]> git.ipfire.org Git - thirdparty/systemd.git/blame - NEWS
Merge pull request #9153 from poettering/private-mounts
[thirdparty/systemd.git] / NEWS
CommitLineData
d657c51f 1systemd System and Service Manager
220a21d3 2
019cb3ab
SH
3CHANGES WITH 239 in spe:
4
5 * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
6 builtin may name network interfaces differently than in previous
5cadf58e
ZJS
7 versions. SR-IOV virtual functions and NPAR partitions with PCI
8 function numbers of 8 and above will be named more predictably, and
9 udev may generate names based on the PCI slot number in some cases
019cb3ab
SH
10 where it previously did not.
11
6e2d744b
YW
12 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
13 systemd-logind.service. Since v235, IPAddressDeny=any has been set to
14 the unit. So, it is expected that the default behavior of systemd-logind
15 is not changed. However, if distribution packagers or administrators
16 disabled or modified IPAddressDeny= setting by a drop-in config file,
17 then it may be necessary to update the file to re-enable AF_INET and
18 AF_INET6 to support network user name services, e.g. NIS.
19
e0eee477
YW
20 * When the RestrictNamespaces= unit property is specified multiple times,
21 then the specified types are merged now. Previously, only the last
22 assignment was used. So, if distribution packagers or administrators
23 modified the setting by a drop-in config file, then it may be necessary
24 to update the file.
25
1fc83d09
LP
26 * When OnFailure= is used in combination with Restart= on a service
27 unit, then the specified units will no longer be triggered on
28 failures that result in restarting. Previously, the specified units
29 would be activated each time the unit failed, even when the unit was
30 going to be restarted automatically. This behaviour contradicted the
31 documentation. With this release the code is adjusted to match the
32 documentation.
33
41a4c3ec
LP
34 * systemd-tmpfiles will now print a notice whenever it encounters
35 tmpfiles.d/ lines referencing the /var/run/ directory. It will
36 recommend reworking them to use the /run/ directory instead (for
5cadf58e
ZJS
37 which /var/run/ is simply a symlinked compatibility alias). This way
38 systemd-tmpfiles can properly detect line conflicts and merge lines
39 referencing the same file by two paths, without having to access
40 them.
41a4c3ec 41
ce55bd5e
ZJS
42 * systemctl disable/unmask/preset/preset-all cannot be used with
43 --runtime. Previously this was allowed, but resulted in unintuitive
c7f93e28
ZJS
44 behaviour that wasn't useful. systemctl disable/unmask will now
45 undo both runtime and persistent enablement/masking, i.e. it will
46 remove any relevant symlinks both in /run and /etc.
ce55bd5e 47
5cadf58e
ZJS
48 * sd-boot acquired new loader configuration settings to optionally turn
49 off Windows and MacOS boot partition discovery as well as
50 reboot-into-firmware menu items. It is also able to pick a better
51 screen resolution for HiDPI systems, and now provides loader
41a4c3ec
LP
52 configuration settings to change the resolution explicitly.
53
c086ce8c
LP
54 * systemd-resolved now supports DNS-over-TLS ("PrivateDNS"). It's still
55 turned off by default, use PrivateDNS=yes to turn it on in
56 resolved.conf. We intend to make this the default as soon as couple
57 of additional techniques for optimizing the initial latency caused by
58 establishing a TLS/TCP connection are implemented.
59
41a4c3ec
LP
60 * The systemd-resolve tool has been renamed to resolvectl (it also
61 remains available under the old name, for compatibility), and its
62 interface is now verb-based, similar in style to the other <xyz>ctl
5cadf58e
ZJS
63 tools, such as systemctl or loginctl.
64
65 * The resolvectl/systemd-resolve tool also provides 'resolveconf'
bc99dac5 66 compatibility. It may be symlinked under the 'resolveconf' name, in
5cadf58e
ZJS
67 which case it will take arguments and input compatible with the
68 Debian and FreeBSD resolvconf tool.
41a4c3ec
LP
69
70 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
71 where the system initially suspends, and after a time-out resumes and
72 hibernates again.
73
74 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
75 set the client will only send a DUID as client identifier.
76
77 * The nss-systemd glibc NSS module will now enumerate dynamic users and
78 groups in effect. Previously, it could resolve UIDs/GIDs to user
79 names/groups and vice versa, but did not support enumeration.
80
81 * journald's Compress= configuration setting now optionally accepts a
82 byte threshold value. All journal objects larger than this threshold
83 will be compressed, smaller ones will not. Previously this threshold
84 was not configurable and set to 512.
85
5cadf58e
ZJS
86 * A new system.conf setting NoNewPrivileges= is now available which may
87 be used to turn off acquisition of new privileges system-wide
88 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
89 for all its children). Note that turning this option on means setuid
90 binaries and file system capabilities lose their special powers.
91 While turning on this option is a big step towards a more secure
92 system, doing so is likely to break numerous pre-existing UNIX tools,
93 in particular su and sudo.
41a4c3ec
LP
94
95 * A new service systemd-time-sync-wait.service has been added. If
96 enabled it will delay the time-sync.target unit at boot until time
bc99dac5 97 synchronization has been received from the network. This
41a4c3ec
LP
98 functionality is useful on systems lacking a local RTC or where it is
99 acceptable that the boot process shall be delayed by external network
100 services.
101
102 * When hibernating, systemd will now inform the kernel of the image
103 write offset, on kernels new enough to support this. This means swap
104 files should work for hibernation now.
105
5cadf58e
ZJS
106 * When loading unit files, systemd will now look for drop-in unit files
107 extensions in additional places. Previously, for a unit file name
41a4c3ec
LP
108 "foo-bar-baz.service" it would look for dropin files in
109 "foo-bar-baz.service.d/*.conf". Now, it will also look in
110 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
111 service name truncated after all inner dashes. This scheme allows
112 writing drop-ins easily that apply to a whole set of unit files at
113 once. It's particularly useful for mount and slice units (as their
5cadf58e
ZJS
114 naming is prefix based), but is also useful for service and other
115 units, for packages that install multiple unit files at once,
41a4c3ec 116 following a strict naming regime of beginning the unit file name with
5cadf58e
ZJS
117 the package's name. Two new specifiers are now supported in unit
118 files to match this: %j and %J are replaced by the part of the unit
119 name following the last dash.
120
121 * Unit files and other configuration files that support specifier
122 expansion now understand another two new specifiers: %T and %V will
123 resolve to /tmp and /var/tmp respectively, or whatever temporary
124 directory has been set for the calling user.
41a4c3ec
LP
125
126 * The ExecStart= lines of unit files are no longer required to
127 reference absolute paths. If non-absolute paths are specified the
128 specified binary name is searched within the service manager's
5cadf58e
ZJS
129 built-in $PATH, which may be queried with 'systemd-path
130 search-binaries-default'. It's generally recommended to continue to
131 use absolute paths for all binaries specified in unit files.
41a4c3ec 132
c7f93e28
ZJS
133 * Units gained a new load state "bad-setting", which is used when a
134 unit file was loaded, but contained fatal errors which prevent it
135 from being started (for example, an ExecStart= path which references
136 a non-existent executable).
137
41a4c3ec
LP
138 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
139 support alternative debuggers, for example lldb. The old name
140 continues to be available however, for compatibility reasons. Use the
5cadf58e
ZJS
141 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
142 to pick an alternative debugger instead of the default gdb.
41a4c3ec
LP
143
144 * systemctl and the other tools will now output escape sequences that
145 generate proper clickable hyperlinks in various terminal emulators
146 where useful (for example, in the "systemctl status" output you can
147 now click on the unit file name to quickly open it in the
148 editor/viewer of your choice). Note that not all terminal emulators
149 support this functionality yet, but many do. Unfortunately, the
150 "less" pager doesn't support this yet, hence this functionality is
151 currently automatically turned off when a pager is started (which
152 happens quite often due to auto-paging). We hope to remove this
5cadf58e
ZJS
153 limitation as soon as "less" learns these escape sequences. This new
154 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
155 environment variable. For details on these escape sequences see:
41a4c3ec
LP
156 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
157
158 * networkd's .network files now support a new IPv6MTUBytes= option for
159 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
160 option in the [Route] section to configure the MTU to use for
161 specific routes. It also gained support for configuration of the DHCP
162 "UserClass" option through the new UserClass= setting. It gained
163 three new options in the new [CAN] section for configuring CAN
164 networks. The MULTICAST and ALLMULTI interface flags may now be
165 controlled explicitly with the new Multicast= and AllMulticast=
166 settings.
167
168 * networkd will now automatically make use of the kernel's route
169 expiration feature, if it is available.
170
5cadf58e
ZJS
171 * udevd's .link files now support setting the number of receive and
172 transmit channels, using the RxChannels=, TxChannels=,
173 OtherChannels=, CombinedChannels= settings.
174
175 * Support for UDPSegmentationOffload= has been removed, given its
176 limited support in hardware, and waning software support.
41a4c3ec
LP
177
178 * networkd's .netdev files now support creating "netdevsim" interfaces.
179
180 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
181 to query the unit belonging to a specific kernel control group.
182
5cadf58e 183 * systemd-analyze gained a new verb "cat-config", which may be used to
41a4c3ec
LP
184 dump the contents of any configuration file, with all its matching
185 drop-in files added in, and honouring the usual search and masking
186 logic applied to systemd configuration files. For example use
187 "systemd-analyze cat-config systemd/system.conf" to get the complete
188 system configuration file of systemd how it would be loaded by PID 1
5cadf58e
ZJS
189 itself. Similar to this, various tools such as systemd-tmpfiles or
190 systemd-sysusers, gained a new option "--cat-config", which does the
41a4c3ec
LP
191 corresponding operation for their own configuration settings. For
192 example, "systemd-tmpfiles --cat-config" will now output the full
193 list of tmpfiles.d/ lines in place.
194
195 * timedatectl gained two new verbs "timesync-status" (to show the
196 current NTP synchronization state of systemd-timesyncd) and
197 "show-timesync" (to show bus properties of systemd-timesyncd).
198
199 * systemd-timesyncd gained a bus interface on which it exposes details
200 about its state.
201
202 * systemd-nspawn gained a new --rlimit= switch for setting initial
203 resource limits for the container payload. There's a new switch
5cadf58e 204 --hostname= to explicitly override the container's hostname. A new
41a4c3ec
LP
205 --no-new-privileges= switch may be used to control the
206 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
207 --oom-score-adjust= switch controls the OOM scoring adjustment value
208 for the payload. The new --cpu-affinity= switch controls the CPU
209 affinity of the container payload. The new --resolv-conf= switch
210 allows more detailed control of /etc/resolv.conf handling of the
5cadf58e 211 container. Similarly, the new --timezone= switch allows more detailed
41a4c3ec
LP
212 control of /etc/localtime handling of the container.
213
5cadf58e 214 * systemd-detect-virt gained a new --list switch, which will print a
41a4c3ec
LP
215 list of all currently known VM and container environments.
216
5cadf58e 217 * Support for "Portable Services" has been added, see
41a4c3ec 218 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
5cadf58e
ZJS
219 experimental, but this is expected to change soon. Reflecting this
220 experimental state, the "portablectl" binary is not installed into
41a4c3ec
LP
221 /usr/bin yet. The binary has to be called with the full path
222 /usr/lib/systemd/portablectl instead.
223
224 * journalctl's and systemctl's -o switch now knows a new log output
225 mode "with-unit". The output it generates is very similar to the
226 regular "short" mode, but displays the unit name instead of the
227 syslog tag for each log line. Also, the date is shown with timezone
228 information. This mode is probably more useful than the classic
229 "short" output mode for most purposes, except where pixel-perfect
230 compatibility with classic /var/log/messages formatting is required.
231
232 * A new --dump-bus-properties switch has been added to the systemd
233 binary, which may be used to dump all supported D-Bus properties.
c7f93e28
ZJS
234 (Options which are still supported, but are deprecated, are *not*
235 shown.)
41a4c3ec 236
41a4c3ec
LP
237 * sd-bus gained a set of new calls:
238 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
239 enable/disable the "floating" state of a bus slot object,
240 i.e. whether the slot object pins the bus it is allocated for into
241 memory or if the bus slot object gets disconnected when the bus goes
242 away. sd_bus_open_with_description(),
243 sd_bus_open_user_with_description(),
244 sd_bus_open_system_with_description() may be used to allocate bus
245 objects and set their description string already during allocation.
246
247 * sd-event gained support for watching inotify events from the event
248 loop, in an efficient way, sharing inotify handles between multiple
249 users. For this a new function sd_event_add_inotify() has been added.
250
251 * sd-event and sd-bus gained support for calling special user-supplied
252 destructor functions for userdata pointers associated with
c7f93e28
ZJS
253 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
254 functions sd_bus_slot_set_destroy_callback,
255 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
256 sd_bus_track_get_destroy_callback,
257 sd_event_source_set_destroy_callback,
258 sd_event_source_get_destroy_callback have been added.
41a4c3ec
LP
259
260 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
261
262 * PID 1 will now automatically reschedule .timer units whenever the
5cadf58e 263 local timezone changes. (They previously got rescheduled
41a4c3ec
LP
264 automatically when the system clock changed.)
265
266 * New documentation has been added to document cgroups delegation,
267 portable services and the various code quality tools we have set up:
268
269 https://github.com/systemd/systemd/blob/master/doc/CGROUP_DELEGATION.md
270 https://github.com/systemd/systemd/blob/master/doc/PORTABLE_SERVICES.md
271 https://github.com/systemd/systemd/blob/master/doc/CODE_QUALITY.md
272
273 * pam_systemd will now look for PAM userdata fields systemd.memory_max,
274 systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
275 earlier PAM modules. The data in these fields is used to initialize
276 the session scope's resource properties. Thus external PAM modules
277 may now configure per-session limits, for example sourced from
278 external user databases.
279
280 * socket units with Accept=yes will now maintain a "refused" counter in
281 addition to the existing "accepted" counter, counting connections
282 refused due to the enforced limits.
283
284 * The "systemd-path search-binaries-default" command may now be use to
285 query the default, built-in $PATH PID 1 will pass to the services it
286 manages.
287
c49a7cbd
LP
288 * A new unit file setting PrivateMounts= has been added. It's a boolean
289 option. If enabled the unit's processes are invoked in their own file
290 system namespace. Note that this behaviour is also implied if any
291 other file system namespacing options (such as PrivateTmp=,
292 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
293 primarily useful for services that do not use any of the other file
294 system namespacing options. One such service is systemd-udevd.service
295 wher this is now used by default.
296
41a4c3ec
LP
297 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
298 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Antique, Arnaud
299 Rebillout, Brian J. Murrell, Bruno Vernay, Chris Lesiak, Christian
300 Brauner, Christian Hesse, Daniel Dao, Daniel Lin, Danylo Korostil,
301 Davide Cavalca, David Tardon, Dimitri John Ledkov, Dmitriy Geels,
302 Douglas Christman, Elia Geretto, emelenas, Evegeny Vereshchagin, Evgeny
303 Vereshchagin, Felipe Sateler, Feng Sun, Filipe Brandenburger, Franck
304 Bui, futpib, Giuseppe Scrivano, Guillem Jover, guixxx, Hans de Goede,
305 Henrique Dante de Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko,
306 Ivan Shapovalov, James Cowgill, Jan Janssen, Jan Synacek, Jared
307 Kazimir, João Paulo Rechi Vita, Joost Heitbrink, juergbi, Jui-Chi Ricky
308 Liang, Kai-Heng Feng, Karol Augustin, Krzysztof Nowicki, Lauri
309 Tirkkonen, Lennart Poettering, Leonard, Long Li, Luca Boccassi, Lucas
310 Werkmeister, Marcel Hoppe, Marc Kleine-Budde, Mario Limonciello, Martin
311 Jansa, Martin Wilck, Mathieu Malaterre, Matteo F. Vescovi, Michael
312 Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal Sekletar,
313 Mike Gilbert, Mikhail Kasimov, Milan, Milan Broz, mourikwa, Muhammet
314 Kara, Nicolas Boichat, Omer Katz, Paride Legovini, Paul Menzel, Paul
315 Milliken, Peter A. Bigot, Peter Hutterer, Peter Jones, Philip Sequeira,
316 Philip Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de
317 Araujo, Rosen Penev, rubensa, Ryan Gonzalez, Salvo 'LtWorf' Tomaselli,
318 Sebastian Reichel, Sergio Lindo Mansilla, Stefan Schweter, Stephen
319 Hemminger, Stuart Hayes, Susant Sahani, Sylvain Plantefève, Thomas
320 H. P. Andersen, Tobias Jungel, Tomasz Torcz, Vito Caputo, Will Dietz,
321 Will Thompson, xginn8, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
322
323 — Berlin, 2018-06-XX
324
c657bff1 325CHANGES WITH 238:
e0c46a73
LP
326
327 * The MemoryAccounting= unit property now defaults to on. After
328 discussions with the upstream control group maintainers we learnt
329 that the negative impact of cgroup memory accounting on current
330 kernels is finally relatively minimal, so that it should be safe to
444d5863
ZJS
331 enable this by default without affecting system performance. Besides
332 memory accounting only task accounting is turned on by default, all
333 other forms of resource accounting (CPU, IO, IP) remain off for now,
334 because it's not clear yet that their impact is small enough to move
335 from opt-in to opt-out. We recommend downstreams to leave memory
07a35e84 336 accounting on by default if kernel 4.14 or higher is primarily
444d5863
ZJS
337 used. On very resource constrained systems or when support for old
338 kernels is a necessity, -Dmemory-accounting-default=false can be used
339 to revert this change.
e0c46a73 340
313c32c3
ZJS
341 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
342 %udev_rules_update) and the journal catalog (%journal_catalog_update)
343 from the upgrade scriptlets of individual packages now do nothing.
344 Transfiletriggers have been added which will perform those updates
345 once at the end of the transaction.
346
347 Similar transfiletriggers have been added to execute any sysctl.d
348 and binfmt.d rules. Thus, it should be unnecessary to provide any
349 scriptlets to execute this configuration from package installation
350 scripts.
351
352 * systemd-sysusers gained a mode where the configuration to execute is
353 specified on the command line, but this configuration is not executed
354 directly, but instead it is merged with the configuration on disk,
355 and the result is executed. This is useful for package installation
356 scripts which want to create the user before installing any files on
357 disk (in case some of those files are owned by that user), while
358 still allowing local admin overrides.
359
07a35e84 360 This functionality is exposed to rpm scriptlets through a new
313c32c3
ZJS
361 %sysusers_create_package macro. Old %sysusers_create and
362 %sysusers_create_inline macros are deprecated.
363
364 A transfiletrigger for sysusers.d configuration is now installed,
07a35e84 365 which means that it should be unnecessary to call systemd-sysusers from
313c32c3
ZJS
366 package installation scripts, unless the package installs any files
367 owned by those newly-created users, in which case
368 %sysusers_create_package should be used.
369
370 * Analogous change has been done for systemd-tmpfiles: it gained a mode
371 where the command-line configuration is merged with the configuration
372 on disk. This is exposed as the new %tmpfiles_create_package macro,
373 and %tmpfiles_create is deprecated. A transfiletrigger is installed
374 for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
375 from package installation scripts.
376
377 * sysusers.d configuration for a user may now also specify the group
378 number, in addition to the user number ("u username 123:456"), or
379 without the user number ("u username -:456").
380
381 * Configution items for systemd-sysusers can now be specified as
382 positional arguments when the new --inline switch is used.
383
384 * The login shell of users created through sysusers.d may now be
385 specified (previously, it was always /bin/sh for root and
386 /sbin/nologin for other users).
387
388 * systemd-analyze gained a new --global switch to look at global user
389 configuration. It also gained a unit-paths verb to list the unit load
390 paths that are compiled into systemd (which can be used with
391 --systemd, --user, or --global).
392
393 * udevadm trigger gained a new --settle/-w option to wait for any
394 triggered events to finish (but just those, and not any other events
395 which are triggered meanwhile).
396
397 * The action that systemd-logind takes when the lid is closed and the
398 machine is connected to external power can now be configured using
399 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
400 was determined by HandleLidSwitch=, and, for backwards compatibility,
401 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
402
403 * journalctl will periodically call sd_journal_process() to make it
404 resilient against inotify queue overruns when journal files are
405 rotated very quickly.
406
407 * Two new functions in libsystemd — sd_bus_get_n_queued_read and
408 sd_bus_get_n_queued_write — may be used to check the number of
409 pending bus messages.
410
411 * systemd gained a new
412 org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
413 which can be used to migrate foreign processes to scope and service
414 units. The primary user for this new API is systemd itself: the
415 systemd --user instance uses this call of the systemd --system
416 instance to migrate processes if it itself gets the request to
417 migrate processes and the kernel refuses this due to access
418 restrictions. Thanks to this "systemd-run --scope --user …" works
419 again in pure cgroups v2 environments when invoked from the user
420 session scope.
421
422 * A new TemporaryFileSystem= setting can be used to mask out part of
423 the real file system tree with tmpfs mounts. This may be combined
424 with BindPaths= and BindReadOnlyPaths= to hide files or directories
425 not relevant to the unit, while still allowing some paths lower in
426 the tree to be accessed.
427
428 ProtectHome=tmpfs may now be used to hide user home and runtime
429 directories from units, in a way that is mostly equivalent to
430 "TemporaryFileSystem=/home /run/user /root".
431
432 * Non-service units are now started with KeyringMode=shared by default.
433 This means that mount and swapon and other mount tools have access
434 to keys in the main keyring.
435
436 * /sys/fs/bpf is now mounted automatically.
437
438 * QNX virtualization is now detected by systemd-detect-virt and may
439 be used in ConditionVirtualization=.
440
441 * IPAccounting= may now be enabled also for slice units.
442
443 * A new -Dsplit-bin= build configuration switch may be used to specify
444 whether bin and sbin directories are merged, or if they should be
445 included separately in $PATH and various listings of executable
446 directories. The build configuration scripts will try to autodetect
447 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
448 system, but distributions are encouraged to configure this
449 explicitly.
450
451 * A new -Dok-color= build configuration switch may be used to change
452 the colour of "OK" status messages.
453
454 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
455 PrivateNetwork=yes was buggy in previous versions of systemd. This
456 means that after the upgrade and daemon-reexec, any such units must
457 be restarted.
458
459 * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
460 will not exclude read-only files owned by root from cleanup.
461
c657bff1
ZJS
462 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
463 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
464 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
465 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
466 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
467 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
468 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
469 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
470 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
471 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
472 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
473 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
474 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
475 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
476 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
477 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
478
479 — Warsaw, 2018-03-05
480
82c8e3e6 481CHANGES WITH 237:
2b0c59ba
MP
482
483 * Some keyboards come with a zoom see-saw or rocker which until now got
484 mapped to the Linux "zoomin/out" keys in hwdb. However, these
485 keycodes are not recognized by any major desktop. They now produce
486 Up/Down key events so that they can be used for scrolling.
487
49e87292
LP
488 * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
489 slightly: previously, if an argument was specified for lines of this
490 type (i.e. the right-most column was set) this string was appended to
491 existing files each time systemd-tmpfiles was run. This behaviour was
492 different from what the documentation said, and not particularly
493 useful, as repeated systemd-tmpfiles invocations would not be
494 idempotent and grow such files without bounds. With this release
495 behaviour has been altered slightly, to match what the documentation
496 says: lines of this type only have an effect if the indicated files
497 don't exist yet, and only then the argument string is written to the
498 file.
499
82c8e3e6
LP
500 * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
501 systemd-tmpfiles behaviour: previously, read-only files owned by root
502 were always excluded from the file "aging" algorithm (i.e. the
503 automatic clean-up of directories like /tmp based on
504 atime/mtime/ctime). We intend to drop this restriction, and age files
505 by default even when owned by root and read-only. This behaviour was
506 inherited from older tools, but there have been requests to remove
507 it, and it's not obvious why this restriction was made in the first
508 place. Please speak up now, if you are aware of software that reqires
509 this behaviour, otherwise we'll remove the restriction in v238.
510
95894b91
LP
511 * A new environment variable $SYSTEMD_OFFLINE is now understood by
512 systemctl. It takes a boolean argument. If on, systemctl assumes it
513 operates on an "offline" OS tree, and will not attempt to talk to the
514 service manager. Previously, this mode was implicitly enabled if a
515 chroot() environment was detected, and this new environment variable
516 now provides explicit control.
517
1a0cd2c7
ZJS
518 * .path and .socket units may now be created transiently, too.
519 Previously only service, mount, automount and timer units were
95894b91
LP
520 supported as transient units. The systemd-run tool has been updated
521 to expose this new functionality, you may hence use it now to bind
522 arbitrary commands to path or socket activation on-the-fly from the
1a0cd2c7
ZJS
523 command line. Moreover, almost all properties are now exposed for the
524 unit types that already supported transient operation.
95894b91
LP
525
526 * The systemd-mount command gained support for a new --owner= parameter
527 which takes a user name, which is then resolved and included in uid=
528 and gid= mount options string of the file system to mount.
529
530 * A new unit condition ConditionControlGroupController= has been added
531 that checks whether a specific cgroup controller is available.
532
533 * Unit files, udev's .link files, and systemd-networkd's .netdev and
534 .network files all gained support for a new condition
535 ConditionKernelVersion= for checking against specific kernel
536 versions.
537
538 * In systemd-networkd, the [IPVLAN] section in .netdev files gained
6cddc792 539 support for configuring device flags in the Flags= setting. In the
95894b91
LP
540 same files, the [Tunnel] section gained support for configuring
541 AllowLocalRemote=. The [Route] section in .network files gained
542 support for configuring InitialCongestionWindow=,
543 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
544 understands RapidCommit=.
545
546 * systemd-networkd's DHCPv6 support gained support for Prefix
547 Delegation.
548
549 * sd-bus gained support for a new "watch-bind" feature. When this
550 feature is enabled, an sd_bus connection may be set up to connect to
551 an AF_UNIX socket in the file system as soon as it is created. This
552 functionality is useful for writing early-boot services that
553 automatically connect to the system bus as soon as it is started,
554 without ugly time-based polling. systemd-networkd and
555 systemd-resolved have been updated to make use of this
556 functionality. busctl exposes this functionality in a new
557 --watch-bind= command line switch.
558
559 * sd-bus will now optionally synthesize a local "Connected" signal as
560 soon as a D-Bus connection is set up fully. This message mirrors the
561 already existing "Disconnected" signal which is synthesized when the
562 connection is terminated. This signal is generally useful but
563 particularly handy in combination with the "watch-bind" feature
564 described above. Synthesizing of this message has to be requested
565 explicitly through the new API call sd_bus_set_connected_signal(). In
566 addition a new call sd_bus_is_ready() has been added that checks
caf2a2d8 567 whether a connection is fully set up (i.e. between the "Connected" and
95894b91
LP
568 "Disconnected" signals).
569
570 * sd-bus gained two new calls sd_bus_request_name_async() and
571 sd_bus_release_name_async() for asynchronously registering bus
572 names. Similar, there is now sd_bus_add_match_async() for installing
573 a signal match asynchronously. All of systemd's own services have
574 been updated to make use of these calls. Doing these operations
575 asynchronously has two benefits: it reduces the risk of deadlocks in
576 case of cyclic dependencies between bus services, and it speeds up
577 service initialization since synchronization points for bus
578 round-trips are removed.
579
580 * sd-bus gained two new calls sd_bus_match_signal() and
581 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
582 and sd_bus_add_match_async() but instead of taking a D-Bus match
583 string take match fields as normal function parameters.
584
585 * sd-bus gained two new calls sd_bus_set_sender() and
586 sd_bus_message_set_sender() for setting the sender name of outgoing
587 messages (either for all outgoing messages or for just one specific
588 one). These calls are only useful in direct connections as on
589 brokered connections the broker fills in the sender anyway,
590 overwriting whatever the client filled in.
591
592 * sd-event gained a new pseudo-handle that may be specified on all API
593 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
594 used this refers to the default event loop object of the calling
595 thread. Note however that this does not implicitly allocate one —
6cddc792
CR
596 which has to be done prior by using sd_event_default(). Similarly
597 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
95894b91
LP
598 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
599 to the default bus of the specified type of the calling thread. Here
600 too this does not implicitly allocate bus connection objects, this
601 has to be done prior with sd_bus_default() and friends.
602
603 * sd-event gained a new call pair
6cddc792
CR
604 sd_event_source_{get|set}_io_fd_own(). This may be used to request
605 automatic closure of the file descriptor an IO event source watches
95894b91
LP
606 when the event source is destroyed.
607
608 * systemd-networkd gained support for natively configuring WireGuard
609 connections.
610
6cddc792
CR
611 * In previous versions systemd synthesized user records both for the
612 "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
613 internally. In order to simplify distribution-wide renames of the
95894b91
LP
614 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
615 new transitional flag file has been added: if
616 /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
617 user and group record within the systemd codebase is disabled.
618
619 * systemd-notify gained a new --uid= option for selecting the source
620 user/UID to use for notification messages sent to the service
621 manager.
622
31751f7e 623 * journalctl gained a new --grep= option to list only entries in which
e6501af8
ZJS
624 the message matches a certain pattern. By default matching is case
625 insensitive if the pattern is lowercase, and case sensitive
626 otherwise. Option --case-sensitive=yes|no can be used to override
627 this an specify case sensitivity or case insensitivity.
628
56a29112 629 * There's now a "systemd-analyze service-watchdogs" command for printing
508058c9 630 the current state of the service runtime watchdog, and optionally
56a29112 631 enabling or disabling the per-service watchdogs system-wide if given a
508058c9
LP
632 boolean argument (i.e. the concept you configure in WatchdogSec=), for
633 debugging purposes. There's also a kernel command line option
56a29112 634 systemd.service_watchdogs= for controlling the same.
508058c9
LP
635
636 * Two new "log-level" and "log-target" options for systemd-analyze were
bc99dac5 637 added that merge the now deprecated get-log-level, set-log-level and
508058c9
LP
638 get-log-target, set-log-target pairs. The deprecated options are still
639 understood for backwards compatibility. The two new options print the
640 current value when no arguments are given, and set them when a
56a29112 641 level/target is given as an argument.
95894b91 642
508058c9
LP
643 * sysusers.d's "u" lines now optionally accept both a UID and a GID
644 specification, separated by a ":" character, in order to create users
645 where UID and GID do not match.
646
95894b91 647 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
508058c9
LP
648 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
649 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
650 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
651 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
652 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
653 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
654 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
655 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
656 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
657 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
658 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
659 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
660 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
661 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
662 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
663 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
664 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
665 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
666 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
667 Палаузов
668
669 — Brno, 2018-01-28
2b0c59ba 670
a1b2c92d 671CHANGES WITH 236:
195b943d 672
89780840
ZJS
673 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
674 in v235 has been extended to also set the dummy.ko module option
675 numdummies=0, preventing the kernel from automatically creating
676 dummy0. All dummy interfaces must now be explicitly created.
195b943d 677
3925496a
LP
678 * Unknown '%' specifiers in configuration files are now rejected. This
679 applies to units and tmpfiles.d configuration. Any percent characters
680 that are followed by a letter or digit that are not supposed to be
681 interpreted as the beginning of a specifier should be escaped by
682 doubling ("%%"). (So "size=5%" is still accepted, as well as
683 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
684 valid specifiers today.)
751223fe 685
e6b2d948 686 * systemd-resolved now maintains a new dynamic
89780840
ZJS
687 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
688 recommended to make /etc/resolv.conf a symlink to it. This file
689 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
690 includes dynamically acquired search domains, achieving more correct
691 DNS resolution by software that bypasses local DNS APIs such as NSS.
e6b2d948 692
67eb5b38
LP
693 * The "uaccess" udev tag has been dropped from /dev/kvm and
694 /dev/dri/renderD*. These devices now have the 0666 permissions by
695 default (but this may be changed at build-time). /dev/dri/renderD*
696 will now be owned by the "render" group along with /dev/kfd.
697
89780840
ZJS
698 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
699 systemd-journal-gatewayd.service and
700 systemd-journal-upload.service. This means "nss-systemd" must be
701 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
702 services are resolved properly.
67eb5b38 703
3925496a
LP
704 * In /etc/fstab two new mount options are now understood:
705 x-systemd.makefs and x-systemd.growfs. The former has the effect that
706 the configured file system is formatted before it is mounted, the
707 latter that the file system is resized to the full block device size
708 after it is mounted (i.e. if the file system is smaller than the
709 partition it resides on, it's grown). This is similar to the fsck
710 logic in /etc/fstab, and pulls in systemd-makefs@.service and
711 systemd-growfs@.service as necessary, similar to
712 systemd-fsck@.service. Resizing is currently only supported on ext4
713 and btrfs.
714
67eb5b38
LP
715 * In systemd-networkd, the IPv6 RA logic now optionally may announce
716 DNS server and domain information.
717
718 * Support for the LUKS2 on-disk format for encrypted partitions has
719 been added. This requires libcryptsetup2 during compilation and
720 runtime.
721
89780840 722 * The systemd --user instance will now signal "readiness" when its
67eb5b38
LP
723 basic.target unit has been reached, instead of when the run queue ran
724 empty for the first time.
725
8ea2dcb0
ZJS
726 * Tmpfiles.d with user configuration are now also supported.
727 systemd-tmpfiles gained a new --user switch, and snippets placed in
728 ~/.config/user-tmpfiles.d/ and corresponding directories will be
729 executed by systemd-tmpfiles --user running in the new
730 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
731 running in the user session.
732
733 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
734 %S resolves to the top-level state directory (/var/lib for the system
735 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
736 top-level cache directory (/var/cache for the system instance,
737 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
738 logs directory (/var/log for the system instance,
67eb5b38 739 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
8ea2dcb0 740 existing %t specifier, that resolves to the top-level runtime
67eb5b38
LP
741 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
742 user instance).
743
744 * journalctl learnt a new parameter --output-fields= for limiting the
745 set of journal fields to output in verbose and JSON output modes.
746
747 * systemd-timesyncd's configuration file gained a new option
89780840
ZJS
748 RootDistanceMaxSec= for setting the maximum root distance of servers
749 it'll use, as well as the new options PollIntervalMinSec= and
750 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
67eb5b38
LP
751
752 * bootctl gained a new command "list" for listing all available boot
89780840 753 menu items on systems that follow the boot loader specification.
67eb5b38
LP
754
755 * systemctl gained a new --dry-run switch that shows what would be done
756 instead of doing it, and is currently supported by the shutdown and
757 sleep verbs.
758
e9ad86d5 759 * ConditionSecurity= can now detect the TOMOYO security module.
67eb5b38
LP
760
761 * Unit file [Install] sections are now also respected in unit drop-in
89780840 762 files. This is intended to be used by drop-ins under /usr/lib/.
67eb5b38 763
89780840 764 * systemd-firstboot may now also set the initial keyboard mapping.
67eb5b38 765
89780840
ZJS
766 * Udev "changed" events for devices which are exposed as systemd
767 .device units are now propagated to units specified in
768 ReloadPropagatedFrom= as reload requests.
67eb5b38 769
89780840
ZJS
770 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
771 unit template name (i.e. a name in the form of 'foobar@.service',
772 without the instance component between the '@' and - the '.'), then
773 the escaped sysfs path of the device is automatically used as the
774 instance.
67eb5b38
LP
775
776 * SystemCallFilter= in unit files has been extended so that an "errno"
777 can be specified individually for each system call. Example:
778 SystemCallFilter=~uname:EILSEQ.
779
780 * The cgroup delegation logic has been substantially updated. Delegate=
781 now optionally takes a list of controllers (instead of a boolean, as
782 before), which lists the controllers to delegate at least.
783
89780840 784 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
67eb5b38 785
89780840
ZJS
786 * A new LogLevelMax= setting configures the maximum log level any
787 process of the service may log at (i.e. anything with a lesser
788 priority than what is specified is automatically dropped). A new
789 LogExtraFields= setting allows configuration of additional journal
790 fields to attach to all log records generated by any of the unit's
791 processes.
67eb5b38 792
89780840
ZJS
793 * New StandardInputData= and StandardInputText= settings along with the
794 new option StandardInput=data may be used to configure textual or
795 binary data that shall be passed to the executed service process via
796 standard input, encoded in-line in the unit file.
67eb5b38
LP
797
798 * StandardInput=, StandardOutput= and StandardError= may now be used to
799 connect stdin/stdout/stderr of executed processes directly with a
800 file or AF_UNIX socket in the file system, using the new "file:" option.
801
89780840
ZJS
802 * A new unit file option CollectMode= has been added, that allows
803 tweaking the garbage collection logic for units. It may be used to
804 tell systemd to garbage collect units that have failed automatically
805 (normally it only GCs units that exited successfully). systemd-run
806 and systemd-mount expose this new functionality with a new -G option.
807
67eb5b38
LP
808 * "machinectl bind" may now be used to bind mount non-directories
809 (i.e. regularfiles, devices, fifos, sockets).
810
811 * systemd-analyze gained a new verb "calendar" for validating and
812 testing calendar time specifications to use for OnCalendar= in timer
813 units. Besides validating the expression it will calculate the next
814 time the specified expression would elapse.
815
816 * In addition to the pre-existing FailureAction= unit file setting
89780840
ZJS
817 there's now SuccessAction=, for configuring a shutdown action to
818 execute when a unit completes successfully. This is useful in
819 particular inside containers that shall terminate after some workload
820 has been completed. Also, both options are now supported for all unit
821 types, not just services.
67eb5b38
LP
822
823 * networkds's IP rule support gained two new options
dd014eeb 824 IncomingInterface= and OutgoingInterface= for configuring the incoming
67eb5b38
LP
825 and outgoing interfaces of configured rules. systemd-networkd also
826 gained support for "vxcan" network devices.
827
828 * networkd gained a new setting RequiredForOnline=, taking a
829 boolean. If set, systemd-wait-online will take it into consideration
830 when determining that the system is up, otherwise it will ignore the
831 interface for this purpose.
832
833 * The sd_notify() protocol gained support for a new operation: with
834 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
835 store again, ahead of POLLHUP or POLLERR when they are removed
836 anyway.
837
f09eb768
LP
838 * A new document doc/UIDS-GIDS.md has been added to the source tree,
839 that documents the UID/GID range and assignment assumptions and
3925496a
LP
840 requirements of systemd.
841
842 * The watchdog device PID 1 will ping may now be configured through the
843 WatchdogDevice= configuration file setting, or by setting the
844 systemd.watchdog_service= kernel commandline option.
845
846 * systemd-resolved's gained support for registering DNS-SD services on
847 the local network using MulticastDNS. Services may either be
848 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
849 the same dir below /run, /usr/lib), or through its D-Bus API.
850
a327431b
DB
851 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
852 extend the effective start, runtime, and stop time. The service must
853 continue to send EXTEND_TIMEOUT_USEC within the period specified to
854 prevent the service manager from making the service as timedout.
855
ea2a3c9e
LP
856 * systemd-resolved's DNSSEC support gained support for RFC 8080
857 (Ed25519 keys and signatures).
858
a1b2c92d
LP
859 * The systemd-resolve command line tool gained a new set of options
860 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
861 --set-nta= and --revert to configure per-interface DNS configuration
862 dynamically during runtime. It's useful for pushing DNS information
863 into systemd-resolved from DNS hook scripts that various interface
864 managing software supports (such as pppd).
865
866 * systemd-nspawn gained a new --network-namespace-path= command line
867 option, which may be used to make a container join an existing
868 network namespace, by specifying a path to a "netns" file.
869
3925496a
LP
870 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
871 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
872 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
ea2a3c9e 873 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
a1b2c92d
LP
874 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
875 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
876 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
877 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
878 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
879 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
880 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
881 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
882 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
883 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
884 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
885 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
886 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
887 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
888 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
889 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
890 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
891 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
892 Jędrzejewski-Szmek, Zeal Jagannatha
67eb5b38 893
ea2a3c9e 894 — Berlin, 2017-12-14
3754abc5 895
582faeb4
DJL
896CHANGES WITH 235:
897
2bcbffd6
LP
898 * INCOMPATIBILITY: systemd-logind.service and other long-running
899 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
900 communication with the outside. This generally improves security of
901 the system, and is in almost all cases a safe and good choice, as
23d37367 902 these services do not and should not provide any network-facing
2bcbffd6
LP
903 functionality. However, systemd-logind uses the glibc NSS API to
904 query the user database. This creates problems on systems where NSS
905 is set up to directly consult network services for user database
906 lookups. In particular, this creates incompatibilities with the
907 "nss-nis" module, which attempts to directly contact the NIS/YP
908 network servers it is configured for, and will now consistently
909 fail. In such cases, it is possible to turn off IP sandboxing for
910 systemd-logind.service (set IPAddressDeny= in its [Service] section
911 to the empty string, via a .d/ unit file drop-in). Downstream
912 distributions might want to update their nss-nis packaging to include
913 such a drop-in snippet, accordingly, to hide this incompatibility
914 from the user. Another option is to make use of glibc's nscd service
915 to proxy such network requests through a privilege-separated, minimal
916 local caching daemon, or to switch to more modern technologies such
917 sssd, whose NSS hook-ups generally do not involve direct network
918 access. In general, we think it's definitely time to question the
919 implementation choices of nss-nis, i.e. whether it's a good idea
920 today to embed a network-facing loadable module into all local
921 processes that need to query the user database, including the most
922 trivial and benign ones, such as "ls". For more details about
923 IPAddressDeny= see below.
924
fccf5419
LP
925 * A new modprobe.d drop-in is now shipped by default that sets the
926 bonding module option max_bonds=0. This overrides the kernel default,
927 to avoid conflicts and ambiguity as to whether or not bond0 should be
928 managed by systemd-networkd or not. This resolves multiple issues
929 with bond0 properties not being applied, when bond0 is configured
930 with systemd-networkd. Distributors may choose to not package this,
931 however in that case users will be prevented from correctly managing
932 bond0 interface using systemd-networkd.
582faeb4 933
ef5a8cb1 934 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
21723f53
ZJS
935 which print the logging level and target of the system manager. They
936 complement the existing "set-log-level" and "set-log-target" verbs
937 used to change those values.
ef5a8cb1 938
fccf5419
LP
939 * journald.conf gained a new boolean setting ReadKMsg= which defaults
940 to on. If turned off kernel log messages will not be read by
21723f53
ZJS
941 systemd-journald or included in the logs. It also gained a new
942 setting LineMax= for configuring the maximum line length in
943 STDOUT/STDERR log streams. The new default for this value is 48K, up
944 from the previous hardcoded 2048.
fccf5419 945
21723f53
ZJS
946 * A new unit setting RuntimeDirectoryPreserve= has been added, which
947 allows more detailed control of what to do with a runtime directory
948 configured with RuntimeDirectory= (i.e. a directory below /run or
949 $XDG_RUNTIME_DIR) after a unit is stopped.
fccf5419
LP
950
951 * The RuntimeDirectory= setting for units gained support for creating
952 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
953 one top-level directory.
954
955 * Units gained new options StateDirectory=, CacheDirectory=,
956 LogsDirectory= and ConfigurationDirectory= which are closely related
957 to RuntimeDirectory= but manage per-service directories below
21723f53 958 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
fccf5419
LP
959 possible to write unit files which when activated automatically gain
960 properly owned service specific directories in these locations, thus
961 making unit files self-contained and increasing compatibility with
962 stateless systems and factory reset where /etc or /var are
963 unpopulated at boot. Matching these new settings there's also
964 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
965 ConfigurationDirectoryMode= for configuring the access mode of these
75dfbbac
LP
966 directories. These settings are particularly useful in combination
967 with DynamicUser=yes as they provide secure, properly-owned,
968 writable, and stateful locations for storage, excluded from the
969 sandbox that such services live in otherwise.
fccf5419
LP
970
971 * Automake support has been removed from this release. systemd is now
972 Meson-only.
973
974 * systemd-journald will now aggressively cache client metadata during
975 runtime, speeding up log write performance under pressure. This comes
976 at a small price though: as much of the metadata is read
977 asynchronously from /proc/ (and isn't implicitly attached to log
978 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
979 metadata stored alongside a log entry might be slightly
980 out-of-date. Previously it could only be slightly newer than the log
981 message. The time window is small however, and given that the kernel
982 is unlikely to be improved anytime soon in this regard, this appears
983 acceptable to us.
984
985 * nss-myhostname/systemd-resolved will now by default synthesize an
986 A/AAAA resource record for the "_gateway" hostname, pointing to the
987 current default IP gateway. Previously it did that for the "gateway"
988 name, hampering adoption, as some distributions wanted to leave that
989 host name open for local use. The old behaviour may still be
990 requested at build time.
991
992 * systemd-networkd's [Address] section in .network files gained a new
993 Scope= setting for configuring the IP address scope. The [Network]
994 section gained a new boolean setting ConfigureWithoutCarrier= that
995 tells systemd-networkd to ignore link sensing when configuring the
996 device. The [DHCP] section gained a new Anonymize= boolean option for
997 turning on a number of options suggested in RFC 7844. A new
998 [RoutingPolicyRule] section has been added for configuring the IP
999 routing policy. The [Route] section has gained support for a new
1000 Type= setting which permits configuring
1001 blackhole/unreachable/prohibit routes.
1002
1003 * The [VRF] section in .netdev files gained a new Table= setting for
1004 configuring the routing table to use. The [Tunnel] section gained a
1005 new Independent= boolean field for configuring tunnels independent of
1006 an underlying network interface. The [Bridge] section gained a new
1007 GroupForwardMask= option for configuration of propagation of link
1008 local frames between bridge ports.
1009
1010 * The WakeOnLan= setting in .link files gained support for a number of
1011 new modes. A new TCP6SegmentationOffload= setting has been added for
1012 configuring TCP/IPv6 hardware segmentation offload.
1013
1014 * The IPv6 RA sender implementation may now optionally send out RDNSS
21723f53 1015 and RDNSSL records to supply DNS configuration to peers.
fccf5419
LP
1016
1017 * systemd-nspawn gained support for a new --system-call-filter= command
21723f53
ZJS
1018 line option for adding and removing entries in the default system
1019 call filter it applies. Moreover systemd-nspawn has been changed to
fccf5419
LP
1020 implement a system call whitelist instead of a blacklist.
1021
1022 * systemd-run gained support for a new --pipe command line option. If
1023 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
1024 are directly passed on to the activated transient service
21723f53
ZJS
1025 executable. This allows invoking arbitrary processes as systemd
1026 services (for example to take benefit of dependency management,
1027 accounting management, resource management or log management that is
1028 done automatically for services) — while still allowing them to be
fccf5419
LP
1029 integrated in a classic UNIX shell pipeline.
1030
1031 * When a service sends RELOAD=1 via sd_notify() and reload propagation
1032 using ReloadPropagationTo= is configured, a reload is now propagated
1033 to configured units. (Previously this was only done on explicitly
1034 requested reloads, using "systemctl reload" or an equivalent
1035 command.)
1036
1037 * For each service unit a restart counter is now kept: it is increased
1038 each time the service is restarted due to Restart=, and may be
1039 queried using "systemctl show -p NRestarts …".
1040
44898c53
LP
1041 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
1042 @signal and @timer have been added, for usage with SystemCallFilter=
fccf5419
LP
1043 in unit files and the new --system-call-filter= command line option
1044 of systemd-nspawn (see above).
1045
1046 * ExecStart= lines in unit files gained two new modifiers: when a
1047 command line is prefixed with "!" the command will be executed as
1048 configured, except for the credentials applied by
1049 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
1050 "+", but does still apply namespacing options unlike "+". There's
1051 also "!!" now, which is mostly identical, but becomes a NOP on
1052 systems that support ambient capabilities. This is useful to write
1053 unit files that work with ambient capabilities where possible but
1054 automatically fall back to traditional privilege dropping mechanisms
1055 on systems where this is not supported.
1056
1057 * ListenNetlink= settings in socket units now support RDMA netlink
1058 sockets.
1059
1060 * A new unit file setting LockPersonality= has been added which permits
1061 locking down the chosen execution domain ("personality") of a service
1062 during runtime.
1063
1064 * A new special target "getty-pre.target" has been added, which is
1065 ordered before all text logins, and may be used to order services
21723f53 1066 before textual logins acquire access to the console.
fccf5419
LP
1067
1068 * systemd will now attempt to load the virtio-rng.ko kernel module very
1069 early on if a VM environment supporting this is detected. This should
1070 improve entropy during early boot in virtualized environments.
1071
1072 * A _netdev option is now supported in /etc/crypttab that operates in a
1073 similar way as the same option in /etc/fstab: it permits configuring
21723f53
ZJS
1074 encrypted devices that need to be ordered after the network is up.
1075 Following this logic, two new special targets
fccf5419 1076 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
21723f53
ZJS
1077 added that are to cryptsetup.target what remote-fs.target and
1078 remote-fs-pre.target are to local-fs.target.
fccf5419
LP
1079
1080 * Service units gained a new UnsetEnvironment= setting which permits
21723f53
ZJS
1081 unsetting specific environment variables for services that are
1082 normally passed to it (for example in order to mask out locale
fccf5419
LP
1083 settings for specific services that can't deal with it).
1084
1085 * Units acquired a new boolean option IPAccounting=. When turned on, IP
1086 traffic accounting (packet count as well as byte count) is done for
1087 the service, and shown as part of "systemctl status" or "systemd-run
1088 --wait".
1089
1090 * Service units acquired two new options IPAddressAllow= and
1091 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
1092 for configuring a simple IP access control list for all sockets of
1093 the unit. These options are available also on .slice and .socket
1094 units, permitting flexible access list configuration for individual
1095 services as well as groups of services (as defined by a slice unit),
1096 including system-wide. Note that IP ACLs configured this way are
1097 enforced on every single IPv4 and IPv6 socket created by any process
1098 of the service unit, and apply to ingress as well as egress traffic.
1099
21723f53 1100 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
608f70e6 1101 structured log message is generated each time the unit is stopped,
fccf5419
LP
1102 containing information about the consumed resources of this
1103 invocation.
1104
1105 * A new setting KeyringMode= has been added to unit files, which may be
1106 used to control how the kernel keyring is set up for executed
1107 processes.
1108
e06fafb2
LP
1109 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
1110 "systemctl kexec" and "systemctl exit" are now always asynchronous in
1111 behaviour (that is: these commands return immediately after the
21723f53
ZJS
1112 operation was enqueued instead of waiting for the operation to
1113 complete). Previously, "systemctl poweroff" and "systemctl reboot"
e06fafb2
LP
1114 were asynchronous on systems using systemd-logind (i.e. almost
1115 always, and like they were on sysvinit), and the other three commands
1116 were unconditionally synchronous. With this release this is cleaned
1117 up, and callers will see the same asynchronous behaviour on all
1118 systems for all five operations.
1119
1120 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
1121 the system.
1122
fccf5419
LP
1123 * .timer units now accept calendar specifications in other timezones
1124 than UTC or the local timezone.
1125
f6e64b78 1126 * The tmpfiles snippet var.conf has been changed to create
21723f53
ZJS
1127 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
1128 the "utmp" group already, and it appears to be generally understood
1129 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
1130 databases. Previously this was implemented correctly for all these
1131 databases excepts btmp, which has been opened up like this now
1132 too. Note that while the other databases are world-readable
1133 (i.e. 0644), btmp is not and remains more restrictive.
f6e64b78 1134
d55b0463
LP
1135 * The systemd-resolve tool gained a new --reset-server-features
1136 switch. When invoked like this systemd-resolved will forget
1137 everything it learnt about the features supported by the configured
1138 upstream DNS servers, and restarts the feature probing logic on the
cf84484a
LP
1139 next resolver look-up for them at the highest feature level
1140 again.
1141
1142 * The status dump systemd-resolved sends to the logs upon receiving
1143 SIGUSR1 now also includes information about all DNS servers it is
1144 configured to use, and the features levels it probed for them.
d55b0463 1145
fccf5419
LP
1146 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
1147 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
76451c1d
LP
1148 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
1149 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
1150 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
1151 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
1152 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
1153 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
1154 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
1155 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
1156 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
1157 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
1158 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
1159 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
1160 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
1161 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
1162 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
1163 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
1164 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
1165 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
fccf5419 1166
c1719d8b 1167 — Berlin, 2017-10-06
fccf5419 1168
4b4da299
LP
1169CHANGES WITH 234:
1170
1171 * Meson is now supported as build system in addition to Automake. It is
1172 our plan to remove Automake in one of our next releases, so that
1173 Meson becomes our exclusive build system. Hence, please start using
1174 the Meson build system in your downstream packaging. There's plenty
1175 of documentation around how to use Meson, the extremely brief
1176 summary:
1177
1178 ./autogen.sh && ./configure && make && sudo make install
1179
1180 becomes:
1181
1182 meson build && ninja -C build && sudo ninja -C build install
1183
1184 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
1185 which permits configuring a timeout on the time a job is
1186 running. This is particularly useful for setting timeouts on jobs for
1187 .device units.
1188
1189 * Unit files gained two new options ConditionUser= and ConditionGroup=
1190 for conditionalizing units based on the identity of the user/group
1191 running a systemd user instance.
1192
1193 * systemd-networkd now understands a new FlowLabel= setting in the
1194 [VXLAN] section of .network files, as well as a Priority= in
1195 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
1196 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
1197 gained support for configuration of GENEVE links, and IPv6 address
1198 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
1199
9f09a95a 1200 * .link files now understand a new Port= setting.
4b4da299
LP
1201
1202 * systemd-networkd's DHCP support gained support for DHCP option 119
1203 (domain search list).
1204
1205 * systemd-networkd gained support for serving IPv6 address ranges using
bc99dac5 1206 the Router Advertisement protocol. The new .network configuration
4b4da299
LP
1207 section [IPv6Prefix] may be used to configure the ranges to
1208 serve. This is implemented based on a new, minimal, native server
1209 implementation of RA.
1210
1211 * journalctl's --output= switch gained support for a new parameter
1212 "short-iso-precise" for a mode where timestamps are shown as precise
1213 ISO date values.
1214
1215 * systemd-udevd's "net_id" builtin may now generate stable network
1216 interface names from IBM PowerVM VIO devices as well as ACPI platform
1217 devices.
1218
1219 * MulticastDNS support in systemd-resolved may now be explicitly
1220 enabled/disabled using the new MulticastDNS= configuration file
1221 option.
1222
1223 * systemd-resolved may now optionally use libidn2 instead of the libidn
7f7ab228
ZJS
1224 for processing internationalized domain names. Support for libidn2
1225 should be considered experimental and should not be enabled by
1226 default yet.
4b4da299
LP
1227
1228 * "machinectl pull-tar" and related call may now do verification of
1229 downloaded images using SUSE-style .sha256 checksum files in addition
1230 to the already existing support for validating using Ubuntu-style
1231 SHA256SUMS files.
1232
1233 * sd-bus gained support for a new sd_bus_message_appendv() call which
1234 is va_list equivalent of sd_bus_message_append().
1235
1236 * sd-boot gained support for validating images using SHIM/MOK.
1237
1238 * The SMACK code learnt support for "onlycap".
1239
1240 * systemd-mount --umount is now much smarter in figuring out how to
1241 properly unmount a device given its mount or device path.
5486a31d
ZJS
1242
1243 * The code to call libnss_dns as a fallback from libnss_resolve when
1244 the communication with systemd-resolved fails was removed. This
1245 fallback was redundant and interfered with the [!UNAVAIL=return]
1246 suffix. See nss-resolve(8) for the recommended configuration.
1247
9f09a95a
ZJS
1248 * systemd-logind may now be restarted without losing state. It stores
1249 the file descriptors for devices it manages in the system manager
38d93385 1250 using the FDSTORE= mechanism. Please note that further changes in
9f09a95a
ZJS
1251 other components may be required to make use of this (for example
1252 Xorg has code to listen for stops of systemd-logind and terminate
1253 itself when logind is stopped or restarted, in order to avoid using
1254 stale file descriptors for graphical devices, which is now
1255 counterproductive and must be reverted in order for restarts of
1256 systemd-logind to be safe. See
1257 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
1258
9d8813b3
YW
1259 * All kernel install plugins are called with the environment variable
1260 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
1261 /etc/machine-id. If the file is missing or empty, the variable is
1262 empty and BOOT_DIR_ABS is the path of a temporary directory which is
38d93385 1263 removed after all the plugins exit. So, if KERNEL_INSTALL_MACHINE_ID
9d8813b3
YW
1264 is empty, all plugins should not put anything in BOOT_DIR_ABS.
1265
184d2c15 1266 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
ac172e52
LP
1267 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
1268 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
1269 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
1270 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
184d2c15
LP
1271 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
1272 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
1273 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
1274 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
ac172e52
LP
1275 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
1276 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
1277 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
1278 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
1279 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
1280 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
1281 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
1282 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
1283 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
1284 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
1285 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
1286 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
1287 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
1288 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
1289 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
1290 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
184d2c15
LP
1291 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
1292 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
ac172e52
LP
1293 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
1294 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
1295 Георгиевски
4b4da299 1296
ac172e52 1297 — Berlin, 2017-07-12
4b4da299 1298
a2b53448 1299CHANGES WITH 233:
d08ee7cb 1300
23eb30b3
ZJS
1301 * The "hybrid" control group mode has been modified to improve
1302 compatibility with "legacy" cgroups-v1 setups. Specifically, the
1303 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
1304 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
1305 cgroups-v1 hierarchy), the only externally visible change being that
1306 the cgroups-v2 hierarchy is also mounted, to
1307 /sys/fs/cgroup/unified. This should provide a large degree of
1308 compatibility with "legacy" cgroups-v1, while taking benefit of the
1309 better management capabilities of cgroups-v2.
1310
1311 * The default control group setup mode may be selected both a boot-time
1312 via a set of kernel command line parameters (specifically:
1313 systemd.unified_cgroup_hierarchy= and
1314 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
1315 default selected on the configure command line
1316 (--with-default-hierarchy=). The upstream default is "hybrid"
1317 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
1318 this will change in a future systemd version to be "unified" (pure
1319 cgroups-v2 mode). The third option for the compile time option is
1320 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
1321 distributions to default to "hybrid" mode for release distributions,
1322 starting with v233. We recommend "unified" for development
1323 distributions (specifically: distributions such as Fedora's rawhide)
1324 as that's where things are headed in the long run. Use "legacy" for
1325 greatest stability and compatibility only.
1326
1327 * Note one current limitation of "unified" and "hybrid" control group
1328 setup modes: the kernel currently does not permit the systemd --user
1329 instance (i.e. unprivileged code) to migrate processes between two
1330 disconnected cgroup subtrees, even if both are managed and owned by
1331 the user. This effectively means "systemd-run --user --scope" doesn't
1332 work when invoked from outside of any "systemd --user" service or
1333 scope. Specifically, it is not supported from session scopes. We are
1334 working on fixing this in a future systemd version. (See #3388 for
1335 further details about this.)
1336
fb7c4eff
MG
1337 * DBus policy files are now installed into /usr rather than /etc. Make
1338 sure your system has dbus >= 1.9.18 running before upgrading to this
1339 version, or override the install path with --with-dbuspolicydir= .
1340
23eb30b3
ZJS
1341 * All python scripts shipped with systemd (specifically: the various
1342 tests written in Python) now require Python 3.
1343
d60c5270 1344 * systemd unit tests can now run standalone (without the source or
4dfe64f8
ZJS
1345 build directories), and can be installed into /usr/lib/systemd/tests/
1346 with 'make install-tests'.
1347
23eb30b3
ZJS
1348 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
1349 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
1350 kernel.
1351
1352 * Support for the %c, %r, %R specifiers in unit files has been
1353 removed. Specifiers are not supposed to be dependent on configuration
1354 in the unit file itself (so that they resolve the same regardless
1355 where used in the unit files), but these specifiers were influenced
1356 by the Slice= option.
1357
5cfc0a84
LP
1358 * The shell invoked by debug-shell.service now defaults to /bin/sh in
1359 all cases. If distributions want to use a different shell for this
1360 purpose (for example Fedora's /sbin/sushell) they need to specify
1361 this explicitly at configure time using --with-debug-shell=.
1362
2bcc3309
FB
1363 * The confirmation spawn prompt has been reworked to offer the
1364 following choices:
1365
b0eb2944 1366 (c)ontinue, proceed without asking anymore
dd6f9ac0 1367 (D)ump, show the state of the unit
2bcc3309 1368 (f)ail, don't execute the command and pretend it failed
d172b175 1369 (h)elp
eedf223a 1370 (i)nfo, show a short summary of the unit
56fde33a 1371 (j)obs, show jobs that are in progress
2bcc3309
FB
1372 (s)kip, don't execute the command and pretend it succeeded
1373 (y)es, execute the command
1374
1375 The 'n' choice for the confirmation spawn prompt has been removed,
1376 because its meaning was confusing.
1377
d08ee7cb
LP
1378 The prompt may now also be redirected to an alternative console by
1379 specifying the console as parameter to systemd.confirm_spawn=.
1380
8e458bfe
JW
1381 * Services of Type=notify require a READY=1 notification to be sent
1382 during startup. If no such message is sent, the service now fails,
1383 even if the main process exited with a successful exit code.
1384
85266f9b
LP
1385 * Services that fail to start up correctly now always have their
1386 ExecStopPost= commands executed. Previously, they'd enter "failed"
1387 state directly, without executing these commands.
1388
baf32786
MP
1389 * The option MulticastDNS= of network configuration files has acquired
1390 an actual implementation. With MulticastDNS=yes a host can resolve
23eb30b3 1391 names of remote hosts and reply to mDNS A and AAAA requests.
fa8b4499 1392
631b676b
LP
1393 * When units are about to be started an additional check is now done to
1394 ensure that all dependencies of type BindsTo= (when used in
1395 combination with After=) have been started.
1396
d08ee7cb
LP
1397 * systemd-analyze gained a new verb "syscall-filter" which shows which
1398 system call groups are defined for the SystemCallFilter= unit file
23eb30b3 1399 setting, and which system calls they contain.
d08ee7cb
LP
1400
1401 * A new system call filter group "@filesystem" has been added,
23eb30b3 1402 consisting of various file system related system calls. Group
d08ee7cb 1403 "@reboot" has been added, covering reboot, kexec and shutdown related
23eb30b3 1404 calls. Finally, group "@swap" has been added covering swap
d08ee7cb
LP
1405 configuration related calls.
1406
1407 * A new unit file option RestrictNamespaces= has been added that may be
1408 used to restrict access to the various process namespace types the
1409 Linux kernel provides. Specifically, it may be used to take away the
23eb30b3
ZJS
1410 right for a service unit to create additional file system, network,
1411 user, and other namespaces. This sandboxing option is particularly
1412 relevant due to the high amount of recently discovered namespacing
1413 related vulnerabilities in the kernel.
d08ee7cb 1414
23eb30b3
ZJS
1415 * systemd-udev's .link files gained support for a new AutoNegotiation=
1416 setting for configuring Ethernet auto-negotiation.
d08ee7cb
LP
1417
1418 * systemd-networkd's .network files gained support for a new
1419 ListenPort= setting in the [DHCP] section to explicitly configure the
1420 UDP client port the DHCP client shall listen on.
1421
23eb30b3
ZJS
1422 * .network files gained a new Unmanaged= boolean setting for explicitly
1423 excluding one or more interfaces from management by systemd-networkd.
1424
1425 * The systemd-networkd ProxyARP= option has been renamed to
1426 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
1427 renamed to ReduceARPProxy=. The old names continue to be available
1428 for compatibility.
1429
1430 * systemd-networkd gained support for configuring IPv6 Proxy NDP
1431 addresses via the new IPv6ProxyNDPAddress= .network file setting.
1432
1433 * systemd-networkd's bonding device support gained support for two new
1434 configuration options ActiveSlave= and PrimarySlave=.
1435
1436 * The various options in the [Match] section of .network files gained
1437 support for negative matching.
1438
d08ee7cb
LP
1439 * New systemd-specific mount options are now understood in /etc/fstab:
1440
1441 x-systemd.mount-timeout= may be used to configure the maximum
1442 permitted runtime of the mount command.
1443
1444 x-systemd.device-bound may be set to bind a mount point to its
1445 backing device unit, in order to automatically remove a mount point
1446 if its backing device is unplugged. This option may also be
1447 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
1448 on the block device, which is now automatically set for all CDROM
1449 drives, so that mounted CDs are automatically unmounted when they are
1450 removed from the drive.
1451
23eb30b3
ZJS
1452 x-systemd.after= and x-systemd.before= may be used to explicitly
1453 order a mount after or before another unit or mount point.
d08ee7cb
LP
1454
1455 * Enqueued start jobs for device units are now automatically garbage
1456 collected if there are no jobs waiting for them anymore.
1457
23eb30b3
ZJS
1458 * systemctl list-jobs gained two new switches: with --after, for every
1459 queued job the jobs it's waiting for are shown; with --before the
1460 jobs which it's blocking are shown.
d08ee7cb
LP
1461
1462 * systemd-nspawn gained support for ephemeral boots from disk images
1463 (or in other words: --ephemeral and --image= may now be
1464 combined). Moreover, ephemeral boots are now supported for normal
1465 directories, even if the backing file system is not btrfs. Of course,
baf32786
MP
1466 if the file system does not support file system snapshots or
1467 reflinks, the initial copy operation will be relatively expensive, but
1468 this should still be suitable for many use cases.
d08ee7cb
LP
1469
1470 * Calendar time specifications in .timer units now support
1471 specifications relative to the end of a month by using "~" instead of
1472 "-" as separator between month and day. For example, "*-02~03" means
23eb30b3 1473 "the third last day in February". In addition a new syntax for
d08ee7cb
LP
1474 repeated events has been added using the "/" character. For example,
1475 "9..17/2:00" means "every two hours from 9am to 5pm".
1476
1477 * systemd-socket-proxyd gained a new parameter --connections-max= for
1478 configuring the maximum number of concurrent connections.
1479
23eb30b3
ZJS
1480 * sd-id128 gained a new API for generating unique IDs for the host in a
1481 way that does not leak the machine ID. Specifically,
d08ee7cb 1482 sd_id128_get_machine_app_specific() derives an ID based on the
baf32786 1483 machine ID a in well-defined, non-reversible, stable way. This is
d08ee7cb
LP
1484 useful whenever an identifier for the host is needed but where the
1485 identifier shall not be useful to identify the system beyond the
1486 scope of the application itself. (Internally this uses HMAC-SHA256 as
1487 keyed hash function using the machine ID as input.)
1488
1489 * NotifyAccess= gained a new supported value "exec". When set
1490 notifications are accepted from all processes systemd itself invoked,
1491 including all control processes.
1492
1493 * .nspawn files gained support for defining overlay mounts using the
1494 Overlay= and OverlayReadOnly= options. Previously this functionality
1495 was only available on the systemd-nspawn command line.
1496
1497 * systemd-nspawn's --bind= and --overlay= options gained support for
1498 bind/overlay mounts whose source lies within the container tree by
1499 prefixing the source path with "+".
1500
1501 * systemd-nspawn's --bind= and --overlay= options gained support for
1502 automatically allocating a temporary source directory in /var/tmp
1503 that is removed when the container dies. Specifically, if the source
1504 directory is specified as empty string this mechanism is selected. An
1505 example usage is --overlay=+/var::/var, which creates an overlay
1506 mount based on the original /var contained in the image, overlayed
1507 with a temporary directory in the host's /var/tmp. This way changes
1508 to /var are automatically flushed when the container shuts down.
1509
baf32786
MP
1510 * systemd-nspawn --image= option does now permit raw file system block
1511 devices (in addition to images containing partition tables, as
1512 before).
d08ee7cb
LP
1513
1514 * The disk image dissection logic in systemd-nspawn gained support for
1515 automatically setting up LUKS encrypted as well as Verity protected
1516 partitions. When a container is booted from an encrypted image the
1517 passphrase is queried at start-up time. When a container with Verity
1518 data is started, the root hash is search in a ".roothash" file
1519 accompanying the disk image (alternatively, pass the root hash via
1520 the new --root-hash= command line option).
1521
1522 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
1523 be used to dissect disk images the same way as systemd-nspawn does
1524 it, following the Bootable Partition Specification. It may even be
1525 used to mount disk images with complex partition setups (including
1526 LUKS and Verity partitions) to a local host directory, in order to
1527 inspect them. This tool is not considered public API (yet), and is
1528 thus not installed into /usr/bin. Please do not rely on its
3b31c466 1529 existence, since it might go away or be changed in later systemd
d08ee7cb
LP
1530 versions.
1531
1532 * A new generator "systemd-verity-generator" has been added, similar in
baf32786 1533 style to "systemd-cryptsetup-generator", permitting automatic setup of
d08ee7cb
LP
1534 Verity root partitions when systemd boots up. In order to make use of
1535 this your partition setup should follow the Discoverable Partitions
1536 Specification, and the GPT partition ID of the root file system
1537 partition should be identical to the upper 128bit of the Verity root
1538 hash. The GPT partition ID of the Verity partition protecting it
1539 should be the lower 128bit of the Verity root hash. If the partition
1540 image follows this model it is sufficient to specify a single
1541 "roothash=" kernel command line argument to both configure which root
1542 image and verity partition to use as well as the root hash for
1543 it. Note that systemd-nspawn's Verity support follows the same
1544 semantics, meaning that disk images with proper Verity data in place
1545 may be booted in containers with systemd-nspawn as well as on
1546 physical systems via the verity generator. Also note that the "mkosi"
1547 tool available at https://github.com/systemd/mkosi has been updated
1548 to generate Verity protected disk images following this scheme. In
1549 fact, it has been updated to generate disk images that optionally
1550 implement a complete UEFI SecureBoot trust chain, involving a signed
1551 kernel and initrd image that incorporates such a root hash as well as
1552 a Verity-enabled root partition.
1553
d08ee7cb
LP
1554 * The hardware database (hwdb) udev supports has been updated to carry
1555 accelerometer quirks.
1556
1557 * All system services are now run with a fresh kernel keyring set up
1558 for them. The invocation ID is stored by default in it, thus
1559 providing a safe, non-overridable way to determine the invocation
1560 ID of each service.
1561
1562 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
1563 options for bind mounting arbitrary paths in a service-specific
1564 way. When these options are used, arbitrary host or service files and
1565 directories may be mounted to arbitrary locations in the service's
1566 view.
1567
1568 * Documentation has been added that lists all of systemd's low-level
1569 environment variables:
1570
f09eb768 1571 https://github.com/systemd/systemd/blob/master/doc/ENVIRONMENT.md
d08ee7cb
LP
1572
1573 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
1574 whether a specific socket file descriptor matches a specified socket
1575 address.
1576
1577 * systemd-firstboot has been updated to check for the
1578 systemd.firstboot= kernel command line option. It accepts a boolean
1579 and when set to false the first boot questions are skipped.
1580
d08ee7cb 1581 * systemd-fstab-generator has been updated to check for the
23eb30b3
ZJS
1582 systemd.volatile= kernel command line option, which either takes an
1583 optional boolean parameter or the special value "state". If used the
1584 system may be booted in a "volatile" boot mode. Specifically,
1585 "systemd.volatile" is used, the root directory will be mounted as
d08ee7cb 1586 tmpfs, and only /usr is mounted from the actual root file system. If
23eb30b3
ZJS
1587 "systemd.volatile=state" is used, the root directory will be mounted
1588 as usual, but /var is mounted as tmpfs. This concept provides similar
d08ee7cb
LP
1589 functionality as systemd-nspawn's --volatile= option, but provides it
1590 on physical boots. Use this option for implementing stateless
1591 systems, or testing systems with all state and/or configuration reset
1592 to the defaults. (Note though that many distributions are not
23eb30b3 1593 prepared to boot up without a populated /etc or /var, though.)
d08ee7cb
LP
1594
1595 * systemd-gpt-auto-generator gained support for LUKS encrypted root
1596 partitions. Previously it only supported LUKS encrypted partitions
1597 for all other uses, except for the root partition itself.
1598
1599 * Socket units gained support for listening on AF_VSOCK sockets for
1600 communication in virtualized QEMU environments.
1601
1602 * The "configure" script gained a new option --with-fallback-hostname=
1603 for specifying the fallback hostname to use if none is configured in
1604 /etc/hostname. For example, by specifying
1605 --with-fallback-hostname=fedora it is possible to default to a
23eb30b3 1606 hostname of "fedora" on pristine installations.
d08ee7cb
LP
1607
1608 * systemd-cgls gained support for a new --unit= switch for listing only
1609 the control groups of a specific unit. Similar --user-unit= has been
1610 added for listing only the control groups of a specific user unit.
1611
1612 * systemd-mount gained a new --umount switch for unmounting a mount or
1613 automount point (and all mount/automount points below it).
1614
1615 * systemd will now refuse full configuration reloads (via systemctl
1616 daemon-reload and related calls) unless at least 16MiB of free space
1617 are available in /run. This is a safety precaution in order to ensure
1618 that generators can safely operate after the reload completed.
1619
1620 * A new unit file option RootImage= has been added, which has a similar
1621 effect as RootDirectory= but mounts the service's root directory from
1622 a disk image instead of plain directory. This logic reuses the same
1623 image dissection and mount logic that systemd-nspawn already uses,
1624 and hence supports any disk images systemd-nspawn supports, including
1625 those following the Discoverable Partition Specification, as well as
1626 Verity enabled images. This option enables systemd to run system
1627 services directly off disk images acting as resource bundles,
1628 possibly even including full integrity data.
1629
1630 * A new MountAPIVFS= unit file option has been added, taking a boolean
baf32786 1631 argument. If enabled /proc, /sys and /dev (collectively called the
d08ee7cb
LP
1632 "API VFS") will be mounted for the service. This is only relevant if
1633 RootDirectory= or RootImage= is used for the service, as these mounts
1634 are of course in place in the host mount namespace anyway.
1635
1636 * systemd-nspawn gained support for a new --pivot-root= switch. If
1637 specified the root directory within the container image is pivoted to
1638 the specified mount point, while the original root disk is moved to a
1639 different place. This option enables booting of ostree images
1640 directly with systemd-nspawn.
1641
d08ee7cb 1642 * The systemd build scripts will no longer complain if the NTP server
23eb30b3 1643 addresses are not changed from the defaults. Google now supports
d08ee7cb
LP
1644 these NTP servers officially. We still recommend downstreams to
1645 properly register an NTP pool with the NTP pool project though.
1646
c1ec34d1 1647 * coredumpctl gained a new "--reverse" option for printing the list
d08ee7cb
LP
1648 of coredumps in reverse order.
1649
23eb30b3
ZJS
1650 * coredumpctl will now show additional information about truncated and
1651 inaccessible coredumps, as well as coredumps that are still being
1652 processed. It also gained a new --quiet switch for suppressing
1653 additional informational message in its output.
1654
1655 * coredumpctl gained support for only showing coredumps newer and/or
1656 older than specific timestamps, using the new --since= and --until=
1657 options, reminiscent of journalctl's options by the same name.
1658
d08ee7cb 1659 * The systemd-coredump logic has been improved so that it may be reused
23eb30b3 1660 to collect backtraces in non-compiled languages, for example in
d08ee7cb
LP
1661 scripting languages such as Python.
1662
1663 * machinectl will now show the UID shift of local containers, if user
1664 namespacing is enabled for them.
1665
baf32786 1666 * systemd will now optionally run "environment generator" binaries at
d08ee7cb
LP
1667 configuration load time. They may be used to add environment
1668 variables to the environment block passed to services invoked. One
baf32786 1669 user environment generator is shipped by default that sets up
23eb30b3
ZJS
1670 environment variables based on files dropped into /etc/environment.d
1671 and ~/.config/environment.d/.
d08ee7cb 1672
a2b53448
LP
1673 * systemd-resolved now includes the new, recently published 2017 DNSSEC
1674 root key (KSK).
1675
a2b53448
LP
1676 * hostnamed has been updated to report a new chassis type of
1677 "convertible" to cover "foldable" laptops that can both act as a
1678 tablet and as a laptop, such as various Lenovo Yoga devices.
1679
d08ee7cb
LP
1680 Contributions from: Adrián López, Alexander Galanin, Alexander
1681 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
1682 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
1683 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
1684 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
1685 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
1686 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
1687 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
1688 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
a2b53448
LP
1689 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
1690 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
1691 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
1692 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
1693 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
1694 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
1695 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
1696 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
1697 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
1698 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
1699 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
1700 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
1701 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
1702 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
1703 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
1704 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
1705 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
1706 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
1707 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
1708 Тихонов
1709
1710 — Berlin, 2017-03-01
d08ee7cb 1711
54b24597 1712CHANGES WITH 232:
76153ad4 1713
05f426d2
LP
1714 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
1715 RestrictAddressFamilies= enabled. These sandboxing options should
1716 generally be compatible with the various external udev call-out
1717 binaries we are aware of, however there may be exceptions, in
1718 particular when exotic languages for these call-outs are used. In
1719 this case, consider turning off these settings locally.
1720
4ffe2479
ZJS
1721 * The new RemoveIPC= option can be used to remove IPC objects owned by
1722 the user or group of a service when that service exits.
1723
6fa44114 1724 * The new ProtectKernelModules= option can be used to disable explicit
4c37970d
LP
1725 load and unload operations of kernel modules by a service. In
1726 addition access to /usr/lib/modules is removed if this option is set.
6fa44114 1727
4a77c53d
ZJS
1728 * ProtectSystem= option gained a new value "strict", which causes the
1729 whole file system tree with the exception of /dev, /proc, and /sys,
1730 to be remounted read-only for a service.
1731
e49e2c25 1732 * The new ProtectKernelTunables= option can be used to disable
4a77c53d
ZJS
1733 modification of configuration files in /sys and /proc by a service.
1734 Various directories and files are remounted read-only, so access is
1735 restricted even if the file permissions would allow it.
1736
6fa44114 1737 * The new ProtectControlGroups= option can be used to disable write
4a77c53d
ZJS
1738 access by a service to /sys/fs/cgroup.
1739
1740 * Various systemd services have been hardened with
1741 ProtectKernelTunables=yes, ProtectControlGroups=yes,
1742 RestrictAddressFamilies=.
1743
4ffe2479
ZJS
1744 * Support for dynamically creating users for the lifetime of a service
1745 has been added. If DynamicUser=yes is specified, user and group IDs
1746 will be allocated from the range 61184..65519 for the lifetime of the
1747 service. They can be resolved using the new nss-systemd.so NSS
1748 module. The module must be enabled in /etc/nsswitch.conf. Services
1749 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
1750 any resources allocated by the service will be cleaned up when the
4a77c53d
ZJS
1751 service exits. They also have ProtectHome=read-only and
1752 ProtectSystem=strict enabled, so they are not able to make any
1753 permanent modifications to the system.
4ffe2479 1754
171ae2cd 1755 * The nss-systemd module also always resolves root and nobody, making
4ffe2479 1756 it possible to have no /etc/passwd or /etc/group files in minimal
171ae2cd 1757 container or chroot environments.
4ffe2479
ZJS
1758
1759 * Services may be started with their own user namespace using the new
171ae2cd
LP
1760 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
1761 under which the service is running are mapped. All other users are
1762 mapped to nobody.
4ffe2479
ZJS
1763
1764 * Support for the cgroup namespace has been added to systemd-nspawn. If
1765 supported by kernel, the container system started by systemd-nspawn
1766 will have its own view of the cgroup hierarchy. This new behaviour
1767 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
1768
1769 * The new MemorySwapMax= option can be used to limit the maximum swap
1770 usage under the unified cgroup hierarchy.
1771
1772 * Support for the CPU controller in the unified cgroup hierarchy has
1773 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
1774 options. This controller requires out-of-tree patches for the kernel
1775 and the support is provisional.
1776
171ae2cd
LP
1777 * Mount and automount units may now be created transiently
1778 (i.e. dynamically at runtime via the bus API, instead of requiring
1779 unit files in the file system).
1780
1781 * systemd-mount is a new tool which may mount file systems – much like
1782 mount(8), optionally pulling in additional dependencies through
1783 transient .mount and .automount units. For example, this tool
1784 automatically runs fsck on a backing block device before mounting,
1785 and allows the automount logic to be used dynamically from the
1786 command line for establishing mount points. This tool is particularly
1787 useful when dealing with removable media, as it will ensure fsck is
1788 run – if necessary – before the first access and that the file system
1789 is quickly unmounted after each access by utilizing the automount
1790 logic. This maximizes the chance that the file system on the
1791 removable media stays in a clean state, and if it isn't in a clean
1792 state is fixed automatically.
4ffe2479
ZJS
1793
1794 * LazyUnmount=yes option for mount units has been added to expose the
1795 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
1796 option.
1797
1798 * /efi will be used as the mount point of the EFI boot partition, if
1799 the directory is present, and the mount point was not configured
1800 through other means (e.g. fstab). If /efi directory does not exist,
1801 /boot will be used as before. This makes it easier to automatically
1802 mount the EFI partition on systems where /boot is used for something
1803 else.
1804
171ae2cd
LP
1805 * When operating on GPT disk images for containers, systemd-nspawn will
1806 now mount the ESP to /boot or /efi according to the same rules as PID
1807 1 running on a host. This allows tools like "bootctl" to operate
1808 correctly within such containers, in order to make container images
1809 bootable on physical systems.
1810
4a77c53d 1811 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
4ffe2479
ZJS
1812
1813 * Two new user session targets have been added to support running
1814 graphical sessions under the systemd --user instance:
1815 graphical-session.target and graphical-session-pre.target. See
1816 systemd.special(7) for a description of how those targets should be
1817 used.
1818
1819 * The vconsole initialization code has been significantly reworked to
d4c08299 1820 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
4ffe2479
ZJS
1821 support unicode keymaps. Font and keymap configuration will now be
1822 copied to all allocated virtual consoles.
1823
05ecf467 1824 * FreeBSD's bhyve virtualization is now detected.
4ffe2479 1825
d4c08299 1826 * Information recorded in the journal for core dumps now includes the
4ffe2479
ZJS
1827 contents of /proc/mountinfo and the command line of the process at
1828 the top of the process hierarchy (which is usually the init process
1829 of the container).
1830
171ae2cd 1831 * systemd-journal-gatewayd learned the --directory= option to serve
4ffe2479
ZJS
1832 files from the specified location.
1833
1834 * journalctl --root=… can be used to peruse the journal in the
1835 /var/log/ directories inside of a container tree. This is similar to
1836 the existing --machine= option, but does not require the container to
1837 be active.
1838
1839 * The hardware database has been extended to support
1840 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
1841 trackball devices.
1842
1843 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
1844 specify the click rate for mice which include a horizontal wheel with
1845 a click rate that is different than the one for the vertical wheel.
1846
1847 * systemd-run gained a new --wait option that makes service execution
171ae2cd
LP
1848 synchronous. (Specifically, the command will not return until the
1849 specified service binary exited.)
4ffe2479 1850
171ae2cd 1851 * systemctl gained a new --wait option that causes the start command to
4a77c53d
ZJS
1852 wait until the units being started have terminated again.
1853
171ae2cd 1854 * A new journal output mode "short-full" has been added which displays
4ffe2479 1855 timestamps with abbreviated English day names and adds a timezone
171ae2cd
LP
1856 suffix. Those timestamps include more information than the default
1857 "short" output mode, and can be passed directly to journalctl's
1858 --since= and --until= options.
4ffe2479
ZJS
1859
1860 * /etc/resolv.conf will be bind-mounted into containers started by
1861 systemd-nspawn, if possible, so any changes to resolv.conf contents
1862 are automatically propagated to the container.
1863
1864 * The number of instances for socket-activated services originating
171ae2cd
LP
1865 from a single IP address can be limited with
1866 MaxConnectionsPerSource=, extending the existing setting of
1867 MaxConnections=.
4ffe2479 1868
4a77c53d
ZJS
1869 * systemd-networkd gained support for vcan ("Virtual CAN") interface
1870 configuration.
1871
1872 * .netdev and .network configuration can now be extended through
1873 drop-ins.
1874
4ffe2479
ZJS
1875 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
1876 Segmentation Offload, Generic Receive Offload, Large Receive Offload
1877 can be enabled and disabled using the new UDPSegmentationOffload=,
1878 TCPSegmentationOffload=, GenericSegmentationOffload=,
1879 GenericReceiveOffload=, LargeReceiveOffload= options in the
1880 [Link] section of .link files.
1881
171ae2cd
LP
1882 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
1883 Port VLAN ID can be configured for bridge devices using the new STP=,
1884 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
1885 section of .netdev files.
4ffe2479 1886
171ae2cd 1887 * The route table to which routes received over DHCP or RA should be
4a77c53d
ZJS
1888 added can be configured with the new RouteTable= option in the [DHCP]
1889 and [IPv6AcceptRA] sections of .network files.
1890
171ae2cd 1891 * The Address Resolution Protocol can be disabled on links managed by
4ffe2479
ZJS
1892 systemd-networkd using the ARP=no setting in the [Link] section of
1893 .network files.
1894
171ae2cd
LP
1895 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
1896 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
1897 encode information about the result and exit codes of the current
1898 service runtime cycle.
4ffe2479 1899
4a77c53d 1900 * systemd-sysctl will now configure kernel parameters in the order
1f4f4cf7 1901 they occur in the configuration files. This matches what sysctl
4a77c53d
ZJS
1902 has been traditionally doing.
1903
1904 * kernel-install "plugins" that are executed to perform various
1905 tasks after a new kernel is added and before an old one is removed
1906 can now return a special value to terminate the procedure and
1907 prevent any later plugins from running.
1908
76153ad4 1909 * Journald's SplitMode=login setting has been deprecated. It has been
d4c08299 1910 removed from documentation, and its use is discouraged. In a future
76153ad4
ZJS
1911 release it will be completely removed, and made equivalent to current
1912 default of SplitMode=uid.
1913
4a77c53d
ZJS
1914 * Storage=both option setting in /etc/systemd/coredump.conf has been
1915 removed. With fast LZ4 compression storing the core dump twice is not
1916 useful.
1917
4ffe2479
ZJS
1918 * The --share-system systemd-nspawn option has been replaced with an
1919 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
1920 this functionality is discouraged. In addition the variables
1921 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
1922 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
1923 individual namespaces.
1924
171ae2cd
LP
1925 * "machinectl list" now shows the IP address of running containers in
1926 the output, as well as OS release information.
1927
1928 * "loginctl list" now shows the TTY of each session in the output.
1929
1930 * sd-bus gained new API calls sd_bus_track_set_recursive(),
1931 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
1932 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
1933 tracking objects in a "recursive" mode, where a single client can be
1934 counted multiple times, if it takes multiple references.
1935
1936 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
bc99dac5 1937 sd_bus_get_exit_on_disconnect(). They may be used to make a
171ae2cd
LP
1938 process using sd-bus automatically exit if the bus connection is
1939 severed.
1940
1941 * Bus clients of the service manager may now "pin" loaded units into
1942 memory, by taking an explicit reference on them. This is useful to
1943 ensure the client can retrieve runtime data about the service even
1944 after the service completed execution. Taking such a reference is
1945 available only for privileged clients and should be helpful to watch
1946 running services in a race-free manner, and in particular collect
1947 information about exit statuses and results.
1948
4c37970d
LP
1949 * The nss-resolve module has been changed to strictly return UNAVAIL
1950 when communication via D-Bus with resolved failed, and NOTFOUND when
1951 a lookup completed but was negative. This means it is now possible to
1952 neatly configure fallbacks using nsswitch.conf result checking
1953 expressions. Taking benefit of this, the new recommended
1954 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
1955
1956 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
1957
1958 * A new setting CtrlAltDelBurstAction= has been added to
1959 /etc/systemd/system.conf which may be used to configure the precise
1960 behaviour if the user on the console presses Ctrl-Alt-Del more often
1961 than 7 times in 2s. Previously this would unconditionally result in
1962 an expedited, immediate reboot. With this new setting the precise
1963 operation may be configured in more detail, and also turned off
1964 entirely.
1965
1966 * In .netdev files two new settings RemoteChecksumTx= and
1967 RemoteChecksumRx= are now understood that permit configuring the
1968 remote checksumming logic for VXLAN networks.
1969
1970 * The service manager learnt a new "invocation ID" concept for invoked
1971 services. Each runtime cycle of a service will get a new invocation
1972 ID (a 128bit random UUID) assigned that identifies the current
1973 run of the service uniquely and globally. A new invocation ID
1974 is generated each time a service starts up. The journal will store
1975 the invocation ID of a service along with any logged messages, thus
1976 making the invocation ID useful for matching the online runtime of a
1977 service with the offline log data it generated in a safe way without
1978 relying on synchronized timestamps. In many ways this new service
1979 invocation ID concept is similar to the kernel's boot ID concept that
1980 uniquely and globally identifies the runtime of each boot. The
1981 invocation ID of a service is passed to the service itself via an
1982 environment variable ($INVOCATION_ID). A new bus call
1983 GetUnitByInvocationID() has been added that is similar to GetUnit()
1984 but instead of retrieving the bus path for a unit by its name
1985 retrieves it by its invocation ID. The returned path is valid only as
1986 long as the passed invocation ID is current.
1987
1988 * systemd-resolved gained a new "DNSStubListener" setting in
1989 resolved.conf. It either takes a boolean value or the special values
1990 "udp" and "tcp", and configures whether to enable the stub DNS
1991 listener on 127.0.0.53:53.
1992
1993 * IP addresses configured via networkd may now carry additional
1994 configuration settings supported by the kernel. New options include:
1995 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
1996 PrefixRoute=, AutoJoin=.
1997
1998 * The PAM configuration fragment file for "user@.service" shipped with
1999 systemd (i.e. the --user instance of systemd) has been stripped to
2000 the minimum necessary to make the system boot. Previously, it
2001 contained Fedora-specific stanzas that did not apply to other
2002 distributions. It is expected that downstream distributions add
2003 additional configuration lines, matching their needs to this file,
2004 using it only as rough template of what systemd itself needs. Note
2005 that this reduced fragment does not even include an invocation of
2006 pam_limits which most distributions probably want to add, even though
2007 systemd itself does not need it. (There's also the new build time
2008 option --with-pamconfdir=no to disable installation of the PAM
2009 fragment entirely.)
2010
2011 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
2012 capability is now also dropped from its set (in addition to
2013 CAP_SYS_MKNOD as before).
2014
2015 * In service unit files it is now possible to connect a specific named
2016 file descriptor with stdin/stdout/stdout of an executed service. The
2017 name may be specified in matching .socket units using the
2018 FileDescriptorName= setting.
2019
2020 * A number of journal settings may now be configured on the kernel
2021 command line. Specifically, the following options are now understood:
2022 systemd.journald.max_level_console=,
2023 systemd.journald.max_level_store=,
2024 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
2025 systemd.journald.max_level_wall=.
2026
2027 * "systemctl is-enabled --full" will now show by which symlinks a unit
2028 file is enabled in the unit dependency tree.
2029
b4eed568
LP
2030 * Support for VeraCrypt encrypted partitions has been added to the
2031 "cryptsetup" logic and /etc/crypttab.
2032
2033 * systemd-detect-virt gained support for a new --private-users switch
2034 that checks whether the invoking processes are running inside a user
2035 namespace. Similar, a new special value "private-users" for the
2036 existing ConditionVirtualization= setting has been added, permitting
2037 skipping of specific units in user namespace environments.
2038
07393b6e
LP
2039 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
2040 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
2041 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
2042 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
2043 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
2044 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
2045 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
2046 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
2047 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
2048 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
2049 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
2050 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
2051 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
2052 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
2053 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
2054 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
2055 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
2056 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
2057 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
2058 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
2059 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
2060 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
2061 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
2062 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
2063 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
2064 Jędrzejewski-Szmek, Zeal Jagannatha
2065
54b24597 2066 — Santa Fe, 2016-11-03
07393b6e 2067
5cd118ba
MP
2068CHANGES WITH 231:
2069
fcd30826
LP
2070 * In service units the various ExecXYZ= settings have been extended
2071 with an additional special character as first argument of the
43eb109a 2072 assigned value: if the character '+' is used the specified command
fcd30826
LP
2073 line it will be run with full privileges, regardless of User=,
2074 Group=, CapabilityBoundingSet= and similar options. The effect is
2075 similar to the existing PermissionsStartOnly= option, but allows
2076 configuration of this concept for each executed command line
2077 independently.
2078
2079 * Services may now alter the service watchdog timeout at runtime by
2080 sending a WATCHDOG_USEC= message via sd_notify().
2081
2082 * MemoryLimit= and related unit settings now optionally take percentage
2083 specifications. The percentage is taken relative to the amount of
2084 physical memory in the system (or in case of containers, the assigned
2085 amount of memory). This allows scaling service resources neatly with
771de3f5 2086 the amount of RAM available on the system. Similarly, systemd-logind's
fcd30826
LP
2087 RuntimeDirectorySize= option now also optionally takes percentage
2088 values.
2089
2090 * In similar fashion TasksMax= takes percentage values now, too. The
2091 value is taken relative to the configured maximum number of processes
2092 on the system. The per-service task maximum has been changed to 15%
2093 using this functionality. (Effectively this is an increase of 512 →
2094 4915 for service units, given the kernel's default pid_max setting.)
2095
2096 * Calendar time specifications in .timer units now understand a ".."
2097 syntax for time ranges. Example: "4..7:10" may now be used for
2098 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
2099 7:10am every day.
2100
2101 * The InaccessableDirectories=, ReadOnlyDirectories= and
2102 ReadWriteDirectories= unit file settings have been renamed to
2103 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
2104 applied to all kinds of file nodes, and not just directories, with
2105 the exception of symlinks. Specifically these settings may now be
2106 used on block and character device nodes, UNIX sockets and FIFOS as
2107 well as regular files. The old names of these settings remain
2108 available for compatibility.
2109
2110 * systemd will now log about all service processes it kills forcibly
2111 (using SIGKILL) because they remained after the clean shutdown phase
2112 of the service completed. This should help identifying services that
2113 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
2114 systemd-logind's configuration a similar log message is generated for
2115 processes killed at the end of each session due to this setting.
2116
2117 * systemd will now set the $JOURNAL_STREAM environment variable for all
2118 services whose stdout/stderr are connected to the Journal (which
2119 effectively means by default: all services). The variable contains
2120 the device and inode number of the file descriptor used for
2121 stdout/stderr. This may be used by invoked programs to detect whether
2122 their stdout/stderr is connected to the Journal, in which case they
2123 can switch over to direct Journal communication, thus being able to
2124 pass extended, structured metadata along with their log messages. As
2125 one example, this is now used by glib's logging primitives.
2126
2127 * When using systemd's default tmp.mount unit for /tmp, the mount point
2128 will now be established with the "nosuid" and "nodev" options. This
2129 avoids privilege escalation attacks that put traps and exploits into
2130 /tmp. However, this might cause problems if you e. g. put container
5cd118ba
MP
2131 images or overlays into /tmp; if you need this, override tmp.mount's
2132 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
2133 desired options.
2134
fcd30826
LP
2135 * systemd now supports the "memory" cgroup controller also on
2136 cgroupsv2.
2137
2138 * The systemd-cgtop tool now optionally takes a control group path as
2139 command line argument. If specified, the control group list shown is
2140 limited to subgroups of that group.
2141
2142 * The SystemCallFilter= unit file setting gained support for
2143 pre-defined, named system call filter sets. For example
2144 SystemCallFilter=@clock is now an effective way to make all clock
771de3f5 2145 changing-related system calls unavailable to a service. A number of
fcd30826
LP
2146 similar pre-defined groups are defined. Writing system call filters
2147 for system services is simplified substantially with this new
2148 concept. Accordingly, all of systemd's own, long-running services now
2149 enable system call filtering based on this, by default.
2150
2151 * A new service setting MemoryDenyWriteExecute= has been added, taking
2152 a boolean value. If turned on, a service may no longer create memory
2153 mappings that are writable and executable at the same time. This
2154 enhances security for services where this is enabled as it becomes
2155 harder to dynamically write and then execute memory in exploited
2156 service processes. This option has been enabled for all of systemd's
2157 own long-running services.
2158
2159 * A new RestrictRealtime= service setting has been added, taking a
2160 boolean argument. If set the service's processes may no longer
2161 acquire realtime scheduling. This improves security as realtime
2162 scheduling may otherwise be used to easily freeze the system.
2163
2164 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
2165 value. This may be used for requesting that the system manager inside
2166 of the container reports start-up completion to nspawn which then
2167 propagates this notification further to the service manager
2168 supervising nspawn itself. A related option NotifyReady= in .nspawn
2169 files has been added too. This functionality allows ordering of the
2170 start-up of multiple containers using the usual systemd ordering
2171 primitives.
2172
2173 * machinectl gained a new command "stop" that is an alias for
2174 "terminate".
2175
2176 * systemd-resolved gained support for contacting DNS servers on
2177 link-local IPv6 addresses.
2178
2179 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
2180 its caches. A method call for requesting the same operation has been
2181 added to the bus API too, and is made available via "systemd-resolve
2182 --flush-caches".
2183
771de3f5 2184 * systemd-resolve gained a new --status switch. If passed a brief
fcd30826
LP
2185 summary of the used DNS configuration with per-interface information
2186 is shown.
2187
2188 * resolved.conf gained a new Cache= boolean option, defaulting to
2189 on. If turned off local DNS caching is disabled. This comes with a
2190 performance penalty in particular when DNSSEC is enabled. Note that
771de3f5 2191 resolved disables its internal caching implicitly anyway, when the
fcd30826
LP
2192 configured DNS server is on a host-local IP address such as ::1 or
2193 127.0.0.1, thus automatically avoiding double local caching.
2194
2195 * systemd-resolved now listens on the local IP address 127.0.0.53:53
2196 for DNS requests. This improves compatibility with local programs
2197 that do not use the libc NSS or systemd-resolved's bus APIs for name
2198 resolution. This minimal DNS service is only available to local
2199 programs and does not implement the full DNS protocol, but enough to
2200 cover local DNS clients. A new, static resolv.conf file, listing just
2201 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
2202 now recommended to make /etc/resolv.conf a symlink to this file in
2203 order to route all DNS lookups to systemd-resolved, regardless if
2204 done via NSS, the bus API or raw DNS packets. Note that this local
2205 DNS service is not as fully featured as the libc NSS or
2206 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
2207 used to deliver link-local address information (as this implies
2208 sending a local interface index along), LLMNR/mDNS support via this
2209 interface is severely restricted. It is thus strongly recommended for
2210 all applications to use the libc NSS API or native systemd-resolved
2211 bus API instead.
2212
2213 * systemd-networkd's bridge support learned a new setting
2214 VLANFiltering= for controlling VLAN filtering. Moreover a new section
2215 in .network files has been added for configuring VLAN bridging in
2216 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
2217
2218 * systemd-networkd's IPv6 Router Advertisement code now makes use of
2219 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
2220 now be acquired without relying on DHCPv6. Two new options
2221 UseDomains= and UseDNS= have been added to configure this behaviour.
2222
2223 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
2224 renamed IPv6AcceptRA=, without altering its behaviour. The old
2225 setting name remains available for compatibility reasons.
2226
2227 * The systemd-networkd VTI/VTI6 tunneling support gained new options
2228 Key=, InputKey= and OutputKey=.
2229
2230 * systemd-networkd gained support for VRF ("Virtual Routing Function")
2231 interface configuration.
2232
2233 * "systemctl edit" may now be used to create new unit files by
2234 specifying the --force switch.
2235
2236 * sd-event gained a new function sd_event_get_iteration() for
2237 requesting the current iteration counter of the event loop. It starts
2238 at zero and is increased by one with each event loop iteration.
2239
43a569a1
ZJS
2240 * A new rpm macro %systemd_ordering is provided by the macros.systemd
2241 file. It can be used in lieu of %systemd_requires in packages which
2242 don't use any systemd functionality and are intended to be installed
2243 in minimal containers without systemd present. This macro provides
ce830873 2244 ordering dependencies to ensure that if the package is installed in
43a569a1
ZJS
2245 the same rpm transaction as systemd, systemd will be installed before
2246 the scriptlets for the package are executed, allowing unit presets
2247 to be handled.
2248
2249 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
2250 been added to simplify packaging of generators.
2251
2252 * The os-release file gained VERSION_CODENAME field for the
2253 distribution nickname (e.g. VERSION_CODENAME=woody).
2254
2255 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
2256 can be set to disable parsing of metadata and the creation
2257 of persistent symlinks for that device.
2258
0f1da52b
LP
2259 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
2260 to make them available to logged-in users has been reverted.
2261
2262 * Much of the common code of the various systemd components is now
2263 built into an internal shared library libsystemd-shared-231.so
2264 (incorporating the systemd version number in the name, to be updated
2265 with future releases) that the components link to. This should
2266 decrease systemd footprint both in memory during runtime and on
2267 disk. Note that the shared library is not for public use, and is
ead6bd25 2268 neither API nor ABI stable, but is likely to change with every new
1ecbf32f
ZJS
2269 released update. Packagers need to make sure that binaries
2270 linking to libsystemd-shared.so are updated in step with the
2271 library.
43a569a1 2272
fcd30826
LP
2273 * Configuration for "mkosi" is now part of the systemd
2274 repository. mkosi is a tool to easily build legacy-free OS images,
2275 and is available on github: https://github.com/systemd/mkosi. If
2276 "mkosi" is invoked in the build tree a new raw OS image is generated
2277 incorporating the systemd sources currently being worked on and a
2278 clean, fresh distribution installation. The generated OS image may be
ce830873 2279 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
fcd30826
LP
2280 UEFI PC. This functionality is particularly useful to easily test
2281 local changes made to systemd in a pristine, defined environment. See
f09eb768 2282 doc/HACKING for details.
ceeddf79 2283
4ffe2479
ZJS
2284 * configure learned the --with-support-url= option to specify the
2285 distribution's bugtracker.
2286
38b383d9
LP
2287 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
2288 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
2289 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
2290 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
2291 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
2292 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
2293 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
2294 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
2295 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
2296 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
2297 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
2298 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
2299 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
2300 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
2301 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
2302 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
771de3f5
ZJS
2303 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
2304 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
38b383d9 2305 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
5cd118ba 2306
38b383d9 2307 — Berlin, 2016-07-25
5cd118ba 2308
46e40fab 2309CHANGES WITH 230:
7f6e8043 2310
61ecb465
LP
2311 * DNSSEC is now turned on by default in systemd-resolved (in
2312 "allow-downgrade" mode), but may be turned off during compile time by
2313 passing "--with-default-dnssec=no" to "configure" (and of course,
2314 during runtime with DNSSEC= in resolved.conf). We recommend
2315 downstreams to leave this on at least during development cycles and
2316 report any issues with the DNSSEC logic upstream. We are very
2317 interested in collecting feedback about the DNSSEC validator and its
2318 limitations in the wild. Note however, that DNSSEC support is
2319 probably nothing downstreams should turn on in stable distros just
96d49011 2320 yet, as it might create incompatibilities with a few DNS servers and
61ecb465
LP
2321 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
2322 automatically whenever we detect such incompatible setups, but there
2323 might be systems we do not cover yet. Hence: please help us testing
2324 the DNSSEC code, leave this on where you can, report back, but then
2325 again don't consider turning this on in your stable, LTS or
e40a326c
LP
2326 production release just yet. (Note that you have to enable
2327 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
2328 and its DNSSEC mode for host name resolution from local
2329 applications.)
61ecb465 2330
96515dbf 2331 * systemd-resolve conveniently resolves DANE records with the --tlsa
e40a326c 2332 option and OPENPGPKEY records with the --openpgp option. It also
e75690c3 2333 supports dumping raw DNS record data via the new --raw= switch.
96515dbf 2334
97e5530c
ZJS
2335 * systemd-logind will now by default terminate user processes that are
2336 part of the user session scope unit (session-XX.scope) when the user
977f2bea 2337 logs out. This behavior is controlled by the KillUserProcesses=
e40a326c
LP
2338 setting in logind.conf, and the previous default of "no" is now
2339 changed to "yes". This means that user sessions will be properly
2340 cleaned up after, but additional steps are necessary to allow
2341 intentionally long-running processes to survive logout.
97e5530c
ZJS
2342
2343 While the user is logged in at least once, user@.service is running,
2344 and any service that should survive the end of any individual login
2345 session can be started at a user service or scope using systemd-run.
e40a326c 2346 systemd-run(1) man page has been extended with an example which shows
8951eaec 2347 how to run screen in a scope unit underneath user@.service. The same
e40a326c 2348 command works for tmux.
97e5530c
ZJS
2349
2350 After the user logs out of all sessions, user@.service will be
2351 terminated too, by default, unless the user has "lingering" enabled.
2352 To effectively allow users to run long-term tasks even if they are
152199f2
ZJS
2353 logged out, lingering must be enabled for them. See loginctl(1) for
2354 details. The default polkit policy was modified to allow users to
2355 set lingering for themselves without authentication.
7f6e8043 2356
95365a57 2357 Previous defaults can be restored at compile time by the
e40a326c 2358 --without-kill-user-processes option to "configure".
7f6e8043 2359
e75690c3
ZJS
2360 * systemd-logind gained new configuration settings SessionsMax= and
2361 InhibitorsMax=, both with a default of 8192. It will not register new
188d3082 2362 user sessions or inhibitors above this limit.
e75690c3
ZJS
2363
2364 * systemd-logind will now reload configuration on SIGHUP.
2365
96515dbf 2366 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
e40a326c 2367 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
25b0e6cb
LP
2368 enable. Also, support for the "io" cgroup controller in the unified
2369 hierarchy has been added, so that the "memory", "pids" and "io" are
2370 now the controllers that are supported on the unified hierarchy.
e40a326c 2371
96515dbf
ZJS
2372 WARNING: it is not possible to use previous systemd versions with
2373 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
2374 is necessary to also update systemd in the initramfs if using the
e40a326c 2375 unified hierarchy. An updated SELinux policy is also required.
96515dbf 2376
e40a326c
LP
2377 * LLDP support has been extended, and both passive (receive-only) and
2378 active (sender) modes are supported. Passive mode ("routers-only") is
8951eaec
ZJS
2379 enabled by default in systemd-networkd. Active LLDP mode is enabled
2380 by default for containers on the internal network. The "networkctl
e40a326c
LP
2381 lldp" command may be used to list information gathered. "networkctl
2382 status" will also show basic LLDP information on connected peers now.
96515dbf 2383
e40a326c
LP
2384 * The IAID and DUID unique identifier sent in DHCP requests may now be
2385 configured for the system and each .network file managed by
e75690c3
ZJS
2386 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
2387
2388 * systemd-networkd gained support for configuring proxy ARP support for
2389 each interface, via the ProxyArp= setting in .network files. It also
2390 gained support for configuring the multicast querier feature of
2391 bridge devices, via the new MulticastQuerier= setting in .netdev
2392 files. Similarly, snooping on the IGMP traffic can be controlled
2393 via the new setting MulticastSnooping=.
2394
2395 A new setting PreferredLifetime= has been added for addresses
2396 configured in .network file to configure the lifetime intended for an
2397 address.
2398
2399 The systemd-networkd DHCP server gained the option EmitRouter=, which
2400 defaults to yes, to configure whether the DHCP Option 3 (Router)
2401 should be emitted.
96515dbf 2402
e40a326c 2403 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
97e5530c
ZJS
2404 systemd-socket-activate and installed into /usr/bin. It is now fully
2405 supported.
2406
e40a326c
LP
2407 * systemd-journald now uses separate threads to flush changes to disk
2408 when closing journal files, thus reducing impact of slow disk I/O on
2409 logging performance.
96515dbf 2410
e75690c3
ZJS
2411 * The sd-journal API gained two new calls
2412 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
2413 can be used to open journal files using file descriptors instead of
2414 file or directory paths. sd_journal_open_container() has been
2415 deprecated, sd_journal_open_directory_fd() should be used instead
2416 with the flag SD_JOURNAL_OS_ROOT.
2417
2418 * journalctl learned a new output mode "-o short-unix" that outputs log
2419 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
2420 UTC). It also gained support for a new --no-hostname setting to
2421 suppress the hostname column in the family of "short" output modes.
2422
e40a326c
LP
2423 * systemd-ask-password now optionally skips printing of the password to
2424 stdout with --no-output which can be useful in scripts.
96515dbf
ZJS
2425
2426 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
2427 (devices tagged with ID_MAKER_TOOL) are now tagged with
2428 "uaccess" and are available to logged in users.
2429
e75690c3 2430 * The DeviceAllow= unit setting now supports specifiers (with "%").
e40a326c
LP
2431
2432 * "systemctl show" gained a new --value switch, which allows print a
2433 only the contents of a specific unit property, without also printing
8951eaec
ZJS
2434 the property's name. Similar support was added to "show*" verbs
2435 of loginctl and machinectl that output "key=value" lists.
e40a326c 2436
e75690c3
ZJS
2437 * A new unit type "generated" was added for files dynamically generated
2438 by generator tools. Similarly, a new unit type "transient" is used
2439 for unit files created using the runtime API. "systemctl enable" will
2440 refuse to operate on such files.
2441
e40a326c
LP
2442 * A new command "systemctl revert" has been added that may be used to
2443 revert to the vendor version of a unit file, in case local changes
2444 have been made by adding drop-ins or overriding the unit file.
2445
2446 * "machinectl clean" gained a new verb to automatically remove all or
2447 just hidden container images.
2448
e40a326c
LP
2449 * systemd-tmpfiles gained support for a new line type "e" for emptying
2450 directories, if they exist, without creating them if they don't.
2451
e40a326c
LP
2452 * systemd-nspawn gained support for automatically patching the UID/GIDs
2453 of the owners and the ACLs of all files and directories in a
2454 container tree to match the UID/GID user namespacing range selected
2455 for the container invocation. This mode is enabled via the new
24597ee0
ZJS
2456 --private-users-chown switch. It also gained support for
2457 automatically choosing a free, previously unused UID/GID range when
2458 starting a container, via the new --private-users=pick setting (which
2459 implies --private-users-chown). Together, these options for the first
2460 time make user namespacing for nspawn containers fully automatic and
2461 thus deployable. The systemd-nspawn@.service template unit file has
2462 been changed to use this functionality by default.
e40a326c 2463
25b0e6cb
LP
2464 * systemd-nspawn gained a new --network-zone= switch, that allows
2465 creating ad-hoc virtual Ethernet links between multiple containers,
2466 that only exist as long as at least one container referencing them is
2467 running. This allows easy connecting of multiple containers with a
2468 common link that implements an Ethernet broadcast domain. Each of
2469 these network "zones" may be named relatively freely by the user, and
2470 may be referenced by any number of containers, but each container may
2471 only reference one of these "zones". On the lower level, this is
2472 implemented by an automatically managed bridge network interface for
2473 each zone, that is created when the first container referencing its
2474 zone is created and removed when the last one referencing its zone
2475 terminates.
2476
e40a326c 2477 * The default start timeout may now be configured on the kernel command
8951eaec
ZJS
2478 line via systemd.default_timeout_start_sec=. It was already
2479 configurable via the DefaultTimeoutStartSec= option in
2480 /etc/systemd/system.conf.
e40a326c 2481
030bd839 2482 * Socket units gained a new TriggerLimitIntervalSec= and
e40a326c
LP
2483 TriggerLimitBurst= setting to configure a limit on the activation
2484 rate of the socket unit.
2485
2486 * The LimitNICE= setting now optionally takes normal UNIX nice values
2487 in addition to the raw integer limit value. If the specified
2488 parameter is prefixed with "+" or "-" and is in the range -20..19 the
2489 value is understood as UNIX nice value. If not prefixed like this it
2490 is understood as raw RLIMIT_NICE limit.
2491
999a43f8
LP
2492 * Note that the effect of the PrivateDevices= unit file setting changed
2493 slightly with this release: the per-device /dev file system will be
2494 mounted read-only from this version on, and will have "noexec"
188d3082 2495 set. This (minor) change of behavior might cause some (exceptional)
999a43f8
LP
2496 legacy software to break, when PrivateDevices=yes is set for its
2497 service. Please leave PrivateDevices= off if you run into problems
2498 with this.
2499
e75690c3
ZJS
2500 * systemd-bootchart has been split out to a separate repository:
2501 https://github.com/systemd/systemd-bootchart
2502
2503 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
2504 merged into the kernel in its current form.
2505
2506 * The compatibility libraries libsystemd-daemon.so,
2507 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
2508 which have been deprecated since systemd-209 have been removed along
2509 with the corresponding pkg-config files. All symbols provided by
2510 those libraries are provided by libsystemd.so.
2511
2512 * The Capabilities= unit file setting has been removed (it is ignored
2513 for backwards compatibility). AmbientCapabilities= and
2514 CapabilityBoundingSet= should be used instead.
2515
4f9020fa
DR
2516 * A new special target has been added, initrd-root-device.target,
2517 which creates a synchronization point for dependencies of the root
2518 device in early userspace. Initramfs builders must ensure that this
2519 target is now included in early userspace.
2520
e75690c3
ZJS
2521 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
2522 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
2523 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
2524 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
2525 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
2526 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
2527 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
2528 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
77ff6022
CG
2529 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
2530 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
e75690c3
ZJS
2531 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
2532 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
2533 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
77ff6022
CG
2534 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
2535 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
2536 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
e75690c3
ZJS
2537 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
2538 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
2539 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
2540 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
2541 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
2542 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
46e40fab
ZJS
2543 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
2544 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
2545 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
2546 Jędrzejewski-Szmek
e40a326c 2547
46e40fab 2548 — Fairfax, 2016-05-21
96515dbf 2549
61f32bff
MP
2550CHANGES WITH 229:
2551
d5f8b295
LP
2552 * The systemd-resolved DNS resolver service has gained a substantial
2553 set of new features, most prominently it may now act as a DNSSEC
2554 validating stub resolver. DNSSEC mode is currently turned off by
ed5f8840
ZJS
2555 default, but is expected to be turned on by default in one of the
2556 next releases. For now, we invite everybody to test the DNSSEC logic
2557 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
2558 service also gained a full set of D-Bus interfaces, including calls
2559 to configure DNS and DNSSEC settings per link (for use by external
2560 network management software). systemd-resolved and systemd-networkd
2561 now distinguish between "search" and "routing" domains. The former
2562 are used to qualify single-label names, the latter are used purely
2563 for routing lookups within certain domains to specific links.
2564 resolved now also synthesizes RRs for all entries from /etc/hosts.
d5f8b295
LP
2565
2566 * The systemd-resolve tool (which is a client utility for
ed5f8840
ZJS
2567 systemd-resolved) has been improved considerably and is now fully
2568 supported and documented. Hence it has moved from /usr/lib/systemd to
2569 /usr/bin.
d5f8b295
LP
2570
2571 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
2572 devices.
2573
a7c723c0
LP
2574 * The coredump collection logic has been reworked: when a coredump is
2575 collected it is now written to disk, compressed and processed
2576 (including stacktrace extraction) from a new instantiated service
2577 systemd-coredump@.service, instead of directly from the
2578 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
2579 processing large coredumps can take up a substantial amount of
2580 resources and time, and this previously happened entirely outside of
2581 systemd's service supervision. With the new logic the core_pattern
2582 hook only does minimal metadata collection before passing off control
2583 to the new instantiated service, which is configured with a time
2584 limit, a nice level and other settings to minimize negative impact on
2585 the rest of the system. Also note that the new logic will honour the
2586 RLIMIT_CORE setting of the crashed process, which now allows users
2587 and processes to turn off coredumping for their processes by setting
2588 this limit.
2589
2590 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
2591 and all forked processes by default. Previously, PID 1 would leave
2592 the setting at "0" for all processes, as set by the kernel. Note that
2593 the resource limit traditionally has no effect on the generated
2594 coredumps on the system if the /proc/sys/kernel/core_pattern hook
2595 logic is used. Since the limit is now honoured (see above) its
2596 default has been changed so that the coredumping logic is enabled by
2597 default for all processes, while allowing specific opt-out.
2598
2599 * When the stacktrace is extracted from processes of system users, this
2600 is now done as "systemd-coredump" user, in order to sandbox this
2601 potentially security sensitive parsing operation. (Note that when
2602 processing coredumps of normal users this is done under the user ID
2603 of process that crashed, as before.) Packagers should take notice
2604 that it is now necessary to create the "systemd-coredump" system user
2605 and group at package installation time.
2606
d5f8b295
LP
2607 * The systemd-activate socket activation testing tool gained support
2608 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
2609 and --seqpacket switches. It also has been extended to support both
2610 new-style and inetd-style file descriptor passing. Use the new
2611 --inetd switch to request inetd-style file descriptor passing.
2612
8968aea0
MP
2613 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
2614 variable, which takes a boolean value. If set to false, ANSI color
d5f8b295
LP
2615 output is disabled in the tools even when run on a terminal that
2616 supports it.
2617
2618 * The VXLAN support in networkd now supports two new settings
2619 DestinationPort= and PortRange=.
2620
2621 * A new systemd.machine_id= kernel command line switch has been added,
2622 that may be used to set the machine ID in /etc/machine-id if it is
2623 not initialized yet. This command line option has no effect if the
2624 file is already initialized.
2625
2626 * systemd-nspawn gained a new --as-pid2 switch that invokes any
2627 specified command line as PID 2 rather than PID 1 in the
ed5f8840
ZJS
2628 container. In this mode PID 1 is a minimal stub init process that
2629 implements the special POSIX and Linux semantics of PID 1 regarding
2630 signal and child process management. Note that this stub init process
2631 is implemented in nspawn itself and requires no support from the
2632 container image. This new logic is useful to support running
2633 arbitrary commands in the container, as normal processes are
d5f8b295
LP
2634 generally not prepared to run as PID 1.
2635
2636 * systemd-nspawn gained a new --chdir= switch for setting the current
2637 working directory for the process started in the container.
2638
ed5f8840
ZJS
2639 * "journalctl /dev/sda" will now output all kernel log messages for
2640 specified device from the current boot, in addition to all devices
2641 that are parents of it. This should make log output about devices
2642 pretty useful, as long as kernel drivers attach enough metadata to
2643 the log messages. (The usual SATA drivers do.)
d5f8b295
LP
2644
2645 * The sd-journal API gained two new calls
2646 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
2647 that report whether log data from /run or /var has been found.
2648
2649 * journalctl gained a new switch "--fields" that prints all journal
2650 record field names currently in use in the journal. This is backed
2651 by two new sd-journal API calls sd_journal_enumerate_fields() and
2652 sd_journal_restart_fields().
2653
2654 * Most configurable timeouts in systemd now expect an argument of
8968aea0
MP
2655 "infinity" to turn them off, instead of "0" as before. The semantics
2656 from now on is that a timeout of "0" means "now", and "infinity"
2657 means "never". To maintain backwards compatibility, "0" continues to
2658 turn off previously existing timeout settings.
d5f8b295
LP
2659
2660 * "systemctl reload-or-try-restart" has been renamed to "systemctl
8968aea0
MP
2661 try-reload-or-restart" to clarify what it actually does: the "try"
2662 logic applies to both reloading and restarting, not just restarting.
2663 The old name continues to be accepted for compatibility.
2664
2665 * On boot-up, when PID 1 detects that the system clock is behind the
2666 release date of the systemd version in use, the clock is now set
2667 to the latter. Previously, this was already done in timesyncd, in order
d5f8b295
LP
2668 to avoid running with clocks set to the various clock epochs such as
2669 1902, 1938 or 1970. With this change the logic is now done in PID 1
2670 in addition to timesyncd during early boot-up, so that it is enforced
2671 before the first process is spawned by systemd. Note that the logic
2672 in timesyncd remains, as it is more comprehensive and ensures
ed5f8840 2673 clock monotonicity by maintaining a persistent timestamp file in
d5f8b295
LP
2674 /var. Since /var is generally not available in earliest boot or the
2675 initrd, this part of the logic remains in timesyncd, and is not done
2676 by PID 1.
2677
50f48ad3
DM
2678 * Support for tweaking details in net_cls.class_id through the
2679 NetClass= configuration directive has been removed, as the kernel
2680 people have decided to deprecate that controller in cgroup v2.
2681 Userspace tools such as nftables are moving over to setting rules
2682 that are specific to the full cgroup path of a task, which obsoletes
2683 these controllers anyway. The NetClass= directive is kept around for
2684 legacy compatibility reasons. For a more in-depth description of the
2685 kernel change, please refer to the respective upstream commit:
2686
2687 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
2688
d5f8b295 2689 * A new service setting RuntimeMaxSec= has been added that may be used
8968aea0 2690 to specify a maximum runtime for a service. If the timeout is hit, the
d5f8b295
LP
2691 service is terminated and put into a failure state.
2692
8968aea0
MP
2693 * A new service setting AmbientCapabilities= has been added. It allows
2694 configuration of additional Linux process capabilities that are
2695 passed to the activated processes. This is only available on very
d5f8b295
LP
2696 recent kernels.
2697
2698 * The process resource limit settings in service units may now be used
2699 to configure hard and soft limits individually.
2700
8968aea0 2701 * The various libsystemd APIs such as sd-bus or sd-event now publicly
ed5f8840
ZJS
2702 expose support for gcc's __attribute__((cleanup())) C extension.
2703 Specifically, for many object destructor functions alternative
2704 versions have been added that have names suffixed with "p" and take a
2705 pointer to a pointer to the object to destroy, instead of just a
2706 pointer to the object itself. This is useful because these destructor
2707 functions may be used directly as parameters to the cleanup
2708 construct. Internally, systemd has been a heavy user of this GCC
2709 extension for a long time, and with this change similar support is
2710 now available to consumers of the library outside of systemd. Note
8968aea0 2711 that by using this extension in your sources compatibility with old
ed5f8840
ZJS
2712 and strictly ANSI compatible C compilers is lost. However, all gcc or
2713 LLVM versions of recent years support this extension.
d5f8b295
LP
2714
2715 * Timer units gained support for a new setting RandomizedDelaySec= that
8968aea0
MP
2716 allows configuring some additional randomized delay to the configured
2717 time. This is useful to spread out timer events to avoid load peaks in
2718 clusters or larger setups.
d5f8b295
LP
2719
2720 * Calendar time specifications now support sub-second accuracy.
2721
2722 * Socket units now support listening on SCTP and UDP-lite protocol
2723 sockets.
2724
2725 * The sd-event API now comes with a full set of man pages.
2726
2727 * Older versions of systemd contained experimental support for
2728 compressing journal files and coredumps with the LZ4 compressor that
2729 was not compatible with the lz4 binary (due to API limitations of the
2730 lz4 library). This support has been removed; only support for files
2731 compatible with the lz4 binary remains. This LZ4 logic is now
2732 officially supported and no longer considered experimental.
2733
2734 * The dkr image import logic has been removed again from importd. dkr's
2735 micro-services focus doesn't fit into the machine image focus of
2736 importd, and quickly got out of date with the upstream dkr API.
2737
2738 * Creation of the /run/lock/lockdev/ directory was dropped from
2739 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
61f32bff
MP
2740 been available for many years. If you still need this, you need to
2741 create your own tmpfiles.d config file with:
d5f8b295
LP
2742
2743 d /run/lock/lockdev 0775 root lock -
61f32bff 2744
dd95b381
LP
2745 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
2746 and RebootArgument= have been moved from the [Service] section of
2747 unit files to [Unit], and they are now supported on all unit types,
2748 not just service units. Of course, systemd will continue to
2749 understand these settings also at the old location, in order to
2750 maintain compatibility.
2751
3545ab35
LP
2752 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
2753 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
2754 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
2755 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
2756 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
2757 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
2758 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
2759 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
2760 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
2761 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
2762 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
2763 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
2764 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
2765 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
2766 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
2767 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
2768 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
2769 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
2770 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
2771
ccddd104 2772 — Berlin, 2016-02-11
61f32bff 2773
a11c7ea5
LP
2774CHANGES WITH 228:
2775
a11c7ea5
LP
2776 * A number of properties previously only settable in unit
2777 files are now also available as properties to set when
2778 creating transient units programmatically via the bus, as it
2779 is exposed with systemd-run's --property=
2780 setting. Specifically, these are: SyslogIdentifier=,
2781 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
2782 EnvironmentFile=, ReadWriteDirectories=,
2783 ReadOnlyDirectories=, InaccessibleDirectories=,
2784 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
2785
28c85daf
LP
2786 * When creating transient services via the bus API it is now
2787 possible to pass in a set of file descriptors to use as
2788 STDIN/STDOUT/STDERR for the invoked process.
a11c7ea5 2789
f1f8a5a5
LP
2790 * Slice units may now be created transiently via the bus APIs,
2791 similar to the way service and scope units may already be
2792 created transiently.
2793
a11c7ea5
LP
2794 * Wherever systemd expects a calendar timestamp specification
2795 (like in journalctl's --since= and --until= switches) UTC
2796 timestamps are now supported. Timestamps suffixed with "UTC"
2797 are now considered to be in Universal Time Coordinated
2798 instead of the local timezone. Also, timestamps may now
815bb5bd 2799 optionally be specified with sub-second accuracy. Both of
a11c7ea5
LP
2800 these additions also apply to recurring calendar event
2801 specification, such as OnCalendar= in timer units.
2802
28c85daf
LP
2803 * journalctl gained a new "--sync" switch that asks the
2804 journal daemon to write all so far unwritten log messages to
2805 disk and sync the files, before returning.
2806
a11c7ea5
LP
2807 * systemd-tmpfiles learned two new line types "q" and "Q" that
2808 operate like "v", but also set up a basic btrfs quota
2809 hierarchy when used on a btrfs file system with quota
2810 enabled.
2811
f1f8a5a5
LP
2812 * tmpfiles' "v", "q" and "Q" will now create a plain directory
2813 instead of a subvolume (even on a btrfs file system) if the
2814 root directory is a plain directory, and not a
2815 subvolume. This should simplify things with certain chroot()
2816 environments which are not aware of the concept of btrfs
2817 subvolumes.
2818
a11c7ea5
LP
2819 * systemd-detect-virt gained a new --chroot switch to detect
2820 whether execution takes place in a chroot() environment.
2821
28c85daf 2822 * CPUAffinity= now takes CPU index ranges in addition to
a11c7ea5
LP
2823 individual indexes.
2824
28c85daf
LP
2825 * The various memory-related resource limit settings (such as
2826 LimitAS=) now understand the usual K, M, G, ... suffixes to
2827 the base of 1024 (IEC). Similar, the time-related resource
2828 limit settings understand the usual min, h, day, ...
2829 suffixes now.
2830
f1f8a5a5
LP
2831 * There's a new system.conf setting DefaultTasksMax= to
2832 control the default TasksMax= setting for services and
2833 scopes running on the system. (TasksMax= is the primary
2834 setting that exposes the "pids" cgroup controller on systemd
2835 and was introduced in the previous systemd release.) The
2836 setting now defaults to 512, which means services that are
2837 not explicitly configured otherwise will only be able to
2838 create 512 processes or threads at maximum, from this
2839 version on. Note that this means that thread- or
2840 process-heavy services might need to be reconfigured to set
2841 TasksMax= to a higher value. It is sufficient to set
2842 TasksMax= in these specific unit files to a higher value, or
2843 even "infinity". Similar, there's now a logind.conf setting
2844 UserTasksMax= that defaults to 4096 and limits the total
2845 number of processes or tasks each user may own
2846 concurrently. nspawn containers also have the TasksMax=
2847 value set by default now, to 8192. Note that all of this
2848 only has an effect if the "pids" cgroup controller is
2849 enabled in the kernel. The general benefit of these changes
2850 should be a more robust and safer system, that provides a
2851 certain amount of per-service fork() bomb protection.
2852
28c85daf
LP
2853 * systemd-nspawn gained the new --network-veth-extra= switch
2854 to define additional and arbitrarily-named virtual Ethernet
2855 links between the host and the container.
2856
2857 * A new service execution setting PassEnvironment= has been
2858 added that allows importing select environment variables
2859 from PID1's environment block into the environment block of
2860 the service.
2861
ddb4b0d3 2862 * Timer units gained support for a new RemainAfterElapse=
595bfe7d 2863 setting which takes a boolean argument. It defaults to on,
ddb4b0d3
LP
2864 exposing behaviour unchanged to previous releases. If set to
2865 off, timer units are unloaded after they elapsed if they
2866 cannot elapse again. This is particularly useful for
2867 transient timer units, which shall not stay around longer
2868 than until they first elapse.
2869
a11c7ea5 2870 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
28c85daf
LP
2871 default now (the kernel default is 16). This is beneficial
2872 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
a11c7ea5
LP
2873 allows substantially larger numbers of queued
2874 datagrams. This should increase the capability of systemd to
2875 parallelize boot-up, as logging and sd_notify() are unlikely
2876 to stall execution anymore. If you need to change the value
2877 from the new defaults, use the usual sysctl.d/ snippets.
2878
28c85daf
LP
2879 * The compression framing format used by the journal or
2880 coredump processing has changed to be in line with what the
2881 official LZ4 tools generate. LZ4 compression support in
2882 systemd was considered unsupported previously, as the format
2883 was not compatible with the normal tools. With this release
2884 this has changed now, and it is hence safe for downstream
2885 distributions to turn it on. While not compressing as well
815bb5bd 2886 as the XZ, LZ4 is substantially faster, which makes
28c85daf
LP
2887 it a good default choice for the compression logic in the
2888 journal and in coredump handling.
a11c7ea5 2889
28c85daf
LP
2890 * Any reference to /etc/mtab has been dropped from
2891 systemd. The file has been obsolete since a while, but
2892 systemd refused to work on systems where it was incorrectly
815bb5bd 2893 set up (it should be a symlink or non-existent). Please make
28c85daf
LP
2894 sure to update to util-linux 2.27.1 or newer in conjunction
2895 with this systemd release, which also drops any reference to
2896 /etc/mtab. If you maintain a distribution make sure that no
2897 software you package still references it, as this is a
2898 likely source of bugs. There's also a glibc bug pending,
2899 asking for removal of any reference to this obsolete file:
2900
2901 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
a11c7ea5 2902
d5bd92bb
LP
2903 Note that only util-linux versions built with
2904 --enable-libmount-force-mountinfo are supported.
2905
a11c7ea5
LP
2906 * Support for the ".snapshot" unit type has been removed. This
2907 feature turned out to be little useful and little used, and
2908 has now been removed from the core and from systemctl.
2909
b9e2f7eb
LP
2910 * The dependency types RequiresOverridable= and
2911 RequisiteOverridable= have been removed from systemd. They
2912 have been used only very sparingly to our knowledge and
2913 other options that provide a similar effect (such as
2914 systemctl --mode=ignore-dependencies) are much more useful
2915 and commonly used. Moreover, they were only half-way
2916 implemented as the option to control behaviour regarding
2917 these dependencies was never added to systemctl. By removing
2918 these dependency types the execution engine becomes a bit
2919 simpler. Unit files that use these dependencies should be
2920 changed to use the non-Overridable dependency types
2921 instead. In fact, when parsing unit files with these
2922 options, that's what systemd will automatically convert them
2923 too, but it will also warn, asking users to fix the unit
2924 files accordingly. Removal of these dependency types should
2925 only affect a negligible number of unit files in the wild.
2926
2927 * Behaviour of networkd's IPForward= option changed
2928 (again). It will no longer maintain a per-interface setting,
2929 but propagate one way from interfaces where this is enabled
2930 to the global kernel setting. The global setting will be
2931 enabled when requested by a network that is set up, but
2932 never be disabled again. This change was made to make sure
2933 IPv4 and IPv6 behaviour regarding packet forwarding is
2934 similar (as the Linux IPv6 stack does not support
2935 per-interface control of this setting) and to minimize
2936 surprises.
2937
28c85daf
LP
2938 * In unit files the behaviour of %u, %U, %h, %s has
2939 changed. These specifiers will now unconditionally resolve
2940 to the various user database fields of the user that the
2941 systemd instance is running as, instead of the user
2942 configured in the specific unit via User=. Note that this
2943 effectively doesn't change much, as resolving of these
2944 specifiers was already turned off in the --system instance
2945 of systemd, as we cannot do NSS lookups from PID 1. In the
2946 --user instance of systemd these specifiers where correctly
2947 resolved, but hardly made any sense, since the user instance
2948 lacks privileges to do user switches anyway, and User= is
ce830873 2949 hence useless. Moreover, even in the --user instance of
28c85daf
LP
2950 systemd behaviour was awkward as it would only take settings
2951 from User= assignment placed before the specifier into
2952 account. In order to unify and simplify the logic around
2953 this the specifiers will now always resolve to the
2954 credentials of the user invoking the manager (which in case
2955 of PID 1 is the root user).
2956
2957 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
2958 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
2959 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
f1f8a5a5
LP
2960 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
2961 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
2962 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
2963 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
2964 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
2965 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
2966 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
2967 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
2968 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
2969 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
2970 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
2971 Jędrzejewski-Szmek
28c85daf 2972
ccddd104 2973 — Berlin, 2015-11-18
a11c7ea5 2974
c97e586d
DM
2975CHANGES WITH 227:
2976
2977 * systemd now depends on util-linux v2.27. More specifically,
2978 the newly added mount monitor feature in libmount now
2979 replaces systemd's former own implementation.
2980
2981 * libmount mandates /etc/mtab not to be regular file, and
2982 systemd now enforces this condition at early boot.
2983 /etc/mtab has been deprecated and warned about for a very
2984 long time, so systems running systemd should already have
2985 stopped having this file around as anything else than a
2986 symlink to /proc/self/mounts.
2987
d046fb93
LP
2988 * Support for the "pids" cgroup controller has been added. It
2989 allows accounting the number of tasks in a cgroup and
c97e586d
DM
2990 enforcing limits on it. This adds two new setting
2991 TasksAccounting= and TasksMax= to each unit, as well as a
6fd5517b 2992 global option DefaultTasksAccounting=.
c97e586d
DM
2993
2994 * Support for the "net_cls" cgroup controller has been added.
fe08a30b
LP
2995 It allows assigning a net class ID to each task in the
2996 cgroup, which can then be used in firewall rules and traffic
2997 shaping configurations. Note that the kernel netfilter net
2998 class code does not currently work reliably for ingress
2999 packets on unestablished sockets.
c97e586d
DM
3000
3001 This adds a new config directive called NetClass= to CGroup
6fd5517b 3002 enabled units. Allowed values are positive numbers for fixed
c97e586d
DM
3003 assignments and "auto" for picking a free value
3004 automatically.
3005
21d86c61
DM
3006 * 'systemctl is-system-running' now returns 'offline' if the
3007 system is not booted with systemd. This command can now be
3008 used as a substitute for 'systemd-notify --booted'.
3009
3010 * Watchdog timeouts have been increased to 3 minutes for all
3011 in-tree service files. Apparently, disk IO issues are more
3012 frequent than we hoped, and user reported >1 minute waiting
3013 for disk IO.
3014
3015 * 'machine-id-commit' functionality has been merged into
3016 'machine-id-setup --commit'. The separate binary has been
3017 removed.
3018
d046fb93
LP
3019 * The WorkingDirectory= directive in unit files may now be set
3020 to the special value '~'. In this case, the working
3021 directory is set to the home directory of the user
3022 configured in User=.
21d86c61 3023
fe08a30b
LP
3024 * "machinectl shell" will now open the shell in the home
3025 directory of the selected user by default.
3026
21d86c61 3027 * The CrashChVT= configuration file setting is renamed to
d046fb93
LP
3028 CrashChangeVT=, following our usual logic of not
3029 abbreviating unnecessarily. The old directive is still
3030 supported for compat reasons. Also, this directive now takes
3031 an integer value between 1 and 63, or a boolean value. The
3032 formerly supported '-1' value for disabling stays around for
3033 compat reasons.
21d86c61 3034
fe08a30b 3035 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
8b5f9d15 3036 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
fe08a30b
LP
3037 RootDirectory= properties can now be set for transient
3038 units.
3039
3040 * The systemd-analyze tool gained a new "set-log-target" verb
3041 to change the logging target the system manager logs to
3042 dynamically during runtime. This is similar to how
3043 "systemd-analyze set-log-level" already changes the log
3044 level.
3045
3046 * In nspawn /sys is now mounted as tmpfs, with only a selected
3047 set of subdirectories mounted in from the real sysfs. This
3048 enhances security slightly, and is useful for ensuring user
3049 namespaces work correctly.
3050
3051 * Support for USB FunctionFS activation has been added. This
3052 allows implementation of USB gadget services that are
3053 activated as soon as they are requested, so that they don't
595bfe7d 3054 have to run continuously, similar to classic socket
fe08a30b
LP
3055 activation.
3056
3057 * The "systemctl exit" command now optionally takes an
3058 additional parameter that sets the exit code to return from
3059 the systemd manager when exiting. This is only relevant when
3060 running the systemd user instance, or when running the
3061 system instance in a container.
3062
3063 * sd-bus gained the new API calls sd_bus_path_encode_many()
3064 and sd_bus_path_decode_many() that allow easy encoding and
3065 decoding of multiple identifier strings inside a D-Bus
3066 object path. Another new call sd_bus_default_flush_close()
3067 has been added to flush and close per-thread default
3068 connections.
3069
3070 * systemd-cgtop gained support for a -M/--machine= switch to
3071 show the control groups within a certain container only.
3072
3073 * "systemctl kill" gained support for an optional --fail
3074 switch. If specified the requested operation will fail of no
3075 processes have been killed, because the unit had no
3076 processes attached, or similar.
3077
bdba9227
DM
3078 * A new systemd.crash_reboot=1 kernel command line option has
3079 been added that triggers a reboot after crashing. This can
3080 also be set through CrashReboot= in systemd.conf.
3081
3082 * The RuntimeDirectory= setting now understands unit
3083 specifiers like %i or %f.
3084
ce830873 3085 * A new (still internal) library API sd-ipv4acd has been added,
fe08a30b
LP
3086 that implements address conflict detection for IPv4. It's
3087 based on code from sd-ipv4ll, and will be useful for
3088 detecting DHCP address conflicts.
3089
bdba9227
DM
3090 * File descriptors passed during socket activation may now be
3091 named. A new API sd_listen_fds_with_names() is added to
a8eaaee7 3092 access the names. The default names may be overridden,
bdba9227
DM
3093 either in the .socket file using the FileDescriptorName=
3094 parameter, or by passing FDNAME= when storing the file
3095 descriptors using sd_notify().
fe08a30b 3096
d046fb93
LP
3097 * systemd-networkd gained support for:
3098
0053598f 3099 - Setting the IPv6 Router Advertisement settings via
edf4126f 3100 IPv6AcceptRouterAdvertisements= in .network files.
d046fb93
LP
3101
3102 - Configuring the HelloTimeSec=, MaxAgeSec= and
3103 ForwardDelaySec= bridge parameters in .netdev files.
3104
3105 - Configuring PreferredSource= for static routes in
edf4126f 3106 .network files.
fe08a30b 3107
bdba9227
DM
3108 * The "ask-password" framework used to query for LUKS harddisk
3109 passwords or SSL passwords during boot gained support for
3110 caching passwords in the kernel keyring, if it is
3111 available. This makes sure that the user only has to type in
3112 a passphrase once if there are multiple objects to unlock
3113 with the same one. Previously, such password caching was
3114 available only when Plymouth was used; this moves the
3115 caching logic into the systemd codebase itself. The
3116 "systemd-ask-password" utility gained a new --keyname=
3117 switch to control which kernel keyring key to use for
3118 caching a password in. This functionality is also useful for
3119 enabling display managers such as gdm to automatically
3120 unlock the user's GNOME keyring if its passphrase, the
3121 user's password and the harddisk password are the same, if
3122 gdm-autologin is used.
fe08a30b
LP
3123
3124 * When downloading tar or raw images using "machinectl
3125 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
3126 file is now also downloaded, if it is available and stored
3127 next to the image file.
c97e586d 3128
91d0d699
LP
3129 * Units of type ".socket" gained a new boolean setting
3130 Writable= which is only useful in conjunction with
3131 ListenSpecial=. If true, enables opening the specified
3132 special file in O_RDWR mode rather than O_RDONLY mode.
3133
3134 * systemd-rfkill has been reworked to become a singleton
3135 service that is activated through /dev/rfkill on each rfkill
3136 state change and saves the settings to disk. This way,
3137 systemd-rfkill is now compatible with devices that exist
3138 only intermittendly, and even restores state if the previous
3139 system shutdown was abrupt rather than clean.
3140
d046fb93
LP
3141 * The journal daemon gained support for vacuuming old journal
3142 files controlled by the number of files that shall remain,
3143 in addition to the already existing control by size and by
3144 date. This is useful as journal interleaving performance
6dd6a9c4 3145 degrades with too many separate journal files, and allows
d046fb93
LP
3146 putting an effective limit on them. The new setting defaults
3147 to 100, but this may be changed by setting SystemMaxFiles=
3148 and RuntimeMaxFiles= in journald.conf. Also, the
3149 "journalctl" tool gained the new --vacuum-files= switch to
3150 manually vacuum journal files to leave only the specified
3151 number of files in place.
c48eb61f 3152
bdba9227
DM
3153 * udev will now create /dev/disk/by-path links for ATA devices
3154 on kernels where that is supported.
c30f086f 3155
efce0ffe 3156 * Galician, Serbian, Turkish and Korean translations were added.
c97e586d 3157
61e6771c
LP
3158 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
3159 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
3160 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
3161 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
3162 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
3163 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
3164 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
3165 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
3166 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
3167 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
3168 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
3169 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
3170 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
3171 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
3172 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
3173 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
3174 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
3175 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
3176
ccddd104 3177 — Berlin, 2015-10-07
c97e586d 3178
c9912c5e
DH
3179CHANGES WITH 226:
3180
5e8d4254
LP
3181 * The DHCP implementation of systemd-networkd gained a set of
3182 new features:
3183
3184 - The DHCP server now supports emitting DNS and NTP
3185 information. It may be enabled and configured via
3186 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
3187 and NTP information is enabled, but no servers are
3188 configured, the corresponding uplink information (if there
3189 is any) is propagated.
3190
3191 - Server and client now support transmission and reception
3192 of timezone information. It can be configured via the
3193 newly introduced network options UseTimezone=,
3194 EmitTimezone=, and Timezone=. Transmission of timezone
3195 information is enabled between host and containers by
3196 default now: the container will change its local timezone
3197 to what the host has set.
3198
3199 - Lease timeouts can now be configured via
3200 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
3201
3202 - The DHCP server improved on the stability of
3203 leases. Clients are more likely to get the same lease
3204 information back, even if the server loses state.
3205
3206 - The DHCP server supports two new configuration options to
3207 control the lease address pool metrics, PoolOffset= and
3208 PoolSize=.
3209
3210 * The encapsulation limit of tunnels in systemd-networkd may
3211 now be configured via 'EncapsulationLimit='. It allows
3212 modifying the maximum additional levels of encapsulation
3213 that are permitted to be prepended to a packet.
3214
3215 * systemd now supports the concept of user buses replacing
3216 session buses, if used with dbus-1.10 (and enabled via dbus
3217 --enable-user-session). It previously only supported this on
3218 kdbus-enabled systems, and this release expands this to
3219 'dbus-daemon' systems.
3220
3221 * systemd-networkd now supports predictable interface names
3222 for virtio devices.
3223
3224 * systemd now optionally supports the new Linux kernel
3225 "unified" control group hierarchy. If enabled via the kernel
3226 command-line option 'systemd.unified_cgroup_hierarchy=1',
3227 systemd will try to mount the unified cgroup hierarchy
3228 directly on /sys/fs/cgroup. If not enabled, or not
3229 available, systemd will fall back to the legacy cgroup
3230 hierarchy setup, as before. Host system and containers can
3231 mix and match legacy and unified hierarchies as they
856ca72b 3232 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
5e8d4254
LP
3233 environment variable to individually select the hierarchy to
3234 use for executed containers. By default, nspawn will use the
3235 unified hierarchy for the containers if the host uses the
3236 unified hierarchy, and the legacy hierarchy otherwise.
3237 Please note that at this point the unified hierarchy is an
3238 experimental kernel feature and is likely to change in one
3239 of the next kernel releases. Therefore, it should not be
3240 enabled by default in downstream distributions yet. The
3241 minimum required kernel version for the unified hierarchy to
3242 work is 4.2. Note that when the unified hierarchy is used
3243 for the first time delegated access to controllers is
3244 safe. Because of this systemd-nspawn containers will get
3245 access to controllers now, as will systemd user
3246 sessions. This means containers and user sessions may now
3247 manage their own resources, partitioning up what the system
3248 grants them.
3249
3250 * A new special scope unit "init.scope" has been introduced
3251 that encapsulates PID 1 of the system. It may be used to
3252 determine resource usage and enforce resource limits on PID
3253 1 itself. PID 1 hence moved out of the root of the control
3254 group tree.
3255
3256 * The cgtop tool gained support for filtering out kernel
3257 threads when counting tasks in a control group. Also, the
3258 count of processes is now recursively summed up by
3259 default. Two options -k and --recursive= have been added to
3260 revert to old behaviour. The tool has also been updated to
3261 work correctly in containers now.
3262
3263 * systemd-nspawn's --bind= and --bind-ro= options have been
3264 extended to allow creation of non-recursive bind mounts.
3265
c626bf1d
DM
3266 * libsystemd gained two new calls sd_pid_get_cgroup() and
3267 sd_peer_get_cgroup() which return the control group path of
5e8d4254
LP
3268 a process or peer of a connected AF_UNIX socket. This
3269 function call is particularly useful when implementing
3270 delegated subtrees support in the control group hierarchy.
3271
3272 * The "sd-event" event loop API of libsystemd now supports
3273 correct dequeuing of real-time signals, without losing
3274 signal events.
3275
3276 * When systemd requests a PolicyKit decision when managing
3277 units it will now add additional fields to the request,
3278 including unit name and desired operation. This enables more
3279 powerful PolicyKit policies, that make decisions depending
3280 on these parameters.
c9912c5e 3281
47f5a38c
LP
3282 * nspawn learnt support for .nspawn settings files, that may
3283 accompany the image files or directories of containers, and
3284 may contain additional settings for the container. This is
3285 an alternative to configuring container parameters via the
3286 nspawn command line.
3287
2f77decc
LP
3288 Contributions from: Cristian Rodríguez, Daniel Mack, David
3289 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
3290 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
3291 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
3292 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
3293 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
3294 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
23d08d1b 3295 Andersen, Tom Gundersen, Torstein Husebø
c9912c5e 3296
ccddd104 3297 — Berlin, 2015-09-08
c9912c5e 3298
ec5249a2
DM
3299CHANGES WITH 225:
3300
5e8d4254
LP
3301 * machinectl gained a new verb 'shell' which opens a fresh
3302 shell on the target container or the host. It is similar to
3303 the existing 'login' command of machinectl, but spawns the
3304 shell directly without prompting for username or
3305 password. The pseudo machine '.host' now refers to the local
3306 host and is used by default. Hence, 'machinectl shell' can
3307 be used as replacement for 'su -' which spawns a session as
3308 a fresh systemd unit in a way that is fully isolated from
3309 the originating session.
3310
3311 * systemd-networkd learned to cope with private-zone DHCP
3312 options and allows other programs to query the values.
3313
3314 * SELinux access control when enabling/disabling units is no
3315 longer enforced with this release. The previous
3316 implementation was incorrect, and a new corrected
3317 implementation is not yet available. As unit file operations
3318 are still protected via PolicyKit and D-Bus policy this is
3319 not a security problem. Yet, distributions which care about
3320 optimal SELinux support should probably not stabilize on
3321 this release.
3322
3323 * sd-bus gained support for matches of type "arg0has=", that
3324 test for membership of strings in string arrays sent in bus
3325 messages.
3326
3327 * systemd-resolved now dumps the contents of its DNS and LLMNR
3328 caches to the logs on reception of the SIGUSR1 signal. This
3329 is useful to debug DNS behaviour.
3330
3331 * The coredumpctl tool gained a new --directory= option to
3332 operate on journal files in a specific directory.
3333
3334 * "systemctl reboot" and related commands gained a new
3335 "--message=" option which may be used to set a free-text
3336 wall message when shutting down or rebooting the
3337 system. This message is also logged, which is useful for
3338 figuring out the reason for a reboot or shutdown a
3339 posteriori.
3340
3341 * The "systemd-resolve-host" tool's -i switch now takes
3342 network interface numbers as alternative to interface names.
3343
3344 * A new unit file setting for services has been introduced:
3345 UtmpMode= allows configuration of how precisely systemd
3346 handles utmp and wtmp entries for the service if this is
3347 enabled. This allows writing services that appear similar to
3348 user sessions in the output of the "w", "who", "last" and
3349 "lastlog" tools.
3350
3351 * systemd-resolved will now locally synthesize DNS resource
3352 records for the "localhost" and "gateway" domains as well as
3353 the local hostname. This should ensure that clients querying
3354 RRs via resolved will get similar results as those going via
3355 NSS, if nss-myhostname is enabled.
3356
3357 Contributions from: Alastair Hughes, Alex Crawford, Daniel
3358 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
3359 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
3360 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
3361 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
3362 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
3363 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
3364 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
3365 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
3366 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
3367 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
3368 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
e1439a14 3369
ccddd104 3370 — Berlin, 2015-08-27
ec5249a2 3371
11811e85
DH
3372CHANGES WITH 224:
3373
10fa421c
DH
3374 * The systemd-efi-boot-generator functionality was merged into
3375 systemd-gpt-auto-generator.
3376
5e8d4254
LP
3377 * systemd-networkd now supports Group Policy for vxlan
3378 devices. It can be enabled via the new boolean configuration
3379 option called 'GroupPolicyExtension='.
10fa421c 3380
11811e85
DH
3381 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
3382 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
3383 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
3384
ccddd104 3385 — Berlin, 2015-07-31
11811e85 3386
e57eaef8
DH
3387CHANGES WITH 223:
3388
3389 * The python-systemd code has been removed from the systemd repository.
3390 A new repository has been created which accommodates the code from
3391 now on, and we kindly ask distributions to create a separate package
3392 for this: https://github.com/systemd/python-systemd
3393
01608bc8 3394 * The systemd daemon will now reload its main configuration
e57eaef8
DH
3395 (/etc/systemd/system.conf) on daemon-reload.
3396
3397 * sd-dhcp now exposes vendor specific extensions via
3398 sd_dhcp_lease_get_vendor_specific().
3399
931618d0
DM
3400 * systemd-networkd gained a number of new configuration options.
3401
3402 - A new boolean configuration option for TAP devices called
37d54b93 3403 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
931618d0
DM
3404 device, thus allowing to send and receive GSO packets.
3405
3406 - A new tunnel configuration option called 'CopyDSCP='.
3407 If enabled, the DSCP field of ip6 tunnels is copied into the
3408 decapsulated packet.
3409
3410 - A set of boolean bridge configuration options were added.
3411 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
3412 and 'UnicastFlood=' are now parsed by networkd and applied to the
3413 respective bridge link device via the respective IFLA_BRPORT_*
3414 netlink attribute.
3415
3416 - A new string configuration option to override the hostname sent
3417 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
3418 is true, networkd will use the configured hostname instead of the
3419 system hostname when sending DHCP requests.
3420
3421 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
3422 networkd will configure the IPv6 flow-label of the tunnel device
3423 according to RFC2460.
e57eaef8 3424
f5f113f6
DH
3425 - The 'macvtap' virtual network devices are now supported, similar to
3426 the already supported 'macvlan' devices.
3427
e57eaef8 3428 * systemd-resolved now implements RFC5452 to improve resilience against
01608bc8 3429 cache poisoning. Additionally, source port randomization is enabled
e57eaef8
DH
3430 by default to further protect against DNS spoofing attacks.
3431
3432 * nss-mymachines now supports translating UIDs and GIDs of running
3433 containers with user-namespaces enabled. If a container 'foo'
3434 translates a host uid 'UID' to the container uid 'TUID', then
3435 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
3436 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
3437 mapped as 'vg-foo-TGID'.
3438
3439 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
e4e66993
DH
3440 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
3441 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
3442 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
3443 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
3444 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
3445 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
3446 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
3447 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
3448 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
3449
ccddd104 3450 — Berlin, 2015-07-29
e57eaef8 3451
0db83ad7 3452CHANGES WITH 222:
5541c889 3453
861b02eb
KS
3454 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
3455 There are no known issues with current sysfs, and udev does not need
3456 or should be used to work around such bugs.
3457
3458 * udev does no longer enable USB HID power management. Several reports
3459 indicate, that some devices cannot handle that setting.
0db83ad7
DH
3460
3461 * The udev accelerometer helper was removed. The functionality
3462 is now fully included in iio-sensor-proxy. But this means,
3463 older iio-sensor-proxy versions will no longer provide
3464 accelerometer/orientation data with this systemd version.
3465 Please upgrade iio-sensor-proxy to version 1.0.
3466
5541c889
DH
3467 * networkd gained a new configuration option IPv6PrivacyExtensions=
3468 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
3469 for Stateless Address") on selected networks.
3470
9b361114
DM
3471 * For the sake of fewer build-time dependencies and less code in the
3472 main repository, the python bindings are about to be removed in the
3473 next release. A new repository has been created which accommodates
3474 the code from now on, and we kindly ask distributions to create a
3475 separate package for this. The removal will take place in v223.
3476
3477 https://github.com/systemd/python-systemd
3478
0db83ad7
DH
3479 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
3480 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
3481 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
3482 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
5541c889
DH
3483 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
3484 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
0db83ad7
DH
3485 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
3486 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
2d1ca112
DH
3487 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
3488 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
0db83ad7 3489
ccddd104 3490 — Berlin, 2015-07-07
0db83ad7 3491
0f0467e6
MP
3492CHANGES WITH 221:
3493
470e72d4 3494 * The sd-bus.h and sd-event.h APIs have now been declared
5f92d24f 3495 stable and have been added to the official interface of
470e72d4
LP
3496 libsystemd.so. sd-bus implements an alternative D-Bus client
3497 library, that is relatively easy to use, very efficient and
3498 supports both classic D-Bus as well as kdbus as transport
3499 backend. sd-event is a generic event loop abstraction that
3500 is built around Linux epoll, but adds features such as event
0aee49d5 3501 prioritization or efficient timer handling. Both APIs are good
470e72d4
LP
3502 choices for C programs looking for a bus and/or event loop
3503 implementation that is minimal and does not have to be
5f92d24f 3504 portable to other kernels.
0f0467e6 3505
470e72d4
LP
3506 * kdbus support is no longer compile-time optional. It is now
3507 always built-in. However, it can still be disabled at
3508 runtime using the kdbus=0 kernel command line setting, and
c6551464 3509 that setting may be changed to default to off, by specifying
470e72d4
LP
3510 --disable-kdbus at build-time. Note though that the kernel
3511 command line setting has no effect if the kdbus.ko kernel
3512 module is not installed, in which case kdbus is (obviously)
3513 also disabled. We encourage all downstream distributions to
0aee49d5 3514 begin testing kdbus by adding it to the kernel images in the
470e72d4
LP
3515 development distributions, and leaving kdbus support in
3516 systemd enabled.
0f0467e6 3517
470e72d4
LP
3518 * The minimal required util-linux version has been bumped to
3519 2.26.
3520
3521 * Support for chkconfig (--enable-chkconfig) was removed in
0aee49d5 3522 favor of calling an abstraction tool
470e72d4
LP
3523 /lib/systemd/systemd-sysv-install. This needs to be
3524 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
3525 in README for details.
3526
3527 * If there's a systemd unit and a SysV init script for the
3528 same service name, and the user executes "systemctl enable"
3529 for it (or a related call), then this will now enable both
3530 (or execute the related operation on both), not just the
3531 unit.
3532
3533 * The libudev API documentation has been converted from gtkdoc
3534 into man pages.
3535
3536 * gudev has been removed from the systemd tree, it is now an
3537 external project.
3538
3539 * The systemd-cgtop tool learnt a new --raw switch to generate
0aee49d5 3540 "raw" (machine parsable) output.
470e72d4
LP
3541
3542 * networkd's IPForwarding= .network file setting learnt the
3543 new setting "kernel", which ensures that networkd does not
3544 change the IP forwarding sysctl from the default kernel
3545 state.
3546
3547 * The systemd-logind bus API now exposes a new boolean
3548 property "Docked" that reports whether logind considers the
3549 system "docked", i.e. connected to a docking station or not.
3550
3551 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
3552 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
3553 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
3554 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
3555 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
3556 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
3557 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
3558 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
3559 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
3560 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
3561 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
b912e251
LP
3562 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
3563 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
3564 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
3565 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
3566 Fink, Zbigniew Jędrzejewski-Szmek
470e72d4 3567
ccddd104 3568 — Berlin, 2015-06-19
0f0467e6 3569
481a0aa2
LP
3570CHANGES WITH 220:
3571
f7a73a25
DH
3572 * The gudev library has been extracted into a separate repository
3573 available at: https://git.gnome.org/browse/libgudev/
3574 It is now managed as part of the Gnome project. Distributions
3575 are recommended to pass --disable-gudev to systemd and use
3576 gudev from the Gnome project instead. gudev is still included
3577 in systemd, for now. It will be removed soon, though. Please
3578 also see the announcement-thread on systemd-devel:
56cadcb6 3579 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
f7a73a25 3580
481a0aa2
LP
3581 * systemd now exposes a CPUUsageNSec= property for each
3582 service unit on the bus, that contains the overall consumed
3583 CPU time of a service (the sum of what each process of the
3584 service consumed). This value is only available if
3585 CPUAccounting= is turned on for a service, and is then shown
3586 in the "systemctl status" output.
3587
3588 * Support for configuring alternative mappings of the old SysV
3589 runlevels to systemd targets has been removed. They are now
29d1fcb4 3590 hardcoded in a way that runlevels 2, 3, 4 all map to
481a0aa2
LP
3591 multi-user.target and 5 to graphical.target (which
3592 previously was already the default behaviour).
3593
3594 * The auto-mounter logic gained support for mount point
3595 expiry, using a new TimeoutIdleSec= setting in .automount
3596 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
3597
3598 * The EFI System Partition (ESP) as mounted to /boot by
3599 systemd-efi-boot-generator will now be unmounted
29d1fcb4 3600 automatically after 2 minutes of not being used. This should
481a0aa2
LP
3601 minimize the risk of ESP corruptions.
3602
3603 * New /etc/fstab options x-systemd.requires= and
3604 x-systemd.requires-mounts-for= are now supported to express
3605 additional dependencies for mounts. This is useful for
3606 journalling file systems that support external journal
3607 devices or overlay file systems that require underlying file
3608 systems to be mounted.
3609
3610 * systemd does not support direct live-upgrades (via systemctl
3611 daemon-reexec) from versions older than v44 anymore. As no
3612 distribution we are aware of shipped such old versions in a
3613 stable release this should not be problematic.
3614
3615 * When systemd forks off a new per-connection service instance
3616 it will now set the $REMOTE_ADDR environment variable to the
3617 remote IP address, and $REMOTE_PORT environment variable to
3618 the remote IP port. This behaviour is similar to the
3619 corresponding environment variables defined by CGI.
3620
3621 * systemd-networkd gained support for uplink failure
3622 detection. The BindCarrier= option allows binding interface
3623 configuration dynamically to the link sense of other
3624 interfaces. This is useful to achieve behaviour like in
3625 network switches.
3626
3627 * systemd-networkd gained support for configuring the DHCP
3628 client identifier to use when requesting leases.
3629
3630 * systemd-networkd now has a per-network UseNTP= option to
3631 configure whether NTP server information acquired via DHCP
3632 is passed on to services like systemd-timesyncd.
3633
3634 * systemd-networkd gained support for vti6 tunnels.
3635
1579dd2c
LP
3636 * Note that systemd-networkd manages the sysctl variable
3637 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
3638 it is configured for since v219. The variable controls IP
3639 forwarding, and is a per-interface alternative to the global
3640 /proc/sys/net/ipv[46]/ip_forward. This setting is
3641 configurable in the IPForward= option, which defaults to
3642 "no". This means if networkd is used for an interface it is
3643 no longer sufficient to set the global sysctl option to turn
3644 on IP forwarding! Instead, the .network file option
3645 IPForward= needs to be turned on! Note that the
3646 implementation of this behaviour was broken in v219 and has
3647 been fixed in v220.
3648
481a0aa2
LP
3649 * Many bonding and vxlan options are now configurable in
3650 systemd-networkd.
3651
3652 * systemd-nspawn gained a new --property= setting to set unit
3653 properties for the container scope. This is useful for
ce830873 3654 setting resource parameters (e.g. "CPUShares=500") on
481a0aa2
LP
3655 containers started from the command line.
3656
3657 * systemd-nspawn gained a new --private-users= switch to make
3658 use of user namespacing available on recent Linux kernels.
3659
3660 * systemd-nspawn may now be called as part of a shell pipeline
3661 in which case the pipes used for stdin and stdout are passed
3662 directly to the process invoked in the container, without
3663 indirection via a pseudo tty.
3664
3665 * systemd-nspawn gained a new switch to control the UNIX
3666 signal to use when killing the init process of the container
3667 when shutting down.
3668
3669 * systemd-nspawn gained a new --overlay= switch for mounting
3670 overlay file systems into the container using the new kernel
3671 overlayfs support.
3672
3673 * When a container image is imported via systemd-importd and
3674 the host file system is not btrfs, a loopback block device
3675 file is created in /var/lib/machines.raw with a btrfs file
3676 system inside. It is then mounted to /var/lib/machines to
3677 enable btrfs features for container management. The loopback
3678 file and btrfs file system is grown as needed when container
3679 images are imported via systemd-importd.
3680
3681 * systemd-machined/systemd-importd gained support for btrfs
3682 quota, to enforce container disk space limits on disk. This
3683 is exposed in "machinectl set-limit".
3684
3685 * systemd-importd now can import containers from local .tar,
3686 .raw and .qcow2 images, and export them to .tar and .raw. It
3687 can also import dkr v2 images now from the network (on top
3688 of v1 as before).
3689
3690 * systemd-importd gained support for verifying downloaded
3691 images with gpg2 (previously only gpg1 was supported).
3692
3693 * systemd-machined, systemd-logind, systemd: most bus calls
3694 are now accessible to unprivileged processes via
3695 PolicyKit. Also, systemd-logind will now allow users to kill
3696 their own sessions without further privileges or
3697 authorization.
3698
3699 * systemd-shutdownd has been removed. This service was
3700 previously responsible for implementing scheduled shutdowns
3701 as exposed in /usr/bin/shutdown's time parameter. This
3702 functionality has now been moved into systemd-logind and is
3703 accessible via a bus interface.
3704
3705 * "systemctl reboot" gained a new switch --firmware-setup that
3706 can be used to reboot into the EFI firmware setup, if that
3707 is available. systemd-logind now exposes an API on the bus
3708 to trigger such reboots, in case graphical desktop UIs want
3709 to cover this functionality.
3710
3711 * "systemctl enable", "systemctl disable" and "systemctl mask"
1579dd2c 3712 now support a new "--now" switch. If specified the units
481a0aa2
LP
3713 that are enabled will also be started, and the ones
3714 disabled/masked also stopped.
3715
3716 * The Gummiboot EFI boot loader tool has been merged into
1a2d5fbe
DH
3717 systemd, and renamed to "systemd-boot". The bootctl tool has been
3718 updated to support systemd-boot.
481a0aa2
LP
3719
3720 * An EFI kernel stub has been added that may be used to create
3721 kernel EFI binaries that contain not only the actual kernel,
3722 but also an initrd, boot splash, command line and OS release
3723 information. This combined binary can then be signed as a
3724 single image, so that the firmware can verify it all in one
1a2d5fbe 3725 step. systemd-boot has special support for EFI binaries created
481a0aa2
LP
3726 like this and can extract OS release information from them
3727 and show them in the boot menu. This functionality is useful
3728 to implement cryptographically verified boot schemes.
3729
3730 * Optional support has been added to systemd-fsck to pass
3731 fsck's progress report to an AF_UNIX socket in the file
3732 system.
3733
3734 * udev will no longer create device symlinks for all block
3735 devices by default. A blacklist for excluding special block
3736 devices from this logic has been turned into a whitelist
3737 that requires picking block devices explicitly that require
3738 device symlinks.
3739
3740 * A new (currently still internal) API sd-device.h has been
3741 added to libsystemd. This modernized API is supposed to
3742 replace libudev eventually. In fact, already much of libudev
3743 is now just a wrapper around sd-device.h.
3744
3745 * A new hwdb database for storing metadata about pointing
3746 stick devices has been added.
3747
3748 * systemd-tmpfiles gained support for setting file attributes
3749 similar to the "chattr" tool with new 'h' and 'H' lines.
3750
3751 * systemd-journald will no longer unconditionally set the
3752 btrfs NOCOW flag on new journal files. This is instead done
3753 with tmpfiles snippet using the new 'h' line type. This
3754 allows easy disabling of this logic, by masking the
3755 journal-nocow.conf tmpfiles file.
3756
3757 * systemd-journald will now translate audit message types to
3758 human readable identifiers when writing them to the
3759 journal. This should improve readability of audit messages.
3760
3761 * The LUKS logic gained support for the offset= and skip=
3762 options in /etc/crypttab, as previously implemented by
3763 Debian.
3764
3765 * /usr/lib/os-release gained a new optional field VARIANT= for
3766 distributions that support multiple variants (such as a
3767 desktop edition, a server edition, ...)
3768
3769 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
3770 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
3771 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
3772 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
3773 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
3774 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
3775 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
3776 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
3777 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
3778 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
3779 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
3780 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
3781 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
3782 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
3783 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
3784 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
3785 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
3786 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
3787 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
3788 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
3789 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
3790 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
3791 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
3792 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
3793 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
3794 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
3795 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
3796
ccddd104 3797 — Berlin, 2015-05-22
481a0aa2 3798
615aaf41
LP
3799CHANGES WITH 219:
3800
615aaf41
LP
3801 * Introduce a new API "sd-hwdb.h" for querying the hardware
3802 metadata database. With this minimal interface one can query
3803 and enumerate the udev hwdb, decoupled from the old libudev
3804 library. libudev's interface for this is now only a wrapper
3805 around sd-hwdb. A new tool systemd-hwdb has been added to
3806 interface with and update the database.
3807
3808 * When any of systemd's tools copies files (for example due to
3809 tmpfiles' C lines) a btrfs reflink will attempted first,
3810 before bytewise copying is done.
3811
3812 * systemd-nspawn gained a new --ephemeral switch. When
3813 specified a btrfs snapshot is taken of the container's root
3814 directory, and immediately removed when the container
3815 terminates again. Thus, a container can be started whose
3816 changes never alter the container's root directory, and are
3817 lost on container termination. This switch can also be used
3818 for starting a container off the root file system of the
3819 host without affecting the host OS. This switch is only
3820 available on btrfs file systems.
3821
3822 * systemd-nspawn gained a new --template= switch. It takes the
3823 path to a container tree to use as template for the tree
7edecf21 3824 specified via --directory=, should that directory be
615aaf41
LP
3825 missing. This allows instantiating containers dynamically,
3826 on first run. This switch is only available on btrfs file
3827 systems.
3828
3829 * When a .mount unit refers to a mount point on which multiple
3830 mounts are stacked, and the .mount unit is stopped all of
3831 the stacked mount points will now be unmounted until no
3832 mount point remains.
3833
3834 * systemd now has an explicit notion of supported and
3835 unsupported unit types. Jobs enqueued for unsupported unit
3836 types will now fail with an "unsupported" error code. More
3837 specifically .swap, .automount and .device units are not
3838 supported in containers, .busname units are not supported on
3839 non-kdbus systems. .swap and .automount are also not
3840 supported if their respective kernel compile time options
3841 are disabled.
3842
3843 * machinectl gained support for two new "copy-from" and
3844 "copy-to" commands for copying files from a running
3845 container to the host or vice versa.
3846
3847 * machinectl gained support for a new "bind" command to bind
3848 mount host directories into local containers. This is
3849 currently only supported for nspawn containers.
3850
3851 * networkd gained support for configuring bridge forwarding
3852 database entries (fdb) from .network files.
3853
3854 * A new tiny daemon "systemd-importd" has been added that can
3855 download container images in tar, raw, qcow2 or dkr formats,
3856 and make them available locally in /var/lib/machines, so
3857 that they can run as nspawn containers. The daemon can GPG
3858 verify the downloads (not supported for dkr, since it has no
3859 provisions for verifying downloads). It will transparently
3860 decompress bz2, xz, gzip compressed downloads if necessary,
3861 and restore sparse files on disk. The daemon uses privilege
3862 separation to ensure the actual download logic runs with
94e5ba37 3863 fewer privileges than the daemon itself. machinectl has
615aaf41
LP
3864 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
3865 make the functionality of importd available to the
3866 user. With this in place the Fedora and Ubuntu "Cloud"
3867 images can be downloaded and booted as containers unmodified
3868 (the Fedora images lack the appropriate GPG signature files
3869 currently, so they cannot be verified, but this will change
3870 soon, hopefully). Note that downloading images is currently
3871 only fully supported on btrfs.
3872
3873 * machinectl is now able to list container images found in
3874 /var/lib/machines, along with some metadata about sizes of
3875 disk and similar. If the directory is located on btrfs and
3876 quota is enabled, this includes quota display. A new command
3877 "image-status" has been added that shows additional
3878 information about images.
3879
3880 * machinectl is now able to clone container images
3881 efficiently, if the underlying file system (btrfs) supports
f59dba26 3882 it, with the new "machinectl clone" command. It also
615aaf41
LP
3883 gained commands for renaming and removing images, as well as
3884 marking them read-only or read-write (supported also on
3885 legacy file systems).
3886
3887 * networkd gained support for collecting LLDP network
3888 announcements, from hardware that supports this. This is
3889 shown in networkctl output.
3890
3891 * systemd-run gained support for a new -t (--pty) switch for
3892 invoking a binary on a pty whose input and output is
3893 connected to the invoking terminal. This allows executing
3894 processes as system services while interactively
3895 communicating with them via the terminal. Most interestingly
3896 this is supported across container boundaries. Invoking
3897 "systemd-run -t /bin/bash" is an alternative to running a
3898 full login session, the difference being that the former
3899 will not register a session, nor go through the PAM session
3900 setup.
3901
3902 * tmpfiles gained support for a new "v" line type for creating
3903 btrfs subvolumes. If the underlying file system is a legacy
3904 file system, this automatically degrades to creating a
3905 normal directory. Among others /var/lib/machines is now
3906 created like this at boot, should it be missing.
3907
3908 * The directory /var/lib/containers/ has been deprecated and
3909 been replaced by /var/lib/machines. The term "machines" has
3910 been used in the systemd context as generic term for both
3911 VMs and containers, and hence appears more appropriate for
3912 this, as the directory can also contain raw images bootable
3913 via qemu/kvm.
3914
3915 * systemd-nspawn when invoked with -M but without --directory=
3916 or --image= is now capable of searching for the container
3917 root directory, subvolume or disk image automatically, in
3918 /var/lib/machines. systemd-nspawn@.service has been updated
3919 to make use of this, thus allowing it to be used for raw
3920 disk images, too.
3921
3922 * A new machines.target unit has been introduced that is
3923 supposed to group all containers/VMs invoked as services on
3924 the system. systemd-nspawn@.service has been updated to
3925 integrate with that.
3926
3927 * machinectl gained a new "start" command, for invoking a
3928 container as a service. "machinectl start foo" is mostly
3929 equivalent to "systemctl start systemd-nspawn@foo.service",
3930 but handles escaping in a nicer way.
3931
3932 * systemd-nspawn will now mount most of the cgroupfs tree
3933 read-only into each container, with the exception of the
3934 container's own subtree in the name=systemd hierarchy.
3935
3936 * journald now sets the special FS_NOCOW file flag for its
3937 journal files. This should improve performance on btrfs, by
3938 avoiding heavy fragmentation when journald's write-pattern
3939 is used on COW file systems. It degrades btrfs' data
3940 integrity guarantees for the files to the same levels as for
3941 ext3/ext4 however. This should be OK though as journald does
3942 its own data integrity checks and all its objects are
3943 checksummed on disk. Also, journald should handle btrfs disk
3944 full events a lot more gracefully now, by processing SIGBUS
3945 errors, and not relying on fallocate() anymore.
3946
3947 * When journald detects that journal files it is writing to
3948 have been deleted it will immediately start new journal
3949 files.
3950
3951 * systemd now provides a way to store file descriptors
4c37970d 3952 per-service in PID 1. This is useful for daemons to ensure
615aaf41 3953 that fds they require are not lost during a daemon
94e5ba37 3954 restart. The fds are passed to the daemon on the next
615aaf41
LP
3955 invocation in the same way socket activation fds are
3956 passed. This is now used by journald to ensure that the
3957 various sockets connected to all the system's stdout/stderr
3958 are not lost when journald is restarted. File descriptors
3959 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
3960 an extension to sd_notify(). Note that a limit is enforced
3961 on the number of fds a service can store in PID 1, and it
3962 defaults to 0, so that no fds may be stored, unless this is
3963 explicitly turned on.
3964
3965 * The default TERM variable to use for units connected to a
3966 terminal, when no other value is explicitly is set is now
3967 vt220 rather than vt102. This should be fairly safe still,
3968 but allows PgUp/PgDn work.
3969
3970 * The /etc/crypttab option header= as known from Debian is now
3971 supported.
3972
3973 * "loginctl user-status" and "loginctl session-status" will
3974 now show the last 10 lines of log messages of the
3975 user/session following the status output. Similar,
3976 "machinectl status" will show the last 10 log lines
3977 associated with a virtual machine or container
3978 service. (Note that this is usually not the log messages
3979 done in the VM/container itself, but simply what the
3980 container manager logs. For nspawn this includes all console
3981 output however.)
3982
3983 * "loginctl session-status" without further argument will now
3984 show the status of the session of the caller. Similar,
3985 "lock-session", "unlock-session", "activate",
3986 "enable-linger", "disable-linger" may now be called without
3987 session/user parameter in which case they apply to the
3988 caller's session/user.
3989
3990 * An X11 session scriptlet is now shipped that uploads
3991 $DISPLAY and $XAUTHORITY into the environment of the systemd
3992 --user daemon if a session begins. This should improve
3993 compatibility with X11 enabled applications run as systemd
3994 user services.
3995
3996 * Generators are now subject to masking via /etc and /run, the
3997 same way as unit files.
3998
3999 * networkd .network files gained support for configuring
4000 per-link IPv4/IPv6 packet forwarding as well as IPv4
4001 masquerading. This is by default turned on for veth links to
4002 containers, as registered by systemd-nspawn. This means that
4003 nspawn containers run with --network-veth will now get
4004 automatic routed access to the host's networks without any
4005 further configuration or setup, as long as networkd runs on
4006 the host.
4007
4008 * systemd-nspawn gained the --port= (-p) switch to expose TCP
4009 or UDP posts of a container on the host. With this in place
4010 it is possible to run containers with private veth links
4011 (--network-veth), and have their functionality exposed on
4012 the host as if their services were running directly on the
4013 host.
4014
dd2fd155 4015 * systemd-nspawn's --network-veth switch now gained a short
615aaf41
LP
4016 version "-n", since with the changes above it is now truly
4017 useful out-of-the-box. The systemd-nspawn@.service has been
4018 updated to make use of it too by default.
4019
4020 * systemd-nspawn will now maintain a per-image R/W lock, to
4021 ensure that the same image is not started more than once
4022 writable. (It's OK to run an image multiple times
4023 simultaneously in read-only mode.)
4024
4025 * systemd-nspawn's --image= option is now capable of
4026 dissecting and booting MBR and GPT disk images that contain
4027 only a single active Linux partition. Previously it
4028 supported only GPT disk images with proper GPT type
4029 IDs. This allows running cloud images from major
4030 distributions directly with systemd-nspawn, without
4031 modification.
4032
4033 * In addition to collecting mouse dpi data in the udev
4034 hardware database, there's now support for collecting angle
4035 information for mouse scroll wheels. The database is
7edecf21 4036 supposed to guarantee similar scrolling behavior on mice
615aaf41
LP
4037 that it knows about. There's also support for collecting
4038 information about Touchpad types.
4039
4040 * udev's input_id built-in will now also collect touch screen
4041 dimension data and attach it to probed devices.
4042
4043 * /etc/os-release gained support for a Distribution Privacy
4044 Policy link field.
4045
4046 * networkd gained support for creating "ipvlan", "gretap",
4047 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
4048
4049 * systemd-tmpfiles gained support for "a" lines for setting
4050 ACLs on files.
4051
4052 * systemd-nspawn will now mount /tmp in the container to
4053 tmpfs, automatically.
4054
4055 * systemd now exposes the memory.usage_in_bytes cgroup
4056 attribute and shows it for each service in the "systemctl
4057 status" output, if available.
4058
4059 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
4060 immediate reboot is triggered. This useful if shutdown is
4061 hung and is unable to complete, to expedite the
4062 operation. Note that this kind of reboot will still unmount
4063 all file systems, and hence should not result in fsck being
4064 run on next reboot.
4065
4066 * A .device unit for an optical block device will now be
4067 considered active only when a medium is in the drive. Also,
4068 mount units are now bound to their backing devices thus
4069 triggering automatic unmounting when devices become
4070 unavailable. With this in place systemd will now
4071 automatically unmount left-over mounts when a CD-ROM is
4072 ejected or an USB stick is yanked from the system.
4073
4074 * networkd-wait-online now has support for waiting for
4075 specific interfaces only (with globbing), and for giving up
4076 after a configurable timeout.
4077
4078 * networkd now exits when idle. It will be automatically
4079 restarted as soon as interfaces show up, are removed or
4080 change state. networkd will stay around as long as there is
4081 at least one DHCP state machine or similar around, that keep
4082 it non-idle.
4083
4084 * networkd may now configure IPv6 link-local addressing in
4085 addition to IPv4 link-local addressing.
4086
4087 * The IPv6 "token" for use in SLAAC may now be configured for
4088 each .network interface in networkd.
4089
4090 * Routes configured with networkd may now be assigned a scope
4091 in .network files.
4092
4093 * networkd's [Match] sections now support globbing and lists
4094 of multiple space-separated matches per item.
4095
11ea2781 4096 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
d2c643c6
LP
4097 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
4098 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
4099 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
4100 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
4101 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
4102 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
4103 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
4104 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
4105 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
4106 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
4107 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
4108 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
4109 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
4110 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
11ea2781
LP
4111 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
4112 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
4113 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
4114 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
4115 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
4116 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
4117 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
d2c643c6
LP
4118 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
4119 Hoffmann, Zbigniew Jędrzejewski-Szmek
11ea2781 4120
ccddd104 4121 — Berlin, 2015-02-16
11ea2781 4122
d4f5a1f4
DH
4123CHANGES WITH 218:
4124
f9e00a9f
LP
4125 * When querying unit file enablement status (for example via
4126 "systemctl is-enabled"), a new state "indirect" is now known
4127 which indicates that a unit might not be enabled itself, but
c7683ffb 4128 another unit listed in its Also= setting might be.
f9e00a9f
LP
4129
4130 * Similar to the various existing ConditionXYZ= settings for
b938cb90 4131 units, there are now matching AssertXYZ= settings. While
f9e00a9f
LP
4132 failing conditions cause a unit to be skipped, but its job
4133 to succeed, failing assertions declared like this will cause
4134 a unit start operation and its job to fail.
4135
4136 * hostnamed now knows a new chassis type "embedded".
4137
4138 * systemctl gained a new "edit" command. When used on a unit
b938cb90 4139 file, this allows extending unit files with .d/ drop-in
f9e00a9f
LP
4140 configuration snippets or editing the full file (after
4141 copying it from /usr/lib to /etc). This will invoke the
4142 user's editor (as configured with $EDITOR), and reload the
4143 modified configuration after editing.
4144
4145 * "systemctl status" now shows the suggested enablement state
4146 for a unit, as declared in the (usually vendor-supplied)
4147 system preset files.
4148
4149 * nss-myhostname will now resolve the single-label host name
4150 "gateway" to the locally configured default IP routing
4151 gateways, ordered by their metrics. This assigns a stable
4152 name to the used gateways, regardless which ones are
4153 currently configured. Note that the name will only be
4154 resolved after all other name sources (if nss-myhostname is
4155 configured properly) and should hence not negatively impact
4156 systems that use the single-label host name "gateway" in
4157 other contexts.
4158
4159 * systemd-inhibit now allows filtering by mode when listing
4160 inhibitors.
4161
122676c9 4162 * Scope and service units gained a new "Delegate" boolean
b938cb90 4163 property, which, when set, allows processes running inside the
122676c9
LP
4164 unit to further partition resources. This is primarily
4165 useful for systemd user instances as well as container
4166 managers.
f9e00a9f
LP
4167
4168 * journald will now pick up audit messages directly from
4169 the kernel, and log them like any other log message. The
4170 audit fields are split up and fully indexed. This means that
4171 journalctl in many ways is now a (nicer!) alternative to
4172 ausearch, the traditional audit client. Note that this
b938cb90 4173 implements only a minimal audit client. If you want the
f9e00a9f
LP
4174 special audit modes like reboot-on-log-overflow, please use
4175 the traditional auditd instead, which can be used in
4176 parallel to journald.
4177
4178 * The ConditionSecurity= unit file option now understands the
4179 special string "audit" to check whether auditing is
4180 available.
4181
4182 * journalctl gained two new commands --vacuum-size= and
4183 --vacuum-time= to delete old journal files until the
a8eaaee7 4184 remaining ones take up no more than the specified size on disk,
f9e00a9f
LP
4185 or are not older than the specified time.
4186
4187 * A new, native PPPoE library has been added to sd-network,
4188 systemd's library of light-weight networking protocols. This
4189 library will be used in a future version of networkd to
4190 enable PPPoE communication without an external pppd daemon.
4191
4192 * The busctl tool now understands a new "capture" verb that
4193 works similar to "monitor", but writes a packet capture
4194 trace to STDOUT that can be redirected to a file which is
4195 compatible with libcap's capture file format. This can then
4196 be loaded in Wireshark and similar tools to inspect bus
4197 communication.
4198
4199 * The busctl tool now understands a new "tree" verb that shows
4200 the object trees of a specific service on the bus, or of all
4201 services.
4202
4203 * The busctl tool now understands a new "introspect" verb that
4204 shows all interfaces and members of objects on the bus,
4205 including their signature and values. This is particularly
4206 useful to get more information about bus objects shown by
4207 the new "busctl tree" command.
4208
4209 * The busctl tool now understands new verbs "call",
4210 "set-property" and "get-property" for invoking bus method
4211 calls, setting and getting bus object properties in a
4212 friendly way.
4213
4214 * busctl gained a new --augment-creds= argument that controls
4215 whether the tool shall augment credential information it
4216 gets from the bus with data from /proc, in a possibly
4217 race-ful way.
4218
4219 * nspawn's --link-journal= switch gained two new values
4220 "try-guest" and "try-host" that work like "guest" and
17c29493 4221 "host", but do not fail if the host has no persistent
f9e00a9f
LP
4222 journalling enabled. -j is now equivalent to
4223 --link-journal=try-guest.
4224
4225 * macvlan network devices created by nspawn will now have
4226 stable MAC addresses.
4227
4228 * A new SmackProcessLabel= unit setting has been added, which
4229 controls the SMACK security label processes forked off by
4230 the respective unit shall use.
4231
d4f5a1f4
DH
4232 * If compiled with --enable-xkbcommon, systemd-localed will
4233 verify x11 keymap settings by compiling the given keymap. It
4234 will spew out warnings if the compilation fails. This
4235 requires libxkbcommon to be installed.
4236
b938cb90 4237 * When a coredump is collected, a larger number of metadata
f9e00a9f 4238 fields is now collected and included in the journal records
b938cb90 4239 created for it. More specifically, control group membership,
f9e00a9f
LP
4240 environment variables, memory maps, working directory,
4241 chroot directory, /proc/$PID/status, and a list of open file
4242 descriptors is now stored in the log entry.
4243
17c29493 4244 * The udev hwdb now contains DPI information for mice. For
f9e00a9f
LP
4245 details see:
4246
4247 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
4248
4249 * All systemd programs that read standalone configuration
4250 files in /etc now also support a corresponding series of
997b2b43
JT
4251 .conf.d configuration directories in /etc/, /run/,
4252 /usr/local/lib/, /usr/lib/, and (if configured with
4253 --enable-split-usr) /lib/. In particular, the following
4254 configuration files now have corresponding configuration
4255 directories: system.conf user.conf, logind.conf,
4256 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
4257 resolved.conf, timesyncd.conf, journal-remote.conf, and
4258 journal-upload.conf. Note that distributions should use the
4259 configuration directories in /usr/lib/; the directories in
4260 /etc/ are reserved for the system administrator.
4261
f9e00a9f
LP
4262 * systemd-rfkill will no longer take the rfkill device name
4263 into account when storing rfkill state on disk, as the name
4264 might be dynamically assigned and not stable. Instead, the
4265 ID_PATH udev variable combined with the rfkill type (wlan,
4266 bluetooth, ...) is used.
4267
4268 * A new service systemd-machine-id-commit.service has been
4269 added. When used on systems where /etc is read-only during
4270 boot, and /etc/machine-id is not initialized (but an empty
4271 file), this service will copy the temporary machine ID
4272 created as replacement into /etc after the system is fully
4273 booted up. This is useful for systems that are freshly
4274 installed with a non-initialized machine ID, but should get
4275 a fixed machine ID for subsequent boots.
4276
4277 * networkd's .netdev files now provide a large set of
a8eaaee7 4278 configuration parameters for VXLAN devices. Similarly, the
f9e00a9f
LP
4279 bridge port cost parameter is now configurable in .network
4280 files. There's also new support for configuring IP source
4281 routing. networkd .link files gained support for a new
4282 OriginalName= match that is useful to match against the
4283 original interface name the kernel assigned. .network files
4284 may include MTU= and MACAddress= fields for altering the MTU
4285 and MAC address while being connected to a specific network
4286 interface.
4287
4288 * The LUKS logic gained supported for configuring
4289 UUID-specific key files. There's also new support for naming
4290 LUKS device from the kernel command line, using the new
4291 luks.name= argument.
4292
4293 * Timer units may now be transiently created via the bus API
4294 (this was previously already available for scope and service
4295 units). In addition it is now possible to create multiple
4296 transient units at the same time with a single bus call. The
4297 "systemd-run" tool has been updated to make use of this for
4298 running commands on a specified time, in at(1)-style.
4299
4300 * tmpfiles gained support for "t" lines, for assigning
4301 extended attributes to files. Among other uses this may be
4302 used to assign SMACK labels to files.
4303
13e92f39
LP
4304 Contributions from: Alin Rauta, Alison Chaiken, Andrej
4305 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
4306 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
4307 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
4308 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
4309 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
4310 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
4311 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
4312 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
4313 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
4314 Michael Biebl, Michael Chapman, Michael Marineau, Michal
7da81d33
LP
4315 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
4316 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
4317 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
4318 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
4319 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
4320 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
4321 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
13e92f39 4322
ccddd104 4323 — Berlin, 2014-12-10
f9e00a9f 4324
b62a309a
ZJS
4325CHANGES WITH 217:
4326
78b6b7ce
LP
4327 * journalctl gained the new options -t/--identifier= to match
4328 on the syslog identifier (aka "tag"), as well as --utc to
4329 show log timestamps in the UTC timezone. journalctl now also
4330 accepts -n/--lines=all to disable line capping in a pager.
b62a309a 4331
a65b8245
ZJS
4332 * journalctl gained a new switch, --flush, that synchronously
4333 flushes logs from /run/log/journal to /var/log/journal if
4334 persistent storage is enabled. systemd-journal-flush.service
4335 now waits until the operation is complete.
2a97b03b 4336
b62a309a
ZJS
4337 * Services can notify the manager before they start a reload
4338 (by sending RELOADING=1) or shutdown (by sending
4bdc60cb
LP
4339 STOPPING=1). This allows the manager to track and show the
4340 internal state of daemons and closes a race condition when
78b6b7ce 4341 the process is still running but has closed its D-Bus
4bdc60cb 4342 connection.
b62a309a 4343
78b6b7ce
LP
4344 * Services with Type=oneshot do not have to have any ExecStart
4345 commands anymore.
b62a309a
ZJS
4346
4347 * User units are now loaded also from
4348 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
4349 /run/systemd/user directory that was already previously
4350 supported, but is under the control of the user.
4351
4ffd29fd
LP
4352 * Job timeouts (i.e. time-outs on the time a job that is
4353 queued stays in the run queue) can now optionally result in
4354 immediate reboot or power-off actions (JobTimeoutAction= and
4355 JobTimeoutRebootArgument=). This is useful on ".target"
4356 units, to limit the maximum time a target remains
4357 undispatched in the run queue, and to trigger an emergency
4358 operation in such a case. This is now used by default to
4359 turn off the system if boot-up (as defined by everything in
4360 basic.target) hangs and does not complete for at least
4361 15min. Also, if power-off or reboot hang for at least 30min
4362 an immediate power-off/reboot operation is triggered. This
4363 functionality is particularly useful to increase reliability
4364 on embedded devices, but also on laptops which might
4365 accidentally get powered on when carried in a backpack and
4366 whose boot stays stuck in a hard disk encryption passphrase
4367 question.
4368
b62a309a
ZJS
4369 * systemd-logind can be configured to also handle lid switch
4370 events even when the machine is docked or multiple displays
4371 are attached (HandleLidSwitchDocked= option).
4372
4373 * A helper binary and a service have been added which can be
4374 used to resume from hibernation in the initramfs. A
4375 generator will parse the resume= option on the kernel
81c7dd89 4376 command line to trigger resume.
b62a309a 4377
78b6b7ce
LP
4378 * A user console daemon systemd-consoled has been
4379 added. Currently, it is a preview, and will so far open a
4380 single terminal on each session of the user marked as
09077149 4381 Desktop=systemd-console.
b62a309a
ZJS
4382
4383 * Route metrics can be specified for DHCP routes added by
4384 systemd-networkd.
4385
ba8df74b 4386 * The SELinux context of socket-activated services can be set
78b6b7ce 4387 from the information provided by the networking stack
b62a309a
ZJS
4388 (SELinuxContextFromNet= option).
4389
4390 * Userspace firmware loading support has been removed and
4391 the minimum supported kernel version is thus bumped to 3.7.
4392
4393 * Timeout for udev workers has been increased from 1 to 3
4394 minutes, but a warning will be printed after 1 minute to
4395 help diagnose kernel modules that take a long time to load.
4396
78b6b7ce 4397 * Udev rules can now remove tags on devices with TAG-="foobar".
b62a309a 4398
4bdc60cb 4399 * systemd's readahead implementation has been removed. In many
f6d1de85 4400 circumstances it didn't give expected benefits even for
b62a309a 4401 rotational disk drives and was becoming less relevant in the
78b6b7ce
LP
4402 age of SSDs. As none of the developers has been using
4403 rotating media anymore, and nobody stepped up to actively
4404 maintain this component of systemd it has now been removed.
b62a309a 4405
c4ac9900 4406 * Swap units can use Options= to specify discard options.
b62a309a
ZJS
4407 Discard options specified for swaps in /etc/fstab are now
4408 respected.
4409
4410 * Docker containers are now detected as a separate type of
4411 virtualization.
4412
4413 * The Password Agent protocol gained support for queries where
ba8df74b 4414 the user input is shown, useful e.g. for user names.
78b6b7ce
LP
4415 systemd-ask-password gained a new --echo option to turn that
4416 on.
b62a309a 4417
e6c253e3
MS
4418 * The default sysctl.d/ snippets will now set:
4419
4420 net.core.default_qdisc = fq_codel
4421
ba8df74b
KS
4422 This selects Fair Queuing Controlled Delay as the default
4423 queuing discipline for network interfaces. fq_codel helps
e6c253e3
MS
4424 fight the network bufferbloat problem. It is believed to be
4425 a good default with no tuning required for most workloads.
4426 Downstream distributions may override this choice. On 10Gbit
4427 servers that do not do forwarding, "fq" may perform better.
4428 Systems without a good clocksource should use "pfifo_fast".
4429
4bdc60cb
LP
4430 * If kdbus is enabled during build a new option BusPolicy= is
4431 available for service units, that allows locking all service
4432 processes into a stricter bus policy, in order to limit
4433 access to various bus services, or even hide most of them
4434 from the service's view entirely.
4435
4436 * networkctl will now show the .network and .link file
4437 networkd has applied to a specific interface.
4438
4439 * sd-login gained a new API call sd_session_get_desktop() to
4440 query which desktop environment has been selected for a
4441 session.
4442
4443 * UNIX utmp support is now compile-time optional to support
4444 legacy-free systems.
4445
78b6b7ce
LP
4446 * systemctl gained two new commands "add-wants" and
4447 "add-requires" for pulling in units from specific targets
4448 easily.
4449
4450 * If the word "rescue" is specified on the kernel command line
4451 the system will now boot into rescue mode (aka
4452 rescue.target), which was previously available only by
4453 specifying "1" or "systemd.unit=rescue.target" on the kernel
4454 command line. This new kernel command line option nicely
4455 mirrors the already existing "emergency" kernel command line
4456 option.
4457
4458 * New kernel command line options mount.usr=, mount.usrflags=,
d4474c41 4459 mount.usrfstype= have been added that match root=, rootflags=,
78b6b7ce
LP
4460 rootfstype= but allow mounting a specific file system to
4461 /usr.
4462
f6d1de85 4463 * The $NOTIFY_SOCKET is now also passed to control processes of
78b6b7ce
LP
4464 services, not only the main process.
4465
4466 * This version reenables support for fsck's -l switch. This
4467 means at least version v2.25 of util-linux is required for
4468 operation, otherwise dead-locks on device nodes may
4469 occur. Again: you need to update util-linux to at least
4470 v2.25 when updating systemd to v217.
4471
3769415e
TT
4472 * The "multi-seat-x" tool has been removed from systemd, as
4473 its functionality has been integrated into X servers 1.16,
4474 and the tool is hence redundant. It is recommended to update
4475 display managers invoking this tool to simply invoke X
4476 directly from now on, again.
4477
fae9332b
LP
4478 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
4479 message flag has been added for all of systemd's PolicyKit
4480 authenticated method calls has been added. In particular
4481 this now allows optional interactive authorization via
ba8df74b 4482 PolicyKit for many of PID1's privileged operations such as
fae9332b
LP
4483 unit file enabling and disabling.
4484
cfa1571b
LP
4485 * "udevadm hwdb --update" learnt a new switch "--usr" for
4486 placing the rebuilt hardware database in /usr instead of
4487 /etc. When used only hardware database entries stored in
4488 /usr will be used, and any user database entries in /etc are
4489 ignored. This functionality is useful for vendors to ship a
4490 pre-built database on systems where local configuration is
4491 unnecessary or unlikely.
4492
7e63dd10
LP
4493 * Calendar time specifications in .timer units now also
4494 understand the strings "semi-annually", "quarterly" and
ba8df74b 4495 "minutely" as shortcuts (in addition to the preexisting
7e63dd10
LP
4496 "anually", "hourly", ...).
4497
d4474c41
TG
4498 * systemd-tmpfiles will now correctly create files in /dev
4499 at boot which are marked for creation only at boot. It is
4500 recommended to always create static device nodes with 'c!'
4501 and 'b!', so that they are created only at boot and not
4502 overwritten at runtime.
4503
3b187c5c
LP
4504 * When the watchdog logic is used for a service (WatchdogSec=)
4505 and the watchdog timeout is hit the service will now be
4506 terminated with SIGABRT (instead of just SIGTERM), in order
4507 to make sure a proper coredump and backtrace is
4508 generated. This ensures that hanging services will result in
4509 similar coredump/backtrace behaviour as services that hit a
4510 segmentation fault.
4511
4b08dd87
LP
4512 Contributions from: Andreas Henriksson, Andrei Borzenkov,
4513 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
4514 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
4515 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
4516 Herrmann, David Sommerseth, David Strauss, Emil Renner
4517 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
4518 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
4519 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
4520 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
4521 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
4522 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
4523 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
4524 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
4525 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
4526 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
4527 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
4528 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
4529 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
4530 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
4531 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
4532 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
13e92f39 4533 Jędrzejewski-Szmek
4b08dd87 4534
ccddd104 4535 — Berlin, 2014-10-28
4b08dd87 4536
b72ddf0f 4537CHANGES WITH 216:
b2ca0d63
LP
4538
4539 * timedated no longer reads NTP implementation unit names from
b72ddf0f 4540 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
b2ca0d63
LP
4541 implementations should add a
4542
b72ddf0f 4543 Conflicts=systemd-timesyncd.service
b2ca0d63
LP
4544
4545 to their unit files to take over and replace systemd's NTP
4546 default functionality.
4547
4548 * systemd-sysusers gained a new line type "r" for configuring
4549 which UID/GID ranges to allocate system users/groups
4550 from. Lines of type "u" may now add an additional column
4551 that specifies the home directory for the system user to be
4552 created. Also, systemd-sysusers may now optionally read user
4553 information from STDIN instead of a file. This is useful for
4554 invoking it from RPM preinst scriptlets that need to create
4555 users before the first RPM file is installed since these
4556 files might need to be owned by them. A new
4557 %sysusers_create_inline RPM macro has been introduced to do
4558 just that. systemd-sysusers now updates the shadow files as
4559 well as the user/group databases, which should enhance
4560 compatibility with certain tools like grpck.
4561
4562 * A number of bus APIs of PID 1 now optionally consult
5f02e26c 4563 PolicyKit to permit access for otherwise unprivileged
b2ca0d63
LP
4564 clients under certain conditions. Note that this currently
4565 doesn't support interactive authentication yet, but this is
4566 expected to be added eventually, too.
4567
4568 * /etc/machine-info now has new fields for configuring the
4569 deployment environment of the machine, as well as the
4570 location of the machine. hostnamectl has been updated with
4571 new command to update these fields.
4572
4573 * systemd-timesyncd has been updated to automatically acquire
4574 NTP server information from systemd-networkd, which might
4575 have been discovered via DHCP.
4576
4577 * systemd-resolved now includes a caching DNS stub resolver
4578 and a complete LLMNR name resolution implementation. A new
daa05349
AB
4579 NSS module "nss-resolve" has been added which can be used
4580 instead of glibc's own "nss-dns" to resolve hostnames via
b2ca0d63
LP
4581 systemd-resolved. Hostnames, addresses and arbitrary RRs may
4582 be resolved via systemd-resolved D-Bus APIs. In contrast to
4583 the glibc internal resolver systemd-resolved is aware of
4584 multi-homed system, and keeps DNS server and caches separate
5f02e26c 4585 and per-interface. Queries are sent simultaneously on all
b2ca0d63
LP
4586 interfaces that have DNS servers configured, in order to
4587 properly handle VPNs and local LANs which might resolve
4588 separate sets of domain names. systemd-resolved may acquire
a1a4a25e 4589 DNS server information from systemd-networkd automatically,
b2ca0d63
LP
4590 which in turn might have discovered them via DHCP. A tool
4591 "systemd-resolve-host" has been added that may be used to
4592 query the DNS logic in resolved. systemd-resolved implements
4593 IDNA and automatically uses IDNA or UTF-8 encoding depending
4594 on whether classic DNS or LLMNR is used as transport. In the
4595 next releases we intend to add a DNSSEC and mDNS/DNS-SD
4596 implementation to systemd-resolved.
4597
4598 * A new NSS module nss-mymachines has been added, that
4599 automatically resolves the names of all local registered
4600 containers to their respective IP addresses.
4601
4602 * A new client tool "networkctl" for systemd-networkd has been
4603 added. It currently is entirely passive and will query
4604 networking configuration from udev, rtnetlink and networkd,
5f02e26c 4605 and present it to the user in a very friendly
b2ca0d63
LP
4606 way. Eventually, we hope to extend it to become a full
4607 control utility for networkd.
4608
4609 * .socket units gained a new DeferAcceptSec= setting that
4610 controls the kernels' TCP_DEFER_ACCEPT sockopt for
a8eaaee7 4611 TCP. Similarly, support for controlling TCP keep-alive
b2ca0d63
LP
4612 settings has been added (KeepAliveTimeSec=,
4613 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
4614 turning off Nagle's algorithm on TCP has been added
4615 (NoDelay=).
4616
a1a4a25e 4617 * logind learned a new session type "web", for use in projects
b2ca0d63
LP
4618 like Cockpit which register web clients as PAM sessions.
4619
4620 * timer units with at least one OnCalendar= setting will now
46ae28d8 4621 be started only after time-sync.target has been
b2ca0d63
LP
4622 reached. This way they will not elapse before the system
4623 clock has been corrected by a local NTP client or
4624 similar. This is particular useful on RTC-less embedded
4625 machines, that come up with an invalid system clock.
4626
4627 * systemd-nspawn's --network-veth= switch should now result in
4628 stable MAC addresses for both the outer and the inner side
4629 of the link.
4630
4631 * systemd-nspawn gained a new --volatile= switch for running
4632 container instances with /etc or /var unpopulated.
4633
4634 * The kdbus client code has been updated to use the new Linux
4635 3.17 memfd subsystem instead of the old kdbus-specific one.
4636
4637 * systemd-networkd's DHCP client and server now support
01da80b1
LP
4638 FORCERENEW. There are also new configuration options to
4639 configure the vendor client identifier and broadcast mode
4640 for DHCP.
b2ca0d63
LP
4641
4642 * systemd will no longer inform the kernel about the current
4643 timezone, as this is necessarily incorrect and racy as the
4644 kernel has no understanding of DST and similar
4645 concepts. This hence means FAT timestamps will be always
4646 considered UTC, similar to what Android is already
4647 doing. Also, when the RTC is configured to the local time
4648 (rather than UTC) systemd will never synchronize back to it,
4649 as this might confuse Windows at a later boot.
4650
4651 * systemd-analyze gained a new command "verify" for offline
4652 validation of unit files.
4653
4654 * systemd-networkd gained support for a couple of additional
4655 settings for bonding networking setups. Also, the metric for
4656 statically configured routes may now be configured. For
4657 network interfaces where this is appropriate the peer IP
4658 address may now be configured.
4659
26568403
TG
4660 * systemd-networkd's DHCP client will no longer request
4661 broadcasting by default, as this tripped up some networks.
4662 For hardware where broadcast is required the feature should
4663 be switched back on using RequestBroadcast=yes.
4664
4665 * systemd-networkd will now set up IPv4LL addresses (when
4666 enabled) even if DHCP is configured successfully.
4667
4668 * udev will now default to respect network device names given
4669 by the kernel when the kernel indicates that these are
4670 predictable. This behavior can be tweaked by changing
4671 NamePolicy= in the relevant .link file.
4672
b2ca0d63
LP
4673 * A new library systemd-terminal has been added that
4674 implements full TTY stream parsing and rendering. This
4675 library is supposed to be used later on for implementing a
4676 full userspace VT subsystem, replacing the current kernel
4677 implementation.
4678
4679 * A new tool systemd-journal-upload has been added to push
4680 journal data to a remote system running
4681 systemd-journal-remote.
4682
4683 * journald will no longer forward all local data to another
4684 running syslog daemon. This change has been made because
4685 rsyslog (which appears to be the most commonly used syslog
4686 implementation these days) no longer makes use of this, and
4687 instead pulls the data out of the journal on its own. Since
5f02e26c 4688 forwarding the messages to a non-existent syslog server is
b2ca0d63
LP
4689 more expensive than we assumed we have now turned this
4690 off. If you run a syslog server that is not a recent rsyslog
4691 version, you have to turn this option on again
4692 (ForwardToSyslog= in journald.conf).
4693
4694 * journald now optionally supports the LZ4 compressor for
4695 larger journal fields. This compressor should perform much
4696 better than XZ which was the previous default.
4697
4698 * machinectl now shows the IP addresses of local containers,
4699 if it knows them, plus the interface name of the container.
4700
4701 * A new tool "systemd-escape" has been added that makes it
4702 easy to escape strings to build unit names and similar.
4703
4704 * sd_notify() messages may now include a new ERRNO= field
4705 which is parsed and collected by systemd and shown among the
4706 "systemctl status" output for a service.
4707
4708 * A new component "systemd-firstboot" has been added that
4709 queries the most basic systemd information (timezone,
a1a4a25e 4710 hostname, root password) interactively on first
b2ca0d63
LP
4711 boot. Alternatively it may also be used to provision these
4712 things offline on OS images installed into directories.
4713
01da80b1
LP
4714 * The default sysctl.d/ snippets will now set
4715
4716 net.ipv4.conf.default.promote_secondaries=1
4717
4718 This has the benefit of no flushing secondary IP addresses
4719 when primary addresses are removed.
4720
b2ca0d63
LP
4721 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
4722 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
4723 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
4724 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
4725 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
4726 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
4727 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
4728 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
4729 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
4730 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
4731 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
4732 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
4733 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
4734 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
4735 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
4736
ccddd104 4737 — Berlin, 2014-08-19
b72ddf0f 4738
3dff3e00 4739CHANGES WITH 215:
24a2bf4c
LP
4740
4741 * A new tool systemd-sysusers has been added. This tool
4742 creates system users and groups in /etc/passwd and
4743 /etc/group, based on static declarative system user/group
4744 definitions in /usr/lib/sysusers.d/. This is useful to
4745 enable factory resets and volatile systems that boot up with
4746 an empty /etc directory, and thus need system users and
4747 groups created during early boot. systemd now also ships
4748 with two default sysusers.d/ files for the most basic
4749 users and groups systemd and the core operating system
4750 require.
4751
4752 * A new tmpfiles snippet has been added that rebuilds the
4753 essential files in /etc on boot, should they be missing.
4754
4755 * A directive for ensuring automatic clean-up of
4756 /var/cache/man/ has been removed from the default
4757 configuration. This line should now be shipped by the man
4758 implementation. The necessary change has been made to the
4759 man-db implementation. Note that you need to update your man
4760 implementation to one that ships this line, otherwise no
4761 automatic clean-up of /var/cache/man will take place.
4762
4763 * A new condition ConditionNeedsUpdate= has been added that
4764 may conditionalize services to only run when /etc or /var
4765 are "older" than the vendor operating system resources in
4766 /usr. This is useful for reconstructing or updating /etc
4767 after an offline update of /usr or a factory reset, on the
4768 next reboot. Services that want to run once after such an
4769 update or reset should use this condition and order
4770 themselves before the new systemd-update-done.service, which
4771 will mark the two directories as fully updated. A number of
4772 service files have been added making use of this, to rebuild
4773 the udev hardware database, the journald message catalog and
4774 dynamic loader cache (ldconfig). The systemd-sysusers tool
4775 described above also makes use of this now. With this in
4776 place it is now possible to start up a minimal operating
ce1dde29 4777 system with /etc empty cleanly. For more information on the
24a2bf4c
LP
4778 concepts involved see this recent blog story:
4779
4780 http://0pointer.de/blog/projects/stateless.html
4781
4782 * A new system group "input" has been introduced, and all
4783 input device nodes get this group assigned. This is useful
4784 for system-level software to get access to input devices. It
3dff3e00
KS
4785 complements what is already done for "audio" and "video".
4786
24a2bf4c
LP
4787 * systemd-networkd learnt minimal DHCPv4 server support in
4788 addition to the existing DHCPv4 client support. It also
4789 learnt DHCPv6 client and IPv6 Router Solicitation client
4790 support. The DHCPv4 client gained support for static routes
4791 passed in from the server. Note that the [DHCPv4] section
4792 known in older systemd-networkd versions has been renamed to
4793 [DHCP] and is now also used by the DHCPv6 client. Existing
c7435cc9
LP
4794 .network files using settings of this section should be
4795 updated, though compatibility is maintained. Optionally, the
4796 client hostname may now be sent to the DHCP server.
24a2bf4c 4797
c7435cc9
LP
4798 * networkd gained support for vxlan virtual networks as well
4799 as tun/tap and dummy devices.
24a2bf4c
LP
4800
4801 * networkd gained support for automatic allocation of address
4802 ranges for interfaces from a system-wide pool of
4803 addresses. This is useful for dynamically managing a large
4804 number of interfaces with a single network configuration
4805 file. In particular this is useful to easily assign
4806 appropriate IP addresses to the veth links of a large number
4807 of nspawn instances.
4808
4809 * RPM macros for processing sysusers, sysctl and binfmt
4810 drop-in snippets at package installation time have been
4811 added.
4812
4813 * The /etc/os-release file should now be placed in
4814 /usr/lib/os-release. The old location is automatically
4815 created as symlink. /usr/lib is the more appropriate
4816 location of this file, since it shall actually describe the
4817 vendor operating system shipped in /usr, and not the
4818 configuration stored in /etc.
4819
4820 * .mount units gained a new boolean SloppyOptions= setting
4821 that maps to mount(8)'s -s option which enables permissive
4822 parsing of unknown mount options.
4823
4824 * tmpfiles learnt a new "L+" directive which creates a symlink
4825 but (unlike "L") deletes a pre-existing file first, should
4826 it already exist and not already be the correct
a8eaaee7 4827 symlink. Similarly, "b+", "c+" and "p+" directives have been
24a2bf4c
LP
4828 added as well, which create block and character devices, as
4829 well as fifos in the filesystem, possibly removing any
4830 pre-existing files of different types.
4831
4832 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
4833 'argument' field (which so far specified the source to
ce1dde29 4834 symlink/copy the files from) is now optional. If omitted the
24a2bf4c
LP
4835 same file os copied from /usr/share/factory/ suffixed by the
4836 full destination path. This is useful for populating /etc
4837 with essential files, by copying them from vendor defaults
4838 shipped in /usr/share/factory/etc.
4839
4840 * A new command "systemctl preset-all" has been added that
4841 applies the service preset settings to all installed unit
4842 files. A new switch --preset-mode= has been added that
4843 controls whether only enable or only disable operations
4844 shall be executed.
4845
4846 * A new command "systemctl is-system-running" has been added
4847 that allows checking the overall state of the system, for
ce1dde29 4848 example whether it is fully up and running.
24a2bf4c
LP
4849
4850 * When the system boots up with an empty /etc, the equivalent
4851 to "systemctl preset-all" is executed during early boot, to
4852 make sure all default services are enabled after a factory
4853 reset.
4854
4855 * systemd now contains a minimal preset file that enables the
4856 most basic services systemd ships by default.
4857
4858 * Unit files' [Install] section gained a new DefaultInstance=
4859 field for defining the default instance to create if a
4860 template unit is enabled with no instance specified.
4861
4862 * A new passive target cryptsetup-pre.target has been added
4863 that may be used by services that need to make they run and
4864 finish before the first LUKS cryptographic device is set up.
4865
4866 * The /dev/loop-control and /dev/btrfs-control device nodes
4867 are now owned by the "disk" group by default, opening up
4868 access to this group.
4869
4870 * systemd-coredump will now automatically generate a
4871 stack trace of all core dumps taking place on the system,
4872 based on elfutils' libdw library. This stack trace is logged
4873 to the journal.
4874
4875 * systemd-coredump may now optionally store coredumps directly
4876 on disk (in /var/lib/systemd/coredump, possibly compressed),
4877 instead of storing them unconditionally in the journal. This
4878 mode is the new default. A new configuration file
4879 /etc/systemd/coredump.conf has been added to configure this
4880 and other parameters of systemd-coredump.
4881
4882 * coredumpctl gained a new "info" verb to show details about a
4883 specific coredump. A new switch "-1" has also been added
4884 that makes sure to only show information about the most
4885 recent entry instead of all entries. Also, as the tool is
4886 generally useful now the "systemd-" prefix of the binary
4887 name has been removed. Distributions that want to maintain
4888 compatibility with the old name should add a symlink from
4889 the old name to the new name.
4890
4891 * journald's SplitMode= now defaults to "uid". This makes sure
ce1dde29 4892 that unprivileged users can access their own coredumps with
24a2bf4c
LP
4893 coredumpctl without restrictions.
4894
4895 * New kernel command line options "systemd.wants=" (for
4896 pulling an additional unit during boot), "systemd.mask="
4897 (for masking a specific unit for the boot), and
4898 "systemd.debug-shell" (for enabling the debug shell on tty9)
4899 have been added. This is implemented in the new generator
4900 "systemd-debug-generator".
4901
4902 * systemd-nspawn will now by default filter a couple of
4903 syscalls for containers, among them those required for
4904 kernel module loading, direct x86 IO port access, swap
4905 management, and kexec. Most importantly though
4906 open_by_handle_at() is now prohibited for containers,
4907 closing a hole similar to a recently discussed vulnerability
4908 in docker regarding access to files on file hierarchies the
b938cb90
JE
4909 container should normally not have access to. Note that, for
4910 nspawn, we generally make no security claims anyway (and
24a2bf4c
LP
4911 this is explicitly documented in the man page), so this is
4912 just a fix for one of the most obvious problems.
4913
4914 * A new man page file-hierarchy(7) has been added that
4915 contains a minimized, modernized version of the file system
4916 layout systemd expects, similar in style to the FHS
c7435cc9
LP
4917 specification or hier(5). A new tool systemd-path(1) has
4918 been added to query many of these paths for the local
4919 machine and user.
24a2bf4c
LP
4920
4921 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
4922 longer done. Since the directory now has a per-user size
4923 limit, and is cleaned on logout this appears unnecessary,
4924 in particular since this now brings the lifecycle of this
4925 directory closer in line with how IPC objects are handled.
4926
4927 * systemd.pc now exports a number of additional directories,
4928 including $libdir (which is useful to identify the library
4929 path for the primary architecture of the system), and a
4930 couple of drop-in directories.
4931
3058e017
TLSC
4932 * udev's predictable network interface names now use the dev_port
4933 sysfs attribute, introduced in linux 3.15 instead of dev_id to
4934 distinguish between ports of the same PCI function. dev_id should
4935 only be used for ports using the same HW address, hence the need
4936 for dev_port.
4937
c7435cc9
LP
4938 * machined has been updated to export the OS version of a
4939 container (read from /etc/os-release and
4940 /usr/lib/os-release) on the bus. This is now shown in
4941 "machinectl status" for a machine.
4942
4943 * A new service setting RestartForceExitStatus= has been
4944 added. If configured to a set of exit signals or process
4945 return values, the service will be restarted when the main
4946 daemon process exits with any of them, regardless of the
4947 Restart= setting.
4948
4949 * systemctl's -H switch for connecting to remote systemd
4950 machines has been extended so that it may be used to
4951 directly connect to a specific container on the
4952 host. "systemctl -H root@foobar:waldi" will now connect as
4953 user "root" to host "foobar", and then proceed directly to
4954 the container named "waldi". Note that currently you have to
4955 authenticate as user "root" for this to work, as entering
4956 containers is a privileged operation.
4957
4958 Contributions from: Andreas Henriksson, Benjamin Steinwender,
4959 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
4960 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
4961 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
4962 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
4963 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
4964 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
4965 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
4966 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
4967 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
4968 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
4969 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
4970
ccddd104 4971 — Berlin, 2014-07-03
c7435cc9 4972
4196a3ea
KS
4973CHANGES WITH 214:
4974
4975 * As an experimental feature, udev now tries to lock the
4976 disk device node (flock(LOCK_SH|LOCK_NB)) while it
4977 executes events for the disk or any of its partitions.
4978 Applications like partitioning programs can lock the
4979 disk device node (flock(LOCK_EX)) and claim temporary
4980 device ownership that way; udev will entirely skip all event
4981 handling for this disk and its partitions. If the disk
4982 was opened for writing, the close will trigger a partition
4983 table rescan in udev's "watch" facility, and if needed
71449caf 4984 synthesize "change" events for the disk and all its partitions.
8d0e0ddd 4985 This is now unconditionally enabled, and if it turns out to
4196a3ea 4986 cause major problems, we might turn it on only for specific
45df8656 4987 devices, or might need to disable it entirely. Device Mapper
4196a3ea
KS
4988 devices are excluded from this logic.
4989
04e91da2
LP
4990 * We temporarily dropped the "-l" switch for fsck invocations,
4991 since they collide with the flock() logic above. util-linux
4992 upstream has been changed already to avoid this conflict,
4993 and we will readd "-l" as soon as util-linux with this
4994 change has been released.
4995
4996 * The dependency on libattr has been removed. Since a long
8d0e0ddd 4997 time, the extended attribute calls have moved to glibc, and
04e91da2
LP
4998 libattr is thus unnecessary.
4999
ce830873 5000 * Virtualization detection works without privileges now. This
04e91da2
LP
5001 means the systemd-detect-virt binary no longer requires
5002 CAP_SYS_PTRACE file capabilities, and our daemons can run
71449caf 5003 with fewer privileges.
04e91da2
LP
5004
5005 * systemd-networkd now runs under its own "systemd-network"
5006 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
5007 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
5008 loses the ability to write to files owned by root this way.
5009
a8eaaee7 5010 * Similarly, systemd-resolved now runs under its own
04e91da2
LP
5011 "systemd-resolve" user with no capabilities remaining.
5012
a8eaaee7 5013 * Similarly, systemd-bus-proxyd now runs under its own
04e91da2
LP
5014 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
5015
5016 * systemd-networkd gained support for setting up "veth"
a8eaaee7 5017 virtual Ethernet devices for container connectivity, as well
04e91da2
LP
5018 as GRE and VTI tunnels.
5019
5020 * systemd-networkd will no longer automatically attempt to
5021 manually load kernel modules necessary for certain tunnel
8d0e0ddd 5022 transports. Instead, it is assumed the kernel loads them
04e91da2
LP
5023 automatically when required. This only works correctly on
5024 very new kernels. On older kernels, please consider adding
c54bed5d 5025 the kernel modules to /etc/modules-load.d/ as a work-around.
04e91da2 5026
cd14eda3 5027 * The resolv.conf file systemd-resolved generates has been
8d0e0ddd
JE
5028 moved to /run/systemd/resolve/. If you have a symlink from
5029 /etc/resolv.conf, it might be necessary to correct it.
cd14eda3 5030
ef392da6 5031 * Two new service settings, ProtectHome= and ProtectSystem=,
8d0e0ddd 5032 have been added. When enabled, they will make the user data
04e91da2
LP
5033 (such as /home) inaccessible or read-only and the system
5034 (such as /usr) read-only, for specific services. This allows
5035 very light-weight per-service sandboxing to avoid
5036 modifications of user data or system files from
5037 services. These two new switches have been enabled for all
5038 of systemd's long-running services, where appropriate.
5039
5040 * Socket units gained new SocketUser= and SocketGroup=
5041 settings to set the owner user and group of AF_UNIX sockets
5042 and FIFOs in the file system.
5043
8d0e0ddd 5044 * Socket units gained a new RemoveOnStop= setting. If enabled,
04e91da2
LP
5045 all FIFOS and sockets in the file system will be removed
5046 when the specific socket unit is stopped.
5047
5048 * Socket units gained a new Symlinks= setting. It takes a list
5049 of symlinks to create to file system sockets or FIFOs
45df8656 5050 created by the specific Unix sockets. This is useful to
71449caf 5051 manage symlinks to socket nodes with the same life-cycle as
04e91da2
LP
5052 the socket itself.
5053
5054 * The /dev/log socket and /dev/initctl FIFO have been moved to
5055 /run, and have been replaced by symlinks. This allows
5056 connecting to these facilities even if PrivateDevices=yes is
5057 used for a service (which makes /dev/log itself unavailable,
5058 but /run is left). This also has the benefit of ensuring
5059 that /dev only contains device nodes, directories and
5060 symlinks, and nothing else.
5061
5062 * sd-daemon gained two new calls sd_pid_notify() and
5063 sd_pid_notifyf(). They are similar to sd_notify() and
5064 sd_notifyf(), but allow overriding of the source PID of
5065 notification messages if permissions permit this. This is
5066 useful to send notify messages on behalf of a different
5067 process (for example, the parent process). The
5068 systemd-notify tool has been updated to make use of this
5069 when sending messages (so that notification messages now
5070 originate from the shell script invoking systemd-notify and
5071 not the systemd-notify process itself. This should minimize
5072 a race where systemd fails to associate notification
5073 messages to services when the originating process already
5074 vanished.
5075
5076 * A new "on-abnormal" setting for Restart= has been added. If
8d0e0ddd 5077 set, it will result in automatic restarts on all "abnormal"
04e91da2
LP
5078 reasons for a process to exit, which includes unclean
5079 signals, core dumps, timeouts and watchdog timeouts, but
5080 does not include clean and unclean exit codes or clean
5081 signals. Restart=on-abnormal is an alternative for
5082 Restart=on-failure for services that shall be able to
5083 terminate and avoid restarts on certain errors, by
5084 indicating so with an unclean exit code. Restart=on-failure
5085 or Restart=on-abnormal is now the recommended setting for
5086 all long-running services.
5087
5088 * If the InaccessibleDirectories= service setting points to a
5089 mount point (or if there are any submounts contained within
5090 it), it is now attempted to completely unmount it, to make
5091 the file systems truly unavailable for the respective
5092 service.
5093
5094 * The ReadOnlyDirectories= service setting and
5095 systemd-nspawn's --read-only parameter are now recursively
5096 applied to all submounts, too.
5097
5098 * Mount units may now be created transiently via the bus APIs.
5099
5100 * The support for SysV and LSB init scripts has been removed
5101 from the systemd daemon itself. Instead, it is now
5102 implemented as a generator that creates native systemd units
5103 from these scripts when needed. This enables us to remove a
5104 substantial amount of legacy code from PID 1, following the
5105 fact that many distributions only ship a very small number
5106 of LSB/SysV init scripts nowadays.
5107
cc98b302 5108 * Privileged Xen (dom0) domains are not considered
04e91da2
LP
5109 virtualization anymore by the virtualization detection
5110 logic. After all, they generally have unrestricted access to
71449caf 5111 the hardware and usually are used to manage the unprivileged
04e91da2
LP
5112 (domU) domains.
5113
5114 * systemd-tmpfiles gained a new "C" line type, for copying
5115 files or entire directories.
5116
5117 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
8d0e0ddd
JE
5118 lines. So far, they have been non-globbing versions of the
5119 latter, and have thus been redundant. In future, it is
5120 recommended to only use "z". "m" has hence been removed
04e91da2
LP
5121 from the documentation, even though it stays supported.
5122
5123 * A tmpfiles snippet to recreate the most basic structure in
5124 /var has been added. This is enough to create the /var/run →
5125 /run symlink and create a couple of structural
5126 directories. This allows systems to boot up with an empty or
8d0e0ddd
JE
5127 volatile /var. Of course, while with this change, the core OS
5128 now is capable with dealing with a volatile /var, not all
04e91da2 5129 user services are ready for it. However, we hope that sooner
8d0e0ddd 5130 or later, many service daemons will be changed upstream so
04e91da2
LP
5131 that they are able to automatically create their necessary
5132 directories in /var at boot, should they be missing. This is
5133 the first step to allow state-less systems that only require
5134 the vendor image for /usr to boot.
5135
5136 * systemd-nspawn has gained a new --tmpfs= switch to mount an
5137 empty tmpfs instance to a specific directory. This is
5138 particularly useful for making use of the automatic
5139 reconstruction of /var (see above), by passing --tmpfs=/var.
5140
5141 * Access modes specified in tmpfiles snippets may now be
5142 prefixed with "~", which indicates that they shall be masked
daa05349 5143 by whether the existing file or directory is currently
8d0e0ddd 5144 writable, readable or executable at all. Also, if specified,
04e91da2
LP
5145 the sgid/suid/sticky bits will be masked for all
5146 non-directories.
5147
5148 * A new passive target unit "network-pre.target" has been
5149 added which is useful for services that shall run before any
5150 network is configured, for example firewall scripts.
5151
4c0d13bd
LP
5152 * The "floppy" group that previously owned the /dev/fd*
5153 devices is no longer used. The "disk" group is now used
5154 instead. Distributions should probably deprecate usage of
5155 this group.
5156
dc1d6c02
LP
5157 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
5158 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
5159 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
5160 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
5161 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
5162 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
5163 Jędrzejewski-Szmek
5164
ccddd104 5165 — Berlin, 2014-06-11
dc1d6c02 5166
6936cd89
LP
5167CHANGES WITH 213:
5168
5169 * A new "systemd-timesyncd" daemon has been added for
69beda1f 5170 synchronizing the system clock across the network. It
6936cd89 5171 implements an SNTP client. In contrast to NTP
8d0e0ddd 5172 implementations such as chrony or the NTP reference server,
6936cd89 5173 this only implements a client side, and does not bother with
c9679c65
LP
5174 the full NTP complexity, focusing only on querying time from
5175 one remote server and synchronizing the local clock to
6936cd89 5176 it. Unless you intend to serve NTP to networked clients or
8d0e0ddd 5177 want to connect to local hardware clocks, this simple NTP
6936cd89
LP
5178 client should be more than appropriate for most
5179 installations. The daemon runs with minimal privileges, and
5180 has been hooked up with networkd to only operate when
5181 network connectivity is available. The daemon saves the
5182 current clock to disk every time a new NTP sync has been
5183 acquired, and uses this to possibly correct the system clock
69beda1f 5184 early at bootup, in order to accommodate for systems that
6936cd89 5185 lack an RTC such as the Raspberry Pi and embedded devices,
8d0e0ddd 5186 and to make sure that time monotonically progresses on these
c9679c65 5187 systems, even if it is not always correct. To make use of
8d0e0ddd 5188 this daemon, a new system user and group "systemd-timesync"
c9679c65 5189 needs to be created on installation of systemd.
6936cd89 5190
69beda1f
KS
5191 * The queue "seqnum" interface of libudev has been disabled, as
5192 it was generally incompatible with device namespacing as
6936cd89
LP
5193 sequence numbers of devices go "missing" if the devices are
5194 part of a different namespace.
5195
5196 * "systemctl list-timers" and "systemctl list-sockets" gained
5197 a --recursive switch for showing units of these types also
499b604b
ZJS
5198 for all local containers, similar in style to the already
5199 supported --recursive switch for "systemctl list-units".
6936cd89
LP
5200
5201 * A new RebootArgument= setting has been added for service
5202 units, which may be used to specify a kernel reboot argument
499b604b 5203 to use when triggering reboots with StartLimitAction=.
6936cd89
LP
5204
5205 * A new FailureAction= setting has been added for service
5206 units which may be used to specify an operation to trigger
499b604b 5207 when a service fails. This works similarly to
8d0e0ddd 5208 StartLimitAction=, but unlike it, controls what is done
6936cd89
LP
5209 immediately rather than only after several attempts to
5210 restart the service in question.
5211
5212 * hostnamed got updated to also expose the kernel name,
499b604b
ZJS
5213 release, and version on the bus. This is useful for
5214 executing commands like hostnamectl with the -H switch.
5215 systemd-analyze makes use of this to properly display
5216 details when running non-locally.
6936cd89
LP
5217
5218 * The bootchart tool can now show cgroup information in the
5219 graphs it generates.
5220
5221 * The CFS CPU quota cgroup attribute is now exposed for
5222 services. The new CPUQuota= switch has been added for this
5223 which takes a percentage value. Setting this will have the
5224 result that a service may never get more CPU time than the
5225 specified percentage, even if the machine is otherwise idle.
5226
5227 * systemd-networkd learned IPIP and SIT tunnel support.
5228
5229 * LSB init scripts exposing a dependency on $network will now
5230 get a dependency on network-online.target rather than simply
5231 network.target. This should bring LSB handling closer to
5232 what it was on SysV systems.
5233
5234 * A new fsck.repair= kernel option has been added to control
5235 how fsck shall deal with unclean file systems at boot.
5236
5237 * The (.ini) configuration file parser will now silently
5238 ignore sections whose name begins with "X-". This may be
5239 used to maintain application-specific extension sections in unit
5240 files.
5241
5242 * machined gained a new API to query the IP addresses of
5243 registered containers. "machinectl status" has been updated
5244 to show these addresses in its output.
5245
5246 * A new call sd_uid_get_display() has been added to the
5247 sd-login APIs for querying the "primary" session of a
5248 user. The "primary" session of the user is elected from the
5249 user's sessions and generally a graphical session is
5250 preferred over a text one.
5251
5252 * A minimal systemd-resolved daemon has been added. It
5253 currently simply acts as a companion to systemd-networkd and
5254 manages resolv.conf based on per-interface DNS
5255 configuration, possibly supplied via DHCP. In the long run
5256 we hope to extend this into a local DNSSEC enabled DNS and
5257 mDNS cache.
5258
68dd0956
TG
5259 * The systemd-networkd-wait-online tool is now enabled by
5260 default. It will delay network-online.target until a network
5261 connection has been configured. The tool primarily integrates
5262 with networkd, but will also make a best effort to make sense
5263 of network configuration performed in some other way.
5264
6936cd89 5265 * Two new service options StartupCPUShares= and
499b604b 5266 StartupBlockIOWeight= have been added that work similarly to
6936cd89 5267 CPUShares= and BlockIOWeight= however only apply during
69beda1f 5268 system startup. This is useful to prioritize certain services
6936cd89
LP
5269 differently during bootup than during normal runtime.
5270
8e7acf67
LP
5271 * hostnamed has been changed to prefer the statically
5272 configured hostname in /etc/hostname (unless set to
5273 'localhost' or empty) over any dynamic one supplied by
8d0e0ddd 5274 dhcp. With this change, the rules for picking the hostname
8e7acf67
LP
5275 match more closely the rules of other configuration settings
5276 where the local administrator's configuration in /etc always
5277 overrides any other settings.
5278
5279 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
6936cd89
LP
5280 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
5281 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
5282 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
5283 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
5284 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
5285 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
5286 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
5287 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
8e7acf67
LP
5288 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
5289 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
5290 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
5291 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
5292 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
5293 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
5294 Lindskog, WaLyong Cho, Will Woods, Zbigniew
6936cd89
LP
5295 Jędrzejewski-Szmek
5296
ccddd104 5297 — Beijing, 2014-05-28
6936cd89 5298
51c61cda
LP
5299CHANGES WITH 212:
5300
5301 * When restoring the screen brightness at boot, stay away from
5302 the darkest setting or from the lowest 5% of the available
5303 range, depending on which is the larger value of both. This
5304 should effectively protect the user from rebooting into a
5305 black screen, should the brightness have been set to minimum
5306 by accident.
5307
5308 * sd-login gained a new sd_machine_get_class() call to
5309 determine the class ("vm" or "container") of a machine
5310 registered with machined.
5311
5312 * sd-login gained new calls
5313 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
5314 to query the identity of the peer of a local AF_UNIX
499b604b 5315 connection. They operate similarly to their sd_pid_get_xyz()
51c61cda
LP
5316 counterparts.
5317
5318 * PID 1 will now maintain a system-wide system state engine
5319 with the states "starting", "running", "degraded",
5320 "maintenance", "stopping". These states are bound to system
5321 startup, normal runtime, runtime with at least one failed
5322 service, rescue/emergency mode and system shutdown. This
5323 state is shown in the "systemctl status" output when no unit
5324 name is passed. It is useful to determine system state, in
5325 particularly when doing so for many systems or containers at
5326 once.
5327
5328 * A new command "list-machines" has been added to "systemctl"
5329 that lists all local OS containers and shows their system
5330 state (see above), if systemd runs inside of them.
5331
5332 * systemctl gained a new "-r" switch to recursively enumerate
5333 units on all local containers, when used with the
5334 "list-unit" command (which is the default one that is
5335 executed when no parameters are specified).
5336
5337 * The GPT automatic partition discovery logic will now honour
5338 two GPT partition flags: one may be set on a partition to
5339 cause it to be mounted read-only, and the other may be set
5340 on a partition to ignore it during automatic discovery.
5341
5342 * Two new GPT type UUIDs have been added for automatic root
70a44afe 5343 partition discovery, for 32-bit and 64-bit ARM. This is not
51c61cda
LP
5344 particularly useful for discovering the root directory on
5345 these architectures during bare-metal boots (since UEFI is
5346 not common there), but still very useful to allow booting of
5347 ARM disk images in nspawn with the -i option.
5348
5349 * MAC addresses of interfaces created with nspawn's
5350 --network-interface= switch will now be generated from the
5351 machine name, and thus be stable between multiple invocations
5352 of the container.
5353
5354 * logind will now automatically remove all IPC objects owned
5355 by a user if she or he fully logs out. This makes sure that
5356 users who are logged out cannot continue to consume IPC
5357 resources. This covers SysV memory, semaphores and message
5358 queues as well as POSIX shared memory and message
b8bde116
JE
5359 queues. Traditionally, SysV and POSIX IPC had no life-cycle
5360 limits. With this functionality, that is corrected. This may
5361 be turned off by using the RemoveIPC= switch of logind.conf.
51c61cda
LP
5362
5363 * The systemd-machine-id-setup and tmpfiles tools gained a
5364 --root= switch to operate on a specific root directory,
5365 instead of /.
5366
5367 * journald can now forward logged messages to the TTYs of all
5368 logged in users ("wall"). This is the default for all
5369 emergency messages now.
5370
5371 * A new tool systemd-journal-remote has been added to stream
5372 journal log messages across the network.
5373
5374 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
5375 controller trees are mounted into it. Note that the
5376 directories mounted beneath it are not read-only. This is a
5377 security measure and is particularly useful because glibc
5378 actually includes a search logic to pick any tmpfs it can
5379 find to implement shm_open() if /dev/shm is not available
5380 (which it might very well be in namespaced setups).
5381
5382 * machinectl gained a new "poweroff" command to cleanly power
5383 down a local OS container.
5384
5385 * The PrivateDevices= unit file setting will now also drop the
5386 CAP_MKNOD capability from the capability bound set, and
5387 imply DevicePolicy=closed.
5388
5389 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
5390 comprehensively on all long-running systemd services where
5391 this is appropriate.
5392
5393 * systemd-udevd will now run in a disassociated mount
b8bde116 5394 namespace. To mount directories from udev rules, make sure to
51c61cda
LP
5395 pull in mount units via SYSTEMD_WANTS properties.
5396
5397 * The kdbus support gained support for uploading policy into
5398 the kernel. sd-bus gained support for creating "monitoring"
5399 connections that can eavesdrop into all bus communication
5400 for debugging purposes.
5401
5402 * Timestamps may now be specified in seconds since the UNIX
5403 epoch Jan 1st, 1970 by specifying "@" followed by the value
5404 in seconds.
5405
5406 * Native tcpwrap support in systemd has been removed. tcpwrap
5407 is old code, not really maintained anymore and has serious
5408 shortcomings, and better options such as firewalls
5409 exist. For setups that require tcpwrap usage, please
5410 consider invoking your socket-activated service via tcpd,
5411 like on traditional inetd.
5412
5413 * A new system.conf configuration option
5414 DefaultTimerAccuracySec= has been added that controls the
5415 default AccuracySec= setting of .timer units.
5416
b8bde116 5417 * Timer units gained a new WakeSystem= switch. If enabled,
51c61cda
LP
5418 timers configured this way will cause the system to resume
5419 from system suspend (if the system supports that, which most
5420 do these days).
5421
b8bde116 5422 * Timer units gained a new Persistent= switch. If enabled,
51c61cda
LP
5423 timers configured this way will save to disk when they have
5424 been last triggered. This information is then used on next
5425 reboot to possible execute overdue timer events, that
d28315e4
JE
5426 could not take place because the system was powered off.
5427 This enables simple anacron-like behaviour for timer units.
51c61cda
LP
5428
5429 * systemctl's "list-timers" will now also list the time a
5430 timer unit was last triggered in addition to the next time
5431 it will be triggered.
5432
5433 * systemd-networkd will now assign predictable IPv4LL
5434 addresses to its local interfaces.
5435
5436 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
5437 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
5438 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
5439 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
5440 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
5441 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
5442 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
5443 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
5444 Jędrzejewski-Szmek
5445
ccddd104 5446 — Berlin, 2014-03-25
51c61cda 5447
699b6b34
LP
5448CHANGES WITH 211:
5449
5450 * A new unit file setting RestrictAddressFamilies= has been
5451 added to restrict which socket address families unit
5452 processes gain access to. This takes address family names
5453 like "AF_INET" or "AF_UNIX", and is useful to minimize the
5454 attack surface of services via exotic protocol stacks. This
5455 is built on seccomp system call filters.
5456
5457 * Two new unit file settings RuntimeDirectory= and
5458 RuntimeDirectoryMode= have been added that may be used to
5459 manage a per-daemon runtime directories below /run. This is
5460 an alternative for setting up directory permissions with
5461 tmpfiles snippets, and has the advantage that the runtime
5462 directory's lifetime is bound to the daemon runtime and that
5463 the daemon starts up with an empty directory each time. This
5464 is particularly useful when writing services that drop
f1721625 5465 privileges using the User= or Group= setting.
699b6b34
LP
5466
5467 * The DeviceAllow= unit setting now supports globbing for
5468 matching against device group names.
5469
5470 * The systemd configuration file system.conf gained new
5471 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
5472 DefaultMemoryAccounting= to globally turn on/off accounting
5473 for specific resources (cgroups) for all units. These
22e7062d 5474 settings may still be overridden individually in each unit
699b6b34
LP
5475 though.
5476
5477 * systemd-gpt-auto-generator is now able to discover /srv and
5478 root partitions in addition to /home and swap partitions. It
5479 also supports LUKS-encrypted partitions now. With this in
b8bde116 5480 place, automatic discovery of partitions to mount following
699b6b34 5481 the Discoverable Partitions Specification
56cadcb6 5482 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
699b6b34
LP
5483 is now a lot more complete. This allows booting without
5484 /etc/fstab and without root= on the kernel command line on
b8bde116 5485 systems prepared appropriately.
699b6b34
LP
5486
5487 * systemd-nspawn gained a new --image= switch which allows
5488 booting up disk images and Linux installations on any block
5489 device that follow the Discoverable Partitions Specification
5490 (see above). This means that installations made with
5491 appropriately updated installers may now be started and
5492 deployed using container managers, completely
5493 unmodified. (We hope that libvirt-lxc will add support for
5494 this feature soon, too.)
5495
5496 * systemd-nspawn gained a new --network-macvlan= setting to
5497 set up a private macvlan interface for the
499b604b 5498 container. Similarly, systemd-networkd gained a new
699b6b34
LP
5499 Kind=macvlan setting in .netdev files.
5500
5501 * systemd-networkd now supports configuring local addresses
5502 using IPv4LL.
5503
5504 * A new tool systemd-network-wait-online has been added to
5505 synchronously wait for network connectivity using
5506 systemd-networkd.
5507
5508 * The sd-bus.h bus API gained a new sd_bus_track object for
5509 tracking the life-cycle of bus peers. Note that sd-bus.h is
5510 still not a public API though (unless you specify
5511 --enable-kdbus on the configure command line, which however
5512 voids your warranty and you get no API stability guarantee).
5513
5514 * The $XDG_RUNTIME_DIR runtime directories for each user are
5515 now individual tmpfs instances, which has the benefit of
5516 introducing separate pools for each user, with individual
4ef6e535 5517 size limits, and thus making sure that unprivileged clients
699b6b34
LP
5518 can no longer negatively impact the system or other users by
5519 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
5520 RuntimeDirectorySize= has been introduced that allows
5521 controlling the default size limit for all users. It
5522 defaults to 10% of the available physical memory. This is no
5523 replacement for quotas on tmpfs though (which the kernel
5524 still does not support), as /dev/shm and /tmp are still
4ef6e535 5525 shared resources used by both the system and unprivileged
699b6b34
LP
5526 users.
5527
5528 * logind will now automatically turn off automatic suspending
5529 on laptop lid close when more than one display is
5530 connected. This was previously expected to be implemented
5531 individually in desktop environments (such as GNOME),
5532 however has been added to logind now, in order to fix a
5533 boot-time race where a desktop environment might not have
5534 been started yet and thus not been able to take an inhibitor
5535 lock at the time where logind already suspends the system
5536 due to a closed lid.
5537
5538 * logind will now wait at least 30s after each system
5539 suspend/resume cycle, and 3min after system boot before
5540 suspending the system due to a closed laptop lid. This
5541 should give USB docking stations and similar enough time to
4ef6e535 5542 be probed and configured after system resume and boot in
699b6b34
LP
5543 order to then act as suspend blocker.
5544
5545 * systemd-run gained a new --property= setting which allows
5546 initialization of resource control properties (and others)
5547 for the created scope or service unit. Example: "systemd-run
5548 --property=BlockIOWeight=10 updatedb" may be used to run
5549 updatedb at a low block IO scheduling weight.
5550
5551 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
5552 now also work in --scope mode.
5553
5554 * When systemd is compiled with kdbus support, basic support
5555 for enforced policies is now in place. (Note that enabling
5556 kdbus still voids your warranty and no API compatibility
5557 promises are made.)
5558
5559 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
5560 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
5561 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
5562 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
5563 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
5564 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
5565 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
5566 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
5567 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
5568 Zbigniew Jędrzejewski-Szmek
5569
ccddd104 5570 — Berlin, 2014-03-12
699b6b34 5571
43c71255
LP
5572CHANGES WITH 210:
5573
5574 * systemd will now relabel /dev after loading the SMACK policy
5575 according to SMACK rules.
5576
67dd87c5 5577 * A new unit file option AppArmorProfile= has been added to
43c71255
LP
5578 set the AppArmor profile for the processes of a unit.
5579
5580 * A new condition check ConditionArchitecture= has been added
5581 to conditionalize units based on the system architecture, as
5582 reported by uname()'s "machine" field.
5583
5584 * systemd-networkd now supports matching on the system
5585 virtualization, architecture, kernel command line, host name
5586 and machine ID.
5587
ed28905e 5588 * logind is now a lot more aggressive when suspending the
43c71255 5589 machine due to a closed laptop lid. Instead of acting only
b8bde116 5590 on the lid close action, it will continuously watch the lid
43c71255
LP
5591 status and act on it. This is useful for laptops where the
5592 power button is on the outside of the chassis so that it can
ed28905e 5593 be reached without opening the lid (such as the Lenovo
b8bde116 5594 Yoga). On those machines, logind will now immediately
ed28905e 5595 re-suspend the machine if the power button has been
43c71255
LP
5596 accidentally pressed while the laptop was suspended and in a
5597 backpack or similar.
5598
5599 * logind will now watch SW_DOCK switches and inhibit reaction
5600 to the lid switch if it is pressed. This means that logind
d27893ef 5601 will not suspend the machine anymore if the lid is closed
949138cc 5602 and the system is docked, if the laptop supports SW_DOCK
43c71255
LP
5603 notifications via the input layer. Note that ACPI docking
5604 stations do not generate this currently. Also note that this
5605 logic is usually not fully sufficient and Desktop
5606 Environments should take a lid switch inhibitor lock when an
5607 external display is connected, as systemd will not watch
5608 this on its own.
5609
5610 * nspawn will now make use of the devices cgroup controller by
5611 default, and only permit creation of and access to the usual
5612 API device nodes like /dev/null or /dev/random, as well as
5613 access to (but not creation of) the pty devices.
5614
5615 * We will now ship a default .network file for
5616 systemd-networkd that automatically configures DHCP for
5617 network interfaces created by nspawn's --network-veth or
5618 --network-bridge= switches.
5619
5620 * systemd will now understand the usual M, K, G, T suffixes
5621 according to SI conventions (i.e. to the base 1000) when
5622 referring to throughput and hardware metrics. It will stay
5623 with IEC conventions (i.e. to the base 1024) for software
5624 metrics, according to what is customary according to
5625 Wikipedia. We explicitly document which base applies for
5626 each configuration option.
5627
5628 * The DeviceAllow= setting in unit files now supports a syntax
ed28905e 5629 to whitelist an entire group of devices node majors at once,
43c71255 5630 based on the /proc/devices listing. For example, with the
b8bde116 5631 string "char-pts", it is now possible to whitelist all
43c71255
LP
5632 current and future pseudo-TTYs at once.
5633
5634 * sd-event learned a new "post" event source. Event sources of
5635 this type are triggered by the dispatching of any event
5636 source of a type that is not "post". This is useful for
5637 implementing clean-up and check event sources that are
5638 triggered by other work being done in the program.
5639
5640 * systemd-networkd is no longer statically enabled, but uses
5641 the usual [Install] sections so that it can be
5642 enabled/disabled using systemctl. It still is enabled by
5643 default however.
5644
b8bde116 5645 * When creating a veth interface pair with systemd-nspawn, the
43c71255
LP
5646 host side will now be prefixed with "vb-" if
5647 --network-bridge= is used, and with "ve-" if --network-veth
b8bde116 5648 is used. This way, it is easy to distinguish these cases on
43c71255
LP
5649 the host, for example to apply different configuration to
5650 them with systemd-networkd.
5651
d27893ef
LP
5652 * The compatibility libraries for libsystemd-journal.so,
5653 libsystem-id128.so, libsystemd-login.so and
5654 libsystemd-daemon.so do not make use of IFUNC
b8bde116 5655 anymore. Instead, we now build libsystemd.so multiple times
d27893ef
LP
5656 under these alternative names. This means that the footprint
5657 is drastically increased, but given that these are
b8bde116 5658 transitional compatibility libraries, this should not matter
d27893ef
LP
5659 much. This change has been made necessary to support the ARM
5660 platform for these compatibility libraries, as the ARM
d28315e4 5661 toolchain is not really at the same level as the toolchain
ed28905e 5662 for other architectures like x86 and does not support
d27893ef
LP
5663 IFUNC. Please make sure to use --enable-compat-libs only
5664 during a transitional period!
5665
13b28d82 5666 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
43c71255
LP
5667 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
5668 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
5669 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
5670 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
5671 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
5672 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
5673 Zbigniew Jędrzejewski-Szmek
5674
ccddd104 5675 — Berlin, 2014-02-24
43c71255 5676
e49b5aad
LP
5677CHANGES WITH 209:
5678
5679 * A new component "systemd-networkd" has been added that can
5680 be used to configure local network interfaces statically or
8b7d0494
JSJ
5681 via DHCP. It is capable of bringing up bridges, VLANs, and
5682 bonding. Currently, no hook-ups for interactive network
4670e9d5 5683 configuration are provided. Use this for your initrd,
8b7d0494
JSJ
5684 container, embedded, or server setup if you need a simple,
5685 yet powerful, network configuration solution. This
4670e9d5 5686 configuration subsystem is quite nifty, as it allows wildcard
1e190502 5687 hotplug matching in interfaces. For example, with a single
4670e9d5 5688 configuration snippet, you can configure that all Ethernet
1e190502
ZJS
5689 interfaces showing up are automatically added to a bridge,
5690 or similar. It supports link-sensing and more.
e49b5aad
LP
5691
5692 * A new tool "systemd-socket-proxyd" has been added which can
4c2413bf 5693 act as a bidirectional proxy for TCP sockets. This is
e49b5aad
LP
5694 useful for adding socket activation support to services that
5695 do not actually support socket activation, including virtual
4c2413bf 5696 machines and the like.
e49b5aad
LP
5697
5698 * Add a new tool to save/restore rfkill state on
5699 shutdown/boot.
5700
8b7d0494
JSJ
5701 * Save/restore state of keyboard backlights in addition to
5702 display backlights on shutdown/boot.
e49b5aad
LP
5703
5704 * udev learned a new SECLABEL{} construct to label device
5705 nodes with a specific security label when they appear. For
4c2413bf 5706 now, only SECLABEL{selinux} is supported, but the syntax is
e49b5aad
LP
5707 prepared for additional security frameworks.
5708
5709 * udev gained a new scheme to configure link-level attributes
5710 from files in /etc/systemd/network/*.link. These files can
8b7d0494 5711 match against MAC address, device path, driver name and type,
4c2413bf 5712 and will apply attributes like the naming policy, link speed,
8b7d0494 5713 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
e49b5aad
LP
5714 address assignment policy (randomized, ...).
5715
dfb08b05
ZJS
5716 * The configuration of network interface naming rules for
5717 "permanent interface names" has changed: a new NamePolicy=
5718 setting in the [Link] section of .link files determines the
a8eaaee7 5719 priority of possible naming schemes (onboard, slot, MAC,
dfb08b05
ZJS
5720 path). The default value of this setting is determined by
5721 /usr/lib/net/links/99-default.link. Old
5722 80-net-name-slot.rules udev configuration file has been
5723 removed, so local configuration overriding this file should
ce830873 5724 be adapted to override 99-default.link instead.
dfb08b05 5725
e49b5aad 5726 * When the User= switch is used in a unit file, also
4c2413bf 5727 initialize $SHELL= based on the user database entry.
e49b5aad
LP
5728
5729 * systemd no longer depends on libdbus. All communication is
5730 now done with sd-bus, systemd's low-level bus library
5731 implementation.
5732
5733 * kdbus support has been added to PID 1 itself. When kdbus is
4c2413bf 5734 enabled, this causes PID 1 to set up the system bus and
e49b5aad
LP
5735 enable support for a new ".busname" unit type that
5736 encapsulates bus name activation on kdbus. It works a little
5737 bit like ".socket" units, except for bus names. A new
5738 generator has been added that converts classic dbus1 service
5739 activation files automatically into native systemd .busname
5740 and .service units.
5741
5742 * sd-bus: add a light-weight vtable implementation that allows
5743 defining objects on the bus with a simple static const
5744 vtable array of its methods, signals and properties.
5745
8b7d0494 5746 * systemd will not generate or install static dbus
e49b5aad 5747 introspection data anymore to /usr/share/dbus-1/interfaces,
1e190502 5748 as the precise format of these files is unclear, and
e49b5aad
LP
5749 nothing makes use of it.
5750
5751 * A proxy daemon is now provided to proxy clients connecting
5752 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
5753 compatibility with classic D-Bus.
5754
5755 * A bus driver implementation has been added that supports the
5756 classic D-Bus bus driver calls on kdbus, also for
5757 compatibility purposes.
5758
5759 * A new API "sd-event.h" has been added that implements a
5760 minimal event loop API built around epoll. It provides a
5761 couple of features that direct epoll usage is lacking:
b9761003 5762 prioritization of events, scales to large numbers of timer
e49b5aad
LP
5763 events, per-event timer slack (accuracy), system-wide
5764 coalescing of timer events, exit handlers, watchdog
5765 supervision support using systemd's sd_notify() API, child
5766 process handling.
5767
5768 * A new API "sd-rntl.h" has been added that provides an API
5769 around the route netlink interface of the kernel, similar in
5770 style to "sd-bus.h".
5771
7e95eda5
PF
5772 * A new API "sd-dhcp-client.h" has been added that provides a
5773 small DHCPv4 client-side implementation. This is used by
e49b5aad
LP
5774 "systemd-networkd".
5775
4c2413bf 5776 * There is a new kernel command line option
8b7d0494
JSJ
5777 "systemd.restore_state=0|1". When set to "0", none of the
5778 systemd tools will restore saved runtime state to hardware
5779 devices. More specifically, the rfkill and backlight states
5780 are not restored.
e49b5aad
LP
5781
5782 * The FsckPassNo= compatibility option in mount/service units
5783 has been removed. The fstab generator will now add the
5784 necessary dependencies automatically, and does not require
5785 PID1's support for that anymore.
5786
8b7d0494 5787 * journalctl gained a new switch, --list-boots, that lists
e49b5aad
LP
5788 recent boots with their times and boot IDs.
5789
5790 * The various tools like systemctl, loginctl, timedatectl,
5791 busctl, systemd-run, ... have gained a new switch "-M" to
5792 connect to a specific, local OS container (as direct
5793 connection, without requiring SSH). This works on any
5794 container that is registered with machined, such as those
5795 created by libvirt-lxc or nspawn.
5796
5797 * systemd-run and systemd-analyze also gained support for "-H"
4c2413bf 5798 to connect to remote hosts via SSH. This is particularly
8b7d0494
JSJ
5799 useful for systemd-run because it enables queuing of jobs
5800 onto remote systems.
e49b5aad
LP
5801
5802 * machinectl gained a new command "login" to open a getty
5803 login in any local container. This works with any container
5804 that is registered with machined (such as those created by
8e420494 5805 libvirt-lxc or nspawn), and which runs systemd inside.
e49b5aad
LP
5806
5807 * machinectl gained a new "reboot" command that may be used to
5808 trigger a reboot on a specific container that is registered
5809 with machined. This works on any container that runs an init
5810 system of some kind.
5811
5812 * systemctl gained a new "list-timers" command to print a nice
5813 listing of installed timer units with the times they elapse
5814 next.
5815
5816 * Alternative reboot() parameters may now be specified on the
5817 "systemctl reboot" command line and are passed to the
5818 reboot() system call.
5819
5820 * systemctl gained a new --job-mode= switch to configure the
5821 mode to queue a job with. This is a more generic version of
8b7d0494 5822 --fail, --irreversible, and --ignore-dependencies, which are
e49b5aad
LP
5823 still available but not advertised anymore.
5824
e49b5aad
LP
5825 * /etc/systemd/system.conf gained new settings to configure
5826 various default timeouts of units, as well as the default
b9761003 5827 start limit interval and burst. These may still be overridden
e49b5aad
LP
5828 within each Unit.
5829
270f1624
LP
5830 * PID1 will now export on the bus profile data of the security
5831 policy upload process (such as the SELinux policy upload to
8e420494 5832 the kernel).
e49b5aad 5833
4670e9d5 5834 * journald: when forwarding logs to the console, include
1e190502
ZJS
5835 timestamps (following the setting in
5836 /sys/module/printk/parameters/time).
e49b5aad
LP
5837
5838 * OnCalendar= in timer units now understands the special
5839 strings "yearly" and "annually". (Both are equivalent)
5840
5841 * The accuracy of timer units is now configurable with the new
5842 AccuracySec= setting. It defaults to 1min.
5843
5844 * A new dependency type JoinsNamespaceOf= has been added that
5845 allows running two services within the same /tmp and network
5846 namespace, if PrivateNetwork= or PrivateTmp= are used.
5847
5848 * A new command "cat" has been added to systemctl. It outputs
5849 the original unit file of a unit, and concatenates the
1e190502
ZJS
5850 contents of additional "drop-in" unit file snippets, so that
5851 the full configuration is shown.
e49b5aad
LP
5852
5853 * systemctl now supports globbing on the various "list-xyz"
5854 commands, like "list-units" or "list-sockets", as well as on
1e190502
ZJS
5855 those commands which take multiple unit names.
5856
5857 * journalctl's --unit= switch gained support for globbing.
e49b5aad
LP
5858
5859 * All systemd daemons now make use of the watchdog logic so
5860 that systemd automatically notices when they hang.
5861
4c2413bf 5862 * If the $container_ttys environment variable is set,
e49b5aad
LP
5863 getty-generator will automatically spawn a getty for each
5864 listed tty. This is useful for container managers to request
5865 login gettys to be spawned on as many ttys as needed.
5866
5867 * %h, %s, %U specifier support is not available anymore when
5868 used in unit files for PID 1. This is because NSS calls are
5869 not safe from PID 1. They stay available for --user
5870 instances of systemd, and as special case for the root user.
5871
e49b5aad
LP
5872 * loginctl gained a new "--no-legend" switch to turn off output
5873 of the legend text.
5874
5875 * The "sd-login.h" API gained three new calls:
5876 sd_session_is_remote(), sd_session_get_remote_user(),
5877 sd_session_get_remote_host() to query information about
5878 remote sessions.
5879
8e420494
LP
5880 * The udev hardware database now also carries vendor/product
5881 information of SDIO devices.
e49b5aad
LP
5882
5883 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
5884 determine whether watchdog notifications are requested by
5885 the system manager.
5886
1e190502 5887 * Socket-activated per-connection services now include a
e49b5aad
LP
5888 short description of the connection parameters in the
5889 description.
5890
4c2413bf 5891 * tmpfiles gained a new "--boot" option. When this is not used,
e49b5aad 5892 only lines where the command character is not suffixed with
4670e9d5 5893 "!" are executed. When this option is specified, those
1e190502
ZJS
5894 options are executed too. This partitions tmpfiles
5895 directives into those that can be safely executed at any
5896 time, and those which should be run only at boot (for
5897 example, a line that creates /run/nologin).
e49b5aad 5898
c0c5af00 5899 * A new API "sd-resolve.h" has been added which provides a simple
4c2413bf 5900 asynchronous wrapper around glibc NSS host name resolution
e49b5aad 5901 calls, such as getaddrinfo(). In contrast to glibc's
4c2413bf
JE
5902 getaddrinfo_a(), it does not use signals. In contrast to most
5903 other asynchronous name resolution libraries, this one does
5904 not reimplement DNS, but reuses NSS, so that alternate
e49b5aad 5905 host name resolution systems continue to work, such as mDNS,
8b7d0494 5906 LDAP, etc. This API is based on libasyncns, but it has been
e49b5aad
LP
5907 cleaned up for inclusion in systemd.
5908
6300b3ec
LP
5909 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
5910 "sd-daemon.h" are no longer found in individual libraries
5911 libsystemd-journal.so, libsystemd-login.so,
5912 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
8b7d0494
JSJ
5913 merged them into a single library, libsystemd.so, which
5914 provides all symbols. The reason for this is cyclic
e49b5aad 5915 dependencies, as these libraries tend to use each other's
d28315e4 5916 symbols. So far, we have managed to workaround that by linking
6300b3ec
LP
5917 a copy of a good part of our code into each of these
5918 libraries again and again, which, however, makes certain
5919 things hard to do, like sharing static variables. Also, it
5920 substantially increases footprint. With this change, there
5921 is only one library for the basic APIs systemd
5922 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
5923 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
5924 library as well, however are subject to the --enable-kdbus
5925 switch (see below). Note that "sd-dhcp-client.h" is not part
5926 of this library (this is because it only consumes, never
5927 provides, services of/to other APIs). To make the transition
8b7d0494 5928 easy from the separate libraries to the unified one, we
4c2413bf 5929 provide the --enable-compat-libs compile-time switch which
e49b5aad
LP
5930 will generate stub libraries that are compatible with the
5931 old ones but redirect all calls to the new one.
5932
8b7d0494 5933 * All of the kdbus logic and the new APIs "sd-bus.h",
e49b5aad 5934 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
8b7d0494
JSJ
5935 and "sd-utf8.h" are compile-time optional via the
5936 "--enable-kdbus" switch, and they are not compiled in by
5937 default. To make use of kdbus, you have to explicitly enable
4c2413bf 5938 the switch. Note however, that neither the kernel nor the
e49b5aad
LP
5939 userspace API for all of this is considered stable yet. We
5940 want to maintain the freedom to still change the APIs for
4c2413bf 5941 now. By specifying this build-time switch, you acknowledge
e49b5aad 5942 that you are aware of the instability of the current
ad42cf73
KS
5943 APIs.
5944
5945 * Also, note that while kdbus is pretty much complete,
e49b5aad 5946 it lacks one thing: proper policy support. This means you
8b7d0494 5947 can build a fully working system with all features; however,
4c2413bf
JE
5948 it will be highly insecure. Policy support will be added in
5949 one of the next releases, at the same time that we will
5950 declare the APIs stable.
e49b5aad 5951
81c7dd89 5952 * When the kernel command line argument "kdbus" is specified,
ad42cf73 5953 systemd will automatically load the kdbus.ko kernel module. At
8b7d0494 5954 this stage of development, it is only useful for testing kdbus
ad42cf73 5955 and should not be used in production. Note: if "--enable-kdbus"
8b7d0494 5956 is specified, and the kdbus.ko kernel module is available, and
ad42cf73
KS
5957 "kdbus" is added to the kernel command line, the entire system
5958 runs with kdbus instead of dbus-daemon, with the above mentioned
5959 problem of missing the system policy enforcement. Also a future
5960 version of kdbus.ko or a newer systemd will not be compatible with
5961 each other, and will unlikely be able to boot the machine if only
5962 one of them is updated.
5963
e49b5aad 5964 * systemctl gained a new "import-environment" command which
4c2413bf 5965 uploads the caller's environment (or parts thereof) into the
e49b5aad
LP
5966 service manager so that it is inherited by services started
5967 by the manager. This is useful to upload variables like
5968 $DISPLAY into the user service manager.
5969
5970 * A new PrivateDevices= switch has been added to service units
5971 which allows running a service with a namespaced /dev
5972 directory that does not contain any device nodes for
4c2413bf 5973 physical devices. More specifically, it only includes devices
8b7d0494 5974 such as /dev/null, /dev/urandom, and /dev/zero which are API
e49b5aad
LP
5975 entry points.
5976
5977 * logind has been extended to support behaviour like VT
5978 switching on seats that do not support a VT. This makes
5979 multi-session available on seats that are not the first seat
5980 (seat0), and on systems where kernel support for VTs has
8b7d0494 5981 been disabled at compile-time.
e49b5aad
LP
5982
5983 * If a process holds a delay lock for system sleep or shutdown
1e190502 5984 and fails to release it in time, we will now log its
e49b5aad
LP
5985 identity. This makes it easier to identify processes that
5986 cause slow suspends or power-offs.
5987
1e190502
ZJS
5988 * When parsing /etc/crypttab, support for a new key-slot=
5989 option as supported by Debian is added. It allows indicating
5990 which LUKS slot to use on disk, speeding up key loading.
e49b5aad 5991
000b1ba5 5992 * The sd_journal_sendv() API call has been checked and
1e190502
ZJS
5993 officially declared to be async-signal-safe so that it may
5994 be invoked from signal handlers for logging purposes.
e49b5aad
LP
5995
5996 * Boot-time status output is now enabled automatically after a
5997 short timeout if boot does not progress, in order to give
8e420494 5998 the user an indication what she or he is waiting for.
1e190502
ZJS
5999
6000 * The boot-time output has been improved to show how much time
6001 remains until jobs expire.
e49b5aad
LP
6002
6003 * The KillMode= switch in service units gained a new possible
8b7d0494 6004 value "mixed". If set, and the unit is shut down, then the
e49b5aad 6005 initial SIGTERM signal is sent only to the main daemon
8e420494 6006 process, while the following SIGKILL signal is sent to
e49b5aad
LP
6007 all remaining processes of the service.
6008
4c2413bf
JE
6009 * When a scope unit is registered, a new property "Controller"
6010 may be set. If set to a valid bus name, systemd will send a
e49b5aad
LP
6011 RequestStop() signal to this name when it would like to shut
6012 down the scope. This may be used to hook manager logic into
6013 the shutdown logic of scope units. Also, scope units may now
8b7d0494 6014 be put in a special "abandoned" state, in which case the
e49b5aad
LP
6015 manager process which created them takes no further
6016 responsibilities for it.
6017
1e190502 6018 * When reading unit files, systemd will now verify
e49b5aad
LP
6019 the access mode of these files, and warn about certain
6020 suspicious combinations. This has been added to make it
6021 easier to track down packaging bugs where unit files are
6022 marked executable or world-writable.
6023
6024 * systemd-nspawn gained a new "--setenv=" switch to set
8b7d0494 6025 container-wide environment variables. The similar option in
1e190502
ZJS
6026 systemd-activate was renamed from "--environment=" to
6027 "--setenv=" for consistency.
e49b5aad
LP
6028
6029 * systemd-nspawn has been updated to create a new kdbus domain
6030 for each container that is invoked, thus allowing each
b9761003 6031 container to have its own set of system and user buses,
8b7d0494 6032 independent of the host.
e49b5aad
LP
6033
6034 * systemd-nspawn gained a new --drop-capability= switch to run
6035 the container with less capabilities than the default. Both
b9761003 6036 --drop-capability= and --capability= now take the special
e49b5aad
LP
6037 string "all" for dropping or keeping all capabilities.
6038
6039 * systemd-nspawn gained new switches for executing containers
6040 with specific SELinux labels set.
6041
6042 * systemd-nspawn gained a new --quiet switch to not generate
6043 any additional output but the container's own console
6044 output.
6045
6046 * systemd-nspawn gained a new --share-system switch to run a
6047 container without PID namespacing enabled.
6048
6049 * systemd-nspawn gained a new --register= switch to control
1e190502 6050 whether the container is registered with systemd-machined or
8e420494 6051 not. This is useful for containers that do not run full
e49b5aad
LP
6052 OS images, but only specific apps.
6053
6054 * systemd-nspawn gained a new --keep-unit which may be used
8b7d0494 6055 when invoked as the only program from a service unit, and
e49b5aad 6056 results in registration of the unit service itself in
1e190502 6057 systemd-machined, instead of a newly opened scope unit.
e49b5aad
LP
6058
6059 * systemd-nspawn gained a new --network-interface= switch for
6060 moving arbitrary interfaces to the container. The new
4c2413bf 6061 --network-veth switch creates a virtual Ethernet connection
8b7d0494
JSJ
6062 between host and container. The new --network-bridge=
6063 switch then allows assigning the host side of this virtual
6064 Ethernet connection to a bridge device.
e49b5aad 6065
6afc95b7
LP
6066 * systemd-nspawn gained a new --personality= switch for
6067 setting the kernel personality for the container. This is
70a44afe 6068 useful when running a 32-bit container on a 64-bit host. A
b8bde116
JE
6069 similar option Personality= is now also available for service
6070 units to use.
6afc95b7 6071
e49b5aad
LP
6072 * logind will now also track a "Desktop" identifier for each
6073 session which encodes the desktop environment of it. This is
6074 useful for desktop environments that want to identify
6075 multiple running sessions of itself easily.
6076
6077 * A new SELinuxContext= setting for service units has been
6078 added that allows setting a specific SELinux execution
6079 context for a service.
6080
6081 * Most systemd client tools will now honour $SYSTEMD_LESS for
6082 settings of the "less" pager. By default, these tools will
8b7d0494
JSJ
6083 override $LESS to allow certain operations to work, such as
6084 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
e49b5aad
LP
6085 influence this logic.
6086
6087 * systemd's "seccomp" hook-up has been changed to make use of
6088 the libseccomp library instead of using its own
6089 implementation. This has benefits for portability among
6090 other things.
6091
4c2413bf 6092 * For usage together with SystemCallFilter=, a new
8b7d0494 6093 SystemCallErrorNumber= setting has been introduced that
b8bde116
JE
6094 allows configuration of a system error number to be returned
6095 on filtered system calls, instead of immediately killing the
e49b5aad
LP
6096 process. Also, SystemCallArchitectures= has been added to
6097 limit access to system calls of a particular architecture
6098 (in order to turn off support for unused secondary
4c2413bf 6099 architectures). There is also a global
8b7d0494 6100 SystemCallArchitectures= setting in system.conf now to turn
e49b5aad
LP
6101 off support for non-native system calls system-wide.
6102
210054d7
KS
6103 * systemd requires a kernel with a working name_to_handle_at(),
6104 please see the kernel config requirements in the README file.
6105
e49b5aad
LP
6106 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
6107 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
6108 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
6109 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
6110 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
6111 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
6112 Elia Pinto, Florian Weimer, George McCollister, Goffredo
6113 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
6114 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
6115 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
6116 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
6117 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
6118 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
6119 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
6120 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
6121 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
6122 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
6123 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
6124 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
6125 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
6126 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
6127 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
6128 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
6129 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
6130
ccddd104 6131 — Berlin, 2014-02-20
e49b5aad 6132
cd4010b3
LP
6133CHANGES WITH 208:
6134
6135 * logind has gained support for facilitating privileged input
6136 and drm device access for unprivileged clients. This work is
6137 useful to allow Wayland display servers (and similar
6138 programs, such as kmscon) to run under the user's ID and
6139 access input and drm devices which are normally
6140 protected. When this is used (and the kernel is new enough)
6141 logind will "mute" IO on the file descriptors passed to
6142 Wayland as long as it is in the background and "unmute" it
6143 if it returns into the foreground. This allows secure
6144 session switching without allowing background sessions to
6145 eavesdrop on input and display data. This also introduces
6146 session switching support if VT support is turned off in the
6147 kernel, and on seats that are not seat0.
6148
6149 * A new kernel command line option luks.options= is understood
06b643e7 6150 now which allows specifying LUKS options for usage for LUKS
cd4010b3
LP
6151 encrypted partitions specified with luks.uuid=.
6152
6153 * tmpfiles.d(5) snippets may now use specifier expansion in
6154 path names. More specifically %m, %b, %H, %v, are now
6155 replaced by the local machine id, boot id, hostname, and
6156 kernel version number.
6157
6158 * A new tmpfiles.d(5) command "m" has been introduced which
6159 may be used to change the owner/group/access mode of a file
d28315e4 6160 or directory if it exists, but do nothing if it does not.
cd4010b3
LP
6161
6162 * This release removes high-level support for the
6163 MemorySoftLimit= cgroup setting. The underlying kernel
6164 cgroup attribute memory.soft_limit= is currently badly
6165 designed and likely to be removed from the kernel API in its
d28315e4 6166 current form, hence we should not expose it for now.
cd4010b3
LP
6167
6168 * The memory.use_hierarchy cgroup attribute is now enabled for
6169 all cgroups systemd creates in the memory cgroup
6170 hierarchy. This option is likely to be come the built-in
cc98b302
TH
6171 default in the kernel anyway, and the non-hierarchical mode
6172 never made much sense in the intrinsically hierarchical
cd4010b3
LP
6173 cgroup system.
6174
6175 * A new field _SYSTEMD_SLICE= is logged along with all journal
6176 messages containing the slice a message was generated
6177 from. This is useful to allow easy per-customer filtering of
6178 logs among other things.
6179
6180 * systemd-journald will no longer adjust the group of journal
6181 files it creates to the "systemd-journal" group. Instead we
6182 rely on the journal directory to be owned by the
6183 "systemd-journal" group, and its setgid bit set, so that the
6184 kernel file system layer will automatically enforce that
6185 journal files inherit this group assignment. The reason for
6186 this change is that we cannot allow NSS look-ups from
6187 journald which would be necessary to resolve
6188 "systemd-journal" to a numeric GID, because this might
6189 create deadlocks if NSS involves synchronous queries to
6190 other daemons (such as nscd, or sssd) which in turn are
6191 logging clients of journald and might block on it, which
6192 would then dead lock. A tmpfiles.d(5) snippet included in
6193 systemd will make sure the setgid bit and group are
6194 properly set on the journal directory if it exists on every
6195 boot. However, we recommend adjusting it manually after
6196 upgrades too (or from RPM scriptlets), so that the change is
6197 not delayed until next reboot.
6198
6199 * Backlight and random seed files in /var/lib/ have moved into
6200 the /var/lib/systemd/ directory, in order to centralize all
6201 systemd generated files in one directory.
6202
6203 * Boot time performance measurements (as displayed by
6204 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
6205 performance information if that's available to determine how
6206 much time BIOS and boot loader initialization required. With
6207 a sufficiently new BIOS you hence no longer need to boot
6208 with Gummiboot to get access to such information.
6209
6210 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
6211 Cristian Rodríguez, Dave Reisner, David Herrmann, David
6212 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
6213 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
6214 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
6215 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
6216 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
6217
ccddd104 6218 — Berlin, 2013-10-02
cd4010b3 6219
4f0be680
LP
6220CHANGES WITH 207:
6221
6222 * The Restart= option for services now understands a new
f3a165b0 6223 on-watchdog setting, which will restart the service
4f0be680
LP
6224 automatically if the service stops sending out watchdog keep
6225 alive messages (as configured with WatchdogSec=).
6226
6227 * The getty generator (which is responsible for bringing up a
6228 getty on configured serial consoles) will no longer only
6229 start a getty on the primary kernel console but on all
6230 others, too. This makes the order in which console= is
6231 specified on the kernel command line less important.
6232
6233 * libsystemd-logind gained a new sd_session_get_vt() call to
6234 retrieve the VT number of a session.
6235
6236 * If the option "tries=0" is set for an entry of /etc/crypttab
6237 its passphrase is queried indefinitely instead of any
6238 maximum number of tries.
6239
6240 * If a service with a configure PID file terminates its PID
6241 file will now be removed automatically if it still exists
6242 afterwards. This should put an end to stale PID files.
6243
6244 * systemd-run will now also take relative binary path names
6245 for execution and no longer insists on absolute paths.
6246
6247 * InaccessibleDirectories= and ReadOnlyDirectories= now take
6248 paths that are optionally prefixed with "-" to indicate that
d28315e4 6249 it should not be considered a failure if they do not exist.
4f0be680 6250
f3a165b0
KS
6251 * journalctl -o (and similar commands) now understands a new
6252 output mode "short-precise", it is similar to "short" but
4f0be680
LP
6253 shows timestamps with usec accuracy.
6254
6255 * The option "discard" (as known from Debian) is now
6256 synonymous to "allow-discards" in /etc/crypttab. In fact,
387abf80 6257 "discard" is preferred now (since it is easier to remember
4f0be680
LP
6258 and type).
6259
f3a165b0 6260 * Some licensing clean-ups were made, so that more code is now
4f0be680
LP
6261 LGPL-2.1 licensed than before.
6262
6263 * A minimal tool to save/restore the display backlight
6264 brightness across reboots has been added. It will store the
f3a165b0 6265 backlight setting as late as possible at shutdown, and
4f0be680
LP
6266 restore it as early as possible during reboot.
6267
6268 * A logic to automatically discover and enable home and swap
6269 partitions on GPT disks has been added. With this in place
6270 /etc/fstab becomes optional for many setups as systemd can
6271 discover certain partitions located on the root disk
6272 automatically. Home partitions are recognized under their
6273 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
6274 partitions are recognized under their GPT type ID
6275 0657fd6da4ab43c484e50933c84b4f4f.
6276
6277 * systemd will no longer pass any environment from the kernel
6278 or initrd to system services. If you want to set an
6279 environment for all services, do so via the kernel command
6280 line systemd.setenv= assignment.
6281
387abf80
LP
6282 * The systemd-sysctl tool no longer natively reads the file
6283 /etc/sysctl.conf. If desired, the file should be symlinked
6284 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
6285 legacy support by a symlink rather than built-in code, it
6286 also makes the otherwise hidden order of application of the
6287 different files visible. (Note that this partly reverts to a
6288 pre-198 application order of sysctl knobs!)
04bf3c1a 6289
4f0be680
LP
6290 * The "systemctl set-log-level" and "systemctl dump" commands
6291 have been moved to systemd-analyze.
6292
6293 * systemd-run learned the new --remain-after-exit switch,
6294 which causes the scope unit not to be cleaned up
6295 automatically after the process terminated.
6296
6297 * tmpfiles learned a new --exclude-prefix= switch to exclude
6298 certain paths from operation.
6299
6300 * journald will now automatically flush all messages to disk
f47ad593
ZJS
6301 as soon as a message at the log level CRIT, ALERT or EMERG
6302 is received.
4f0be680
LP
6303
6304 Contributions from: Andrew Cook, Brandon Philips, Christian
6305 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
6306 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
6307 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
6308 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
6309 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
6310 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
6311 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
6312 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
6313 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
6314 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
6315 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
6316 William Giokas, Zbigniew Jędrzejewski-Szmek
6317
ccddd104 6318 — Berlin, 2013-09-13
4f0be680 6319
408f281b
LP
6320CHANGES WITH 206:
6321
6322 * The documentation has been updated to cover the various new
6323 concepts introduced with 205.
6324
6325 * Unit files now understand the new %v specifier which
6326 resolves to the kernel version string as returned by "uname
6327 -r".
6328
6329 * systemctl now supports filtering the unit list output by
6330 load state, active state and sub state, using the new
33b521be 6331 --state= parameter.
408f281b
LP
6332
6333 * "systemctl status" will now show the results of the
6334 condition checks (like ConditionPathExists= and similar) of
6335 the last start attempts of the unit. They are also logged to
6336 the journal.
6337
6338 * "journalctl -b" may now be used to look for boot output of a
6339 specific boot. Try "journalctl -b -1" for the previous boot,
6340 but the syntax is substantially more powerful.
6341
6342 * "journalctl --show-cursor" has been added which prints the
6343 cursor string the last shown log line. This may then be used
6344 with the new "journalctl --after-cursor=" switch to continue
6345 browsing logs from that point on.
6346
6347 * "journalctl --force" may now be used to force regeneration
6348 of an FSS key.
6349
251cc819
LP
6350 * Creation of "dead" device nodes has been moved from udev
6351 into kmod and tmpfiles. Previously, udev would read the kmod
6352 databases to pre-generate dead device nodes based on meta
6353 information contained in kernel modules, so that these would
6354 be auto-loaded on access rather then at boot. As this
d28315e4 6355 does not really have much to do with the exposing actual
251cc819
LP
6356 kernel devices to userspace this has always been slightly
6357 alien in the udev codebase. Following the new scheme kmod
6358 will now generate a runtime snippet for tmpfiles from the
6359 module meta information and it now is tmpfiles' job to the
6360 create the nodes. This also allows overriding access and
6361 other parameters for the nodes using the usual tmpfiles
6362 facilities. As side effect this allows us to remove the
6363 CAP_SYS_MKNOD capability bit from udevd entirely.
6364
6365 * logind's device ACLs may now be applied to these "dead"
6366 devices nodes too, thus finally allowing managed access to
ce830873 6367 devices such as /dev/snd/sequencer without loading the
251cc819 6368 backing module right-away.
408f281b
LP
6369
6370 * A new RPM macro has been added that may be used to apply
6371 tmpfiles configuration during package installation.
6372
6373 * systemd-detect-virt and ConditionVirtualization= now can
6374 detect User-Mode-Linux machines (UML).
6375
251cc819
LP
6376 * journald will now implicitly log the effective capabilities
6377 set of processes in the message metadata.
408f281b
LP
6378
6379 * systemd-cryptsetup has gained support for TrueCrypt volumes.
6380
6381 * The initrd interface has been simplified (more specifically,
6382 support for passing performance data via environment
6383 variables and fsck results via files in /run has been
6384 removed). These features were non-essential, and are
6385 nowadays available in a much nicer way by having systemd in
6386 the initrd serialize its state and have the hosts systemd
6387 deserialize it again.
6388
28f5c779
KS
6389 * The udev "keymap" data files and tools to apply keyboard
6390 specific mappings of scan to key codes, and force-release
6391 scan code lists have been entirely replaced by a udev
6392 "keyboard" builtin and a hwdb data file.
408f281b 6393
251cc819
LP
6394 * systemd will now honour the kernel's "quiet" command line
6395 argument also during late shutdown, resulting in a
6396 completely silent shutdown when used.
6397
6398 * There's now an option to control the SO_REUSEPORT socket
6399 option in .socket units.
6400
6401 * Instance units will now automatically get a per-template
6402 subslice of system.slice unless something else is explicitly
6403 configured. For example, instances of sshd@.service will now
6404 implicitly be placed in system-sshd.slice rather than
6405 system.slice as before.
6406
6407 * Test coverage support may now be enabled at build time.
6408
6409 Contributions from: Dave Reisner, Frederic Crozat, Harald
6410 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
6411 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
6412 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
6413 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
6414 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
6415 Giokas, Zbigniew Jędrzejewski-Szmek
6416
ccddd104 6417 — Berlin, 2013-07-23
4f0be680 6418
00aa832b
LP
6419CHANGES WITH 205:
6420
6421 * Two new unit types have been introduced:
6422
6423 Scope units are very similar to service units, however, are
ccddd104 6424 created out of pre-existing processes — instead of PID 1
00aa832b
LP
6425 forking off the processes. By using scope units it is
6426 possible for system services and applications to group their
6427 own child processes (worker processes) in a powerful way
6428 which then maybe used to organize them, or kill them
6429 together, or apply resource limits on them.
6430
6431 Slice units may be used to partition system resources in an
cc98b302 6432 hierarchical fashion and then assign other units to them. By
00aa832b
LP
6433 default there are now three slices: system.slice (for all
6434 system services), user.slice (for all user sessions),
6435 machine.slice (for VMs and containers).
6436
6437 Slices and scopes have been introduced primarily in
6438 context of the work to move cgroup handling to a
6439 single-writer scheme, where only PID 1
6440 creates/removes/manages cgroups.
6441
6442 * There's a new concept of "transient" units. In contrast to
6443 normal units these units are created via an API at runtime,
6444 not from configuration from disk. More specifically this
6445 means it is now possible to run arbitrary programs as
6446 independent services, with all execution parameters passed
6447 in via bus APIs rather than read from disk. Transient units
6448 make systemd substantially more dynamic then it ever was,
6449 and useful as a general batch manager.
6450
6451 * logind has been updated to make use of scope and slice units
6452 for managing user sessions. As a user logs in he will get
6453 his own private slice unit, to which all sessions are added
6454 as scope units. We also added support for automatically
6455 adding an instance of user@.service for the user into the
6456 slice. Effectively logind will no longer create cgroup
6457 hierarchies on its own now, it will defer entirely to PID 1
6458 for this by means of scope, service and slice units. Since
6459 user sessions this way become entities managed by PID 1
6460 the output of "systemctl" is now a lot more comprehensive.
6461
6462 * A new mini-daemon "systemd-machined" has been added which
6463 may be used by virtualization managers to register local
6464 VMs/containers. nspawn has been updated accordingly, and
6465 libvirt will be updated shortly. machined will collect a bit
6466 of meta information about the VMs/containers, and assign
6467 them their own scope unit (see above). The collected
6468 meta-data is then made available via the "machinectl" tool,
6469 and exposed in "ps" and similar tools. machined/machinectl
6470 is compile-time optional.
6471
6472 * As discussed earlier, the low-level cgroup configuration
6473 options ControlGroup=, ControlGroupModify=,
6474 ControlGroupPersistent=, ControlGroupAttribute= have been
6475 removed. Please use high-level attribute settings instead as
6476 well as slice units.
6477
6478 * A new bus call SetUnitProperties() has been added to alter
6479 various runtime parameters of a unit. This is primarily
6480 useful to alter cgroup parameters dynamically in a nice way,
6481 but will be extended later on to make more properties
6482 modifiable at runtime. systemctl gained a new set-properties
6483 command that wraps this call.
6484
6485 * A new tool "systemd-run" has been added which can be used to
6486 run arbitrary command lines as transient services or scopes,
6487 while configuring a number of settings via the command
6488 line. This tool is currently very basic, however already
6489 very useful. We plan to extend this tool to even allow
6490 queuing of execution jobs with time triggers from the
6491 command line, similar in fashion to "at".
6492
6493 * nspawn will now inform the user explicitly that kernels with
6494 audit enabled break containers, and suggest the user to turn
6495 off audit.
6496
6497 * Support for detecting the IMA and AppArmor security
6498 frameworks with ConditionSecurity= has been added.
6499
6500 * journalctl gained a new "-k" switch for showing only kernel
1fda0ab5
ZJS
6501 messages, mimicking dmesg output; in addition to "--user"
6502 and "--system" switches for showing only user's own logs
6503 and system logs.
00aa832b
LP
6504
6505 * systemd-delta can now show information about drop-in
6506 snippets extending unit files.
6507
6508 * libsystemd-bus has been substantially updated but is still
6509 not available as public API.
6510
6511 * systemd will now look for the "debug" argument on the kernel
499b604b 6512 command line and enable debug logging, similar to what
00aa832b
LP
6513 "systemd.log_level=debug" already did before.
6514
6515 * "systemctl set-default", "systemctl get-default" has been
6516 added to configure the default.target symlink, which
6517 controls what to boot into by default.
6518
1fda0ab5
ZJS
6519 * "systemctl set-log-level" has been added as a convenient
6520 way to raise and lower systemd logging threshold.
6521
00aa832b
LP
6522 * "systemd-analyze plot" will now show the time the various
6523 generators needed for execution, as well as information
6524 about the unit file loading.
6525
00aa832b
LP
6526 * libsystemd-journal gained a new sd_journal_open_files() call
6527 for opening specific journal files. journactl also gained a
6528 new switch to expose this new functionality. Previously we
6529 only supported opening all files from a directory, or all
6530 files from the system, as opening individual files only is
6531 racy due to journal file rotation.
6532
6533 * systemd gained the new DefaultEnvironment= setting in
6534 /etc/systemd/system.conf to set environment variables for
6535 all services.
6536
6537 * If a privileged process logs a journal message with the
6538 OBJECT_PID= field set, then journald will automatically
6539 augment this with additional OBJECT_UID=, OBJECT_GID=,
6540 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
6541 system services want to log events about specific client
6542 processes. journactl/systemctl has been updated to make use
6543 of this information if all log messages regarding a specific
6544 unit is requested.
6545
6546 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
6547 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
6548 Reisner, David Coppa, David King, David Strauss, Eelco
6549 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
6550 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
6551 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
6552 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
6553 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
6554 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
6555 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
6556 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
6557 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
6558 Łukasz Stelmach, 장동준
6559
606c24e3
LP
6560CHANGES WITH 204:
6561
6562 * The Python bindings gained some minimal support for the APIs
6563 exposed by libsystemd-logind.
6564
6565 * ConditionSecurity= gained support for detecting SMACK. Since
6566 this condition already supports SELinux and AppArmor we only
6567 miss IMA for this. Patches welcome!
6568
6569 Contributions from: Karol Lewandowski, Lennart Poettering,
6570 Zbigniew Jędrzejewski-Szmek
6571
2f3fcf85
LP
6572CHANGES WITH 203:
6573
6574 * systemd-nspawn will now create /etc/resolv.conf if
6575 necessary, before bind-mounting the host's file onto it.
6576
6577 * systemd-nspawn will now store meta information about a
6578 container on the container's cgroup as extended attribute
6579 fields, including the root directory.
6580
6581 * The cgroup hierarchy has been reworked in many ways. All
6582 objects any of the components systemd creates in the cgroup
b82eed9a 6583 tree are now suffixed. More specifically, user sessions are
2f3fcf85
LP
6584 now placed in cgroups suffixed with ".session", users in
6585 cgroups suffixed with ".user", and nspawn containers in
6586 cgroups suffixed with ".nspawn". Furthermore, all cgroup
6587 names are now escaped in a simple scheme to avoid collision
6588 of userspace object names with kernel filenames. This work
6589 is preparation for making these objects relocatable in the
6590 cgroup tree, in order to allow easy resource partitioning of
6591 these objects without causing naming conflicts.
6592
6593 * systemctl list-dependencies gained the new switches
6594 --plain, --reverse, --after and --before.
6595
6596 * systemd-inhibit now shows the process name of processes that
6597 have taken an inhibitor lock.
6598
6599 * nss-myhostname will now also resolve "localhost"
6600 implicitly. This makes /etc/hosts an optional file and
6601 nicely handles that on IPv6 ::1 maps to both "localhost" and
6602 the local hostname.
6603
6604 * libsystemd-logind.so gained a new call
6605 sd_get_machine_names() to enumerate running containers and
6606 VMs (currently only supported by very new libvirt and
6607 nspawn). sd_login_monitor can now be used to watch
6608 VMs/containers coming and going.
6609
6610 * .include is not allowed recursively anymore, and only in
6611 unit files. Usually it is better to use drop-in snippets in
6612 .d/*.conf anyway, as introduced with systemd 198.
6613
6614 * systemd-analyze gained a new "critical-chain" command that
6615 determines the slowest chain of units run during system
6616 boot-up. It is very useful for tracking down where
6617 optimizing boot time is the most beneficial.
6618
6619 * systemd will no longer allow manipulating service paths in
6620 the name=systemd:/system cgroup tree using ControlGroup= in
6621 units. (But is still fine with it in all other dirs.)
6622
6623 * There's a new systemd-nspawn@.service service file that may
6624 be used to easily run nspawn containers as system
6625 services. With the container's root directory in
6626 /var/lib/container/foobar it is now sufficient to run
6627 "systemctl start systemd-nspawn@foobar.service" to boot it.
6628
6629 * systemd-cgls gained a new parameter "--machine" to list only
6630 the processes within a certain container.
6631
6632 * ConditionSecurity= now can check for "apparmor". We still
6633 are lacking checks for SMACK and IMA for this condition
6634 check though. Patches welcome!
6635
6636 * A new configuration file /etc/systemd/sleep.conf has been
6637 added that may be used to configure which kernel operation
6638 systemd is supposed to execute when "suspend", "hibernate"
6639 or "hybrid-sleep" is requested. This makes the new kernel
6640 "freeze" state accessible to the user.
6641
6642 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
6643 the passed argument if applicable.
6644
6645 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
6646 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
6647 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
6648 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
6649 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
6650 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
6651 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
6652 Jędrzejewski-Szmek
6653
ef3b5246
LP
6654CHANGES WITH 202:
6655
6656 * The output of 'systemctl list-jobs' got some polishing. The
6657 '--type=' argument may now be passed more than once. A new
6658 command 'systemctl list-sockets' has been added which shows
6659 a list of kernel sockets systemd is listening on with the
6660 socket units they belong to, plus the units these socket
6661 units activate.
6662
6663 * The experimental libsystemd-bus library got substantial
6664 updates to work in conjunction with the (also experimental)
6665 kdbus kernel project. It works well enough to exchange
6666 messages with some sophistication. Note that kdbus is not
6667 ready yet, and the library is mostly an elaborate test case
6668 for now, and not installable.
6669
6670 * systemd gained a new unit 'systemd-static-nodes.service'
6671 that generates static device nodes earlier during boot, and
6672 can run in conjunction with udev.
6673
6674 * libsystemd-login gained a new call sd_pid_get_user_unit()
6675 to retrieve the user systemd unit a process is running
6676 in. This is useful for systems where systemd is used as
6677 session manager.
6678
6679 * systemd-nspawn now places all containers in the new /machine
6680 top-level cgroup directory in the name=systemd
6681 hierarchy. libvirt will soon do the same, so that we get a
6682 uniform separation of /system, /user and /machine for system
6683 services, user processes and containers/virtual
6684 machines. This new cgroup hierarchy is also useful to stick
6685 stable names to specific container instances, which can be
7c04ad2d 6686 recognized later this way (this name may be controlled
ef3b5246
LP
6687 via systemd-nspawn's new -M switch). libsystemd-login also
6688 gained a new call sd_pid_get_machine_name() to retrieve the
6689 name of the container/VM a specific process belongs to.
6690
6691 * bootchart can now store its data in the journal.
6692
6693 * libsystemd-journal gained a new call
6694 sd_journal_add_conjunction() for AND expressions to the
6695 matching logic. This can be used to express more complex
6696 logical expressions.
6697
6698 * journactl can now take multiple --unit= and --user-unit=
6699 switches.
6700
6701 * The cryptsetup logic now understands the "luks.key=" kernel
6702 command line switch for specifying a file to read the
7c04ad2d 6703 decryption key from. Also, if a configured key file is not
ef3b5246
LP
6704 found the tool will now automatically fall back to prompting
6705 the user.
6706
cbeabcfb
ZJS
6707 * Python systemd.journal module was updated to wrap recently
6708 added functions from libsystemd-journal. The interface was
6709 changed to bring the low level interface in s.j._Reader
6710 closer to the C API, and the high level interface in
6711 s.j.Reader was updated to wrap and convert all data about
6712 an entry.
6713
ef3b5246
LP
6714 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
6715 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
6716 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
6717 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
6718 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
6719 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
6720
d3a86981
LP
6721CHANGES WITH 201:
6722
6723 * journalctl --update-catalog now understands a new --root=
6724 option to operate on catalogs found in a different root
6725 directory.
6726
6727 * During shutdown after systemd has terminated all running
6728 services a final killing loop kills all remaining left-over
6729 processes. We will now print the name of these processes
6730 when we send SIGKILL to them, since this usually indicates a
6731 problem.
6732
6733 * If /etc/crypttab refers to password files stored on
6734 configured mount points automatic dependencies will now be
6735 generated to ensure the specific mount is established first
6736 before the key file is attempted to be read.
6737
6738 * 'systemctl status' will now show information about the
6739 network sockets a socket unit is listening on.
6740
6741 * 'systemctl status' will also shown information about any
6742 drop-in configuration file for units. (Drop-In configuration
6743 files in this context are files such as
6744 /etc/systemd/systemd/foobar.service.d/*.conf)
6745
6746 * systemd-cgtop now optionally shows summed up CPU times of
6747 cgroups. Press '%' while running cgtop to switch between
6748 percentage and absolute mode. This is useful to determine
6749 which cgroups use up the most CPU time over the entire
6750 runtime of the system. systemd-cgtop has also been updated
6751 to be 'pipeable' for processing with further shell tools.
6752
6753 * 'hostnamectl set-hostname' will now allow setting of FQDN
6754 hostnames.
6755
6756 * The formatting and parsing of time span values has been
6757 changed. The parser now understands fractional expressions
6758 such as "5.5h". The formatter will now output fractional
6759 expressions for all time spans under 1min, i.e. "5.123456s"
6760 rather than "5s 123ms 456us". For time spans under 1s
6761 millisecond values are shown, for those under 1ms
6762 microsecond values are shown. This should greatly improve
6763 all time-related output of systemd.
6764
6765 * libsystemd-login and libsystemd-journal gained new
6766 functions for querying the poll() events mask and poll()
6767 timeout value for integration into arbitrary event
6768 loops.
6769
6770 * localectl gained the ability to list available X11 keymaps
6771 (models, layouts, variants, options).
6772
6773 * 'systemd-analyze dot' gained the ability to filter for
6774 specific units via shell-style globs, to create smaller,
d28315e4 6775 more useful graphs. I.e. it is now possible to create simple
d3a86981
LP
6776 graphs of all the dependencies between only target units, or
6777 of all units that Avahi has dependencies with.
6778
6779 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
6780 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
6781 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
6782 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
6783 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
6784 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
6785 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
6786
9ca3c17f
LP
6787CHANGES WITH 200:
6788
6789 * The boot-time readahead implementation for rotating media
6790 will now read the read-ahead data in multiple passes which
6791 consist of all read requests made in equidistant time
6792 intervals. This means instead of strictly reading read-ahead
6793 data in its physical order on disk we now try to find a
6794 middle ground between physical and access time order.
6795
6796 * /etc/os-release files gained a new BUILD_ID= field for usage
6797 on operating systems that provide continuous builds of OS
6798 images.
6799
6800 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
6801 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
6802 William Douglas, Zbigniew Jędrzejewski-Szmek
6803
35911459
LP
6804CHANGES WITH 199:
6805
6806 * systemd-python gained an API exposing libsystemd-daemon.
6807
6808 * The SMACK setup logic gained support for uploading CIPSO
6809 security policy.
6810
6811 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
6812 ReadOnlyDirectories= and InaccessibleDirectories= has
6813 changed. The private /tmp and /var/tmp directories are now
6814 shared by all processes of a service (which means
6815 ExecStartPre= may now leave data in /tmp that ExecStart= of
6816 the same service can still access). When a service is
6817 stopped its temporary directories are immediately deleted
a87197f5 6818 (normal clean-up with tmpfiles is still done in addition to
35911459
LP
6819 this though).
6820
6821 * By default, systemd will now set a couple of sysctl
6822 variables in the kernel: the safe sysrq options are turned
6823 on, IP route verification is turned on, and source routing
6824 disabled. The recently added hardlink and softlink
6825 protection of the kernel is turned on. These settings should
6826 be reasonably safe, and good defaults for all new systems.
6827
6828 * The predictable network naming logic may now be turned off
a87197f5 6829 with a new kernel command line switch: net.ifnames=0.
35911459
LP
6830
6831 * A new libsystemd-bus module has been added that implements a
6832 pretty complete D-Bus client library. For details see:
6833
56cadcb6 6834 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
35911459 6835
c20d8298 6836 * journald will now explicitly flush the journal files to disk
a87197f5
ZJS
6837 at the latest 5min after each write. The file will then also
6838 be marked offline until the next write. This should increase
6839 reliability in case of a crash. The synchronization delay
6840 can be configured via SyncIntervalSec= in journald.conf.
35911459
LP
6841
6842 * There's a new remote-fs-setup.target unit that can be used
6843 to pull in specific services when at least one remote file
6844 system is to be mounted.
6845
6846 * There are new targets timers.target and paths.target as
6847 canonical targets to pull user timer and path units in
6848 from. This complements sockets.target with a similar
6849 purpose for socket units.
6850
6a7d3d68
LP
6851 * libudev gained a new call udev_device_set_attribute_value()
6852 to set sysfs attributes of a device.
6853
a87197f5
ZJS
6854 * The udev daemon now sets the default number of worker
6855 processes executed in parallel based on the number of available
c20d8298 6856 CPUs instead of the amount of available RAM. This is supposed
ab06eef8 6857 to provide a more reliable default and limit a too aggressive
ce830873 6858 parallelism for setups with 1000s of devices connected.
c20d8298 6859
35911459
LP
6860 Contributions from: Auke Kok, Colin Walters, Cristian
6861 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
6862 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
6863 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
6864 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
6865 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
6866 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
6867 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
6868 Zbigniew Jędrzejewski-Szmek
6869
85d68397
LP
6870CHANGES WITH 198:
6871
6872 * Configuration of unit files may now be extended via drop-in
6873 files without having to edit/override the unit files
6874 themselves. More specifically, if the administrator wants to
6875 change one value for a service file foobar.service he can
6876 now do so by dropping in a configuration snippet into
ad88e758 6877 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
85d68397
LP
6878 will load all these snippets and apply them on top of the
6879 main unit configuration file, possibly extending or
6880 overriding its settings. Using these drop-in snippets is
40e21da8
KS
6881 generally nicer than the two earlier options for changing
6882 unit files locally: copying the files from
85d68397
LP
6883 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
6884 them there; or creating a new file in /etc/systemd/system/
6885 that incorporates the original one via ".include". Drop-in
6886 snippets into these .d/ directories can be placed in any
fd868975 6887 directory systemd looks for units in, and the usual
85d68397
LP
6888 overriding semantics between /usr/lib, /etc and /run apply
6889 for them too.
6890
6891 * Most unit file settings which take lists of items can now be
6aa8d43a 6892 reset by assigning the empty string to them. For example,
85d68397
LP
6893 normally, settings such as Environment=FOO=BAR append a new
6894 environment variable assignment to the environment block,
6895 each time they are used. By assigning Environment= the empty
6896 string the environment block can be reset to empty. This is
6897 particularly useful with the .d/*.conf drop-in snippets
156f7d09
KS
6898 mentioned above, since this adds the ability to reset list
6899 settings from vendor unit files via these drop-ins.
85d68397
LP
6900
6901 * systemctl gained a new "list-dependencies" command for
6902 listing the dependencies of a unit recursively.
6903
40e21da8 6904 * Inhibitors are now honored and listed by "systemctl
85d68397
LP
6905 suspend", "systemctl poweroff" (and similar) too, not only
6906 GNOME. These commands will also list active sessions by
6907 other users.
6908
6909 * Resource limits (as exposed by the various control group
6910 controllers) can now be controlled dynamically at runtime
6911 for all units. More specifically, you can now use a command
6912 like "systemctl set-cgroup-attr foobar.service cpu.shares
6913 2000" to alter the CPU shares a specific service gets. These
6aa8d43a 6914 settings are stored persistently on disk, and thus allow the
85d68397
LP
6915 administrator to easily adjust the resource usage of
6916 services with a few simple commands. This dynamic resource
6aa8d43a 6917 management logic is also available to other programs via the
85d68397
LP
6918 bus. Almost any kernel cgroup attribute and controller is
6919 supported.
6920
6921 * systemd-vconsole-setup will now copy all font settings to
6aa8d43a
LP
6922 all allocated VTs, where it previously applied them only to
6923 the foreground VT.
85d68397
LP
6924
6925 * libsystemd-login gained the new sd_session_get_tty() API
6926 call.
6927
6aa8d43a
LP
6928 * This release drops support for a few legacy or
6929 distribution-specific LSB facility names when parsing init
6930 scripts: $x-display-manager, $mail-transfer-agent,
85d68397
LP
6931 $mail-transport-agent, $mail-transfer-agent, $smtp,
6932 $null. Also, the mail-transfer-agent.target unit backing
6933 this has been removed. Distributions which want to retain
6aa8d43a
LP
6934 compatibility with this should carry the burden for
6935 supporting this themselves and patch support for these back
6936 in, if they really need to. Also, the facilities $syslog and
6937 $local_fs are now ignored, since systemd does not support
6938 early-boot LSB init scripts anymore, and these facilities
6939 are implied anyway for normal services. syslog.target has
6940 also been removed.
85d68397 6941
40e21da8 6942 * There are new bus calls on PID1's Manager object for
6aa8d43a 6943 cancelling jobs, and removing snapshot units. Previously,
85d68397
LP
6944 both calls were only available on the Job and Snapshot
6945 objects themselves.
6946
6947 * systemd-journal-gatewayd gained SSL support.
6948
6949 * The various "environment" files, such as /etc/locale.conf
6950 now support continuation lines with a backslash ("\") as
499b604b 6951 last character in the line, similarly in style (but different)
85d68397
LP
6952 to how this is supported in shells.
6953
6954 * For normal user processes the _SYSTEMD_USER_UNIT= field is
6955 now implicitly appended to every log entry logged. systemctl
6956 has been updated to filter by this field when operating on a
6957 user systemd instance.
6958
6959 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
6960 CAP_AUDIT_CONTROL capabilities to the capabilities set for
6961 the container. This makes it easier to boot unmodified
6962 Fedora systems in a container, which however still requires
6963 audit=0 to be passed on the kernel command line. Auditing in
6964 kernel and userspace is unfortunately still too broken in
6965 context of containers, hence we recommend compiling it out
6966 of the kernel or using audit=0. Hopefully this will be fixed
6967 one day for good in the kernel.
6968
6969 * nspawn gained the new --bind= and --bind-ro= parameters to
6970 bind mount specific directories from the host into the
6971 container.
6972
40e21da8 6973 * nspawn will now mount its own devpts file system instance
6aa8d43a 6974 into the container, in order not to leak pty devices from
85d68397
LP
6975 the host into the container.
6976
6977 * systemd will now read the firmware boot time performance
6aa8d43a
LP
6978 information from the EFI variables, if the used boot loader
6979 supports this, and takes it into account for boot performance
6980 analysis via "systemd-analyze". This is currently supported
6981 only in conjunction with Gummiboot, but could be supported
6982 by other boot loaders too. For details see:
85d68397 6983
56cadcb6 6984 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
85d68397
LP
6985
6986 * A new generator has been added that automatically mounts the
6987 EFI System Partition (ESP) to /boot, if that directory
6aa8d43a
LP
6988 exists, is empty, and no other file system has been
6989 configured to be mounted there.
85d68397
LP
6990
6991 * logind will now send out PrepareForSleep(false) out
6992 unconditionally, after coming back from suspend. This may be
6993 used by applications as asynchronous notification for
6994 system resume events.
6995
6996 * "systemctl unlock-sessions" has been added, that allows
6997 unlocking the screens of all user sessions at once, similar
499b604b 6998 to how "systemctl lock-sessions" already locked all users
40e21da8 6999 sessions. This is backed by a new D-Bus call UnlockSessions().
85d68397
LP
7000
7001 * "loginctl seat-status" will now show the master device of a
7002 seat. (i.e. the device of a seat that needs to be around for
7003 the seat to be considered available, usually the graphics
7004 card).
7005
7006 * tmpfiles gained a new "X" line type, that allows
7007 configuration of files and directories (with wildcards) that
7008 shall be excluded from automatic cleanup ("aging").
7009
bf933560
KS
7010 * udev default rules set the device node permissions now only
7011 at "add" events, and do not change them any longer with a
7012 later "change" event.
85d68397
LP
7013
7014 * The log messages for lid events and power/sleep keypresses
7015 now carry a message ID.
7016
7017 * We now have a substantially larger unit test suite, but this
7018 continues to be work in progress.
7019
7020 * udevadm hwdb gained a new --root= parameter to change the
7021 root directory to operate relative to.
7022
40e21da8
KS
7023 * logind will now issue a background sync() request to the kernel
7024 early at shutdown, so that dirty buffers are flushed to disk early
85d68397
LP
7025 instead of at the last moment, in order to optimize shutdown
7026 times a little.
7027
7028 * A new bootctl tool has been added that is an interface for
7029 certain boot loader operations. This is currently a preview
7030 and is likely to be extended into a small mechanism daemon
7031 like timedated, localed, hostnamed, and can be used by
7032 graphical UIs to enumerate available boot options, and
7033 request boot into firmware operations.
7034
7035 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
7036 the rest of the package. It also has been updated to work
7037 correctly in initrds.
7038
7039 * Policykit previously has been runtime optional, and is now
7040 also compile time optional via a configure switch.
7041
7042 * systemd-analyze has been reimplemented in C. Also "systemctl
7043 dot" has moved into systemd-analyze.
7044
7045 * "systemctl status" with no further parameters will now print
7046 the status of all active or failed units.
7047
7048 * Operations such as "systemctl start" can now be executed
7049 with a new mode "--irreversible" which may be used to queue
7050 operations that cannot accidentally be reversed by a later
6aa8d43a 7051 job queuing. This is by default used to make shutdown
85d68397
LP
7052 requests more robust.
7053
7054 * The Python API of systemd now gained a new module for
7055 reading journal files.
7056
7057 * A new tool kernel-install has been added that can install
7058 kernel images according to the Boot Loader Specification:
7059
56cadcb6 7060 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
85d68397
LP
7061
7062 * Boot time console output has been improved to provide
6aa8d43a 7063 animated boot time output for hanging jobs.
85d68397
LP
7064
7065 * A new tool systemd-activate has been added which can be used
7066 to test socket activation with, directly from the command
7067 line. This should make it much easier to test and debug
7068 socket activation in daemons.
7069
7070 * journalctl gained a new "--reverse" (or -r) option to show
7071 journal output in reverse order (i.e. newest line first).
7072
43447fb7
LP
7073 * journalctl gained a new "--pager-end" (or -e) option to jump
7074 to immediately jump to the end of the journal in the
7075 pager. This is only supported in conjunction with "less".
7076
85d68397 7077 * journalctl gained a new "--user-unit=" option, that works
499b604b 7078 similarly to "--unit=" but filters for user units rather than
85d68397
LP
7079 system units.
7080
7081 * A number of unit files to ease adoption of systemd in
7082 initrds has been added. This moves some minimal logic from
7083 the various initrd implementations into systemd proper.
7084
7085 * The journal files are now owned by a new group
7086 "systemd-journal", which exists specifically to allow access
7087 to the journal, and nothing else. Previously, we used the
6aa8d43a 7088 "adm" group for that, which however possibly covers more
85d68397
LP
7089 than just journal/log file access. This new group is now
7090 already used by systemd-journal-gatewayd to ensure this
7091 daemon gets access to the journal files and as little else
7092 as possible. Note that "make install" will also set FS ACLs
7093 up for /var/log/journal to give "adm" and "wheel" read
7094 access to it, in addition to "systemd-journal" which owns
7095 the journal files. We recommend that packaging scripts also
6aa8d43a 7096 add read access to "adm" + "wheel" to /var/log/journal, and
85d68397
LP
7097 all existing/future journal files. To normal users and
7098 administrators little changes, however packagers need to
7099 ensure to create the "systemd-journal" system group at
7100 package installation time.
7101
7102 * The systemd-journal-gatewayd now runs as unprivileged user
7103 systemd-journal-gateway:systemd-journal-gateway. Packaging
7104 scripts need to create these system user/group at
7105 installation time.
7106
7107 * timedated now exposes a new boolean property CanNTP that
7108 indicates whether a local NTP service is available or not.
7109
7110 * systemd-detect-virt will now also detect xen PVs
7111
40e21da8
KS
7112 * The pstore file system is now mounted by default, if it is
7113 available.
85d68397 7114
1aed4590
LP
7115 * In addition to the SELinux and IMA policies we will now also
7116 load SMACK policies at early boot.
7117
85d68397
LP
7118 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
7119 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
7120 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
7121 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
7122 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
7123 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
7124 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
7125 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
7126 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
7127 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
7128 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
7129 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
7130 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
7131 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
7132
8ad26859
LP
7133CHANGES WITH 197:
7134
7135 * Timer units now support calendar time events in addition to
7136 monotonic time events. That means you can now trigger a unit
7137 based on a calendar time specification such as "Thu,Fri
7138 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
7139 or fifth day of any month of the year 2013, given that it is
7140 a thursday or friday. This brings timer event support
7141 considerably closer to cron's capabilities. For details on
7142 the supported calendar time specification language see
7143 systemd.time(7).
7144
7145 * udev now supports a number of different naming policies for
7146 network interfaces for predictable names, and a combination
7147 of these policies is now the default. Please see this wiki
7148 document for details:
7149
56cadcb6 7150 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
8ad26859
LP
7151
7152 * Auke Kok's bootchart implementation has been added to the
d28315e4
JE
7153 systemd tree. It is an optional component that can graph the
7154 boot in quite some detail. It is one of the best bootchart
8ad26859
LP
7155 implementations around and minimal in its code and
7156 dependencies.
7157
7158 * nss-myhostname has been integrated into the systemd source
7159 tree. nss-myhostname guarantees that the local hostname
7160 always stays resolvable via NSS. It has been a weak
7161 requirement of systemd-hostnamed since a long time, and
7162 since its code is actually trivial we decided to just
7163 include it in systemd's source tree. It can be turned off
7164 with a configure switch.
7165
7166 * The read-ahead logic is now capable of properly detecting
7167 whether a btrfs file system is on SSD or rotating media, in
7168 order to optimize the read-ahead scheme. Previously, it was
7169 only capable of detecting this on traditional file systems
7170 such as ext4.
7171
7172 * In udev, additional device properties are now read from the
7173 IAB in addition to the OUI database. Also, Bluetooth company
7174 identities are attached to the devices as well.
7175
7176 * In service files %U may be used as specifier that is
7177 replaced by the configured user name of the service.
7178
7179 * nspawn may now be invoked without a controlling TTY. This
7180 makes it suitable for invocation as its own service. This
7181 may be used to set up a simple containerized server system
7182 using only core OS tools.
7183
7184 * systemd and nspawn can now accept socket file descriptors
7185 when they are started for socket activation. This enables
7186 implementation of socket activated nspawn
7187 containers. i.e. think about autospawning an entire OS image
7188 when the first SSH or HTTP connection is received. We expect
7189 that similar functionality will also be added to libvirt-lxc
7190 eventually.
7191
7192 * journalctl will now suppress ANSI color codes when
7193 presenting log data.
7194
7195 * systemctl will no longer show control group information for
ce830873 7196 a unit if the control group is empty anyway.
8ad26859
LP
7197
7198 * logind can now automatically suspend/hibernate/shutdown the
7199 system on idle.
7200
7201 * /etc/machine-info and hostnamed now also expose the chassis
7202 type of the system. This can be used to determine whether
7203 the local system is a laptop, desktop, handset or
7204 tablet. This information may either be configured by the
7205 user/vendor or is automatically determined from ACPI and DMI
7206 information if possible.
7207
7208 * A number of PolicyKit actions are now bound together with
7209 "imply" rules. This should simplify creating UIs because
7210 many actions will now authenticate similar ones as well.
7211
7212 * Unit files learnt a new condition ConditionACPower= which
7213 may be used to conditionalize a unit depending on whether an
7214 AC power source is connected or not, of whether the system
7215 is running on battery power.
7216
7217 * systemctl gained a new "is-failed" verb that may be used in
7218 shell scripts and suchlike to check whether a specific unit
7219 is in the "failed" state.
7220
7221 * The EnvironmentFile= setting in unit files now supports file
7222 globbing, and can hence be used to easily read a number of
7223 environment files at once.
7224
7225 * systemd will no longer detect and recognize specific
7226 distributions. All distribution-specific #ifdeffery has been
7227 removed, systemd is now fully generic and
7228 distribution-agnostic. Effectively, not too much is lost as
7229 a lot of the code is still accessible via explicit configure
7230 switches. However, support for some distribution specific
7231 legacy configuration file formats has been dropped. We
7232 recommend distributions to simply adopt the configuration
7233 files everybody else uses now and convert the old
7234 configuration from packaging scripts. Most distributions
7235 already did that. If that's not possible or desirable,
7236 distributions are welcome to forward port the specific
7237 pieces of code locally from the git history.
7238
7239 * When logging a message about a unit systemd will now always
7240 log the unit name in the message meta data.
7241
7242 * localectl will now also discover system locale data that is
7243 not stored in locale archives, but directly unpacked.
7244
7245 * logind will no longer unconditionally use framebuffer
7246 devices as seat masters, i.e. as devices that are required
7247 to be existing before a seat is considered preset. Instead,
7248 it will now look for all devices that are tagged as
b938cb90
JE
7249 "seat-master" in udev. By default, framebuffer devices will
7250 be marked as such, but depending on local systems, other
8ad26859
LP
7251 devices might be marked as well. This may be used to
7252 integrate graphics cards using closed source drivers (such
7253 as NVidia ones) more nicely into logind. Note however, that
7254 we recommend using the open source NVidia drivers instead,
7255 and no udev rules for the closed-source drivers will be
7256 shipped from us upstream.
7257
7258 Contributions from: Adam Williamson, Alessandro Crismani, Auke
7259 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
7260 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
7261 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
7262 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
7263 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
7264 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
7265 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
7266 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
7267 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
7268 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
7269 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
7270 Jędrzejewski-Szmek
7271
0428ddb7
LP
7272CHANGES WITH 196:
7273
7274 * udev gained support for loading additional device properties
7275 from an indexed database that is keyed by vendor/product IDs
7276 and similar device identifiers. For the beginning this
7277 "hwdb" is populated with data from the well-known PCI and
7278 USB database, but also includes PNP, ACPI and OID data. In
7279 the longer run this indexed database shall grow into
7280 becoming the one central database for non-essential
7281 userspace device metadata. Previously, data from the PCI/USB
96ec33c0 7282 database was only attached to select devices, since the
0428ddb7 7283 lookup was a relatively expensive operation due to O(n) time
96ec33c0
LP
7284 complexity (with n being the number of entries in the
7285 database). Since this is now O(1), we decided to add in this
7286 data for all devices where this is available, by
0428ddb7
LP
7287 default. Note that the indexed database needs to be rebuilt
7288 when new data files are installed. To achieve this you need
7289 to update your packaging scripts to invoke "udevadm hwdb
7290 --update" after installation of hwdb data files. For
7291 RPM-based distributions we introduced the new
7292 %udev_hwdb_update macro for this purpose.
7293
7294 * The Journal gained support for the "Message Catalog", an
7295 indexed database to link up additional information with
7296 journal entries. For further details please check:
7297
56cadcb6 7298 https://www.freedesktop.org/wiki/Software/systemd/catalog
0428ddb7
LP
7299
7300 The indexed message catalog database also needs to be
7301 rebuilt after installation of message catalog files. Use
7302 "journalctl --update-catalog" for this. For RPM-based
7303 distributions we introduced the %journal_catalog_update
7304 macro for this purpose.
7305
7306 * The Python Journal bindings gained support for the standard
7307 Python logging framework.
7308
7309 * The Journal API gained new functions for checking whether
7310 the underlying file system of a journal file is capable of
7311 properly reporting file change notifications, or whether
7312 applications that want to reflect journal changes "live"
ab06eef8 7313 need to recheck journal files continuously in appropriate
0428ddb7
LP
7314 time intervals.
7315
7316 * It is now possible to set the "age" field for tmpfiles
7317 entries to 0, indicating that files matching this entry
7318 shall always be removed when the directories are cleaned up.
7319
7320 * coredumpctl gained a new "gdb" verb which invokes gdb
7321 right-away on the selected coredump.
7322
7323 * There's now support for "hybrid sleep" on kernels that
7324 support this, in addition to "suspend" and "hibernate". Use
7325 "systemctl hybrid-sleep" to make use of this.
7326
7327 * logind's HandleSuspendKey= setting (and related settings)
7328 now gained support for a new "lock" setting to simply
7329 request the screen lock on all local sessions, instead of
7330 actually executing a suspend or hibernation.
7331
7332 * systemd will now mount the EFI variables file system by
7333 default.
7334
7335 * Socket units now gained support for configuration of the
7336 SMACK security label.
7337
7338 * timedatectl will now output the time of the last and next
7339 daylight saving change.
7340
7341 * We dropped support for various legacy and distro-specific
7342 concepts, such as insserv, early-boot SysV services
7343 (i.e. those for non-standard runlevels such as 'b' or 'S')
7344 or ArchLinux /etc/rc.conf support. We recommend the
7345 distributions who still need support this to either continue
7346 to maintain the necessary patches downstream, or find a
7347 different solution. (Talk to us if you have questions!)
7348
7349 * Various systemd components will now bypass PolicyKit checks
7350 for root and otherwise handle properly if PolicyKit is not
7351 found to be around. This should fix most issues for
7352 PolicyKit-less systems. Quite frankly this should have been
7353 this way since day one. It is absolutely our intention to
7354 make systemd work fine on PolicyKit-less systems, and we
d28315e4 7355 consider it a bug if something does not work as it should if
0428ddb7
LP
7356 PolicyKit is not around.
7357
7358 * For embedded systems it is now possible to build udev and
7359 systemd without blkid and/or kmod support.
7360
7361 * "systemctl switch-root" is now capable of switching root
7362 more than once. I.e. in addition to transitions from the
7363 initrd to the host OS it is now possible to transition to
7364 further OS images from the host. This is useful to implement
7365 offline updating tools.
7366
7367 * Various other additions have been made to the RPM macros
7368 shipped with systemd. Use %udev_rules_update() after
7369 installing new udev rules files. %_udevhwdbdir,
7370 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
7371 %_sysctldir are now available which resolve to the right
7372 directories for packages to place various data files in.
7373
7374 * journalctl gained the new --full switch (in addition to
7375 --all, to disable ellipsation for long messages.
7376
7377 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
7378 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
7379 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
7380 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
7381 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
7382 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
7383 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
7384 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
7385 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
7386
139ee8cc
LP
7387CHANGES WITH 195:
7388
6827101a 7389 * journalctl gained new --since= and --until= switches to
139ee8cc
LP
7390 filter by time. It also now supports nice filtering for
7391 units via --unit=/-u.
7392
6827101a 7393 * Type=oneshot services may use ExecReload= and do the
139ee8cc
LP
7394 right thing.
7395
7396 * The journal daemon now supports time-based rotation and
7397 vacuuming, in addition to the usual disk-space based
7398 rotation.
7399
7400 * The journal will now index the available field values for
7401 each field name. This enables clients to show pretty drop
7402 downs of available match values when filtering. The bash
7403 completion of journalctl has been updated
7404 accordingly. journalctl gained a new switch -F to list all
7405 values a certain field takes in the journal database.
7406
7407 * More service events are now written as structured messages
7408 to the journal, and made recognizable via message IDs.
7409
7410 * The timedated, localed and hostnamed mini-services which
7411 previously only provided support for changing time, locale
7412 and hostname settings from graphical DEs such as GNOME now
7413 also have a minimal (but very useful) text-based client
7414 utility each. This is probably the nicest way to changing
7415 these settings from the command line now, especially since
7416 it lists available options and is fully integrated with bash
7417 completion.
7418
7419 * There's now a new tool "systemd-coredumpctl" to list and
7420 extract coredumps from the journal.
7421
7422 * We now install a README each in /var/log/ and
7423 /etc/rc.d/init.d explaining where the system logs and init
7424 scripts went. This hopefully should help folks who go to
7425 that dirs and look into the otherwise now empty void and
7426 scratch their heads.
7427
7428 * When user-services are invoked (by systemd --user) the
7429 $MANAGERPID env var is set to the PID of systemd.
7430
7431 * SIGRTMIN+24 when sent to a --user instance will now result
7432 in immediate termination of systemd.
7433
7434 * gatewayd received numerous feature additions such as a
7435 "follow" mode, for live syncing and filtering.
7436
7437 * browse.html now allows filtering and showing detailed
7438 information on specific entries. Keyboard navigation and
7439 mouse screen support has been added.
7440
7441 * gatewayd/journalctl now supports HTML5/JSON
7442 Server-Sent-Events as output.
7443
1cb88f2c 7444 * The SysV init script compatibility logic will now
139ee8cc
LP
7445 heuristically determine whether a script supports the
7446 "reload" verb, and only then make this available as
7447 "systemctl reload".
7448
15f47220 7449 * "systemctl status --follow" has been removed, use "journalctl
139ee8cc
LP
7450 -u" instead.
7451
7452 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
7453 have been removed since they are hardly useful to be
7454 configured.
7455
7456 * And I'd like to take the opportunity to specifically mention
7457 Zbigniew for his great contributions. Zbigniew, you rock!
7458
7459 Contributions from: Andrew Eikum, Christian Hesse, Colin
7460 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
4d92e078
LP
7461 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
7462 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
7463 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
7464 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
7465 Jędrzejewski-Szmek, Сковорода Никита Андреевич
139ee8cc 7466
f9b55720
LP
7467CHANGES WITH 194:
7468
7469 * If /etc/vconsole.conf is non-existent or empty we will no
7470 longer load any console font or key map at boot by
7471 default. Instead the kernel defaults will be left
7472 intact. This is definitely the right thing to do, as no
7473 configuration should mean no configuration, and hard-coding
7474 font names that are different on all archs is probably a bad
7475 idea. Also, the kernel default key map and font should be
7476 good enough for most cases anyway, and mostly identical to
7477 the userspace fonts/key maps we previously overloaded them
7478 with. If distributions want to continue to default to a
7479 non-kernel font or key map they should ship a default
7480 /etc/vconsole.conf with the appropriate contents.
7481
7482 Contributions from: Colin Walters, Daniel J Walsh, Dave
7483 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
7484 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
7485
597c52cf
LP
7486CHANGES WITH 193:
7487
7488 * journalctl gained a new --cursor= switch to show entries
7489 starting from the specified location in the journal.
7490
7491 * We now enforce a size limit on journal entry fields exported
7492 with "-o json" in journalctl. Fields larger than 4K will be
7493 assigned null. This can be turned off with --all.
7494
7495 * An (optional) journal gateway daemon is now available as
7496 "systemd-journal-gatewayd.service". This service provides
7497 access to the journal via HTTP and JSON. This functionality
7498 will be used to implement live log synchronization in both
7499 pull and push modes, but has various other users too, such
7500 as easy log access for debugging of embedded devices. Right
7501 now it is already useful to retrieve the journal via HTTP:
7502
7503 # systemctl start systemd-journal-gatewayd.service
7504 # wget http://localhost:19531/entries
7505
7506 This will download the journal contents in a
7507 /var/log/messages compatible format. The same as JSON:
7508
7509 # curl -H"Accept: application/json" http://localhost:19531/entries
7510
7511 This service is also accessible via a web browser where a
7512 single static HTML5 app is served that uses the JSON logic
7513 to enable the user to do some basic browsing of the
7514 journal. This will be extended later on. Here's an example
7515 screenshot of this app in its current state:
7516
7517 http://0pointer.de/public/journal-gatewayd
7518
7519 Contributions from: Kay Sievers, Lennart Poettering, Robert
7520 Milasan, Tom Gundersen
7521
075d4ecb
LP
7522CHANGES WITH 192:
7523
7524 * The bash completion logic is now available for journalctl
7525 too.
7526
d28315e4 7527 * We do not mount the "cpuset" controller anymore together with
075d4ecb
LP
7528 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
7529 started if no parameters are assigned to it. "cpuset" hence
61233823 7530 broke code that assumed it could create "cpu" groups and
075d4ecb
LP
7531 just start them.
7532
7533 * journalctl -f will now subscribe to terminal size changes,
7534 and line break accordingly.
7535
597c52cf
LP
7536 Contributions from: Dave Reisner, Kay Sievers, Lennart
7537 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
075d4ecb 7538
b6a86739
LP
7539CHANGES WITH 191:
7540
7541 * nspawn will now create a symlink /etc/localtime in the
7542 container environment, copying the host's timezone
7543 setting. Previously this has been done via a bind mount, but
7544 since symlinks cannot be bind mounted this has now been
7545 changed to create/update the appropriate symlink.
7546
7547 * journalctl -n's line number argument is now optional, and
7548 will default to 10 if omitted.
7549
7550 * journald will now log the maximum size the journal files may
7551 take up on disk. This is particularly useful if the default
7552 built-in logic of determining this parameter from the file
7553 system size is used. Use "systemctl status
6563b535 7554 systemd-journald.service" to see this information.
b6a86739
LP
7555
7556 * The multi-seat X wrapper tool has been stripped down. As X
7557 is now capable of enumerating graphics devices via udev in a
7558 seat-aware way the wrapper is not strictly necessary
7559 anymore. A stripped down temporary stop-gap is still shipped
7560 until the upstream display managers have been updated to
7561 fully support the new X logic. Expect this wrapper to be
6563b535 7562 removed entirely in one of the next releases.
b6a86739
LP
7563
7564 * HandleSleepKey= in logind.conf has been split up into
7565 HandleSuspendKey= and HandleHibernateKey=. The old setting
6563b535 7566 is not available anymore. X11 and the kernel are
45afd519 7567 distinguishing between these keys and we should too. This
b6a86739
LP
7568 also means the inhibition lock for these keys has been split
7569 into two.
7570
597c52cf
LP
7571 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
7572 Poettering, Lukas Nykryn, Václav Pavlín
b6a86739 7573
0c11f949
LP
7574CHANGES WITH 190:
7575
d28315e4 7576 * Whenever a unit changes state we will now log this to the
0c11f949
LP
7577 journal and show along the unit's own log output in
7578 "systemctl status".
7579
7580 * ConditionPathIsMountPoint= can now properly detect bind
7581 mount points too. (Previously, a bind mount of one file
8d0256b7 7582 system to another place in the same file system could not be
0c11f949
LP
7583 detected as mount, since they shared struct stat's st_dev
7584 field.)
7585
7586 * We will now mount the cgroup controllers cpu, cpuacct,
7587 cpuset and the controllers net_cls, net_prio together by
7588 default.
7589
7590 * nspawn containers will now have a virtualized boot
7591 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
7592 over with a randomized ID at container initialization). This
7593 has the effect of making "journalctl -b" do the right thing
7594 in a container.
7595
7596 * The JSON output journal serialization has been updated not
7597 to generate "endless" list objects anymore, but rather one
7598 JSON object per line. This is more in line how most JSON
7599 parsers expect JSON objects. The new output mode
7600 "json-pretty" has been added to provide similar output, but
7601 neatly aligned for readability by humans.
7602
7603 * We dropped all explicit sync() invocations in the shutdown
7604 code. The kernel does this implicitly anyway in the kernel
7605 reboot() syscall. halt(8)'s -n option is now a compatibility
7606 no-op.
7607
7608 * We now support virtualized reboot() in containers, as
7609 supported by newer kernels. We will fall back to exit() if
7610 CAP_SYS_REBOOT is not available to the container. Also,
7611 nspawn makes use of this now and will actually reboot the
7612 container if the containerized OS asks for that.
7613
7614 * journalctl will only show local log output by default
7615 now. Use --merge (-m) to show remote log output, too.
7616
7617 * libsystemd-journal gained the new sd_journal_get_usage()
7618 call to determine the current disk usage of all journal
7619 files. This is exposed in the new "journalctl --disk-usage"
7620 command.
7621
7622 * journald gained a new configuration setting SplitMode= in
7623 journald.conf which may be used to control how user journals
7624 are split off. See journald.conf(5) for details.
7625
7626 * A new condition type ConditionFileNotEmpty= has been added.
7627
7628 * tmpfiles' "w" lines now support file globbing, to write
7629 multiple files at once.
7630
7631 * We added Python bindings for the journal submission
7632 APIs. More Python APIs for a number of selected APIs will
7633 likely follow. Note that we intend to add native bindings
7634 only for the Python language, as we consider it common
7635 enough to deserve bindings shipped within systemd. There are
7636 various projects outside of systemd that provide bindings
7637 for languages such as PHP or Lua.
7638
a98d5d64
LP
7639 * Many conditions will now resolve specifiers such as %i. In
7640 addition, PathChanged= and related directives of .path units
7641 now support specifiers as well.
0c11f949
LP
7642
7643 * There's now a new RPM macro definition for the system preset
7644 dir: %_presetdir.
7645
d28315e4 7646 * journald will now warn if it ca not forward a message to the
dca348bc 7647 syslog daemon because its socket is full.
0c11f949
LP
7648
7649 * timedated will no longer write or process /etc/timezone,
7650 except on Debian. As we do not support late mounted /usr
7651 anymore /etc/localtime always being a symlink is now safe,
7652 and hence the information in /etc/timezone is not necessary
7653 anymore.
7654
aaccc32c 7655 * logind will now always reserve one VT for a text getty (VT6
0c11f949
LP
7656 by default). Previously if more than 6 X sessions where
7657 started they took up all the VTs with auto-spawned gettys,
7658 so that no text gettys were available anymore.
7659
7660 * udev will now automatically inform the btrfs kernel logic
7661 about btrfs RAID components showing up. This should make
7662 simple hotplug based btrfs RAID assembly work.
7663
7664 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
7665 (but not for its children which will stay at the kernel
7666 default). This should allow setups with a lot more listening
7667 sockets.
7668
7669 * systemd will now always pass the configured timezone to the
7670 kernel at boot. timedated will do the same when the timezone
7671 is changed.
7672
7673 * logind's inhibition logic has been updated. By default,
7674 logind will now handle the lid switch, the power and sleep
7675 keys all the time, even in graphical sessions. If DEs want
7676 to handle these events on their own they should take the new
7677 handle-power-key, handle-sleep-key and handle-lid-switch
f131770b 7678 inhibitors during their runtime. A simple way to achieve
0c11f949
LP
7679 that is to invoke the DE wrapped in an invocation of:
7680
7681 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
7682
7683 * Access to unit operations is now checked via SELinux taking
7684 the unit file label and client process label into account.
7685
aad803af
LP
7686 * systemd will now notify the administrator in the journal
7687 when he over-mounts a non-empty directory.
7688
7689 * There are new specifiers that are resolved in unit files,
7690 for the host name (%H), the machine ID (%m) and the boot ID
7691 (%b).
7692
b6a86739 7693 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
0c11f949
LP
7694 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
7695 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
7696 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
7697 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
7698 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
7699 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
7700
38a60d71
LP
7701CHANGES WITH 189:
7702
7703 * Support for reading structured kernel messages from
7704 /dev/kmsg has now been added and is enabled by default.
7705
7706 * Support for reading kernel messages from /proc/kmsg has now
7707 been removed. If you want kernel messages in the journal
7708 make sure to run a recent kernel (>= 3.5) that supports
7709 reading structured messages from /dev/kmsg (see
7710 above). /proc/kmsg is now exclusive property of classic
7711 syslog daemons again.
7712
7713 * The libudev API gained the new
7714 udev_device_new_from_device_id() call.
7715
7716 * The logic for file system namespace (ReadOnlyDirectory=,
7717 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
7718 require pivot_root() anymore. This means fewer temporary
7719 directories are created below /tmp for this feature.
7720
7721 * nspawn containers will now see and receive all submounts
7722 made on the host OS below the root file system of the
7723 container.
7724
7725 * Forward Secure Sealing is now supported for Journal files,
7726 which provide cryptographical sealing of journal files so
7727 that attackers cannot alter log history anymore without this
7728 being detectable. Lennart will soon post a blog story about
7729 this explaining it in more detail.
7730
7731 * There are two new service settings RestartPreventExitStatus=
7732 and SuccessExitStatus= which allow configuration of exit
7733 status (exit code or signal) which will be excepted from the
7734 restart logic, resp. consider successful.
7735
7736 * journalctl gained the new --verify switch that can be used
7737 to check the integrity of the structure of journal files and
7738 (if Forward Secure Sealing is enabled) the contents of
7739 journal files.
7740
7741 * nspawn containers will now be run with /dev/stdin, /dev/fd/
7742 and similar symlinks pre-created. This makes running shells
7743 as container init process a lot more fun.
7744
7745 * The fstab support can now handle PARTUUID= and PARTLABEL=
7746 entries.
7747
7748 * A new ConditionHost= condition has been added to match
7749 against the hostname (with globs) and machine ID. This is
7750 useful for clusters where a single OS image is used to
7751 provision a large number of hosts which shall run slightly
7752 different sets of services.
7753
7754 * Services which hit the restart limit will now be placed in a
7755 failure state.
7756
b6a86739 7757 Contributions from: Bertram Poettering, Dave Reisner, Huang
38a60d71
LP
7758 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
7759 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
7760
c269cec3
LP
7761CHANGES WITH 188:
7762
7763 * When running in --user mode systemd will now become a
7764 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
7765 tree a lot more organized.
7766
7767 * A new PartOf= unit dependency type has been introduced that
7768 may be used to group services in a natural way.
7769
7770 * "systemctl enable" may now be used to enable instances of
7771 services.
7772
7773 * journalctl now prints error log levels in red, and
7774 warning/notice log levels in bright white. It also supports
7775 filtering by log level now.
7776
7777 * cgtop gained a new -n switch (similar to top), to configure
7778 the maximum number of iterations to run for. It also gained
7779 -b, to run in batch mode (accepting no input).
7780
ab06eef8 7781 * The suffix ".service" may now be omitted on most systemctl
c269cec3
LP
7782 command lines involving service unit names.
7783
7784 * There's a new bus call in logind to lock all sessions, as
7785 well as a loginctl verb for it "lock-sessions".
7786
7787 * libsystemd-logind.so gained a new call sd_journal_perror()
7788 that works similar to libc perror() but logs to the journal
7789 and encodes structured information about the error number.
7790
7791 * /etc/crypttab entries now understand the new keyfile-size=
7792 option.
7793
7794 * shutdown(8) now can send a (configurable) wall message when
7795 a shutdown is cancelled.
7796
7797 * The mount propagation mode for the root file system will now
7798 default to "shared", which is useful to make containers work
7799 nicely out-of-the-box so that they receive new mounts from
7800 the host. This can be undone locally by running "mount
7801 --make-rprivate /" if needed.
7802
7803 * The prefdm.service file has been removed. Distributions
7804 should maintain this unit downstream if they intend to keep
7805 it around. However, we recommend writing normal unit files
7806 for display managers instead.
7807
7808 * Since systemd is a crucial part of the OS we will now
7809 default to a number of compiler switches that improve
7810 security (hardening) such as read-only relocations, stack
7811 protection, and suchlike.
7812
7813 * The TimeoutSec= setting for services is now split into
7814 TimeoutStartSec= and TimeoutStopSec= to allow configuration
7815 of individual time outs for the start and the stop phase of
7816 the service.
7817
7818 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
7819 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
7820 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
7821 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
7822 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
7823 Gundersen, Zbigniew Jędrzejewski-Szmek
7824
c4f1b862
LP
7825CHANGES WITH 187:
7826
7827 * The journal and id128 C APIs are now fully documented as man
7828 pages.
7829
7830 * Extra safety checks have been added when transitioning from
7831 the initial RAM disk to the main system to avoid accidental
7832 data loss.
7833
c269cec3 7834 * /etc/crypttab entries now understand the new keyfile-offset=
c4f1b862
LP
7835 option.
7836
7837 * systemctl -t can now be used to filter by unit load state.
7838
7839 * The journal C API gained the new sd_journal_wait() call to
7840 make writing synchronous journal clients easier.
7841
7842 * journalctl gained the new -D switch to show journals from a
7843 specific directory.
7844
7845 * journalctl now displays a special marker between log
7846 messages of two different boots.
7847
7848 * The journal is now explicitly flushed to /var via a service
7849 systemd-journal-flush.service, rather than implicitly simply
7850 by seeing /var/log/journal to be writable.
7851
7852 * journalctl (and the journal C APIs) can now match for much
7853 more complex expressions, with alternatives and
7854 disjunctions.
7855
7856 * When transitioning from the initial RAM disk to the main
7857 system we will now kill all processes in a killing spree to
7858 ensure no processes stay around by accident.
7859
7860 * Three new specifiers may be used in unit files: %u, %h, %s
7861 resolve to the user name, user home directory resp. user
7862 shell. This is useful for running systemd user instances.
7863
7864 * We now automatically rotate journal files if their data
7865 object hash table gets a fill level > 75%. We also size the
7866 hash table based on the configured maximum file size. This
7867 together should lower hash collisions drastically and thus
7868 speed things up a bit.
7869
7870 * journalctl gained the new "--header" switch to introspect
7871 header data of journal files.
7872
7873 * A new setting SystemCallFilters= has been added to services
7874 which may be used to apply blacklists or whitelists to
7875 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
7876
7877 * nspawn gained a new --link-journal= switch (and quicker: -j)
7878 to link the container journal with the host. This makes it
7879 very easy to centralize log viewing on the host for all
7880 guests while still keeping the journal files separated.
7881
7882 * Many bugfixes and optimizations
7883
7884 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
7885 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
7886 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
7887 Jędrzejewski-Szmek
7888
b5b4c94a
LP
7889CHANGES WITH 186:
7890
7891 * Several tools now understand kernel command line arguments,
7892 which are only read when run in an initial RAM disk. They
7893 usually follow closely their normal counterparts, but are
7894 prefixed with rd.
7895
7896 * There's a new tool to analyze the readahead files that are
7897 automatically generated at boot. Use:
7898
7899 /usr/lib/systemd/systemd-readahead analyze /.readahead
7900
7901 * We now provide an early debug shell on tty9 if this enabled. Use:
7902
d1f9edaf 7903 systemctl enable debug-shell.service
b5b4c94a
LP
7904
7905 * All plymouth related units have been moved into the Plymouth
7906 package. Please make sure to upgrade your Plymouth version
7907 as well.
7908
7909 * systemd-tmpfiles now supports getting passed the basename of
7910 a configuration file only, in which case it will look for it
7911 in all appropriate directories automatically.
7912
7913 * udevadm info now takes a /dev or /sys path as argument, and
7914 does the right thing. Example:
7915
7916 udevadm info /dev/sda
7917 udevadm info /sys/class/block/sda
7918
7919 * systemctl now prints a warning if a unit is stopped but a
7920 unit that might trigger it continues to run. Example: a
7921 service is stopped but the socket that activates it is left
7922 running.
7923
7924 * "systemctl status" will now mention if the log output was
7925 shortened due to rotation since a service has been started.
7926
7927 * The journal API now exposes functions to determine the
7928 "cutoff" times due to rotation.
7929
7930 * journald now understands SIGUSR1 and SIGUSR2 for triggering
7931 immediately flushing of runtime logs to /var if possible,
7932 resp. for triggering immediate rotation of the journal
7933 files.
7934
7935 * It is now considered an error if a service is attempted to
7936 be stopped that is not loaded.
7937
7938 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
7939
7940 * systemd-analyze now supports Python 3
7941
7942 * tmpfiles now supports cleaning up directories via aging
7943 where the first level dirs are always kept around but
7944 directories beneath it automatically aged. This is enabled
7945 by prefixing the age field with '~'.
7946
7947 * Seat objects now expose CanGraphical, CanTTY properties
7948 which is required to deal with very fast bootups where the
7949 display manager might be running before the graphics drivers
7950 completed initialization.
7951
7952 * Seat objects now expose a State property.
7953
7954 * We now include RPM macros for service enabling/disabling
7955 based on the preset logic. We recommend RPM based
7956 distributions to make use of these macros if possible. This
7957 makes it simpler to reuse RPM spec files across
7958 distributions.
7959
7960 * We now make sure that the collected systemd unit name is
7961 always valid when services log to the journal via
7962 STDOUT/STDERR.
7963
7964 * There's a new man page kernel-command-line(7) detailing all
7965 command line options we understand.
7966
7967 * The fstab generator may now be disabled at boot by passing
7968 fstab=0 on the kernel command line.
7969
91ac7425 7970 * A new kernel command line option modules-load= is now understood
b5b4c94a
LP
7971 to load a specific kernel module statically, early at boot.
7972
7973 * Unit names specified on the systemctl command line are now
7974 automatically escaped as needed. Also, if file system or
7975 device paths are specified they are automatically turned
7976 into the appropriate mount or device unit names. Example:
7977
7978 systemctl status /home
7979 systemctl status /dev/sda
7980
7981 * The SysVConsole= configuration option has been removed from
7982 system.conf parsing.
7983
7984 * The SysV search path is no longer exported on the D-Bus
7985 Manager object.
7986
ce830873 7987 * The Names= option has been removed from unit file parsing.
b5b4c94a
LP
7988
7989 * There's a new man page bootup(7) detailing the boot process.
7990
7991 * Every unit and every generator we ship with systemd now
7992 comes with full documentation. The self-explanatory boot is
7993 complete.
7994
7995 * A couple of services gained "systemd-" prefixes in their
7996 name if they wrap systemd code, rather than only external
7997 code. Among them fsck@.service which is now
7998 systemd-fsck@.service.
7999
8000 * The HaveWatchdog property has been removed from the D-Bus
8001 Manager object.
8002
8003 * systemd.confirm_spawn= on the kernel command line should now
8004 work sensibly.
8005
8006 * There's a new man page crypttab(5) which details all options
8007 we actually understand.
8008
8009 * systemd-nspawn gained a new --capability= switch to pass
8010 additional capabilities to the container.
8011
8012 * timedated will now read known NTP implementation unit names
5b00c016 8013 from /usr/lib/systemd/ntp-units.d/*.list,
b5b4c94a
LP
8014 systemd-timedated-ntp.target has been removed.
8015
8016 * journalctl gained a new switch "-b" that lists log data of
8017 the current boot only.
8018
8019 * The notify socket is in the abstract namespace again, in
8020 order to support daemons which chroot() at start-up.
8021
8022 * There is a new Storage= configuration option for journald
8023 which allows configuration of where log data should go. This
8024 also provides a way to disable journal logging entirely, so
8025 that data collected is only forwarded to the console, the
8026 kernel log buffer or another syslog implementation.
8027
c4f1b862 8028 * Many bugfixes and optimizations
b5b4c94a 8029
2d938ac7
LP
8030 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
8031 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
8032 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
8033 Shawn Landden, Tom Gundersen
b5b4c94a 8034
2d197285 8035CHANGES WITH 185:
b6a86739 8036
2d197285
KS
8037 * "systemctl help <unit>" now shows the man page if one is
8038 available.
8039
8040 * Several new man pages have been added.
8041
b5b4c94a
LP
8042 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
8043 MaxLevelConsole= can now be specified in
8044 journald.conf. These options allow reducing the amount of
8045 data stored on disk or forwarded by the log level.
2d197285 8046
b5b4c94a
LP
8047 * TimerSlackNSec= can now be specified in system.conf for
8048 PID1. This allows system-wide power savings.
2d197285
KS
8049
8050 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
8051 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
8052 Matthias Clasen
8053
4c8cd173 8054CHANGES WITH 184:
b6a86739 8055
4c8cd173
LP
8056 * logind is now capable of (optionally) handling power and
8057 sleep keys as well as the lid switch.
8058
8059 * journalctl now understands the syntax "journalctl
8060 /usr/bin/avahi-daemon" to get all log output of a specific
8061 daemon.
8062
8063 * CapabilityBoundingSet= in system.conf now also influences
8064 the capability bound set of usermode helpers of the kernel.
8065
8066 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
8067 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
8068 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
8069 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
8070
ea5943d3 8071CHANGES WITH 183:
b6a86739 8072
187076d4
LP
8073 * Note that we skipped 139 releases here in order to set the
8074 new version to something that is greater than both udev's
8075 and systemd's most recent version number.
8076
194bbe33
KS
8077 * udev: all udev sources are merged into the systemd source tree now.
8078 All future udev development will happen in the systemd tree. It
8079 is still fully supported to use the udev daemon and tools without
8080 systemd running, like in initramfs or other init systems. Building
8081 udev though, will require the *build* of the systemd tree, but
ea5943d3 8082 udev can be properly *run* without systemd.
07cd4fc1 8083
91cf7e5c 8084 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
f13b388f
KS
8085 should be used to create dead device nodes as workarounds for broken
8086 subsystems.
64661ee7 8087
2d13da88
KS
8088 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
8089 no longer supported. udev_monitor_new_from_netlink() needs to be
8090 used to subscribe to events.
8091
194bbe33
KS
8092 * udev: when udevd is started by systemd, processes which are left
8093 behind by forking them off of udev rules, are unconditionally cleaned
8094 up and killed now after the event handling has finished. Services or
8095 daemons must be started as systemd services. Services can be
ea5943d3 8096 pulled-in by udev to get started, but they can no longer be directly
194bbe33
KS
8097 forked by udev rules.
8098
f13b388f
KS
8099 * udev: the daemon binary is called systemd-udevd now and installed
8100 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
8101 to adapt to that, create symlink, or rename the binary after building
8102 it.
8103
ea5943d3 8104 * libudev no longer provides these symbols:
c1959569
KS
8105 udev_monitor_from_socket()
8106 udev_queue_get_failed_list_entry()
8107 udev_get_{dev,sys,run}_path()
ea5943d3 8108 The versions number was bumped and symbol versioning introduced.
c1959569 8109
ea5943d3 8110 * systemd-loginctl and systemd-journalctl have been renamed
9ae9afce 8111 to loginctl and journalctl to match systemctl.
18b754d3
KS
8112
8113 * The config files: /etc/systemd/systemd-logind.conf and
8114 /etc/systemd/systemd-journald.conf have been renamed to
8115 logind.conf and journald.conf. Package updates should rename
8116 the files to the new names on upgrade.
8117
ea5943d3
LP
8118 * For almost all files the license is now LGPL2.1+, changed
8119 from the previous GPL2.0+. Exceptions are some minor stuff
8120 of udev (which will be changed to LGPL2.1 eventually, too),
8121 and the MIT licensed sd-daemon.[ch] library that is suitable
8122 to be used as drop-in files.
8123
8124 * systemd and logind now handle system sleep states, in
49f43d5f 8125 particular suspending and hibernating.
ea5943d3
LP
8126
8127 * logind now implements a sleep/shutdown/idle inhibiting logic
8128 suitable for a variety of uses. Soonishly Lennart will blog
8129 about this in more detail.
8130
8131 * var-run.mount and var-lock.mount are no longer provided
ce830873 8132 (which previously bind mounted these directories to their new
ea5943d3
LP
8133 places). Distributions which have not converted these
8134 directories to symlinks should consider stealing these files
8135 from git history and add them downstream.
8136
8137 * We introduced the Documentation= field for units and added
8138 this to all our shipped units. This is useful to make it
3943231c 8139 easier to explore the boot and the purpose of the various
ea5943d3
LP
8140 units.
8141
8142 * All smaller setup units (such as
8143 systemd-vconsole-setup.service) now detect properly if they
8144 are run in a container and are skipped when
8145 appropriate. This guarantees an entirely noise-free boot in
8146 Linux container environments such as systemd-nspawn.
8147
8148 * A framework for implementing offline system updates is now
8149 integrated, for details see:
c6749ba5 8150 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
ea5943d3
LP
8151
8152 * A new service type Type=idle is available now which helps us
8153 avoiding ugly interleaving of getty output and boot status
8154 messages.
8155
439d6dfd
LP
8156 * There's now a system-wide CapabilityBoundingSet= option to
8157 globally reduce the set of capabilities for the
ea5943d3
LP
8158 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
8159 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
8160 even CAP_NET_ADMIN system-wide for secure systems.
8161
8162 * There are now system-wide DefaultLimitXXX= options to
8163 globally change the defaults of the various resource limits
8164 for all units started by PID 1.
8165
8166 * Harald Hoyer's systemd test suite has been integrated into
8167 systemd which allows easy testing of systemd builds in qemu
8168 and nspawn. (This is really awesome! Ask us for details!)
8169
3943231c
LP
8170 * The fstab parser is now implemented as generator, not inside
8171 of PID 1 anymore.
ea5943d3
LP
8172
8173 * systemctl will now warn you if .mount units generated from
8174 /etc/fstab are out of date due to changes in fstab that
d28315e4 8175 have not been read by systemd yet.
ea5943d3
LP
8176
8177 * systemd is now suitable for usage in initrds. Dracut has
8178 already been updated to make use of this. With this in place
8179 initrds get a slight bit faster but primarily are much
8180 easier to introspect and debug since "systemctl status" in
8181 the host system can be used to introspect initrd services,
8182 and the journal from the initrd is kept around too.
8183
8184 * systemd-delta has been added, a tool to explore differences
8185 between user/admin configuration and vendor defaults.
8186
8187 * PrivateTmp= now affects both /tmp and /var/tmp.
8188
8189 * Boot time status messages are now much prettier and feature
8190 proper english language. Booting up systemd has never been
8191 so sexy.
8192
8193 * Read-ahead pack files now include the inode number of all
8194 files to pre-cache. When the inode changes the pre-caching
8195 is not attempted. This should be nicer to deal with updated
8196 packages which might result in changes of read-ahead
8197 patterns.
8198
8199 * We now temporaritly lower the kernel's read_ahead_kb variable
8200 when collecting read-ahead data to ensure the kernel's
8201 built-in read-ahead does not add noise to our measurements
8202 of necessary blocks to pre-cache.
8203
8204 * There's now RequiresMountsFor= to add automatic dependencies
8205 for all mounts necessary for a specific file system path.
8206
8207 * MountAuto= and SwapAuto= have been removed from
8208 system.conf. Mounting file systems at boot has to take place
8209 in systemd now.
8210
8211 * nspawn now learned a new switch --uuid= to set the machine
8212 ID on the command line.
8213
f8c0a2cb 8214 * nspawn now learned the -b switch to automatically search
ea5943d3
LP
8215 for an init system.
8216
8217 * vt102 is now the default TERM for serial TTYs, upgraded from
8218 vt100.
8219
8220 * systemd-logind now works on VT-less systems.
8221
8222 * The build tree has been reorganized. The individual
3943231c 8223 components now have directories of their own.
ea5943d3
LP
8224
8225 * A new condition type ConditionPathIsReadWrite= is now available.
8226
8227 * nspawn learned the new -C switch to create cgroups for the
8228 container in other hierarchies.
8229
8230 * We now have support for hardware watchdogs, configurable in
8231 system.conf.
8232
8233 * The scheduled shutdown logic now has a public API.
8234
8235 * We now mount /tmp as tmpfs by default, but this can be
8236 masked and /etc/fstab can override it.
8237
d28315e4 8238 * Since udisks does not make use of /media anymore we are not
ea5943d3
LP
8239 mounting a tmpfs on it anymore.
8240
8241 * journalctl gained a new --local switch to only interleave
8242 locally generated journal files.
8243
8244 * We can now load the IMA policy at boot automatically.
8245
8246 * The GTK tools have been split off into a systemd-ui.
8247
79849bf9
LP
8248 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
8249 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
8250 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
8251 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
8252 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
8253 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
8254 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
8255 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
8256 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
8257 Gundersen
8258
16f1239e 8259CHANGES WITH 44:
b6a86739 8260
16f1239e
LP
8261 * This is mostly a bugfix release
8262
8263 * Support optional initialization of the machine ID from the
8264 KVM or container configured UUID.
8265
8266 * Support immediate reboots with "systemctl reboot -ff"
8267
8268 * Show /etc/os-release data in systemd-analyze output
8269
ab06eef8 8270 * Many bugfixes for the journal, including endianness fixes and
16f1239e
LP
8271 ensuring that disk space enforcement works
8272
ce830873 8273 * sd-login.h is C++ compatible again
16f1239e
LP
8274
8275 * Extend the /etc/os-release format on request of the Debian
8276 folks
8277
8278 * We now refuse non-UTF8 strings used in various configuration
d28315e4 8279 and unit files. This is done to ensure we do not pass invalid
16f1239e
LP
8280 data over D-Bus or expose it elsewhere.
8281
8282 * Register Mimo USB Screens as suitable for automatic seat
8283 configuration
8284
8285 * Read SELinux client context from journal clients in a race
8286 free fashion
8287
8288 * Reorder configuration file lookup order. /etc now always
8289 overrides /run in order to allow the administrator to always
b938cb90 8290 and unconditionally override vendor-supplied or
16f1239e
LP
8291 automatically generated data.
8292
8293 * The various user visible bits of the journal now have man
8294 pages. We still lack man pages for the journal API calls
8295 however.
8296
8297 * We now ship all man pages in HTML format again in the
8298 tarball.
8299
8300 Contributions from: Dave Reisner, Dirk Eibach, Frederic
8301 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
8302 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
8303 Reding
8304
437b7dee 8305CHANGES WITH 43:
b6a86739 8306
437b7dee
LP
8307 * This is mostly a bugfix release
8308
8309 * systems lacking /etc/os-release are no longer supported.
8310
8311 * Various functionality updates to libsystemd-login.so
8312
45afd519 8313 * Track class of PAM logins to distinguish greeters from
437b7dee
LP
8314 normal user logins.
8315
8316 Contributions from: Kay Sievers, Lennart Poettering, Michael
8317 Biebl
8318
204fa33c 8319CHANGES WITH 42:
b6a86739 8320
204fa33c
LP
8321 * This is an important bugfix release for v41.
8322
8323 * Building man pages is now optional which should be useful
8324 for those building systemd from git but unwilling to install
8325 xsltproc.
8326
8327 * Watchdog support for supervising services is now usable. In
8328 a future release support for hardware watchdogs
8329 (i.e. /dev/watchdog) will be added building on this.
8330
8331 * Service start rate limiting is now configurable and can be
8332 turned off per service. When a start rate limit is hit a
8333 reboot can automatically be triggered.
8334
8335 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
8336
8337 Contributions from: Benjamin Franzke, Bill Nottingham,
8338 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
8339 Schmidt, Michał Górny, Piotr Drąg
8340
e0d25329 8341CHANGES WITH 41:
b6a86739 8342
e0d25329
KS
8343 * The systemd binary is installed /usr/lib/systemd/systemd now;
8344 An existing /sbin/init symlink needs to be adapted with the
8345 package update.
8346
b13df964
LP
8347 * The code that loads kernel modules has been ported to invoke
8348 libkmod directly, instead of modprobe. This means we do not
8349 support systems with module-init-tools anymore.
8350
8351 * Watchdog support is now already useful, but still not
8352 complete.
8353
8354 * A new kernel command line option systemd.setenv= is
8355 understood to set system wide environment variables
8356 dynamically at boot.
8357
e9c1ea9d 8358 * We now limit the set of capabilities of systemd-journald.
ccd07a08 8359
353e12c2
LP
8360 * We now set SIGPIPE to ignore by default, since it only is
8361 useful in shell pipelines, and has little use in general
8362 code. This can be disabled with IgnoreSIPIPE=no in unit
8363 files.
8364
b13df964
LP
8365 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
8366 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
8367 William Douglas
8368
d26e4270 8369CHANGES WITH 40:
b6a86739 8370
d26e4270
LP
8371 * This is mostly a bugfix release
8372
8373 * We now expose the reason why a service failed in the
8374 "Result" D-Bus property.
8375
8376 * Rudimentary service watchdog support (will be completed over
8377 the next few releases.)
8378
8379 * When systemd forks off in order execute some service we will
8380 now immediately changes its argv[0] to reflect which process
8381 it will execute. This is useful to minimize the time window
8382 with a generic argv[0], which makes bootcharts more useful
8383
b13df964
LP
8384 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
8385 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
8386 Mike Kazantsev, Ray Strode
8387
220a21d3 8388CHANGES WITH 39:
b6a86739 8389
220a21d3
LP
8390 * This is mostly a test release, but incorporates many
8391 bugfixes.
8392
8393 * New systemd-cgtop tool to show control groups by their
8394 resource usage.
8395
8396 * Linking against libacl for ACLs is optional again. If
8397 disabled, support tracking device access for active logins
8398 goes becomes unavailable, and so does access to the user
8399 journals by the respective users.
8400
8401 * If a group "adm" exists, journal files are automatically
8402 owned by them, thus allow members of this group full access
8403 to the system journal as well as all user journals.
8404
8405 * The journal now stores the SELinux context of the logging
8406 client for all entries.
8407
8408 * Add C++ inclusion guards to all public headers
8409
8410 * New output mode "cat" in the journal to print only text
8411 messages, without any meta data like date or time.
8412
8413 * Include tiny X server wrapper as a temporary stop-gap to
8414 teach XOrg udev display enumeration. This is used by display
8415 managers such as gdm, and will go away as soon as XOrg
8416 learned native udev hotplugging for display devices.
8417
8418 * Add new systemd-cat tool for executing arbitrary programs
8419 with STDERR/STDOUT connected to the journal. Can also act as
8420 BSD logger replacement, and does so by default.
8421
8422 * Optionally store all locally generated coredumps in the
8423 journal along with meta data.
8424
8425 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
8426 writing short strings to files (for usage for /sys), and for
8427 creating symlinks, character and block device nodes.
8428
8429 * New unit file option ControlGroupPersistent= to make cgroups
8430 persistent, following the mechanisms outlined in
56cadcb6 8431 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
220a21d3
LP
8432
8433 * Support multiple local RTCs in a sane way
8434
8435 * No longer monopolize IO when replaying readahead data on
8436 rotating disks, since we might starve non-file-system IO to
8437 death, since fanotify() will not see accesses done by blkid,
8438 or fsck.
8439
d28315e4 8440 * Do not show kernel threads in systemd-cgls anymore, unless
220a21d3
LP
8441 requested with new -k switch.
8442
8443 Contributions from: Dan Horák, Kay Sievers, Lennart
8444 Poettering, Michal Schmidt
8445
8446CHANGES WITH 38:
b6a86739 8447
220a21d3
LP
8448 * This is mostly a test release, but incorporates many
8449 bugfixes.
8450
8451 * The git repository moved to:
8452 git://anongit.freedesktop.org/systemd/systemd
8453 ssh://git.freedesktop.org/git/systemd/systemd
8454
8455 * First release with the journal
8456 http://0pointer.de/blog/projects/the-journal.html
8457
8458 * The journal replaces both systemd-kmsg-syslogd and
8459 systemd-stdout-bridge.
8460
8461 * New sd_pid_get_unit() API call in libsystemd-logind
8462
8463 * Many systemadm clean-ups
8464
8465 * Introduce remote-fs-pre.target which is ordered before all
8466 remote mounts and may be used to start services before all
8467 remote mounts.
8468
8469 * Added Mageia support
8470
8471 * Add bash completion for systemd-loginctl
8472
8473 * Actively monitor PID file creation for daemons which exit in
8474 the parent process before having finished writing the PID
8475 file in the daemon process. Daemons which do this need to be
8476 fixed (i.e. PID file creation must have finished before the
8477 parent exits), but we now react a bit more gracefully to them.
8478
8479 * Add colourful boot output, mimicking the well-known output
8480 of existing distributions.
8481
8482 * New option PassCredentials= for socket units, for
8483 compatibility with a recent kernel ABI breakage.
8484
8485 * /etc/rc.local is now hooked in via a generator binary, and
8486 thus will no longer act as synchronization point during
8487 boot.
8488
8489 * systemctl list-unit-files now supports --root=.
8490
8491 * systemd-tmpfiles now understands two new commands: z, Z for
8492 relabelling files according to the SELinux database. This is
8493 useful to apply SELinux labels to specific files in /sys,
8494 among other things.
8495
8496 * Output of SysV services is now forwarded to both the console
8497 and the journal by default, not only just the console.
8498
8499 * New man pages for all APIs from libsystemd-login.
8500
ce830873 8501 * The build tree got reorganized and the build system is a
220a21d3
LP
8502 lot more modular allowing embedded setups to specifically
8503 select the components of systemd they are interested in.
8504
8505 * Support for Linux systems lacking the kernel VT subsystem is
8506 restored.
8507
8508 * configure's --with-rootdir= got renamed to
8509 --with-rootprefix= to follow the naming used by udev and
8510 kmod
8511
d28315e4 8512 * Unless specified otherwise we will now install to /usr instead
220a21d3
LP
8513 of /usr/local by default.
8514
8515 * Processes with '@' in argv[0][0] are now excluded from the
8516 final shut-down killing spree, following the logic explained
8517 in:
56cadcb6 8518 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
220a21d3
LP
8519
8520 * All processes remaining in a service cgroup when we enter
8521 the START or START_PRE states are now killed with
8522 SIGKILL. That means it is no longer possible to spawn
8523 background processes from ExecStart= lines (which was never
8524 supported anyway, and bad style).
8525
8526 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
8527 reloading of units together.
8528
4c8cd173 8529 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
220a21d3
LP
8530 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
8531 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
8532 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
8533 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek