support for
231 FIDO2 or PKCS#11 authentication, as a secure fallback in case the
232 security tokens are lost. Recovery keys may be entered wherever the
233 system asks for a password.
235 * systemd-homed now maintains a "dirty" flag for each LUKS encrypted
236 home directory which indicates that a home directory has not been
237 deactivated cleanly when offline. This flag is useful to identify
238 home directories for which the offline discard logic did not run when
239 offlining, and where it would be a good idea to log in again to catch
242 * systemctl gained a new parameter --timestamp= which may be used to
243 change the style in which timestamps are output, i.e. whether to show
244 them in local timezone or UTC, or whether to show µs granularity.
246 * Alibaba's "pouch" container manager is now detected by
247 systemd-detect-virt, ConditionVirtualization= and similar
248 constructs. Similar, they now also recognize IBM PowerVM machine
251 * systemd-nspawn has been reworked to use the /run/host/incoming/ as
252 place to use for propagating external mounts into the
253 container. Similarly /run/host/notify is now used as the socket path
254 for container payloads to communicate with the container manager
255 using sd_notify(). The container manager now uses the
256 /run/host/inaccessible/ directory to place "inaccessible" file nodes
257 of all relevant types which may be used by the container payload as
258 bind mount source to over-mount inodes to make them inaccessible.
259 /run/host/container-manager will now be initialized with the same
260 string as the $container environment variable passed to the
261 container's PID 1. /run/host/container-uuid will be initialized with
262 the same string as $container_uuid. This means the /run/host/
263 hierarchy is now the primary way to make host resources available to
264 the container. The Container Interface documents these new files and
267 https://systemd.io/CONTAINER_INTERFACE
269 * Support for the "ConditionNull=" unit file condition has been
270 deprecated and undocumented for 6 years. systemd started to warn
271 about its use 1.5 years ago. It has now been removed entirely.
273 * sd-bus.h gained a new API call sd_bus_error_has_names(), which takes
274 a sd_bus_error struct and a list of error names, and checks if the
275 error matches one of these names. It's a convenience wrapper that is
276 useful in cases where multiple errors shall be handled the same way.
278 * A new system call filter list "@known" has been added, that contains
279 all system calls known at the time systemd was built.
281 * Behaviour of system call filter allow lists has changed slightly:
282 system calls that are contained in @known will result in a EPERM by
283 default, while those not contained in it result in ENOSYS. This
284 should improve compatibility because known system calls will thus be
285 communicated as prohibited, while unknown (and thus newer ones) will
286 be communicated as not implemented, which hopefully has the greatest
287 chance of triggering the right fallback code paths in client
290 * "systemd-analyze syscall-filter" will now show two separate sections
291 at the bottom of the output: system calls known during systemd build
292 time but not included in any of the filter groups shown above, and
293 system calls defined on the local kernel but known during systemd
296 * If the $SYSTEMD_LOG_SECCOMP=1 environment variable is set for
297 systemd-nspawn all system call filter violations will be logged by
298 the kernel (audit). This is useful for tracking down system calls
299 invoked by container payloads that are prohibited by the container's
300 system call filter policy.
302 * Two new unit file settings ProtectProc= and ProcSubset= have been
303 added that expose the hidepid= and subset= mount options of procfs.
304 All processes of the unit will only see processes in /proc that are
305 are owned by the unit's user. This is an important new sandboxing
306 option that is recommended to be set on all system services. All
307 long-running system services that are included in systemd itself set
308 this option now. This option is only supported on kernel 5.8 and
309 above, since the hidepid= option supported on older kernels was not a
310 per-mount option but actually applied to the whole PID namespace.
312 * Socket units gained a new boolean setting FlushPending=. If enabled
313 all pending socket data/connections are flushed whenever the socket
314 unit enters the "listening" state, i.e. after the associated service
317 * The unit file setting NUMAMask= gained a new "all" value: when used,
318 all existing NUMA nodes are added to the NUMA mask.
320 * A new "credentials" logic has been added to system services. This is
321 a simple mechanism to pass privileged data to services in a safe and
322 secure way. It's supposed to be used to pass per-service secret data
323 such as passwords or cryptographic keys but also associated less
324 private information such as user names, certificates, and similar to
325 system services. Each credential is identified by a short user-chosen
326 name and may contain arbitrary binary data. Two new unit file
327 settings have been added: SetCredential= and LoadCredential=. The
328 former allows setting a credential to a literal string, the latter
329 sets a credential to the contents of a file (or data read from a
330 user-chosen AF_UNIX stream socket). Credentials are passed to the
331 service via a special credentials directory, one file for each
332 credential. The path to the credentials directory is passed in a new
333 $CREDENTIALS_DIRECTORY environment variable. Since the credentials
334 are passed in the file system they may be easily referenced in
335 ExecStart= command lines too, thus no explicit support for the
336 credentials logic in daemons is required (though ideally daemons
337 would look for the bits they need in $CREDENTIALS_DIRECTORY
338 themselves automatically, if set). The $CREDENTIALS_DIRECTORY is
339 backed by unswappable memory if privileges allow it, immutable if
340 privileges allow it, is accessible only to the service's UID, and is
341 automatically destroyed when the service stops.
343 * systemd-nspawn supports the same credentials logic. It can both
344 consume credentials passed to it via the aforementioned
345 $CREDENTIALS_DIRECTORY protocol as well as pass these credentials on
346 to its payload. The service manager/PID 1 has been updated to match
347 this: it can also accept credentials from the container manager that
348 invokes it (in fact: any process that invokes it), and passes them on
349 to its services. Thus, credentials can be propagated recursively down
350 the tree: from a system's service manager to a systemd-nspawn
351 service, to the service manager that runs as container payload and to
352 the service it runs below. Credentials may also be added on the
353 systemd-nspawn command line, using new --set-credential= and
354 --load-credential= command line switches that match the
355 aforementioned service settings.
357 * systemd-repart gained new settings Format=, Encrypt=, CopyFiles= in
358 the partition drop-ins which may be used to format/LUKS
359 encrypt/populate any created partitions. The partitions are
360 encrypted/formatted/populated before they are registered in the
361 partition table, so that they appear atomically: either the
362 partitions do not exist yet or they exist fully encrypted, formatted,
363 and populated — there is no time window where they are
364 "half-initialized". Thus the system is robust to abrupt shutdown: if
365 the tool is terminated half-way during its operations on next boot it
366 will start from the beginning.
368 * systemd-repart's --size= operation gained a new "auto" value. If
369 specified, and operating on a loopback file it is automatically sized
370 to the minimal size the size constraints permit. This is useful to
371 use "systemd-repart" as an image builder for minimally sized images.
373 * systemd-resolved now gained a third IPC interface for requesting name
374 resolution: besides D-Bus and local DNS to 127.0.0.53 a Varlink
375 interface is now supported. The nss-resolve NSS module has been
376 modified to use this new interface instead of D-Bus. Using Varlink
377 has a major benefit over D-Bus: it works without a broker service,
378 and thus already during earliest boot, before the dbus daemon has
379 been started. This means name resolution via systemd-resolved now
380 works at the same time systemd-networkd operates: from earliest boot
381 on, including in the initrd.
383 * systemd-resolved gained support for a new DNSStubListenerExtra=
384 configuration file setting which may be used to specify additional IP
385 addresses the built-in DNS stub shall listen on, in addition to the
386 main one on 127.0.0.53:53.
388 * Name lookups issued via systemd-resolved's D-Bus and Varlink
389 interfaces (and thus also via glibc NSS if nss-resolve is used) will
390 now honour a trailing dot in the hostname: if specified the search
391 path logic is turned off. Thus "resolvectl query foo." is now
392 equivalent to "resolvectl query --search=off foo.".
394 * systemd-resolved gained a new D-Bus property "ResolvConfMode" that
395 exposes how /etc/resolv.conf is currently managed: by resolved (and
396 in which mode if so) or another subsystem. "resolvctl" will display
397 this property in its status output.
399 * The resolv.conf snippets systemd-resolved provides will now set "."
400 as the search domain if no other search domain is known. This turns
401 off the derivation of an implicit search domain by nss-dns for the
402 hostname, when the hostname is set to an FQDN. This change is done to
403 make nss-dns using resolv.conf provided by systemd-resolved behave
404 more similarly to nss-resolve.
406 * systemd-tmpfiles' file "aging" logic (i.e. the automatic clean-up of
407 /tmp/ and /var/tmp/ based on file timestamps) now looks at the
408 "birth" time (btime) of a file in addition to the atime, mtime, and
411 * systemd-analyze gained a new verb "capability" that lists all known
412 capabilities by the systemd build and by the kernel.
414 * If a file /usr/lib/clock-epoch exists, PID 1 will read its mtime and
415 advance the system clock to it at boot if it is noticed to be before
416 that time. Previously, PID 1 would only advance the time to an epoch
417 time that is set during build-time. With this new file OS builders
418 can change this epoch timestamp on individual OS images without
419 having to rebuild systemd.
421 * systemd-logind will now listen to the KEY_RESTART key from the Linux
422 input layer and reboot the system if it is pressed, similarly to how
423 it already handles KEY_POWER, KEY_SUSPEND or KEY_SLEEP. KEY_RESTART
424 was originally defined in the Multimedia context (to restart playback
425 of a song or film), but is now primarily used in various embedded
426 devices for "Reboot" buttons. Accordingly, systemd-logind will now
427 honour it as such. This may configured in more detail via the new
428 HandleRebootKey= and RebootKeyIgnoreInhibited=.
430 * systemd-nspawn/systemd-machined will now reconstruct hardlinks when
431 copying OS trees, for example in "systemd-nspawn --ephemeral",
432 "systemd-nspawn --template=", "machinectl clone" and similar. This is
433 useful when operating with OSTree images, which use hardlinks heavily
434 throughout, and where such copies previously resulting in "exploding"
437 * systemd-nspawn's --console= setting gained support for a new
438 "autopipe" value, which is identical to "interactive" when invoked on
439 a TTY, and "pipe" otherwise.
441 * systemd-networkd's .network files gained support for explicitly
442 configuring the multicast membership entries of bridge devices in the
443 [BridgeMDB] section. It also gained support for the PIE queuing
444 discipline in the [FlowQueuePIE] sections.
446 * systemd-networkd's .netdev files may now be used to create "BareUDP"
447 tunnels, configured in the new [BareUDP] setting. VXLAN tunnels may
448 now be marked to be independent of any underlying network interface
449 via the new Independent= boolean setting.
451 * systemd-networkd's Gateway= setting in .network files now accepts the
452 special values _dhcp4 and _ipv6ra to configure additional, locally
453 defined, explicit routes to the gateway acquired via DHCP or IPv6
454 Router Advertisements.
456 * systemd-networkd's [IPv6PrefixDelegation] section and
457 IPv6PrefixDelegation= options have been renamed as [IPv6SendRA] and
458 IPv6SendRA= (the old names are still accepted for backwards
461 * systemctl gained support for two new verbs: "service-log-level" and
462 "service-log-target" may be used on services that implement the
463 generic org.freedesktop.LogControl1 D-Bus interface to dynamically
464 adjust the log level and target. All of systemd's long-running
465 services support this now, but ideally all system services would
466 implement this interface to make the system more uniformly
469 * The SystemCallErrorNumber= unit file setting now accepts the new
470 "kill" and "log" actions, in addition to arbitrary error number
471 specifications as before. If "kill" the the processes are killed on
472 the event, if "log" the offending system call is audit logged.
474 * A new SystemCallLog= unit file setting has been added that accepts a
475 list of system calls that shall be logged about (audit).
477 * The OS image dissection logic (as used by RootImage= in unit files or
478 systemd-nspawn's --image= switch) has gained support for identifying
479 and mounting explicit /usr/ partitions, which are now defined in the
480 discoverable partition specification. This should be useful for
481 environments where the root file system is
482 generated/formatted/populated dynamically on first boot and combined
483 with an immutable /usr/ tree that is supplied by the vendor.
485 * In the final phase of shutdown, within the systemd-shutdown binary
486 we'll now try to detach MD devices (i.e software RAID) in addition to
487 loopback block devices and DM devices as before. This is supposed to
488 be a safety net only, in order to increase robustness if things go
489 wrong. Storage subsystems are expected to properly detach their
490 storage volumes during regular shutdown already (or in case of
491 storage backing the root file system: in the initrd hook we return to
494 * If the SYSTEMD_LOG_TID environment variable is set all systemd tools
495 will now log the thread ID in their log output. This is useful when
496 working with heavily threaded programs.
498 * If the SYSTEMD_RDRAND environment variable is set to "0", systemd will
499 not use the RDRAND CPU instruction. This is useful in environments
500 such as replay debuggers where non-deterministic behaviour is not
503 * The autopaging logic in systemd's various tools (such as systemctl)
504 has been updated to turn on "secure" mode in "less"
505 (i.e. $LESSECURE=1) if execution in a "sudo" environment is
506 detected. This disables invoking external programs from the pager,
507 via the pipe logic. This behaviour may be overridden via the new
508 $SYSTEMD_PAGERSECURE environment variable.
510 * Units which have resource limits (.service, .mount, .swap, .slice,
511 .socket, and .slice) gained new configuration settings
512 ManagedOOMSwap=, ManagedOOMMemoryPressure=, and
513 ManagedOOMMemoryPressureLimitPercent= that specify resource pressure
514 limits and optional action taken by systemd-oomd.
516 * A new service systemd-oomd has been added. It monitors resource
517 contention for selected parts of the unit hierarchy using the PSI
518 information reported by the kernel, and kills processes when memory
519 or swap pressure is above configured limits. This service is only
520 enabled in developer mode (see below) and should be considered a
521 preview in this release. Behaviour details and option names are
522 subject to change without the usual backwards-compatibility promises.
524 * A new helper oomctl has been added to introspect systemd-oomd state.
525 If also is only available in developer mode and should be considered
526 a preview without the usual backwards-compatibility promises.
528 * New meson option -Dcompat-mutable-uid-boundaries= has been added. If
529 enabled, systemd reads the system UID boundaries from /etc/login.defs
530 at runtime, instead of using the built-in values selected during
531 build. This is an option to improve compatibility for upgrades from
532 old systems. It's strongly recommended not to make use of this
533 functionality on new systems (or even enable it during build), as it
534 makes something runtime-configurable that is mostly an implementation
535 detail of the OS, and permits avoidable differences in deployments
536 that create all kinds of problems in the long run.
538 * New meson option '-Dmode=developer|release' has been added. When
539 'developer', additional checks and features are enabled that are
540 relevant during upstream development, e.g. verification that
541 semi-automatically-generated documentation has been properly updated
542 following API changes. Those checks are considered hints for
543 developers and are not actionable in downstream builds. In addition,
544 extra features that are not ready for general consumption may be
545 enabled in developer mode. It is thus recommended to set
546 '-Dmode=release' in end-user and distro builds.
548 * systemd-cryptsetup gained support for processing detached LUKS
549 headers specified on the kernel command line via the header=
550 parameter of the luks.options= kernel command line option. The same
551 device/path syntax as for key files is supported for header files
554 * The "net_id" built-in of udev has been updated to ignore ACPI _SUN
555 slot index data for devices that are connected through a PCI bridge
556 where the _SUN index is associated with the bridge instead of the
557 network device itself. Previously this would create ambiguous device
558 naming if multiple network interfaces were connected to the same PCI
559 bridge. Since this is a naming scheme incompatibility on systems that
560 possess hardware like this it has been introduced as new naming
561 scheme "v247". The previous scheme can be selected via the
562 "net.naming-scheme=v245" kernel command line parameter.
564 * ConditionFirstBoot= semantics have been modified to be safe towards
565 abnormal system power-off during first boot. Specifically, the
566 "systemd-machine-id-commit.service" service now acts as boot
567 milestone indicating when the first boot process is sufficiently
568 complete in order to not consider the next following boot also a
569 first boot. If the system is reset before this unit is reached the
570 first time, the next boot will still be considered a first boot; once
571 it has been reached, no further boots will be considered a first
572 boot. The "first-boot-complete.target" unit now acts as official hook
573 point to order against this. If a service shall be run on every boot
574 until the first boot fully succeeds it may thus be ordered before
575 this target unit (and pull it in) and carry ConditionFirstBoot=
578 * bootctl's set-default and set-oneshot commands now accept the three
579 special strings "@default", "@oneshot", "@current" in place of a boot
580 entry id. These strings are resolved to the current default and
581 oneshot boot loader entry, as well as the currently booted one. Thus
582 a command "bootctl set-default @current" may be used to make the
583 currently boot menu item the new default for all subsequent boots.
585 * A new generic target unit "initrd-cryptsetup.target" has been added
586 that is supposed to pull in all encrypted volumes that shall be set
587 up during the initrd phase. It takes the place of "cryptsetup.target"
588 and "remote-cryptsetup.target" that exist during the host boot
589 phase. In other words, the new "initrd-cryptsetup.target" is supposed
590 to take the role for "initrd-fs.target", but for encrypted volumes.
592 * "systemctl edit" has been updated to show the original effective unit
593 contents in commented form in the text editor.
595 Contributions from: Adolfo Jayme Barrientos, afg, Alec Moskvin, Alyssa
596 Ross, Amitanand.Chikorde, Andrew Hangsleben, Anita Zhang, Ansgar
597 Burchardt, Arian van Putten, Aurelien Jarno, Axel Rasmussen, bauen1,
598 Beniamino Galvani, Benjamin Berg, Bjørn Mork, brainrom, Chandradeep
599 Dey, Charles Lee, Chris Down, Christian Göttsche, Clemens Gruber, Daan
600 De Meyer, Daniele Medri, Daniel Mack, Dan Streetman, David Tardon,
601 Dimitri John Ledkov, Dmitry Borodaenko, Elias Probst, Elisei Roca,
602 ErrantSpore, Etienne Doms, Fabrice Fontaine, fangxiuning, Felix
603 Riemann, Florian Klink, Franck Bui, Frantisek Sumsal, George Rawlinson,
604 germanztz, Gibeom Gwon, Glen Whitney, Gogo Gogsi, Göran Uddeborg, Grant
605 Mathews, Hans de Goede, Hans Ulrich Niedermann, Haochen Tong, Harald
606 Seiler, huangyong, Hubert Kario, Ikey Doherty, Jan Chren, Jan Schlüter,
607 Jérémy Nouhaud, Joerg Behrmann, Jonathan Lebon, Juergen Hoetzel, Julien
608 Humbert, Kai-Chuan Hsieh, Kairui Song, Kamil Dudka, Kir Kolyshkin, Kyle
609 Huey, Kyle Russell, Lennart Poettering, lichangze, Luca Boccassi, Lucas
610 Werkmeister, Marc Kleine-Budde, Marco Wang, Marti Raudsepp,
611 masmullin2000, Máté Pozsgay, Matt Fenwick, Michael Biebl, Michael
612 Scherer, Michal Koutný, Michal Sekletár, Michal Suchanek, Mikael
613 Szreder, Milo Casagrande, mirabilos, Mitsuha_QuQ, mog422, Nazar
614 Vinnichuk, Nicholas Narsing, Nicolas Fella, Njibhu, nl6720, Oğuz Ersen,
615 Olivier Le Moal, Ondrej Kozina, Pass Automated Testing Suite, Pat
616 Coulthard, Pedro Ruiz, Peter Hutterer, Phaedrus Leeds, PhoenixDiscord,
617 Piotr Drąg, Plan C, Renaud Métrich, Robert Marko, Ronan Pigott, Roy
618 Chen (陳彥廷), RussianNeuroMancer, Samanta Navarro, Samuel BF,
619 scootergrisen, Steve Dodd, Susant Sahani, Tobias Hunger, Tobias
620 Kaufmann, Topi Miettinen, Vito Caputo, Weblate, Wen Yang, williamvds,
621 Yuri Chornoivan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Zmicer
622 Turok, Дамјан Георгиевски
628 * The service manager gained basic support for cgroup v2 freezer. Units
629 can now be suspended or resumed either using new systemctl verbs,
630 freeze and thaw respectively, or via D-Bus.
632 * PID 1 may now automatically load pre-compiled AppArmor policies from
633 /etc/apparmor/earlypolicy during early boot.
635 * The CPUAffinity= setting in service unit files now supports a new
636 special value "numa" that causes the CPU affinity masked to be set
637 based on the NUMA mask.
639 * systemd will now log about all left-over processes remaining in a
640 unit when the unit is stopped. It will now warn about services using
641 KillMode=none, as this is generally an unsafe thing to make use of.
643 * Two new unit file settings
644 ConditionPathIsEncrypted=/AssertPathIsEncrypted= have been
645 added. They may be used to check whether a specific file system path
646 resides on a block device that is encrypted on the block level
647 (i.e. using dm-crypt/LUKS).
649 * Another pair of new settings ConditionEnvironment=/AssertEnvironment=
650 has been added that may be used for simple environment checks. This
651 is particularly useful when passing in environment variables from a
652 container manager (or from PAM in case of the systemd --user
655 * .service unit files now accept a new setting CoredumpFilter= which
656 allows configuration of the memory sections coredumps of the
657 service's processes shall include.
659 * .mount units gained a new ReadWriteOnly= boolean option. If set
660 it will not be attempted to mount a file system read-only if mounting
661 in read-write mode doesn't succeed. An option x-systemd.rw-only is
662 available in /etc/fstab to control the same.
664 * .socket units gained a new boolean setting PassPacketInfo=. If
665 enabled, the kernel will attach additional per-packet metadata to all
666 packets read from the socket, as an ancillary message. This controls
667 the IP_PKTINFO, IPV6_RECVPKTINFO, NETLINK_PKTINFO socket options,
668 depending on socket type.
670 * .service units gained a new setting RootHash= which may be used to
671 specify the root hash for verity enabled disk images which are
672 specified in RootImage=. RootVerity= may be used to specify a path to
673 the Verity data matching a RootImage= file system. (The latter is
674 only useful for images that do not contain the Verity data embedded
675 into the same image that carries a GPT partition table following the
676 Discoverable Partition Specification). Similarly, systemd-nspawn
677 gained a new switch --verity-data= that takes a path to a file with
678 the verity data of the disk image supplied in --image=, if the image
679 doesn't contain the verity data itself.
681 * .service units gained a new setting RootHashSignature= which takes
682 either a base64 encoded PKCS#7 signature of the root hash specified
683 with RootHash=, or a path to a file to read the signature from. This
684 allows validation of the root hash against public keys available in
685 the kernel keyring, and is only supported on recent kernels
686 (>= 5.4)/libcryptsetup (>= 2.30). A similar switch has been added to
687 systemd-nspawn and systemd-dissect (--root-hash-sig=). Support for
688 this mechanism has also been added to systemd-veritysetup.
690 * .service unit files gained two new options
691 TimeoutStartFailureMode=/TimeoutStopFailureMode= that may be used to
692 tune behaviour if a start or stop timeout is hit, i.e. whether to
693 terminate the service with SIGTERM, SIGABRT or SIGKILL.
695 * Most options in systemd that accept hexadecimal values prefixed with
696 0x in additional to the usual decimal notation now also support octal
697 notation when the 0o prefix is used and binary notation if the 0b
700 * Various command line parameters and configuration file settings that
701 configure key or certificate files now optionally take paths to
702 AF_UNIX sockets in the file system. If configured that way a stream
703 connection is made to the socket and the required data read from
704 it. This is a simple and natural extension to the existing regular
705 file logic, and permits other software to provide keys or
706 certificates via simple IPC services, for example when unencrypted
707 storage on disk is not desired. Specifically, systemd-networkd's
708 Wireguard and MACSEC key file settings as well as
709 systemd-journal-gatewayd's and systemd-journal-remote's PEM
710 key/certificate parameters support this now.
712 * Unit files, tmpfiles.d/ snippets, sysusers.d/ snippets and other
713 configuration files that support specifier expansion learnt six new
714 specifiers: %a resolves to the current architecture, %o/%w/%B/%W
715 resolve to the various ID fields from /etc/os-release, %l resolves to
716 the "short" hostname of the system, i.e. the hostname configured in
717 the kernel truncated at the first dot.
719 * Support for the .include syntax in unit files has been removed. The
720 concept has been obsolete for 6 years and we started warning about
721 its pending removal 2 years ago (also see NEWS file below). It's
724 * StandardError= and StandardOutput= in unit files no longer support
725 the "syslog" and "syslog-console" switches. They were long removed
726 from the documentation, but will now result in warnings when used,
727 and be converted to "journal" and "journal+console" automatically.
729 * If the service setting User= is set to the "nobody" user, a warning
730 message is now written to the logs (but the value is nonetheless
731 accepted). Setting User=nobody is unsafe, since the primary purpose
732 of the "nobody" user is to own all files whose owner cannot be mapped
733 locally. It's in particular used by the NFS subsystem and in user
734 namespacing. By running a service under this user's UID it might get
735 read and even write access to all these otherwise unmappable files,
736 which is quite likely a major security problem.
738 * tmpfs mounts automatically created by systemd (/tmp, /run, /dev/shm,
739 and others) now have a size and inode limits applied (50% of RAM for
740 /tmp and /dev/shm, 10% of RAM for other mounts, etc.). Please note
741 that the implicit kernel default is 50% too, so there is no change
742 in the size limit for /tmp and /dev/shm.
744 * nss-mymachines lost support for resolution of users and groups, and
745 now only does resolution of hostnames. This functionality is now
746 provided by nss-systemd. Thus, the 'mymachines' entry should be
747 removed from the 'passwd:' and 'group:' lines in /etc/nsswitch.conf
748 (and 'systemd' added if it is not already there).
750 * A new kernel command line option systemd.hostname= has been added
751 that allows controlling the hostname that is initialized early during
754 * A kernel command line option "udev.blockdev_read_only" has been
755 added. If specified all hardware block devices that show up are
756 immediately marked as read-only by udev. This option is useful for
757 making sure that a specific boot under no circumstances modifies data
758 on disk. Use "blockdev --setrw" to undo the effect of this, per
761 * A new boolean kernel command line option systemd.swap= has been
762 added, which may be used to turn off automatic activation of swap
763 devices listed in /etc/fstab.
765 * New kernel command line options systemd.condition-needs-update= and
766 systemd.condition-first-boot= have been added, which override the
767 result of the ConditionNeedsUpdate= and ConditionFirstBoot=
770 * A new kernel command line option systemd.clock-usec= has been added
771 that allows setting the system clock to the specified time in µs
772 since Jan 1st, 1970 early during boot. This is in particular useful
773 in order to make test cases more reliable.
775 * The fs.suid_dumpable sysctl is set to 2 / "suidsafe". This allows
776 systemd-coredump to save core files for suid processes. When saving
777 the core file, systemd-coredump will use the effective uid and gid of
778 the process that faulted.
780 * The /sys/module/kernel/parameters/crash_kexec_post_notifiers file is
781 now automatically set to "Y" at boot, in order to enable pstore
782 generation for collection with systemd-pstore.
784 * We provide a set of udev rules to enable auto-suspend on PCI and USB
785 devices that were tested to correctly support it. Previously, this
786 was distributed as a set of udev rules, but has now been replaced by
787 by a set of hwdb entries (and a much shorter udev rule to take action
788 if the device modalias matches one of the new hwdb entries).
790 As before, entries are periodically imported from the database
791 maintained by the ChromiumOS project. If you have a device that
792 supports auto-suspend correctly and where it should be enabled by
793 default, please submit a patch that adds it to the database (see
794 /usr/lib/udev/hwdb.d/60-autosuspend.hwdb).
796 * systemd-udevd gained the new configuration option timeout_signal= as well
797 as a corresponding kernel command line option udev.timeout_signal=.
798 The option can be used to configure the UNIX signal that the main
799 daemon sends to the worker processes on timeout. Setting the signal
800 to SIGABRT is useful for debugging.
802 * .link files managed by systemd-udevd gained options RxFlowControl=,
803 TxFlowControl=, AutoNegotiationFlowControl= in the [Link] section, in
804 order to configure various flow control parameters. They also gained
805 RxMiniBufferSize= and RxJumboBufferSize= in order to configure jumbo
806 frame ring buffer sizes.
808 * networkd.conf gained a new boolean setting ManageForeignRoutes=. If
809 enabled systemd-networkd manages all routes configured by other tools.
811 * .network files managed by systemd-networkd gained a new section
812 [SR-IOV], in order to configure SR-IOV capable network devices.
814 * systemd-networkd's [IPv6Prefix] section in .network files gained a
815 new boolean setting Assign=. If enabled an address from the prefix is
816 automatically assigned to the interface.
818 * systemd-networkd gained a new section [DHCPv6PrefixDelegation] which
819 controls delegated prefixes assigned by DHCPv6 client. The section
820 has three settings: SubnetID=, Assign=, and Token=. The setting
821 SubnetID= allows explicit configuration of the preferred subnet that
822 systemd-networkd's Prefix Delegation logic assigns to interfaces. If
823 Assign= is enabled (which is the default) an address from any acquired
824 delegated prefix is automatically chosen and assigned to the
825 interface. The setting Token= specifies an optional address generation
828 * systemd-networkd's [Network] section gained a new setting
829 IPv4AcceptLocal=. If enabled the interface accepts packets with local
832 * systemd-networkd gained support for configuring the HTB queuing
833 discipline in the [HierarchyTokenBucket] and
834 [HierarchyTokenBucketClass] sections. Similar the "pfifo" qdisc may
835 be configured in the [PFIFO] section, "GRED" in
836 [GenericRandomEarlyDetection], "SFB" in [StochasticFairBlue], "cake"
837 in [CAKE], "PIE" in [PIE], "DRR" in [DeficitRoundRobinScheduler] and
838 [DeficitRoundRobinSchedulerClass], "BFIFO" in [BFIFO],
839 "PFIFOHeadDrop" in [PFIFOHeadDrop], "PFIFOFast" in [PFIFOFast], "HHF"
840 in [HeavyHitterFilter], "ETS" in [EnhancedTransmissionSelection] and
841 "QFQ" in [QuickFairQueueing] and [QuickFairQueueingClass].
843 * systemd-networkd gained support for a new Termination= setting in the
844 [CAN] section for configuring the termination resistor. It also
845 gained a new ListenOnly= setting for controlling whether to only
846 listen on CAN interfaces, without interfering with traffic otherwise
847 (which is useful for debugging/monitoring CAN network
848 traffic). DataBitRate=, DataSamplePoint=, FDMode=, FDNonISO= have
849 been added to configure various CAN-FD aspects.
851 * systemd-networkd's [DHCPv6] section gained a new option WithoutRA=.
852 When enabled, DHCPv6 will be attempted right-away without requiring an
853 Router Advertisement packet suggesting it first (i.e. without the 'M'
854 or 'O' flags set). The [IPv6AcceptRA] section gained a boolean option
855 DHCPv6Client= that may be used to turn off the DHCPv6 client even if
856 the RA packets suggest it.
858 * systemd-networkd's [DHCPv4] section gained a new setting UseGateway=
859 which may be used to turn off use of the gateway information provided
860 by the DHCP lease. A new FallbackLeaseLifetimeSec= setting may be
861 used to configure how to process leases that lack a lifetime option.
863 * systemd-networkd's [DHCPv4] and [DHCPServer] sections gained a new
864 setting SendVendorOption= allowing configuration of additional vendor
865 options to send in the DHCP requests/responses. The [DHCPv6] section
866 gained a new SendOption= setting for sending arbitrary DHCP
867 options. RequestOptions= has been added to request arbitrary options
868 from the server. UserClass= has been added to set the DHCP user class
871 * systemd-networkd's [DHCPServer] section gained a new set of options
872 EmitPOP3=/POP3=, EmitSMTP=/SMTP=, EmitLPR=/LPR= for including server
873 information about these three protocols in the DHCP lease. It also
874 gained support for including "MUD" URLs ("Manufacturer Usage
875 Description"). Support for "MUD" URLs was also added to the LLDP
876 stack, configurable in the [LLDP] section in .network files.
878 * The Mode= settings in [MACVLAN] and [MACVTAP] now support 'source'
879 mode. Also, the sections now support a new setting SourceMACAddress=.
881 * systemd-networkd's .netdev files now support a new setting
882 VLANProtocol= in the [Bridge] section that allows configuration of
883 the VLAN protocol to use.
885 * systemd-networkd supports a new Group= setting in the [Link] section
886 of the .network files, to control the link group.
888 * systemd-networkd's [Network] section gained a new
889 IPv6LinkLocalAddressGenerationMode= setting, which specifies how IPv6
890 link local address is generated.
892 * A new default .network file is now shipped that matches TUN/TAP
893 devices that begin with "vt-" in their name. Such interfaces will
894 have IP routing onto the host links set up automatically. This is
895 supposed to be used by VM managers to trivially acquire a network
896 interface which is fully set up for host communication, simply by
897 carefully picking an interface name to use.
899 * systemd-networkd's [DHCPv6] section gained a new setting RouteMetric=
900 which sets the route priority for routes specified by the DHCP server.
902 * systemd-networkd's [DHCPv6] section gained a new setting VendorClass=
903 which configures the vendor class information sent to DHCP server.
905 * The BlackList= settings in .network files' [DHCPv4] and
906 [IPv6AcceptRA] sections have been renamed DenyList=. The old names
907 are still understood to provide compatibility.
909 * networkctl gained the new "forcerenew" command for forcing all DHCP
910 server clients to renew their lease. The interface "status" output
911 will now show numerous additional fields of information about an
912 interface. There are new "up" and "down" commands to bring specific
913 interfaces up or down.
915 * systemd-resolved's DNS= configuration option now optionally accepts a
916 port number (after ":") and a host name (after "#"). When the host
917 name is specified, the DNS-over-TLS certificate is validated to match
918 the specified hostname. Additionally, in case of IPv6 addresses, an
919 interface may be specified (after "%").
921 * systemd-resolved may be configured to forward single-label DNS names.
922 This is not standard-conformant, but may make sense in setups where
923 public DNS servers are not used.
925 * systemd-resolved's DNS-over-TLS support gained SNI validation.
927 * systemd-nspawn's --resolv-conf= switch gained a number of new
928 supported values. Specifically, options starting with "replace-" are
929 like those prefixed "copy-" but replace any existing resolv.conf
930 file. And options ending in "-uplink" and "-stub" can now be used to
931 propagate other flavours of resolv.conf into the container (as
932 defined by systemd-resolved).
934 * The various programs included in systemd can now optionally output
935 their log messages on stderr prefixed with a timestamp, controlled by
936 the $SYSTEMD_LOG_TIME environment variable.
938 * systemctl gained a new "-P" switch that is a shortcut for "--value
941 * "systemctl list-units" and "systemctl list-machines" no longer hide
942 their first output column with --no-legend. To hide the first column,
945 * "systemctl reboot" takes the option "--reboot-argument=".
946 The optional positional argument to "systemctl reboot" is now
947 being deprecated in favor of this option.
949 * systemd-run gained a new switch --slice-inherit. If specified the
950 unit it generates is placed in the same slice as the systemd-run
953 * systemd-journald gained support for zstd compression of large fields
954 in journal files. The hash tables in journal files have been hardened
955 against hash collisions. This is an incompatible change and means
956 that journal files created with new systemd versions are not readable
957 with old versions. If the $SYSTEMD_JOURNAL_KEYED_HASH boolean
958 environment variable for systemd-journald.service is set to 0 this
959 new hardening functionality may be turned off, so that generated
960 journal files remain compatible with older journalctl
963 * journalctl will now include a clickable link in the default output for
964 each log message for which an URL with further documentation is
965 known. This is only supported on terminal emulators that support
966 clickable hyperlinks, and is turned off if a pager is used (since
967 "less" still doesn't support hyperlinks,
968 unfortunately). Documentation URLs may be included in log messages
969 either by including a DOCUMENTATION= journal field in it, or by
970 associating a journal message catalog entry with the log message's
971 MESSAGE_ID, which then carries a "Documentation:" tag.
973 * journald.conf gained a new boolean setting Audit= that may be used to
974 control whether systemd-journald will enable audit during
977 * when systemd-journald's log stream is broken up into multiple lines
978 because the PID of the sender changed this is indicated in the
979 generated log records via the _LINE_BREAK=pid-change field.
981 * journalctl's "-o cat" output mode will now show one or more journal
982 fields specified with --output-fields= instead of unconditionally
983 MESSAGE=. This is useful to retrieve a very specific set of fields
984 without any decoration.
986 * The sd-journal.h API gained two new functions:
987 sd_journal_enumerate_available_unique() and
988 sd_journal_enumerate_available_data() that operate like their
989 counterparts that lack the _available_ in the name, but skip items
990 that cannot be read and processed by the local implementation
991 (i.e. are compressed in an unsupported format or such),
993 * coredumpctl gained a new --file= switch, matching the same one in
994 journalctl: a specific journal file may be specified to read the
997 * coredumps collected by systemd-coredump may now be compressed using
1000 * systemd-binfmt gained a new switch --unregister for unregistering all
1001 registered entries at once. This is now invoked automatically at
1002 shutdown, so that binary formats registered with the "F" flag will
1003 not block clean file system unmounting.
1005 * systemd-notify's --pid= switch gained new values: "parent", "self",
1006 "auto" for controlling which PID to send to the service manager: the
1007 systemd-notify process' PID, or the one of the process invoking it.
1009 * systemd-logind's Session bus object learnt a new method call
1010 SetType() for temporarily updating the session type of an already
1011 allocated session. This is useful for upgrading tty sessions to
1012 graphical ones once a compositor is invoked.
1014 * systemd-socket-proxy gained a new switch --exit-idle-time= for
1015 configuring an exit-on-idle time.
1017 * systemd-repart's --empty= setting gained a new value "create". If
1018 specified a new empty regular disk image file is created under the
1019 specified name. Its size may be specified with the new --size=
1020 option. The latter is also supported without the "create" mode, in
1021 order to grow existing disk image files to the specified size. These
1022 two new options are useful when creating or manipulating disk images
1023 instead of operating on actual block devices.
1025 * systemd-repart drop-ins now support a new UUID= setting to control
1026 the UUID to assign to a newly created partition.
1028 * systemd-repart's SizeMin= per-partition parameter now defaults to 10M
1031 * systemd-repart's Label= setting now support the usual, simple
1032 specifier expansion.
1034 * systemd-homed's LUKS backend gained the ability to discard empty file
1035 system blocks automatically when the user logs out. This is enabled
1036 by default to ensure that home directories take minimal space when
1037 logged out but get full size guarantees when logged in. This may be
1038 controlled with the new --luks-offline-discard= switch to homectl.
1040 * If systemd-homed detects that /home/ is encrypted as a whole it will
1041 now default to the directory or subvolume backends instead of the
1042 LUKS backend, in order to avoid double encryption. The default
1043 storage and file system may now be configured explicitly, too, via
1044 the new /etc/systemd/homed.conf configuration file.
1046 * systemd-homed now supports unlocking home directories with FIDO2
1047 security tokens that support the 'hmac-secret' extension, in addition
1048 to the existing support for PKCS#11 security token unlocking
1049 support. Note that many recent hardware security tokens support both
1050 interfaces. The FIDO2 support is accessible via homectl's
1051 --fido2-device= option.
1053 * homectl's --pkcs11-uri= setting now accepts two special parameters:
1054 if "auto" is specified and only one suitable PKCS#11 security token
1055 is plugged in, its URL is automatically determined and enrolled for
1056 unlocking the home directory. If "list" is specified a brief table of
1057 suitable PKCS#11 security tokens is shown. Similar, the new
1058 --fido2-device= option also supports these two special values, for
1059 automatically selecting and listing suitable FIDO2 devices.
1061 * The /etc/crypttab tmp option now optionally takes an argument
1062 selecting the file system to use. Moreover, the default is now
1063 changed from ext2 to ext4.
1065 * There's a new /etc/crypttab option "keyfile-erase". If specified the
1066 key file listed in the same line is removed after use, regardless if
1067 volume activation was successful or not. This is useful if the key
1068 file is only acquired transiently at runtime and shall be erased
1069 before the system continues to boot.
1071 * There's also a new /etc/crypttab option "try-empty-password". If
1072 specified, before asking the user for a password it is attempted to
1073 unlock the volume with an empty password. This is useful for
1074 installing encrypted images whose password shall be set on first boot
1075 instead of at installation time.
1077 * systemd-cryptsetup will now attempt to load the keys to unlock
1078 volumes with automatically from files in
1079 /etc/cryptsetup-keys.d/<volume>.key and
1080 /run/cryptsetup-keys.d/<volume>.key, if any of these files exist.
1082 * systemd-cryptsetup may now activate Microsoft BitLocker volumes via
1083 /etc/crypttab, during boot.
1085 * logind.conf gained a new RuntimeDirectoryInodesMax= setting to
1086 control the inode limit for the per-user $XDG_RUNTIME_DIR tmpfs
1089 * A new generator systemd-xdg-autostart-generator has been added. It
1090 generates systemd unit files from XDG autostart .desktop files, and
1091 may be used to let the systemd user instance manage services that are
1092 started automatically as part of the desktop session.
1094 * "bootctl" gained a new verb "reboot-to-firmware" that may be used
1095 to query and change the firmware's 'reboot into firmware' setup flag.
1097 * systemd-firstboot gained a new switch --kernel-command-line= that may
1098 be used to initialize the /etc/kernel/cmdline file of the image. It
1099 also gained a new switch --root-password-hashed= which is like
1100 --root-password= but accepts a pre-hashed UNIX password as
1101 argument. The new option --delete-root-password may be used to unset
1102 any password for the root user (dangerous!). The --root-shell= switch
1103 may be used to control the shell to use for the root account. A new
1104 --force option may be used to override any already set settings with
1105 the parameters specified on the command line (by default, the tool
1106 will not override what has already been set before, i.e. is purely
1109 * systemd-firstboot gained support for a new --image= switch, which is
1110 similar to --root= but accepts the path to a disk image file, on
1111 which it then operates.
1113 * A new sd-path.h API has been added to libsystemd. It provides a
1114 simple API for retrieving various search paths and primary
1115 directories for various resources.
1117 * A new call sd_notify_barrier() has been added to the sd-daemon.h
1118 API. The call will block until all previously sent sd_notify()
1119 messages have been processed by the service manager. This is useful
1120 to remove races caused by a process already having disappeared at the
1121 time a notification message is processed by the service manager,
1122 making correct attribution impossible. The systemd-notify tool will
1123 now make use of this call implicitly, but this can be turned off again
1124 via the new --no-block switch.
1126 * When sending a file descriptor (fd) to the service manager to keep
1127 track of, using the sd_notify() mechanism, a new parameter FDPOLL=0
1128 may be specified. If passed the service manager will refrain from
1129 poll()ing on the file descriptor. Traditionally (and when the
1130 parameter is not specified), the service manager will poll it for
1131 POLLHUP or POLLERR events, and immediately close the fds in that
1134 * The service manager (PID1) gained a new D-Bus method call
1135 SetShowStatus() which may be used to control whether it shall show
1136 boot-time status output on the console. This method has a similar
1137 effect to sending SIGRTMIN+20/SIGRTMIN+21 to PID 1.
1139 * The sd-bus API gained a number of convenience functions that take
1140 va_list arguments rather than "...". For example, there's now
1141 sd_bus_call_methodv() to match sd_bus_call_method(). Those calls make
1142 it easier to build wrappers that accept variadic arguments and want
1143 to pass a ready va_list structure to sd-bus.
1145 * sd-bus vtable entries can have a new SD_BUS_VTABLE_ABSOLUTE_OFFSET
1146 flag which alters how the userdata pointer to pass to the callbacks
1147 is determined. When the flag is set, the offset field is converted
1148 as-is into a pointer, without adding it to the object pointer the
1149 vtable is associated with.
1151 * sd-bus now exposes four new functions:
1152 sd_bus_interface_name_is_valid() + sd_bus_service_name_is_valid() +
1153 sd_bus_member_name_is_valid() + sd_bus_object_path_is_valid() will
1154 validate strings to check if they qualify as various D-Bus concepts.
1156 * The sd-bus API gained the SD_BUS_METHOD_WITH_ARGS(),
1157 SD_BUS_METHOD_WITH_ARGS_OFFSET() and SD_BUS_SIGNAL_WITH_ARGS() macros
1158 that simplify adding argument names to D-Bus methods and signals.
1160 * The man pages for the sd-bus and sd-hwdb APIs have been completed.
1162 * Various D-Bus APIs of systemd daemons now have man pages that
1163 document the methods, signals and properties.
1165 * The expectations on user/group name syntax are now documented in
1166 detail; documentation on how classic home directories may be
1167 converted into home directories managed by homed has been added;
1168 documentation regarding integration of homed/userdb functionality in
1169 desktops has been added:
1171 https://systemd.io/USER_NAMES
1172 https://systemd.io/CONVERTING_TO_HOMED
1173 https://systemd.io/USERDB_AND_DESKTOPS
1175 * Documentation for the on-disk Journal file format has been updated
1176 and has now moved to:
1178 https://systemd.io/JOURNAL_FILE_FORMAT
1180 * The interface for containers (https://systemd.io/CONTAINER_INTERFACE)
1181 has been extended by a set of environment variables that expose
1182 select fields from the host's os-release file to the container
1183 payload. Similarly, host's os-release files can be mounted into the
1184 container underneath /run/host. Together, those mechanisms provide a
1185 standardized way to expose information about the host to the
1186 container payload. Both interfaces are implemented in systemd-nspawn.
1188 * All D-Bus services shipped in systemd now implement the generic
1189 LogControl1 D-Bus API which allows clients to change log level +
1190 target of the service during runtime.
1192 * Only relevant for developers: the mkosi.default symlink has been
1193 dropped from version control. Please create a symlink to one of the
1194 distribution-specific defaults in .mkosi/ based on your preference.
1196 Contributions from: 24bisquitz, Adam Nielsen, Alan Perry, Alexander
1197 Malafeev, Amitanand.Chikorde, Alin Popa, Alvin Šipraga, Amos Bird,
1198 Andreas Rammhold, AndreRH, Andrew Doran, Anita Zhang, Ankit Jain,
1199 antznin, Arnaud Ferraris, Arthur Moraes do Lago, Arusekk, Balaji
1200 Punnuru, Balint Reczey, Bastien Nocera, bemarek, Benjamin Berg,
1201 Benjamin Dahlhoff, Benjamin Robin, Chris Down, Chris Kerr, Christian
1202 Göttsche, Christian Hesse, Christian Oder, Ciprian Hacman, Clinton Roy,
1203 codicodi, Corey Hinshaw, Daan De Meyer, Dana Olson, Dan Callaghan,
1204 Daniel Fullmer, Daniel Rusek, Dan Streetman, Dave Reisner, David
1205 Edmundson, David Wood, Denis Pronin, Diego Escalante Urrelo, Dimitri
1206 John Ledkov, dolphrundgren, duguxy, Einsler Lee, Elisei Roca, Emmanuel
1207 Garette, Eric Anderson, Eric DeVolder, Evgeny Vereshchagin,
1208 ExtinctFire, fangxiuning, Ferran Pallarès Roca, Filipe Brandenburger,
1209 Filippo Falezza, Finn, Florian Klink, Florian Mayer, Franck Bui,
1210 Frantisek Sumsal, gaurav, Georg Müller, Gergely Polonkai, Giedrius
1211 Statkevičius, Gigadoc2, gogogogi, Gaurav Singh, gzjsgdsb, Hans de
1212 Goede, Haochen Tong, ianhi, ignapk, Jakov Smolic, James T. Lee, Jan
1213 Janssen, Jan Klötzke, Jan Palus, Jay Burger, Jeremy Cline, Jérémy
1214 Rosen, Jian-Hong Pan, Jiri Slaby, Joel Shapiro, Joerg Behrmann, Jörg
1215 Thalheim, Jouke Witteveen, Kai-Heng Feng, Kenny Levinsen, Kevin
1216 Kuehler, Kumar Kartikeya Dwivedi, layderv, laydervus, Lénaïc Huard,
1217 Lennart Poettering, Lidong Zhong, Luca Boccassi, Luca BRUNO, Lucas
1218 Werkmeister, Lukas Klingsbo, Lukáš Nykrýn, Łukasz Stelmach, Maciej
1219 S. Szmigiero, MadMcCrow, Marc-André Lureau, Marcel Holtmann, Marc
1220 Kleine-Budde, Martin Hundebøll, Matthew Leeds, Matt Ranostay, Maxim
1221 Fomin, MaxVerevkin, Michael Biebl, Michael Chapman, Michael Gubbels,
1222 Michael Marley, Michał Bartoszkiewicz, Michal Koutný, Michal Sekletár,
1223 Mike Gilbert, Mike Kazantsev, Mikhail Novosyolov, ml, Motiejus Jakštys,
1224 nabijaczleweli, nerdopolis, Niccolò Maggioni, Niklas Hambüchen, Norbert
1225 Lange, Paul Cercueil, pelzvieh, Peter Hutterer, Piero La Terza, Pieter
1226 Lexis, Piotr Drąg, Rafael Fontenelle, Richard Petri, Ronan Pigott, Ross
1227 Lagerwall, Rubens Figueiredo, satmandu, Sean-StarLabs, Sebastian
1228 Jennen, sterlinghughes, Surhud More, Susant Sahani, szb512, Thomas
1229 Haller, Tobias Hunger, Tom, Tomáš Pospíšek, Tomer Shechner, Tom Hughes,
1230 Topi Miettinen, Tudor Roman, Uwe Kleine-König, Valery0xff, Vito Caputo,
1231 Vladimir Panteleev, Vladyslav Tronko, Wen Yang, Yegor Vialov, Yigal
1232 Korman, Yi Gao, YmrDtnJu, Yuri Chornoivan, Yu Watanabe, Zbigniew
1233 Jędrzejewski-Szmek, Zhu Li, Дамјан Георгиевски, наб
1235 – Warsaw, 2020-07-30
1239 * A new tool "systemd-repart" has been added, that operates as an
1240 idempotent declarative repartitioner for GPT partition tables.
1241 Specifically, a set of partitions that must or may exist can be
1242 configured via drop-in files, and during every boot the partition
1243 table on disk is compared with these files, creating missing
1244 partitions or growing existing ones based on configurable relative
1245 and absolute size constraints. The tool is strictly incremental,
1246 i.e. does not delete, shrink or move partitions, but only adds and
1247 grows them. The primary use-case is OS images that ship in minimized
1248 form, that on first boot are grown to the size of the underlying
1249 block device or augmented with additional partitions. For example,
1250 the root partition could be extended to cover the whole disk, or a
1251 swap or /home partitions could be added on first boot. It can also be
1252 used for systems that use an A/B update scheme but ship images with
1253 just the A partition, with B added on first boot. The tool is
1254 primarily intended to be run in the initrd, shortly before
1255 transitioning into the host OS, but can also be run after the
1256 transition took place. It automatically discovers the disk backing
1257 the root file system, and should hence not require any additional
1258 configuration besides the partition definition drop-ins. If no
1259 configuration drop-ins are present, no action is taken.
1261 * A new component "userdb" has been added, along with a small daemon
1262 "systemd-userdbd.service" and a client tool "userdbctl". The framework
1263 allows defining rich user and group records in a JSON format,
1264 extending on the classic "struct passwd" and "struct group"
1265 structures. Various components in systemd have been updated to
1266 process records in this format, including systemd-logind and
1267 pam-systemd. The user records are intended to be extensible, and
1268 allow setting various resource management, security and runtime
1269 parameters that shall be applied to processes and sessions of the
1270 user as they log in. This facility is intended to allow associating
1271 such metadata directly with user/group records so that they can be
1272 produced, extended and consumed in unified form. We hope that
1273 eventually frameworks such as sssd will generate records this way, so
1274 that for the first time resource management and various other
1275 per-user settings can be configured in LDAP directories and then
1276 provided to systemd (specifically to systemd-logind and pam-system)
1277 to apply on login. For further details see:
1279 https://systemd.io/USER_RECORD
1280 https://systemd.io/GROUP_RECORD
1281 https://systemd.io/USER_GROUP_API
1283 * A small new service systemd-homed.service has been added, that may be
1284 used to securely manage home directories with built-in encryption.
1285 The complete user record data is unified with the home directory,
1286 thus making home directories naturally migratable. Its primary
1287 back-end is based on LUKS volumes, but fscrypt, plain directories,
1288 and other storage schemes are also supported. This solves a couple of
1289 problems we saw with traditional ways to manage home directories, in
1290 particular when it comes to encryption. For further discussion of
1291 this, see the video of Lennart's talk at AllSystemsGo! 2019:
1293 https://media.ccc.de/v/ASG2019-164-reinventing-home-directories
1295 For further details about the format and expectations on home
1296 directories this new daemon makes, see:
1298 https://systemd.io/HOME_DIRECTORY
1300 * systemd-journald is now multi-instantiable. In addition to the main
1301 instance systemd-journald.service there's now a template unit
1302 systemd-journald@.service, with each instance defining a new named
1303 log 'namespace' (whose name is specified via the instance part of the
1304 unit name). A new unit file setting LogNamespace= has been added,
1305 taking such a namespace name, that assigns services to the specified
1306 log namespaces. As each log namespace is serviced by its own
1307 independent journal daemon, this functionality may be used to improve
1308 performance and increase isolation of applications, at the price of
1309 losing global message ordering. Each instance of journald has a
1310 separate set of configuration files, with possibly different disk
1311 usage limitations and other settings.
1313 journalctl now takes a new option --namespace= to show logs from a
1314 specific log namespace. The sd-journal.h API gained
1315 sd_journal_open_namespace() for opening the log stream of a specific
1316 log namespace. systemd-journald also gained the ability to exit on
1317 idle, which is useful in the context of log namespaces, as this means
1318 log daemons for log namespaces can be activated automatically on
1319 demand and will stop automatically when no longer used, minimizing
1322 * When systemd-tmpfiles copies a file tree using the 'C' line type it
1323 will now label every copied file according to the SELinux database.
1325 * When systemd/PID 1 detects it is used in the initrd it will now boot
1326 into initrd.target rather than default.target by default. This should
1327 make it simpler to build initrds with systemd as for many cases the
1328 only difference between a host OS image and an initrd image now is
1329 the presence of the /etc/initrd-release file.
1331 * A new kernel command line option systemd.cpu_affinity= is now
1332 understood. It's equivalent to the CPUAffinity= option in
1333 /etc/systemd/system.conf and allows setting the CPU mask for PID 1
1334 itself and the default for all other processes.
1336 * When systemd/PID 1 is reloaded (with systemctl daemon-reload or
1337 equivalent), the SELinux database is now reloaded, ensuring that
1338 sockets and other file system objects are generated taking the new
1339 database into account.
1341 * systemd/PID 1 accepts a new "systemd.show-status=error" setting, and
1342 "quiet" has been changed to imply that instead of
1343 "systemd.show-status=auto". In this mode, only messages about errors
1344 and significant delays in boot are shown on the console.
1346 * The sd-event.h API gained native support for the new Linux "pidfd"
1347 concept. This permits watching processes using file descriptors
1348 instead of PID numbers, which fixes a number of races and makes
1349 process supervision more robust and efficient. All of systemd's
1350 components will now use pidfds if the kernel supports it for process
1351 watching, with the exception of PID 1 itself, unfortunately. We hope
1352 to move PID 1 to exclusively using pidfds too eventually, but this
1353 requires some more kernel work first. (Background: PID 1 watches
1354 processes using waitid() with the P_ALL flag, and that does not play
1355 together nicely with pidfds yet.)
1357 * Closely related to this, the sd-event.h API gained two new calls
1358 sd_event_source_send_child_signal() (for sending a signal to a
1359 watched process) and sd_event_source_get_child_process_own() (for
1360 marking a process so that it is killed automatically whenever the
1361 event source watching it is freed).
1363 * systemd-networkd gained support for configuring Token Bucket Filter
1364 (TBF) parameters in its qdisc configuration support. Similarly,
1365 support for Stochastic Fairness Queuing (SFQ), Controlled-Delay
1366 Active Queue Management (CoDel), and Fair Queue (FQ) has been added.
1368 * systemd-networkd gained support for Intermediate Functional Block
1369 (IFB) network devices.
1371 * systemd-networkd gained support for configuring multi-path IP routes,
1372 using the new MultiPathRoute= setting in the [Route] section.
1374 * systemd-networkd's DHCPv4 client has been updated to support a new
1375 SendDecline= option. If enabled, duplicate address detection is done
1376 after a DHCP offer is received from the server. If a conflict is
1377 detected, the address is declined. The DHCPv4 client also gained
1378 support for a new RouteMTUBytes= setting that allows to configure the
1379 MTU size to be used for routes generated from DHCPv4 leases.
1381 * The PrefixRoute= setting in systemd-networkd's [Address] section of
1382 .network files has been deprecated, and replaced by AddPrefixRoute=,
1383 with its sense inverted.
1385 * The Gateway= setting of [Route] sections of .network files gained
1386 support for a special new value "_dhcp". If set, the configured
1387 static route uses the gateway host configured via DHCP.
1389 * New User= and SuppressPrefixLength= settings have been implemented
1390 for the [RoutingPolicyRule] section of .network files to configure
1391 source routing based on UID ranges and prefix length, respectively.
1393 * sd-bus gained a new API call sd_bus_message_sensitive() that marks a
1394 D-Bus message object as "sensitive". Those objects are erased from
1395 memory when they are freed. This concept is intended to be used for
1396 messages that contain security sensitive data. A new flag
1397 SD_BUS_VTABLE_SENSITIVE has been introduced as well to mark methods
1398 in sd-bus vtables, causing any incoming and outgoing messages of
1399 those methods to be implicitly marked as "sensitive".
1401 * sd-bus gained a new API call sd_bus_message_dump() for dumping the
1402 contents of a message (or parts thereof) to standard output for
1405 * systemd-sysusers gained support for creating users with the primary
1406 group named differently than the user.
1408 * systemd-growfs (i.e. the x-systemd.growfs mount option in /etc/fstab)
1409 gained support for growing XFS partitions. Previously it supported
1410 only ext4 and btrfs partitions.
1412 * The support for /etc/crypttab gained a new x-initrd.attach option. If
1413 set, the specified encrypted volume is unlocked already in the
1414 initrd. This concept corresponds to the x-initrd.mount option in
1417 * systemd-cryptsetup gained native support for unlocking encrypted
1418 volumes utilizing PKCS#11 smartcards, i.e. for example to bind
1419 encryption of volumes to YubiKeys. This is exposed in the new
1420 pkcs11-uri= option in /etc/crypttab.
1422 * The /etc/fstab support in systemd now supports two new mount options
1423 x-systemd.{required,wanted}-by=, for explicitly configuring the units
1424 that the specified mount shall be pulled in by, in place of
1425 the usual local-fs.target/remote-fs.target.
1427 * The https://systemd.io/ web site has been relaunched, directly
1428 populated with most of the documentation included in the systemd
1429 repository. systemd also acquired a new logo, thanks to Tobias
1432 * systemd-udevd gained support for managing "alternative" network
1433 interface names, as supported by new Linux kernels. For the first
1434 time this permits assigning multiple (and longer!) names to a network
1435 interface. systemd-udevd will now by default assign the names
1436 generated via all supported naming schemes to each interface. This
1437 may be further tweaked with .link files and the AlternativeName= and
1438 AlternativeNamesPolicy= settings. Other components of systemd have
1439 been updated to support the new alternative names wherever
1440 appropriate. For example, systemd-nspawn will now generate
1441 alternative interface names for the host-facing side of container
1442 veth links based on the full container name without truncation.
1444 * systemd-nspawn interface naming logic has been updated in another way
1445 too: if the main interface name (i.e. as opposed to new-style
1446 "alternative" names) based on the container name is truncated, a
1447 simple hashing scheme is used to give different interface names to
1448 multiple containers whose names all begin with the same prefix. Since
1449 this changes the primary interface names pointing to containers if
1450 truncation happens, the old scheme may still be requested by
1451 selecting an older naming scheme, via the net.naming-scheme= kernel
1452 command line option.
1454 * PrivateUsers= in service files now works in services run by the
1455 systemd --user per-user instance of the service manager.
1457 * A new per-service sandboxing option ProtectClock= has been added that
1458 locks down write access to the system clock. It takes away device
1459 node access to /dev/rtc as well as the system calls that set the
1460 system clock and the CAP_SYS_TIME and CAP_WAKE_ALARM capabilities.
1461 Note that this option does not affect access to auxiliary services
1462 that allow changing the clock, for example access to
1465 * The systemd-id128 tool gained a new "show" verb for listing or
1466 resolving a number of well-known UUIDs/128bit IDs, currently mostly
1467 GPT partition table types.
1469 * The Discoverable Partitions Specification has been updated to support
1470 /var and /var/tmp partition discovery. Support for this has been
1471 added to systemd-gpt-auto-generator. For details see:
1473 https://systemd.io/DISCOVERABLE_PARTITIONS
1475 * "systemctl list-unit-files" has been updated to show a new column
1476 with the suggested enablement state based on the vendor preset files
1477 for the respective units.
1479 * "systemctl" gained a new option "--with-dependencies". If specified
1480 commands such as "systemctl status" or "systemctl cat" will now show
1481 all specified units along with all units they depend on.
1483 * networkctl gained support for showing per-interface logs in its
1486 * systemd-networkd-wait-online gained support for specifying the maximum
1487 operational state to wait for, and to wait for interfaces to
1490 * The [Match] section of .link and .network files now supports a new
1491 option PermanentMACAddress= which may be used to check against the
1492 permanent MAC address of a network device even if a randomized MAC
1495 * The [TrafficControlQueueingDiscipline] section in .network files has
1496 been renamed to [NetworkEmulator] with the "NetworkEmulator" prefix
1497 dropped from the individual setting names.
1499 * Any .link and .network files that have an empty [Match] section (this
1500 also includes empty and commented-out files) will now be
1501 rejected. systemd-udev and systemd-networkd started warning about
1502 such files in version 243.
1504 * systemd-logind will now validate access to the operation of changing
1505 the virtual terminal via a polkit action. By default, only users
1506 with at least one session on a local VT are granted permission.
1508 * When systemd sets up PAM sessions that invoked service processes
1509 shall run in, the pam_setcred() API is now invoked, thus permitting
1510 PAM modules to set additional credentials for the processes.
1512 * portablectl attach/detach verbs now accept --now and --enable options
1513 to combine attachment with enablement and invocation, or detachment
1514 with stopping and disablement.
1516 * UPGRADE ISSUE: a bug where some jobs were trimmed as redundant was
1517 fixed, which in turn exposed bugs in unit configuration of services
1518 which have Type=oneshot and should only run once, but do not have
1519 RemainAfterExit=yes set. Without RemainAfterExit=yes, a one-shot
1520 service may be started again after exiting successfully, for example
1521 as a dependency in another transaction. Affected services included
1522 some internal systemd services (most notably
1523 systemd-vconsole-setup.service, which was updated to have
1524 RemainAfterExit=yes), and plymouth-start.service. Please ensure that
1525 plymouth has been suitably updated or patched before upgrading to
1526 this systemd release. See
1527 https://bugzilla.redhat.com/show_bug.cgi?id=1807771 for some
1528 additional discussion.
1530 Contributions from: AJ Bagwell, Alin Popa, Andreas Rammhold, Anita
1531 Zhang, Ansgar Burchardt, Antonio Russo, Arian van Putten, Ashley Davis,
1532 Balint Reczey, Bart Willems, Bastien Nocera, Benjamin Dahlhoff, Charles
1533 (Chas) Williams, cheese1, Chris Down, Chris Murphy, Christian Ehrhardt,
1534 Christian Göttsche, cvoinf, Daan De Meyer, Daniele Medri, Daniel Rusek,
1535 Daniel Shahaf, Dann Frazier, Dan Streetman, Dariusz Gadomski, David
1536 Michael, Dimitri John Ledkov, Emmanuel Bourg, Evgeny Vereshchagin,
1537 ezst036, Felipe Sateler, Filipe Brandenburger, Florian Klink, Franck
1538 Bui, Fran Dieguez, Frantisek Sumsal, Greg "GothAck" Miell, Guilhem
1539 Lettron, Guillaume Douézan-Grard, Hans de Goede, HATAYAMA Daisuke, Iain
1540 Lane, James Buren, Jan Alexander Steffens (heftig), Jérémy Rosen, Jin
1541 Park, Jun'ichi Nomura, Kai Krakow, Kevin Kuehler, Kevin P. Fleming,
1542 Lennart Poettering, Leonid Bloch, Leonid Evdokimov, lothrond, Luca
1543 Boccassi, Lukas K, Lynn Kirby, Mario Limonciello, Mark Deneen, Matthew
1544 Leeds, Michael Biebl, Michal Koutný, Michal Sekletár, Mike Auty, Mike
1545 Gilbert, mtron, nabijaczleweli, Naïm Favier, Nate Jones, Norbert Lange,
1546 Oliver Giles, Paul Davey, Paul Menzel, Peter Hutterer, Piotr Drąg, Rafa
1547 Couto, Raphael, rhn, Robert Scheck, Rocka, Romain Naour, Ryan Attard,
1548 Sascha Dewald, Shengjing Zhu, Slava Kardakov, Spencer Michaels, Sylvain
1549 Plantefeve, Stanislav Angelovič, Susant Sahani, Thomas Haller, Thomas
1550 Schmitt, Timo Schlüßler, Timo Wilken, Tobias Bernard, Tobias Klauser,
1551 Tobias Stoeckmann, Topi Miettinen, tsia, WataruMatsuoka, Wieland
1552 Hoffmann, Wilhelm Schuster, Will Fleming, xduugu, Yong Cong Sin, Yuri
1553 Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zeyu
1556 – Warsaw, 2020-03-06
1560 * Support for the cpuset cgroups v2 controller has been added.
1561 Processes may be restricted to specific CPUs using the new
1562 AllowedCPUs= setting, and to specific memory NUMA nodes using the new
1563 AllowedMemoryNodes= setting.
1565 * The signal used in restart jobs (as opposed to e.g. stop jobs) may
1566 now be configured using a new RestartKillSignal= setting. This
1567 allows units which signals to request termination to implement
1568 different behaviour when stopping in preparation for a restart.
1570 * "systemctl clean" may now be used also for socket, mount, and swap
1573 * systemd will also read configuration options from the EFI variable
1574 SystemdOptions. This may be used to configure systemd behaviour when
1575 modifying the kernel command line is inconvenient, but configuration
1576 on disk is read too late, for example for the options related to
1577 cgroup hierarchy setup. 'bootctl systemd-efi-options' may be used to
1578 set the EFI variable.
1580 * systemd will now disable printk ratelimits in early boot. This should
1581 allow us to capture more logs from the early boot phase where normal
1582 storage is not available and the kernel ring buffer is used for
1583 logging. Configuration on the kernel command line has higher priority
1584 and overrides the systemd setting.
1586 systemd programs which log to /dev/kmsg directly use internal
1587 ratelimits to prevent runaway logging. (Normally this is only used
1588 during early boot, so in practice this change has very little
1591 * Unit files now support top level dropin directories of the form
1592 <unit_type>.d/ (e.g. service.d/) that may be used to add configuration
1593 that affects all corresponding unit files.
1595 * systemctl gained support for 'stop --job-mode=triggering' which will
1596 stop the specified unit and any units which could trigger it.
1598 * Unit status display now includes units triggering and triggered by
1599 the unit being shown.
1601 * The RuntimeMaxSec= setting is now supported by scopes, not just
1602 .service units. This is particularly useful for PAM sessions which
1603 create a scope unit for the user login. systemd.runtime_max_sec=
1604 setting may used with the pam_systemd module to limit the duration
1605 of the PAM session, for example for time-limited logins.
1607 * A new @pkey system call group is now defined to make it easier to
1608 allow-list memory protection syscalls for containers and services
1609 which need to use them.
1611 * systemd-udevd: removed the 30s timeout for killing stale workers on
1612 exit. systemd-udevd now waits for workers to finish. The hard-coded
1613 exit timeout of 30s was too short for some large installations, where
1614 driver initialization could be prematurely interrupted during initrd
1615 processing if the root file system had been mounted and init was
1616 preparing to switch root. If udevd is run without systemd and workers
1617 are hanging while udevd receives an exit signal, udevd will now exit
1618 when udev.event_timeout is reached for the last hanging worker. With
1619 systemd, the exit timeout can additionally be configured using
1620 TimeoutStopSec= in systemd-udevd.service.
1622 * udev now provides a program (fido_id) that identifies FIDO CTAP1
1623 ("U2F")/CTAP2 security tokens based on the usage declared in their
1624 report and descriptor and outputs suitable environment variables.
1625 This replaces the externally maintained allow lists of all known
1626 security tokens that were used previously.
1628 * Automatically generated autosuspend udev rules for allow-listed
1629 devices have been imported from the Chromium OS project. This should
1630 improve power saving with many more devices.
1632 * udev gained a new "CONST{key}=value" setting that allows matching
1633 against system-wide constants without forking a helper binary.
1634 Currently "arch" and "virt" keys are supported.
1636 * udev now opens CDROMs in non-exclusive mode when querying their
1637 capabilities. This should fix issues where other programs trying to
1638 use the CDROM cannot gain access to it, but carries a risk of
1639 interfering with programs writing to the disk, if they did not open
1640 the device in exclusive mode as they should.
1642 * systemd-networkd does not create a default route for IPv4 link local
1643 addressing anymore. The creation of the route was unexpected and was
1644 breaking routing in various cases, but people who rely on it being
1645 created implicitly will need to adjust. Such a route may be requested
1646 with DefaultRouteOnDevice=yes.
1648 Similarly, systemd-networkd will not assign a link-local IPv6 address
1649 when IPv6 link-local routing is not enabled.
1651 * Receive and transmit buffers may now be configured on links with
1652 the new RxBufferSize= and TxBufferSize= settings.
1654 * systemd-networkd may now advertise additional IPv6 routes. A new
1655 [IPv6RoutePrefix] section with Route= and LifetimeSec= options is
1658 * systemd-networkd may now configure "next hop" routes using the
1659 [NextHop] section and Gateway= and Id= settings.
1661 * systemd-networkd will now retain DHCP config on restarts by default
1662 (but this may be overridden using the KeepConfiguration= setting).
1663 The default for SendRelease= has been changed to true.
1665 * The DHCPv4 client now uses the OPTION_INFORMATION_REFRESH_TIME option
1666 received from the server.
1668 The client will use the received SIP server list if UseSIP=yes is
1671 The client may be configured to request specific options from the
1672 server using a new RequestOptions= setting.
1674 The client may be configured to send arbitrary options to the server
1675 using a new SendOption= setting.
1677 A new IPServiceType= setting has been added to configure the "IP
1678 service type" value used by the client.
1680 * The DHCPv6 client learnt a new PrefixDelegationHint= option to
1681 request prefix hints in the DHCPv6 solicitation.
1683 * The DHCPv4 server may be configured to send arbitrary options using
1684 a new SendOption= setting.
1686 * The DHCPv4 server may now be configured to emit SIP server list using
1687 the new EmitSIP= and SIP= settings.
1689 * systemd-networkd and networkctl may now renew DHCP leases on demand.
1690 networkctl has a new 'networkctl renew' verb.
1692 * systemd-networkd may now reconfigure links on demand. networkctl
1693 gained two new verbs: "reload" will reload the configuration, and
1694 "reconfigure DEVICE…" will reconfigure one or more devices.
1696 * .network files may now match on SSID and BSSID of a wireless network,
1697 i.e. the access point name and hardware address using the new SSID=
1698 and BSSID= options. networkctl will display the current SSID and
1699 BSSID for wireless links.
1701 .network files may also match on the wireless network type using the
1702 new WLANInterfaceType= option.
1704 * systemd-networkd now includes default configuration that enables
1705 link-local addressing when connected to an ad-hoc wireless network.
1707 * systemd-networkd may configure the Traffic Control queueing
1708 disciplines in the kernel using the new
1709 [TrafficControlQueueingDiscipline] section and Parent=,
1710 NetworkEmulatorDelaySec=, NetworkEmulatorDelayJitterSec=,
1711 NetworkEmulatorPacketLimit=, NetworkEmulatorLossRate=,
1712 NetworkEmulatorDuplicateRate= settings.
1714 * systemd-tmpfiles gained a new w+ setting to append to files.
1716 * systemd-analyze dump will now report when the memory configuration in
1717 the kernel does not match what systemd has configured (usually,
1718 because some external program has modified the kernel configuration
1721 * systemd-analyze gained a new --base-time= switch instructs the
1722 'calendar' verb to resolve times relative to that timestamp instead
1723 of the present time.
1725 * journalctl --update-catalog now produces deterministic output (making
1726 reproducible image builds easier).
1728 * A new devicetree-overlay setting is now documented in the Boot Loader
1731 * The default value of the WatchdogSec= setting used in systemd
1732 services (the ones bundled with the project itself) may be set at
1733 configuration time using the -Dservice-watchdog= setting. If set to
1734 empty, the watchdogs will be disabled.
1736 * systemd-resolved validates IP addresses in certificates now when GnuTLS
1739 * libcryptsetup >= 2.0.1 is now required.
1741 * A configuration option -Duser-path= may be used to override the $PATH
1742 used by the user service manager. The default is again to use the same
1743 path as the system manager.
1745 * The systemd-id128 tool gained a new switch "-u" (or "--uuid") for
1746 outputting the 128bit IDs in UUID format (i.e. in the "canonical
1749 * Service units gained a new sandboxing option ProtectKernelLogs= which
1750 makes sure the program cannot get direct access to the kernel log
1751 buffer anymore, i.e. the syslog() system call (not to be confused
1752 with the API of the same name in libc, which is not affected), the
1753 /proc/kmsg and /dev/kmsg nodes and the CAP_SYSLOG capability are made
1754 inaccessible to the service. It's recommended to enable this setting
1755 for all services that should not be able to read from or write to the
1756 kernel log buffer, which are probably almost all.
1758 Contributions from: Aaron Plattner, Alcaro, Anita Zhang, Balint Reczey,
1759 Bastien Nocera, Baybal Ni, Benjamin Bouvier, Benjamin Gilbert, Carlo
1760 Teubner, cbzxt, Chen Qi, Chris Down, Christian Rebischke, Claudio
1761 Zumbo, ClydeByrdIII, crashfistfight, Cyprien Laplace, Daniel Edgecumbe,
1762 Daniel Gorbea, Daniel Rusek, Daniel Stuart, Dan Streetman, David
1763 Pedersen, David Tardon, Dimitri John Ledkov, Dominique Martinet, Donald
1764 A. Cupp Jr, Evgeny Vereshchagin, Fabian Henneke, Filipe Brandenburger,
1765 Franck Bui, Frantisek Sumsal, Georg Müller, Hans de Goede, Haochen
1766 Tong, HATAYAMA Daisuke, Iwan Timmer, Jan Janssen, Jan Kundrát, Jan
1767 Synacek, Jan Tojnar, Jay Strict, Jérémy Rosen, Jóhann B. Guðmundsson,
1768 Jonas Jelten, Jonas Thelemann, Justin Trudell, J. Xing, Kai-Heng Feng,
1769 Kenneth D'souza, Kevin Becker, Kevin Kuehler, Lennart Poettering,
1770 Léonard Gérard, Lorenz Bauer, Luca Boccassi, Maciej Stanczew, Mario
1771 Limonciello, Marko Myllynen, Mark Stosberg, Martin Wilck, matthiasroos,
1772 Michael Biebl, Michael Olbrich, Michael Tretter, Michal Sekletar,
1773 Michal Sekletár, Michal Suchanek, Mike Gilbert, Mike Kazantsev, Nicolas
1774 Douma, nikolas, Norbert Lange, pan93412, Pascal de Bruijn, Paul Menzel,
1775 Pavel Hrdina, Peter Wu, Philip Withnall, Piotr Drąg, Rafael Fontenelle,
1776 Renaud Métrich, Riccardo Schirone, RoadrunnerWMC, Ronan Pigott, Ryan
1777 Attard, Sebastian Wick, Serge, Siddharth Chandrasekara, Steve Ramage,
1778 Steve Traylen, Susant Sahani, Thibault Nélis, Tim Teichmann, Tom
1779 Fitzhenry, Tommy J, Torsten Hilbrich, Vito Caputo, ypf791, Yu Watanabe,
1780 Zach Smith, Zbigniew Jędrzejewski-Szmek
1782 – Warsaw, 2019-11-29
1786 * This release enables unprivileged programs (i.e. requiring neither
1787 setuid nor file capabilities) to send ICMP Echo (i.e. ping) requests
1788 by turning on the "net.ipv4.ping_group_range" sysctl of the Linux
1789 kernel for the whole UNIX group range, i.e. all processes. This
1790 change should be reasonably safe, as the kernel support for it was
1791 specifically implemented to allow safe access to ICMP Echo for
1792 processes lacking any privileges. If this is not desirable, it can be
1793 disabled again by setting the parameter to "1 0".
1795 * Previously, filters defined with SystemCallFilter= would have the
1796 effect that any calling of an offending system call would terminate
1797 the calling thread. This behaviour never made much sense, since
1798 killing individual threads of unsuspecting processes is likely to
1799 create more problems than it solves. With this release the default
1800 action changed from killing the thread to killing the whole
1801 process. For this to work correctly both a kernel version (>= 4.14)
1802 and a libseccomp version (>= 2.4.0) supporting this new seccomp
1803 action is required. If an older kernel or libseccomp is used the old
1804 behaviour continues to be used. This change does not affect any
1805 services that have no system call filters defined, or that use
1806 SystemCallErrorNumber= (and thus see EPERM or another error instead
1807 of being killed when calling an offending system call). Note that
1808 systemd documentation always claimed that the whole process is
1809 killed. With this change behaviour is thus adjusted to match the
1812 * On 64 bit systems, the "kernel.pid_max" sysctl is now bumped to
1813 4194304 by default, i.e. the full 22bit range the kernel allows, up
1814 from the old 16bit range. This should improve security and
1815 robustness, as PID collisions are made less likely (though certainly
1816 still possible). There are rumours this might create compatibility
1817 problems, though at this moment no practical ones are known to
1818 us. Downstream distributions are hence advised to undo this change in
1819 their builds if they are concerned about maximum compatibility, but
1820 for everybody else we recommend leaving the value bumped. Besides
1821 improving security and robustness this should also simplify things as
1822 the maximum number of allowed concurrent tasks was previously bounded
1823 by both "kernel.pid_max" and "kernel.threads-max" and now effectively
1824 only a single knob is left ("kernel.threads-max"). There have been
1825 concerns that usability is affected by this change because larger PID
1826 numbers are harder to type, but we believe the change from 5 digits
1827 to 7 digits doesn't hamper usability.
1829 * MemoryLow= and MemoryMin= gained hierarchy-aware counterparts,
1830 DefaultMemoryLow= and DefaultMemoryMin=, which can be used to
1831 hierarchically set default memory protection values for a particular
1832 subtree of the unit hierarchy.
1834 * Memory protection directives can now take a value of zero, allowing
1835 explicit opting out of a default value propagated by an ancestor.
1837 * systemd now defaults to the "unified" cgroup hierarchy setup during
1838 build-time, i.e. -Ddefault-hierarchy=unified is now the build-time
1839 default. Previously, -Ddefault-hierarchy=hybrid was the default. This
1840 change reflects the fact that cgroupsv2 support has matured
1841 substantially in both systemd and in the kernel, and is clearly the
1842 way forward. Downstream production distributions might want to
1843 continue to use -Ddefault-hierarchy=hybrid (or even =legacy) for
1844 their builds as unfortunately the popular container managers have not
1845 caught up with the kernel API changes.
1847 * Man pages are not built by default anymore (html pages were already
1848 disabled by default), to make development builds quicker. When
1849 building systemd for a full installation with documentation, meson
1850 should be called with -Dman=true and/or -Dhtml=true as appropriate.
1851 The default was changed based on the assumption that quick one-off or
1852 repeated development builds are much more common than full optimized
1853 builds for installation, and people need to pass various other
1854 options to when doing "proper" builds anyway, so the gain from making
1855 development builds quicker is bigger than the one time disruption for
1858 Two scripts are created in the *build* directory to generate and
1859 preview man and html pages on demand, e.g.:
1861 build/man/man systemctl
1862 build/man/html systemd.index
1864 * libidn2 is used by default if both libidn2 and libidn are installed.
1865 Please use -Dlibidn=true if libidn is preferred.
1867 * The D-Bus "wire format" of the CPUAffinity= attribute is changed on
1868 big-endian machines. Before, bytes were written and read in native
1869 machine order as exposed by the native libc __cpu_mask interface.
1870 Now, little-endian order is always used (CPUs 0–7 are described by
1871 bits 0–7 in byte 0, CPUs 8–15 are described by byte 1, and so on).
1872 This change fixes D-Bus calls that cross endianness boundary.
1874 The presentation format used for CPUAffinity= by "systemctl show" and
1875 "systemd-analyze dump" is changed to present CPU indices instead of
1876 the raw __cpu_mask bitmask. For example, CPUAffinity=0-1 would be
1877 shown as CPUAffinity=03000000000000000000000000000… (on
1878 little-endian) or CPUAffinity=00000000000000300000000000000… (on
1879 64-bit big-endian), and is now shown as CPUAffinity=0-1, matching the
1880 input format. The maximum integer that will be printed in the new
1881 format is 8191 (four digits), while the old format always used a very
1882 long number (with the length varying by architecture), so they can be
1883 unambiguously distinguished.
1885 * /usr/sbin/halt.local is no longer supported. Implementation in
1886 distributions was inconsistent and it seems this functionality was
1889 To replace this functionality, users should:
1890 - either define a new unit and make it a dependency of final.target
1891 (systemctl add-wants final.target my-halt-local.service)
1892 - or move the shutdown script to /usr/lib/systemd/system-shutdown/
1893 and ensure that it accepts "halt", "poweroff", "reboot", and
1894 "kexec" as an argument, see the description in systemd-shutdown(8).
1896 * When a [Match] section in .link or .network file is empty (contains
1897 no match patterns), a warning will be emitted. Please add any "match
1898 all" pattern instead, e.g. OriginalName=* or Name=* in case all
1899 interfaces should really be matched.
1901 * A new setting NUMAPolicy= may be used to set process memory
1902 allocation policy. This setting can be specified in
1903 /etc/systemd/system.conf and hence will set the default policy for
1904 PID1. The default policy can be overridden on a per-service
1905 basis. The related setting NUMAMask= is used to specify NUMA node
1906 mask that should be associated with the selected policy.
1908 * PID 1 will now listen to Out-Of-Memory (OOM) events the kernel
1909 generates when processes it manages are reaching their memory limits,
1910 and will place their units in a special state, and optionally kill or
1911 stop the whole unit.
1913 * The service manager will now expose bus properties for the IO
1914 resources used by units. This information is also shown in "systemctl
1915 status" now (for services that have IOAccounting=yes set). Moreover,
1916 the IO accounting data is included in the resource log message
1917 generated whenever a unit stops.
1919 * Units may now configure an explicit timeout to wait for when killed
1920 with SIGABRT, for example when a service watchdog is hit. Previously,
1921 the regular TimeoutStopSec= timeout was applied in this case too —
1922 now a separate timeout may be set using TimeoutAbortSec=.
1924 * Services may now send a special WATCHDOG=trigger message with
1925 sd_notify() to trigger an immediate "watchdog missed" event, and thus
1926 trigger service termination. This is useful both for testing watchdog
1927 handling, but also for defining error paths in services, that shall
1928 be handled the same way as watchdog events.
1930 * There are two new per-unit settings IPIngressFilterPath= and
1931 IPEgressFilterPath= which allow configuration of a BPF program
1932 (usually by specifying a path to a program uploaded to /sys/fs/bpf/)
1933 to apply to the IP packet ingress/egress path of all processes of a
1934 unit. This is useful to allow running systemd services with BPF
1935 programs set up externally.
1937 * systemctl gained a new "clean" verb for removing the state, cache,
1938 runtime or logs directories of a service while it is terminated. The
1939 new verb may also be used to remove the state maintained on disk for
1940 timer units that have Persistent= configured.
1942 * During the last phase of shutdown systemd will now automatically
1943 increase the log level configured in the "kernel.printk" sysctl so
1944 that any relevant loggable events happening during late shutdown are
1945 made visible. Previously, loggable events happening so late during
1946 shutdown were generally lost if the "kernel.printk" sysctl was set to
1947 high thresholds, as regular logging daemons are terminated at that
1948 time and thus nothing is written to disk.
1950 * If processes terminated during the last phase of shutdown do not exit
1951 quickly systemd will now show their names after a short time, to make
1952 debugging easier. After a longer timeout they are forcibly killed,
1955 * journalctl (and the other tools that display logs) will now highlight
1956 warnings in yellow (previously, both LOG_NOTICE and LOG_WARNING where
1957 shown in bright bold, now only LOG_NOTICE is). Moreover, audit logs
1958 are now shown in blue color, to separate them visually from regular
1959 logs. References to configuration files are now turned into clickable
1960 links on terminals that support that.
1962 * systemd-journald will now stop logging to /var/log/journal during
1963 shutdown when /var/ is on a separate mount, so that it can be
1964 unmounted safely during shutdown.
1966 * systemd-resolved gained support for a new 'strict' DNS-over-TLS mode.
1968 * systemd-resolved "Cache=" configuration option in resolved.conf has
1969 been extended to also accept the 'no-negative' value. Previously,
1970 only a boolean option was allowed (yes/no), having yes as the
1971 default. If this option is set to 'no-negative', negative answers are
1972 not cached while the old cache heuristics are used positive answers.
1973 The default remains unchanged.
1975 * The predictable naming scheme for network devices now supports
1976 generating predictable names for "netdevsim" devices.
1978 Moreover, the "en" prefix was dropped from the ID_NET_NAME_ONBOARD
1981 Those two changes form a new net.naming-policy-scheme= entry.
1982 Distributions which want to preserve naming stability may want to set
1983 the -Ddefault-net-naming-scheme= configuration option.
1985 * systemd-networkd now supports MACsec, nlmon, IPVTAP and Xfrm
1986 interfaces natively.
1988 * systemd-networkd's bridge FDB support now allows configuration of a
1989 destination address for each entry (Destination=), as well as the
1990 VXLAN VNI (VNI=), as well as an option to declare what an entry is
1991 associated with (AssociatedWith=).
1993 * systemd-networkd's DHCPv4 support now understands a new MaxAttempts=
1994 option for configuring the maximum number of DHCP lease requests. It
1995 also learnt a new BlackList= option for deny-listing DHCP servers (a
1996 similar setting has also been added to the IPv6 RA client), as well
1997 as a SendRelease= option for configuring whether to send a DHCP
1998 RELEASE message when terminating.
2000 * systemd-networkd's DHCPv4 and DHCPv6 stacks can now be configured
2001 separately in the [DHCPv4] and [DHCPv6] sections.
2003 * systemd-networkd's DHCP support will now optionally create an
2004 implicit host route to the DNS server specified in the DHCP lease, in
2005 addition to the routes listed explicitly in the lease. This should
2006 ensure that in multi-homed systems DNS traffic leaves the systems on
2007 the interface that acquired the DNS server information even if other
2008 routes such as default routes exist. This behaviour may be turned on
2009 with the new RoutesToDNS= option.
2011 * systemd-networkd's VXLAN support gained a new option
2012 GenericProtocolExtension= for enabling VXLAN Generic Protocol
2013 Extension support, as well as IPDoNotFragment= for setting the IP
2014 "Don't fragment" bit on outgoing packets. A similar option has been
2015 added to the GENEVE support.
2017 * In systemd-networkd's [Route] section you may now configure
2018 FastOpenNoCookie= for configuring per-route TCP fast-open support, as
2019 well as TTLPropagate= for configuring Label Switched Path (LSP) TTL
2020 propagation. The Type= setting now supports local, broadcast,
2021 anycast, multicast, any, xresolve routes, too.
2023 * systemd-networkd's [Network] section learnt a new option
2024 DefaultRouteOnDevice= for automatically configuring a default route
2025 onto the network device.
2027 * systemd-networkd's bridging support gained two new options ProxyARP=
2028 and ProxyARPWifi= for configuring proxy ARP behaviour as well as
2029 MulticastRouter= for configuring multicast routing behaviour. A new
2030 option MulticastIGMPVersion= may be used to change bridge's multicast
2031 Internet Group Management Protocol (IGMP) version.
2033 * systemd-networkd's FooOverUDP support gained the ability to configure
2034 local and peer IP addresses via Local= and Peer=. A new option
2035 PeerPort= may be used to configure the peer's IP port.
2037 * systemd-networkd's TUN support gained a new setting VnetHeader= for
2038 tweaking Generic Segment Offload support.
2040 * The address family for policy rules may be specified using the new
2041 Family= option in the [RoutingPolicyRule] section.
2043 * networkctl gained a new "delete" command for removing virtual network
2044 devices, as well as a new "--stats" switch for showing device
2047 * networkd.conf gained a new setting SpeedMeter= and
2048 SpeedMeterIntervalSec=, to measure bitrate of network interfaces. The
2049 measured speed may be shown by 'networkctl status'.
2051 * "networkctl status" now displays MTU and queue lengths, and more
2052 detailed information about VXLAN and bridge devices.
2054 * systemd-networkd's .network and .link files gained a new Property=
2055 setting in the [Match] section, to match against devices with
2056 specific udev properties.
2058 * systemd-networkd's tunnel support gained a new option
2059 AssignToLoopback= for selecting whether to use the loopback device
2060 "lo" as underlying device.
2062 * systemd-networkd's MACAddress= setting in the [Neighbor] section has
2063 been renamed to LinkLayerAddress=, and it now allows configuration of
2066 * systemd-networkd's handling of the kernel's disable_ipv6 sysctl is
2067 simplified: systemd-networkd will disable the sysctl (enable IPv6) if
2068 IPv6 configuration (static or DHCPv6) was found for a given
2069 interface. It will not touch the sysctl otherwise.
2071 * The order of entries is $PATH used by the user manager instance was
2072 changed to put bin/ entries before the corresponding sbin/ entries.
2073 It is recommended to not rely on this order, and only ever have one
2074 binary with a given name in the system paths under /usr.
2076 * A new tool systemd-network-generator has been added that may generate
2077 .network, .netdev and .link files from IP configuration specified on
2078 the kernel command line in the format used by Dracut.
2080 * The CriticalConnection= setting in .network files is now deprecated,
2081 and replaced by a new KeepConfiguration= setting which allows more
2082 detailed configuration of the IP configuration to keep in place.
2084 * systemd-analyze gained a few new verbs:
2086 - "systemd-analyze timestamp" parses and converts timestamps. This is
2087 similar to the existing "systemd-analyze calendar" command which
2088 does the same for recurring calendar events.
2090 - "systemd-analyze timespan" parses and converts timespans (i.e.
2091 durations as opposed to points in time).
2093 - "systemd-analyze condition" will parse and test ConditionXYZ=
2096 - "systemd-analyze exit-status" will parse and convert exit status
2097 codes to their names and back.
2099 - "systemd-analyze unit-files" will print a list of all unit
2100 file paths and unit aliases.
2102 * SuccessExitStatus=, RestartPreventExitStatus=, and
2103 RestartForceExitStatus= now accept exit status names (e.g. "DATAERR"
2104 is equivalent to "65"). Those exit status name mappings may be
2105 displayed with the systemd-analyze exit-status verb describe above.
2107 * systemd-logind now exposes a per-session SetBrightness() bus call,
2108 which may be used to securely change the brightness of a kernel
2109 brightness device, if it belongs to the session's seat. By using this
2110 call unprivileged clients can make changes to "backlight" and "leds"
2111 devices securely with strict requirements on session membership.
2112 Desktop environments may use this to generically make brightness
2113 changes to such devices without shipping private SUID binaries or
2114 udev rules for that purpose.
2116 * "udevadm info" gained a --wait-for-initialization switch to wait for
2117 a device to be initialized.
2119 * systemd-hibernate-resume-generator will now look for resumeflags= on
2120 the kernel command line, which is similar to rootflags= and may be
2121 used to configure device timeout for the hibernation device.
2123 * sd-event learnt a new API call sd_event_source_disable_unref() for
2124 disabling and unref'ing an event source in a single function. A
2125 related call sd_event_source_disable_unrefp() has been added for use
2126 with gcc's cleanup extension.
2128 * The sd-id128.h public API gained a new definition
2129 SD_ID128_UUID_FORMAT_STR for formatting a 128bit ID in UUID format
2132 * "busctl introspect" gained a new switch --xml-interface for dumping
2133 XML introspection data unmodified.
2135 * PID 1 may now show the unit name instead of the unit description
2136 string in its status output during boot. This may be configured in
2137 the StatusUnitFormat= setting in /etc/systemd/system.conf or the
2138 kernel command line option systemd.status_unit_format=.
2140 * PID 1 now understands a new option KExecWatchdogSec= in
2141 /etc/systemd/system.conf to set a watchdog timeout for kexec reboots.
2142 Previously watchdog functionality was only available for regular
2143 reboots. The new setting defaults to off, because we don't know in
2144 the general case if the watchdog will be reset after kexec (some
2145 drivers do reset it, but not all), and the new userspace might not be
2146 configured to handle the watchdog.
2148 Moreover, the old ShutdownWatchdogSec= setting has been renamed to
2149 RebootWatchdogSec= to more clearly communicate what it is about. The
2150 old name is still accepted for compatibility.
2152 * The systemd.debug_shell kernel command line option now optionally
2153 takes a tty name to spawn the debug shell on, which allows a
2154 different tty to be selected than the built-in default.
2156 * Service units gained a new ExecCondition= setting which will run
2157 before ExecStartPre= and either continue execution of the unit (for
2158 clean exit codes), stop execution without marking the unit failed
2159 (for exit codes 1 through 254), or stop execution and fail the unit
2160 (for exit code 255 or abnormal termination).
2162 * A new service systemd-pstore.service has been added that pulls data
2163 from /sys/fs/pstore/ and saves it to /var/lib/pstore for later
2166 * timedatectl gained new verbs for configuring per-interface NTP
2167 service configuration for systemd-timesyncd.
2169 * "localectl list-locales" won't list non-UTF-8 locales anymore. It's
2170 2019. (You can set non-UTF-8 locales though, if you know their name.)
2172 * If variable assignments in sysctl.d/ files are prefixed with "-" any
2173 failures to apply them are now ignored.
2175 * systemd-random-seed.service now optionally credits entropy when
2176 applying the seed to the system. Set $SYSTEMD_RANDOM_SEED_CREDIT to
2177 true for the service to enable this behaviour, but please consult the
2178 documentation first, since this comes with a couple of caveats.
2180 * systemd-random-seed.service is now a synchronization point for full
2181 initialization of the kernel's entropy pool. Services that require
2182 /dev/urandom to be correctly initialized should be ordered after this
2185 * The systemd-boot boot loader has been updated to optionally maintain
2186 a random seed file in the EFI System Partition (ESP). During the boot
2187 phase, this random seed is read and updated with a new seed
2188 cryptographically derived from it. Another derived seed is passed to
2189 the OS. The latter seed is then credited to the kernel's entropy pool
2190 very early during userspace initialization (from PID 1). This allows
2191 systems to boot up with a fully initialized kernel entropy pool from
2192 earliest boot on, and thus entirely removes all entropy pool
2193 initialization delays from systems using systemd-boot. Special care
2194 is taken to ensure different seeds are derived on system images
2195 replicated to multiple systems. "bootctl status" will show whether
2196 a seed was received from the boot loader.
2198 * bootctl gained two new verbs:
2200 - "bootctl random-seed" will generate the file in ESP and an EFI
2201 variable to allow a random seed to be passed to the OS as described
2204 - "bootctl is-installed" checks whether systemd-boot is currently
2207 * bootctl will warn if it detects that boot entries are misconfigured
2208 (for example if the kernel image was removed without purging the
2211 * A new document has been added describing systemd's use and support
2212 for the kernel's entropy pool subsystem:
2214 https://systemd.io/RANDOM_SEEDS
2216 * When the system is hibernated the swap device to write the
2217 hibernation image to is now automatically picked from all available
2218 swap devices, preferring the swap device with the highest configured
2219 priority over all others, and picking the device with the most free
2220 space if there are multiple devices with the highest priority.
2222 * /etc/crypttab support has learnt a new keyfile-timeout= per-device
2223 option that permits selecting the timeout how long to wait for a
2224 device with an encryption key before asking for the password.
2226 * IOWeight= has learnt to properly set the IO weight when using the
2227 BFQ scheduler officially found in kernels 5.0+.
2229 * A new mailing list has been created for reporting of security issues:
2230 systemd-security@redhat.com. For mode details, see
2231 https://systemd.io/CONTRIBUTING#security-vulnerability-reports.
2233 Contributions from: Aaron Barany, Adrian Bunk, Alan Jenkins, Albrecht
2234 Lohofener, Andrej Valek, Anita Zhang, Arian van Putten, Balint Reczey,
2235 Bastien Nocera, Ben Boeckel, Benjamin Robin, camoz, Chen Qi, Chris
2236 Chiu, Chris Down, Christian Göttsche, Christian Kellner, Clinton Roy,
2237 Connor Reeder, Daniel Black, Daniel Lublin, Daniele Medri, Dan
2238 Streetman, Dave Reisner, Dave Ross, David Art, David Tardon, Debarshi
2239 Ray, Dimitri John Ledkov, Dominick Grift, Donald Buczek, Douglas
2240 Christman, Eric DeVolder, EtherGraf, Evgeny Vereshchagin, Feldwor,
2241 Felix Riemann, Florian Dollinger, Francesco Pennica, Franck Bui,
2242 Frantisek Sumsal, Franz Pletz, frederik, Hans de Goede, Iago López
2243 Galeiras, Insun Pyo, Ivan Shapovalov, Iwan Timmer, Jack, Jakob
2244 Unterwurzacher, Jan Chren, Jan Klötzke, Jan Losinski, Jan Pokorný, Jan
2245 Synacek, Jan-Michael Brummer, Jeka Pats, Jeremy Soller, Jérémy Rosen,
2246 Jiri Pirko, Joe Lin, Joerg Behrmann, Joe Richey, Jóhann B. Guðmundsson,
2247 Johannes Christ, Johannes Schmitz, Jonathan Rouleau, Jorge Niedbalski,
2248 Jörg Thalheim, Kai Krakow, Kai Lüke, Karel Zak, Kashyap Chamarthy,
2249 Krayushkin Konstantin, Lennart Poettering, Lubomir Rintel, Luca
2250 Boccassi, Luís Ferreira, Marc-André Lureau, Markus Felten, Martin Pitt,
2251 Matthew Leeds, Mattias Jernberg, Michael Biebl, Michael Olbrich,
2252 Michael Prokop, Michael Stapelberg, Michael Zhivich, Michal Koutný,
2253 Michal Sekletar, Mike Gilbert, Milan Broz, Miroslav Lichvar, mpe85,
2254 Mr-Foo, Network Silence, Oliver Harley, pan93412, Paul Menzel, pEJipE,
2255 Peter A. Bigot, Philip Withnall, Piotr Drąg, Rafael Fontenelle, Robert
2256 Scheck, Roberto Santalla, Ronan Pigott, root, RussianNeuroMancer,
2257 Sebastian Jennen, shinygold, Shreyas Behera, Simon Schricker, Susant
2258 Sahani, Thadeu Lima de Souza Cascardo, Theo Ouzhinski, Thiebaud
2259 Weksteen, Thomas Haller, Thomas Weißschuh, Tomas Mraz, Tommi Rantala,
2260 Topi Miettinen, VD-Lycos, ven, Vladimir Yerilov, Wieland Hoffmann,
2261 William A. Kennington III, William Wold, Xi Ruoyao, Yuri Chornoivan,
2262 Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zhang Xianwei
2264 – Camerino, 2019-09-03
2268 * In .link files, MACAddressPolicy=persistent (the default) is changed
2269 to cover more devices. For devices like bridges, tun, tap, bond, and
2270 similar interfaces that do not have other identifying information,
2271 the interface name is used as the basis for persistent seed for MAC
2272 and IPv4LL addresses. The way that devices that were handled
2273 previously is not changed, and this change is about covering more
2274 devices then previously by the "persistent" policy.
2276 MACAddressPolicy=random may be used to force randomized MACs and
2277 IPv4LL addresses for a device if desired.
2279 Hint: the log output from udev (at debug level) was enhanced to
2280 clarify what policy is followed and which attributes are used.
2281 `SYSTEMD_LOG_LEVEL=debug udevadm test-builtin net_setup_link /sys/class/net/<name>`
2282 may be used to view this.
2284 Hint: if a bridge interface is created without any slaves, and gains
2285 a slave later, then now the bridge does not inherit slave's MAC.
2286 To inherit slave's MAC, for example, create the following file:
2288 # /etc/systemd/network/98-bridge-inherit-mac.link
2293 MACAddressPolicy=none
2296 * The .device units generated by systemd-fstab-generator and other
2297 generators do not automatically pull in the corresponding .mount unit
2298 as a Wants= dependency. This means that simply plugging in the device
2299 will not cause the mount unit to be started automatically. But please
2300 note that the mount unit may be started for other reasons, in
2301 particular if it is part of local-fs.target, and any unit which
2302 (transitively) depends on local-fs.target is started.
2304 * networkctl list/status/lldp now accept globbing wildcards for network
2305 interface names to match against all existing interfaces.
2307 * The $PIDFILE environment variable is set to point the absolute path
2308 configured with PIDFile= for processes of that service.
2310 * The fallback DNS server list was augmented with Cloudflare public DNS
2311 servers. Use `-Ddns-servers=` to set a different fallback.
2313 * A new special target usb-gadget.target will be started automatically
2314 when a USB Device Controller is detected (which means that the system
2315 is a USB peripheral).
2317 * A new unit setting CPUQuotaPeriodSec= assigns the time period
2318 relatively to which the CPU time quota specified by CPUQuota= is
2321 * A new unit setting ProtectHostname= may be used to prevent services
2322 from modifying hostname information (even if they otherwise would
2323 have privileges to do so).
2325 * A new unit setting NetworkNamespacePath= may be used to specify a
2326 namespace for service or socket units through a path referring to a
2327 Linux network namespace pseudo-file.
2329 * The PrivateNetwork= setting and JoinsNamespaceOf= dependencies now
2330 have an effect on .socket units: when used the listening socket is
2331 created within the configured network namespace instead of the host
2334 * ExecStart= command lines in unit files may now be prefixed with ':'
2335 in which case environment variable substitution is
2336 disabled. (Supported for the other ExecXYZ= settings, too.)
2338 * .timer units gained two new boolean settings OnClockChange= and
2339 OnTimezoneChange= which may be used to also trigger a unit when the
2340 system clock is changed or the local timezone is
2341 modified. systemd-run has been updated to make these options easily
2342 accessible from the command line for transient timers.
2344 * Two new conditions for units have been added: ConditionMemory= may be
2345 used to conditionalize a unit based on installed system
2346 RAM. ConditionCPUs= may be used to conditionalize a unit based on
2347 installed CPU cores.
2349 * The @default system call filter group understood by SystemCallFilter=
2350 has been updated to include the new rseq() system call introduced in
2353 * A new time-set.target has been added that indicates that the system
2354 time has been set from a local source (possibly imprecise). The
2355 existing time-sync.target is stronger and indicates that the time has
2356 been synchronized with a precise external source. Services where
2357 approximate time is sufficient should use the new target.
2359 * "systemctl start" (and related commands) learnt a new
2360 --show-transaction option. If specified brief information about all
2361 jobs queued because of the requested operation is shown.
2363 * systemd-networkd recognizes a new operation state 'enslaved', used
2364 (instead of 'degraded' or 'carrier') for interfaces which form a
2365 bridge, bond, or similar, and an new 'degraded-carrier' operational
2366 state used for the bond or bridge master interface when one of the
2367 enslaved devices is not operational.
2369 * .network files learnt the new IgnoreCarrierLoss= option for leaving
2370 networks configured even if the carrier is lost.
2372 * The RequiredForOnline= setting in .network files may now specify a
2373 minimum operational state required for the interface to be considered
2374 "online" by systemd-networkd-wait-online. Related to this
2375 systemd-networkd-wait-online gained a new option --operational-state=
2376 to configure the same, and its --interface= option was updated to
2377 optionally also take an operational state specific for an interface.
2379 * systemd-networkd-wait-online gained a new setting --any for waiting
2380 for only one of the requested interfaces instead of all of them.
2382 * systemd-networkd now implements L2TP tunnels.
2384 * Two new .network settings UseAutonomousPrefix= and UseOnLinkPrefix=
2385 may be used to cause autonomous and onlink prefixes received in IPv6
2386 Router Advertisements to be ignored.
2388 * New MulticastFlood=, NeighborSuppression=, and Learning= .network
2389 file settings may be used to tweak bridge behaviour.
2391 * The new TripleSampling= option in .network files may be used to
2392 configure CAN triple sampling.
2394 * A new .netdev settings PrivateKeyFile= and PresharedKeyFile= may be
2395 used to point to private or preshared key for a WireGuard interface.
2397 * /etc/crypttab now supports the same-cpu-crypt and
2398 submit-from-crypt-cpus options to tweak encryption work scheduling
2401 * systemd-tmpfiles will now take a BSD file lock before operating on a
2402 contents of directory. This may be used to temporarily exclude
2403 directories from aging by taking the same lock (useful for example
2404 when extracting a tarball into /tmp or /var/tmp as a privileged user,
2405 which might create files with really old timestamps, which
2406 nevertheless should not be deleted). For further details, see:
2408 https://systemd.io/TEMPORARY_DIRECTORIES
2410 * systemd-tmpfiles' h line type gained support for the
2411 FS_PROJINHERIT_FL ('P') file attribute (introduced in kernel 4.5),
2412 controlling project quota inheritance.
2414 * sd-boot and bootctl now implement support for an Extended Boot Loader
2415 (XBOOTLDR) partition, that is intended to be mounted to /boot, in
2416 addition to the ESP partition mounted to /efi or /boot/efi.
2417 Configuration file fragments, kernels, initrds and other EFI images
2418 to boot will be loaded from both the ESP and XBOOTLDR partitions.
2419 The XBOOTLDR partition was previously described by the Boot Loader
2420 Specification, but implementation was missing in sd-boot. Support for
2421 this concept allows using the sd-boot boot loader in more
2422 conservative scenarios where the boot loader itself is placed in the
2423 ESP but the kernels to boot (and their metadata) in a separate
2426 * A system may now be booted with systemd.volatile=overlay on the
2427 kernel command line, which causes the root file system to be set up
2428 an overlayfs mount combining the root-only root directory with a
2429 writable tmpfs. In this setup, the underlying root device is not
2430 modified, and any changes are lost at reboot.
2432 * Similar, systemd-nspawn can now boot containers with a volatile
2433 overlayfs root with the new --volatile=overlay switch.
2435 * systemd-nspawn can now consume OCI runtime bundles using a new
2436 --oci-bundle= option. This implementation is fully usable, with most
2437 features in the specification implemented, but since this a lot of
2438 new code and functionality, this feature should most likely not
2439 be used in production yet.
2441 * systemd-nspawn now supports various options described by the OCI
2442 runtime specification on the command-line and in .nspawn files:
2443 --inaccessible=/Inaccessible= may be used to mask parts of the file
2444 system tree, --console=/--pipe may be used to configure how standard
2445 input, output, and error are set up.
2447 * busctl learned the `emit` verb to generate D-Bus signals.
2449 * systemd-analyze cat-config may be used to gather and display
2450 configuration spread over multiple files, for example system and user
2451 presets, tmpfiles.d, sysusers.d, udev rules, etc.
2453 * systemd-analyze calendar now takes an optional new parameter
2454 --iterations= which may be used to show a maximum number of iterations
2455 the specified expression will elapse next.
2457 * The sd-bus C API gained support for naming method parameters in the
2460 * systemd-logind gained D-Bus APIs to specify the "reboot parameter"
2461 the reboot() system call expects.
2463 * journalctl learnt a new --cursor-file= option that points to a file
2464 from which a cursor should be loaded in the beginning and to which
2465 the updated cursor should be stored at the end.
2467 * ACRN hypervisor and Windows Subsystem for Linux (WSL) are now
2468 detected by systemd-detect-virt (and may also be used in
2469 ConditionVirtualization=).
2471 * The behaviour of systemd-logind may now be modified with environment
2472 variables $SYSTEMD_REBOOT_TO_FIRMWARE_SETUP,
2473 $SYSTEMD_REBOOT_TO_BOOT_LOADER_MENU, and
2474 $SYSTEMD_REBOOT_TO_BOOT_LOADER_ENTRY. They cause logind to either
2475 skip the relevant operation completely (when set to false), or to
2476 create a flag file in /run/systemd (when set to true), instead of
2477 actually commencing the real operation when requested. The presence
2478 of /run/systemd/reboot-to-firmware-setup,
2479 /run/systemd/reboot-to-boot-loader-menu, and
2480 /run/systemd/reboot-to-boot-loader-entry, may be used by alternative
2481 boot loader implementations to replace some steps logind performs
2482 during reboot with their own operations.
2484 * systemctl can be used to request a reboot into the boot loader menu
2485 or a specific boot loader entry with the new --boot-load-menu= and
2486 --boot-loader-entry= options to a reboot command. (This requires a
2487 boot loader that supports this, for example sd-boot.)
2489 * kernel-install will no longer unconditionally create the output
2490 directory (e.g. /efi/<machine-id>/<kernel-version>) for boot loader
2491 snippets, but will do only if the machine-specific parent directory
2492 (i.e. /efi/<machine-id>/) already exists. bootctl has been modified
2493 to create this parent directory during sd-boot installation.
2495 This makes it easier to use kernel-install with plugins which support
2496 a different layout of the bootloader partitions (for example grub2).
2498 * During package installation (with `ninja install`), we would create
2499 symlinks for getty@tty1.service, systemd-networkd.service,
2500 systemd-networkd.socket, systemd-resolved.service,
2501 remote-cryptsetup.target, remote-fs.target,
2502 systemd-networkd-wait-online.service, and systemd-timesyncd.service
2503 in /etc, as if `systemctl enable` was called for those units, to make
2504 the system usable immediately after installation. Now this is not
2505 done anymore, and instead calling `systemctl preset-all` is
2506 recommended after the first installation of systemd.
2508 * A new boolean sandboxing option RestrictSUIDSGID= has been added that
2509 is built on seccomp. When turned on creation of SUID/SGID files is
2512 * The NoNewPrivileges= and the new RestrictSUIDSGID= options are now
2513 implied if DynamicUser= is turned on for a service. This hardens
2514 these services, so that they neither can benefit from nor create
2515 SUID/SGID executables. This is a minor compatibility breakage, given
2516 that when DynamicUser= was first introduced SUID/SGID behaviour was
2517 unaffected. However, the security benefit of these two options is
2518 substantial, and the setting is still relatively new, hence we opted
2519 to make it mandatory for services with dynamic users.
2521 Contributions from: Adam Jackson, Alexander Tsoy, Andrey Yashkin,
2522 Andrzej Pietrasiewicz, Anita Zhang, Balint Reczey, Beniamino Galvani,
2523 Ben Iofel, Benjamin Berg, Benjamin Dahlhoff, Chris, Chris Morin,
2524 Christopher Wong, Claudius Ellsel, Clemens Gruber, dana, Daniel Black,
2525 Davide Cavalca, David Michael, David Rheinsberg, emersion, Evgeny
2526 Vereshchagin, Filipe Brandenburger, Franck Bui, Frantisek Sumsal,
2527 Giacinto Cifelli, Hans de Goede, Hugo Kindel, Ignat Korchagin, Insun
2528 Pyo, Jan Engelhardt, Jonas Dorel, Jonathan Lebon, Jonathon Kowalski,
2529 Jörg Sommer, Jörg Thalheim, Jussi Pakkanen, Kai-Heng Feng, Lennart
2530 Poettering, Lubomir Rintel, Luís Ferreira, Martin Pitt, Matthias
2531 Klumpp, Michael Biebl, Michael Niewöhner, Michael Olbrich, Michal
2532 Sekletar, Mike Lothian, Paul Menzel, Piotr Drąg, Riccardo Schirone,
2533 Robin Elvedi, Roman Kulikov, Ronald Tschalär, Ross Burton, Ryan
2534 Gonzalez, Sebastian Krzyszkowiak, Stephane Chazelas, StKob, Susant
2535 Sahani, Sylvain Plantefève, Szabolcs Fruhwald, Taro Yamada, Theo
2536 Ouzhinski, Thomas Haller, Tobias Jungel, Tom Yan, Tony Asleson, Topi
2537 Miettinen, unixsysadmin, Van Laser, Vesa Jääskeläinen, Yu, Li-Yu,
2538 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
2540 — Warsaw, 2019-04-11
2544 * The default locale can now be configured at compile time. Otherwise,
2545 a suitable default will be selected automatically (one of C.UTF-8,
2546 en_US.UTF-8, and C).
2548 * The version string shown by systemd and other tools now includes the
2549 git commit hash when built from git. An override may be specified
2550 during compilation, which is intended to be used by distributions to
2551 include the package release information.
2553 * systemd-cat can now filter standard input and standard error streams
2554 for different syslog priorities using the new --stderr-priority=
2557 * systemd-journald and systemd-journal-remote reject entries which
2558 contain too many fields (CVE-2018-16865) and set limits on the
2559 process' command line length (CVE-2018-16864).
2561 * $DBUS_SESSION_BUS_ADDRESS environment variable is set by pam_systemd
2564 * A new network device NamePolicy "keep" is implemented for link files,
2565 and used by default in 99-default.link (the fallback configuration
2566 provided by systemd). With this policy, if the network device name
2567 was already set by userspace, the device will not be renamed again.
2568 This matches the naming scheme that was implemented before
2569 systemd-240. If naming-scheme < 240 is specified, the "keep" policy
2570 is also enabled by default, even if not specified. Effectively, this
2571 means that if naming-scheme >= 240 is specified, network devices will
2572 be renamed according to the configuration, even if they have been
2573 renamed already, if "keep" is not specified as the naming policy in
2574 the .link file. The 99-default.link file provided by systemd includes
2575 "keep" for backwards compatibility, but it is recommended for user
2576 installed .link files to *not* include it.
2578 The "kernel" policy, which keeps kernel names declared to be
2579 "persistent", now works again as documented.
2581 * kernel-install script now optionally takes the paths to one or more
2582 initrd files, and passes them to all plugins.
2584 * The mincore() system call has been dropped from the @system-service
2585 system call filter group, as it is pretty exotic and may potentially
2586 used for side-channel attacks.
2588 * -fPIE is dropped from compiler and linker options. Please specify
2589 -Db_pie=true option to meson to build position-independent
2590 executables. Note that the meson option is supported since meson-0.49.
2592 * The fs.protected_regular and fs.protected_fifos sysctls, which were
2593 added in Linux 4.19 to make some data spoofing attacks harder, are
2594 now enabled by default. While this will hopefully improve the
2595 security of most installations, it is technically a backwards
2596 incompatible change; to disable these sysctls again, place the
2597 following lines in /etc/sysctl.d/60-protected.conf or a similar file:
2599 fs.protected_regular = 0
2600 fs.protected_fifos = 0
2602 Note that the similar hardlink and symlink protection has been
2603 enabled since v199, and may be disabled likewise.
2605 * The files read from the EnvironmentFile= setting in unit files now
2606 parse backslashes inside quotes literally, matching the behaviour of
2609 * udevadm trigger, udevadm control, udevadm settle and udevadm monitor
2610 now automatically become NOPs when run in a chroot() environment.
2612 * The tmpfiles.d/ "C" line type will now copy directory trees not only
2613 when the destination is so far missing, but also if it already exists
2614 as a directory and is empty. This is useful to cater for systems
2615 where directory trees are put together from multiple separate mount
2616 points but otherwise empty.
2618 * A new function sd_bus_close_unref() (and the associated
2619 sd_bus_close_unrefp()) has been added to libsystemd, that combines
2620 sd_bus_close() and sd_bus_unref() in one.
2622 * udevadm control learnt a new option for --ping for testing whether a
2623 systemd-udevd instance is running and reacting.
2625 * udevadm trigger learnt a new option for --wait-daemon for waiting
2626 systemd-udevd daemon to be initialized.
2628 Contributions from: Aaron Plattner, Alberts Muktupāvels, Alex Mayer,
2629 Ayman Bagabas, Beniamino Galvani, Burt P, Chris Down, Chris Lamb, Chris
2630 Morin, Christian Hesse, Claudius Ellsel, dana, Daniel Axtens, Daniele
2631 Medri, Dave Reisner, David Santamaría Rogado, Diego Canuhe, Dimitri
2632 John Ledkov, Evgeny Vereshchagin, Fabrice Fontaine, Filipe
2633 Brandenburger, Franck Bui, Frantisek Sumsal, govwin, Hans de Goede,
2634 James Hilliard, Jan Engelhardt, Jani Uusitalo, Jan Janssen, Jan
2635 Synacek, Jonathan McDowell, Jonathan Roemer, Jonathon Kowalski, Joost
2636 Heitbrink, Jörg Thalheim, Lance, Lennart Poettering, Louis Taylor,
2637 Lucas Werkmeister, Mantas Mikulėnas, Marc-Antoine Perennou,
2638 marvelousblack, Michael Biebl, Michael Sloan, Michal Sekletar, Mike
2639 Auty, Mike Gilbert, Mikhail Kasimov, Neil Brown, Niklas Hambüchen,
2640 Patrick Williams, Paul Seyfert, Peter Hutterer, Philip Withnall, Roger
2641 James, Ronnie P. Thomas, Ryan Gonzalez, Sam Morris, Stephan Edel,
2642 Stephan Gerhold, Susant Sahani, Taro Yamada, Thomas Haller, Topi
2643 Miettinen, YiFei Zhu, YmrDtnJu, YunQiang Su, Yu Watanabe, Zbigniew
2644 Jędrzejewski-Szmek, zsergeant77, Дамјан Георгиевски
2646 — Berlin, 2019-02-14
2650 * NoNewPrivileges=yes has been set for all long-running services
2651 implemented by systemd. Previously, this was problematic due to
2652 SELinux (as this would also prohibit the transition from PID1's label
2653 to the service's label). This restriction has since been lifted, but
2654 an SELinux policy update is required.
2655 (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.)
2657 * DynamicUser=yes is dropped from systemd-networkd.service,
2658 systemd-resolved.service and systemd-timesyncd.service, which was
2659 enabled in v239 for systemd-networkd.service and systemd-resolved.service,
2660 and since v236 for systemd-timesyncd.service. The users and groups
2661 systemd-network, systemd-resolve and systemd-timesync are created
2662 by systemd-sysusers again. Distributors or system administrators
2663 may need to create these users and groups if they not exist (or need
2664 to re-enable DynamicUser= for those units) while upgrading systemd.
2665 Also, the clock file for systemd-timesyncd may need to move from
2666 /var/lib/private/systemd/timesync/clock to /var/lib/systemd/timesync/clock.
2668 * When unit files are loaded from disk, previously systemd would
2669 sometimes (depending on the unit loading order) load units from the
2670 target path of symlinks in .wants/ or .requires/ directories of other
2671 units. This meant that unit could be loaded from different paths
2672 depending on whether the unit was requested explicitly or as a
2673 dependency of another unit, not honouring the priority of directories
2674 in search path. It also meant that it was possible to successfully
2675 load and start units which are not found in the unit search path, as
2676 long as they were requested as a dependency and linked to from
2677 .wants/ or .requires/. The target paths of those symlinks are not
2678 used for loading units anymore and the unit file must be found in
2681 * A new service type has been added: Type=exec. It's very similar to
2682 Type=simple but ensures the service manager will wait for both fork()
2683 and execve() of the main service binary to complete before proceeding
2684 with follow-up units. This is primarily useful so that the manager
2685 propagates any errors in the preparation phase of service execution
2686 back to the job that requested the unit to be started. For example,
2687 consider a service that has ExecStart= set to a file system binary
2688 that doesn't exist. With Type=simple starting the unit would be
2689 considered instantly successful, as only fork() has to complete
2690 successfully and the manager does not wait for execve(), and hence
2691 its failure is seen "too late". With the new Type=exec service type
2692 starting the unit will fail, as the manager will wait for the
2693 execve() and notice its failure, which is then propagated back to the
2696 NOTE: with the next release 241 of systemd we intend to change the
2697 systemd-run tool to default to Type=exec for transient services
2698 started by it. This should be mostly safe, but in specific corner
2699 cases might result in problems, as the systemd-run tool will then
2700 block on NSS calls (such as user name look-ups due to User=) done
2701 between the fork() and execve(), which under specific circumstances
2702 might cause problems. It is recommended to specify "-p Type=simple"
2703 explicitly in the few cases where this applies. For regular,
2704 non-transient services (i.e. those defined with unit files on disk)
2705 we will continue to default to Type=simple.
2707 * The Linux kernel's current default RLIMIT_NOFILE resource limit for
2708 userspace processes is set to 1024 (soft) and 4096
2709 (hard). Previously, systemd passed this on unmodified to all
2710 processes it forked off. With this systemd release the hard limit
2711 systemd passes on is increased to 512K, overriding the kernel's
2712 defaults and substantially increasing the number of simultaneous file
2713 descriptors unprivileged userspace processes can allocate. Note that
2714 the soft limit remains at 1024 for compatibility reasons: the
2715 traditional UNIX select() call cannot deal with file descriptors >=
2716 1024 and increasing the soft limit globally might thus result in
2717 programs unexpectedly allocating a high file descriptor and thus
2718 failing abnormally when attempting to use it with select() (of
2719 course, programs shouldn't use select() anymore, and prefer
2720 poll()/epoll, but the call unfortunately remains undeservedly popular
2721 at this time). This change reflects the fact that file descriptor
2722 handling in the Linux kernel has been optimized in more recent
2723 kernels and allocating large numbers of them should be much cheaper
2724 both in memory and in performance than it used to be. Programs that
2725 want to take benefit of the increased limit have to "opt-in" into
2726 high file descriptors explicitly by raising their soft limit. Of
2727 course, when they do that they must acknowledge that they cannot use
2728 select() anymore (and neither can any shared library they use — or
2729 any shared library used by any shared library they use and so on).
2730 Which default hard limit is most appropriate is of course hard to
2731 decide. However, given reports that ~300K file descriptors are used
2732 in real-life applications we believe 512K is sufficiently high as new
2733 default for now. Note that there are also reports that using very
2734 high hard limits (e.g. 1G) is problematic: some software allocates
2735 large arrays with one element for each potential file descriptor
2736 (Java, …) — a high hard limit thus triggers excessively large memory
2737 allocations in these applications. Hopefully, the new default of 512K
2738 is a good middle ground: higher than what real-life applications
2739 currently need, and low enough for avoid triggering excessively large
2740 allocations in problematic software. (And yes, somebody should fix
2743 * The fs.nr_open and fs.file-max sysctls are now automatically bumped
2744 to the highest possible values, as separate accounting of file
2745 descriptors is no longer necessary, as memcg tracks them correctly as
2746 part of the memory accounting anyway. Thus, from the four limits on
2747 file descriptors currently enforced (fs.file-max, fs.nr_open,
2748 RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two,
2749 and keep only the latter two. A set of build-time options
2750 (-Dbump-proc-sys-fs-file-max=false and -Dbump-proc-sys-fs-nr-open=false)
2751 has been added to revert this change in behaviour, which might be
2752 an option for systems that turn off memcg in the kernel.
2754 * When no /etc/locale.conf file exists (and hence no locale settings
2755 are in place), systemd will now use the "C.UTF-8" locale by default,
2756 and set LANG= to it. This locale is supported by various
2757 distributions including Fedora, with clear indications that upstream
2758 glibc is going to make it available too. This locale enables UTF-8
2759 mode by default, which appears appropriate for 2018.
2761 * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by
2762 default. This effectively switches the RFC3704 Reverse Path filtering
2763 from Strict mode to Loose mode. This is more appropriate for hosts
2764 that have multiple links with routes to the same networks (e.g.
2765 a client with a Wi-Fi and Ethernet both connected to the internet).
2767 Consult the kernel documentation for details on this sysctl:
2768 https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt
2770 * CPUAccounting=yes no longer enables the CPU controller when using
2771 kernel 4.15+ and the unified cgroup hierarchy, as required accounting
2772 statistics are now provided independently from the CPU controller.
2774 * Support for disabling a particular cgroup controller within a sub-tree
2775 has been added through the DisableControllers= directive.
2777 * cgroup_no_v1=all on the kernel command line now also implies
2778 using the unified cgroup hierarchy, unless one explicitly passes
2779 systemd.unified_cgroup_hierarchy=0 on the kernel command line.
2781 * The new "MemoryMin=" unit file property may now be used to set the
2782 memory usage protection limit of processes invoked by the unit. This
2783 controls the cgroup v2 memory.min attribute. Similarly, the new
2784 "IODeviceLatencyTargetSec=" property has been added, wrapping the new
2785 cgroup v2 io.latency cgroup property for configuring per-service I/O
2788 * systemd now supports the cgroup v2 devices BPF logic, as counterpart
2789 to the cgroup v1 "devices" cgroup controller.
2791 * systemd-escape now is able to combine --unescape with --template. It
2792 also learnt a new option --instance for extracting and unescaping the
2793 instance part of a unit name.
2795 * sd-bus now provides the sd_bus_message_readv() which is similar to
2796 sd_bus_message_read() but takes a va_list object. The pair
2797 sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout()
2798 has been added for configuring the default method call timeout to
2799 use. sd_bus_error_move() may be used to efficiently move the contents
2800 from one sd_bus_error structure to another, invalidating the
2801 source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may
2802 be used to control whether a bus connection object is automatically
2803 flushed when an sd-event loop is exited.
2805 * When processing classic BSD syslog log messages, journald will now
2806 save the original time-stamp string supplied in the new
2807 SYSLOG_TIMESTAMP= journal field. This permits consumers to
2808 reconstruct the original BSD syslog message more correctly.
2810 * StandardOutput=/StandardError= in service files gained support for
2811 new "append:…" parameters, for connecting STDOUT/STDERR of a service
2812 to a file, and appending to it.
2814 * The signal to use as last step of killing of unit processes is now
2815 configurable. Previously it was hard-coded to SIGKILL, which may now
2816 be overridden with the new KillSignal= setting. Note that this is the
2817 signal used when regular termination (i.e. SIGTERM) does not suffice.
2818 Similarly, the signal used when aborting a program in case of a
2819 watchdog timeout may now be configured too (WatchdogSignal=).
2821 * The XDG_SESSION_DESKTOP environment variable may now be configured in
2822 the pam_systemd argument line, using the new desktop= switch. This is
2823 useful to initialize it properly from a display manager without
2824 having to touch C code.
2826 * Most configuration options that previously accepted percentage values
2827 now also accept permille values with the '‰' suffix (instead of '%').
2829 * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for
2832 * systemd-resolved's configuration file resolved.conf gained a new
2833 option ReadEtcHosts= which may be used to turn off processing and
2834 honoring /etc/hosts entries.
2836 * The "--wait" switch may now be passed to "systemctl
2837 is-system-running", in which case the tool will synchronously wait
2838 until the system finished start-up.
2840 * hostnamed gained a new bus call to determine the DMI product UUID.
2842 * On x86-64 systemd will now prefer using the RDRAND processor
2843 instruction over /dev/urandom whenever it requires randomness that
2844 neither has to be crypto-grade nor should be reproducible. This
2845 should substantially reduce the amount of entropy systemd requests
2846 from the kernel during initialization on such systems, though not
2847 reduce it to zero. (Why not zero? systemd still needs to allocate
2848 UUIDs and such uniquely, which require high-quality randomness.)
2850 * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP
2851 tunnels. It also gained a new option ForceDHCPv6PDOtherInformation=
2852 for forcing the "Other Information" bit in IPv6 RA messages. The
2853 bonding logic gained four new options AdActorSystemPriority=,
2854 AdUserPortKey=, AdActorSystem= for configuring various 802.3ad
2855 aspects, and DynamicTransmitLoadBalancing= for enabling dynamic
2856 shuffling of flows. The tunnel logic gained a new
2857 IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid
2858 Deployment. The policy rule logic gained four new options IPProtocol=,
2859 SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained
2860 support for the MulticastToUnicast= option. networkd also gained
2861 support for configuring static IPv4 ARP or IPv6 neighbor entries.
2863 * .preset files (as read by 'systemctl preset') may now be used to
2864 instantiate services.
2866 * /etc/crypttab now understands the sector-size= option to configure
2867 the sector size for an encrypted partition.
2869 * Key material for encrypted disks may now be placed on a formatted
2870 medium, and referenced from /etc/crypttab by the UUID of the file
2871 system, followed by "=" suffixed by the path to the key file.
2873 * The "collect" udev component has been removed without replacement, as
2874 it is neither used nor maintained.
2876 * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=,
2877 LogsDirectory=, ConfigurationDirectory= settings are used in a
2878 service the executed processes will now receive a set of environment
2879 variables containing the full paths of these directories.
2880 Specifically, RUNTIME_DIRECTORY=, STATE_DIRECTORY, CACHE_DIRECTORY,
2881 LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options
2882 are used. Note that these options may be used multiple times per
2883 service in which case the resulting paths will be concatenated and
2884 separated by colons.
2886 * Predictable interface naming has been extended to cover InfiniBand
2887 NICs. They will be exposed with an "ib" prefix.
2889 * tmpfiles.d/ line types may now be suffixed with a '-' character, in
2890 which case the respective line failing is ignored.
2892 * .link files may now be used to configure the equivalent to the
2893 "ethtool advertise" commands.
2895 * The sd-device.h and sd-hwdb.h APIs are now exported, as an
2896 alternative to libudev.h. Previously, the latter was just an internal
2897 wrapper around the former, but now these two APIs are exposed
2900 * sd-id128.h gained a new function sd_id128_get_boot_app_specific()
2901 which calculates an app-specific boot ID similar to how
2902 sd_id128_get_machine_app_specific() generates an app-specific machine
2905 * A new tool systemd-id128 has been added that can be used to determine
2906 and generate various 128bit IDs.
2908 * /etc/os-release gained two new standardized fields DOCUMENTATION_URL=
2911 * systemd-hibernate-resume-generator will now honor the "noresume"
2912 kernel command line option, in which case it will bypass resuming
2913 from any hibernated image.
2915 * The systemd-sleep.conf configuration file gained new options
2916 AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=,
2917 AllowHybridSleep= for prohibiting specific sleep modes even if the
2918 kernel exports them.
2920 * portablectl is now officially supported and has thus moved to
2923 * bootctl learnt the two new commands "set-default" and "set-oneshot"
2924 for setting the default boot loader item to boot to (either
2925 persistently or only for the next boot). This is currently only
2926 compatible with sd-boot, but may be implemented on other boot loaders
2927 too, that follow the boot loader interface. The updated interface is
2928 now documented here:
2930 https://systemd.io/BOOT_LOADER_INTERFACE
2932 * A new kernel command line option systemd.early_core_pattern= is now
2933 understood which may be used to influence the core_pattern PID 1
2934 installs during early boot.
2936 * busctl learnt two new options -j and --json= for outputting method
2937 call replies, properties and monitoring output in JSON.
2939 * journalctl's JSON output now supports simple ANSI coloring as well as
2940 a new "json-seq" mode for generating RFC7464 output.
2942 * Unit files now support the %g/%G specifiers that resolve to the UNIX
2943 group/GID of the service manager runs as, similar to the existing
2944 %u/%U specifiers that resolve to the UNIX user/UID.
2946 * systemd-logind learnt a new global configuration option
2947 UserStopDelaySec= that may be set in logind.conf. It specifies how
2948 long the systemd --user instance shall remain started after a user
2949 logs out. This is useful to speed up repetitive re-connections of the
2950 same user, as it means the user's service manager doesn't have to be
2951 stopped/restarted on each iteration, but can be reused between
2952 subsequent options. This setting defaults to 10s. systemd-logind also
2953 exports two new properties on its Manager D-Bus objects indicating
2954 whether the system's lid is currently closed, and whether the system
2957 * systemd gained support for a generic boot counting logic, which
2958 generically permits automatic reverting to older boot loader entries
2959 if newer updated ones don't work. The boot loader side is implemented
2960 in sd-boot, but is kept open for other boot loaders too. For details
2963 https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT
2965 * The SuccessAction=/FailureAction= unit file settings now learnt two
2966 new parameters: "exit" and "exit-force", which result in immediate
2967 exiting of the service manager, and are only useful in systemd --user
2968 and container environments.
2970 * Unit files gained support for a pair of options
2971 FailureActionExitStatus=/SuccessActionExitStatus= for configuring the
2972 exit status to use as service manager exit status when
2973 SuccessAction=/FailureAction= is set to exit or exit-force.
2975 * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service
2976 options may now be used to configure the log rate limiting applied by
2977 journald per-service.
2979 * systemd-analyze gained a new verb "timespan" for parsing and
2980 normalizing time span values (i.e. strings like "5min 7s 8us").
2982 * systemd-analyze also gained a new verb "security" for analyzing the
2983 security and sand-boxing settings of services in order to determine an
2984 "exposure level" for them, indicating whether a service would benefit
2985 from more sand-boxing options turned on for them.
2987 * "systemd-analyze syscall-filter" will now also show system calls
2988 supported by the local kernel but not included in any of the defined
2991 * .nspawn files now understand the Ephemeral= setting, matching the
2992 --ephemeral command line switch.
2994 * sd-event gained the new APIs sd_event_source_get_floating() and
2995 sd_event_source_set_floating() for controlling whether a specific
2996 event source is "floating", i.e. destroyed along with the even loop
2999 * Unit objects on D-Bus gained a new "Refs" property that lists all
3000 clients that currently have a reference on the unit (to ensure it is
3003 * The JoinControllers= option in system.conf is no longer supported, as
3004 it didn't work correctly, is hard to support properly, is legacy (as
3005 the concept only exists on cgroup v1) and apparently wasn't used.
3007 * Journal messages that are generated whenever a unit enters the failed
3008 state are now tagged with a unique MESSAGE_ID. Similarly, messages
3009 generated whenever a service process exits are now made recognizable,
3010 too. A tagged message is also emitted whenever a unit enters the
3011 "dead" state on success.
3013 * systemd-run gained a new switch --working-directory= for configuring
3014 the working directory of the service to start. A shortcut -d is
3015 equivalent, setting the working directory of the service to the
3016 current working directory of the invoking program. The new --shell
3017 (or just -S) option has been added for invoking the $SHELL of the
3018 caller as a service, and implies --pty --same-dir --wait --collect
3019 --service-type=exec. Or in other words, "systemd-run -S" is now the
3020 quickest way to quickly get an interactive in a fully clean and
3021 well-defined system service context.
3023 * machinectl gained a new verb "import-fs" for importing an OS tree
3024 from a directory. Moreover, when a directory or tarball is imported
3025 and single top-level directory found with the OS itself below the OS
3026 tree is automatically mangled and moved one level up.
3028 * systemd-importd will no longer set up an implicit btrfs loop-back
3029 file system on /var/lib/machines. If one is already set up, it will
3030 continue to be used.
3032 * A new generator "systemd-run-generator" has been added. It will
3033 synthesize a unit from one or more program command lines included in
3034 the kernel command line. This is very useful in container managers
3037 # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"'
3039 This will run "systemd-nspawn" on an image, invoke the specified
3040 command line and immediately shut down the container again, returning
3041 the command line's exit code.
3043 * The block device locking logic is now documented:
3045 https://systemd.io/BLOCK_DEVICE_LOCKING
3047 * loginctl and machinectl now optionally output the various tables in
3048 JSON using the --output= switch. It is our intention to add similar
3049 support to systemctl and all other commands.
3051 * udevadm's query and trigger verb now optionally take a .device unit
3054 * systemd-udevd's network naming logic now understands a new
3055 net.naming-scheme= kernel command line switch, which may be used to
3056 pick a specific version of the naming scheme. This helps stabilizing
3057 interface names even as systemd/udev are updated and the naming logic
3060 * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and
3061 SD_ID128_ALLF to test if a 128bit ID is set to all 0xFF bytes, and to
3062 initialize one to all 0xFF.
3064 * After loading the SELinux policy systemd will now recursively relabel
3065 all files and directories listed in
3066 /run/systemd/relabel-extra.d/*.relabel (which should be simple
3067 newline separated lists of paths) in addition to the ones it already
3068 implicitly relabels in /run, /dev and /sys. After the relabelling is
3069 completed the *.relabel files (and /run/systemd/relabel-extra.d/) are
3070 removed. This is useful to permit initrds (i.e. code running before
3071 the SELinux policy is in effect) to generate files in the host
3072 filesystem safely and ensure that the correct label is applied during
3073 the transition to the host OS.
3075 * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding
3076 mknod() handling in user namespaces. Previously mknod() would always
3077 fail with EPERM in user namespaces. Since 4.18 mknod() will succeed
3078 but device nodes generated that way cannot be opened, and attempts to
3079 open them result in EPERM. This breaks the "graceful fallback" logic
3080 in systemd's PrivateDevices= sand-boxing option. This option is
3081 implemented defensively, so that when systemd detects it runs in a
3082 restricted environment (such as a user namespace, or an environment
3083 where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD)
3084 where device nodes cannot be created the effect of PrivateDevices= is
3085 bypassed (following the logic that 2nd-level sand-boxing is not
3086 essential if the system systemd runs in is itself already sand-boxed
3087 as a whole). This logic breaks with 4.18 in container managers where
3088 user namespacing is used: suddenly PrivateDevices= succeeds setting
3089 up a private /dev/ file system containing devices nodes — but when
3090 these are opened they don't work.
3092 At this point it is recommended that container managers utilizing
3093 user namespaces that intend to run systemd in the payload explicitly
3094 block mknod() with seccomp or similar, so that the graceful fallback
3097 We are very sorry for the breakage and the requirement to change
3098 container configurations for newer kernels. It's purely caused by an
3099 incompatible kernel change. The relevant kernel developers have been
3100 notified about this userspace breakage quickly, but they chose to
3103 * PermissionsStartOnly= setting is deprecated (but is still supported
3104 for backwards compatibility). The same functionality is provided by
3105 the more flexible "+", "!", and "!!" prefixes to ExecStart= and other
3108 * $DBUS_SESSION_BUS_ADDRESS environment variable is not set by
3109 pam_systemd anymore.
3111 * The naming scheme for network devices was changed to always rename
3112 devices, even if they were already renamed by userspace. The "kernel"
3113 policy was changed to only apply as a fallback, if no other naming
3116 * The requirements to build systemd is bumped to meson-0.46 and
3119 Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander
3120 Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson,
3121 Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov,
3122 asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt
3123 Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen
3124 Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius
3125 Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn
3126 Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner,
3127 David Anderson, Davide Cavalca, David Leeds, David Malcolm, David
3128 Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus,
3129 Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters,
3130 Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad,
3131 Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank
3132 Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe
3133 Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit
3134 Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan
3135 Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld,
3136 javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi
3137 Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens,
3138 Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi,
3139 Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry
3140 Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz
3141 Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier,
3142 Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin
3143 Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko
3144 Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck,
3145 Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich,
3146 Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal
3147 Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal
3148 Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby,
3149 Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł
3150 Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller,
3151 Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez,
3152 Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam
3153 Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher,
3154 Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee
3155 (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen
3156 Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim,
3157 Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas
3158 Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias
3159 Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore
3160 Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech
3161 Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward,
3162 Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe,
3163 Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein
3165 — Warsaw, 2018-12-21
3169 * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
3170 builtin will name network interfaces differently than in previous
3171 versions for virtual network interfaces created with SR-IOV and NPAR
3172 and for devices where the PCI network controller device does not have
3173 a slot number associated.
3175 SR-IOV virtual devices are now named based on the name of the parent
3176 interface, with a suffix of "v<N>", where <N> is the virtual device
3177 number. Previously those virtual devices were named as if completely
3180 The ninth and later NPAR virtual devices will be named following the
3181 scheme used for the first eight NPAR partitions. Previously those
3182 devices were not renamed and the kernel default (eth<n>) was used.
3184 "net_id" will also generate names for PCI devices where the PCI
3185 network controller device does not have an associated slot number
3186 itself, but one of its parents does. Previously those devices were
3187 not renamed and the kernel default (eth<n>) was used.
3189 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
3190 systemd-logind.service. Since v235, IPAddressDeny=any has been set to
3191 the unit. So, it is expected that the default behavior of
3192 systemd-logind is not changed. However, if distribution packagers or
3193 administrators disabled or modified IPAddressDeny= setting by a
3194 drop-in config file, then it may be necessary to update the file to
3195 re-enable AF_INET and AF_INET6 to support network user name services,
3198 * When the RestrictNamespaces= unit property is specified multiple
3199 times, then the specified types are merged now. Previously, only the
3200 last assignment was used. So, if distribution packagers or
3201 administrators modified the setting by a drop-in config file, then it
3202 may be necessary to update the file.
3204 * When OnFailure= is used in combination with Restart= on a service
3205 unit, then the specified units will no longer be triggered on
3206 failures that result in restarting. Previously, the specified units
3207 would be activated each time the unit failed, even when the unit was
3208 going to be restarted automatically. This behaviour contradicted the
3209 documentation. With this release the code is adjusted to match the
3212 * systemd-tmpfiles will now print a notice whenever it encounters
3213 tmpfiles.d/ lines referencing the /var/run/ directory. It will
3214 recommend reworking them to use the /run/ directory instead (for
3215 which /var/run/ is simply a symlinked compatibility alias). This way
3216 systemd-tmpfiles can properly detect line conflicts and merge lines
3217 referencing the same file by two paths, without having to access
3220 * systemctl disable/unmask/preset/preset-all cannot be used with
3221 --runtime. Previously this was allowed, but resulted in unintuitive
3222 behaviour that wasn't useful. systemctl disable/unmask will now undo
3223 both runtime and persistent enablement/masking, i.e. it will remove
3224 any relevant symlinks both in /run and /etc.
3226 * Note that all long-running system services shipped with systemd will
3227 now default to a system call allow list (rather than a deny list, as
3228 before). In particular, systemd-udevd will now enforce one too. For
3229 most cases this should be safe, however downstream distributions
3230 which disabled sandboxing of systemd-udevd (specifically the
3231 MountFlags= setting), might want to disable this security feature
3232 too, as the default allow-listing will prohibit all mount, swap,
3233 reboot and clock changing operations from udev rules.
3235 * sd-boot acquired new loader configuration settings to optionally turn
3236 off Windows and MacOS boot partition discovery as well as
3237 reboot-into-firmware menu items. It is also able to pick a better
3238 screen resolution for HiDPI systems, and now provides loader
3239 configuration settings to change the resolution explicitly.
3241 * systemd-resolved now supports DNS-over-TLS. It's still
3242 turned off by default, use DNSOverTLS=opportunistic to turn it on in
3243 resolved.conf. We intend to make this the default as soon as couple
3244 of additional techniques for optimizing the initial latency caused by
3245 establishing a TLS/TCP connection are implemented.
3247 * systemd-resolved.service and systemd-networkd.service now set
3248 DynamicUser=yes. The users systemd-resolve and systemd-network are
3249 not created by systemd-sysusers anymore.
3251 NOTE: This has a chance of breaking nss-ldap and similar NSS modules
3252 that embed a network facing module into any process using getpwuid()
3253 or related call: the dynamic allocation of the user ID for
3254 systemd-resolved.service means the service manager has to check NSS
3255 if the user name is already taken when forking off the service. Since
3256 the user in the common case won't be defined in /etc/passwd the
3257 lookup is likely to trigger nss-ldap which in turn might use NSS to
3258 ask systemd-resolved for hostname lookups. This will hence result in
3259 a deadlock: a user name lookup in order to start
3260 systemd-resolved.service will result in a hostname lookup for which
3261 systemd-resolved.service needs to be started already. There are
3262 multiple ways to work around this problem: pre-allocate the
3263 "systemd-resolve" user on such systems, so that nss-ldap won't be
3264 triggered; or use a different NSS package that doesn't do networking
3265 in-process but provides a local asynchronous name cache; or configure
3266 the NSS package to avoid lookups for UIDs in the range `pkg-config
3267 systemd --variable=dynamicuidmin` … `pkg-config systemd
3268 --variable=dynamicuidmax`, so that it does not consider itself
3269 authoritative for the same UID range systemd allocates dynamic users
3272 * The systemd-resolve tool has been renamed to resolvectl (it also
3273 remains available under the old name, for compatibility), and its
3274 interface is now verb-based, similar in style to the other <xyz>ctl
3275 tools, such as systemctl or loginctl.
3277 * The resolvectl/systemd-resolve tool also provides 'resolvconf'
3278 compatibility. It may be symlinked under the 'resolvconf' name, in
3279 which case it will take arguments and input compatible with the
3280 Debian and FreeBSD resolvconf tool.
3282 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
3283 where the system initially suspends, and after a timeout resumes and
3286 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
3287 set the client will only send a DUID as client identifier.
3289 * The nss-systemd glibc NSS module will now enumerate dynamic users and
3290 groups in effect. Previously, it could resolve UIDs/GIDs to user
3291 names/groups and vice versa, but did not support enumeration.
3293 * journald's Compress= configuration setting now optionally accepts a
3294 byte threshold value. All journal objects larger than this threshold
3295 will be compressed, smaller ones will not. Previously this threshold
3296 was not configurable and set to 512.
3298 * A new system.conf setting NoNewPrivileges= is now available which may
3299 be used to turn off acquisition of new privileges system-wide
3300 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
3301 for all its children). Note that turning this option on means setuid
3302 binaries and file system capabilities lose their special powers.
3303 While turning on this option is a big step towards a more secure
3304 system, doing so is likely to break numerous pre-existing UNIX tools,
3305 in particular su and sudo.
3307 * A new service systemd-time-sync-wait.service has been added. If
3308 enabled it will delay the time-sync.target unit at boot until time
3309 synchronization has been received from the network. This
3310 functionality is useful on systems lacking a local RTC or where it is
3311 acceptable that the boot process shall be delayed by external network
3314 * When hibernating, systemd will now inform the kernel of the image
3315 write offset, on kernels new enough to support this. This means swap
3316 files should work for hibernation now.
3318 * When loading unit files, systemd will now look for drop-in unit files
3319 extensions in additional places. Previously, for a unit file name
3320 "foo-bar-baz.service" it would look for dropin files in
3321 "foo-bar-baz.service.d/*.conf". Now, it will also look in
3322 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
3323 service name truncated after all inner dashes. This scheme allows
3324 writing drop-ins easily that apply to a whole set of unit files at
3325 once. It's particularly useful for mount and slice units (as their
3326 naming is prefix based), but is also useful for service and other
3327 units, for packages that install multiple unit files at once,
3328 following a strict naming regime of beginning the unit file name with
3329 the package's name. Two new specifiers are now supported in unit
3330 files to match this: %j and %J are replaced by the part of the unit
3331 name following the last dash.
3333 * Unit files and other configuration files that support specifier
3334 expansion now understand another three new specifiers: %T and %V will
3335 resolve to /tmp and /var/tmp respectively, or whatever temporary
3336 directory has been set for the calling user. %E will expand to either
3337 /etc (for system units) or $XDG_CONFIG_HOME (for user units).
3339 * The ExecStart= lines of unit files are no longer required to
3340 reference absolute paths. If non-absolute paths are specified the
3341 specified binary name is searched within the service manager's
3342 built-in $PATH, which may be queried with 'systemd-path
3343 search-binaries-default'. It's generally recommended to continue to
3344 use absolute paths for all binaries specified in unit files.
3346 * Units gained a new load state "bad-setting", which is used when a
3347 unit file was loaded, but contained fatal errors which prevent it
3348 from being started (for example, a service unit has been defined
3349 lacking both ExecStart= and ExecStop= lines).
3351 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
3352 support alternative debuggers, for example lldb. The old name
3353 continues to be available however, for compatibility reasons. Use the
3354 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
3355 to pick an alternative debugger instead of the default gdb.
3357 * systemctl and the other tools will now output escape sequences that
3358 generate proper clickable hyperlinks in various terminal emulators
3359 where useful (for example, in the "systemctl status" output you can
3360 now click on the unit file name to quickly open it in the
3361 editor/viewer of your choice). Note that not all terminal emulators
3362 support this functionality yet, but many do. Unfortunately, the
3363 "less" pager doesn't support this yet, hence this functionality is
3364 currently automatically turned off when a pager is started (which
3365 happens quite often due to auto-paging). We hope to remove this
3366 limitation as soon as "less" learns these escape sequences. This new
3367 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
3368 environment variable. For details on these escape sequences see:
3369 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
3371 * networkd's .network files now support a new IPv6MTUBytes= option for
3372 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
3373 option in the [Route] section to configure the MTU to use for
3374 specific routes. It also gained support for configuration of the DHCP
3375 "UserClass" option through the new UserClass= setting. It gained
3376 three new options in the new [CAN] section for configuring CAN
3377 networks. The MULTICAST and ALLMULTI interface flags may now be
3378 controlled explicitly with the new Multicast= and AllMulticast=
3381 * networkd will now automatically make use of the kernel's route
3382 expiration feature, if it is available.
3384 * udevd's .link files now support setting the number of receive and
3385 transmit channels, using the RxChannels=, TxChannels=,
3386 OtherChannels=, CombinedChannels= settings.
3388 * Support for UDPSegmentationOffload= has been removed, given its
3389 limited support in hardware, and waning software support.
3391 * networkd's .netdev files now support creating "netdevsim" interfaces.
3393 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
3394 to query the unit belonging to a specific kernel control group.
3396 * systemd-analyze gained a new verb "cat-config", which may be used to
3397 dump the contents of any configuration file, with all its matching
3398 drop-in files added in, and honouring the usual search and masking
3399 logic applied to systemd configuration files. For example use
3400 "systemd-analyze cat-config systemd/system.conf" to get the complete
3401 system configuration file of systemd how it would be loaded by PID 1
3402 itself. Similar to this, various tools such as systemd-tmpfiles or
3403 systemd-sysusers, gained a new option "--cat-config", which does the
3404 corresponding operation for their own configuration settings. For
3405 example, "systemd-tmpfiles --cat-config" will now output the full
3406 list of tmpfiles.d/ lines in place.
3408 * timedatectl gained three new verbs: "show" shows bus properties of
3409 systemd-timedated, "timesync-status" shows the current NTP
3410 synchronization state of systemd-timesyncd, and "show-timesync"
3411 shows bus properties of systemd-timesyncd.
3413 * systemd-timesyncd gained a bus interface on which it exposes details
3416 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
3417 understood by systemd-timedated. It takes a colon-separated list of
3418 unit names of NTP client services. The list is used by
3419 "timedatectl set-ntp".
3421 * systemd-nspawn gained a new --rlimit= switch for setting initial
3422 resource limits for the container payload. There's a new switch
3423 --hostname= to explicitly override the container's hostname. A new
3424 --no-new-privileges= switch may be used to control the
3425 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
3426 --oom-score-adjust= switch controls the OOM scoring adjustment value
3427 for the payload. The new --cpu-affinity= switch controls the CPU
3428 affinity of the container payload. The new --resolv-conf= switch
3429 allows more detailed control of /etc/resolv.conf handling of the
3430 container. Similarly, the new --timezone= switch allows more detailed
3431 control of /etc/localtime handling of the container.
3433 * systemd-detect-virt gained a new --list switch, which will print a
3434 list of all currently known VM and container environments.
3436 * Support for "Portable Services" has been added, see
3437 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
3438 experimental, but this is expected to change soon. Reflecting this
3439 experimental state, the "portablectl" binary is not installed into
3440 /usr/bin yet. The binary has to be called with the full path
3441 /usr/lib/systemd/portablectl instead.
3443 * journalctl's and systemctl's -o switch now knows a new log output
3444 mode "with-unit". The output it generates is very similar to the
3445 regular "short" mode, but displays the unit name instead of the
3446 syslog tag for each log line. Also, the date is shown with timezone
3447 information. This mode is probably more useful than the classic
3448 "short" output mode for most purposes, except where pixel-perfect
3449 compatibility with classic /var/log/messages formatting is required.
3451 * A new --dump-bus-properties switch has been added to the systemd
3452 binary, which may be used to dump all supported D-Bus properties.
3453 (Options which are still supported, but are deprecated, are *not*
3456 * sd-bus gained a set of new calls:
3457 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
3458 enable/disable the "floating" state of a bus slot object,
3459 i.e. whether the slot object pins the bus it is allocated for into
3460 memory or if the bus slot object gets disconnected when the bus goes
3461 away. sd_bus_open_with_description(),
3462 sd_bus_open_user_with_description(),
3463 sd_bus_open_system_with_description() may be used to allocate bus
3464 objects and set their description string already during allocation.
3466 * sd-event gained support for watching inotify events from the event
3467 loop, in an efficient way, sharing inotify handles between multiple
3468 users. For this a new function sd_event_add_inotify() has been added.
3470 * sd-event and sd-bus gained support for calling special user-supplied
3471 destructor functions for userdata pointers associated with
3472 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
3473 functions sd_bus_slot_set_destroy_callback,
3474 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
3475 sd_bus_track_get_destroy_callback,
3476 sd_event_source_set_destroy_callback,
3477 sd_event_source_get_destroy_callback have been added.
3479 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
3481 * PID 1 will now automatically reschedule .timer units whenever the
3482 local timezone changes. (They previously got rescheduled
3483 automatically when the system clock changed.)
3485 * New documentation has been added to document cgroups delegation,
3486 portable services and the various code quality tools we have set up:
3488 https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md
3489 https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md
3490 https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md
3492 * The Boot Loader Specification has been added to the source tree.
3494 https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md
3496 While moving it into our source tree we have updated it and further
3497 changes are now accepted through the usual github PR workflow.
3499 * pam_systemd will now look for PAM userdata fields systemd.memory_max,
3500 systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
3501 earlier PAM modules. The data in these fields is used to initialize
3502 the session scope's resource properties. Thus external PAM modules
3503 may now configure per-session limits, for example sourced from
3504 external user databases.
3506 * socket units with Accept=yes will now maintain a "refused" counter in
3507 addition to the existing "accepted" counter, counting connections
3508 refused due to the enforced limits.
3510 * The "systemd-path search-binaries-default" command may now be use to
3511 query the default, built-in $PATH PID 1 will pass to the services it
3514 * A new unit file setting PrivateMounts= has been added. It's a boolean
3515 option. If enabled the unit's processes are invoked in their own file
3516 system namespace. Note that this behaviour is also implied if any
3517 other file system namespacing options (such as PrivateTmp=,
3518 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
3519 primarily useful for services that do not use any of the other file
3520 system namespacing options. One such service is systemd-udevd.service
3521 where this is now used by default.
3523 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
3524 when the system is booted in UEFI "secure mode".
3526 * A new unit "system-update-pre.target" is added, which defines an
3527 optional synchronization point for offline system updates, as
3528 implemented by the pre-existing "system-update.target" unit. It
3529 allows ordering services before the service that executes the actual
3530 update process in a generic way.
3532 * Systemd now emits warnings whenever .include syntax is used.
3534 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
3535 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
3536 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
3537 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
3538 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
3539 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
3540 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
3541 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
3542 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
3543 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
3544 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
3545 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
3546 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
3547 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
3548 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
3549 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
3550 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
3551 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
3552 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
3553 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
3554 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
3555 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
3556 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
3557 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
3558 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
3559 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
3560 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
3561 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
3562 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
3564 — Berlin, 2018-06-22
3568 * The MemoryAccounting= unit property now defaults to on. After
3569 discussions with the upstream control group maintainers we learnt
3570 that the negative impact of cgroup memory accounting on current
3571 kernels is finally relatively minimal, so that it should be safe to
3572 enable this by default without affecting system performance. Besides
3573 memory accounting only task accounting is turned on by default, all
3574 other forms of resource accounting (CPU, IO, IP) remain off for now,
3575 because it's not clear yet that their impact is small enough to move
3576 from opt-in to opt-out. We recommend downstreams to leave memory
3577 accounting on by default if kernel 4.14 or higher is primarily
3578 used. On very resource constrained systems or when support for old
3579 kernels is a necessity, -Dmemory-accounting-default=false can be used
3580 to revert this change.
3582 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
3583 %udev_rules_update) and the journal catalog (%journal_catalog_update)
3584 from the upgrade scriptlets of individual packages now do nothing.
3585 Transfiletriggers have been added which will perform those updates
3586 once at the end of the transaction.
3588 Similar transfiletriggers have been added to execute any sysctl.d
3589 and binfmt.d rules. Thus, it should be unnecessary to provide any
3590 scriptlets to execute this configuration from package installation
3593 * systemd-sysusers gained a mode where the configuration to execute is
3594 specified on the command line, but this configuration is not executed
3595 directly, but instead it is merged with the configuration on disk,
3596 and the result is executed. This is useful for package installation
3597 scripts which want to create the user before installing any files on
3598 disk (in case some of those files are owned by that user), while
3599 still allowing local admin overrides.
3601 This functionality is exposed to rpm scriptlets through a new
3602 %sysusers_create_package macro. Old %sysusers_create and
3603 %sysusers_create_inline macros are deprecated.
3605 A transfiletrigger for sysusers.d configuration is now installed,
3606 which means that it should be unnecessary to call systemd-sysusers from
3607 package installation scripts, unless the package installs any files
3608 owned by those newly-created users, in which case
3609 %sysusers_create_package should be used.
3611 * Analogous change has been done for systemd-tmpfiles: it gained a mode
3612 where the command-line configuration is merged with the configuration
3613 on disk. This is exposed as the new %tmpfiles_create_package macro,
3614 and %tmpfiles_create is deprecated. A transfiletrigger is installed
3615 for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
3616 from package installation scripts.
3618 * sysusers.d configuration for a user may now also specify the group
3619 number, in addition to the user number ("u username 123:456"), or
3620 without the user number ("u username -:456").
3622 * Configution items for systemd-sysusers can now be specified as
3623 positional arguments when the new --inline switch is used.
3625 * The login shell of users created through sysusers.d may now be
3626 specified (previously, it was always /bin/sh for root and
3627 /sbin/nologin for other users).
3629 * systemd-analyze gained a new --global switch to look at global user
3630 configuration. It also gained a unit-paths verb to list the unit load
3631 paths that are compiled into systemd (which can be used with
3632 --systemd, --user, or --global).
3634 * udevadm trigger gained a new --settle/-w option to wait for any
3635 triggered events to finish (but just those, and not any other events
3636 which are triggered meanwhile).
3638 * The action that systemd-logind takes when the lid is closed and the
3639 machine is connected to external power can now be configured using
3640 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
3641 was determined by HandleLidSwitch=, and, for backwards compatibility,
3642 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
3644 * journalctl will periodically call sd_journal_process() to make it
3645 resilient against inotify queue overruns when journal files are
3646 rotated very quickly.
3648 * Two new functions in libsystemd — sd_bus_get_n_queued_read and
3649 sd_bus_get_n_queued_write — may be used to check the number of
3650 pending bus messages.
3652 * systemd gained a new
3653 org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
3654 which can be used to migrate foreign processes to scope and service
3655 units. The primary user for this new API is systemd itself: the
3656 systemd --user instance uses this call of the systemd --system
3657 instance to migrate processes if it itself gets the request to
3658 migrate processes and the kernel refuses this due to access
3659 restrictions. Thanks to this "systemd-run --scope --user …" works
3660 again in pure cgroup v2 environments when invoked from the user
3663 * A new TemporaryFileSystem= setting can be used to mask out part of
3664 the real file system tree with tmpfs mounts. This may be combined
3665 with BindPaths= and BindReadOnlyPaths= to hide files or directories
3666 not relevant to the unit, while still allowing some paths lower in
3667 the tree to be accessed.
3669 ProtectHome=tmpfs may now be used to hide user home and runtime
3670 directories from units, in a way that is mostly equivalent to
3671 "TemporaryFileSystem=/home /run/user /root".
3673 * Non-service units are now started with KeyringMode=shared by default.
3674 This means that mount and swapon and other mount tools have access
3675 to keys in the main keyring.
3677 * /sys/fs/bpf is now mounted automatically.
3679 * QNX virtualization is now detected by systemd-detect-virt and may
3680 be used in ConditionVirtualization=.
3682 * IPAccounting= may now be enabled also for slice units.
3684 * A new -Dsplit-bin= build configuration switch may be used to specify
3685 whether bin and sbin directories are merged, or if they should be
3686 included separately in $PATH and various listings of executable
3687 directories. The build configuration scripts will try to autodetect
3688 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
3689 system, but distributions are encouraged to configure this
3692 * A new -Dok-color= build configuration switch may be used to change
3693 the colour of "OK" status messages.
3695 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
3696 PrivateNetwork=yes was buggy in previous versions of systemd. This
3697 means that after the upgrade and daemon-reexec, any such units must
3700 * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
3701 will not exclude read-only files owned by root from cleanup.
3703 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
3704 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
3705 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
3706 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
3707 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
3708 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
3709 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
3710 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
3711 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
3712 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
3713 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
3714 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
3715 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
3716 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
3717 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
3718 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
3720 — Warsaw, 2018-03-05
3724 * Some keyboards come with a zoom see-saw or rocker which until now got
3725 mapped to the Linux "zoomin/out" keys in hwdb. However, these
3726 keycodes are not recognized by any major desktop. They now produce
3727 Up/Down key events so that they can be used for scrolling.
3729 * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
3730 slightly: previously, if an argument was specified for lines of this
3731 type (i.e. the right-most column was set) this string was appended to
3732 existing files each time systemd-tmpfiles was run. This behaviour was
3733 different from what the documentation said, and not particularly
3734 useful, as repeated systemd-tmpfiles invocations would not be
3735 idempotent and grow such files without bounds. With this release
3736 behaviour has been altered to match what the documentation says:
3737 lines of this type only have an effect if the indicated files don't
3738 exist yet, and only then the argument string is written to the file.
3740 * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
3741 systemd-tmpfiles behaviour: previously, read-only files owned by root
3742 were always excluded from the file "aging" algorithm (i.e. the
3743 automatic clean-up of directories like /tmp based on
3744 atime/mtime/ctime). We intend to drop this restriction, and age files
3745 by default even when owned by root and read-only. This behaviour was
3746 inherited from older tools, but there have been requests to remove
3747 it, and it's not obvious why this restriction was made in the first
3748 place. Please speak up now, if you are aware of software that reqires
3749 this behaviour, otherwise we'll remove the restriction in v238.
3751 * A new environment variable $SYSTEMD_OFFLINE is now understood by
3752 systemctl. It takes a boolean argument. If on, systemctl assumes it
3753 operates on an "offline" OS tree, and will not attempt to talk to the
3754 service manager. Previously, this mode was implicitly enabled if a
3755 chroot() environment was detected, and this new environment variable
3756 now provides explicit control.
3758 * .path and .socket units may now be created transiently, too.
3759 Previously only service, mount, automount and timer units were
3760 supported as transient units. The systemd-run tool has been updated
3761 to expose this new functionality, you may hence use it now to bind
3762 arbitrary commands to path or socket activation on-the-fly from the
3763 command line. Moreover, almost all properties are now exposed for the
3764 unit types that already supported transient operation.
3766 * The systemd-mount command gained support for a new --owner= parameter
3767 which takes a user name, which is then resolved and included in uid=
3768 and gid= mount options string of the file system to mount.
3770 * A new unit condition ConditionControlGroupController= has been added
3771 that checks whether a specific cgroup controller is available.
3773 * Unit files, udev's .link files, and systemd-networkd's .netdev and
3774 .network files all gained support for a new condition
3775 ConditionKernelVersion= for checking against specific kernel
3778 * In systemd-networkd, the [IPVLAN] section in .netdev files gained
3779 support for configuring device flags in the Flags= setting. In the
3780 same files, the [Tunnel] section gained support for configuring
3781 AllowLocalRemote=. The [Route] section in .network files gained
3782 support for configuring InitialCongestionWindow=,
3783 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
3784 understands RapidCommit=.
3786 * systemd-networkd's DHCPv6 support gained support for Prefix
3789 * sd-bus gained support for a new "watch-bind" feature. When this
3790 feature is enabled, an sd_bus connection may be set up to connect to
3791 an AF_UNIX socket in the file system as soon as it is created. This
3792 functionality is useful for writing early-boot services that
3793 automatically connect to the system bus as soon as it is started,
3794 without ugly time-based polling. systemd-networkd and
3795 systemd-resolved have been updated to make use of this
3796 functionality. busctl exposes this functionality in a new
3797 --watch-bind= command line switch.
3799 * sd-bus will now optionally synthesize a local "Connected" signal as
3800 soon as a D-Bus connection is set up fully. This message mirrors the
3801 already existing "Disconnected" signal which is synthesized when the
3802 connection is terminated. This signal is generally useful but
3803 particularly handy in combination with the "watch-bind" feature
3804 described above. Synthesizing of this message has to be requested
3805 explicitly through the new API call sd_bus_set_connected_signal(). In
3806 addition a new call sd_bus_is_ready() has been added that checks
3807 whether a connection is fully set up (i.e. between the "Connected" and
3808 "Disconnected" signals).
3810 * sd-bus gained two new calls sd_bus_request_name_async() and
3811 sd_bus_release_name_async() for asynchronously registering bus
3812 names. Similar, there is now sd_bus_add_match_async() for installing
3813 a signal match asynchronously. All of systemd's own services have
3814 been updated to make use of these calls. Doing these operations
3815 asynchronously has two benefits: it reduces the risk of deadlocks in
3816 case of cyclic dependencies between bus services, and it speeds up
3817 service initialization since synchronization points for bus
3818 round-trips are removed.
3820 * sd-bus gained two new calls sd_bus_match_signal() and
3821 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
3822 and sd_bus_add_match_async() but instead of taking a D-Bus match
3823 string take match fields as normal function parameters.
3825 * sd-bus gained two new calls sd_bus_set_sender() and
3826 sd_bus_message_set_sender() for setting the sender name of outgoing
3827 messages (either for all outgoing messages or for just one specific
3828 one). These calls are only useful in direct connections as on
3829 brokered connections the broker fills in the sender anyway,
3830 overwriting whatever the client filled in.
3832 * sd-event gained a new pseudo-handle that may be specified on all API
3833 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
3834 used this refers to the default event loop object of the calling
3835 thread. Note however that this does not implicitly allocate one —
3836 which has to be done prior by using sd_event_default(). Similarly
3837 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
3838 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
3839 to the default bus of the specified type of the calling thread. Here
3840 too this does not implicitly allocate bus connection objects, this
3841 has to be done prior with sd_bus_default() and friends.
3843 * sd-event gained a new call pair
3844 sd_event_source_{get|set}_io_fd_own(). This may be used to request
3845 automatic closure of the file descriptor an IO event source watches
3846 when the event source is destroyed.
3848 * systemd-networkd gained support for natively configuring WireGuard
3851 * In previous versions systemd synthesized user records both for the
3852 "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
3853 internally. In order to simplify distribution-wide renames of the
3854 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
3855 new transitional flag file has been added: if
3856 /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
3857 user and group record within the systemd codebase is disabled.
3859 * systemd-notify gained a new --uid= option for selecting the source
3860 user/UID to use for notification messages sent to the service
3863 * journalctl gained a new --grep= option to list only entries in which
3864 the message matches a certain pattern. By default matching is case
3865 insensitive if the pattern is lowercase, and case sensitive
3866 otherwise. Option --case-sensitive=yes|no can be used to override
3867 this an specify case sensitivity or case insensitivity.
3869 * There's now a "systemd-analyze service-watchdogs" command for printing
3870 the current state of the service runtime watchdog, and optionally
3871 enabling or disabling the per-service watchdogs system-wide if given a
3872 boolean argument (i.e. the concept you configure in WatchdogSec=), for
3873 debugging purposes. There's also a kernel command line option
3874 systemd.service_watchdogs= for controlling the same.
3876 * Two new "log-level" and "log-target" options for systemd-analyze were
3877 added that merge the now deprecated get-log-level, set-log-level and
3878 get-log-target, set-log-target pairs. The deprecated options are still
3879 understood for backwards compatibility. The two new options print the
3880 current value when no arguments are given, and set them when a
3881 level/target is given as an argument.
3883 * sysusers.d's "u" lines now optionally accept both a UID and a GID
3884 specification, separated by a ":" character, in order to create users
3885 where UID and GID do not match.
3887 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
3888 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
3889 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
3890 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
3891 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
3892 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
3893 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
3894 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
3895 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
3896 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
3897 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
3898 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
3899 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
3900 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
3901 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
3902 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
3903 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
3904 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
3905 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
3906 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
3913 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
3914 in v235 has been extended to also set the dummy.ko module option
3915 numdummies=0, preventing the kernel from automatically creating
3916 dummy0. All dummy interfaces must now be explicitly created.
3918 * Unknown '%' specifiers in configuration files are now rejected. This
3919 applies to units and tmpfiles.d configuration. Any percent characters
3920 that are followed by a letter or digit that are not supposed to be
3921 interpreted as the beginning of a specifier should be escaped by
3922 doubling ("%%"). (So "size=5%" is still accepted, as well as
3923 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
3924 valid specifiers today.)
3926 * systemd-resolved now maintains a new dynamic
3927 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
3928 recommended to make /etc/resolv.conf a symlink to it. This file
3929 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
3930 includes dynamically acquired search domains, achieving more correct
3931 DNS resolution by software that bypasses local DNS APIs such as NSS.
3933 * The "uaccess" udev tag has been dropped from /dev/kvm and
3934 /dev/dri/renderD*. These devices now have the 0666 permissions by
3935 default (but this may be changed at build-time). /dev/dri/renderD*
3936 will now be owned by the "render" group along with /dev/kfd.
3938 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
3939 systemd-journal-gatewayd.service and
3940 systemd-journal-upload.service. This means "nss-systemd" must be
3941 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
3942 services are resolved properly.
3944 * In /etc/fstab two new mount options are now understood:
3945 x-systemd.makefs and x-systemd.growfs. The former has the effect that
3946 the configured file system is formatted before it is mounted, the
3947 latter that the file system is resized to the full block device size
3948 after it is mounted (i.e. if the file system is smaller than the
3949 partition it resides on, it's grown). This is similar to the fsck
3950 logic in /etc/fstab, and pulls in systemd-makefs@.service and
3951 systemd-growfs@.service as necessary, similar to
3952 systemd-fsck@.service. Resizing is currently only supported on ext4
3955 * In systemd-networkd, the IPv6 RA logic now optionally may announce
3956 DNS server and domain information.
3958 * Support for the LUKS2 on-disk format for encrypted partitions has
3959 been added. This requires libcryptsetup2 during compilation and
3962 * The systemd --user instance will now signal "readiness" when its
3963 basic.target unit has been reached, instead of when the run queue ran
3964 empty for the first time.
3966 * Tmpfiles.d with user configuration are now also supported.
3967 systemd-tmpfiles gained a new --user switch, and snippets placed in
3968 ~/.config/user-tmpfiles.d/ and corresponding directories will be
3969 executed by systemd-tmpfiles --user running in the new
3970 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
3971 running in the user session.
3973 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
3974 %S resolves to the top-level state directory (/var/lib for the system
3975 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
3976 top-level cache directory (/var/cache for the system instance,
3977 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
3978 logs directory (/var/log for the system instance,
3979 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
3980 existing %t specifier, that resolves to the top-level runtime
3981 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
3984 * journalctl learnt a new parameter --output-fields= for limiting the
3985 set of journal fields to output in verbose and JSON output modes.
3987 * systemd-timesyncd's configuration file gained a new option
3988 RootDistanceMaxSec= for setting the maximum root distance of servers
3989 it'll use, as well as the new options PollIntervalMinSec= and
3990 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
3992 * bootctl gained a new command "list" for listing all available boot
3993 menu items on systems that follow the boot loader specification.
3995 * systemctl gained a new --dry-run switch that shows what would be done
3996 instead of doing it, and is currently supported by the shutdown and
3999 * ConditionSecurity= can now detect the TOMOYO security module.
4001 * Unit file [Install] sections are now also respected in unit drop-in
4002 files. This is intended to be used by drop-ins under /usr/lib/.
4004 * systemd-firstboot may now also set the initial keyboard mapping.
4006 * Udev "changed" events for devices which are exposed as systemd
4007 .device units are now propagated to units specified in
4008 ReloadPropagatedFrom= as reload requests.
4010 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
4011 unit template name (i.e. a name in the form of 'foobar@.service',
4012 without the instance component between the '@' and - the '.'), then
4013 the escaped sysfs path of the device is automatically used as the
4016 * SystemCallFilter= in unit files has been extended so that an "errno"
4017 can be specified individually for each system call. Example:
4018 SystemCallFilter=~uname:EILSEQ.
4020 * The cgroup delegation logic has been substantially updated. Delegate=
4021 now optionally takes a list of controllers (instead of a boolean, as
4022 before), which lists the controllers to delegate at least.
4024 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
4026 * A new LogLevelMax= setting configures the maximum log level any
4027 process of the service may log at (i.e. anything with a lesser
4028 priority than what is specified is automatically dropped). A new
4029 LogExtraFields= setting allows configuration of additional journal
4030 fields to attach to all log records generated by any of the unit's
4033 * New StandardInputData= and StandardInputText= settings along with the
4034 new option StandardInput=data may be used to configure textual or
4035 binary data that shall be passed to the executed service process via
4036 standard input, encoded in-line in the unit file.
4038 * StandardInput=, StandardOutput= and StandardError= may now be used to
4039 connect stdin/stdout/stderr of executed processes directly with a
4040 file or AF_UNIX socket in the file system, using the new "file:" option.
4042 * A new unit file option CollectMode= has been added, that allows
4043 tweaking the garbage collection logic for units. It may be used to
4044 tell systemd to garbage collect units that have failed automatically
4045 (normally it only GCs units that exited successfully). systemd-run
4046 and systemd-mount expose this new functionality with a new -G option.
4048 * "machinectl bind" may now be used to bind mount non-directories
4049 (i.e. regularfiles, devices, fifos, sockets).
4051 * systemd-analyze gained a new verb "calendar" for validating and
4052 testing calendar time specifications to use for OnCalendar= in timer
4053 units. Besides validating the expression it will calculate the next
4054 time the specified expression would elapse.
4056 * In addition to the pre-existing FailureAction= unit file setting
4057 there's now SuccessAction=, for configuring a shutdown action to
4058 execute when a unit completes successfully. This is useful in
4059 particular inside containers that shall terminate after some workload
4060 has been completed. Also, both options are now supported for all unit
4061 types, not just services.
4063 * networkds's IP rule support gained two new options
4064 IncomingInterface= and OutgoingInterface= for configuring the incoming
4065 and outgoing interfaces of configured rules. systemd-networkd also
4066 gained support for "vxcan" network devices.
4068 * networkd gained a new setting RequiredForOnline=, taking a
4069 boolean. If set, systemd-wait-online will take it into consideration
4070 when determining that the system is up, otherwise it will ignore the
4071 interface for this purpose.
4073 * The sd_notify() protocol gained support for a new operation: with
4074 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
4075 store again, ahead of POLLHUP or POLLERR when they are removed
4078 * A new document doc/UIDS-GIDS.md has been added to the source tree,
4079 that documents the UID/GID range and assignment assumptions and
4080 requirements of systemd.
4082 * The watchdog device PID 1 will ping may now be configured through the
4083 WatchdogDevice= configuration file setting, or by setting the
4084 systemd.watchdog_service= kernel commandline option.
4086 * systemd-resolved's gained support for registering DNS-SD services on
4087 the local network using MulticastDNS. Services may either be
4088 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
4089 the same dir below /run, /usr/lib), or through its D-Bus API.
4091 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
4092 extend the effective start, runtime, and stop time. The service must
4093 continue to send EXTEND_TIMEOUT_USEC within the period specified to
4094 prevent the service manager from making the service as timedout.
4096 * systemd-resolved's DNSSEC support gained support for RFC 8080
4097 (Ed25519 keys and signatures).
4099 * The systemd-resolve command line tool gained a new set of options
4100 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
4101 --set-nta= and --revert to configure per-interface DNS configuration
4102 dynamically during runtime. It's useful for pushing DNS information
4103 into systemd-resolved from DNS hook scripts that various interface
4104 managing software supports (such as pppd).
4106 * systemd-nspawn gained a new --network-namespace-path= command line
4107 option, which may be used to make a container join an existing
4108 network namespace, by specifying a path to a "netns" file.
4110 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
4111 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
4112 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
4113 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
4114 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
4115 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
4116 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
4117 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
4118 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
4119 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
4120 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
4121 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
4122 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
4123 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
4124 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
4125 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
4126 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
4127 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
4128 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
4129 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
4130 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
4131 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
4132 Jędrzejewski-Szmek, Zeal Jagannatha
4134 — Berlin, 2017-12-14
4138 * INCOMPATIBILITY: systemd-logind.service and other long-running
4139 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
4140 communication with the outside. This generally improves security of
4141 the system, and is in almost all cases a safe and good choice, as
4142 these services do not and should not provide any network-facing
4143 functionality. However, systemd-logind uses the glibc NSS API to
4144 query the user database. This creates problems on systems where NSS
4145 is set up to directly consult network services for user database
4146 lookups. In particular, this creates incompatibilities with the
4147 "nss-nis" module, which attempts to directly contact the NIS/YP
4148 network servers it is configured for, and will now consistently
4149 fail. In such cases, it is possible to turn off IP sandboxing for
4150 systemd-logind.service (set IPAddressDeny= in its [Service] section
4151 to the empty string, via a .d/ unit file drop-in). Downstream
4152 distributions might want to update their nss-nis packaging to include
4153 such a drop-in snippet, accordingly, to hide this incompatibility
4154 from the user. Another option is to make use of glibc's nscd service
4155 to proxy such network requests through a privilege-separated, minimal
4156 local caching daemon, or to switch to more modern technologies such
4157 sssd, whose NSS hook-ups generally do not involve direct network
4158 access. In general, we think it's definitely time to question the
4159 implementation choices of nss-nis, i.e. whether it's a good idea
4160 today to embed a network-facing loadable module into all local
4161 processes that need to query the user database, including the most
4162 trivial and benign ones, such as "ls". For more details about
4163 IPAddressDeny= see below.
4165 * A new modprobe.d drop-in is now shipped by default that sets the
4166 bonding module option max_bonds=0. This overrides the kernel default,
4167 to avoid conflicts and ambiguity as to whether or not bond0 should be
4168 managed by systemd-networkd or not. This resolves multiple issues
4169 with bond0 properties not being applied, when bond0 is configured
4170 with systemd-networkd. Distributors may choose to not package this,
4171 however in that case users will be prevented from correctly managing
4172 bond0 interface using systemd-networkd.
4174 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
4175 which print the logging level and target of the system manager. They
4176 complement the existing "set-log-level" and "set-log-target" verbs
4177 used to change those values.
4179 * journald.conf gained a new boolean setting ReadKMsg= which defaults
4180 to on. If turned off kernel log messages will not be read by
4181 systemd-journald or included in the logs. It also gained a new
4182 setting LineMax= for configuring the maximum line length in
4183 STDOUT/STDERR log streams. The new default for this value is 48K, up
4184 from the previous hardcoded 2048.
4186 * A new unit setting RuntimeDirectoryPreserve= has been added, which
4187 allows more detailed control of what to do with a runtime directory
4188 configured with RuntimeDirectory= (i.e. a directory below /run or
4189 $XDG_RUNTIME_DIR) after a unit is stopped.
4191 * The RuntimeDirectory= setting for units gained support for creating
4192 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
4193 one top-level directory.
4195 * Units gained new options StateDirectory=, CacheDirectory=,
4196 LogsDirectory= and ConfigurationDirectory= which are closely related
4197 to RuntimeDirectory= but manage per-service directories below
4198 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
4199 possible to write unit files which when activated automatically gain
4200 properly owned service specific directories in these locations, thus
4201 making unit files self-contained and increasing compatibility with
4202 stateless systems and factory reset where /etc or /var are
4203 unpopulated at boot. Matching these new settings there's also
4204 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
4205 ConfigurationDirectoryMode= for configuring the access mode of these
4206 directories. These settings are particularly useful in combination
4207 with DynamicUser=yes as they provide secure, properly-owned,
4208 writable, and stateful locations for storage, excluded from the
4209 sandbox that such services live in otherwise.
4211 * Automake support has been removed from this release. systemd is now
4214 * systemd-journald will now aggressively cache client metadata during
4215 runtime, speeding up log write performance under pressure. This comes
4216 at a small price though: as much of the metadata is read
4217 asynchronously from /proc/ (and isn't implicitly attached to log
4218 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
4219 metadata stored alongside a log entry might be slightly
4220 out-of-date. Previously it could only be slightly newer than the log
4221 message. The time window is small however, and given that the kernel
4222 is unlikely to be improved anytime soon in this regard, this appears
4225 * nss-myhostname/systemd-resolved will now by default synthesize an
4226 A/AAAA resource record for the "_gateway" hostname, pointing to the
4227 current default IP gateway. Previously it did that for the "gateway"
4228 name, hampering adoption, as some distributions wanted to leave that
4229 hostname open for local use. The old behaviour may still be
4230 requested at build time.
4232 * systemd-networkd's [Address] section in .network files gained a new
4233 Scope= setting for configuring the IP address scope. The [Network]
4234 section gained a new boolean setting ConfigureWithoutCarrier= that
4235 tells systemd-networkd to ignore link sensing when configuring the
4236 device. The [DHCP] section gained a new Anonymize= boolean option for
4237 turning on a number of options suggested in RFC 7844. A new
4238 [RoutingPolicyRule] section has been added for configuring the IP
4239 routing policy. The [Route] section has gained support for a new
4240 Type= setting which permits configuring
4241 blackhole/unreachable/prohibit routes.
4243 * The [VRF] section in .netdev files gained a new Table= setting for
4244 configuring the routing table to use. The [Tunnel] section gained a
4245 new Independent= boolean field for configuring tunnels independent of
4246 an underlying network interface. The [Bridge] section gained a new
4247 GroupForwardMask= option for configuration of propagation of link
4248 local frames between bridge ports.
4250 * The WakeOnLan= setting in .link files gained support for a number of
4251 new modes. A new TCP6SegmentationOffload= setting has been added for
4252 configuring TCP/IPv6 hardware segmentation offload.
4254 * The IPv6 RA sender implementation may now optionally send out RDNSS
4255 and RDNSSL records to supply DNS configuration to peers.
4257 * systemd-nspawn gained support for a new --system-call-filter= command
4258 line option for adding and removing entries in the default system
4259 call filter it applies. Moreover systemd-nspawn has been changed to
4260 implement a system call allow list instead of a deny list.
4262 * systemd-run gained support for a new --pipe command line option. If
4263 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
4264 are directly passed on to the activated transient service
4265 executable. This allows invoking arbitrary processes as systemd
4266 services (for example to take benefit of dependency management,
4267 accounting management, resource management or log management that is
4268 done automatically for services) — while still allowing them to be
4269 integrated in a classic UNIX shell pipeline.
4271 * When a service sends RELOAD=1 via sd_notify() and reload propagation
4272 using ReloadPropagationTo= is configured, a reload is now propagated
4273 to configured units. (Previously this was only done on explicitly
4274 requested reloads, using "systemctl reload" or an equivalent
4277 * For each service unit a restart counter is now kept: it is increased
4278 each time the service is restarted due to Restart=, and may be
4279 queried using "systemctl show -p NRestarts …".
4281 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
4282 @signal and @timer have been added, for usage with SystemCallFilter=
4283 in unit files and the new --system-call-filter= command line option
4284 of systemd-nspawn (see above).
4286 * ExecStart= lines in unit files gained two new modifiers: when a
4287 command line is prefixed with "!" the command will be executed as
4288 configured, except for the credentials applied by
4289 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
4290 "+", but does still apply namespacing options unlike "+". There's
4291 also "!!" now, which is mostly identical, but becomes a NOP on
4292 systems that support ambient capabilities. This is useful to write
4293 unit files that work with ambient capabilities where possible but
4294 automatically fall back to traditional privilege dropping mechanisms
4295 on systems where this is not supported.
4297 * ListenNetlink= settings in socket units now support RDMA netlink
4300 * A new unit file setting LockPersonality= has been added which permits
4301 locking down the chosen execution domain ("personality") of a service
4304 * A new special target "getty-pre.target" has been added, which is
4305 ordered before all text logins, and may be used to order services
4306 before textual logins acquire access to the console.
4308 * systemd will now attempt to load the virtio-rng.ko kernel module very
4309 early on if a VM environment supporting this is detected. This should
4310 improve entropy during early boot in virtualized environments.
4312 * A _netdev option is now supported in /etc/crypttab that operates in a
4313 similar way as the same option in /etc/fstab: it permits configuring
4314 encrypted devices that need to be ordered after the network is up.
4315 Following this logic, two new special targets
4316 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
4317 added that are to cryptsetup.target what remote-fs.target and
4318 remote-fs-pre.target are to local-fs.target.
4320 * Service units gained a new UnsetEnvironment= setting which permits
4321 unsetting specific environment variables for services that are
4322 normally passed to it (for example in order to mask out locale
4323 settings for specific services that can't deal with it).
4325 * Units acquired a new boolean option IPAccounting=. When turned on, IP
4326 traffic accounting (packet count as well as byte count) is done for
4327 the service, and shown as part of "systemctl status" or "systemd-run
4330 * Service units acquired two new options IPAddressAllow= and
4331 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
4332 for configuring a simple IP access control list for all sockets of
4333 the unit. These options are available also on .slice and .socket
4334 units, permitting flexible access list configuration for individual
4335 services as well as groups of services (as defined by a slice unit),
4336 including system-wide. Note that IP ACLs configured this way are
4337 enforced on every single IPv4 and IPv6 socket created by any process
4338 of the service unit, and apply to ingress as well as egress traffic.
4340 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
4341 structured log message is generated each time the unit is stopped,
4342 containing information about the consumed resources of this
4345 * A new setting KeyringMode= has been added to unit files, which may be
4346 used to control how the kernel keyring is set up for executed
4349 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
4350 "systemctl kexec" and "systemctl exit" are now always asynchronous in
4351 behaviour (that is: these commands return immediately after the
4352 operation was enqueued instead of waiting for the operation to
4353 complete). Previously, "systemctl poweroff" and "systemctl reboot"
4354 were asynchronous on systems using systemd-logind (i.e. almost
4355 always, and like they were on sysvinit), and the other three commands
4356 were unconditionally synchronous. With this release this is cleaned
4357 up, and callers will see the same asynchronous behaviour on all
4358 systems for all five operations.
4360 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
4363 * .timer units now accept calendar specifications in other timezones
4364 than UTC or the local timezone.
4366 * The tmpfiles snippet var.conf has been changed to create
4367 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
4368 the "utmp" group already, and it appears to be generally understood
4369 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
4370 databases. Previously this was implemented correctly for all these
4371 databases excepts btmp, which has been opened up like this now
4372 too. Note that while the other databases are world-readable
4373 (i.e. 0644), btmp is not and remains more restrictive.
4375 * The systemd-resolve tool gained a new --reset-server-features
4376 switch. When invoked like this systemd-resolved will forget
4377 everything it learnt about the features supported by the configured
4378 upstream DNS servers, and restarts the feature probing logic on the
4379 next resolver look-up for them at the highest feature level
4382 * The status dump systemd-resolved sends to the logs upon receiving
4383 SIGUSR1 now also includes information about all DNS servers it is
4384 configured to use, and the features levels it probed for them.
4386 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
4387 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
4388 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
4389 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
4390 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
4391 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
4392 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
4393 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
4394 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
4395 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
4396 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
4397 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
4398 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
4399 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
4400 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
4401 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
4402 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
4403 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
4404 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
4405 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
4407 — Berlin, 2017-10-06
4411 * Meson is now supported as build system in addition to Automake. It is
4412 our plan to remove Automake in one of our next releases, so that
4413 Meson becomes our exclusive build system. Hence, please start using
4414 the Meson build system in your downstream packaging. There's plenty
4415 of documentation around how to use Meson, the extremely brief
4418 ./autogen.sh && ./configure && make && sudo make install
4422 meson build && ninja -C build && sudo ninja -C build install
4424 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
4425 which permits configuring a timeout on the time a job is
4426 running. This is particularly useful for setting timeouts on jobs for
4429 * Unit files gained two new options ConditionUser= and ConditionGroup=
4430 for conditionalizing units based on the identity of the user/group
4431 running a systemd user instance.
4433 * systemd-networkd now understands a new FlowLabel= setting in the
4434 [VXLAN] section of .network files, as well as a Priority= in
4435 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
4436 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
4437 gained support for configuration of GENEVE links, and IPv6 address
4438 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
4440 * .link files now understand a new Port= setting.
4442 * systemd-networkd's DHCP support gained support for DHCP option 119
4443 (domain search list).
4445 * systemd-networkd gained support for serving IPv6 address ranges using
4446 the Router Advertisement protocol. The new .network configuration
4447 section [IPv6Prefix] may be used to configure the ranges to
4448 serve. This is implemented based on a new, minimal, native server
4449 implementation of RA.
4451 * journalctl's --output= switch gained support for a new parameter
4452 "short-iso-precise" for a mode where timestamps are shown as precise
4455 * systemd-udevd's "net_id" builtin may now generate stable network
4456 interface names from IBM PowerVM VIO devices as well as ACPI platform
4459 * MulticastDNS support in systemd-resolved may now be explicitly
4460 enabled/disabled using the new MulticastDNS= configuration file
4463 * systemd-resolved may now optionally use libidn2 instead of the libidn
4464 for processing internationalized domain names. Support for libidn2
4465 should be considered experimental and should not be enabled by
4468 * "machinectl pull-tar" and related call may now do verification of
4469 downloaded images using SUSE-style .sha256 checksum files in addition
4470 to the already existing support for validating using Ubuntu-style
4473 * sd-bus gained support for a new sd_bus_message_appendv() call which
4474 is va_list equivalent of sd_bus_message_append().
4476 * sd-boot gained support for validating images using SHIM/MOK.
4478 * The SMACK code learnt support for "onlycap".
4480 * systemd-mount --umount is now much smarter in figuring out how to
4481 properly unmount a device given its mount or device path.
4483 * The code to call libnss_dns as a fallback from libnss_resolve when
4484 the communication with systemd-resolved fails was removed. This
4485 fallback was redundant and interfered with the [!UNAVAIL=return]
4486 suffix. See nss-resolve(8) for the recommended configuration.
4488 * systemd-logind may now be restarted without losing state. It stores
4489 the file descriptors for devices it manages in the system manager
4490 using the FDSTORE= mechanism. Please note that further changes in
4491 other components may be required to make use of this (for example
4492 Xorg has code to listen for stops of systemd-logind and terminate
4493 itself when logind is stopped or restarted, in order to avoid using
4494 stale file descriptors for graphical devices, which is now
4495 counterproductive and must be reverted in order for restarts of
4496 systemd-logind to be safe. See
4497 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
4499 * All kernel-install plugins are called with the environment variable
4500 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
4501 /etc/machine-id. If the machine ID could not be determined,
4502 $KERNEL_INSTALL_MACHINE_ID will be empty. Plugins should not put
4503 anything in the entry directory (passed as the second argument) if
4504 $KERNEL_INSTALL_MACHINE_ID is empty. For backwards compatibility, a
4505 temporary directory is passed as the entry directory and removed
4506 after all the plugins exit.
4508 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
4509 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
4510 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
4511 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
4512 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
4513 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
4514 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
4515 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
4516 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
4517 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
4518 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
4519 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
4520 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
4521 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
4522 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
4523 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
4524 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
4525 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
4526 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
4527 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
4528 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
4529 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
4530 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
4531 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
4532 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
4533 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
4534 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
4535 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
4536 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
4539 — Berlin, 2017-07-12
4543 * The "hybrid" control group mode has been modified to improve
4544 compatibility with "legacy" cgroups-v1 setups. Specifically, the
4545 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
4546 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
4547 cgroups-v1 hierarchy), the only externally visible change being that
4548 the cgroups-v2 hierarchy is also mounted, to
4549 /sys/fs/cgroup/unified. This should provide a large degree of
4550 compatibility with "legacy" cgroups-v1, while taking benefit of the
4551 better management capabilities of cgroups-v2.
4553 * The default control group setup mode may be selected both a boot-time
4554 via a set of kernel command line parameters (specifically:
4555 systemd.unified_cgroup_hierarchy= and
4556 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
4557 default selected on the configure command line
4558 (--with-default-hierarchy=). The upstream default is "hybrid"
4559 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
4560 this will change in a future systemd version to be "unified" (pure
4561 cgroups-v2 mode). The third option for the compile time option is
4562 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
4563 distributions to default to "hybrid" mode for release distributions,
4564 starting with v233. We recommend "unified" for development
4565 distributions (specifically: distributions such as Fedora's rawhide)
4566 as that's where things are headed in the long run. Use "legacy" for
4567 greatest stability and compatibility only.
4569 * Note one current limitation of "unified" and "hybrid" control group
4570 setup modes: the kernel currently does not permit the systemd --user
4571 instance (i.e. unprivileged code) to migrate processes between two
4572 disconnected cgroup subtrees, even if both are managed and owned by
4573 the user. This effectively means "systemd-run --user --scope" doesn't
4574 work when invoked from outside of any "systemd --user" service or
4575 scope. Specifically, it is not supported from session scopes. We are
4576 working on fixing this in a future systemd version. (See #3388 for
4577 further details about this.)
4579 * DBus policy files are now installed into /usr rather than /etc. Make
4580 sure your system has dbus >= 1.9.18 running before upgrading to this
4581 version, or override the install path with --with-dbuspolicydir= .
4583 * All python scripts shipped with systemd (specifically: the various
4584 tests written in Python) now require Python 3.
4586 * systemd unit tests can now run standalone (without the source or
4587 build directories), and can be installed into /usr/lib/systemd/tests/
4588 with 'make install-tests'.
4590 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
4591 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
4594 * Support for the %c, %r, %R specifiers in unit files has been
4595 removed. Specifiers are not supposed to be dependent on configuration
4596 in the unit file itself (so that they resolve the same regardless
4597 where used in the unit files), but these specifiers were influenced
4598 by the Slice= option.
4600 * The shell invoked by debug-shell.service now defaults to /bin/sh in
4601 all cases. If distributions want to use a different shell for this
4602 purpose (for example Fedora's /sbin/sushell) they need to specify
4603 this explicitly at configure time using --with-debug-shell=.
4605 * The confirmation spawn prompt has been reworked to offer the
4608 (c)ontinue, proceed without asking anymore
4609 (D)ump, show the state of the unit
4610 (f)ail, don't execute the command and pretend it failed
4612 (i)nfo, show a short summary of the unit
4613 (j)obs, show jobs that are in progress
4614 (s)kip, don't execute the command and pretend it succeeded
4615 (y)es, execute the command
4617 The 'n' choice for the confirmation spawn prompt has been removed,
4618 because its meaning was confusing.
4620 The prompt may now also be redirected to an alternative console by
4621 specifying the console as parameter to systemd.confirm_spawn=.
4623 * Services of Type=notify require a READY=1 notification to be sent
4624 during startup. If no such message is sent, the service now fails,
4625 even if the main process exited with a successful exit code.
4627 * Services that fail to start up correctly now always have their
4628 ExecStopPost= commands executed. Previously, they'd enter "failed"
4629 state directly, without executing these commands.
4631 * The option MulticastDNS= of network configuration files has acquired
4632 an actual implementation. With MulticastDNS=yes a host can resolve
4633 names of remote hosts and reply to mDNS A and AAAA requests.
4635 * When units are about to be started an additional check is now done to
4636 ensure that all dependencies of type BindsTo= (when used in
4637 combination with After=) have been started.
4639 * systemd-analyze gained a new verb "syscall-filter" which shows which
4640 system call groups are defined for the SystemCallFilter= unit file
4641 setting, and which system calls they contain.
4643 * A new system call filter group "@filesystem" has been added,
4644 consisting of various file system related system calls. Group
4645 "@reboot" has been added, covering reboot, kexec and shutdown related
4646 calls. Finally, group "@swap" has been added covering swap
4647 configuration related calls.
4649 * A new unit file option RestrictNamespaces= has been added that may be
4650 used to restrict access to the various process namespace types the
4651 Linux kernel provides. Specifically, it may be used to take away the
4652 right for a service unit to create additional file system, network,
4653 user, and other namespaces. This sandboxing option is particularly
4654 relevant due to the high amount of recently discovered namespacing
4655 related vulnerabilities in the kernel.
4657 * systemd-udev's .link files gained support for a new AutoNegotiation=
4658 setting for configuring Ethernet auto-negotiation.
4660 * systemd-networkd's .network files gained support for a new
4661 ListenPort= setting in the [DHCP] section to explicitly configure the
4662 UDP client port the DHCP client shall listen on.
4664 * .network files gained a new Unmanaged= boolean setting for explicitly
4665 excluding one or more interfaces from management by systemd-networkd.
4667 * The systemd-networkd ProxyARP= option has been renamed to
4668 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
4669 renamed to ReduceARPProxy=. The old names continue to be available
4672 * systemd-networkd gained support for configuring IPv6 Proxy NDP
4673 addresses via the new IPv6ProxyNDPAddress= .network file setting.
4675 * systemd-networkd's bonding device support gained support for two new
4676 configuration options ActiveSlave= and PrimarySlave=.
4678 * The various options in the [Match] section of .network files gained
4679 support for negative matching.
4681 * New systemd-specific mount options are now understood in /etc/fstab:
4683 x-systemd.mount-timeout= may be used to configure the maximum
4684 permitted runtime of the mount command.
4686 x-systemd.device-bound may be set to bind a mount point to its
4687 backing device unit, in order to automatically remove a mount point
4688 if its backing device is unplugged. This option may also be
4689 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
4690 on the block device, which is now automatically set for all CDROM
4691 drives, so that mounted CDs are automatically unmounted when they are
4692 removed from the drive.
4694 x-systemd.after= and x-systemd.before= may be used to explicitly
4695 order a mount after or before another unit or mount point.
4697 * Enqueued start jobs for device units are now automatically garbage
4698 collected if there are no jobs waiting for them anymore.
4700 * systemctl list-jobs gained two new switches: with --after, for every
4701 queued job the jobs it's waiting for are shown; with --before the
4702 jobs which it's blocking are shown.
4704 * systemd-nspawn gained support for ephemeral boots from disk images
4705 (or in other words: --ephemeral and --image= may now be
4706 combined). Moreover, ephemeral boots are now supported for normal
4707 directories, even if the backing file system is not btrfs. Of course,
4708 if the file system does not support file system snapshots or
4709 reflinks, the initial copy operation will be relatively expensive, but
4710 this should still be suitable for many use cases.
4712 * Calendar time specifications in .timer units now support
4713 specifications relative to the end of a month by using "~" instead of
4714 "-" as separator between month and day. For example, "*-02~03" means
4715 "the third last day in February". In addition a new syntax for
4716 repeated events has been added using the "/" character. For example,
4717 "9..17/2:00" means "every two hours from 9am to 5pm".
4719 * systemd-socket-proxyd gained a new parameter --connections-max= for
4720 configuring the maximum number of concurrent connections.
4722 * sd-id128 gained a new API for generating unique IDs for the host in a
4723 way that does not leak the machine ID. Specifically,
4724 sd_id128_get_machine_app_specific() derives an ID based on the
4725 machine ID a in well-defined, non-reversible, stable way. This is
4726 useful whenever an identifier for the host is needed but where the
4727 identifier shall not be useful to identify the system beyond the
4728 scope of the application itself. (Internally this uses HMAC-SHA256 as
4729 keyed hash function using the machine ID as input.)
4731 * NotifyAccess= gained a new supported value "exec". When set
4732 notifications are accepted from all processes systemd itself invoked,
4733 including all control processes.
4735 * .nspawn files gained support for defining overlay mounts using the
4736 Overlay= and OverlayReadOnly= options. Previously this functionality
4737 was only available on the systemd-nspawn command line.
4739 * systemd-nspawn's --bind= and --overlay= options gained support for
4740 bind/overlay mounts whose source lies within the container tree by
4741 prefixing the source path with "+".
4743 * systemd-nspawn's --bind= and --overlay= options gained support for
4744 automatically allocating a temporary source directory in /var/tmp
4745 that is removed when the container dies. Specifically, if the source
4746 directory is specified as empty string this mechanism is selected. An
4747 example usage is --overlay=+/var::/var, which creates an overlay
4748 mount based on the original /var contained in the image, overlaid
4749 with a temporary directory in the host's /var/tmp. This way changes
4750 to /var are automatically flushed when the container shuts down.
4752 * systemd-nspawn --image= option does now permit raw file system block
4753 devices (in addition to images containing partition tables, as
4756 * The disk image dissection logic in systemd-nspawn gained support for
4757 automatically setting up LUKS encrypted as well as Verity protected
4758 partitions. When a container is booted from an encrypted image the
4759 passphrase is queried at start-up time. When a container with Verity
4760 data is started, the root hash is search in a ".roothash" file
4761 accompanying the disk image (alternatively, pass the root hash via
4762 the new --root-hash= command line option).
4764 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
4765 be used to dissect disk images the same way as systemd-nspawn does
4766 it, following the Bootable Partition Specification. It may even be
4767 used to mount disk images with complex partition setups (including
4768 LUKS and Verity partitions) to a local host directory, in order to
4769 inspect them. This tool is not considered public API (yet), and is
4770 thus not installed into /usr/bin. Please do not rely on its
4771 existence, since it might go away or be changed in later systemd
4774 * A new generator "systemd-verity-generator" has been added, similar in
4775 style to "systemd-cryptsetup-generator", permitting automatic setup of
4776 Verity root partitions when systemd boots up. In order to make use of
4777 this your partition setup should follow the Discoverable Partitions
4778 Specification, and the GPT partition ID of the root file system
4779 partition should be identical to the upper 128bit of the Verity root
4780 hash. The GPT partition ID of the Verity partition protecting it
4781 should be the lower 128bit of the Verity root hash. If the partition
4782 image follows this model it is sufficient to specify a single
4783 "roothash=" kernel command line argument to both configure which root
4784 image and verity partition to use as well as the root hash for
4785 it. Note that systemd-nspawn's Verity support follows the same
4786 semantics, meaning that disk images with proper Verity data in place
4787 may be booted in containers with systemd-nspawn as well as on
4788 physical systems via the verity generator. Also note that the "mkosi"
4789 tool available at https://github.com/systemd/mkosi has been updated
4790 to generate Verity protected disk images following this scheme. In
4791 fact, it has been updated to generate disk images that optionally
4792 implement a complete UEFI SecureBoot trust chain, involving a signed
4793 kernel and initrd image that incorporates such a root hash as well as
4794 a Verity-enabled root partition.
4796 * The hardware database (hwdb) udev supports has been updated to carry
4797 accelerometer quirks.
4799 * All system services are now run with a fresh kernel keyring set up
4800 for them. The invocation ID is stored by default in it, thus
4801 providing a safe, non-overridable way to determine the invocation
4804 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
4805 options for bind mounting arbitrary paths in a service-specific
4806 way. When these options are used, arbitrary host or service files and
4807 directories may be mounted to arbitrary locations in the service's
4810 * Documentation has been added that lists all of systemd's low-level
4811 environment variables:
4813 https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md
4815 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
4816 whether a specific socket file descriptor matches a specified socket
4819 * systemd-firstboot has been updated to check for the
4820 systemd.firstboot= kernel command line option. It accepts a boolean
4821 and when set to false the first boot questions are skipped.
4823 * systemd-fstab-generator has been updated to check for the
4824 systemd.volatile= kernel command line option, which either takes an
4825 optional boolean parameter or the special value "state". If used the
4826 system may be booted in a "volatile" boot mode. Specifically,
4827 "systemd.volatile" is used, the root directory will be mounted as
4828 tmpfs, and only /usr is mounted from the actual root file system. If
4829 "systemd.volatile=state" is used, the root directory will be mounted
4830 as usual, but /var is mounted as tmpfs. This concept provides similar
4831 functionality as systemd-nspawn's --volatile= option, but provides it
4832 on physical boots. Use this option for implementing stateless
4833 systems, or testing systems with all state and/or configuration reset
4834 to the defaults. (Note though that many distributions are not
4835 prepared to boot up without a populated /etc or /var, though.)
4837 * systemd-gpt-auto-generator gained support for LUKS encrypted root
4838 partitions. Previously it only supported LUKS encrypted partitions
4839 for all other uses, except for the root partition itself.
4841 * Socket units gained support for listening on AF_VSOCK sockets for
4842 communication in virtualized QEMU environments.
4844 * The "configure" script gained a new option --with-fallback-hostname=
4845 for specifying the fallback hostname to use if none is configured in
4846 /etc/hostname. For example, by specifying
4847 --with-fallback-hostname=fedora it is possible to default to a
4848 hostname of "fedora" on pristine installations.
4850 * systemd-cgls gained support for a new --unit= switch for listing only
4851 the control groups of a specific unit. Similar --user-unit= has been
4852 added for listing only the control groups of a specific user unit.
4854 * systemd-mount gained a new --umount switch for unmounting a mount or
4855 automount point (and all mount/automount points below it).
4857 * systemd will now refuse full configuration reloads (via systemctl
4858 daemon-reload and related calls) unless at least 16MiB of free space
4859 are available in /run. This is a safety precaution in order to ensure
4860 that generators can safely operate after the reload completed.
4862 * A new unit file option RootImage= has been added, which has a similar
4863 effect as RootDirectory= but mounts the service's root directory from
4864 a disk image instead of plain directory. This logic reuses the same
4865 image dissection and mount logic that systemd-nspawn already uses,
4866 and hence supports any disk images systemd-nspawn supports, including
4867 those following the Discoverable Partition Specification, as well as
4868 Verity enabled images. This option enables systemd to run system
4869 services directly off disk images acting as resource bundles,
4870 possibly even including full integrity data.
4872 * A new MountAPIVFS= unit file option has been added, taking a boolean
4873 argument. If enabled /proc, /sys and /dev (collectively called the
4874 "API VFS") will be mounted for the service. This is only relevant if
4875 RootDirectory= or RootImage= is used for the service, as these mounts
4876 are of course in place in the host mount namespace anyway.
4878 * systemd-nspawn gained support for a new --pivot-root= switch. If
4879 specified the root directory within the container image is pivoted to
4880 the specified mount point, while the original root disk is moved to a
4881 different place. This option enables booting of ostree images
4882 directly with systemd-nspawn.
4884 * The systemd build scripts will no longer complain if the NTP server
4885 addresses are not changed from the defaults. Google now supports
4886 these NTP servers officially. We still recommend downstreams to
4887 properly register an NTP pool with the NTP pool project though.
4889 * coredumpctl gained a new "--reverse" option for printing the list
4890 of coredumps in reverse order.
4892 * coredumpctl will now show additional information about truncated and
4893 inaccessible coredumps, as well as coredumps that are still being
4894 processed. It also gained a new --quiet switch for suppressing
4895 additional informational message in its output.
4897 * coredumpctl gained support for only showing coredumps newer and/or
4898 older than specific timestamps, using the new --since= and --until=
4899 options, reminiscent of journalctl's options by the same name.
4901 * The systemd-coredump logic has been improved so that it may be reused
4902 to collect backtraces in non-compiled languages, for example in
4903 scripting languages such as Python.
4905 * machinectl will now show the UID shift of local containers, if user
4906 namespacing is enabled for them.
4908 * systemd will now optionally run "environment generator" binaries at
4909 configuration load time. They may be used to add environment
4910 variables to the environment block passed to services invoked. One
4911 user environment generator is shipped by default that sets up
4912 environment variables based on files dropped into /etc/environment.d
4913 and ~/.config/environment.d/.
4915 * systemd-resolved now includes the new, recently published 2017 DNSSEC
4918 * hostnamed has been updated to report a new chassis type of
4919 "convertible" to cover "foldable" laptops that can both act as a
4920 tablet and as a laptop, such as various Lenovo Yoga devices.
4922 Contributions from: Adrián López, Alexander Galanin, Alexander
4923 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
4924 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
4925 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
4926 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
4927 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
4928 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
4929 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
4930 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
4931 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
4932 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
4933 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
4934 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
4935 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
4936 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
4937 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
4938 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
4939 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
4940 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
4941 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
4942 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
4943 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
4944 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
4945 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
4946 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
4947 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
4948 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
4949 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
4952 — Berlin, 2017-03-01
4956 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
4957 RestrictAddressFamilies= enabled. These sandboxing options should
4958 generally be compatible with the various external udev call-out
4959 binaries we are aware of, however there may be exceptions, in
4960 particular when exotic languages for these call-outs are used. In
4961 this case, consider turning off these settings locally.
4963 * The new RemoveIPC= option can be used to remove IPC objects owned by
4964 the user or group of a service when that service exits.
4966 * The new ProtectKernelModules= option can be used to disable explicit
4967 load and unload operations of kernel modules by a service. In
4968 addition access to /usr/lib/modules is removed if this option is set.
4970 * ProtectSystem= option gained a new value "strict", which causes the
4971 whole file system tree with the exception of /dev, /proc, and /sys,
4972 to be remounted read-only for a service.
4974 * The new ProtectKernelTunables= option can be used to disable
4975 modification of configuration files in /sys and /proc by a service.
4976 Various directories and files are remounted read-only, so access is
4977 restricted even if the file permissions would allow it.
4979 * The new ProtectControlGroups= option can be used to disable write
4980 access by a service to /sys/fs/cgroup.
4982 * Various systemd services have been hardened with
4983 ProtectKernelTunables=yes, ProtectControlGroups=yes,
4984 RestrictAddressFamilies=.
4986 * Support for dynamically creating users for the lifetime of a service
4987 has been added. If DynamicUser=yes is specified, user and group IDs
4988 will be allocated from the range 61184..65519 for the lifetime of the
4989 service. They can be resolved using the new nss-systemd.so NSS
4990 module. The module must be enabled in /etc/nsswitch.conf. Services
4991 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
4992 any resources allocated by the service will be cleaned up when the
4993 service exits. They also have ProtectHome=read-only and
4994 ProtectSystem=strict enabled, so they are not able to make any
4995 permanent modifications to the system.
4997 * The nss-systemd module also always resolves root and nobody, making
4998 it possible to have no /etc/passwd or /etc/group files in minimal
4999 container or chroot environments.
5001 * Services may be started with their own user namespace using the new
5002 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
5003 under which the service is running are mapped. All other users are
5006 * Support for the cgroup namespace has been added to systemd-nspawn. If
5007 supported by kernel, the container system started by systemd-nspawn
5008 will have its own view of the cgroup hierarchy. This new behaviour
5009 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
5011 * The new MemorySwapMax= option can be used to limit the maximum swap
5012 usage under the unified cgroup hierarchy.
5014 * Support for the CPU controller in the unified cgroup hierarchy has
5015 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
5016 options. This controller requires out-of-tree patches for the kernel
5017 and the support is provisional.
5019 * Mount and automount units may now be created transiently
5020 (i.e. dynamically at runtime via the bus API, instead of requiring
5021 unit files in the file system).
5023 * systemd-mount is a new tool which may mount file systems – much like
5024 mount(8), optionally pulling in additional dependencies through
5025 transient .mount and .automount units. For example, this tool
5026 automatically runs fsck on a backing block device before mounting,
5027 and allows the automount logic to be used dynamically from the
5028 command line for establishing mount points. This tool is particularly
5029 useful when dealing with removable media, as it will ensure fsck is
5030 run – if necessary – before the first access and that the file system
5031 is quickly unmounted after each access by utilizing the automount
5032 logic. This maximizes the chance that the file system on the
5033 removable media stays in a clean state, and if it isn't in a clean
5034 state is fixed automatically.
5036 * LazyUnmount=yes option for mount units has been added to expose the
5037 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
5040 * /efi will be used as the mount point of the EFI boot partition, if
5041 the directory is present, and the mount point was not configured
5042 through other means (e.g. fstab). If /efi directory does not exist,
5043 /boot will be used as before. This makes it easier to automatically
5044 mount the EFI partition on systems where /boot is used for something
5047 * When operating on GPT disk images for containers, systemd-nspawn will
5048 now mount the ESP to /boot or /efi according to the same rules as PID
5049 1 running on a host. This allows tools like "bootctl" to operate
5050 correctly within such containers, in order to make container images
5051 bootable on physical systems.
5053 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
5055 * Two new user session targets have been added to support running
5056 graphical sessions under the systemd --user instance:
5057 graphical-session.target and graphical-session-pre.target. See
5058 systemd.special(7) for a description of how those targets should be
5061 * The vconsole initialization code has been significantly reworked to
5062 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
5063 support unicode keymaps. Font and keymap configuration will now be
5064 copied to all allocated virtual consoles.
5066 * FreeBSD's bhyve virtualization is now detected.
5068 * Information recorded in the journal for core dumps now includes the
5069 contents of /proc/mountinfo and the command line of the process at
5070 the top of the process hierarchy (which is usually the init process
5073 * systemd-journal-gatewayd learned the --directory= option to serve
5074 files from the specified location.
5076 * journalctl --root=… can be used to peruse the journal in the
5077 /var/log/ directories inside of a container tree. This is similar to
5078 the existing --machine= option, but does not require the container to
5081 * The hardware database has been extended to support
5082 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
5085 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
5086 specify the click rate for mice which include a horizontal wheel with
5087 a click rate that is different than the one for the vertical wheel.
5089 * systemd-run gained a new --wait option that makes service execution
5090 synchronous. (Specifically, the command will not return until the
5091 specified service binary exited.)
5093 * systemctl gained a new --wait option that causes the start command to
5094 wait until the units being started have terminated again.
5096 * A new journal output mode "short-full" has been added which displays
5097 timestamps with abbreviated English day names and adds a timezone
5098 suffix. Those timestamps include more information than the default
5099 "short" output mode, and can be passed directly to journalctl's
5100 --since= and --until= options.
5102 * /etc/resolv.conf will be bind-mounted into containers started by
5103 systemd-nspawn, if possible, so any changes to resolv.conf contents
5104 are automatically propagated to the container.
5106 * The number of instances for socket-activated services originating
5107 from a single IP address can be limited with
5108 MaxConnectionsPerSource=, extending the existing setting of
5111 * systemd-networkd gained support for vcan ("Virtual CAN") interface
5114 * .netdev and .network configuration can now be extended through
5117 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
5118 Segmentation Offload, Generic Receive Offload, Large Receive Offload
5119 can be enabled and disabled using the new UDPSegmentationOffload=,
5120 TCPSegmentationOffload=, GenericSegmentationOffload=,
5121 GenericReceiveOffload=, LargeReceiveOffload= options in the
5122 [Link] section of .link files.
5124 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
5125 Port VLAN ID can be configured for bridge devices using the new STP=,
5126 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
5127 section of .netdev files.
5129 * The route table to which routes received over DHCP or RA should be
5130 added can be configured with the new RouteTable= option in the [DHCP]
5131 and [IPv6AcceptRA] sections of .network files.
5133 * The Address Resolution Protocol can be disabled on links managed by
5134 systemd-networkd using the ARP=no setting in the [Link] section of
5137 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
5138 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
5139 encode information about the result and exit codes of the current
5140 service runtime cycle.
5142 * systemd-sysctl will now configure kernel parameters in the order
5143 they occur in the configuration files. This matches what sysctl
5144 has been traditionally doing.
5146 * kernel-install "plugins" that are executed to perform various
5147 tasks after a new kernel is added and before an old one is removed
5148 can now return a special value to terminate the procedure and
5149 prevent any later plugins from running.
5151 * Journald's SplitMode=login setting has been deprecated. It has been
5152 removed from documentation, and its use is discouraged. In a future
5153 release it will be completely removed, and made equivalent to current
5154 default of SplitMode=uid.
5156 * Storage=both option setting in /etc/systemd/coredump.conf has been
5157 removed. With fast LZ4 compression storing the core dump twice is not
5160 * The --share-system systemd-nspawn option has been replaced with an
5161 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
5162 this functionality is discouraged. In addition the variables
5163 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
5164 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
5165 individual namespaces.
5167 * "machinectl list" now shows the IP address of running containers in
5168 the output, as well as OS release information.
5170 * "loginctl list" now shows the TTY of each session in the output.
5172 * sd-bus gained new API calls sd_bus_track_set_recursive(),
5173 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
5174 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
5175 tracking objects in a "recursive" mode, where a single client can be
5176 counted multiple times, if it takes multiple references.
5178 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
5179 sd_bus_get_exit_on_disconnect(). They may be used to make a
5180 process using sd-bus automatically exit if the bus connection is
5183 * Bus clients of the service manager may now "pin" loaded units into
5184 memory, by taking an explicit reference on them. This is useful to
5185 ensure the client can retrieve runtime data about the service even
5186 after the service completed execution. Taking such a reference is
5187 available only for privileged clients and should be helpful to watch
5188 running services in a race-free manner, and in particular collect
5189 information about exit statuses and results.
5191 * The nss-resolve module has been changed to strictly return UNAVAIL
5192 when communication via D-Bus with resolved failed, and NOTFOUND when
5193 a lookup completed but was negative. This means it is now possible to
5194 neatly configure fallbacks using nsswitch.conf result checking
5195 expressions. Taking benefit of this, the new recommended
5196 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
5198 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
5200 * A new setting CtrlAltDelBurstAction= has been added to
5201 /etc/systemd/system.conf which may be used to configure the precise
5202 behaviour if the user on the console presses Ctrl-Alt-Del more often
5203 than 7 times in 2s. Previously this would unconditionally result in
5204 an expedited, immediate reboot. With this new setting the precise
5205 operation may be configured in more detail, and also turned off
5208 * In .netdev files two new settings RemoteChecksumTx= and
5209 RemoteChecksumRx= are now understood that permit configuring the
5210 remote checksumming logic for VXLAN networks.
5212 * The service manager learnt a new "invocation ID" concept for invoked
5213 services. Each runtime cycle of a service will get a new invocation
5214 ID (a 128bit random UUID) assigned that identifies the current
5215 run of the service uniquely and globally. A new invocation ID
5216 is generated each time a service starts up. The journal will store
5217 the invocation ID of a service along with any logged messages, thus
5218 making the invocation ID useful for matching the online runtime of a
5219 service with the offline log data it generated in a safe way without
5220 relying on synchronized timestamps. In many ways this new service
5221 invocation ID concept is similar to the kernel's boot ID concept that
5222 uniquely and globally identifies the runtime of each boot. The
5223 invocation ID of a service is passed to the service itself via an
5224 environment variable ($INVOCATION_ID). A new bus call
5225 GetUnitByInvocationID() has been added that is similar to GetUnit()
5226 but instead of retrieving the bus path for a unit by its name
5227 retrieves it by its invocation ID. The returned path is valid only as
5228 long as the passed invocation ID is current.
5230 * systemd-resolved gained a new "DNSStubListener" setting in
5231 resolved.conf. It either takes a boolean value or the special values
5232 "udp" and "tcp", and configures whether to enable the stub DNS
5233 listener on 127.0.0.53:53.
5235 * IP addresses configured via networkd may now carry additional
5236 configuration settings supported by the kernel. New options include:
5237 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
5238 PrefixRoute=, AutoJoin=.
5240 * The PAM configuration fragment file for "user@.service" shipped with
5241 systemd (i.e. the --user instance of systemd) has been stripped to
5242 the minimum necessary to make the system boot. Previously, it
5243 contained Fedora-specific stanzas that did not apply to other
5244 distributions. It is expected that downstream distributions add
5245 additional configuration lines, matching their needs to this file,
5246 using it only as rough template of what systemd itself needs. Note
5247 that this reduced fragment does not even include an invocation of
5248 pam_limits which most distributions probably want to add, even though
5249 systemd itself does not need it. (There's also the new build time
5250 option --with-pamconfdir=no to disable installation of the PAM
5253 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
5254 capability is now also dropped from its set (in addition to
5255 CAP_SYS_MKNOD as before).
5257 * In service unit files it is now possible to connect a specific named
5258 file descriptor with stdin/stdout/stdout of an executed service. The
5259 name may be specified in matching .socket units using the
5260 FileDescriptorName= setting.
5262 * A number of journal settings may now be configured on the kernel
5263 command line. Specifically, the following options are now understood:
5264 systemd.journald.max_level_console=,
5265 systemd.journald.max_level_store=,
5266 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
5267 systemd.journald.max_level_wall=.
5269 * "systemctl is-enabled --full" will now show by which symlinks a unit
5270 file is enabled in the unit dependency tree.
5272 * Support for VeraCrypt encrypted partitions has been added to the
5273 "cryptsetup" logic and /etc/crypttab.
5275 * systemd-detect-virt gained support for a new --private-users switch
5276 that checks whether the invoking processes are running inside a user
5277 namespace. Similar, a new special value "private-users" for the
5278 existing ConditionVirtualization= setting has been added, permitting
5279 skipping of specific units in user namespace environments.
5281 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
5282 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
5283 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
5284 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
5285 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
5286 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
5287 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
5288 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
5289 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
5290 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
5291 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
5292 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
5293 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
5294 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
5295 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
5296 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
5297 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
5298 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
5299 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
5300 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
5301 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
5302 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
5303 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
5304 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
5305 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
5306 Jędrzejewski-Szmek, Zeal Jagannatha
5308 — Santa Fe, 2016-11-03
5312 * In service units the various ExecXYZ= settings have been extended
5313 with an additional special character as first argument of the
5314 assigned value: if the character '+' is used the specified command
5315 line it will be run with full privileges, regardless of User=,
5316 Group=, CapabilityBoundingSet= and similar options. The effect is
5317 similar to the existing PermissionsStartOnly= option, but allows
5318 configuration of this concept for each executed command line
5321 * Services may now alter the service watchdog timeout at runtime by
5322 sending a WATCHDOG_USEC= message via sd_notify().
5324 * MemoryLimit= and related unit settings now optionally take percentage
5325 specifications. The percentage is taken relative to the amount of
5326 physical memory in the system (or in case of containers, the assigned
5327 amount of memory). This allows scaling service resources neatly with
5328 the amount of RAM available on the system. Similarly, systemd-logind's
5329 RuntimeDirectorySize= option now also optionally takes percentage
5332 * In similar fashion TasksMax= takes percentage values now, too. The
5333 value is taken relative to the configured maximum number of processes
5334 on the system. The per-service task maximum has been changed to 15%
5335 using this functionality. (Effectively this is an increase of 512 →
5336 4915 for service units, given the kernel's default pid_max setting.)
5338 * Calendar time specifications in .timer units now understand a ".."
5339 syntax for time ranges. Example: "4..7:10" may now be used for
5340 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
5343 * The InaccessableDirectories=, ReadOnlyDirectories= and
5344 ReadWriteDirectories= unit file settings have been renamed to
5345 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
5346 applied to all kinds of file nodes, and not just directories, with
5347 the exception of symlinks. Specifically these settings may now be
5348 used on block and character device nodes, UNIX sockets and FIFOS as
5349 well as regular files. The old names of these settings remain
5350 available for compatibility.
5352 * systemd will now log about all service processes it kills forcibly
5353 (using SIGKILL) because they remained after the clean shutdown phase
5354 of the service completed. This should help identifying services that
5355 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
5356 systemd-logind's configuration a similar log message is generated for
5357 processes killed at the end of each session due to this setting.
5359 * systemd will now set the $JOURNAL_STREAM environment variable for all
5360 services whose stdout/stderr are connected to the Journal (which
5361 effectively means by default: all services). The variable contains
5362 the device and inode number of the file descriptor used for
5363 stdout/stderr. This may be used by invoked programs to detect whether
5364 their stdout/stderr is connected to the Journal, in which case they
5365 can switch over to direct Journal communication, thus being able to
5366 pass extended, structured metadata along with their log messages. As
5367 one example, this is now used by glib's logging primitives.
5369 * When using systemd's default tmp.mount unit for /tmp, the mount point
5370 will now be established with the "nosuid" and "nodev" options. This
5371 avoids privilege escalation attacks that put traps and exploits into
5372 /tmp. However, this might cause problems if you e. g. put container
5373 images or overlays into /tmp; if you need this, override tmp.mount's
5374 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
5377 * systemd now supports the "memory" cgroup controller also on
5380 * The systemd-cgtop tool now optionally takes a control group path as
5381 command line argument. If specified, the control group list shown is
5382 limited to subgroups of that group.
5384 * The SystemCallFilter= unit file setting gained support for
5385 pre-defined, named system call filter sets. For example
5386 SystemCallFilter=@clock is now an effective way to make all clock
5387 changing-related system calls unavailable to a service. A number of
5388 similar pre-defined groups are defined. Writing system call filters
5389 for system services is simplified substantially with this new
5390 concept. Accordingly, all of systemd's own, long-running services now
5391 enable system call filtering based on this, by default.
5393 * A new service setting MemoryDenyWriteExecute= has been added, taking
5394 a boolean value. If turned on, a service may no longer create memory
5395 mappings that are writable and executable at the same time. This
5396 enhances security for services where this is enabled as it becomes
5397 harder to dynamically write and then execute memory in exploited
5398 service processes. This option has been enabled for all of systemd's
5399 own long-running services.
5401 * A new RestrictRealtime= service setting has been added, taking a
5402 boolean argument. If set the service's processes may no longer
5403 acquire realtime scheduling. This improves security as realtime
5404 scheduling may otherwise be used to easily freeze the system.
5406 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
5407 value. This may be used for requesting that the system manager inside
5408 of the container reports start-up completion to nspawn which then
5409 propagates this notification further to the service manager
5410 supervising nspawn itself. A related option NotifyReady= in .nspawn
5411 files has been added too. This functionality allows ordering of the
5412 start-up of multiple containers using the usual systemd ordering
5415 * machinectl gained a new command "stop" that is an alias for
5418 * systemd-resolved gained support for contacting DNS servers on
5419 link-local IPv6 addresses.
5421 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
5422 its caches. A method call for requesting the same operation has been
5423 added to the bus API too, and is made available via "systemd-resolve
5426 * systemd-resolve gained a new --status switch. If passed a brief
5427 summary of the used DNS configuration with per-interface information
5430 * resolved.conf gained a new Cache= boolean option, defaulting to
5431 on. If turned off local DNS caching is disabled. This comes with a
5432 performance penalty in particular when DNSSEC is enabled. Note that
5433 resolved disables its internal caching implicitly anyway, when the
5434 configured DNS server is on a host-local IP address such as ::1 or
5435 127.0.0.1, thus automatically avoiding double local caching.
5437 * systemd-resolved now listens on the local IP address 127.0.0.53:53
5438 for DNS requests. This improves compatibility with local programs
5439 that do not use the libc NSS or systemd-resolved's bus APIs for name
5440 resolution. This minimal DNS service is only available to local
5441 programs and does not implement the full DNS protocol, but enough to
5442 cover local DNS clients. A new, static resolv.conf file, listing just
5443 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
5444 now recommended to make /etc/resolv.conf a symlink to this file in
5445 order to route all DNS lookups to systemd-resolved, regardless if
5446 done via NSS, the bus API or raw DNS packets. Note that this local
5447 DNS service is not as fully featured as the libc NSS or
5448 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
5449 used to deliver link-local address information (as this implies
5450 sending a local interface index along), LLMNR/mDNS support via this
5451 interface is severely restricted. It is thus strongly recommended for
5452 all applications to use the libc NSS API or native systemd-resolved
5455 * systemd-networkd's bridge support learned a new setting
5456 VLANFiltering= for controlling VLAN filtering. Moreover a new section
5457 in .network files has been added for configuring VLAN bridging in
5458 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
5460 * systemd-networkd's IPv6 Router Advertisement code now makes use of
5461 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
5462 now be acquired without relying on DHCPv6. Two new options
5463 UseDomains= and UseDNS= have been added to configure this behaviour.
5465 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
5466 renamed IPv6AcceptRA=, without altering its behaviour. The old
5467 setting name remains available for compatibility reasons.
5469 * The systemd-networkd VTI/VTI6 tunneling support gained new options
5470 Key=, InputKey= and OutputKey=.
5472 * systemd-networkd gained support for VRF ("Virtual Routing Function")
5473 interface configuration.
5475 * "systemctl edit" may now be used to create new unit files by
5476 specifying the --force switch.
5478 * sd-event gained a new function sd_event_get_iteration() for
5479 requesting the current iteration counter of the event loop. It starts
5480 at zero and is increased by one with each event loop iteration.
5482 * A new rpm macro %systemd_ordering is provided by the macros.systemd
5483 file. It can be used in lieu of %systemd_requires in packages which
5484 don't use any systemd functionality and are intended to be installed
5485 in minimal containers without systemd present. This macro provides
5486 ordering dependencies to ensure that if the package is installed in
5487 the same rpm transaction as systemd, systemd will be installed before
5488 the scriptlets for the package are executed, allowing unit presets
5491 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
5492 been added to simplify packaging of generators.
5494 * The os-release file gained VERSION_CODENAME field for the
5495 distribution nickname (e.g. VERSION_CODENAME=woody).
5497 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
5498 can be set to disable parsing of metadata and the creation
5499 of persistent symlinks for that device.
5501 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
5502 to make them available to logged-in users has been reverted.
5504 * Much of the common code of the various systemd components is now
5505 built into an internal shared library libsystemd-shared-231.so
5506 (incorporating the systemd version number in the name, to be updated
5507 with future releases) that the components link to. This should
5508 decrease systemd footprint both in memory during runtime and on
5509 disk. Note that the shared library is not for public use, and is
5510 neither API nor ABI stable, but is likely to change with every new
5511 released update. Packagers need to make sure that binaries
5512 linking to libsystemd-shared.so are updated in step with the
5515 * Configuration for "mkosi" is now part of the systemd
5516 repository. mkosi is a tool to easily build legacy-free OS images,
5517 and is available on github: https://github.com/systemd/mkosi. If
5518 "mkosi" is invoked in the build tree a new raw OS image is generated
5519 incorporating the systemd sources currently being worked on and a
5520 clean, fresh distribution installation. The generated OS image may be
5521 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
5522 UEFI PC. This functionality is particularly useful to easily test
5523 local changes made to systemd in a pristine, defined environment. See
5524 doc/HACKING for details.
5526 * configure learned the --with-support-url= option to specify the
5527 distribution's bugtracker.
5529 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
5530 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
5531 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
5532 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
5533 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
5534 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
5535 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
5536 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
5537 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
5538 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
5539 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
5540 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
5541 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
5542 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
5543 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
5544 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
5545 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
5546 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
5547 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
5549 — Berlin, 2016-07-25
5553 * DNSSEC is now turned on by default in systemd-resolved (in
5554 "allow-downgrade" mode), but may be turned off during compile time by
5555 passing "--with-default-dnssec=no" to "configure" (and of course,
5556 during runtime with DNSSEC= in resolved.conf). We recommend
5557 downstreams to leave this on at least during development cycles and
5558 report any issues with the DNSSEC logic upstream. We are very
5559 interested in collecting feedback about the DNSSEC validator and its
5560 limitations in the wild. Note however, that DNSSEC support is
5561 probably nothing downstreams should turn on in stable distros just
5562 yet, as it might create incompatibilities with a few DNS servers and
5563 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
5564 automatically whenever we detect such incompatible setups, but there
5565 might be systems we do not cover yet. Hence: please help us testing
5566 the DNSSEC code, leave this on where you can, report back, but then
5567 again don't consider turning this on in your stable, LTS or
5568 production release just yet. (Note that you have to enable
5569 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
5570 and its DNSSEC mode for hostname resolution from local
5573 * systemd-resolve conveniently resolves DANE records with the --tlsa
5574 option and OPENPGPKEY records with the --openpgp option. It also
5575 supports dumping raw DNS record data via the new --raw= switch.
5577 * systemd-logind will now by default terminate user processes that are
5578 part of the user session scope unit (session-XX.scope) when the user
5579 logs out. This behavior is controlled by the KillUserProcesses=
5580 setting in logind.conf, and the previous default of "no" is now
5581 changed to "yes". This means that user sessions will be properly
5582 cleaned up after, but additional steps are necessary to allow
5583 intentionally long-running processes to survive logout.
5585 While the user is logged in at least once, user@.service is running,
5586 and any service that should survive the end of any individual login
5587 session can be started at a user service or scope using systemd-run.
5588 systemd-run(1) man page has been extended with an example which shows
5589 how to run screen in a scope unit underneath user@.service. The same
5590 command works for tmux.
5592 After the user logs out of all sessions, user@.service will be
5593 terminated too, by default, unless the user has "lingering" enabled.
5594 To effectively allow users to run long-term tasks even if they are
5595 logged out, lingering must be enabled for them. See loginctl(1) for
5596 details. The default polkit policy was modified to allow users to
5597 set lingering for themselves without authentication.
5599 Previous defaults can be restored at compile time by the
5600 --without-kill-user-processes option to "configure".
5602 * systemd-logind gained new configuration settings SessionsMax= and
5603 InhibitorsMax=, both with a default of 8192. It will not register new
5604 user sessions or inhibitors above this limit.
5606 * systemd-logind will now reload configuration on SIGHUP.
5608 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
5609 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
5610 enable. Also, support for the "io" cgroup controller in the unified
5611 hierarchy has been added, so that the "memory", "pids" and "io" are
5612 now the controllers that are supported on the unified hierarchy.
5614 WARNING: it is not possible to use previous systemd versions with
5615 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
5616 is necessary to also update systemd in the initramfs if using the
5617 unified hierarchy. An updated SELinux policy is also required.
5619 * LLDP support has been extended, and both passive (receive-only) and
5620 active (sender) modes are supported. Passive mode ("routers-only") is
5621 enabled by default in systemd-networkd. Active LLDP mode is enabled
5622 by default for containers on the internal network. The "networkctl
5623 lldp" command may be used to list information gathered. "networkctl
5624 status" will also show basic LLDP information on connected peers now.
5626 * The IAID and DUID unique identifier sent in DHCP requests may now be
5627 configured for the system and each .network file managed by
5628 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
5630 * systemd-networkd gained support for configuring proxy ARP support for
5631 each interface, via the ProxyArp= setting in .network files. It also
5632 gained support for configuring the multicast querier feature of
5633 bridge devices, via the new MulticastQuerier= setting in .netdev
5634 files. Similarly, snooping on the IGMP traffic can be controlled
5635 via the new setting MulticastSnooping=.
5637 A new setting PreferredLifetime= has been added for addresses
5638 configured in .network file to configure the lifetime intended for an
5641 The systemd-networkd DHCP server gained the option EmitRouter=, which
5642 defaults to yes, to configure whether the DHCP Option 3 (Router)
5645 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
5646 systemd-socket-activate and installed into /usr/bin. It is now fully
5649 * systemd-journald now uses separate threads to flush changes to disk
5650 when closing journal files, thus reducing impact of slow disk I/O on
5651 logging performance.
5653 * The sd-journal API gained two new calls
5654 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
5655 can be used to open journal files using file descriptors instead of
5656 file or directory paths. sd_journal_open_container() has been
5657 deprecated, sd_journal_open_directory_fd() should be used instead
5658 with the flag SD_JOURNAL_OS_ROOT.
5660 * journalctl learned a new output mode "-o short-unix" that outputs log
5661 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
5662 UTC). It also gained support for a new --no-hostname setting to
5663 suppress the hostname column in the family of "short" output modes.
5665 * systemd-ask-password now optionally skips printing of the password to
5666 stdout with --no-output which can be useful in scripts.
5668 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
5669 (devices tagged with ID_MAKER_TOOL) are now tagged with
5670 "uaccess" and are available to logged in users.
5672 * The DeviceAllow= unit setting now supports specifiers (with "%").
5674 * "systemctl show" gained a new --value switch, which allows print a
5675 only the contents of a specific unit property, without also printing
5676 the property's name. Similar support was added to "show*" verbs
5677 of loginctl and machinectl that output "key=value" lists.
5679 * A new unit type "generated" was added for files dynamically generated
5680 by generator tools. Similarly, a new unit type "transient" is used
5681 for unit files created using the runtime API. "systemctl enable" will
5682 refuse to operate on such files.
5684 * A new command "systemctl revert" has been added that may be used to
5685 revert to the vendor version of a unit file, in case local changes
5686 have been made by adding drop-ins or overriding the unit file.
5688 * "machinectl clean" gained a new verb to automatically remove all or
5689 just hidden container images.
5691 * systemd-tmpfiles gained support for a new line type "e" for emptying
5692 directories, if they exist, without creating them if they don't.
5694 * systemd-nspawn gained support for automatically patching the UID/GIDs
5695 of the owners and the ACLs of all files and directories in a
5696 container tree to match the UID/GID user namespacing range selected
5697 for the container invocation. This mode is enabled via the new
5698 --private-users-chown switch. It also gained support for
5699 automatically choosing a free, previously unused UID/GID range when
5700 starting a container, via the new --private-users=pick setting (which
5701 implies --private-users-chown). Together, these options for the first
5702 time make user namespacing for nspawn containers fully automatic and
5703 thus deployable. The systemd-nspawn@.service template unit file has
5704 been changed to use this functionality by default.
5706 * systemd-nspawn gained a new --network-zone= switch, that allows
5707 creating ad-hoc virtual Ethernet links between multiple containers,
5708 that only exist as long as at least one container referencing them is
5709 running. This allows easy connecting of multiple containers with a
5710 common link that implements an Ethernet broadcast domain. Each of
5711 these network "zones" may be named relatively freely by the user, and
5712 may be referenced by any number of containers, but each container may
5713 only reference one of these "zones". On the lower level, this is
5714 implemented by an automatically managed bridge network interface for
5715 each zone, that is created when the first container referencing its
5716 zone is created and removed when the last one referencing its zone
5719 * The default start timeout may now be configured on the kernel command
5720 line via systemd.default_timeout_start_sec=. It was already
5721 configurable via the DefaultTimeoutStartSec= option in
5722 /etc/systemd/system.conf.
5724 * Socket units gained a new TriggerLimitIntervalSec= and
5725 TriggerLimitBurst= setting to configure a limit on the activation
5726 rate of the socket unit.
5728 * The LimitNICE= setting now optionally takes normal UNIX nice values
5729 in addition to the raw integer limit value. If the specified
5730 parameter is prefixed with "+" or "-" and is in the range -20..19 the
5731 value is understood as UNIX nice value. If not prefixed like this it
5732 is understood as raw RLIMIT_NICE limit.
5734 * Note that the effect of the PrivateDevices= unit file setting changed
5735 slightly with this release: the per-device /dev file system will be
5736 mounted read-only from this version on, and will have "noexec"
5737 set. This (minor) change of behavior might cause some (exceptional)
5738 legacy software to break, when PrivateDevices=yes is set for its
5739 service. Please leave PrivateDevices= off if you run into problems
5742 * systemd-bootchart has been split out to a separate repository:
5743 https://github.com/systemd/systemd-bootchart
5745 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
5746 merged into the kernel in its current form.
5748 * The compatibility libraries libsystemd-daemon.so,
5749 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
5750 which have been deprecated since systemd-209 have been removed along
5751 with the corresponding pkg-config files. All symbols provided by
5752 those libraries are provided by libsystemd.so.
5754 * The Capabilities= unit file setting has been removed (it is ignored
5755 for backwards compatibility). AmbientCapabilities= and
5756 CapabilityBoundingSet= should be used instead.
5758 * A new special target has been added, initrd-root-device.target,
5759 which creates a synchronization point for dependencies of the root
5760 device in early userspace. Initramfs builders must ensure that this
5761 target is now included in early userspace.
5763 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
5764 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
5765 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
5766 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
5767 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
5768 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
5769 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
5770 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
5771 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
5772 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
5773 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
5774 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
5775 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
5776 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
5777 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
5778 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
5779 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
5780 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
5781 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
5782 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
5783 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
5784 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
5785 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
5786 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
5787 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
5790 — Fairfax, 2016-05-21
5794 * The systemd-resolved DNS resolver service has gained a substantial
5795 set of new features, most prominently it may now act as a DNSSEC
5796 validating stub resolver. DNSSEC mode is currently turned off by
5797 default, but is expected to be turned on by default in one of the
5798 next releases. For now, we invite everybody to test the DNSSEC logic
5799 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
5800 service also gained a full set of D-Bus interfaces, including calls
5801 to configure DNS and DNSSEC settings per link (for use by external
5802 network management software). systemd-resolved and systemd-networkd
5803 now distinguish between "search" and "routing" domains. The former
5804 are used to qualify single-label names, the latter are used purely
5805 for routing lookups within certain domains to specific links.
5806 resolved now also synthesizes RRs for all entries from /etc/hosts.
5808 * The systemd-resolve tool (which is a client utility for
5809 systemd-resolved) has been improved considerably and is now fully
5810 supported and documented. Hence it has moved from /usr/lib/systemd to
5813 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
5816 * The coredump collection logic has been reworked: when a coredump is
5817 collected it is now written to disk, compressed and processed
5818 (including stacktrace extraction) from a new instantiated service
5819 systemd-coredump@.service, instead of directly from the
5820 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
5821 processing large coredumps can take up a substantial amount of
5822 resources and time, and this previously happened entirely outside of
5823 systemd's service supervision. With the new logic the core_pattern
5824 hook only does minimal metadata collection before passing off control
5825 to the new instantiated service, which is configured with a time
5826 limit, a nice level and other settings to minimize negative impact on
5827 the rest of the system. Also note that the new logic will honour the
5828 RLIMIT_CORE setting of the crashed process, which now allows users
5829 and processes to turn off coredumping for their processes by setting
5832 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
5833 and all forked processes by default. Previously, PID 1 would leave
5834 the setting at "0" for all processes, as set by the kernel. Note that
5835 the resource limit traditionally has no effect on the generated
5836 coredumps on the system if the /proc/sys/kernel/core_pattern hook
5837 logic is used. Since the limit is now honoured (see above) its
5838 default has been changed so that the coredumping logic is enabled by
5839 default for all processes, while allowing specific opt-out.
5841 * When the stacktrace is extracted from processes of system users, this
5842 is now done as "systemd-coredump" user, in order to sandbox this
5843 potentially security sensitive parsing operation. (Note that when
5844 processing coredumps of normal users this is done under the user ID
5845 of process that crashed, as before.) Packagers should take notice
5846 that it is now necessary to create the "systemd-coredump" system user
5847 and group at package installation time.
5849 * The systemd-activate socket activation testing tool gained support
5850 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
5851 and --seqpacket switches. It also has been extended to support both
5852 new-style and inetd-style file descriptor passing. Use the new
5853 --inetd switch to request inetd-style file descriptor passing.
5855 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
5856 variable, which takes a boolean value. If set to false, ANSI color
5857 output is disabled in the tools even when run on a terminal that
5860 * The VXLAN support in networkd now supports two new settings
5861 DestinationPort= and PortRange=.
5863 * A new systemd.machine_id= kernel command line switch has been added,
5864 that may be used to set the machine ID in /etc/machine-id if it is
5865 not initialized yet. This command line option has no effect if the
5866 file is already initialized.
5868 * systemd-nspawn gained a new --as-pid2 switch that invokes any
5869 specified command line as PID 2 rather than PID 1 in the
5870 container. In this mode PID 1 is a minimal stub init process that
5871 implements the special POSIX and Linux semantics of PID 1 regarding
5872 signal and child process management. Note that this stub init process
5873 is implemented in nspawn itself and requires no support from the
5874 container image. This new logic is useful to support running
5875 arbitrary commands in the container, as normal processes are
5876 generally not prepared to run as PID 1.
5878 * systemd-nspawn gained a new --chdir= switch for setting the current
5879 working directory for the process started in the container.
5881 * "journalctl /dev/sda" will now output all kernel log messages for
5882 specified device from the current boot, in addition to all devices
5883 that are parents of it. This should make log output about devices
5884 pretty useful, as long as kernel drivers attach enough metadata to
5885 the log messages. (The usual SATA drivers do.)
5887 * The sd-journal API gained two new calls
5888 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
5889 that report whether log data from /run or /var has been found.
5891 * journalctl gained a new switch "--fields" that prints all journal
5892 record field names currently in use in the journal. This is backed
5893 by two new sd-journal API calls sd_journal_enumerate_fields() and
5894 sd_journal_restart_fields().
5896 * Most configurable timeouts in systemd now expect an argument of
5897 "infinity" to turn them off, instead of "0" as before. The semantics
5898 from now on is that a timeout of "0" means "now", and "infinity"
5899 means "never". To maintain backwards compatibility, "0" continues to
5900 turn off previously existing timeout settings.
5902 * "systemctl reload-or-try-restart" has been renamed to "systemctl
5903 try-reload-or-restart" to clarify what it actually does: the "try"
5904 logic applies to both reloading and restarting, not just restarting.
5905 The old name continues to be accepted for compatibility.
5907 * On boot-up, when PID 1 detects that the system clock is behind the
5908 release date of the systemd version in use, the clock is now set
5909 to the latter. Previously, this was already done in timesyncd, in order
5910 to avoid running with clocks set to the various clock epochs such as
5911 1902, 1938 or 1970. With this change the logic is now done in PID 1
5912 in addition to timesyncd during early boot-up, so that it is enforced
5913 before the first process is spawned by systemd. Note that the logic
5914 in timesyncd remains, as it is more comprehensive and ensures
5915 clock monotonicity by maintaining a persistent timestamp file in
5916 /var. Since /var is generally not available in earliest boot or the
5917 initrd, this part of the logic remains in timesyncd, and is not done
5920 * Support for tweaking details in net_cls.class_id through the
5921 NetClass= configuration directive has been removed, as the kernel
5922 people have decided to deprecate that controller in cgroup v2.
5923 Userspace tools such as nftables are moving over to setting rules
5924 that are specific to the full cgroup path of a task, which obsoletes
5925 these controllers anyway. The NetClass= directive is kept around for
5926 legacy compatibility reasons. For a more in-depth description of the
5927 kernel change, please refer to the respective upstream commit:
5929 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
5931 * A new service setting RuntimeMaxSec= has been added that may be used
5932 to specify a maximum runtime for a service. If the timeout is hit, the
5933 service is terminated and put into a failure state.
5935 * A new service setting AmbientCapabilities= has been added. It allows
5936 configuration of additional Linux process capabilities that are
5937 passed to the activated processes. This is only available on very
5940 * The process resource limit settings in service units may now be used
5941 to configure hard and soft limits individually.
5943 * The various libsystemd APIs such as sd-bus or sd-event now publicly
5944 expose support for gcc's __attribute__((cleanup())) C extension.
5945 Specifically, for many object destructor functions alternative
5946 versions have been added that have names suffixed with "p" and take a
5947 pointer to a pointer to the object to destroy, instead of just a
5948 pointer to the object itself. This is useful because these destructor
5949 functions may be used directly as parameters to the cleanup
5950 construct. Internally, systemd has been a heavy user of this GCC
5951 extension for a long time, and with this change similar support is
5952 now available to consumers of the library outside of systemd. Note
5953 that by using this extension in your sources compatibility with old
5954 and strictly ANSI compatible C compilers is lost. However, all gcc or
5955 LLVM versions of recent years support this extension.
5957 * Timer units gained support for a new setting RandomizedDelaySec= that
5958 allows configuring some additional randomized delay to the configured
5959 time. This is useful to spread out timer events to avoid load peaks in
5960 clusters or larger setups.
5962 * Calendar time specifications now support sub-second accuracy.
5964 * Socket units now support listening on SCTP and UDP-lite protocol
5967 * The sd-event API now comes with a full set of man pages.
5969 * Older versions of systemd contained experimental support for
5970 compressing journal files and coredumps with the LZ4 compressor that
5971 was not compatible with the lz4 binary (due to API limitations of the
5972 lz4 library). This support has been removed; only support for files
5973 compatible with the lz4 binary remains. This LZ4 logic is now
5974 officially supported and no longer considered experimental.
5976 * The dkr image import logic has been removed again from importd. dkr's
5977 micro-services focus doesn't fit into the machine image focus of
5978 importd, and quickly got out of date with the upstream dkr API.
5980 * Creation of the /run/lock/lockdev/ directory was dropped from
5981 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
5982 been available for many years. If you still need this, you need to
5983 create your own tmpfiles.d config file with:
5985 d /run/lock/lockdev 0775 root lock -
5987 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
5988 and RebootArgument= have been moved from the [Service] section of
5989 unit files to [Unit], and they are now supported on all unit types,
5990 not just service units. Of course, systemd will continue to
5991 understand these settings also at the old location, in order to
5992 maintain compatibility.
5994 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
5995 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
5996 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
5997 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
5998 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
5999 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
6000 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
6001 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
6002 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
6003 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
6004 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
6005 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
6006 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
6007 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
6008 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
6009 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
6010 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
6011 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
6012 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
6014 — Berlin, 2016-02-11
6018 * A number of properties previously only settable in unit
6019 files are now also available as properties to set when
6020 creating transient units programmatically via the bus, as it
6021 is exposed with systemd-run's --property=
6022 setting. Specifically, these are: SyslogIdentifier=,
6023 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
6024 EnvironmentFile=, ReadWriteDirectories=,
6025 ReadOnlyDirectories=, InaccessibleDirectories=,
6026 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
6028 * When creating transient services via the bus API it is now
6029 possible to pass in a set of file descriptors to use as
6030 STDIN/STDOUT/STDERR for the invoked process.
6032 * Slice units may now be created transiently via the bus APIs,
6033 similar to the way service and scope units may already be
6034 created transiently.
6036 * Wherever systemd expects a calendar timestamp specification
6037 (like in journalctl's --since= and --until= switches) UTC
6038 timestamps are now supported. Timestamps suffixed with "UTC"
6039 are now considered to be in Universal Time Coordinated
6040 instead of the local timezone. Also, timestamps may now
6041 optionally be specified with sub-second accuracy. Both of
6042 these additions also apply to recurring calendar event
6043 specification, such as OnCalendar= in timer units.
6045 * journalctl gained a new "--sync" switch that asks the
6046 journal daemon to write all so far unwritten log messages to
6047 disk and sync the files, before returning.
6049 * systemd-tmpfiles learned two new line types "q" and "Q" that
6050 operate like "v", but also set up a basic btrfs quota
6051 hierarchy when used on a btrfs file system with quota
6054 * tmpfiles' "v", "q" and "Q" will now create a plain directory
6055 instead of a subvolume (even on a btrfs file system) if the
6056 root directory is a plain directory, and not a
6057 subvolume. This should simplify things with certain chroot()
6058 environments which are not aware of the concept of btrfs
6061 * systemd-detect-virt gained a new --chroot switch to detect
6062 whether execution takes place in a chroot() environment.
6064 * CPUAffinity= now takes CPU index ranges in addition to
6067 * The various memory-related resource limit settings (such as
6068 LimitAS=) now understand the usual K, M, G, ... suffixes to
6069 the base of 1024 (IEC). Similar, the time-related resource
6070 limit settings understand the usual min, h, day, ...
6073 * There's a new system.conf setting DefaultTasksMax= to
6074 control the default TasksMax= setting for services and
6075 scopes running on the system. (TasksMax= is the primary
6076 setting that exposes the "pids" cgroup controller on systemd
6077 and was introduced in the previous systemd release.) The
6078 setting now defaults to 512, which means services that are
6079 not explicitly configured otherwise will only be able to
6080 create 512 processes or threads at maximum, from this
6081 version on. Note that this means that thread- or
6082 process-heavy services might need to be reconfigured to set
6083 TasksMax= to a higher value. It is sufficient to set
6084 TasksMax= in these specific unit files to a higher value, or
6085 even "infinity". Similar, there's now a logind.conf setting
6086 UserTasksMax= that defaults to 4096 and limits the total
6087 number of processes or tasks each user may own
6088 concurrently. nspawn containers also have the TasksMax=
6089 value set by default now, to 8192. Note that all of this
6090 only has an effect if the "pids" cgroup controller is
6091 enabled in the kernel. The general benefit of these changes
6092 should be a more robust and safer system, that provides a
6093 certain amount of per-service fork() bomb protection.
6095 * systemd-nspawn gained the new --network-veth-extra= switch
6096 to define additional and arbitrarily-named virtual Ethernet
6097 links between the host and the container.
6099 * A new service execution setting PassEnvironment= has been
6100 added that allows importing select environment variables
6101 from PID1's environment block into the environment block of
6104 * Timer units gained support for a new RemainAfterElapse=
6105 setting which takes a boolean argument. It defaults to on,
6106 exposing behaviour unchanged to previous releases. If set to
6107 off, timer units are unloaded after they elapsed if they
6108 cannot elapse again. This is particularly useful for
6109 transient timer units, which shall not stay around longer
6110 than until they first elapse.
6112 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
6113 default now (the kernel default is 16). This is beneficial
6114 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
6115 allows substantially larger numbers of queued
6116 datagrams. This should increase the capability of systemd to
6117 parallelize boot-up, as logging and sd_notify() are unlikely
6118 to stall execution anymore. If you need to change the value
6119 from the new defaults, use the usual sysctl.d/ snippets.
6121 * The compression framing format used by the journal or
6122 coredump processing has changed to be in line with what the
6123 official LZ4 tools generate. LZ4 compression support in
6124 systemd was considered unsupported previously, as the format
6125 was not compatible with the normal tools. With this release
6126 this has changed now, and it is hence safe for downstream
6127 distributions to turn it on. While not compressing as well
6128 as the XZ, LZ4 is substantially faster, which makes
6129 it a good default choice for the compression logic in the
6130 journal and in coredump handling.
6132 * Any reference to /etc/mtab has been dropped from
6133 systemd. The file has been obsolete since a while, but
6134 systemd refused to work on systems where it was incorrectly
6135 set up (it should be a symlink or non-existent). Please make
6136 sure to update to util-linux 2.27.1 or newer in conjunction
6137 with this systemd release, which also drops any reference to
6138 /etc/mtab. If you maintain a distribution make sure that no
6139 software you package still references it, as this is a
6140 likely source of bugs. There's also a glibc bug pending,
6141 asking for removal of any reference to this obsolete file:
6143 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
6145 Note that only util-linux versions built with
6146 --enable-libmount-force-mountinfo are supported.
6148 * Support for the ".snapshot" unit type has been removed. This
6149 feature turned out to be little useful and little used, and
6150 has now been removed from the core and from systemctl.
6152 * The dependency types RequiresOverridable= and
6153 RequisiteOverridable= have been removed from systemd. They
6154 have been used only very sparingly to our knowledge and
6155 other options that provide a similar effect (such as
6156 systemctl --mode=ignore-dependencies) are much more useful
6157 and commonly used. Moreover, they were only half-way
6158 implemented as the option to control behaviour regarding
6159 these dependencies was never added to systemctl. By removing
6160 these dependency types the execution engine becomes a bit
6161 simpler. Unit files that use these dependencies should be
6162 changed to use the non-Overridable dependency types
6163 instead. In fact, when parsing unit files with these
6164 options, that's what systemd will automatically convert them
6165 too, but it will also warn, asking users to fix the unit
6166 files accordingly. Removal of these dependency types should
6167 only affect a negligible number of unit files in the wild.
6169 * Behaviour of networkd's IPForward= option changed
6170 (again). It will no longer maintain a per-interface setting,
6171 but propagate one way from interfaces where this is enabled
6172 to the global kernel setting. The global setting will be
6173 enabled when requested by a network that is set up, but
6174 never be disabled again. This change was made to make sure
6175 IPv4 and IPv6 behaviour regarding packet forwarding is
6176 similar (as the Linux IPv6 stack does not support
6177 per-interface control of this setting) and to minimize
6180 * In unit files the behaviour of %u, %U, %h, %s has
6181 changed. These specifiers will now unconditionally resolve
6182 to the various user database fields of the user that the
6183 systemd instance is running as, instead of the user
6184 configured in the specific unit via User=. Note that this
6185 effectively doesn't change much, as resolving of these
6186 specifiers was already turned off in the --system instance
6187 of systemd, as we cannot do NSS lookups from PID 1. In the
6188 --user instance of systemd these specifiers where correctly
6189 resolved, but hardly made any sense, since the user instance
6190 lacks privileges to do user switches anyway, and User= is
6191 hence useless. Moreover, even in the --user instance of
6192 systemd behaviour was awkward as it would only take settings
6193 from User= assignment placed before the specifier into
6194 account. In order to unify and simplify the logic around
6195 this the specifiers will now always resolve to the
6196 credentials of the user invoking the manager (which in case
6197 of PID 1 is the root user).
6199 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
6200 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
6201 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
6202 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
6203 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
6204 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
6205 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
6206 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
6207 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
6208 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
6209 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
6210 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
6211 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
6212 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
6215 — Berlin, 2015-11-18
6219 * systemd now depends on util-linux v2.27. More specifically,
6220 the newly added mount monitor feature in libmount now
6221 replaces systemd's former own implementation.
6223 * libmount mandates /etc/mtab not to be regular file, and
6224 systemd now enforces this condition at early boot.
6225 /etc/mtab has been deprecated and warned about for a very
6226 long time, so systems running systemd should already have
6227 stopped having this file around as anything else than a
6228 symlink to /proc/self/mounts.
6230 * Support for the "pids" cgroup controller has been added. It
6231 allows accounting the number of tasks in a cgroup and
6232 enforcing limits on it. This adds two new setting
6233 TasksAccounting= and TasksMax= to each unit, as well as a
6234 global option DefaultTasksAccounting=.
6236 * Support for the "net_cls" cgroup controller has been added.
6237 It allows assigning a net class ID to each task in the
6238 cgroup, which can then be used in firewall rules and traffic
6239 shaping configurations. Note that the kernel netfilter net
6240 class code does not currently work reliably for ingress
6241 packets on unestablished sockets.
6243 This adds a new config directive called NetClass= to CGroup
6244 enabled units. Allowed values are positive numbers for fixed
6245 assignments and "auto" for picking a free value
6248 * 'systemctl is-system-running' now returns 'offline' if the
6249 system is not booted with systemd. This command can now be
6250 used as a substitute for 'systemd-notify --booted'.
6252 * Watchdog timeouts have been increased to 3 minutes for all
6253 in-tree service files. Apparently, disk IO issues are more
6254 frequent than we hoped, and user reported >1 minute waiting
6257 * 'machine-id-commit' functionality has been merged into
6258 'machine-id-setup --commit'. The separate binary has been
6261 * The WorkingDirectory= directive in unit files may now be set
6262 to the special value '~'. In this case, the working
6263 directory is set to the home directory of the user
6264 configured in User=.
6266 * "machinectl shell" will now open the shell in the home
6267 directory of the selected user by default.
6269 * The CrashChVT= configuration file setting is renamed to
6270 CrashChangeVT=, following our usual logic of not
6271 abbreviating unnecessarily. The old directive is still
6272 supported for compat reasons. Also, this directive now takes
6273 an integer value between 1 and 63, or a boolean value. The
6274 formerly supported '-1' value for disabling stays around for
6277 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
6278 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
6279 RootDirectory= properties can now be set for transient
6282 * The systemd-analyze tool gained a new "set-log-target" verb
6283 to change the logging target the system manager logs to
6284 dynamically during runtime. This is similar to how
6285 "systemd-analyze set-log-level" already changes the log
6288 * In nspawn /sys is now mounted as tmpfs, with only a selected
6289 set of subdirectories mounted in from the real sysfs. This
6290 enhances security slightly, and is useful for ensuring user
6291 namespaces work correctly.
6293 * Support for USB FunctionFS activation has been added. This
6294 allows implementation of USB gadget services that are
6295 activated as soon as they are requested, so that they don't
6296 have to run continuously, similar to classic socket
6299 * The "systemctl exit" command now optionally takes an
6300 additional parameter that sets the exit code to return from
6301 the systemd manager when exiting. This is only relevant when
6302 running the systemd user instance, or when running the
6303 system instance in a container.
6305 * sd-bus gained the new API calls sd_bus_path_encode_many()
6306 and sd_bus_path_decode_many() that allow easy encoding and
6307 decoding of multiple identifier strings inside a D-Bus
6308 object path. Another new call sd_bus_default_flush_close()
6309 has been added to flush and close per-thread default
6312 * systemd-cgtop gained support for a -M/--machine= switch to
6313 show the control groups within a certain container only.
6315 * "systemctl kill" gained support for an optional --fail
6316 switch. If specified the requested operation will fail of no
6317 processes have been killed, because the unit had no
6318 processes attached, or similar.
6320 * A new systemd.crash_reboot=1 kernel command line option has
6321 been added that triggers a reboot after crashing. This can
6322 also be set through CrashReboot= in systemd.conf.
6324 * The RuntimeDirectory= setting now understands unit
6325 specifiers like %i or %f.
6327 * A new (still internal) library API sd-ipv4acd has been added,
6328 that implements address conflict detection for IPv4. It's
6329 based on code from sd-ipv4ll, and will be useful for
6330 detecting DHCP address conflicts.
6332 * File descriptors passed during socket activation may now be
6333 named. A new API sd_listen_fds_with_names() is added to
6334 access the names. The default names may be overridden,
6335 either in the .socket file using the FileDescriptorName=
6336 parameter, or by passing FDNAME= when storing the file
6337 descriptors using sd_notify().
6339 * systemd-networkd gained support for:
6341 - Setting the IPv6 Router Advertisement settings via
6342 IPv6AcceptRouterAdvertisements= in .network files.
6344 - Configuring the HelloTimeSec=, MaxAgeSec= and
6345 ForwardDelaySec= bridge parameters in .netdev files.
6347 - Configuring PreferredSource= for static routes in
6350 * The "ask-password" framework used to query for LUKS harddisk
6351 passwords or SSL passwords during boot gained support for
6352 caching passwords in the kernel keyring, if it is
6353 available. This makes sure that the user only has to type in
6354 a passphrase once if there are multiple objects to unlock
6355 with the same one. Previously, such password caching was
6356 available only when Plymouth was used; this moves the
6357 caching logic into the systemd codebase itself. The
6358 "systemd-ask-password" utility gained a new --keyname=
6359 switch to control which kernel keyring key to use for
6360 caching a password in. This functionality is also useful for
6361 enabling display managers such as gdm to automatically
6362 unlock the user's GNOME keyring if its passphrase, the
6363 user's password and the harddisk password are the same, if
6364 gdm-autologin is used.
6366 * When downloading tar or raw images using "machinectl
6367 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
6368 file is now also downloaded, if it is available and stored
6369 next to the image file.
6371 * Units of type ".socket" gained a new boolean setting
6372 Writable= which is only useful in conjunction with
6373 ListenSpecial=. If true, enables opening the specified
6374 special file in O_RDWR mode rather than O_RDONLY mode.
6376 * systemd-rfkill has been reworked to become a singleton
6377 service that is activated through /dev/rfkill on each rfkill
6378 state change and saves the settings to disk. This way,
6379 systemd-rfkill is now compatible with devices that exist
6380 only intermittendly, and even restores state if the previous
6381 system shutdown was abrupt rather than clean.
6383 * The journal daemon gained support for vacuuming old journal
6384 files controlled by the number of files that shall remain,
6385 in addition to the already existing control by size and by
6386 date. This is useful as journal interleaving performance
6387 degrades with too many separate journal files, and allows
6388 putting an effective limit on them. The new setting defaults
6389 to 100, but this may be changed by setting SystemMaxFiles=
6390 and RuntimeMaxFiles= in journald.conf. Also, the
6391 "journalctl" tool gained the new --vacuum-files= switch to
6392 manually vacuum journal files to leave only the specified
6393 number of files in place.
6395 * udev will now create /dev/disk/by-path links for ATA devices
6396 on kernels where that is supported.
6398 * Galician, Serbian, Turkish and Korean translations were added.
6400 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
6401 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
6402 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
6403 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
6404 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
6405 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
6406 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
6407 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
6408 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
6409 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
6410 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
6411 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
6412 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
6413 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
6414 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
6415 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
6416 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
6417 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
6419 — Berlin, 2015-10-07
6423 * The DHCP implementation of systemd-networkd gained a set of
6426 - The DHCP server now supports emitting DNS and NTP
6427 information. It may be enabled and configured via
6428 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
6429 and NTP information is enabled, but no servers are
6430 configured, the corresponding uplink information (if there
6431 is any) is propagated.
6433 - Server and client now support transmission and reception
6434 of timezone information. It can be configured via the
6435 newly introduced network options UseTimezone=,
6436 EmitTimezone=, and Timezone=. Transmission of timezone
6437 information is enabled between host and containers by
6438 default now: the container will change its local timezone
6439 to what the host has set.
6441 - Lease timeouts can now be configured via
6442 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
6444 - The DHCP server improved on the stability of
6445 leases. Clients are more likely to get the same lease
6446 information back, even if the server loses state.
6448 - The DHCP server supports two new configuration options to
6449 control the lease address pool metrics, PoolOffset= and
6452 * The encapsulation limit of tunnels in systemd-networkd may
6453 now be configured via 'EncapsulationLimit='. It allows
6454 modifying the maximum additional levels of encapsulation
6455 that are permitted to be prepended to a packet.
6457 * systemd now supports the concept of user buses replacing
6458 session buses, if used with dbus-1.10 (and enabled via dbus
6459 --enable-user-session). It previously only supported this on
6460 kdbus-enabled systems, and this release expands this to
6461 'dbus-daemon' systems.
6463 * systemd-networkd now supports predictable interface names
6466 * systemd now optionally supports the new Linux kernel
6467 "unified" control group hierarchy. If enabled via the kernel
6468 command-line option 'systemd.unified_cgroup_hierarchy=1',
6469 systemd will try to mount the unified cgroup hierarchy
6470 directly on /sys/fs/cgroup. If not enabled, or not
6471 available, systemd will fall back to the legacy cgroup
6472 hierarchy setup, as before. Host system and containers can
6473 mix and match legacy and unified hierarchies as they
6474 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
6475 environment variable to individually select the hierarchy to
6476 use for executed containers. By default, nspawn will use the
6477 unified hierarchy for the containers if the host uses the
6478 unified hierarchy, and the legacy hierarchy otherwise.
6479 Please note that at this point the unified hierarchy is an
6480 experimental kernel feature and is likely to change in one
6481 of the next kernel releases. Therefore, it should not be
6482 enabled by default in downstream distributions yet. The
6483 minimum required kernel version for the unified hierarchy to
6484 work is 4.2. Note that when the unified hierarchy is used
6485 for the first time delegated access to controllers is
6486 safe. Because of this systemd-nspawn containers will get
6487 access to controllers now, as will systemd user
6488 sessions. This means containers and user sessions may now
6489 manage their own resources, partitioning up what the system
6492 * A new special scope unit "init.scope" has been introduced
6493 that encapsulates PID 1 of the system. It may be used to
6494 determine resource usage and enforce resource limits on PID
6495 1 itself. PID 1 hence moved out of the root of the control
6498 * The cgtop tool gained support for filtering out kernel
6499 threads when counting tasks in a control group. Also, the
6500 count of processes is now recursively summed up by
6501 default. Two options -k and --recursive= have been added to
6502 revert to old behaviour. The tool has also been updated to
6503 work correctly in containers now.
6505 * systemd-nspawn's --bind= and --bind-ro= options have been
6506 extended to allow creation of non-recursive bind mounts.
6508 * libsystemd gained two new calls sd_pid_get_cgroup() and
6509 sd_peer_get_cgroup() which return the control group path of
6510 a process or peer of a connected AF_UNIX socket. This
6511 function call is particularly useful when implementing
6512 delegated subtrees support in the control group hierarchy.
6514 * The "sd-event" event loop API of libsystemd now supports
6515 correct dequeuing of real-time signals, without losing
6518 * When systemd requests a polkit decision when managing units it
6519 will now add additional fields to the request, including unit
6520 name and desired operation. This enables more powerful polkit
6521 policies, that make decisions depending on these parameters.
6523 * nspawn learnt support for .nspawn settings files, that may
6524 accompany the image files or directories of containers, and
6525 may contain additional settings for the container. This is
6526 an alternative to configuring container parameters via the
6527 nspawn command line.
6529 Contributions from: Cristian Rodríguez, Daniel Mack, David
6530 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
6531 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
6532 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
6533 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
6534 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
6535 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
6536 Andersen, Tom Gundersen, Torstein Husebø
6538 — Berlin, 2015-09-08
6542 * machinectl gained a new verb 'shell' which opens a fresh
6543 shell on the target container or the host. It is similar to
6544 the existing 'login' command of machinectl, but spawns the
6545 shell directly without prompting for username or
6546 password. The pseudo machine '.host' now refers to the local
6547 host and is used by default. Hence, 'machinectl shell' can
6548 be used as replacement for 'su -' which spawns a session as
6549 a fresh systemd unit in a way that is fully isolated from
6550 the originating session.
6552 * systemd-networkd learned to cope with private-zone DHCP
6553 options and allows other programs to query the values.
6555 * SELinux access control when enabling/disabling units is no
6556 longer enforced with this release. The previous implementation
6557 was incorrect, and a new corrected implementation is not yet
6558 available. As unit file operations are still protected via
6559 polkit and D-Bus policy this is not a security problem. Yet,
6560 distributions which care about optimal SELinux support should
6561 probably not stabilize on this release.
6563 * sd-bus gained support for matches of type "arg0has=", that
6564 test for membership of strings in string arrays sent in bus
6567 * systemd-resolved now dumps the contents of its DNS and LLMNR
6568 caches to the logs on reception of the SIGUSR1 signal. This
6569 is useful to debug DNS behaviour.
6571 * The coredumpctl tool gained a new --directory= option to
6572 operate on journal files in a specific directory.
6574 * "systemctl reboot" and related commands gained a new
6575 "--message=" option which may be used to set a free-text
6576 wall message when shutting down or rebooting the
6577 system. This message is also logged, which is useful for
6578 figuring out the reason for a reboot or shutdown a
6581 * The "systemd-resolve-host" tool's -i switch now takes
6582 network interface numbers as alternative to interface names.
6584 * A new unit file setting for services has been introduced:
6585 UtmpMode= allows configuration of how precisely systemd
6586 handles utmp and wtmp entries for the service if this is
6587 enabled. This allows writing services that appear similar to
6588 user sessions in the output of the "w", "who", "last" and
6591 * systemd-resolved will now locally synthesize DNS resource
6592 records for the "localhost" and "gateway" domains as well as
6593 the local hostname. This should ensure that clients querying
6594 RRs via resolved will get similar results as those going via
6595 NSS, if nss-myhostname is enabled.
6597 Contributions from: Alastair Hughes, Alex Crawford, Daniel
6598 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
6599 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
6600 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
6601 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
6602 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
6603 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
6604 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
6605 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
6606 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
6607 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
6608 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
6610 — Berlin, 2015-08-27
6614 * The systemd-efi-boot-generator functionality was merged into
6615 systemd-gpt-auto-generator.
6617 * systemd-networkd now supports Group Policy for vxlan
6618 devices. It can be enabled via the new boolean configuration
6619 option called 'GroupPolicyExtension='.
6621 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
6622 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
6623 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
6625 — Berlin, 2015-07-31
6629 * The python-systemd code has been removed from the systemd repository.
6630 A new repository has been created which accommodates the code from
6631 now on, and we kindly ask distributions to create a separate package
6632 for this: https://github.com/systemd/python-systemd
6634 * The systemd daemon will now reload its main configuration
6635 (/etc/systemd/system.conf) on daemon-reload.
6637 * sd-dhcp now exposes vendor specific extensions via
6638 sd_dhcp_lease_get_vendor_specific().
6640 * systemd-networkd gained a number of new configuration options.
6642 - A new boolean configuration option for TAP devices called
6643 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
6644 device, thus allowing to send and receive GSO packets.
6646 - A new tunnel configuration option called 'CopyDSCP='.
6647 If enabled, the DSCP field of ip6 tunnels is copied into the
6648 decapsulated packet.
6650 - A set of boolean bridge configuration options were added.
6651 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
6652 and 'UnicastFlood=' are now parsed by networkd and applied to the
6653 respective bridge link device via the respective IFLA_BRPORT_*
6656 - A new string configuration option to override the hostname sent
6657 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
6658 is true, networkd will use the configured hostname instead of the
6659 system hostname when sending DHCP requests.
6661 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
6662 networkd will configure the IPv6 flow-label of the tunnel device
6663 according to RFC2460.
6665 - The 'macvtap' virtual network devices are now supported, similar to
6666 the already supported 'macvlan' devices.
6668 * systemd-resolved now implements RFC5452 to improve resilience against
6669 cache poisoning. Additionally, source port randomization is enabled
6670 by default to further protect against DNS spoofing attacks.
6672 * nss-mymachines now supports translating UIDs and GIDs of running
6673 containers with user-namespaces enabled. If a container 'foo'
6674 translates a host uid 'UID' to the container uid 'TUID', then
6675 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
6676 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
6677 mapped as 'vg-foo-TGID'.
6679 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
6680 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
6681 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
6682 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
6683 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
6684 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
6685 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
6686 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
6687 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
6688 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
6690 — Berlin, 2015-07-29
6694 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
6695 There are no known issues with current sysfs, and udev does not need
6696 or should be used to work around such bugs.
6698 * udev does no longer enable USB HID power management. Several reports
6699 indicate, that some devices cannot handle that setting.
6701 * The udev accelerometer helper was removed. The functionality
6702 is now fully included in iio-sensor-proxy. But this means,
6703 older iio-sensor-proxy versions will no longer provide
6704 accelerometer/orientation data with this systemd version.
6705 Please upgrade iio-sensor-proxy to version 1.0.
6707 * networkd gained a new configuration option IPv6PrivacyExtensions=
6708 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
6709 for Stateless Address") on selected networks.
6711 * For the sake of fewer build-time dependencies and less code in the
6712 main repository, the python bindings are about to be removed in the
6713 next release. A new repository has been created which accommodates
6714 the code from now on, and we kindly ask distributions to create a
6715 separate package for this. The removal will take place in v223.
6717 https://github.com/systemd/python-systemd
6719 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
6720 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
6721 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
6722 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
6723 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
6724 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
6725 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
6726 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
6727 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
6728 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
6730 — Berlin, 2015-07-07
6734 * The sd-bus.h and sd-event.h APIs have now been declared
6735 stable and have been added to the official interface of
6736 libsystemd.so. sd-bus implements an alternative D-Bus client
6737 library, that is relatively easy to use, very efficient and
6738 supports both classic D-Bus as well as kdbus as transport
6739 backend. sd-event is a generic event loop abstraction that
6740 is built around Linux epoll, but adds features such as event
6741 prioritization or efficient timer handling. Both APIs are good
6742 choices for C programs looking for a bus and/or event loop
6743 implementation that is minimal and does not have to be
6744 portable to other kernels.
6746 * kdbus support is no longer compile-time optional. It is now
6747 always built-in. However, it can still be disabled at
6748 runtime using the kdbus=0 kernel command line setting, and
6749 that setting may be changed to default to off, by specifying
6750 --disable-kdbus at build-time. Note though that the kernel
6751 command line setting has no effect if the kdbus.ko kernel
6752 module is not installed, in which case kdbus is (obviously)
6753 also disabled. We encourage all downstream distributions to
6754 begin testing kdbus by adding it to the kernel images in the
6755 development distributions, and leaving kdbus support in
6758 * The minimal required util-linux version has been bumped to
6761 * Support for chkconfig (--enable-chkconfig) was removed in
6762 favor of calling an abstraction tool
6763 /lib/systemd/systemd-sysv-install. This needs to be
6764 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
6765 in README for details.
6767 * If there's a systemd unit and a SysV init script for the
6768 same service name, and the user executes "systemctl enable"
6769 for it (or a related call), then this will now enable both
6770 (or execute the related operation on both), not just the
6773 * The libudev API documentation has been converted from gtkdoc
6776 * gudev has been removed from the systemd tree, it is now an
6779 * The systemd-cgtop tool learnt a new --raw switch to generate
6780 "raw" (machine parsable) output.
6782 * networkd's IPForwarding= .network file setting learnt the
6783 new setting "kernel", which ensures that networkd does not
6784 change the IP forwarding sysctl from the default kernel
6787 * The systemd-logind bus API now exposes a new boolean
6788 property "Docked" that reports whether logind considers the
6789 system "docked", i.e. connected to a docking station or not.
6791 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
6792 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
6793 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
6794 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
6795 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
6796 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
6797 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
6798 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
6799 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
6800 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
6801 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
6802 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
6803 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
6804 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
6805 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
6806 Fink, Zbigniew Jędrzejewski-Szmek
6808 — Berlin, 2015-06-19
6812 * The gudev library has been extracted into a separate repository
6813 available at: https://git.gnome.org/browse/libgudev/
6814 It is now managed as part of the Gnome project. Distributions
6815 are recommended to pass --disable-gudev to systemd and use
6816 gudev from the Gnome project instead. gudev is still included
6817 in systemd, for now. It will be removed soon, though. Please
6818 also see the announcement-thread on systemd-devel:
6819 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
6821 * systemd now exposes a CPUUsageNSec= property for each
6822 service unit on the bus, that contains the overall consumed
6823 CPU time of a service (the sum of what each process of the
6824 service consumed). This value is only available if
6825 CPUAccounting= is turned on for a service, and is then shown
6826 in the "systemctl status" output.
6828 * Support for configuring alternative mappings of the old SysV
6829 runlevels to systemd targets has been removed. They are now
6830 hardcoded in a way that runlevels 2, 3, 4 all map to
6831 multi-user.target and 5 to graphical.target (which
6832 previously was already the default behaviour).
6834 * The auto-mounter logic gained support for mount point
6835 expiry, using a new TimeoutIdleSec= setting in .automount
6836 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
6838 * The EFI System Partition (ESP) as mounted to /boot by
6839 systemd-efi-boot-generator will now be unmounted
6840 automatically after 2 minutes of not being used. This should
6841 minimize the risk of ESP corruptions.
6843 * New /etc/fstab options x-systemd.requires= and
6844 x-systemd.requires-mounts-for= are now supported to express
6845 additional dependencies for mounts. This is useful for
6846 journalling file systems that support external journal
6847 devices or overlay file systems that require underlying file
6848 systems to be mounted.
6850 * systemd does not support direct live-upgrades (via systemctl
6851 daemon-reexec) from versions older than v44 anymore. As no
6852 distribution we are aware of shipped such old versions in a
6853 stable release this should not be problematic.
6855 * When systemd forks off a new per-connection service instance
6856 it will now set the $REMOTE_ADDR environment variable to the
6857 remote IP address, and $REMOTE_PORT environment variable to
6858 the remote IP port. This behaviour is similar to the
6859 corresponding environment variables defined by CGI.
6861 * systemd-networkd gained support for uplink failure
6862 detection. The BindCarrier= option allows binding interface
6863 configuration dynamically to the link sense of other
6864 interfaces. This is useful to achieve behaviour like in
6867 * systemd-networkd gained support for configuring the DHCP
6868 client identifier to use when requesting leases.
6870 * systemd-networkd now has a per-network UseNTP= option to
6871 configure whether NTP server information acquired via DHCP
6872 is passed on to services like systemd-timesyncd.
6874 * systemd-networkd gained support for vti6 tunnels.
6876 * Note that systemd-networkd manages the sysctl variable
6877 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
6878 it is configured for since v219. The variable controls IP
6879 forwarding, and is a per-interface alternative to the global
6880 /proc/sys/net/ipv[46]/ip_forward. This setting is
6881 configurable in the IPForward= option, which defaults to
6882 "no". This means if networkd is used for an interface it is
6883 no longer sufficient to set the global sysctl option to turn
6884 on IP forwarding! Instead, the .network file option
6885 IPForward= needs to be turned on! Note that the
6886 implementation of this behaviour was broken in v219 and has
6889 * Many bonding and vxlan options are now configurable in
6892 * systemd-nspawn gained a new --property= setting to set unit
6893 properties for the container scope. This is useful for
6894 setting resource parameters (e.g. "CPUShares=500") on
6895 containers started from the command line.
6897 * systemd-nspawn gained a new --private-users= switch to make
6898 use of user namespacing available on recent Linux kernels.
6900 * systemd-nspawn may now be called as part of a shell pipeline
6901 in which case the pipes used for stdin and stdout are passed
6902 directly to the process invoked in the container, without
6903 indirection via a pseudo tty.
6905 * systemd-nspawn gained a new switch to control the UNIX
6906 signal to use when killing the init process of the container
6909 * systemd-nspawn gained a new --overlay= switch for mounting
6910 overlay file systems into the container using the new kernel
6913 * When a container image is imported via systemd-importd and
6914 the host file system is not btrfs, a loopback block device
6915 file is created in /var/lib/machines.raw with a btrfs file
6916 system inside. It is then mounted to /var/lib/machines to
6917 enable btrfs features for container management. The loopback
6918 file and btrfs file system is grown as needed when container
6919 images are imported via systemd-importd.
6921 * systemd-machined/systemd-importd gained support for btrfs
6922 quota, to enforce container disk space limits on disk. This
6923 is exposed in "machinectl set-limit".
6925 * systemd-importd now can import containers from local .tar,
6926 .raw and .qcow2 images, and export them to .tar and .raw. It
6927 can also import dkr v2 images now from the network (on top
6930 * systemd-importd gained support for verifying downloaded
6931 images with gpg2 (previously only gpg1 was supported).
6933 * systemd-machined, systemd-logind, systemd: most bus calls are
6934 now accessible to unprivileged processes via polkit. Also,
6935 systemd-logind will now allow users to kill their own sessions
6936 without further privileges or authorization.
6938 * systemd-shutdownd has been removed. This service was
6939 previously responsible for implementing scheduled shutdowns
6940 as exposed in /usr/bin/shutdown's time parameter. This
6941 functionality has now been moved into systemd-logind and is
6942 accessible via a bus interface.
6944 * "systemctl reboot" gained a new switch --firmware-setup that
6945 can be used to reboot into the EFI firmware setup, if that
6946 is available. systemd-logind now exposes an API on the bus
6947 to trigger such reboots, in case graphical desktop UIs want
6948 to cover this functionality.
6950 * "systemctl enable", "systemctl disable" and "systemctl mask"
6951 now support a new "--now" switch. If specified the units
6952 that are enabled will also be started, and the ones
6953 disabled/masked also stopped.
6955 * The Gummiboot EFI boot loader tool has been merged into
6956 systemd, and renamed to "systemd-boot". The bootctl tool has been
6957 updated to support systemd-boot.
6959 * An EFI kernel stub has been added that may be used to create
6960 kernel EFI binaries that contain not only the actual kernel,
6961 but also an initrd, boot splash, command line and OS release
6962 information. This combined binary can then be signed as a
6963 single image, so that the firmware can verify it all in one
6964 step. systemd-boot has special support for EFI binaries created
6965 like this and can extract OS release information from them
6966 and show them in the boot menu. This functionality is useful
6967 to implement cryptographically verified boot schemes.
6969 * Optional support has been added to systemd-fsck to pass
6970 fsck's progress report to an AF_UNIX socket in the file
6973 * udev will no longer create device symlinks for all block devices by
6974 default. A deny list for excluding special block devices from this
6975 logic has been turned into a allow list that requires picking block
6976 devices explicitly that require device symlinks.
6978 * A new (currently still internal) API sd-device.h has been
6979 added to libsystemd. This modernized API is supposed to
6980 replace libudev eventually. In fact, already much of libudev
6981 is now just a wrapper around sd-device.h.
6983 * A new hwdb database for storing metadata about pointing
6984 stick devices has been added.
6986 * systemd-tmpfiles gained support for setting file attributes
6987 similar to the "chattr" tool with new 'h' and 'H' lines.
6989 * systemd-journald will no longer unconditionally set the
6990 btrfs NOCOW flag on new journal files. This is instead done
6991 with tmpfiles snippet using the new 'h' line type. This
6992 allows easy disabling of this logic, by masking the
6993 journal-nocow.conf tmpfiles file.
6995 * systemd-journald will now translate audit message types to
6996 human readable identifiers when writing them to the
6997 journal. This should improve readability of audit messages.
6999 * The LUKS logic gained support for the offset= and skip=
7000 options in /etc/crypttab, as previously implemented by
7003 * /usr/lib/os-release gained a new optional field VARIANT= for
7004 distributions that support multiple variants (such as a
7005 desktop edition, a server edition, ...)
7007 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
7008 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
7009 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
7010 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
7011 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
7012 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
7013 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
7014 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
7015 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
7016 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
7017 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
7018 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
7019 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
7020 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
7021 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
7022 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
7023 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
7024 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
7025 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
7026 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
7027 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
7028 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
7029 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
7030 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
7031 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
7032 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
7033 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
7035 — Berlin, 2015-05-22
7039 * Introduce a new API "sd-hwdb.h" for querying the hardware
7040 metadata database. With this minimal interface one can query
7041 and enumerate the udev hwdb, decoupled from the old libudev
7042 library. libudev's interface for this is now only a wrapper
7043 around sd-hwdb. A new tool systemd-hwdb has been added to
7044 interface with and update the database.
7046 * When any of systemd's tools copies files (for example due to
7047 tmpfiles' C lines) a btrfs reflink will attempted first,
7048 before bytewise copying is done.
7050 * systemd-nspawn gained a new --ephemeral switch. When
7051 specified a btrfs snapshot is taken of the container's root
7052 directory, and immediately removed when the container
7053 terminates again. Thus, a container can be started whose
7054 changes never alter the container's root directory, and are
7055 lost on container termination. This switch can also be used
7056 for starting a container off the root file system of the
7057 host without affecting the host OS. This switch is only
7058 available on btrfs file systems.
7060 * systemd-nspawn gained a new --template= switch. It takes the
7061 path to a container tree to use as template for the tree
7062 specified via --directory=, should that directory be
7063 missing. This allows instantiating containers dynamically,
7064 on first run. This switch is only available on btrfs file
7067 * When a .mount unit refers to a mount point on which multiple
7068 mounts are stacked, and the .mount unit is stopped all of
7069 the stacked mount points will now be unmounted until no
7070 mount point remains.
7072 * systemd now has an explicit notion of supported and
7073 unsupported unit types. Jobs enqueued for unsupported unit
7074 types will now fail with an "unsupported" error code. More
7075 specifically .swap, .automount and .device units are not
7076 supported in containers, .busname units are not supported on
7077 non-kdbus systems. .swap and .automount are also not
7078 supported if their respective kernel compile time options
7081 * machinectl gained support for two new "copy-from" and
7082 "copy-to" commands for copying files from a running
7083 container to the host or vice versa.
7085 * machinectl gained support for a new "bind" command to bind
7086 mount host directories into local containers. This is
7087 currently only supported for nspawn containers.
7089 * networkd gained support for configuring bridge forwarding
7090 database entries (fdb) from .network files.
7092 * A new tiny daemon "systemd-importd" has been added that can
7093 download container images in tar, raw, qcow2 or dkr formats,
7094 and make them available locally in /var/lib/machines, so
7095 that they can run as nspawn containers. The daemon can GPG
7096 verify the downloads (not supported for dkr, since it has no
7097 provisions for verifying downloads). It will transparently
7098 decompress bz2, xz, gzip compressed downloads if necessary,
7099 and restore sparse files on disk. The daemon uses privilege
7100 separation to ensure the actual download logic runs with
7101 fewer privileges than the daemon itself. machinectl has
7102 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
7103 make the functionality of importd available to the
7104 user. With this in place the Fedora and Ubuntu "Cloud"
7105 images can be downloaded and booted as containers unmodified
7106 (the Fedora images lack the appropriate GPG signature files
7107 currently, so they cannot be verified, but this will change
7108 soon, hopefully). Note that downloading images is currently
7109 only fully supported on btrfs.
7111 * machinectl is now able to list container images found in
7112 /var/lib/machines, along with some metadata about sizes of
7113 disk and similar. If the directory is located on btrfs and
7114 quota is enabled, this includes quota display. A new command
7115 "image-status" has been added that shows additional
7116 information about images.
7118 * machinectl is now able to clone container images
7119 efficiently, if the underlying file system (btrfs) supports
7120 it, with the new "machinectl clone" command. It also
7121 gained commands for renaming and removing images, as well as
7122 marking them read-only or read-write (supported also on
7123 legacy file systems).
7125 * networkd gained support for collecting LLDP network
7126 announcements, from hardware that supports this. This is
7127 shown in networkctl output.
7129 * systemd-run gained support for a new -t (--pty) switch for
7130 invoking a binary on a pty whose input and output is
7131 connected to the invoking terminal. This allows executing
7132 processes as system services while interactively
7133 communicating with them via the terminal. Most interestingly
7134 this is supported across container boundaries. Invoking
7135 "systemd-run -t /bin/bash" is an alternative to running a
7136 full login session, the difference being that the former
7137 will not register a session, nor go through the PAM session
7140 * tmpfiles gained support for a new "v" line type for creating
7141 btrfs subvolumes. If the underlying file system is a legacy
7142 file system, this automatically degrades to creating a
7143 normal directory. Among others /var/lib/machines is now
7144 created like this at boot, should it be missing.
7146 * The directory /var/lib/containers/ has been deprecated and
7147 been replaced by /var/lib/machines. The term "machines" has
7148 been used in the systemd context as generic term for both
7149 VMs and containers, and hence appears more appropriate for
7150 this, as the directory can also contain raw images bootable
7153 * systemd-nspawn when invoked with -M but without --directory=
7154 or --image= is now capable of searching for the container
7155 root directory, subvolume or disk image automatically, in
7156 /var/lib/machines. systemd-nspawn@.service has been updated
7157 to make use of this, thus allowing it to be used for raw
7160 * A new machines.target unit has been introduced that is
7161 supposed to group all containers/VMs invoked as services on
7162 the system. systemd-nspawn@.service has been updated to
7163 integrate with that.
7165 * machinectl gained a new "start" command, for invoking a
7166 container as a service. "machinectl start foo" is mostly
7167 equivalent to "systemctl start systemd-nspawn@foo.service",
7168 but handles escaping in a nicer way.
7170 * systemd-nspawn will now mount most of the cgroupfs tree
7171 read-only into each container, with the exception of the
7172 container's own subtree in the name=systemd hierarchy.
7174 * journald now sets the special FS_NOCOW file flag for its
7175 journal files. This should improve performance on btrfs, by
7176 avoiding heavy fragmentation when journald's write-pattern
7177 is used on COW file systems. It degrades btrfs' data
7178 integrity guarantees for the files to the same levels as for
7179 ext3/ext4 however. This should be OK though as journald does
7180 its own data integrity checks and all its objects are
7181 checksummed on disk. Also, journald should handle btrfs disk
7182 full events a lot more gracefully now, by processing SIGBUS
7183 errors, and not relying on fallocate() anymore.
7185 * When journald detects that journal files it is writing to
7186 have been deleted it will immediately start new journal
7189 * systemd now provides a way to store file descriptors
7190 per-service in PID 1. This is useful for daemons to ensure
7191 that fds they require are not lost during a daemon
7192 restart. The fds are passed to the daemon on the next
7193 invocation in the same way socket activation fds are
7194 passed. This is now used by journald to ensure that the
7195 various sockets connected to all the system's stdout/stderr
7196 are not lost when journald is restarted. File descriptors
7197 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
7198 an extension to sd_notify(). Note that a limit is enforced
7199 on the number of fds a service can store in PID 1, and it
7200 defaults to 0, so that no fds may be stored, unless this is
7201 explicitly turned on.
7203 * The default TERM variable to use for units connected to a
7204 terminal, when no other value is explicitly is set is now
7205 vt220 rather than vt102. This should be fairly safe still,
7206 but allows PgUp/PgDn work.
7208 * The /etc/crypttab option header= as known from Debian is now
7211 * "loginctl user-status" and "loginctl session-status" will
7212 now show the last 10 lines of log messages of the
7213 user/session following the status output. Similar,
7214 "machinectl status" will show the last 10 log lines
7215 associated with a virtual machine or container
7216 service. (Note that this is usually not the log messages
7217 done in the VM/container itself, but simply what the
7218 container manager logs. For nspawn this includes all console
7221 * "loginctl session-status" without further argument will now
7222 show the status of the session of the caller. Similar,
7223 "lock-session", "unlock-session", "activate",
7224 "enable-linger", "disable-linger" may now be called without
7225 session/user parameter in which case they apply to the
7226 caller's session/user.
7228 * An X11 session scriptlet is now shipped that uploads
7229 $DISPLAY and $XAUTHORITY into the environment of the systemd
7230 --user daemon if a session begins. This should improve
7231 compatibility with X11 enabled applications run as systemd
7234 * Generators are now subject to masking via /etc and /run, the
7235 same way as unit files.
7237 * networkd .network files gained support for configuring
7238 per-link IPv4/IPv6 packet forwarding as well as IPv4
7239 masquerading. This is by default turned on for veth links to
7240 containers, as registered by systemd-nspawn. This means that
7241 nspawn containers run with --network-veth will now get
7242 automatic routed access to the host's networks without any
7243 further configuration or setup, as long as networkd runs on
7246 * systemd-nspawn gained the --port= (-p) switch to expose TCP
7247 or UDP posts of a container on the host. With this in place
7248 it is possible to run containers with private veth links
7249 (--network-veth), and have their functionality exposed on
7250 the host as if their services were running directly on the
7253 * systemd-nspawn's --network-veth switch now gained a short
7254 version "-n", since with the changes above it is now truly
7255 useful out-of-the-box. The systemd-nspawn@.service has been
7256 updated to make use of it too by default.
7258 * systemd-nspawn will now maintain a per-image R/W lock, to
7259 ensure that the same image is not started more than once
7260 writable. (It's OK to run an image multiple times
7261 simultaneously in read-only mode.)
7263 * systemd-nspawn's --image= option is now capable of
7264 dissecting and booting MBR and GPT disk images that contain
7265 only a single active Linux partition. Previously it
7266 supported only GPT disk images with proper GPT type
7267 IDs. This allows running cloud images from major
7268 distributions directly with systemd-nspawn, without
7271 * In addition to collecting mouse dpi data in the udev
7272 hardware database, there's now support for collecting angle
7273 information for mouse scroll wheels. The database is
7274 supposed to guarantee similar scrolling behavior on mice
7275 that it knows about. There's also support for collecting
7276 information about Touchpad types.
7278 * udev's input_id built-in will now also collect touch screen
7279 dimension data and attach it to probed devices.
7281 * /etc/os-release gained support for a Distribution Privacy
7284 * networkd gained support for creating "ipvlan", "gretap",
7285 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
7287 * systemd-tmpfiles gained support for "a" lines for setting
7290 * systemd-nspawn will now mount /tmp in the container to
7291 tmpfs, automatically.
7293 * systemd now exposes the memory.usage_in_bytes cgroup
7294 attribute and shows it for each service in the "systemctl
7295 status" output, if available.
7297 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
7298 immediate reboot is triggered. This useful if shutdown is
7299 hung and is unable to complete, to expedite the
7300 operation. Note that this kind of reboot will still unmount
7301 all file systems, and hence should not result in fsck being
7304 * A .device unit for an optical block device will now be
7305 considered active only when a medium is in the drive. Also,
7306 mount units are now bound to their backing devices thus
7307 triggering automatic unmounting when devices become
7308 unavailable. With this in place systemd will now
7309 automatically unmount left-over mounts when a CD-ROM is
7310 ejected or an USB stick is yanked from the system.
7312 * networkd-wait-online now has support for waiting for
7313 specific interfaces only (with globbing), and for giving up
7314 after a configurable timeout.
7316 * networkd now exits when idle. It will be automatically
7317 restarted as soon as interfaces show up, are removed or
7318 change state. networkd will stay around as long as there is
7319 at least one DHCP state machine or similar around, that keep
7322 * networkd may now configure IPv6 link-local addressing in
7323 addition to IPv4 link-local addressing.
7325 * The IPv6 "token" for use in SLAAC may now be configured for
7326 each .network interface in networkd.
7328 * Routes configured with networkd may now be assigned a scope
7331 * networkd's [Match] sections now support globbing and lists
7332 of multiple space-separated matches per item.
7334 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
7335 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
7336 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
7337 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
7338 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
7339 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
7340 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
7341 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
7342 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
7343 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
7344 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
7345 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
7346 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
7347 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
7348 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
7349 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
7350 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
7351 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
7352 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
7353 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
7354 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
7355 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
7356 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
7357 Hoffmann, Zbigniew Jędrzejewski-Szmek
7359 — Berlin, 2015-02-16
7363 * When querying unit file enablement status (for example via
7364 "systemctl is-enabled"), a new state "indirect" is now known
7365 which indicates that a unit might not be enabled itself, but
7366 another unit listed in its Also= setting might be.
7368 * Similar to the various existing ConditionXYZ= settings for
7369 units, there are now matching AssertXYZ= settings. While
7370 failing conditions cause a unit to be skipped, but its job
7371 to succeed, failing assertions declared like this will cause
7372 a unit start operation and its job to fail.
7374 * hostnamed now knows a new chassis type "embedded".
7376 * systemctl gained a new "edit" command. When used on a unit
7377 file, this allows extending unit files with .d/ drop-in
7378 configuration snippets or editing the full file (after
7379 copying it from /usr/lib to /etc). This will invoke the
7380 user's editor (as configured with $EDITOR), and reload the
7381 modified configuration after editing.
7383 * "systemctl status" now shows the suggested enablement state
7384 for a unit, as declared in the (usually vendor-supplied)
7385 system preset files.
7387 * nss-myhostname will now resolve the single-label hostname
7388 "gateway" to the locally configured default IP routing
7389 gateways, ordered by their metrics. This assigns a stable
7390 name to the used gateways, regardless which ones are
7391 currently configured. Note that the name will only be
7392 resolved after all other name sources (if nss-myhostname is
7393 configured properly) and should hence not negatively impact
7394 systems that use the single-label hostname "gateway" in
7397 * systemd-inhibit now allows filtering by mode when listing
7400 * Scope and service units gained a new "Delegate" boolean
7401 property, which, when set, allows processes running inside the
7402 unit to further partition resources. This is primarily
7403 useful for systemd user instances as well as container
7406 * journald will now pick up audit messages directly from
7407 the kernel, and log them like any other log message. The
7408 audit fields are split up and fully indexed. This means that
7409 journalctl in many ways is now a (nicer!) alternative to
7410 ausearch, the traditional audit client. Note that this
7411 implements only a minimal audit client. If you want the
7412 special audit modes like reboot-on-log-overflow, please use
7413 the traditional auditd instead, which can be used in
7414 parallel to journald.
7416 * The ConditionSecurity= unit file option now understands the
7417 special string "audit" to check whether auditing is
7420 * journalctl gained two new commands --vacuum-size= and
7421 --vacuum-time= to delete old journal files until the
7422 remaining ones take up no more than the specified size on disk,
7423 or are not older than the specified time.
7425 * A new, native PPPoE library has been added to sd-network,
7426 systemd's library of light-weight networking protocols. This
7427 library will be used in a future version of networkd to
7428 enable PPPoE communication without an external pppd daemon.
7430 * The busctl tool now understands a new "capture" verb that
7431 works similar to "monitor", but writes a packet capture
7432 trace to STDOUT that can be redirected to a file which is
7433 compatible with libcap's capture file format. This can then
7434 be loaded in Wireshark and similar tools to inspect bus
7437 * The busctl tool now understands a new "tree" verb that shows
7438 the object trees of a specific service on the bus, or of all
7441 * The busctl tool now understands a new "introspect" verb that
7442 shows all interfaces and members of objects on the bus,
7443 including their signature and values. This is particularly
7444 useful to get more information about bus objects shown by
7445 the new "busctl tree" command.
7447 * The busctl tool now understands new verbs "call",
7448 "set-property" and "get-property" for invoking bus method
7449 calls, setting and getting bus object properties in a
7452 * busctl gained a new --augment-creds= argument that controls
7453 whether the tool shall augment credential information it
7454 gets from the bus with data from /proc, in a possibly
7457 * nspawn's --link-journal= switch gained two new values
7458 "try-guest" and "try-host" that work like "guest" and
7459 "host", but do not fail if the host has no persistent
7460 journalling enabled. -j is now equivalent to
7461 --link-journal=try-guest.
7463 * macvlan network devices created by nspawn will now have
7464 stable MAC addresses.
7466 * A new SmackProcessLabel= unit setting has been added, which
7467 controls the SMACK security label processes forked off by
7468 the respective unit shall use.
7470 * If compiled with --enable-xkbcommon, systemd-localed will
7471 verify x11 keymap settings by compiling the given keymap. It
7472 will spew out warnings if the compilation fails. This
7473 requires libxkbcommon to be installed.
7475 * When a coredump is collected, a larger number of metadata
7476 fields is now collected and included in the journal records
7477 created for it. More specifically, control group membership,
7478 environment variables, memory maps, working directory,
7479 chroot directory, /proc/$PID/status, and a list of open file
7480 descriptors is now stored in the log entry.
7482 * The udev hwdb now contains DPI information for mice. For
7485 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
7487 * All systemd programs that read standalone configuration
7488 files in /etc now also support a corresponding series of
7489 .conf.d configuration directories in /etc/, /run/,
7490 /usr/local/lib/, /usr/lib/, and (if configured with
7491 --enable-split-usr) /lib/. In particular, the following
7492 configuration files now have corresponding configuration
7493 directories: system.conf user.conf, logind.conf,
7494 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
7495 resolved.conf, timesyncd.conf, journal-remote.conf, and
7496 journal-upload.conf. Note that distributions should use the
7497 configuration directories in /usr/lib/; the directories in
7498 /etc/ are reserved for the system administrator.
7500 * systemd-rfkill will no longer take the rfkill device name
7501 into account when storing rfkill state on disk, as the name
7502 might be dynamically assigned and not stable. Instead, the
7503 ID_PATH udev variable combined with the rfkill type (wlan,
7504 bluetooth, ...) is used.
7506 * A new service systemd-machine-id-commit.service has been
7507 added. When used on systems where /etc is read-only during
7508 boot, and /etc/machine-id is not initialized (but an empty
7509 file), this service will copy the temporary machine ID
7510 created as replacement into /etc after the system is fully
7511 booted up. This is useful for systems that are freshly
7512 installed with a non-initialized machine ID, but should get
7513 a fixed machine ID for subsequent boots.
7515 * networkd's .netdev files now provide a large set of
7516 configuration parameters for VXLAN devices. Similarly, the
7517 bridge port cost parameter is now configurable in .network
7518 files. There's also new support for configuring IP source
7519 routing. networkd .link files gained support for a new
7520 OriginalName= match that is useful to match against the
7521 original interface name the kernel assigned. .network files
7522 may include MTU= and MACAddress= fields for altering the MTU
7523 and MAC address while being connected to a specific network
7526 * The LUKS logic gained supported for configuring
7527 UUID-specific key files. There's also new support for naming
7528 LUKS device from the kernel command line, using the new
7529 luks.name= argument.
7531 * Timer units may now be transiently created via the bus API
7532 (this was previously already available for scope and service
7533 units). In addition it is now possible to create multiple
7534 transient units at the same time with a single bus call. The
7535 "systemd-run" tool has been updated to make use of this for
7536 running commands on a specified time, in at(1)-style.
7538 * tmpfiles gained support for "t" lines, for assigning
7539 extended attributes to files. Among other uses this may be
7540 used to assign SMACK labels to files.
7542 Contributions from: Alin Rauta, Alison Chaiken, Andrej
7543 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
7544 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
7545 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
7546 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
7547 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
7548 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
7549 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
7550 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
7551 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
7552 Michael Biebl, Michael Chapman, Michael Marineau, Michal
7553 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
7554 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
7555 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
7556 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
7557 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
7558 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
7559 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
7561 — Berlin, 2014-12-10
7565 * journalctl gained the new options -t/--identifier= to match
7566 on the syslog identifier (aka "tag"), as well as --utc to
7567 show log timestamps in the UTC timezone. journalctl now also
7568 accepts -n/--lines=all to disable line capping in a pager.
7570 * journalctl gained a new switch, --flush, that synchronously
7571 flushes logs from /run/log/journal to /var/log/journal if
7572 persistent storage is enabled. systemd-journal-flush.service
7573 now waits until the operation is complete.
7575 * Services can notify the manager before they start a reload
7576 (by sending RELOADING=1) or shutdown (by sending
7577 STOPPING=1). This allows the manager to track and show the
7578 internal state of daemons and closes a race condition when
7579 the process is still running but has closed its D-Bus
7582 * Services with Type=oneshot do not have to have any ExecStart
7585 * User units are now loaded also from
7586 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
7587 /run/systemd/user directory that was already previously
7588 supported, but is under the control of the user.
7590 * Job timeouts (i.e. timeouts on the time a job that is
7591 queued stays in the run queue) can now optionally result in
7592 immediate reboot or power-off actions (JobTimeoutAction= and
7593 JobTimeoutRebootArgument=). This is useful on ".target"
7594 units, to limit the maximum time a target remains
7595 undispatched in the run queue, and to trigger an emergency
7596 operation in such a case. This is now used by default to
7597 turn off the system if boot-up (as defined by everything in
7598 basic.target) hangs and does not complete for at least
7599 15min. Also, if power-off or reboot hang for at least 30min
7600 an immediate power-off/reboot operation is triggered. This
7601 functionality is particularly useful to increase reliability
7602 on embedded devices, but also on laptops which might
7603 accidentally get powered on when carried in a backpack and
7604 whose boot stays stuck in a hard disk encryption passphrase
7607 * systemd-logind can be configured to also handle lid switch
7608 events even when the machine is docked or multiple displays
7609 are attached (HandleLidSwitchDocked= option).
7611 * A helper binary and a service have been added which can be
7612 used to resume from hibernation in the initramfs. A
7613 generator will parse the resume= option on the kernel
7614 command line to trigger resume.
7616 * A user console daemon systemd-consoled has been
7617 added. Currently, it is a preview, and will so far open a
7618 single terminal on each session of the user marked as
7619 Desktop=systemd-console.
7621 * Route metrics can be specified for DHCP routes added by
7624 * The SELinux context of socket-activated services can be set
7625 from the information provided by the networking stack
7626 (SELinuxContextFromNet= option).
7628 * Userspace firmware loading support has been removed and
7629 the minimum supported kernel version is thus bumped to 3.7.
7631 * Timeout for udev workers has been increased from 1 to 3
7632 minutes, but a warning will be printed after 1 minute to
7633 help diagnose kernel modules that take a long time to load.
7635 * Udev rules can now remove tags on devices with TAG-="foobar".
7637 * systemd's readahead implementation has been removed. In many
7638 circumstances it didn't give expected benefits even for
7639 rotational disk drives and was becoming less relevant in the
7640 age of SSDs. As none of the developers has been using
7641 rotating media anymore, and nobody stepped up to actively
7642 maintain this component of systemd it has now been removed.
7644 * Swap units can use Options= to specify discard options.
7645 Discard options specified for swaps in /etc/fstab are now
7648 * Docker containers are now detected as a separate type of
7651 * The Password Agent protocol gained support for queries where
7652 the user input is shown, useful e.g. for user names.
7653 systemd-ask-password gained a new --echo option to turn that
7656 * The default sysctl.d/ snippets will now set:
7658 net.core.default_qdisc = fq_codel
7660 This selects Fair Queuing Controlled Delay as the default
7661 queuing discipline for network interfaces. fq_codel helps
7662 fight the network bufferbloat problem. It is believed to be
7663 a good default with no tuning required for most workloads.
7664 Downstream distributions may override this choice. On 10Gbit
7665 servers that do not do forwarding, "fq" may perform better.
7666 Systems without a good clocksource should use "pfifo_fast".
7668 * If kdbus is enabled during build a new option BusPolicy= is
7669 available for service units, that allows locking all service
7670 processes into a stricter bus policy, in order to limit
7671 access to various bus services, or even hide most of them
7672 from the service's view entirely.
7674 * networkctl will now show the .network and .link file
7675 networkd has applied to a specific interface.
7677 * sd-login gained a new API call sd_session_get_desktop() to
7678 query which desktop environment has been selected for a
7681 * UNIX utmp support is now compile-time optional to support
7682 legacy-free systems.
7684 * systemctl gained two new commands "add-wants" and
7685 "add-requires" for pulling in units from specific targets
7688 * If the word "rescue" is specified on the kernel command line
7689 the system will now boot into rescue mode (aka
7690 rescue.target), which was previously available only by
7691 specifying "1" or "systemd.unit=rescue.target" on the kernel
7692 command line. This new kernel command line option nicely
7693 mirrors the already existing "emergency" kernel command line
7696 * New kernel command line options mount.usr=, mount.usrflags=,
7697 mount.usrfstype= have been added that match root=, rootflags=,
7698 rootfstype= but allow mounting a specific file system to
7701 * The $NOTIFY_SOCKET is now also passed to control processes of
7702 services, not only the main process.
7704 * This version reenables support for fsck's -l switch. This
7705 means at least version v2.25 of util-linux is required for
7706 operation, otherwise dead-locks on device nodes may
7707 occur. Again: you need to update util-linux to at least
7708 v2.25 when updating systemd to v217.
7710 * The "multi-seat-x" tool has been removed from systemd, as
7711 its functionality has been integrated into X servers 1.16,
7712 and the tool is hence redundant. It is recommended to update
7713 display managers invoking this tool to simply invoke X
7714 directly from now on, again.
7716 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
7717 message flag has been added for all of systemd's polkit
7718 authenticated method calls has been added. In particular this
7719 now allows optional interactive authorization via polkit for
7720 many of PID1's privileged operations such as unit file
7721 enabling and disabling.
7723 * "udevadm hwdb --update" learnt a new switch "--usr" for
7724 placing the rebuilt hardware database in /usr instead of
7725 /etc. When used only hardware database entries stored in
7726 /usr will be used, and any user database entries in /etc are
7727 ignored. This functionality is useful for vendors to ship a
7728 pre-built database on systems where local configuration is
7729 unnecessary or unlikely.
7731 * Calendar time specifications in .timer units now also
7732 understand the strings "semi-annually", "quarterly" and
7733 "minutely" as shortcuts (in addition to the preexisting
7734 "annually", "hourly", ...).
7736 * systemd-tmpfiles will now correctly create files in /dev
7737 at boot which are marked for creation only at boot. It is
7738 recommended to always create static device nodes with 'c!'
7739 and 'b!', so that they are created only at boot and not
7740 overwritten at runtime.
7742 * When the watchdog logic is used for a service (WatchdogSec=)
7743 and the watchdog timeout is hit the service will now be
7744 terminated with SIGABRT (instead of just SIGTERM), in order
7745 to make sure a proper coredump and backtrace is
7746 generated. This ensures that hanging services will result in
7747 similar coredump/backtrace behaviour as services that hit a
7750 Contributions from: Andreas Henriksson, Andrei Borzenkov,
7751 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
7752 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
7753 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
7754 Herrmann, David Sommerseth, David Strauss, Emil Renner
7755 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
7756 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
7757 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
7758 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
7759 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
7760 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
7761 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
7762 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
7763 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
7764 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
7765 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
7766 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
7767 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
7768 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
7769 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
7770 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
7773 — Berlin, 2014-10-28
7777 * timedated no longer reads NTP implementation unit names from
7778 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
7779 implementations should add a
7781 Conflicts=systemd-timesyncd.service
7783 to their unit files to take over and replace systemd's NTP
7784 default functionality.
7786 * systemd-sysusers gained a new line type "r" for configuring
7787 which UID/GID ranges to allocate system users/groups
7788 from. Lines of type "u" may now add an additional column
7789 that specifies the home directory for the system user to be
7790 created. Also, systemd-sysusers may now optionally read user
7791 information from STDIN instead of a file. This is useful for
7792 invoking it from RPM preinst scriptlets that need to create
7793 users before the first RPM file is installed since these
7794 files might need to be owned by them. A new
7795 %sysusers_create_inline RPM macro has been introduced to do
7796 just that. systemd-sysusers now updates the shadow files as
7797 well as the user/group databases, which should enhance
7798 compatibility with certain tools like grpck.
7800 * A number of bus APIs of PID 1 now optionally consult polkit to
7801 permit access for otherwise unprivileged clients under certain
7802 conditions. Note that this currently doesn't support
7803 interactive authentication yet, but this is expected to be
7804 added eventually, too.
7806 * /etc/machine-info now has new fields for configuring the
7807 deployment environment of the machine, as well as the
7808 location of the machine. hostnamectl has been updated with
7809 new command to update these fields.
7811 * systemd-timesyncd has been updated to automatically acquire
7812 NTP server information from systemd-networkd, which might
7813 have been discovered via DHCP.
7815 * systemd-resolved now includes a caching DNS stub resolver
7816 and a complete LLMNR name resolution implementation. A new
7817 NSS module "nss-resolve" has been added which can be used
7818 instead of glibc's own "nss-dns" to resolve hostnames via
7819 systemd-resolved. Hostnames, addresses and arbitrary RRs may
7820 be resolved via systemd-resolved D-Bus APIs. In contrast to
7821 the glibc internal resolver systemd-resolved is aware of
7822 multi-homed system, and keeps DNS server and caches separate
7823 and per-interface. Queries are sent simultaneously on all
7824 interfaces that have DNS servers configured, in order to
7825 properly handle VPNs and local LANs which might resolve
7826 separate sets of domain names. systemd-resolved may acquire
7827 DNS server information from systemd-networkd automatically,
7828 which in turn might have discovered them via DHCP. A tool
7829 "systemd-resolve-host" has been added that may be used to
7830 query the DNS logic in resolved. systemd-resolved implements
7831 IDNA and automatically uses IDNA or UTF-8 encoding depending
7832 on whether classic DNS or LLMNR is used as transport. In the
7833 next releases we intend to add a DNSSEC and mDNS/DNS-SD
7834 implementation to systemd-resolved.
7836 * A new NSS module nss-mymachines has been added, that
7837 automatically resolves the names of all local registered
7838 containers to their respective IP addresses.
7840 * A new client tool "networkctl" for systemd-networkd has been
7841 added. It currently is entirely passive and will query
7842 networking configuration from udev, rtnetlink and networkd,
7843 and present it to the user in a very friendly
7844 way. Eventually, we hope to extend it to become a full
7845 control utility for networkd.
7847 * .socket units gained a new DeferAcceptSec= setting that
7848 controls the kernels' TCP_DEFER_ACCEPT sockopt for
7849 TCP. Similarly, support for controlling TCP keep-alive
7850 settings has been added (KeepAliveTimeSec=,
7851 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
7852 turning off Nagle's algorithm on TCP has been added
7855 * logind learned a new session type "web", for use in projects
7856 like Cockpit which register web clients as PAM sessions.
7858 * timer units with at least one OnCalendar= setting will now
7859 be started only after time-sync.target has been
7860 reached. This way they will not elapse before the system
7861 clock has been corrected by a local NTP client or
7862 similar. This is particular useful on RTC-less embedded
7863 machines, that come up with an invalid system clock.
7865 * systemd-nspawn's --network-veth= switch should now result in
7866 stable MAC addresses for both the outer and the inner side
7869 * systemd-nspawn gained a new --volatile= switch for running
7870 container instances with /etc or /var unpopulated.
7872 * The kdbus client code has been updated to use the new Linux
7873 3.17 memfd subsystem instead of the old kdbus-specific one.
7875 * systemd-networkd's DHCP client and server now support
7876 FORCERENEW. There are also new configuration options to
7877 configure the vendor client identifier and broadcast mode
7880 * systemd will no longer inform the kernel about the current
7881 timezone, as this is necessarily incorrect and racy as the
7882 kernel has no understanding of DST and similar
7883 concepts. This hence means FAT timestamps will be always
7884 considered UTC, similar to what Android is already
7885 doing. Also, when the RTC is configured to the local time
7886 (rather than UTC) systemd will never synchronize back to it,
7887 as this might confuse Windows at a later boot.
7889 * systemd-analyze gained a new command "verify" for offline
7890 validation of unit files.
7892 * systemd-networkd gained support for a couple of additional
7893 settings for bonding networking setups. Also, the metric for
7894 statically configured routes may now be configured. For
7895 network interfaces where this is appropriate the peer IP
7896 address may now be configured.
7898 * systemd-networkd's DHCP client will no longer request
7899 broadcasting by default, as this tripped up some networks.
7900 For hardware where broadcast is required the feature should
7901 be switched back on using RequestBroadcast=yes.
7903 * systemd-networkd will now set up IPv4LL addresses (when
7904 enabled) even if DHCP is configured successfully.
7906 * udev will now default to respect network device names given
7907 by the kernel when the kernel indicates that these are
7908 predictable. This behavior can be tweaked by changing
7909 NamePolicy= in the relevant .link file.
7911 * A new library systemd-terminal has been added that
7912 implements full TTY stream parsing and rendering. This
7913 library is supposed to be used later on for implementing a
7914 full userspace VT subsystem, replacing the current kernel
7917 * A new tool systemd-journal-upload has been added to push
7918 journal data to a remote system running
7919 systemd-journal-remote.
7921 * journald will no longer forward all local data to another
7922 running syslog daemon. This change has been made because
7923 rsyslog (which appears to be the most commonly used syslog
7924 implementation these days) no longer makes use of this, and
7925 instead pulls the data out of the journal on its own. Since
7926 forwarding the messages to a non-existent syslog server is
7927 more expensive than we assumed we have now turned this
7928 off. If you run a syslog server that is not a recent rsyslog
7929 version, you have to turn this option on again
7930 (ForwardToSyslog= in journald.conf).
7932 * journald now optionally supports the LZ4 compressor for
7933 larger journal fields. This compressor should perform much
7934 better than XZ which was the previous default.
7936 * machinectl now shows the IP addresses of local containers,
7937 if it knows them, plus the interface name of the container.
7939 * A new tool "systemd-escape" has been added that makes it
7940 easy to escape strings to build unit names and similar.
7942 * sd_notify() messages may now include a new ERRNO= field
7943 which is parsed and collected by systemd and shown among the
7944 "systemctl status" output for a service.
7946 * A new component "systemd-firstboot" has been added that
7947 queries the most basic systemd information (timezone,
7948 hostname, root password) interactively on first
7949 boot. Alternatively it may also be used to provision these
7950 things offline on OS images installed into directories.
7952 * The default sysctl.d/ snippets will now set
7954 net.ipv4.conf.default.promote_secondaries=1
7956 This has the benefit of no flushing secondary IP addresses
7957 when primary addresses are removed.
7959 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
7960 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
7961 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
7962 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
7963 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
7964 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
7965 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
7966 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
7967 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
7968 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
7969 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
7970 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
7971 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
7972 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
7973 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
7975 — Berlin, 2014-08-19
7979 * A new tool systemd-sysusers has been added. This tool
7980 creates system users and groups in /etc/passwd and
7981 /etc/group, based on static declarative system user/group
7982 definitions in /usr/lib/sysusers.d/. This is useful to
7983 enable factory resets and volatile systems that boot up with
7984 an empty /etc directory, and thus need system users and
7985 groups created during early boot. systemd now also ships
7986 with two default sysusers.d/ files for the most basic
7987 users and groups systemd and the core operating system
7990 * A new tmpfiles snippet has been added that rebuilds the
7991 essential files in /etc on boot, should they be missing.
7993 * A directive for ensuring automatic clean-up of
7994 /var/cache/man/ has been removed from the default
7995 configuration. This line should now be shipped by the man
7996 implementation. The necessary change has been made to the
7997 man-db implementation. Note that you need to update your man
7998 implementation to one that ships this line, otherwise no
7999 automatic clean-up of /var/cache/man will take place.
8001 * A new condition ConditionNeedsUpdate= has been added that
8002 may conditionalize services to only run when /etc or /var
8003 are "older" than the vendor operating system resources in
8004 /usr. This is useful for reconstructing or updating /etc
8005 after an offline update of /usr or a factory reset, on the
8006 next reboot. Services that want to run once after such an
8007 update or reset should use this condition and order
8008 themselves before the new systemd-update-done.service, which
8009 will mark the two directories as fully updated. A number of
8010 service files have been added making use of this, to rebuild
8011 the udev hardware database, the journald message catalog and
8012 dynamic loader cache (ldconfig). The systemd-sysusers tool
8013 described above also makes use of this now. With this in
8014 place it is now possible to start up a minimal operating
8015 system with /etc empty cleanly. For more information on the
8016 concepts involved see this recent blog story:
8018 http://0pointer.de/blog/projects/stateless.html
8020 * A new system group "input" has been introduced, and all
8021 input device nodes get this group assigned. This is useful
8022 for system-level software to get access to input devices. It
8023 complements what is already done for "audio" and "video".
8025 * systemd-networkd learnt minimal DHCPv4 server support in
8026 addition to the existing DHCPv4 client support. It also
8027 learnt DHCPv6 client and IPv6 Router Solicitation client
8028 support. The DHCPv4 client gained support for static routes
8029 passed in from the server. Note that the [DHCPv4] section
8030 known in older systemd-networkd versions has been renamed to
8031 [DHCP] and is now also used by the DHCPv6 client. Existing
8032 .network files using settings of this section should be
8033 updated, though compatibility is maintained. Optionally, the
8034 client hostname may now be sent to the DHCP server.
8036 * networkd gained support for vxlan virtual networks as well
8037 as tun/tap and dummy devices.
8039 * networkd gained support for automatic allocation of address
8040 ranges for interfaces from a system-wide pool of
8041 addresses. This is useful for dynamically managing a large
8042 number of interfaces with a single network configuration
8043 file. In particular this is useful to easily assign
8044 appropriate IP addresses to the veth links of a large number
8045 of nspawn instances.
8047 * RPM macros for processing sysusers, sysctl and binfmt
8048 drop-in snippets at package installation time have been
8051 * The /etc/os-release file should now be placed in
8052 /usr/lib/os-release. The old location is automatically
8053 created as symlink. /usr/lib is the more appropriate
8054 location of this file, since it shall actually describe the
8055 vendor operating system shipped in /usr, and not the
8056 configuration stored in /etc.
8058 * .mount units gained a new boolean SloppyOptions= setting
8059 that maps to mount(8)'s -s option which enables permissive
8060 parsing of unknown mount options.
8062 * tmpfiles learnt a new "L+" directive which creates a symlink
8063 but (unlike "L") deletes a pre-existing file first, should
8064 it already exist and not already be the correct
8065 symlink. Similarly, "b+", "c+" and "p+" directives have been
8066 added as well, which create block and character devices, as
8067 well as fifos in the filesystem, possibly removing any
8068 pre-existing files of different types.
8070 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
8071 'argument' field (which so far specified the source to
8072 symlink/copy the files from) is now optional. If omitted the
8073 same file os copied from /usr/share/factory/ suffixed by the
8074 full destination path. This is useful for populating /etc
8075 with essential files, by copying them from vendor defaults
8076 shipped in /usr/share/factory/etc.
8078 * A new command "systemctl preset-all" has been added that
8079 applies the service preset settings to all installed unit
8080 files. A new switch --preset-mode= has been added that
8081 controls whether only enable or only disable operations
8084 * A new command "systemctl is-system-running" has been added
8085 that allows checking the overall state of the system, for
8086 example whether it is fully up and running.
8088 * When the system boots up with an empty /etc, the equivalent
8089 to "systemctl preset-all" is executed during early boot, to
8090 make sure all default services are enabled after a factory
8093 * systemd now contains a minimal preset file that enables the
8094 most basic services systemd ships by default.
8096 * Unit files' [Install] section gained a new DefaultInstance=
8097 field for defining the default instance to create if a
8098 template unit is enabled with no instance specified.
8100 * A new passive target cryptsetup-pre.target has been added
8101 that may be used by services that need to make they run and
8102 finish before the first LUKS cryptographic device is set up.
8104 * The /dev/loop-control and /dev/btrfs-control device nodes
8105 are now owned by the "disk" group by default, opening up
8106 access to this group.
8108 * systemd-coredump will now automatically generate a
8109 stack trace of all core dumps taking place on the system,
8110 based on elfutils' libdw library. This stack trace is logged
8113 * systemd-coredump may now optionally store coredumps directly
8114 on disk (in /var/lib/systemd/coredump, possibly compressed),
8115 instead of storing them unconditionally in the journal. This
8116 mode is the new default. A new configuration file
8117 /etc/systemd/coredump.conf has been added to configure this
8118 and other parameters of systemd-coredump.
8120 * coredumpctl gained a new "info" verb to show details about a
8121 specific coredump. A new switch "-1" has also been added
8122 that makes sure to only show information about the most
8123 recent entry instead of all entries. Also, as the tool is
8124 generally useful now the "systemd-" prefix of the binary
8125 name has been removed. Distributions that want to maintain
8126 compatibility with the old name should add a symlink from
8127 the old name to the new name.
8129 * journald's SplitMode= now defaults to "uid". This makes sure
8130 that unprivileged users can access their own coredumps with
8131 coredumpctl without restrictions.
8133 * New kernel command line options "systemd.wants=" (for
8134 pulling an additional unit during boot), "systemd.mask="
8135 (for masking a specific unit for the boot), and
8136 "systemd.debug-shell" (for enabling the debug shell on tty9)
8137 have been added. This is implemented in the new generator
8138 "systemd-debug-generator".
8140 * systemd-nspawn will now by default filter a couple of
8141 syscalls for containers, among them those required for
8142 kernel module loading, direct x86 IO port access, swap
8143 management, and kexec. Most importantly though
8144 open_by_handle_at() is now prohibited for containers,
8145 closing a hole similar to a recently discussed vulnerability
8146 in docker regarding access to files on file hierarchies the
8147 container should normally not have access to. Note that, for
8148 nspawn, we generally make no security claims anyway (and
8149 this is explicitly documented in the man page), so this is
8150 just a fix for one of the most obvious problems.
8152 * A new man page file-hierarchy(7) has been added that
8153 contains a minimized, modernized version of the file system
8154 layout systemd expects, similar in style to the FHS
8155 specification or hier(5). A new tool systemd-path(1) has
8156 been added to query many of these paths for the local
8159 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
8160 longer done. Since the directory now has a per-user size
8161 limit, and is cleaned on logout this appears unnecessary,
8162 in particular since this now brings the lifecycle of this
8163 directory closer in line with how IPC objects are handled.
8165 * systemd.pc now exports a number of additional directories,
8166 including $libdir (which is useful to identify the library
8167 path for the primary architecture of the system), and a
8168 couple of drop-in directories.
8170 * udev's predictable network interface names now use the dev_port
8171 sysfs attribute, introduced in linux 3.15 instead of dev_id to
8172 distinguish between ports of the same PCI function. dev_id should
8173 only be used for ports using the same HW address, hence the need
8176 * machined has been updated to export the OS version of a
8177 container (read from /etc/os-release and
8178 /usr/lib/os-release) on the bus. This is now shown in
8179 "machinectl status" for a machine.
8181 * A new service setting RestartForceExitStatus= has been
8182 added. If configured to a set of exit signals or process
8183 return values, the service will be restarted when the main
8184 daemon process exits with any of them, regardless of the
8187 * systemctl's -H switch for connecting to remote systemd
8188 machines has been extended so that it may be used to
8189 directly connect to a specific container on the
8190 host. "systemctl -H root@foobar:waldi" will now connect as
8191 user "root" to host "foobar", and then proceed directly to
8192 the container named "waldi". Note that currently you have to
8193 authenticate as user "root" for this to work, as entering
8194 containers is a privileged operation.
8196 Contributions from: Andreas Henriksson, Benjamin Steinwender,
8197 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
8198 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
8199 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
8200 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
8201 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
8202 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
8203 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
8204 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
8205 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
8206 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
8207 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
8209 — Berlin, 2014-07-03
8213 * As an experimental feature, udev now tries to lock the
8214 disk device node (flock(LOCK_SH|LOCK_NB)) while it
8215 executes events for the disk or any of its partitions.
8216 Applications like partitioning programs can lock the
8217 disk device node (flock(LOCK_EX)) and claim temporary
8218 device ownership that way; udev will entirely skip all event
8219 handling for this disk and its partitions. If the disk
8220 was opened for writing, the close will trigger a partition
8221 table rescan in udev's "watch" facility, and if needed
8222 synthesize "change" events for the disk and all its partitions.
8223 This is now unconditionally enabled, and if it turns out to
8224 cause major problems, we might turn it on only for specific
8225 devices, or might need to disable it entirely. Device Mapper
8226 devices are excluded from this logic.
8228 * We temporarily dropped the "-l" switch for fsck invocations,
8229 since they collide with the flock() logic above. util-linux
8230 upstream has been changed already to avoid this conflict,
8231 and we will re-add "-l" as soon as util-linux with this
8232 change has been released.
8234 * The dependency on libattr has been removed. Since a long
8235 time, the extended attribute calls have moved to glibc, and
8236 libattr is thus unnecessary.
8238 * Virtualization detection works without privileges now. This
8239 means the systemd-detect-virt binary no longer requires
8240 CAP_SYS_PTRACE file capabilities, and our daemons can run
8241 with fewer privileges.
8243 * systemd-networkd now runs under its own "systemd-network"
8244 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
8245 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
8246 loses the ability to write to files owned by root this way.
8248 * Similarly, systemd-resolved now runs under its own
8249 "systemd-resolve" user with no capabilities remaining.
8251 * Similarly, systemd-bus-proxyd now runs under its own
8252 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
8254 * systemd-networkd gained support for setting up "veth"
8255 virtual Ethernet devices for container connectivity, as well
8256 as GRE and VTI tunnels.
8258 * systemd-networkd will no longer automatically attempt to
8259 manually load kernel modules necessary for certain tunnel
8260 transports. Instead, it is assumed the kernel loads them
8261 automatically when required. This only works correctly on
8262 very new kernels. On older kernels, please consider adding
8263 the kernel modules to /etc/modules-load.d/ as a work-around.
8265 * The resolv.conf file systemd-resolved generates has been
8266 moved to /run/systemd/resolve/. If you have a symlink from
8267 /etc/resolv.conf, it might be necessary to correct it.
8269 * Two new service settings, ProtectHome= and ProtectSystem=,
8270 have been added. When enabled, they will make the user data
8271 (such as /home) inaccessible or read-only and the system
8272 (such as /usr) read-only, for specific services. This allows
8273 very light-weight per-service sandboxing to avoid
8274 modifications of user data or system files from
8275 services. These two new switches have been enabled for all
8276 of systemd's long-running services, where appropriate.
8278 * Socket units gained new SocketUser= and SocketGroup=
8279 settings to set the owner user and group of AF_UNIX sockets
8280 and FIFOs in the file system.
8282 * Socket units gained a new RemoveOnStop= setting. If enabled,
8283 all FIFOS and sockets in the file system will be removed
8284 when the specific socket unit is stopped.
8286 * Socket units gained a new Symlinks= setting. It takes a list
8287 of symlinks to create to file system sockets or FIFOs
8288 created by the specific Unix sockets. This is useful to
8289 manage symlinks to socket nodes with the same lifecycle as
8292 * The /dev/log socket and /dev/initctl FIFO have been moved to
8293 /run, and have been replaced by symlinks. This allows
8294 connecting to these facilities even if PrivateDevices=yes is
8295 used for a service (which makes /dev/log itself unavailable,
8296 but /run is left). This also has the benefit of ensuring
8297 that /dev only contains device nodes, directories and
8298 symlinks, and nothing else.
8300 * sd-daemon gained two new calls sd_pid_notify() and
8301 sd_pid_notifyf(). They are similar to sd_notify() and
8302 sd_notifyf(), but allow overriding of the source PID of
8303 notification messages if permissions permit this. This is
8304 useful to send notify messages on behalf of a different
8305 process (for example, the parent process). The
8306 systemd-notify tool has been updated to make use of this
8307 when sending messages (so that notification messages now
8308 originate from the shell script invoking systemd-notify and
8309 not the systemd-notify process itself. This should minimize
8310 a race where systemd fails to associate notification
8311 messages to services when the originating process already
8314 * A new "on-abnormal" setting for Restart= has been added. If
8315 set, it will result in automatic restarts on all "abnormal"
8316 reasons for a process to exit, which includes unclean
8317 signals, core dumps, timeouts and watchdog timeouts, but
8318 does not include clean and unclean exit codes or clean
8319 signals. Restart=on-abnormal is an alternative for
8320 Restart=on-failure for services that shall be able to
8321 terminate and avoid restarts on certain errors, by
8322 indicating so with an unclean exit code. Restart=on-failure
8323 or Restart=on-abnormal is now the recommended setting for
8324 all long-running services.
8326 * If the InaccessibleDirectories= service setting points to a
8327 mount point (or if there are any submounts contained within
8328 it), it is now attempted to completely unmount it, to make
8329 the file systems truly unavailable for the respective
8332 * The ReadOnlyDirectories= service setting and
8333 systemd-nspawn's --read-only parameter are now recursively
8334 applied to all submounts, too.
8336 * Mount units may now be created transiently via the bus APIs.
8338 * The support for SysV and LSB init scripts has been removed
8339 from the systemd daemon itself. Instead, it is now
8340 implemented as a generator that creates native systemd units
8341 from these scripts when needed. This enables us to remove a
8342 substantial amount of legacy code from PID 1, following the
8343 fact that many distributions only ship a very small number
8344 of LSB/SysV init scripts nowadays.
8346 * Privileged Xen (dom0) domains are not considered
8347 virtualization anymore by the virtualization detection
8348 logic. After all, they generally have unrestricted access to
8349 the hardware and usually are used to manage the unprivileged
8352 * systemd-tmpfiles gained a new "C" line type, for copying
8353 files or entire directories.
8355 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
8356 lines. So far, they have been non-globbing versions of the
8357 latter, and have thus been redundant. In future, it is
8358 recommended to only use "z". "m" has hence been removed
8359 from the documentation, even though it stays supported.
8361 * A tmpfiles snippet to recreate the most basic structure in
8362 /var has been added. This is enough to create the /var/run →
8363 /run symlink and create a couple of structural
8364 directories. This allows systems to boot up with an empty or
8365 volatile /var. Of course, while with this change, the core OS
8366 now is capable with dealing with a volatile /var, not all
8367 user services are ready for it. However, we hope that sooner
8368 or later, many service daemons will be changed upstream so
8369 that they are able to automatically create their necessary
8370 directories in /var at boot, should they be missing. This is
8371 the first step to allow state-less systems that only require
8372 the vendor image for /usr to boot.
8374 * systemd-nspawn has gained a new --tmpfs= switch to mount an
8375 empty tmpfs instance to a specific directory. This is
8376 particularly useful for making use of the automatic
8377 reconstruction of /var (see above), by passing --tmpfs=/var.
8379 * Access modes specified in tmpfiles snippets may now be
8380 prefixed with "~", which indicates that they shall be masked
8381 by whether the existing file or directory is currently
8382 writable, readable or executable at all. Also, if specified,
8383 the sgid/suid/sticky bits will be masked for all
8386 * A new passive target unit "network-pre.target" has been
8387 added which is useful for services that shall run before any
8388 network is configured, for example firewall scripts.
8390 * The "floppy" group that previously owned the /dev/fd*
8391 devices is no longer used. The "disk" group is now used
8392 instead. Distributions should probably deprecate usage of
8395 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
8396 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
8397 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
8398 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
8399 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
8400 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
8403 — Berlin, 2014-06-11
8407 * A new "systemd-timesyncd" daemon has been added for
8408 synchronizing the system clock across the network. It
8409 implements an SNTP client. In contrast to NTP
8410 implementations such as chrony or the NTP reference server,
8411 this only implements a client side, and does not bother with
8412 the full NTP complexity, focusing only on querying time from
8413 one remote server and synchronizing the local clock to
8414 it. Unless you intend to serve NTP to networked clients or
8415 want to connect to local hardware clocks, this simple NTP
8416 client should be more than appropriate for most
8417 installations. The daemon runs with minimal privileges, and
8418 has been hooked up with networkd to only operate when
8419 network connectivity is available. The daemon saves the
8420 current clock to disk every time a new NTP sync has been
8421 acquired, and uses this to possibly correct the system clock
8422 early at bootup, in order to accommodate for systems that
8423 lack an RTC such as the Raspberry Pi and embedded devices,
8424 and to make sure that time monotonically progresses on these
8425 systems, even if it is not always correct. To make use of
8426 this daemon, a new system user and group "systemd-timesync"
8427 needs to be created on installation of systemd.
8429 * The queue "seqnum" interface of libudev has been disabled, as
8430 it was generally incompatible with device namespacing as
8431 sequence numbers of devices go "missing" if the devices are
8432 part of a different namespace.
8434 * "systemctl list-timers" and "systemctl list-sockets" gained
8435 a --recursive switch for showing units of these types also
8436 for all local containers, similar in style to the already
8437 supported --recursive switch for "systemctl list-units".
8439 * A new RebootArgument= setting has been added for service
8440 units, which may be used to specify a kernel reboot argument
8441 to use when triggering reboots with StartLimitAction=.
8443 * A new FailureAction= setting has been added for service
8444 units which may be used to specify an operation to trigger
8445 when a service fails. This works similarly to
8446 StartLimitAction=, but unlike it, controls what is done
8447 immediately rather than only after several attempts to
8448 restart the service in question.
8450 * hostnamed got updated to also expose the kernel name,
8451 release, and version on the bus. This is useful for
8452 executing commands like hostnamectl with the -H switch.
8453 systemd-analyze makes use of this to properly display
8454 details when running non-locally.
8456 * The bootchart tool can now show cgroup information in the
8457 graphs it generates.
8459 * The CFS CPU quota cgroup attribute is now exposed for
8460 services. The new CPUQuota= switch has been added for this
8461 which takes a percentage value. Setting this will have the
8462 result that a service may never get more CPU time than the
8463 specified percentage, even if the machine is otherwise idle.
8465 * systemd-networkd learned IPIP and SIT tunnel support.
8467 * LSB init scripts exposing a dependency on $network will now
8468 get a dependency on network-online.target rather than simply
8469 network.target. This should bring LSB handling closer to
8470 what it was on SysV systems.
8472 * A new fsck.repair= kernel option has been added to control
8473 how fsck shall deal with unclean file systems at boot.
8475 * The (.ini) configuration file parser will now silently ignore
8476 sections whose names begin with "X-". This may be used to maintain
8477 application-specific extension sections in unit files.
8479 * machined gained a new API to query the IP addresses of
8480 registered containers. "machinectl status" has been updated
8481 to show these addresses in its output.
8483 * A new call sd_uid_get_display() has been added to the
8484 sd-login APIs for querying the "primary" session of a
8485 user. The "primary" session of the user is elected from the
8486 user's sessions and generally a graphical session is
8487 preferred over a text one.
8489 * A minimal systemd-resolved daemon has been added. It
8490 currently simply acts as a companion to systemd-networkd and
8491 manages resolv.conf based on per-interface DNS
8492 configuration, possibly supplied via DHCP. In the long run
8493 we hope to extend this into a local DNSSEC enabled DNS and
8496 * The systemd-networkd-wait-online tool is now enabled by
8497 default. It will delay network-online.target until a network
8498 connection has been configured. The tool primarily integrates
8499 with networkd, but will also make a best effort to make sense
8500 of network configuration performed in some other way.
8502 * Two new service options StartupCPUShares= and
8503 StartupBlockIOWeight= have been added that work similarly to
8504 CPUShares= and BlockIOWeight= however only apply during
8505 system startup. This is useful to prioritize certain services
8506 differently during bootup than during normal runtime.
8508 * hostnamed has been changed to prefer the statically
8509 configured hostname in /etc/hostname (unless set to
8510 'localhost' or empty) over any dynamic one supplied by
8511 dhcp. With this change, the rules for picking the hostname
8512 match more closely the rules of other configuration settings
8513 where the local administrator's configuration in /etc always
8514 overrides any other settings.
8516 Contributions from: Ali H. Caliskan, Alison Chaiken, Bas van
8517 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
8518 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
8519 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
8520 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
8521 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
8522 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
8523 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
8524 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
8525 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
8526 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
8527 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
8528 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
8529 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
8530 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
8531 Lindskog, WaLyong Cho, Will Woods, Zbigniew
8534 — Beijing, 2014-05-28
8538 * When restoring the screen brightness at boot, stay away from
8539 the darkest setting or from the lowest 5% of the available
8540 range, depending on which is the larger value of both. This
8541 should effectively protect the user from rebooting into a
8542 black screen, should the brightness have been set to minimum
8545 * sd-login gained a new sd_machine_get_class() call to
8546 determine the class ("vm" or "container") of a machine
8547 registered with machined.
8549 * sd-login gained new calls
8550 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
8551 to query the identity of the peer of a local AF_UNIX
8552 connection. They operate similarly to their sd_pid_get_xyz()
8555 * PID 1 will now maintain a system-wide system state engine
8556 with the states "starting", "running", "degraded",
8557 "maintenance", "stopping". These states are bound to system
8558 startup, normal runtime, runtime with at least one failed
8559 service, rescue/emergency mode and system shutdown. This
8560 state is shown in the "systemctl status" output when no unit
8561 name is passed. It is useful to determine system state, in
8562 particularly when doing so for many systems or containers at
8565 * A new command "list-machines" has been added to "systemctl"
8566 that lists all local OS containers and shows their system
8567 state (see above), if systemd runs inside of them.
8569 * systemctl gained a new "-r" switch to recursively enumerate
8570 units on all local containers, when used with the
8571 "list-unit" command (which is the default one that is
8572 executed when no parameters are specified).
8574 * The GPT automatic partition discovery logic will now honour
8575 two GPT partition flags: one may be set on a partition to
8576 cause it to be mounted read-only, and the other may be set
8577 on a partition to ignore it during automatic discovery.
8579 * Two new GPT type UUIDs have been added for automatic root
8580 partition discovery, for 32-bit and 64-bit ARM. This is not
8581 particularly useful for discovering the root directory on
8582 these architectures during bare-metal boots (since UEFI is
8583 not common there), but still very useful to allow booting of
8584 ARM disk images in nspawn with the -i option.
8586 * MAC addresses of interfaces created with nspawn's
8587 --network-interface= switch will now be generated from the
8588 machine name, and thus be stable between multiple invocations
8591 * logind will now automatically remove all IPC objects owned
8592 by a user if she or he fully logs out. This makes sure that
8593 users who are logged out cannot continue to consume IPC
8594 resources. This covers SysV memory, semaphores and message
8595 queues as well as POSIX shared memory and message
8596 queues. Traditionally, SysV and POSIX IPC had no lifecycle
8597 limits. With this functionality, that is corrected. This may
8598 be turned off by using the RemoveIPC= switch of logind.conf.
8600 * The systemd-machine-id-setup and tmpfiles tools gained a
8601 --root= switch to operate on a specific root directory,
8604 * journald can now forward logged messages to the TTYs of all
8605 logged in users ("wall"). This is the default for all
8606 emergency messages now.
8608 * A new tool systemd-journal-remote has been added to stream
8609 journal log messages across the network.
8611 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
8612 controller trees are mounted into it. Note that the
8613 directories mounted beneath it are not read-only. This is a
8614 security measure and is particularly useful because glibc
8615 actually includes a search logic to pick any tmpfs it can
8616 find to implement shm_open() if /dev/shm is not available
8617 (which it might very well be in namespaced setups).
8619 * machinectl gained a new "poweroff" command to cleanly power
8620 down a local OS container.
8622 * The PrivateDevices= unit file setting will now also drop the
8623 CAP_MKNOD capability from the capability bound set, and
8624 imply DevicePolicy=closed.
8626 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
8627 comprehensively on all long-running systemd services where
8628 this is appropriate.
8630 * systemd-udevd will now run in a disassociated mount
8631 namespace. To mount directories from udev rules, make sure to
8632 pull in mount units via SYSTEMD_WANTS properties.
8634 * The kdbus support gained support for uploading policy into
8635 the kernel. sd-bus gained support for creating "monitoring"
8636 connections that can eavesdrop into all bus communication
8637 for debugging purposes.
8639 * Timestamps may now be specified in seconds since the UNIX
8640 epoch Jan 1st, 1970 by specifying "@" followed by the value
8643 * Native tcpwrap support in systemd has been removed. tcpwrap
8644 is old code, not really maintained anymore and has serious
8645 shortcomings, and better options such as firewalls
8646 exist. For setups that require tcpwrap usage, please
8647 consider invoking your socket-activated service via tcpd,
8648 like on traditional inetd.
8650 * A new system.conf configuration option
8651 DefaultTimerAccuracySec= has been added that controls the
8652 default AccuracySec= setting of .timer units.
8654 * Timer units gained a new WakeSystem= switch. If enabled,
8655 timers configured this way will cause the system to resume
8656 from system suspend (if the system supports that, which most
8659 * Timer units gained a new Persistent= switch. If enabled,
8660 timers configured this way will save to disk when they have
8661 been last triggered. This information is then used on next
8662 reboot to possible execute overdue timer events, that
8663 could not take place because the system was powered off.
8664 This enables simple anacron-like behaviour for timer units.
8666 * systemctl's "list-timers" will now also list the time a
8667 timer unit was last triggered in addition to the next time
8668 it will be triggered.
8670 * systemd-networkd will now assign predictable IPv4LL
8671 addresses to its local interfaces.
8673 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
8674 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
8675 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
8676 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
8677 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
8678 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
8679 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
8680 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
8683 — Berlin, 2014-03-25
8687 * A new unit file setting RestrictAddressFamilies= has been
8688 added to restrict which socket address families unit
8689 processes gain access to. This takes address family names
8690 like "AF_INET" or "AF_UNIX", and is useful to minimize the
8691 attack surface of services via exotic protocol stacks. This
8692 is built on seccomp system call filters.
8694 * Two new unit file settings RuntimeDirectory= and
8695 RuntimeDirectoryMode= have been added that may be used to
8696 manage a per-daemon runtime directories below /run. This is
8697 an alternative for setting up directory permissions with
8698 tmpfiles snippets, and has the advantage that the runtime
8699 directory's lifetime is bound to the daemon runtime and that
8700 the daemon starts up with an empty directory each time. This
8701 is particularly useful when writing services that drop
8702 privileges using the User= or Group= setting.
8704 * The DeviceAllow= unit setting now supports globbing for
8705 matching against device group names.
8707 * The systemd configuration file system.conf gained new
8708 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
8709 DefaultMemoryAccounting= to globally turn on/off accounting
8710 for specific resources (cgroups) for all units. These
8711 settings may still be overridden individually in each unit
8714 * systemd-gpt-auto-generator is now able to discover /srv and
8715 root partitions in addition to /home and swap partitions. It
8716 also supports LUKS-encrypted partitions now. With this in
8717 place, automatic discovery of partitions to mount following
8718 the Discoverable Partitions Specification
8719 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
8720 is now a lot more complete. This allows booting without
8721 /etc/fstab and without root= on the kernel command line on
8722 systems prepared appropriately.
8724 * systemd-nspawn gained a new --image= switch which allows
8725 booting up disk images and Linux installations on any block
8726 device that follow the Discoverable Partitions Specification
8727 (see above). This means that installations made with
8728 appropriately updated installers may now be started and
8729 deployed using container managers, completely
8730 unmodified. (We hope that libvirt-lxc will add support for
8731 this feature soon, too.)
8733 * systemd-nspawn gained a new --network-macvlan= setting to
8734 set up a private macvlan interface for the
8735 container. Similarly, systemd-networkd gained a new
8736 Kind=macvlan setting in .netdev files.
8738 * systemd-networkd now supports configuring local addresses
8741 * A new tool systemd-network-wait-online has been added to
8742 synchronously wait for network connectivity using
8745 * The sd-bus.h bus API gained a new sd_bus_track object for
8746 tracking the lifecycle of bus peers. Note that sd-bus.h is
8747 still not a public API though (unless you specify
8748 --enable-kdbus on the configure command line, which however
8749 voids your warranty and you get no API stability guarantee).
8751 * The $XDG_RUNTIME_DIR runtime directories for each user are
8752 now individual tmpfs instances, which has the benefit of
8753 introducing separate pools for each user, with individual
8754 size limits, and thus making sure that unprivileged clients
8755 can no longer negatively impact the system or other users by
8756 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
8757 RuntimeDirectorySize= has been introduced that allows
8758 controlling the default size limit for all users. It
8759 defaults to 10% of the available physical memory. This is no
8760 replacement for quotas on tmpfs though (which the kernel
8761 still does not support), as /dev/shm and /tmp are still
8762 shared resources used by both the system and unprivileged
8765 * logind will now automatically turn off automatic suspending
8766 on laptop lid close when more than one display is
8767 connected. This was previously expected to be implemented
8768 individually in desktop environments (such as GNOME),
8769 however has been added to logind now, in order to fix a
8770 boot-time race where a desktop environment might not have
8771 been started yet and thus not been able to take an inhibitor
8772 lock at the time where logind already suspends the system
8773 due to a closed lid.
8775 * logind will now wait at least 30s after each system
8776 suspend/resume cycle, and 3min after system boot before
8777 suspending the system due to a closed laptop lid. This
8778 should give USB docking stations and similar enough time to
8779 be probed and configured after system resume and boot in
8780 order to then act as suspend blocker.
8782 * systemd-run gained a new --property= setting which allows
8783 initialization of resource control properties (and others)
8784 for the created scope or service unit. Example: "systemd-run
8785 --property=BlockIOWeight=10 updatedb" may be used to run
8786 updatedb at a low block IO scheduling weight.
8788 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
8789 now also work in --scope mode.
8791 * When systemd is compiled with kdbus support, basic support
8792 for enforced policies is now in place. (Note that enabling
8793 kdbus still voids your warranty and no API compatibility
8796 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
8797 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
8798 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
8799 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
8800 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
8801 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
8802 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
8803 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
8804 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
8805 Zbigniew Jędrzejewski-Szmek
8807 — Berlin, 2014-03-12
8811 * systemd will now relabel /dev after loading the SMACK policy
8812 according to SMACK rules.
8814 * A new unit file option AppArmorProfile= has been added to
8815 set the AppArmor profile for the processes of a unit.
8817 * A new condition check ConditionArchitecture= has been added
8818 to conditionalize units based on the system architecture, as
8819 reported by uname()'s "machine" field.
8821 * systemd-networkd now supports matching on the system
8822 virtualization, architecture, kernel command line, hostname
8825 * logind is now a lot more aggressive when suspending the
8826 machine due to a closed laptop lid. Instead of acting only
8827 on the lid close action, it will continuously watch the lid
8828 status and act on it. This is useful for laptops where the
8829 power button is on the outside of the chassis so that it can
8830 be reached without opening the lid (such as the Lenovo
8831 Yoga). On those machines, logind will now immediately
8832 re-suspend the machine if the power button has been
8833 accidentally pressed while the laptop was suspended and in a
8834 backpack or similar.
8836 * logind will now watch SW_DOCK switches and inhibit reaction
8837 to the lid switch if it is pressed. This means that logind
8838 will not suspend the machine anymore if the lid is closed
8839 and the system is docked, if the laptop supports SW_DOCK
8840 notifications via the input layer. Note that ACPI docking
8841 stations do not generate this currently. Also note that this
8842 logic is usually not fully sufficient and Desktop
8843 Environments should take a lid switch inhibitor lock when an
8844 external display is connected, as systemd will not watch
8847 * nspawn will now make use of the devices cgroup controller by
8848 default, and only permit creation of and access to the usual
8849 API device nodes like /dev/null or /dev/random, as well as
8850 access to (but not creation of) the pty devices.
8852 * We will now ship a default .network file for
8853 systemd-networkd that automatically configures DHCP for
8854 network interfaces created by nspawn's --network-veth or
8855 --network-bridge= switches.
8857 * systemd will now understand the usual M, K, G, T suffixes
8858 according to SI conventions (i.e. to the base 1000) when
8859 referring to throughput and hardware metrics. It will stay
8860 with IEC conventions (i.e. to the base 1024) for software
8861 metrics, according to what is customary according to
8862 Wikipedia. We explicitly document which base applies for
8863 each configuration option.
8865 * The DeviceAllow= setting in unit files now supports a syntax to
8866 allow-list an entire group of devices node majors at once, based on
8867 the /proc/devices listing. For example, with the string "char-pts",
8868 it is now possible to allow-list all current and future pseudo-TTYs
8871 * sd-event learned a new "post" event source. Event sources of
8872 this type are triggered by the dispatching of any event
8873 source of a type that is not "post". This is useful for
8874 implementing clean-up and check event sources that are
8875 triggered by other work being done in the program.
8877 * systemd-networkd is no longer statically enabled, but uses
8878 the usual [Install] sections so that it can be
8879 enabled/disabled using systemctl. It still is enabled by
8882 * When creating a veth interface pair with systemd-nspawn, the
8883 host side will now be prefixed with "vb-" if
8884 --network-bridge= is used, and with "ve-" if --network-veth
8885 is used. This way, it is easy to distinguish these cases on
8886 the host, for example to apply different configuration to
8887 them with systemd-networkd.
8889 * The compatibility libraries for libsystemd-journal.so,
8890 libsystem-id128.so, libsystemd-login.so and
8891 libsystemd-daemon.so do not make use of IFUNC
8892 anymore. Instead, we now build libsystemd.so multiple times
8893 under these alternative names. This means that the footprint
8894 is drastically increased, but given that these are
8895 transitional compatibility libraries, this should not matter
8896 much. This change has been made necessary to support the ARM
8897 platform for these compatibility libraries, as the ARM
8898 toolchain is not really at the same level as the toolchain
8899 for other architectures like x86 and does not support
8900 IFUNC. Please make sure to use --enable-compat-libs only
8901 during a transitional period!
8903 * The .include syntax has been deprecated and is not documented
8904 anymore. Drop-in files in .d directories should be used instead.
8906 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
8907 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
8908 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
8909 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
8910 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
8911 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
8912 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
8913 Zbigniew Jędrzejewski-Szmek
8915 — Berlin, 2014-02-24
8919 * A new component "systemd-networkd" has been added that can
8920 be used to configure local network interfaces statically or
8921 via DHCP. It is capable of bringing up bridges, VLANs, and
8922 bonding. Currently, no hook-ups for interactive network
8923 configuration are provided. Use this for your initrd,
8924 container, embedded, or server setup if you need a simple,
8925 yet powerful, network configuration solution. This
8926 configuration subsystem is quite nifty, as it allows wildcard
8927 hotplug matching in interfaces. For example, with a single
8928 configuration snippet, you can configure that all Ethernet
8929 interfaces showing up are automatically added to a bridge,
8930 or similar. It supports link-sensing and more.
8932 * A new tool "systemd-socket-proxyd" has been added which can
8933 act as a bidirectional proxy for TCP sockets. This is
8934 useful for adding socket activation support to services that
8935 do not actually support socket activation, including virtual
8936 machines and the like.
8938 * Add a new tool to save/restore rfkill state on
8941 * Save/restore state of keyboard backlights in addition to
8942 display backlights on shutdown/boot.
8944 * udev learned a new SECLABEL{} construct to label device
8945 nodes with a specific security label when they appear. For
8946 now, only SECLABEL{selinux} is supported, but the syntax is
8947 prepared for additional security frameworks.
8949 * udev gained a new scheme to configure link-level attributes
8950 from files in /etc/systemd/network/*.link. These files can
8951 match against MAC address, device path, driver name and type,
8952 and will apply attributes like the naming policy, link speed,
8953 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
8954 address assignment policy (randomized, ...).
8956 * The configuration of network interface naming rules for
8957 "permanent interface names" has changed: a new NamePolicy=
8958 setting in the [Link] section of .link files determines the
8959 priority of possible naming schemes (onboard, slot, MAC,
8960 path). The default value of this setting is determined by
8961 /usr/lib/net/links/99-default.link. Old
8962 80-net-name-slot.rules udev configuration file has been
8963 removed, so local configuration overriding this file should
8964 be adapted to override 99-default.link instead.
8966 * When the User= switch is used in a unit file, also
8967 initialize $SHELL= based on the user database entry.
8969 * systemd no longer depends on libdbus. All communication is
8970 now done with sd-bus, systemd's low-level bus library
8973 * kdbus support has been added to PID 1 itself. When kdbus is
8974 enabled, this causes PID 1 to set up the system bus and
8975 enable support for a new ".busname" unit type that
8976 encapsulates bus name activation on kdbus. It works a little
8977 bit like ".socket" units, except for bus names. A new
8978 generator has been added that converts classic dbus1 service
8979 activation files automatically into native systemd .busname
8982 * sd-bus: add a light-weight vtable implementation that allows
8983 defining objects on the bus with a simple static const
8984 vtable array of its methods, signals and properties.
8986 * systemd will not generate or install static dbus
8987 introspection data anymore to /usr/share/dbus-1/interfaces,
8988 as the precise format of these files is unclear, and
8989 nothing makes use of it.
8991 * A proxy daemon is now provided to proxy clients connecting
8992 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
8993 compatibility with classic D-Bus.
8995 * A bus driver implementation has been added that supports the
8996 classic D-Bus bus driver calls on kdbus, also for
8997 compatibility purposes.
8999 * A new API "sd-event.h" has been added that implements a
9000 minimal event loop API built around epoll. It provides a
9001 couple of features that direct epoll usage is lacking:
9002 prioritization of events, scales to large numbers of timer
9003 events, per-event timer slack (accuracy), system-wide
9004 coalescing of timer events, exit handlers, watchdog
9005 supervision support using systemd's sd_notify() API, child
9008 * A new API "sd-rntl.h" has been added that provides an API
9009 around the route netlink interface of the kernel, similar in
9010 style to "sd-bus.h".
9012 * A new API "sd-dhcp-client.h" has been added that provides a
9013 small DHCPv4 client-side implementation. This is used by
9016 * There is a new kernel command line option
9017 "systemd.restore_state=0|1". When set to "0", none of the
9018 systemd tools will restore saved runtime state to hardware
9019 devices. More specifically, the rfkill and backlight states
9022 * The FsckPassNo= compatibility option in mount/service units
9023 has been removed. The fstab generator will now add the
9024 necessary dependencies automatically, and does not require
9025 PID1's support for that anymore.
9027 * journalctl gained a new switch, --list-boots, that lists
9028 recent boots with their times and boot IDs.
9030 * The various tools like systemctl, loginctl, timedatectl,
9031 busctl, systemd-run, ... have gained a new switch "-M" to
9032 connect to a specific, local OS container (as direct
9033 connection, without requiring SSH). This works on any
9034 container that is registered with machined, such as those
9035 created by libvirt-lxc or nspawn.
9037 * systemd-run and systemd-analyze also gained support for "-H"
9038 to connect to remote hosts via SSH. This is particularly
9039 useful for systemd-run because it enables queuing of jobs
9040 onto remote systems.
9042 * machinectl gained a new command "login" to open a getty
9043 login in any local container. This works with any container
9044 that is registered with machined (such as those created by
9045 libvirt-lxc or nspawn), and which runs systemd inside.
9047 * machinectl gained a new "reboot" command that may be used to
9048 trigger a reboot on a specific container that is registered
9049 with machined. This works on any container that runs an init
9050 system of some kind.
9052 * systemctl gained a new "list-timers" command to print a nice
9053 listing of installed timer units with the times they elapse
9056 * Alternative reboot() parameters may now be specified on the
9057 "systemctl reboot" command line and are passed to the
9058 reboot() system call.
9060 * systemctl gained a new --job-mode= switch to configure the
9061 mode to queue a job with. This is a more generic version of
9062 --fail, --irreversible, and --ignore-dependencies, which are
9063 still available but not advertised anymore.
9065 * /etc/systemd/system.conf gained new settings to configure
9066 various default timeouts of units, as well as the default
9067 start limit interval and burst. These may still be overridden
9070 * PID1 will now export on the bus profile data of the security
9071 policy upload process (such as the SELinux policy upload to
9074 * journald: when forwarding logs to the console, include
9075 timestamps (following the setting in
9076 /sys/module/printk/parameters/time).
9078 * OnCalendar= in timer units now understands the special
9079 strings "yearly" and "annually". (Both are equivalent)
9081 * The accuracy of timer units is now configurable with the new
9082 AccuracySec= setting. It defaults to 1min.
9084 * A new dependency type JoinsNamespaceOf= has been added that
9085 allows running two services within the same /tmp and network
9086 namespace, if PrivateNetwork= or PrivateTmp= are used.
9088 * A new command "cat" has been added to systemctl. It outputs
9089 the original unit file of a unit, and concatenates the
9090 contents of additional "drop-in" unit file snippets, so that
9091 the full configuration is shown.
9093 * systemctl now supports globbing on the various "list-xyz"
9094 commands, like "list-units" or "list-sockets", as well as on
9095 those commands which take multiple unit names.
9097 * journalctl's --unit= switch gained support for globbing.
9099 * All systemd daemons now make use of the watchdog logic so
9100 that systemd automatically notices when they hang.
9102 * If the $container_ttys environment variable is set,
9103 getty-generator will automatically spawn a getty for each
9104 listed tty. This is useful for container managers to request
9105 login gettys to be spawned on as many ttys as needed.
9107 * %h, %s, %U specifier support is not available anymore when
9108 used in unit files for PID 1. This is because NSS calls are
9109 not safe from PID 1. They stay available for --user
9110 instances of systemd, and as special case for the root user.
9112 * loginctl gained a new "--no-legend" switch to turn off output
9115 * The "sd-login.h" API gained three new calls:
9116 sd_session_is_remote(), sd_session_get_remote_user(),
9117 sd_session_get_remote_host() to query information about
9120 * The udev hardware database now also carries vendor/product
9121 information of SDIO devices.
9123 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
9124 determine whether watchdog notifications are requested by
9127 * Socket-activated per-connection services now include a
9128 short description of the connection parameters in the
9131 * tmpfiles gained a new "--boot" option. When this is not used,
9132 only lines where the command character is not suffixed with
9133 "!" are executed. When this option is specified, those
9134 options are executed too. This partitions tmpfiles
9135 directives into those that can be safely executed at any
9136 time, and those which should be run only at boot (for
9137 example, a line that creates /run/nologin).
9139 * A new API "sd-resolve.h" has been added which provides a simple
9140 asynchronous wrapper around glibc NSS hostname resolution
9141 calls, such as getaddrinfo(). In contrast to glibc's
9142 getaddrinfo_a(), it does not use signals. In contrast to most
9143 other asynchronous name resolution libraries, this one does
9144 not reimplement DNS, but reuses NSS, so that alternate
9145 hostname resolution systems continue to work, such as mDNS,
9146 LDAP, etc. This API is based on libasyncns, but it has been
9147 cleaned up for inclusion in systemd.
9149 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
9150 "sd-daemon.h" are no longer found in individual libraries
9151 libsystemd-journal.so, libsystemd-login.so,
9152 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
9153 merged them into a single library, libsystemd.so, which
9154 provides all symbols. The reason for this is cyclic
9155 dependencies, as these libraries tend to use each other's
9156 symbols. So far, we have managed to workaround that by linking
9157 a copy of a good part of our code into each of these
9158 libraries again and again, which, however, makes certain
9159 things hard to do, like sharing static variables. Also, it
9160 substantially increases footprint. With this change, there
9161 is only one library for the basic APIs systemd
9162 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
9163 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
9164 library as well, however are subject to the --enable-kdbus
9165 switch (see below). Note that "sd-dhcp-client.h" is not part
9166 of this library (this is because it only consumes, never
9167 provides, services of/to other APIs). To make the transition
9168 easy from the separate libraries to the unified one, we
9169 provide the --enable-compat-libs compile-time switch which
9170 will generate stub libraries that are compatible with the
9171 old ones but redirect all calls to the new one.
9173 * All of the kdbus logic and the new APIs "sd-bus.h",
9174 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
9175 and "sd-utf8.h" are compile-time optional via the
9176 "--enable-kdbus" switch, and they are not compiled in by
9177 default. To make use of kdbus, you have to explicitly enable
9178 the switch. Note however, that neither the kernel nor the
9179 userspace API for all of this is considered stable yet. We
9180 want to maintain the freedom to still change the APIs for
9181 now. By specifying this build-time switch, you acknowledge
9182 that you are aware of the instability of the current
9185 * Also, note that while kdbus is pretty much complete,
9186 it lacks one thing: proper policy support. This means you
9187 can build a fully working system with all features; however,
9188 it will be highly insecure. Policy support will be added in
9189 one of the next releases, at the same time that we will
9190 declare the APIs stable.
9192 * When the kernel command line argument "kdbus" is specified,
9193 systemd will automatically load the kdbus.ko kernel module. At
9194 this stage of development, it is only useful for testing kdbus
9195 and should not be used in production. Note: if "--enable-kdbus"
9196 is specified, and the kdbus.ko kernel module is available, and
9197 "kdbus" is added to the kernel command line, the entire system
9198 runs with kdbus instead of dbus-daemon, with the above mentioned
9199 problem of missing the system policy enforcement. Also a future
9200 version of kdbus.ko or a newer systemd will not be compatible with
9201 each other, and will unlikely be able to boot the machine if only
9202 one of them is updated.
9204 * systemctl gained a new "import-environment" command which
9205 uploads the caller's environment (or parts thereof) into the
9206 service manager so that it is inherited by services started
9207 by the manager. This is useful to upload variables like
9208 $DISPLAY into the user service manager.
9210 * A new PrivateDevices= switch has been added to service units
9211 which allows running a service with a namespaced /dev
9212 directory that does not contain any device nodes for
9213 physical devices. More specifically, it only includes devices
9214 such as /dev/null, /dev/urandom, and /dev/zero which are API
9217 * logind has been extended to support behaviour like VT
9218 switching on seats that do not support a VT. This makes
9219 multi-session available on seats that are not the first seat
9220 (seat0), and on systems where kernel support for VTs has
9221 been disabled at compile-time.
9223 * If a process holds a delay lock for system sleep or shutdown
9224 and fails to release it in time, we will now log its
9225 identity. This makes it easier to identify processes that
9226 cause slow suspends or power-offs.
9228 * When parsing /etc/crypttab, support for a new key-slot=
9229 option as supported by Debian is added. It allows indicating
9230 which LUKS slot to use on disk, speeding up key loading.
9232 * The sd_journal_sendv() API call has been checked and
9233 officially declared to be async-signal-safe so that it may
9234 be invoked from signal handlers for logging purposes.
9236 * Boot-time status output is now enabled automatically after a
9237 short timeout if boot does not progress, in order to give
9238 the user an indication what she or he is waiting for.
9240 * The boot-time output has been improved to show how much time
9241 remains until jobs expire.
9243 * The KillMode= switch in service units gained a new possible
9244 value "mixed". If set, and the unit is shut down, then the
9245 initial SIGTERM signal is sent only to the main daemon
9246 process, while the following SIGKILL signal is sent to
9247 all remaining processes of the service.
9249 * When a scope unit is registered, a new property "Controller"
9250 may be set. If set to a valid bus name, systemd will send a
9251 RequestStop() signal to this name when it would like to shut
9252 down the scope. This may be used to hook manager logic into
9253 the shutdown logic of scope units. Also, scope units may now
9254 be put in a special "abandoned" state, in which case the
9255 manager process which created them takes no further
9256 responsibilities for it.
9258 * When reading unit files, systemd will now verify
9259 the access mode of these files, and warn about certain
9260 suspicious combinations. This has been added to make it
9261 easier to track down packaging bugs where unit files are
9262 marked executable or world-writable.
9264 * systemd-nspawn gained a new "--setenv=" switch to set
9265 container-wide environment variables. The similar option in
9266 systemd-activate was renamed from "--environment=" to
9267 "--setenv=" for consistency.
9269 * systemd-nspawn has been updated to create a new kdbus domain
9270 for each container that is invoked, thus allowing each
9271 container to have its own set of system and user buses,
9272 independent of the host.
9274 * systemd-nspawn gained a new --drop-capability= switch to run
9275 the container with less capabilities than the default. Both
9276 --drop-capability= and --capability= now take the special
9277 string "all" for dropping or keeping all capabilities.
9279 * systemd-nspawn gained new switches for executing containers
9280 with specific SELinux labels set.
9282 * systemd-nspawn gained a new --quiet switch to not generate
9283 any additional output but the container's own console
9286 * systemd-nspawn gained a new --share-system switch to run a
9287 container without PID namespacing enabled.
9289 * systemd-nspawn gained a new --register= switch to control
9290 whether the container is registered with systemd-machined or
9291 not. This is useful for containers that do not run full
9292 OS images, but only specific apps.
9294 * systemd-nspawn gained a new --keep-unit which may be used
9295 when invoked as the only program from a service unit, and
9296 results in registration of the unit service itself in
9297 systemd-machined, instead of a newly opened scope unit.
9299 * systemd-nspawn gained a new --network-interface= switch for
9300 moving arbitrary interfaces to the container. The new
9301 --network-veth switch creates a virtual Ethernet connection
9302 between host and container. The new --network-bridge=
9303 switch then allows assigning the host side of this virtual
9304 Ethernet connection to a bridge device.
9306 * systemd-nspawn gained a new --personality= switch for
9307 setting the kernel personality for the container. This is
9308 useful when running a 32-bit container on a 64-bit host. A
9309 similar option Personality= is now also available for service
9312 * logind will now also track a "Desktop" identifier for each
9313 session which encodes the desktop environment of it. This is
9314 useful for desktop environments that want to identify
9315 multiple running sessions of itself easily.
9317 * A new SELinuxContext= setting for service units has been
9318 added that allows setting a specific SELinux execution
9319 context for a service.
9321 * Most systemd client tools will now honour $SYSTEMD_LESS for
9322 settings of the "less" pager. By default, these tools will
9323 override $LESS to allow certain operations to work, such as
9324 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
9325 influence this logic.
9327 * systemd's "seccomp" hook-up has been changed to make use of
9328 the libseccomp library instead of using its own
9329 implementation. This has benefits for portability among
9332 * For usage together with SystemCallFilter=, a new
9333 SystemCallErrorNumber= setting has been introduced that
9334 allows configuration of a system error number to be returned
9335 on filtered system calls, instead of immediately killing the
9336 process. Also, SystemCallArchitectures= has been added to
9337 limit access to system calls of a particular architecture
9338 (in order to turn off support for unused secondary
9339 architectures). There is also a global
9340 SystemCallArchitectures= setting in system.conf now to turn
9341 off support for non-native system calls system-wide.
9343 * systemd requires a kernel with a working name_to_handle_at(),
9344 please see the kernel config requirements in the README file.
9346 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
9347 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
9348 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
9349 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
9350 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
9351 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
9352 Elia Pinto, Florian Weimer, George McCollister, Goffredo
9353 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
9354 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
9355 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
9356 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
9357 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
9358 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
9359 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
9360 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
9361 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
9362 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
9363 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
9364 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
9365 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
9366 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
9367 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
9368 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
9369 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
9371 — Berlin, 2014-02-20
9375 * logind has gained support for facilitating privileged input
9376 and drm device access for unprivileged clients. This work is
9377 useful to allow Wayland display servers (and similar
9378 programs, such as kmscon) to run under the user's ID and
9379 access input and drm devices which are normally
9380 protected. When this is used (and the kernel is new enough)
9381 logind will "mute" IO on the file descriptors passed to
9382 Wayland as long as it is in the background and "unmute" it
9383 if it returns into the foreground. This allows secure
9384 session switching without allowing background sessions to
9385 eavesdrop on input and display data. This also introduces
9386 session switching support if VT support is turned off in the
9387 kernel, and on seats that are not seat0.
9389 * A new kernel command line option luks.options= is understood
9390 now which allows specifying LUKS options for usage for LUKS
9391 encrypted partitions specified with luks.uuid=.
9393 * tmpfiles.d(5) snippets may now use specifier expansion in
9394 path names. More specifically %m, %b, %H, %v, are now
9395 replaced by the local machine id, boot id, hostname, and
9396 kernel version number.
9398 * A new tmpfiles.d(5) command "m" has been introduced which
9399 may be used to change the owner/group/access mode of a file
9400 or directory if it exists, but do nothing if it does not.
9402 * This release removes high-level support for the
9403 MemorySoftLimit= cgroup setting. The underlying kernel
9404 cgroup attribute memory.soft_limit= is currently badly
9405 designed and likely to be removed from the kernel API in its
9406 current form, hence we should not expose it for now.
9408 * The memory.use_hierarchy cgroup attribute is now enabled for
9409 all cgroups systemd creates in the memory cgroup
9410 hierarchy. This option is likely to be come the built-in
9411 default in the kernel anyway, and the non-hierarchical mode
9412 never made much sense in the intrinsically hierarchical
9415 * A new field _SYSTEMD_SLICE= is logged along with all journal
9416 messages containing the slice a message was generated
9417 from. This is useful to allow easy per-customer filtering of
9418 logs among other things.
9420 * systemd-journald will no longer adjust the group of journal
9421 files it creates to the "systemd-journal" group. Instead we
9422 rely on the journal directory to be owned by the
9423 "systemd-journal" group, and its setgid bit set, so that the
9424 kernel file system layer will automatically enforce that
9425 journal files inherit this group assignment. The reason for
9426 this change is that we cannot allow NSS look-ups from
9427 journald which would be necessary to resolve
9428 "systemd-journal" to a numeric GID, because this might
9429 create deadlocks if NSS involves synchronous queries to
9430 other daemons (such as nscd, or sssd) which in turn are
9431 logging clients of journald and might block on it, which
9432 would then dead lock. A tmpfiles.d(5) snippet included in
9433 systemd will make sure the setgid bit and group are
9434 properly set on the journal directory if it exists on every
9435 boot. However, we recommend adjusting it manually after
9436 upgrades too (or from RPM scriptlets), so that the change is
9437 not delayed until next reboot.
9439 * Backlight and random seed files in /var/lib/ have moved into
9440 the /var/lib/systemd/ directory, in order to centralize all
9441 systemd generated files in one directory.
9443 * Boot time performance measurements (as displayed by
9444 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
9445 performance information if that's available to determine how
9446 much time BIOS and boot loader initialization required. With
9447 a sufficiently new BIOS you hence no longer need to boot
9448 with Gummiboot to get access to such information.
9450 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
9451 Cristian Rodríguez, Dave Reisner, David Herrmann, David
9452 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
9453 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
9454 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
9455 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
9456 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
9458 — Berlin, 2013-10-02
9462 * The Restart= option for services now understands a new
9463 on-watchdog setting, which will restart the service
9464 automatically if the service stops sending out watchdog keep
9465 alive messages (as configured with WatchdogSec=).
9467 * The getty generator (which is responsible for bringing up a
9468 getty on configured serial consoles) will no longer only
9469 start a getty on the primary kernel console but on all
9470 others, too. This makes the order in which console= is
9471 specified on the kernel command line less important.
9473 * libsystemd-logind gained a new sd_session_get_vt() call to
9474 retrieve the VT number of a session.
9476 * If the option "tries=0" is set for an entry of /etc/crypttab
9477 its passphrase is queried indefinitely instead of any
9478 maximum number of tries.
9480 * If a service with a configure PID file terminates its PID
9481 file will now be removed automatically if it still exists
9482 afterwards. This should put an end to stale PID files.
9484 * systemd-run will now also take relative binary path names
9485 for execution and no longer insists on absolute paths.
9487 * InaccessibleDirectories= and ReadOnlyDirectories= now take
9488 paths that are optionally prefixed with "-" to indicate that
9489 it should not be considered a failure if they do not exist.
9491 * journalctl -o (and similar commands) now understands a new
9492 output mode "short-precise", it is similar to "short" but
9493 shows timestamps with usec accuracy.
9495 * The option "discard" (as known from Debian) is now
9496 synonymous to "allow-discards" in /etc/crypttab. In fact,
9497 "discard" is preferred now (since it is easier to remember
9500 * Some licensing clean-ups were made, so that more code is now
9501 LGPL-2.1 licensed than before.
9503 * A minimal tool to save/restore the display backlight
9504 brightness across reboots has been added. It will store the
9505 backlight setting as late as possible at shutdown, and
9506 restore it as early as possible during reboot.
9508 * A logic to automatically discover and enable home and swap
9509 partitions on GPT disks has been added. With this in place
9510 /etc/fstab becomes optional for many setups as systemd can
9511 discover certain partitions located on the root disk
9512 automatically. Home partitions are recognized under their
9513 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
9514 partitions are recognized under their GPT type ID
9515 0657fd6da4ab43c484e50933c84b4f4f.
9517 * systemd will no longer pass any environment from the kernel
9518 or initrd to system services. If you want to set an
9519 environment for all services, do so via the kernel command
9520 line systemd.setenv= assignment.
9522 * The systemd-sysctl tool no longer natively reads the file
9523 /etc/sysctl.conf. If desired, the file should be symlinked
9524 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
9525 legacy support by a symlink rather than built-in code, it
9526 also makes the otherwise hidden order of application of the
9527 different files visible. (Note that this partly reverts to a
9528 pre-198 application order of sysctl knobs!)
9530 * The "systemctl set-log-level" and "systemctl dump" commands
9531 have been moved to systemd-analyze.
9533 * systemd-run learned the new --remain-after-exit switch,
9534 which causes the scope unit not to be cleaned up
9535 automatically after the process terminated.
9537 * tmpfiles learned a new --exclude-prefix= switch to exclude
9538 certain paths from operation.
9540 * journald will now automatically flush all messages to disk
9541 as soon as a message at the log level CRIT, ALERT or EMERG
9544 Contributions from: Andrew Cook, Brandon Philips, Christian
9545 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
9546 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
9547 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
9548 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
9549 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
9550 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
9551 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
9552 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
9553 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
9554 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
9555 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
9556 William Giokas, Zbigniew Jędrzejewski-Szmek
9558 — Berlin, 2013-09-13
9562 * The documentation has been updated to cover the various new
9563 concepts introduced with 205.
9565 * Unit files now understand the new %v specifier which
9566 resolves to the kernel version string as returned by "uname
9569 * systemctl now supports filtering the unit list output by
9570 load state, active state and sub state, using the new
9573 * "systemctl status" will now show the results of the
9574 condition checks (like ConditionPathExists= and similar) of
9575 the last start attempts of the unit. They are also logged to
9578 * "journalctl -b" may now be used to look for boot output of a
9579 specific boot. Try "journalctl -b -1" for the previous boot,
9580 but the syntax is substantially more powerful.
9582 * "journalctl --show-cursor" has been added which prints the
9583 cursor string the last shown log line. This may then be used
9584 with the new "journalctl --after-cursor=" switch to continue
9585 browsing logs from that point on.
9587 * "journalctl --force" may now be used to force regeneration
9590 * Creation of "dead" device nodes has been moved from udev
9591 into kmod and tmpfiles. Previously, udev would read the kmod
9592 databases to pre-generate dead device nodes based on meta
9593 information contained in kernel modules, so that these would
9594 be auto-loaded on access rather then at boot. As this
9595 does not really have much to do with the exposing actual
9596 kernel devices to userspace this has always been slightly
9597 alien in the udev codebase. Following the new scheme kmod
9598 will now generate a runtime snippet for tmpfiles from the
9599 module meta information and it now is tmpfiles' job to the
9600 create the nodes. This also allows overriding access and
9601 other parameters for the nodes using the usual tmpfiles
9602 facilities. As side effect this allows us to remove the
9603 CAP_SYS_MKNOD capability bit from udevd entirely.
9605 * logind's device ACLs may now be applied to these "dead"
9606 devices nodes too, thus finally allowing managed access to
9607 devices such as /dev/snd/sequencer without loading the
9608 backing module right-away.
9610 * A new RPM macro has been added that may be used to apply
9611 tmpfiles configuration during package installation.
9613 * systemd-detect-virt and ConditionVirtualization= now can
9614 detect User-Mode-Linux machines (UML).
9616 * journald will now implicitly log the effective capabilities
9617 set of processes in the message metadata.
9619 * systemd-cryptsetup has gained support for TrueCrypt volumes.
9621 * The initrd interface has been simplified (more specifically,
9622 support for passing performance data via environment
9623 variables and fsck results via files in /run has been
9624 removed). These features were non-essential, and are
9625 nowadays available in a much nicer way by having systemd in
9626 the initrd serialize its state and have the hosts systemd
9627 deserialize it again.
9629 * The udev "keymap" data files and tools to apply keyboard
9630 specific mappings of scan to key codes, and force-release
9631 scan code lists have been entirely replaced by a udev
9632 "keyboard" builtin and a hwdb data file.
9634 * systemd will now honour the kernel's "quiet" command line
9635 argument also during late shutdown, resulting in a
9636 completely silent shutdown when used.
9638 * There's now an option to control the SO_REUSEPORT socket
9639 option in .socket units.
9641 * Instance units will now automatically get a per-template
9642 subslice of system.slice unless something else is explicitly
9643 configured. For example, instances of sshd@.service will now
9644 implicitly be placed in system-sshd.slice rather than
9645 system.slice as before.
9647 * Test coverage support may now be enabled at build time.
9649 Contributions from: Dave Reisner, Frederic Crozat, Harald
9650 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
9651 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
9652 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
9653 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
9654 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
9655 Giokas, Zbigniew Jędrzejewski-Szmek
9657 — Berlin, 2013-07-23
9661 * Two new unit types have been introduced:
9663 Scope units are very similar to service units, however, are
9664 created out of pre-existing processes — instead of PID 1
9665 forking off the processes. By using scope units it is
9666 possible for system services and applications to group their
9667 own child processes (worker processes) in a powerful way
9668 which then maybe used to organize them, or kill them
9669 together, or apply resource limits on them.
9671 Slice units may be used to partition system resources in an
9672 hierarchical fashion and then assign other units to them. By
9673 default there are now three slices: system.slice (for all
9674 system services), user.slice (for all user sessions),
9675 machine.slice (for VMs and containers).
9677 Slices and scopes have been introduced primarily in
9678 context of the work to move cgroup handling to a
9679 single-writer scheme, where only PID 1
9680 creates/removes/manages cgroups.
9682 * There's a new concept of "transient" units. In contrast to
9683 normal units these units are created via an API at runtime,
9684 not from configuration from disk. More specifically this
9685 means it is now possible to run arbitrary programs as
9686 independent services, with all execution parameters passed
9687 in via bus APIs rather than read from disk. Transient units
9688 make systemd substantially more dynamic then it ever was,
9689 and useful as a general batch manager.
9691 * logind has been updated to make use of scope and slice units
9692 for managing user sessions. As a user logs in he will get
9693 his own private slice unit, to which all sessions are added
9694 as scope units. We also added support for automatically
9695 adding an instance of user@.service for the user into the
9696 slice. Effectively logind will no longer create cgroup
9697 hierarchies on its own now, it will defer entirely to PID 1
9698 for this by means of scope, service and slice units. Since
9699 user sessions this way become entities managed by PID 1
9700 the output of "systemctl" is now a lot more comprehensive.
9702 * A new mini-daemon "systemd-machined" has been added which
9703 may be used by virtualization managers to register local
9704 VMs/containers. nspawn has been updated accordingly, and
9705 libvirt will be updated shortly. machined will collect a bit
9706 of meta information about the VMs/containers, and assign
9707 them their own scope unit (see above). The collected
9708 meta-data is then made available via the "machinectl" tool,
9709 and exposed in "ps" and similar tools. machined/machinectl
9710 is compile-time optional.
9712 * As discussed earlier, the low-level cgroup configuration
9713 options ControlGroup=, ControlGroupModify=,
9714 ControlGroupPersistent=, ControlGroupAttribute= have been
9715 removed. Please use high-level attribute settings instead as
9716 well as slice units.
9718 * A new bus call SetUnitProperties() has been added to alter
9719 various runtime parameters of a unit. This is primarily
9720 useful to alter cgroup parameters dynamically in a nice way,
9721 but will be extended later on to make more properties
9722 modifiable at runtime. systemctl gained a new set-properties
9723 command that wraps this call.
9725 * A new tool "systemd-run" has been added which can be used to
9726 run arbitrary command lines as transient services or scopes,
9727 while configuring a number of settings via the command
9728 line. This tool is currently very basic, however already
9729 very useful. We plan to extend this tool to even allow
9730 queuing of execution jobs with time triggers from the
9731 command line, similar in fashion to "at".
9733 * nspawn will now inform the user explicitly that kernels with
9734 audit enabled break containers, and suggest the user to turn
9737 * Support for detecting the IMA and AppArmor security
9738 frameworks with ConditionSecurity= has been added.
9740 * journalctl gained a new "-k" switch for showing only kernel
9741 messages, mimicking dmesg output; in addition to "--user"
9742 and "--system" switches for showing only user's own logs
9745 * systemd-delta can now show information about drop-in
9746 snippets extending unit files.
9748 * libsystemd-bus has been substantially updated but is still
9749 not available as public API.
9751 * systemd will now look for the "debug" argument on the kernel
9752 command line and enable debug logging, similar to what
9753 "systemd.log_level=debug" already did before.
9755 * "systemctl set-default", "systemctl get-default" has been
9756 added to configure the default.target symlink, which
9757 controls what to boot into by default.
9759 * "systemctl set-log-level" has been added as a convenient
9760 way to raise and lower systemd logging threshold.
9762 * "systemd-analyze plot" will now show the time the various
9763 generators needed for execution, as well as information
9764 about the unit file loading.
9766 * libsystemd-journal gained a new sd_journal_open_files() call
9767 for opening specific journal files. journactl also gained a
9768 new switch to expose this new functionality. Previously we
9769 only supported opening all files from a directory, or all
9770 files from the system, as opening individual files only is
9771 racy due to journal file rotation.
9773 * systemd gained the new DefaultEnvironment= setting in
9774 /etc/systemd/system.conf to set environment variables for
9777 * If a privileged process logs a journal message with the
9778 OBJECT_PID= field set, then journald will automatically
9779 augment this with additional OBJECT_UID=, OBJECT_GID=,
9780 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
9781 system services want to log events about specific client
9782 processes. journactl/systemctl has been updated to make use
9783 of this information if all log messages regarding a specific
9786 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
9787 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
9788 Reisner, David Coppa, David King, David Strauss, Eelco
9789 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
9790 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
9791 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
9792 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
9793 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
9794 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
9795 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
9796 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
9797 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
9798 Łukasz Stelmach, 장동준
9802 * The Python bindings gained some minimal support for the APIs
9803 exposed by libsystemd-logind.
9805 * ConditionSecurity= gained support for detecting SMACK. Since
9806 this condition already supports SELinux and AppArmor we only
9807 miss IMA for this. Patches welcome!
9809 Contributions from: Karol Lewandowski, Lennart Poettering,
9810 Zbigniew Jędrzejewski-Szmek
9814 * systemd-nspawn will now create /etc/resolv.conf if
9815 necessary, before bind-mounting the host's file onto it.
9817 * systemd-nspawn will now store meta information about a
9818 container on the container's cgroup as extended attribute
9819 fields, including the root directory.
9821 * The cgroup hierarchy has been reworked in many ways. All
9822 objects any of the components systemd creates in the cgroup
9823 tree are now suffixed. More specifically, user sessions are
9824 now placed in cgroups suffixed with ".session", users in
9825 cgroups suffixed with ".user", and nspawn containers in
9826 cgroups suffixed with ".nspawn". Furthermore, all cgroup
9827 names are now escaped in a simple scheme to avoid collision
9828 of userspace object names with kernel filenames. This work
9829 is preparation for making these objects relocatable in the
9830 cgroup tree, in order to allow easy resource partitioning of
9831 these objects without causing naming conflicts.
9833 * systemctl list-dependencies gained the new switches
9834 --plain, --reverse, --after and --before.
9836 * systemd-inhibit now shows the process name of processes that
9837 have taken an inhibitor lock.
9839 * nss-myhostname will now also resolve "localhost"
9840 implicitly. This makes /etc/hosts an optional file and
9841 nicely handles that on IPv6 ::1 maps to both "localhost" and
9844 * libsystemd-logind.so gained a new call
9845 sd_get_machine_names() to enumerate running containers and
9846 VMs (currently only supported by very new libvirt and
9847 nspawn). sd_login_monitor can now be used to watch
9848 VMs/containers coming and going.
9850 * .include is not allowed recursively anymore, and only in
9851 unit files. Usually it is better to use drop-in snippets in
9852 .d/*.conf anyway, as introduced with systemd 198.
9854 * systemd-analyze gained a new "critical-chain" command that
9855 determines the slowest chain of units run during system
9856 boot-up. It is very useful for tracking down where
9857 optimizing boot time is the most beneficial.
9859 * systemd will no longer allow manipulating service paths in
9860 the name=systemd:/system cgroup tree using ControlGroup= in
9861 units. (But is still fine with it in all other dirs.)
9863 * There's a new systemd-nspawn@.service service file that may
9864 be used to easily run nspawn containers as system
9865 services. With the container's root directory in
9866 /var/lib/container/foobar it is now sufficient to run
9867 "systemctl start systemd-nspawn@foobar.service" to boot it.
9869 * systemd-cgls gained a new parameter "--machine" to list only
9870 the processes within a certain container.
9872 * ConditionSecurity= now can check for "apparmor". We still
9873 are lacking checks for SMACK and IMA for this condition
9874 check though. Patches welcome!
9876 * A new configuration file /etc/systemd/sleep.conf has been
9877 added that may be used to configure which kernel operation
9878 systemd is supposed to execute when "suspend", "hibernate"
9879 or "hybrid-sleep" is requested. This makes the new kernel
9880 "freeze" state accessible to the user.
9882 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
9883 the passed argument if applicable.
9885 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
9886 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
9887 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
9888 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
9889 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
9890 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
9891 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
9896 * The output of 'systemctl list-jobs' got some polishing. The
9897 '--type=' argument may now be passed more than once. A new
9898 command 'systemctl list-sockets' has been added which shows
9899 a list of kernel sockets systemd is listening on with the
9900 socket units they belong to, plus the units these socket
9903 * The experimental libsystemd-bus library got substantial
9904 updates to work in conjunction with the (also experimental)
9905 kdbus kernel project. It works well enough to exchange
9906 messages with some sophistication. Note that kdbus is not
9907 ready yet, and the library is mostly an elaborate test case
9908 for now, and not installable.
9910 * systemd gained a new unit 'systemd-static-nodes.service'
9911 that generates static device nodes earlier during boot, and
9912 can run in conjunction with udev.
9914 * libsystemd-login gained a new call sd_pid_get_user_unit()
9915 to retrieve the user systemd unit a process is running
9916 in. This is useful for systems where systemd is used as
9919 * systemd-nspawn now places all containers in the new /machine
9920 top-level cgroup directory in the name=systemd
9921 hierarchy. libvirt will soon do the same, so that we get a
9922 uniform separation of /system, /user and /machine for system
9923 services, user processes and containers/virtual
9924 machines. This new cgroup hierarchy is also useful to stick
9925 stable names to specific container instances, which can be
9926 recognized later this way (this name may be controlled
9927 via systemd-nspawn's new -M switch). libsystemd-login also
9928 gained a new call sd_pid_get_machine_name() to retrieve the
9929 name of the container/VM a specific process belongs to.
9931 * bootchart can now store its data in the journal.
9933 * libsystemd-journal gained a new call
9934 sd_journal_add_conjunction() for AND expressions to the
9935 matching logic. This can be used to express more complex
9936 logical expressions.
9938 * journactl can now take multiple --unit= and --user-unit=
9941 * The cryptsetup logic now understands the "luks.key=" kernel
9942 command line switch for specifying a file to read the
9943 decryption key from. Also, if a configured key file is not
9944 found the tool will now automatically fall back to prompting
9947 * Python systemd.journal module was updated to wrap recently
9948 added functions from libsystemd-journal. The interface was
9949 changed to bring the low level interface in s.j._Reader
9950 closer to the C API, and the high level interface in
9951 s.j.Reader was updated to wrap and convert all data about
9954 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
9955 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
9956 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
9957 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
9958 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
9959 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
9963 * journalctl --update-catalog now understands a new --root=
9964 option to operate on catalogs found in a different root
9967 * During shutdown after systemd has terminated all running
9968 services a final killing loop kills all remaining left-over
9969 processes. We will now print the name of these processes
9970 when we send SIGKILL to them, since this usually indicates a
9973 * If /etc/crypttab refers to password files stored on
9974 configured mount points automatic dependencies will now be
9975 generated to ensure the specific mount is established first
9976 before the key file is attempted to be read.
9978 * 'systemctl status' will now show information about the
9979 network sockets a socket unit is listening on.
9981 * 'systemctl status' will also shown information about any
9982 drop-in configuration file for units. (Drop-In configuration
9983 files in this context are files such as
9984 /etc/systemd/systemd/foobar.service.d/*.conf)
9986 * systemd-cgtop now optionally shows summed up CPU times of
9987 cgroups. Press '%' while running cgtop to switch between
9988 percentage and absolute mode. This is useful to determine
9989 which cgroups use up the most CPU time over the entire
9990 runtime of the system. systemd-cgtop has also been updated
9991 to be 'pipeable' for processing with further shell tools.
9993 * 'hostnamectl set-hostname' will now allow setting of FQDN
9996 * The formatting and parsing of time span values has been
9997 changed. The parser now understands fractional expressions
9998 such as "5.5h". The formatter will now output fractional
9999 expressions for all time spans under 1min, i.e. "5.123456s"
10000 rather than "5s 123ms 456us". For time spans under 1s
10001 millisecond values are shown, for those under 1ms
10002 microsecond values are shown. This should greatly improve
10003 all time-related output of systemd.
10005 * libsystemd-login and libsystemd-journal gained new
10006 functions for querying the poll() events mask and poll()
10007 timeout value for integration into arbitrary event
10010 * localectl gained the ability to list available X11 keymaps
10011 (models, layouts, variants, options).
10013 * 'systemd-analyze dot' gained the ability to filter for
10014 specific units via shell-style globs, to create smaller,
10015 more useful graphs. I.e. it is now possible to create simple
10016 graphs of all the dependencies between only target units, or
10017 of all units that Avahi has dependencies with.
10019 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
10020 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
10021 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
10022 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
10023 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
10024 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
10025 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
10029 * The boot-time readahead implementation for rotating media
10030 will now read the read-ahead data in multiple passes which
10031 consist of all read requests made in equidistant time
10032 intervals. This means instead of strictly reading read-ahead
10033 data in its physical order on disk we now try to find a
10034 middle ground between physical and access time order.
10036 * /etc/os-release files gained a new BUILD_ID= field for usage
10037 on operating systems that provide continuous builds of OS
10040 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
10041 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
10042 William Douglas, Zbigniew Jędrzejewski-Szmek
10046 * systemd-python gained an API exposing libsystemd-daemon.
10048 * The SMACK setup logic gained support for uploading CIPSO
10051 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
10052 ReadOnlyDirectories= and InaccessibleDirectories= has
10053 changed. The private /tmp and /var/tmp directories are now
10054 shared by all processes of a service (which means
10055 ExecStartPre= may now leave data in /tmp that ExecStart= of
10056 the same service can still access). When a service is
10057 stopped its temporary directories are immediately deleted
10058 (normal clean-up with tmpfiles is still done in addition to
10061 * By default, systemd will now set a couple of sysctl
10062 variables in the kernel: the safe sysrq options are turned
10063 on, IP route verification is turned on, and source routing
10064 disabled. The recently added hardlink and softlink
10065 protection of the kernel is turned on. These settings should
10066 be reasonably safe, and good defaults for all new systems.
10068 * The predictable network naming logic may now be turned off
10069 with a new kernel command line switch: net.ifnames=0.
10071 * A new libsystemd-bus module has been added that implements a
10072 pretty complete D-Bus client library. For details see:
10074 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
10076 * journald will now explicitly flush the journal files to disk
10077 at the latest 5min after each write. The file will then also
10078 be marked offline until the next write. This should increase
10079 reliability in case of a crash. The synchronization delay
10080 can be configured via SyncIntervalSec= in journald.conf.
10082 * There's a new remote-fs-setup.target unit that can be used
10083 to pull in specific services when at least one remote file
10084 system is to be mounted.
10086 * There are new targets timers.target and paths.target as
10087 canonical targets to pull user timer and path units in
10088 from. This complements sockets.target with a similar
10089 purpose for socket units.
10091 * libudev gained a new call udev_device_set_attribute_value()
10092 to set sysfs attributes of a device.
10094 * The udev daemon now sets the default number of worker
10095 processes executed in parallel based on the number of available
10096 CPUs instead of the amount of available RAM. This is supposed
10097 to provide a more reliable default and limit a too aggressive
10098 parallelism for setups with 1000s of devices connected.
10100 Contributions from: Auke Kok, Colin Walters, Cristian
10101 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
10102 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
10103 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
10104 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
10105 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
10106 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
10107 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
10108 Zbigniew Jędrzejewski-Szmek
10112 * Configuration of unit files may now be extended via drop-in
10113 files without having to edit/override the unit files
10114 themselves. More specifically, if the administrator wants to
10115 change one value for a service file foobar.service he can
10116 now do so by dropping in a configuration snippet into
10117 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
10118 will load all these snippets and apply them on top of the
10119 main unit configuration file, possibly extending or
10120 overriding its settings. Using these drop-in snippets is
10121 generally nicer than the two earlier options for changing
10122 unit files locally: copying the files from
10123 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
10124 them there; or creating a new file in /etc/systemd/system/
10125 that incorporates the original one via ".include". Drop-in
10126 snippets into these .d/ directories can be placed in any
10127 directory systemd looks for units in, and the usual
10128 overriding semantics between /usr/lib, /etc and /run apply
10131 * Most unit file settings which take lists of items can now be
10132 reset by assigning the empty string to them. For example,
10133 normally, settings such as Environment=FOO=BAR append a new
10134 environment variable assignment to the environment block,
10135 each time they are used. By assigning Environment= the empty
10136 string the environment block can be reset to empty. This is
10137 particularly useful with the .d/*.conf drop-in snippets
10138 mentioned above, since this adds the ability to reset list
10139 settings from vendor unit files via these drop-ins.
10141 * systemctl gained a new "list-dependencies" command for
10142 listing the dependencies of a unit recursively.
10144 * Inhibitors are now honored and listed by "systemctl
10145 suspend", "systemctl poweroff" (and similar) too, not only
10146 GNOME. These commands will also list active sessions by
10149 * Resource limits (as exposed by the various control group
10150 controllers) can now be controlled dynamically at runtime
10151 for all units. More specifically, you can now use a command
10152 like "systemctl set-cgroup-attr foobar.service cpu.shares
10153 2000" to alter the CPU shares a specific service gets. These
10154 settings are stored persistently on disk, and thus allow the
10155 administrator to easily adjust the resource usage of
10156 services with a few simple commands. This dynamic resource
10157 management logic is also available to other programs via the
10158 bus. Almost any kernel cgroup attribute and controller is
10161 * systemd-vconsole-setup will now copy all font settings to
10162 all allocated VTs, where it previously applied them only to
10165 * libsystemd-login gained the new sd_session_get_tty() API
10168 * This release drops support for a few legacy or
10169 distribution-specific LSB facility names when parsing init
10170 scripts: $x-display-manager, $mail-transfer-agent,
10171 $mail-transport-agent, $mail-transfer-agent, $smtp,
10172 $null. Also, the mail-transfer-agent.target unit backing
10173 this has been removed. Distributions which want to retain
10174 compatibility with this should carry the burden for
10175 supporting this themselves and patch support for these back
10176 in, if they really need to. Also, the facilities $syslog and
10177 $local_fs are now ignored, since systemd does not support
10178 early-boot LSB init scripts anymore, and these facilities
10179 are implied anyway for normal services. syslog.target has
10182 * There are new bus calls on PID1's Manager object for
10183 cancelling jobs, and removing snapshot units. Previously,
10184 both calls were only available on the Job and Snapshot
10185 objects themselves.
10187 * systemd-journal-gatewayd gained SSL support.
10189 * The various "environment" files, such as /etc/locale.conf
10190 now support continuation lines with a backslash ("\") as
10191 last character in the line, similarly in style (but different)
10192 to how this is supported in shells.
10194 * For normal user processes the _SYSTEMD_USER_UNIT= field is
10195 now implicitly appended to every log entry logged. systemctl
10196 has been updated to filter by this field when operating on a
10197 user systemd instance.
10199 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
10200 CAP_AUDIT_CONTROL capabilities to the capabilities set for
10201 the container. This makes it easier to boot unmodified
10202 Fedora systems in a container, which however still requires
10203 audit=0 to be passed on the kernel command line. Auditing in
10204 kernel and userspace is unfortunately still too broken in
10205 context of containers, hence we recommend compiling it out
10206 of the kernel or using audit=0. Hopefully this will be fixed
10207 one day for good in the kernel.
10209 * nspawn gained the new --bind= and --bind-ro= parameters to
10210 bind mount specific directories from the host into the
10213 * nspawn will now mount its own devpts file system instance
10214 into the container, in order not to leak pty devices from
10215 the host into the container.
10217 * systemd will now read the firmware boot time performance
10218 information from the EFI variables, if the used boot loader
10219 supports this, and takes it into account for boot performance
10220 analysis via "systemd-analyze". This is currently supported
10221 only in conjunction with Gummiboot, but could be supported
10222 by other boot loaders too. For details see:
10224 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
10226 * A new generator has been added that automatically mounts the
10227 EFI System Partition (ESP) to /boot, if that directory
10228 exists, is empty, and no other file system has been
10229 configured to be mounted there.
10231 * logind will now send out PrepareForSleep(false) out
10232 unconditionally, after coming back from suspend. This may be
10233 used by applications as asynchronous notification for
10234 system resume events.
10236 * "systemctl unlock-sessions" has been added, that allows
10237 unlocking the screens of all user sessions at once, similar
10238 to how "systemctl lock-sessions" already locked all users
10239 sessions. This is backed by a new D-Bus call UnlockSessions().
10241 * "loginctl seat-status" will now show the master device of a
10242 seat. (i.e. the device of a seat that needs to be around for
10243 the seat to be considered available, usually the graphics
10246 * tmpfiles gained a new "X" line type, that allows
10247 configuration of files and directories (with wildcards) that
10248 shall be excluded from automatic cleanup ("aging").
10250 * udev default rules set the device node permissions now only
10251 at "add" events, and do not change them any longer with a
10252 later "change" event.
10254 * The log messages for lid events and power/sleep keypresses
10255 now carry a message ID.
10257 * We now have a substantially larger unit test suite, but this
10258 continues to be work in progress.
10260 * udevadm hwdb gained a new --root= parameter to change the
10261 root directory to operate relative to.
10263 * logind will now issue a background sync() request to the kernel
10264 early at shutdown, so that dirty buffers are flushed to disk early
10265 instead of at the last moment, in order to optimize shutdown
10268 * A new bootctl tool has been added that is an interface for
10269 certain boot loader operations. This is currently a preview
10270 and is likely to be extended into a small mechanism daemon
10271 like timedated, localed, hostnamed, and can be used by
10272 graphical UIs to enumerate available boot options, and
10273 request boot into firmware operations.
10275 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
10276 the rest of the package. It also has been updated to work
10277 correctly in initrds.
10279 * polkit previously has been runtime optional, and is now also
10280 compile time optional via a configure switch.
10282 * systemd-analyze has been reimplemented in C. Also "systemctl
10283 dot" has moved into systemd-analyze.
10285 * "systemctl status" with no further parameters will now print
10286 the status of all active or failed units.
10288 * Operations such as "systemctl start" can now be executed
10289 with a new mode "--irreversible" which may be used to queue
10290 operations that cannot accidentally be reversed by a later
10291 job queuing. This is by default used to make shutdown
10292 requests more robust.
10294 * The Python API of systemd now gained a new module for
10295 reading journal files.
10297 * A new tool kernel-install has been added that can install
10298 kernel images according to the Boot Loader Specification:
10300 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
10302 * Boot time console output has been improved to provide
10303 animated boot time output for hanging jobs.
10305 * A new tool systemd-activate has been added which can be used
10306 to test socket activation with, directly from the command
10307 line. This should make it much easier to test and debug
10308 socket activation in daemons.
10310 * journalctl gained a new "--reverse" (or -r) option to show
10311 journal output in reverse order (i.e. newest line first).
10313 * journalctl gained a new "--pager-end" (or -e) option to jump
10314 to immediately jump to the end of the journal in the
10315 pager. This is only supported in conjunction with "less".
10317 * journalctl gained a new "--user-unit=" option, that works
10318 similarly to "--unit=" but filters for user units rather than
10321 * A number of unit files to ease adoption of systemd in
10322 initrds has been added. This moves some minimal logic from
10323 the various initrd implementations into systemd proper.
10325 * The journal files are now owned by a new group
10326 "systemd-journal", which exists specifically to allow access
10327 to the journal, and nothing else. Previously, we used the
10328 "adm" group for that, which however possibly covers more
10329 than just journal/log file access. This new group is now
10330 already used by systemd-journal-gatewayd to ensure this
10331 daemon gets access to the journal files and as little else
10332 as possible. Note that "make install" will also set FS ACLs
10333 up for /var/log/journal to give "adm" and "wheel" read
10334 access to it, in addition to "systemd-journal" which owns
10335 the journal files. We recommend that packaging scripts also
10336 add read access to "adm" + "wheel" to /var/log/journal, and
10337 all existing/future journal files. To normal users and
10338 administrators little changes, however packagers need to
10339 ensure to create the "systemd-journal" system group at
10340 package installation time.
10342 * The systemd-journal-gatewayd now runs as unprivileged user
10343 systemd-journal-gateway:systemd-journal-gateway. Packaging
10344 scripts need to create these system user/group at
10347 * timedated now exposes a new boolean property CanNTP that
10348 indicates whether a local NTP service is available or not.
10350 * systemd-detect-virt will now also detect xen PVs
10352 * The pstore file system is now mounted by default, if it is
10355 * In addition to the SELinux and IMA policies we will now also
10356 load SMACK policies at early boot.
10358 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
10359 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
10360 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
10361 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
10362 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
10363 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
10364 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
10365 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
10366 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
10367 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
10368 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
10369 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
10370 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
10371 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
10375 * Timer units now support calendar time events in addition to
10376 monotonic time events. That means you can now trigger a unit
10377 based on a calendar time specification such as "Thu,Fri
10378 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
10379 or fifth day of any month of the year 2013, given that it is
10380 a thursday or friday. This brings timer event support
10381 considerably closer to cron's capabilities. For details on
10382 the supported calendar time specification language see
10385 * udev now supports a number of different naming policies for
10386 network interfaces for predictable names, and a combination
10387 of these policies is now the default. Please see this wiki
10388 document for details:
10390 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
10392 * Auke Kok's bootchart implementation has been added to the
10393 systemd tree. It is an optional component that can graph the
10394 boot in quite some detail. It is one of the best bootchart
10395 implementations around and minimal in its code and
10398 * nss-myhostname has been integrated into the systemd source
10399 tree. nss-myhostname guarantees that the local hostname
10400 always stays resolvable via NSS. It has been a weak
10401 requirement of systemd-hostnamed since a long time, and
10402 since its code is actually trivial we decided to just
10403 include it in systemd's source tree. It can be turned off
10404 with a configure switch.
10406 * The read-ahead logic is now capable of properly detecting
10407 whether a btrfs file system is on SSD or rotating media, in
10408 order to optimize the read-ahead scheme. Previously, it was
10409 only capable of detecting this on traditional file systems
10412 * In udev, additional device properties are now read from the
10413 IAB in addition to the OUI database. Also, Bluetooth company
10414 identities are attached to the devices as well.
10416 * In service files %U may be used as specifier that is
10417 replaced by the configured user name of the service.
10419 * nspawn may now be invoked without a controlling TTY. This
10420 makes it suitable for invocation as its own service. This
10421 may be used to set up a simple containerized server system
10422 using only core OS tools.
10424 * systemd and nspawn can now accept socket file descriptors
10425 when they are started for socket activation. This enables
10426 implementation of socket activated nspawn
10427 containers. i.e. think about autospawning an entire OS image
10428 when the first SSH or HTTP connection is received. We expect
10429 that similar functionality will also be added to libvirt-lxc
10432 * journalctl will now suppress ANSI color codes when
10433 presenting log data.
10435 * systemctl will no longer show control group information for
10436 a unit if the control group is empty anyway.
10438 * logind can now automatically suspend/hibernate/shutdown the
10441 * /etc/machine-info and hostnamed now also expose the chassis
10442 type of the system. This can be used to determine whether
10443 the local system is a laptop, desktop, handset or
10444 tablet. This information may either be configured by the
10445 user/vendor or is automatically determined from ACPI and DMI
10446 information if possible.
10448 * A number of polkit actions are now bound together with "imply"
10449 rules. This should simplify creating UIs because many actions
10450 will now authenticate similar ones as well.
10452 * Unit files learnt a new condition ConditionACPower= which
10453 may be used to conditionalize a unit depending on whether an
10454 AC power source is connected or not, of whether the system
10455 is running on battery power.
10457 * systemctl gained a new "is-failed" verb that may be used in
10458 shell scripts and suchlike to check whether a specific unit
10459 is in the "failed" state.
10461 * The EnvironmentFile= setting in unit files now supports file
10462 globbing, and can hence be used to easily read a number of
10463 environment files at once.
10465 * systemd will no longer detect and recognize specific
10466 distributions. All distribution-specific #ifdeffery has been
10467 removed, systemd is now fully generic and
10468 distribution-agnostic. Effectively, not too much is lost as
10469 a lot of the code is still accessible via explicit configure
10470 switches. However, support for some distribution specific
10471 legacy configuration file formats has been dropped. We
10472 recommend distributions to simply adopt the configuration
10473 files everybody else uses now and convert the old
10474 configuration from packaging scripts. Most distributions
10475 already did that. If that's not possible or desirable,
10476 distributions are welcome to forward port the specific
10477 pieces of code locally from the git history.
10479 * When logging a message about a unit systemd will now always
10480 log the unit name in the message meta data.
10482 * localectl will now also discover system locale data that is
10483 not stored in locale archives, but directly unpacked.
10485 * logind will no longer unconditionally use framebuffer
10486 devices as seat masters, i.e. as devices that are required
10487 to be existing before a seat is considered preset. Instead,
10488 it will now look for all devices that are tagged as
10489 "seat-master" in udev. By default, framebuffer devices will
10490 be marked as such, but depending on local systems, other
10491 devices might be marked as well. This may be used to
10492 integrate graphics cards using closed source drivers (such
10493 as NVidia ones) more nicely into logind. Note however, that
10494 we recommend using the open source NVidia drivers instead,
10495 and no udev rules for the closed-source drivers will be
10496 shipped from us upstream.
10498 Contributions from: Adam Williamson, Alessandro Crismani, Auke
10499 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
10500 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
10501 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
10502 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
10503 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
10504 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
10505 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
10506 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
10507 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
10508 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
10509 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
10514 * udev gained support for loading additional device properties
10515 from an indexed database that is keyed by vendor/product IDs
10516 and similar device identifiers. For the beginning this
10517 "hwdb" is populated with data from the well-known PCI and
10518 USB database, but also includes PNP, ACPI and OID data. In
10519 the longer run this indexed database shall grow into
10520 becoming the one central database for non-essential
10521 userspace device metadata. Previously, data from the PCI/USB
10522 database was only attached to select devices, since the
10523 lookup was a relatively expensive operation due to O(n) time
10524 complexity (with n being the number of entries in the
10525 database). Since this is now O(1), we decided to add in this
10526 data for all devices where this is available, by
10527 default. Note that the indexed database needs to be rebuilt
10528 when new data files are installed. To achieve this you need
10529 to update your packaging scripts to invoke "udevadm hwdb
10530 --update" after installation of hwdb data files. For
10531 RPM-based distributions we introduced the new
10532 %udev_hwdb_update macro for this purpose.
10534 * The Journal gained support for the "Message Catalog", an
10535 indexed database to link up additional information with
10536 journal entries. For further details please check:
10538 https://www.freedesktop.org/wiki/Software/systemd/catalog
10540 The indexed message catalog database also needs to be
10541 rebuilt after installation of message catalog files. Use
10542 "journalctl --update-catalog" for this. For RPM-based
10543 distributions we introduced the %journal_catalog_update
10544 macro for this purpose.
10546 * The Python Journal bindings gained support for the standard
10547 Python logging framework.
10549 * The Journal API gained new functions for checking whether
10550 the underlying file system of a journal file is capable of
10551 properly reporting file change notifications, or whether
10552 applications that want to reflect journal changes "live"
10553 need to recheck journal files continuously in appropriate
10556 * It is now possible to set the "age" field for tmpfiles
10557 entries to 0, indicating that files matching this entry
10558 shall always be removed when the directories are cleaned up.
10560 * coredumpctl gained a new "gdb" verb which invokes gdb
10561 right-away on the selected coredump.
10563 * There's now support for "hybrid sleep" on kernels that
10564 support this, in addition to "suspend" and "hibernate". Use
10565 "systemctl hybrid-sleep" to make use of this.
10567 * logind's HandleSuspendKey= setting (and related settings)
10568 now gained support for a new "lock" setting to simply
10569 request the screen lock on all local sessions, instead of
10570 actually executing a suspend or hibernation.
10572 * systemd will now mount the EFI variables file system by
10575 * Socket units now gained support for configuration of the
10576 SMACK security label.
10578 * timedatectl will now output the time of the last and next
10579 daylight saving change.
10581 * We dropped support for various legacy and distro-specific
10582 concepts, such as insserv, early-boot SysV services
10583 (i.e. those for non-standard runlevels such as 'b' or 'S')
10584 or ArchLinux /etc/rc.conf support. We recommend the
10585 distributions who still need support this to either continue
10586 to maintain the necessary patches downstream, or find a
10587 different solution. (Talk to us if you have questions!)
10589 * Various systemd components will now bypass polkit checks for
10590 root and otherwise handle properly if polkit is not found to
10591 be around. This should fix most issues for polkit-less
10592 systems. Quite frankly this should have been this way since
10593 day one. It is absolutely our intention to make systemd work
10594 fine on polkit-less systems, and we consider it a bug if
10595 something does not work as it should if polkit is not around.
10597 * For embedded systems it is now possible to build udev and
10598 systemd without blkid and/or kmod support.
10600 * "systemctl switch-root" is now capable of switching root
10601 more than once. I.e. in addition to transitions from the
10602 initrd to the host OS it is now possible to transition to
10603 further OS images from the host. This is useful to implement
10604 offline updating tools.
10606 * Various other additions have been made to the RPM macros
10607 shipped with systemd. Use %udev_rules_update() after
10608 installing new udev rules files. %_udevhwdbdir,
10609 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
10610 %_sysctldir are now available which resolve to the right
10611 directories for packages to place various data files in.
10613 * journalctl gained the new --full switch (in addition to
10614 --all, to disable ellipsation for long messages.
10616 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
10617 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
10618 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
10619 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
10620 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
10621 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
10622 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
10623 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
10624 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
10628 * journalctl gained new --since= and --until= switches to
10629 filter by time. It also now supports nice filtering for
10630 units via --unit=/-u.
10632 * Type=oneshot services may use ExecReload= and do the
10635 * The journal daemon now supports time-based rotation and
10636 vacuuming, in addition to the usual disk-space based
10639 * The journal will now index the available field values for
10640 each field name. This enables clients to show pretty drop
10641 downs of available match values when filtering. The bash
10642 completion of journalctl has been updated
10643 accordingly. journalctl gained a new switch -F to list all
10644 values a certain field takes in the journal database.
10646 * More service events are now written as structured messages
10647 to the journal, and made recognizable via message IDs.
10649 * The timedated, localed and hostnamed mini-services which
10650 previously only provided support for changing time, locale
10651 and hostname settings from graphical DEs such as GNOME now
10652 also have a minimal (but very useful) text-based client
10653 utility each. This is probably the nicest way to changing
10654 these settings from the command line now, especially since
10655 it lists available options and is fully integrated with bash
10658 * There's now a new tool "systemd-coredumpctl" to list and
10659 extract coredumps from the journal.
10661 * We now install a README each in /var/log/ and
10662 /etc/rc.d/init.d explaining where the system logs and init
10663 scripts went. This hopefully should help folks who go to
10664 that dirs and look into the otherwise now empty void and
10665 scratch their heads.
10667 * When user-services are invoked (by systemd --user) the
10668 $MANAGERPID env var is set to the PID of systemd.
10670 * SIGRTMIN+24 when sent to a --user instance will now result
10671 in immediate termination of systemd.
10673 * gatewayd received numerous feature additions such as a
10674 "follow" mode, for live syncing and filtering.
10676 * browse.html now allows filtering and showing detailed
10677 information on specific entries. Keyboard navigation and
10678 mouse screen support has been added.
10680 * gatewayd/journalctl now supports HTML5/JSON
10681 Server-Sent-Events as output.
10683 * The SysV init script compatibility logic will now
10684 heuristically determine whether a script supports the
10685 "reload" verb, and only then make this available as
10686 "systemctl reload".
10688 * "systemctl status --follow" has been removed, use "journalctl
10691 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
10692 have been removed since they are hardly useful to be
10695 * And I'd like to take the opportunity to specifically mention
10696 Zbigniew for his great contributions. Zbigniew, you rock!
10698 Contributions from: Andrew Eikum, Christian Hesse, Colin
10699 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
10700 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
10701 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
10702 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
10703 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
10704 Jędrzejewski-Szmek, Сковорода Никита Андреевич
10708 * If /etc/vconsole.conf is non-existent or empty we will no
10709 longer load any console font or key map at boot by
10710 default. Instead the kernel defaults will be left
10711 intact. This is definitely the right thing to do, as no
10712 configuration should mean no configuration, and hard-coding
10713 font names that are different on all archs is probably a bad
10714 idea. Also, the kernel default key map and font should be
10715 good enough for most cases anyway, and mostly identical to
10716 the userspace fonts/key maps we previously overloaded them
10717 with. If distributions want to continue to default to a
10718 non-kernel font or key map they should ship a default
10719 /etc/vconsole.conf with the appropriate contents.
10721 Contributions from: Colin Walters, Daniel J Walsh, Dave
10722 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
10723 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
10727 * journalctl gained a new --cursor= switch to show entries
10728 starting from the specified location in the journal.
10730 * We now enforce a size limit on journal entry fields exported
10731 with "-o json" in journalctl. Fields larger than 4K will be
10732 assigned null. This can be turned off with --all.
10734 * An (optional) journal gateway daemon is now available as
10735 "systemd-journal-gatewayd.service". This service provides
10736 access to the journal via HTTP and JSON. This functionality
10737 will be used to implement live log synchronization in both
10738 pull and push modes, but has various other users too, such
10739 as easy log access for debugging of embedded devices. Right
10740 now it is already useful to retrieve the journal via HTTP:
10742 # systemctl start systemd-journal-gatewayd.service
10743 # wget http://localhost:19531/entries
10745 This will download the journal contents in a
10746 /var/log/messages compatible format. The same as JSON:
10748 # curl -H"Accept: application/json" http://localhost:19531/entries
10750 This service is also accessible via a web browser where a
10751 single static HTML5 app is served that uses the JSON logic
10752 to enable the user to do some basic browsing of the
10753 journal. This will be extended later on. Here's an example
10754 screenshot of this app in its current state:
10756 http://0pointer.de/public/journal-gatewayd
10758 Contributions from: Kay Sievers, Lennart Poettering, Robert
10759 Milasan, Tom Gundersen
10763 * The bash completion logic is now available for journalctl
10766 * We do not mount the "cpuset" controller anymore together with
10767 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
10768 started if no parameters are assigned to it. "cpuset" hence
10769 broke code that assumed it could create "cpu" groups and
10772 * journalctl -f will now subscribe to terminal size changes,
10773 and line break accordingly.
10775 Contributions from: Dave Reisner, Kay Sievers, Lennart
10776 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
10780 * nspawn will now create a symlink /etc/localtime in the
10781 container environment, copying the host's timezone
10782 setting. Previously this has been done via a bind mount, but
10783 since symlinks cannot be bind mounted this has now been
10784 changed to create/update the appropriate symlink.
10786 * journalctl -n's line number argument is now optional, and
10787 will default to 10 if omitted.
10789 * journald will now log the maximum size the journal files may
10790 take up on disk. This is particularly useful if the default
10791 built-in logic of determining this parameter from the file
10792 system size is used. Use "systemctl status
10793 systemd-journald.service" to see this information.
10795 * The multi-seat X wrapper tool has been stripped down. As X
10796 is now capable of enumerating graphics devices via udev in a
10797 seat-aware way the wrapper is not strictly necessary
10798 anymore. A stripped down temporary stop-gap is still shipped
10799 until the upstream display managers have been updated to
10800 fully support the new X logic. Expect this wrapper to be
10801 removed entirely in one of the next releases.
10803 * HandleSleepKey= in logind.conf has been split up into
10804 HandleSuspendKey= and HandleHibernateKey=. The old setting
10805 is not available anymore. X11 and the kernel are
10806 distinguishing between these keys and we should too. This
10807 also means the inhibition lock for these keys has been split
10810 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
10811 Poettering, Lukas Nykryn, Václav Pavlín
10815 * Whenever a unit changes state we will now log this to the
10816 journal and show along the unit's own log output in
10817 "systemctl status".
10819 * ConditionPathIsMountPoint= can now properly detect bind
10820 mount points too. (Previously, a bind mount of one file
10821 system to another place in the same file system could not be
10822 detected as mount, since they shared struct stat's st_dev
10825 * We will now mount the cgroup controllers cpu, cpuacct,
10826 cpuset and the controllers net_cls, net_prio together by
10829 * nspawn containers will now have a virtualized boot
10830 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
10831 over with a randomized ID at container initialization). This
10832 has the effect of making "journalctl -b" do the right thing
10835 * The JSON output journal serialization has been updated not
10836 to generate "endless" list objects anymore, but rather one
10837 JSON object per line. This is more in line how most JSON
10838 parsers expect JSON objects. The new output mode
10839 "json-pretty" has been added to provide similar output, but
10840 neatly aligned for readability by humans.
10842 * We dropped all explicit sync() invocations in the shutdown
10843 code. The kernel does this implicitly anyway in the kernel
10844 reboot() syscall. halt(8)'s -n option is now a compatibility
10847 * We now support virtualized reboot() in containers, as
10848 supported by newer kernels. We will fall back to exit() if
10849 CAP_SYS_REBOOT is not available to the container. Also,
10850 nspawn makes use of this now and will actually reboot the
10851 container if the containerized OS asks for that.
10853 * journalctl will only show local log output by default
10854 now. Use --merge (-m) to show remote log output, too.
10856 * libsystemd-journal gained the new sd_journal_get_usage()
10857 call to determine the current disk usage of all journal
10858 files. This is exposed in the new "journalctl --disk-usage"
10861 * journald gained a new configuration setting SplitMode= in
10862 journald.conf which may be used to control how user journals
10863 are split off. See journald.conf(5) for details.
10865 * A new condition type ConditionFileNotEmpty= has been added.
10867 * tmpfiles' "w" lines now support file globbing, to write
10868 multiple files at once.
10870 * We added Python bindings for the journal submission
10871 APIs. More Python APIs for a number of selected APIs will
10872 likely follow. Note that we intend to add native bindings
10873 only for the Python language, as we consider it common
10874 enough to deserve bindings shipped within systemd. There are
10875 various projects outside of systemd that provide bindings
10876 for languages such as PHP or Lua.
10878 * Many conditions will now resolve specifiers such as %i. In
10879 addition, PathChanged= and related directives of .path units
10880 now support specifiers as well.
10882 * There's now a new RPM macro definition for the system preset
10885 * journald will now warn if it ca not forward a message to the
10886 syslog daemon because its socket is full.
10888 * timedated will no longer write or process /etc/timezone,
10889 except on Debian. As we do not support late mounted /usr
10890 anymore /etc/localtime always being a symlink is now safe,
10891 and hence the information in /etc/timezone is not necessary
10894 * logind will now always reserve one VT for a text getty (VT6
10895 by default). Previously if more than 6 X sessions where
10896 started they took up all the VTs with auto-spawned gettys,
10897 so that no text gettys were available anymore.
10899 * udev will now automatically inform the btrfs kernel logic
10900 about btrfs RAID components showing up. This should make
10901 simple hotplug based btrfs RAID assembly work.
10903 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
10904 (but not for its children which will stay at the kernel
10905 default). This should allow setups with a lot more listening
10908 * systemd will now always pass the configured timezone to the
10909 kernel at boot. timedated will do the same when the timezone
10912 * logind's inhibition logic has been updated. By default,
10913 logind will now handle the lid switch, the power and sleep
10914 keys all the time, even in graphical sessions. If DEs want
10915 to handle these events on their own they should take the new
10916 handle-power-key, handle-sleep-key and handle-lid-switch
10917 inhibitors during their runtime. A simple way to achieve
10918 that is to invoke the DE wrapped in an invocation of:
10920 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
10922 * Access to unit operations is now checked via SELinux taking
10923 the unit file label and client process label into account.
10925 * systemd will now notify the administrator in the journal
10926 when he over-mounts a non-empty directory.
10928 * There are new specifiers that are resolved in unit files,
10929 for the hostname (%H), the machine ID (%m) and the boot ID
10932 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
10933 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
10934 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
10935 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
10936 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
10937 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
10938 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
10942 * Support for reading structured kernel messages from
10943 /dev/kmsg has now been added and is enabled by default.
10945 * Support for reading kernel messages from /proc/kmsg has now
10946 been removed. If you want kernel messages in the journal
10947 make sure to run a recent kernel (>= 3.5) that supports
10948 reading structured messages from /dev/kmsg (see
10949 above). /proc/kmsg is now exclusive property of classic
10950 syslog daemons again.
10952 * The libudev API gained the new
10953 udev_device_new_from_device_id() call.
10955 * The logic for file system namespace (ReadOnlyDirectory=,
10956 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
10957 require pivot_root() anymore. This means fewer temporary
10958 directories are created below /tmp for this feature.
10960 * nspawn containers will now see and receive all submounts
10961 made on the host OS below the root file system of the
10964 * Forward Secure Sealing is now supported for Journal files,
10965 which provide cryptographical sealing of journal files so
10966 that attackers cannot alter log history anymore without this
10967 being detectable. Lennart will soon post a blog story about
10968 this explaining it in more detail.
10970 * There are two new service settings RestartPreventExitStatus=
10971 and SuccessExitStatus= which allow configuration of exit
10972 status (exit code or signal) which will be excepted from the
10973 restart logic, resp. consider successful.
10975 * journalctl gained the new --verify switch that can be used
10976 to check the integrity of the structure of journal files and
10977 (if Forward Secure Sealing is enabled) the contents of
10980 * nspawn containers will now be run with /dev/stdin, /dev/fd/
10981 and similar symlinks pre-created. This makes running shells
10982 as container init process a lot more fun.
10984 * The fstab support can now handle PARTUUID= and PARTLABEL=
10987 * A new ConditionHost= condition has been added to match
10988 against the hostname (with globs) and machine ID. This is
10989 useful for clusters where a single OS image is used to
10990 provision a large number of hosts which shall run slightly
10991 different sets of services.
10993 * Services which hit the restart limit will now be placed in a
10996 Contributions from: Bertram Poettering, Dave Reisner, Huang
10997 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
10998 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
11002 * When running in --user mode systemd will now become a
11003 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
11004 tree a lot more organized.
11006 * A new PartOf= unit dependency type has been introduced that
11007 may be used to group services in a natural way.
11009 * "systemctl enable" may now be used to enable instances of
11012 * journalctl now prints error log levels in red, and
11013 warning/notice log levels in bright white. It also supports
11014 filtering by log level now.
11016 * cgtop gained a new -n switch (similar to top), to configure
11017 the maximum number of iterations to run for. It also gained
11018 -b, to run in batch mode (accepting no input).
11020 * The suffix ".service" may now be omitted on most systemctl
11021 command lines involving service unit names.
11023 * There's a new bus call in logind to lock all sessions, as
11024 well as a loginctl verb for it "lock-sessions".
11026 * libsystemd-logind.so gained a new call sd_journal_perror()
11027 that works similar to libc perror() but logs to the journal
11028 and encodes structured information about the error number.
11030 * /etc/crypttab entries now understand the new keyfile-size=
11033 * shutdown(8) now can send a (configurable) wall message when
11034 a shutdown is cancelled.
11036 * The mount propagation mode for the root file system will now
11037 default to "shared", which is useful to make containers work
11038 nicely out-of-the-box so that they receive new mounts from
11039 the host. This can be undone locally by running "mount
11040 --make-rprivate /" if needed.
11042 * The prefdm.service file has been removed. Distributions
11043 should maintain this unit downstream if they intend to keep
11044 it around. However, we recommend writing normal unit files
11045 for display managers instead.
11047 * Since systemd is a crucial part of the OS we will now
11048 default to a number of compiler switches that improve
11049 security (hardening) such as read-only relocations, stack
11050 protection, and suchlike.
11052 * The TimeoutSec= setting for services is now split into
11053 TimeoutStartSec= and TimeoutStopSec= to allow configuration
11054 of individual time outs for the start and the stop phase of
11057 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
11058 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
11059 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
11060 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
11061 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
11062 Gundersen, Zbigniew Jędrzejewski-Szmek
11066 * The journal and id128 C APIs are now fully documented as man
11069 * Extra safety checks have been added when transitioning from
11070 the initial RAM disk to the main system to avoid accidental
11073 * /etc/crypttab entries now understand the new keyfile-offset=
11076 * systemctl -t can now be used to filter by unit load state.
11078 * The journal C API gained the new sd_journal_wait() call to
11079 make writing synchronous journal clients easier.
11081 * journalctl gained the new -D switch to show journals from a
11082 specific directory.
11084 * journalctl now displays a special marker between log
11085 messages of two different boots.
11087 * The journal is now explicitly flushed to /var via a service
11088 systemd-journal-flush.service, rather than implicitly simply
11089 by seeing /var/log/journal to be writable.
11091 * journalctl (and the journal C APIs) can now match for much
11092 more complex expressions, with alternatives and
11095 * When transitioning from the initial RAM disk to the main
11096 system we will now kill all processes in a killing spree to
11097 ensure no processes stay around by accident.
11099 * Three new specifiers may be used in unit files: %u, %h, %s
11100 resolve to the user name, user home directory resp. user
11101 shell. This is useful for running systemd user instances.
11103 * We now automatically rotate journal files if their data
11104 object hash table gets a fill level > 75%. We also size the
11105 hash table based on the configured maximum file size. This
11106 together should lower hash collisions drastically and thus
11107 speed things up a bit.
11109 * journalctl gained the new "--header" switch to introspect
11110 header data of journal files.
11112 * A new setting SystemCallFilters= has been added to services which may
11113 be used to apply deny lists or allow lists to system calls. This is
11114 based on SECCOMP Mode 2 of Linux 3.5.
11116 * nspawn gained a new --link-journal= switch (and quicker: -j)
11117 to link the container journal with the host. This makes it
11118 very easy to centralize log viewing on the host for all
11119 guests while still keeping the journal files separated.
11121 * Many bugfixes and optimizations
11123 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
11124 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
11125 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
11130 * Several tools now understand kernel command line arguments,
11131 which are only read when run in an initial RAM disk. They
11132 usually follow closely their normal counterparts, but are
11135 * There's a new tool to analyze the readahead files that are
11136 automatically generated at boot. Use:
11138 /usr/lib/systemd/systemd-readahead analyze /.readahead
11140 * We now provide an early debug shell on tty9 if this enabled. Use:
11142 systemctl enable debug-shell.service
11144 * All plymouth related units have been moved into the Plymouth
11145 package. Please make sure to upgrade your Plymouth version
11148 * systemd-tmpfiles now supports getting passed the basename of
11149 a configuration file only, in which case it will look for it
11150 in all appropriate directories automatically.
11152 * udevadm info now takes a /dev or /sys path as argument, and
11153 does the right thing. Example:
11155 udevadm info /dev/sda
11156 udevadm info /sys/class/block/sda
11158 * systemctl now prints a warning if a unit is stopped but a
11159 unit that might trigger it continues to run. Example: a
11160 service is stopped but the socket that activates it is left
11163 * "systemctl status" will now mention if the log output was
11164 shortened due to rotation since a service has been started.
11166 * The journal API now exposes functions to determine the
11167 "cutoff" times due to rotation.
11169 * journald now understands SIGUSR1 and SIGUSR2 for triggering
11170 immediately flushing of runtime logs to /var if possible,
11171 resp. for triggering immediate rotation of the journal
11174 * It is now considered an error if a service is attempted to
11175 be stopped that is not loaded.
11177 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
11179 * systemd-analyze now supports Python 3
11181 * tmpfiles now supports cleaning up directories via aging
11182 where the first level dirs are always kept around but
11183 directories beneath it automatically aged. This is enabled
11184 by prefixing the age field with '~'.
11186 * Seat objects now expose CanGraphical, CanTTY properties
11187 which is required to deal with very fast bootups where the
11188 display manager might be running before the graphics drivers
11189 completed initialization.
11191 * Seat objects now expose a State property.
11193 * We now include RPM macros for service enabling/disabling
11194 based on the preset logic. We recommend RPM based
11195 distributions to make use of these macros if possible. This
11196 makes it simpler to reuse RPM spec files across
11199 * We now make sure that the collected systemd unit name is
11200 always valid when services log to the journal via
11203 * There's a new man page kernel-command-line(7) detailing all
11204 command line options we understand.
11206 * The fstab generator may now be disabled at boot by passing
11207 fstab=0 on the kernel command line.
11209 * A new kernel command line option modules-load= is now understood
11210 to load a specific kernel module statically, early at boot.
11212 * Unit names specified on the systemctl command line are now
11213 automatically escaped as needed. Also, if file system or
11214 device paths are specified they are automatically turned
11215 into the appropriate mount or device unit names. Example:
11217 systemctl status /home
11218 systemctl status /dev/sda
11220 * The SysVConsole= configuration option has been removed from
11221 system.conf parsing.
11223 * The SysV search path is no longer exported on the D-Bus
11226 * The Names= option has been removed from unit file parsing.
11228 * There's a new man page bootup(7) detailing the boot process.
11230 * Every unit and every generator we ship with systemd now
11231 comes with full documentation. The self-explanatory boot is
11234 * A couple of services gained "systemd-" prefixes in their
11235 name if they wrap systemd code, rather than only external
11236 code. Among them fsck@.service which is now
11237 systemd-fsck@.service.
11239 * The HaveWatchdog property has been removed from the D-Bus
11242 * systemd.confirm_spawn= on the kernel command line should now
11245 * There's a new man page crypttab(5) which details all options
11246 we actually understand.
11248 * systemd-nspawn gained a new --capability= switch to pass
11249 additional capabilities to the container.
11251 * timedated will now read known NTP implementation unit names
11252 from /usr/lib/systemd/ntp-units.d/*.list,
11253 systemd-timedated-ntp.target has been removed.
11255 * journalctl gained a new switch "-b" that lists log data of
11256 the current boot only.
11258 * The notify socket is in the abstract namespace again, in
11259 order to support daemons which chroot() at start-up.
11261 * There is a new Storage= configuration option for journald
11262 which allows configuration of where log data should go. This
11263 also provides a way to disable journal logging entirely, so
11264 that data collected is only forwarded to the console, the
11265 kernel log buffer or another syslog implementation.
11267 * Many bugfixes and optimizations
11269 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
11270 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
11271 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
11272 Shawn Landden, Tom Gundersen
11276 * "systemctl help <unit>" now shows the man page if one is
11279 * Several new man pages have been added.
11281 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
11282 MaxLevelConsole= can now be specified in
11283 journald.conf. These options allow reducing the amount of
11284 data stored on disk or forwarded by the log level.
11286 * TimerSlackNSec= can now be specified in system.conf for
11287 PID1. This allows system-wide power savings.
11289 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
11290 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
11295 * logind is now capable of (optionally) handling power and
11296 sleep keys as well as the lid switch.
11298 * journalctl now understands the syntax "journalctl
11299 /usr/bin/avahi-daemon" to get all log output of a specific
11302 * CapabilityBoundingSet= in system.conf now also influences
11303 the capability bound set of usermode helpers of the kernel.
11305 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
11306 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
11307 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
11308 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
11312 * Note that we skipped 139 releases here in order to set the
11313 new version to something that is greater than both udev's
11314 and systemd's most recent version number.
11316 * udev: all udev sources are merged into the systemd source tree now.
11317 All future udev development will happen in the systemd tree. It
11318 is still fully supported to use the udev daemon and tools without
11319 systemd running, like in initramfs or other init systems. Building
11320 udev though, will require the *build* of the systemd tree, but
11321 udev can be properly *run* without systemd.
11323 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
11324 should be used to create dead device nodes as workarounds for broken
11327 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
11328 no longer supported. udev_monitor_new_from_netlink() needs to be
11329 used to subscribe to events.
11331 * udev: when udevd is started by systemd, processes which are left
11332 behind by forking them off of udev rules, are unconditionally cleaned
11333 up and killed now after the event handling has finished. Services or
11334 daemons must be started as systemd services. Services can be
11335 pulled-in by udev to get started, but they can no longer be directly
11336 forked by udev rules.
11338 * udev: the daemon binary is called systemd-udevd now and installed
11339 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
11340 to adapt to that, create symlink, or rename the binary after building
11343 * libudev no longer provides these symbols:
11344 udev_monitor_from_socket()
11345 udev_queue_get_failed_list_entry()
11346 udev_get_{dev,sys,run}_path()
11347 The versions number was bumped and symbol versioning introduced.
11349 * systemd-loginctl and systemd-journalctl have been renamed
11350 to loginctl and journalctl to match systemctl.
11352 * The config files: /etc/systemd/systemd-logind.conf and
11353 /etc/systemd/systemd-journald.conf have been renamed to
11354 logind.conf and journald.conf. Package updates should rename
11355 the files to the new names on upgrade.
11357 * For almost all files the license is now LGPL2.1+, changed
11358 from the previous GPL2.0+. Exceptions are some minor stuff
11359 of udev (which will be changed to LGPL2.1 eventually, too),
11360 and the MIT licensed sd-daemon.[ch] library that is suitable
11361 to be used as drop-in files.
11363 * systemd and logind now handle system sleep states, in
11364 particular suspending and hibernating.
11366 * logind now implements a sleep/shutdown/idle inhibiting logic
11367 suitable for a variety of uses. Soonishly Lennart will blog
11368 about this in more detail.
11370 * var-run.mount and var-lock.mount are no longer provided
11371 (which previously bind mounted these directories to their new
11372 places). Distributions which have not converted these
11373 directories to symlinks should consider stealing these files
11374 from git history and add them downstream.
11376 * We introduced the Documentation= field for units and added
11377 this to all our shipped units. This is useful to make it
11378 easier to explore the boot and the purpose of the various
11381 * All smaller setup units (such as
11382 systemd-vconsole-setup.service) now detect properly if they
11383 are run in a container and are skipped when
11384 appropriate. This guarantees an entirely noise-free boot in
11385 Linux container environments such as systemd-nspawn.
11387 * A framework for implementing offline system updates is now
11388 integrated, for details see:
11389 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
11391 * A new service type Type=idle is available now which helps us
11392 avoiding ugly interleaving of getty output and boot status
11395 * There's now a system-wide CapabilityBoundingSet= option to
11396 globally reduce the set of capabilities for the
11397 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
11398 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
11399 even CAP_NET_ADMIN system-wide for secure systems.
11401 * There are now system-wide DefaultLimitXXX= options to
11402 globally change the defaults of the various resource limits
11403 for all units started by PID 1.
11405 * Harald Hoyer's systemd test suite has been integrated into
11406 systemd which allows easy testing of systemd builds in qemu
11407 and nspawn. (This is really awesome! Ask us for details!)
11409 * The fstab parser is now implemented as generator, not inside
11412 * systemctl will now warn you if .mount units generated from
11413 /etc/fstab are out of date due to changes in fstab that
11414 have not been read by systemd yet.
11416 * systemd is now suitable for usage in initrds. Dracut has
11417 already been updated to make use of this. With this in place
11418 initrds get a slight bit faster but primarily are much
11419 easier to introspect and debug since "systemctl status" in
11420 the host system can be used to introspect initrd services,
11421 and the journal from the initrd is kept around too.
11423 * systemd-delta has been added, a tool to explore differences
11424 between user/admin configuration and vendor defaults.
11426 * PrivateTmp= now affects both /tmp and /var/tmp.
11428 * Boot time status messages are now much prettier and feature
11429 proper english language. Booting up systemd has never been
11432 * Read-ahead pack files now include the inode number of all
11433 files to pre-cache. When the inode changes the pre-caching
11434 is not attempted. This should be nicer to deal with updated
11435 packages which might result in changes of read-ahead
11438 * We now temporaritly lower the kernel's read_ahead_kb variable
11439 when collecting read-ahead data to ensure the kernel's
11440 built-in read-ahead does not add noise to our measurements
11441 of necessary blocks to pre-cache.
11443 * There's now RequiresMountsFor= to add automatic dependencies
11444 for all mounts necessary for a specific file system path.
11446 * MountAuto= and SwapAuto= have been removed from
11447 system.conf. Mounting file systems at boot has to take place
11450 * nspawn now learned a new switch --uuid= to set the machine
11451 ID on the command line.
11453 * nspawn now learned the -b switch to automatically search
11454 for an init system.
11456 * vt102 is now the default TERM for serial TTYs, upgraded from
11459 * systemd-logind now works on VT-less systems.
11461 * The build tree has been reorganized. The individual
11462 components now have directories of their own.
11464 * A new condition type ConditionPathIsReadWrite= is now available.
11466 * nspawn learned the new -C switch to create cgroups for the
11467 container in other hierarchies.
11469 * We now have support for hardware watchdogs, configurable in
11472 * The scheduled shutdown logic now has a public API.
11474 * We now mount /tmp as tmpfs by default, but this can be
11475 masked and /etc/fstab can override it.
11477 * Since udisks does not make use of /media anymore we are not
11478 mounting a tmpfs on it anymore.
11480 * journalctl gained a new --local switch to only interleave
11481 locally generated journal files.
11483 * We can now load the IMA policy at boot automatically.
11485 * The GTK tools have been split off into a systemd-ui.
11487 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
11488 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
11489 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
11490 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
11491 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
11492 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
11493 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
11494 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
11495 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
11500 * This is mostly a bugfix release
11502 * Support optional initialization of the machine ID from the
11503 KVM or container configured UUID.
11505 * Support immediate reboots with "systemctl reboot -ff"
11507 * Show /etc/os-release data in systemd-analyze output
11509 * Many bugfixes for the journal, including endianness fixes and
11510 ensuring that disk space enforcement works
11512 * sd-login.h is C++ compatible again
11514 * Extend the /etc/os-release format on request of the Debian
11517 * We now refuse non-UTF8 strings used in various configuration
11518 and unit files. This is done to ensure we do not pass invalid
11519 data over D-Bus or expose it elsewhere.
11521 * Register Mimo USB Screens as suitable for automatic seat
11524 * Read SELinux client context from journal clients in a race
11527 * Reorder configuration file lookup order. /etc now always
11528 overrides /run in order to allow the administrator to always
11529 and unconditionally override vendor-supplied or
11530 automatically generated data.
11532 * The various user visible bits of the journal now have man
11533 pages. We still lack man pages for the journal API calls
11536 * We now ship all man pages in HTML format again in the
11539 Contributions from: Dave Reisner, Dirk Eibach, Frederic
11540 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
11541 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
11546 * This is mostly a bugfix release
11548 * systems lacking /etc/os-release are no longer supported.
11550 * Various functionality updates to libsystemd-login.so
11552 * Track class of PAM logins to distinguish greeters from
11553 normal user logins.
11555 Contributions from: Kay Sievers, Lennart Poettering, Michael
11560 * This is an important bugfix release for v41.
11562 * Building man pages is now optional which should be useful
11563 for those building systemd from git but unwilling to install
11566 * Watchdog support for supervising services is now usable. In
11567 a future release support for hardware watchdogs
11568 (i.e. /dev/watchdog) will be added building on this.
11570 * Service start rate limiting is now configurable and can be
11571 turned off per service. When a start rate limit is hit a
11572 reboot can automatically be triggered.
11574 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
11576 Contributions from: Benjamin Franzke, Bill Nottingham,
11577 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
11578 Schmidt, Michał Górny, Piotr Drąg
11582 * The systemd binary is installed /usr/lib/systemd/systemd now;
11583 An existing /sbin/init symlink needs to be adapted with the
11586 * The code that loads kernel modules has been ported to invoke
11587 libkmod directly, instead of modprobe. This means we do not
11588 support systems with module-init-tools anymore.
11590 * Watchdog support is now already useful, but still not
11593 * A new kernel command line option systemd.setenv= is
11594 understood to set system wide environment variables
11595 dynamically at boot.
11597 * We now limit the set of capabilities of systemd-journald.
11599 * We now set SIGPIPE to ignore by default, since it only is
11600 useful in shell pipelines, and has little use in general
11601 code. This can be disabled with IgnoreSIPIPE=no in unit
11604 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
11605 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
11610 * This is mostly a bugfix release
11612 * We now expose the reason why a service failed in the
11613 "Result" D-Bus property.
11615 * Rudimentary service watchdog support (will be completed over
11616 the next few releases.)
11618 * When systemd forks off in order execute some service we will
11619 now immediately changes its argv[0] to reflect which process
11620 it will execute. This is useful to minimize the time window
11621 with a generic argv[0], which makes bootcharts more useful
11623 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
11624 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
11625 Mike Kazantsev, Ray Strode
11629 * This is mostly a test release, but incorporates many
11632 * New systemd-cgtop tool to show control groups by their
11635 * Linking against libacl for ACLs is optional again. If
11636 disabled, support tracking device access for active logins
11637 goes becomes unavailable, and so does access to the user
11638 journals by the respective users.
11640 * If a group "adm" exists, journal files are automatically
11641 owned by them, thus allow members of this group full access
11642 to the system journal as well as all user journals.
11644 * The journal now stores the SELinux context of the logging
11645 client for all entries.
11647 * Add C++ inclusion guards to all public headers
11649 * New output mode "cat" in the journal to print only text
11650 messages, without any meta data like date or time.
11652 * Include tiny X server wrapper as a temporary stop-gap to
11653 teach XOrg udev display enumeration. This is used by display
11654 managers such as gdm, and will go away as soon as XOrg
11655 learned native udev hotplugging for display devices.
11657 * Add new systemd-cat tool for executing arbitrary programs
11658 with STDERR/STDOUT connected to the journal. Can also act as
11659 BSD logger replacement, and does so by default.
11661 * Optionally store all locally generated coredumps in the
11662 journal along with meta data.
11664 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
11665 writing short strings to files (for usage for /sys), and for
11666 creating symlinks, character and block device nodes.
11668 * New unit file option ControlGroupPersistent= to make cgroups
11669 persistent, following the mechanisms outlined in
11670 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
11672 * Support multiple local RTCs in a sane way
11674 * No longer monopolize IO when replaying readahead data on
11675 rotating disks, since we might starve non-file-system IO to
11676 death, since fanotify() will not see accesses done by blkid,
11679 * Do not show kernel threads in systemd-cgls anymore, unless
11680 requested with new -k switch.
11682 Contributions from: Dan Horák, Kay Sievers, Lennart
11683 Poettering, Michal Schmidt
11687 * This is mostly a test release, but incorporates many
11690 * The git repository moved to:
11691 git://anongit.freedesktop.org/systemd/systemd
11692 ssh://git.freedesktop.org/git/systemd/systemd
11694 * First release with the journal
11695 http://0pointer.de/blog/projects/the-journal.html
11697 * The journal replaces both systemd-kmsg-syslogd and
11698 systemd-stdout-bridge.
11700 * New sd_pid_get_unit() API call in libsystemd-logind
11702 * Many systemadm clean-ups
11704 * Introduce remote-fs-pre.target which is ordered before all
11705 remote mounts and may be used to start services before all
11708 * Added Mageia support
11710 * Add bash completion for systemd-loginctl
11712 * Actively monitor PID file creation for daemons which exit in
11713 the parent process before having finished writing the PID
11714 file in the daemon process. Daemons which do this need to be
11715 fixed (i.e. PID file creation must have finished before the
11716 parent exits), but we now react a bit more gracefully to them.
11718 * Add colourful boot output, mimicking the well-known output
11719 of existing distributions.
11721 * New option PassCredentials= for socket units, for
11722 compatibility with a recent kernel ABI breakage.
11724 * /etc/rc.local is now hooked in via a generator binary, and
11725 thus will no longer act as synchronization point during
11728 * systemctl list-unit-files now supports --root=.
11730 * systemd-tmpfiles now understands two new commands: z, Z for
11731 relabelling files according to the SELinux database. This is
11732 useful to apply SELinux labels to specific files in /sys,
11733 among other things.
11735 * Output of SysV services is now forwarded to both the console
11736 and the journal by default, not only just the console.
11738 * New man pages for all APIs from libsystemd-login.
11740 * The build tree got reorganized and the build system is a
11741 lot more modular allowing embedded setups to specifically
11742 select the components of systemd they are interested in.
11744 * Support for Linux systems lacking the kernel VT subsystem is
11747 * configure's --with-rootdir= got renamed to
11748 --with-rootprefix= to follow the naming used by udev and
11751 * Unless specified otherwise we will now install to /usr instead
11752 of /usr/local by default.
11754 * Processes with '@' in argv[0][0] are now excluded from the
11755 final shut-down killing spree, following the logic explained
11757 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
11759 * All processes remaining in a service cgroup when we enter
11760 the START or START_PRE states are now killed with
11761 SIGKILL. That means it is no longer possible to spawn
11762 background processes from ExecStart= lines (which was never
11763 supported anyway, and bad style).
11765 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
11766 reloading of units together.
11768 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
11769 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
11770 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
11771 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
11772 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek